Nonlinear hierarchical multiscale modeling of cortical bone considering its nanoscale microstructure

We have used a hierarchical multiscale modeling scheme for the analysis of cortical bone considering it as a nanocomposite. This scheme consists of definition of two boundary value problems, one for macroscale, and another for microscale. The coupling between these scales is done by using the homogenization technique. At every material point in which the constitutive model is needed, a microscale boundary value problem is defined using a macroscopic kinematical quantity and solved. Using the described scheme, we have studied elastic properties of cortical bone considering its nanoscale microstructural constituents with various mineral volume fractions. Since the microstructure of bone consists of mineral platelet with nanometer size embedded in a protein matrix, it is similar to the microstructure of soft matrix nanocomposites reinforced with hard nanostructures. Considering a representative volume element (RVE) of the microstructure of bone as the microscale problem in our hierarchical multiscale modeling scheme, the global behavior of bone is obtained under various macroscopic loading conditions. This scheme may be suitable for modeling arbitrary bone geometries subjected to a variety of loading conditions. Using the presented method, mechanical properties of cortical bone including elastic moduli and Poisson's ratios in two major directions and shear modulus is obtained for different mineral volume fractions.     

Sensitivity analysis and parametric study of elastic properties of an unidirectional mineralized bone fibril-array using mean field methods

The key parameters determining the elastic properties of an unidirectional mineralized bone fibril-array decomposed in two further hierarchical levels are investigated using mean field methods. Modeling of the elastic properties of mineralized micro- and nanostructures requires accurate information about the underlying topology and the constituents’ material properties. These input data are still afflicted by great uncertainties and their influence on computed elastic constants of a bone fibril-array remains unclear. In this work, mean field methods are applied to model mineralized fibrils, the extra-fibrillar matrix and the resulting fibril-array. The isotropic or transverse isotropic elastic constants of these constituents are computed as a function of degree of mineralization, mineral distribution between fibrils and extra-fibrillar matrix, collagen stiffness and fibril volume fraction. The linear sensitivity of the elastic constants was assessed at a default set of the above parameters. The strain ratios between the constituents as well as the axial and transverse indentation moduli of the fibril-array were calculated for comparison with experiments. Results indicate that the degree of mineralization and the collagen stiffness dominate fibril-array elasticity. Interestingly, the stiffness of the extra-fibrillar matrix has a strong influence on transverse and shear moduli of the fibril-array. The axial strain of the intra-fibrillar mineral platelets is 30–90% of the applied fibril strain, depending on mineralization and collagen stiffness. The fibril-to-fibril-array strain ratio is essentially ~1. This study provides an improved insight in the parameters, which govern the fibril-array stiffness of mineralized tissues such as bone.     

Elastic properties of woven bone: effect of mineral content and collagen fibrils orientation

Woven bone is a type of tissue that forms mainly during fracture healing or fetal bone development. Its microstructure can be modeled as a composite with a matrix of mineral (hydroxyapatite) and inclusions of collagen fibrils with a more or less random orientation. In the present study, its elastic properties were estimated as a function of composition (degree of mineralization) and fibril orientation. A self-consistent homogenization scheme considering randomness of inclusions’ orientation was used for this purpose. Lacuno-canalicular porosity in the form of periodically distributed void inclusions was also considered. Assuming collagen fibrils to be uniformly oriented in all directions led to an isotropic tissue with a Young’s modulus \(E = 1.90\) GPa, which is of the same order of magnitude as that of woven bone in fracture calluses. By contrast, assuming fibrils to have a preferential orientation resulted in a Young’s modulus in the preferential direction of 9–16 GPa depending on the mineral content of the tissue. These results are consistent with experimental evidence for woven bone in foetuses, where collagen fibrils are aligned to a certain extent.     

Determination of the heterogeneous anisotropic elastic properties of human femoral bone: From nanoscopic to organ scale

Cortical bone is a multiscale composite material. Its elastic properties are anisotropic and heterogeneous across its cross-section, due to endosteal bone resorption which might affect bone strength. The aim of this paper was to describe a homogenization method leading to the estimation of the variation of the elastic coefficients across the bone cross-section and along the bone longitudinal axis. The method uses the spatial variations of bone porosity and of the degree of mineralization of the bone matrix (DMB) obtained from the analysis of 3-D synchrotron micro-computed tomography images. For all three scales considered (the foam (100 nm), the ultrastructure (5 μm) and the mesoscale (500 μm)), the elastic coefficients were determined using the Eshelby’s inclusion problem. DMB values were used at the scale of the foam. Collagen was introduced at the scale of the ultrastructure and bone porosity was introduced at the mesoscale. The pores were considered as parallel cylinders oriented along the bone axis. Each elastic coefficient was computed for different regions of interest, allowing an estimation of its variations across the bone cross-section and along the bone longitudinal axis. The method was applied to a human femoral neck bone specimen, which is a site of osteoporotic fracture. The computed elastic coefficients for cortical bone were in good agreement with experimental results, but some discrepancies were obtained in the endosteal part (trabecular bone). These results highlight the importance of accounting for the heterogeneity of cortical bone properties across bone cross-section and along bone longitudinal axis.     

Biomimetic mineralization of woven bone-like nanocomposites: role of collagen cross-links

Ideal biomaterials for bone grafts must be biocompatible, osteoconductive, osteoinductive and have appropriate mechanical properties. For this, the development of synthetic bone substitutes mimicking natural bone is desirable, but this requires controllable mineralization of the collagen matrix. In this study, densified collagen films (up to 100 μm thick) were fabricated by a plastic compression technique and cross-linked using carbodiimide. Then, collagen-hydroxyapatite composites were prepared by using a polymer-induced liquid-precursor (PILP) mineralization process. Compared to traditional methods that produce only extrafibrillar hydroxyapatite (HA) clusters on the surface of collagen scaffolds, by using the PILP mineralization process, homogeneous intra- and extrafibrillar minerals were achieved on densified collagen films, leading to a similar nanostructure as bone, and a woven microstructure analogous to woven bone. The role of collagen cross-links on mineralization was examined and it was found that the cross-linked collagen films stimulated the mineralization reaction, which in turn enhanced the mechanical properties (hardness and modulus). The highest value of hardness and elastic modulus was 0.7 ± 0.1 and 9.1 ± 1.4 GPa in the dry state, respectively, which is comparable to that of woven bone. In the wet state, the values were much lower (177 ± 31 and 8 ± 3 MPa) due to inherent microporosity in the films, but still comparable to those of woven bone in the same conditions. Mineralization of collagen films with controllable mineral content and good mechanical properties provide a biomimetic route toward the development of bone substitutes for the next generation of biomaterials. This work also provides insight into understanding the role of collagen fibrils on mineralization.     

Changes in the organic matrix composition of the solid tissues in chick embryogenesis

Biochemical studies have been made on the dynamics of organic matrix of the bone tissue and on its calcification during embryogenesis of chicks. Carbohydrate composition of non-collagen proteins and the degree of mineralization of the bone tissue, as well as utilization of carbohydrates and mineral substances from the egg shell were investigated at various stages of embryonic development. The role of carbohydrates in bone histogenesis and its mineralization is discussed.     

Transmission FT-IR microspectroscopy of mineral phases in calcified tissues

Fourier-transform infrared microspectroscopy (FT-IRM) was used to study bone mineralization processes in an in vivo model and in enamel in osteogenesis imperfecta. Finally, the ability of FT-IRM to map new bone formed in implanted macroporous calcium phosphate biomaterial from sections was reported for the first time. FTIRM allowed the correlation of the microstructure of bone formation in the in vivo model with modifications in carbonate and phosphate environments of the mineral phases during maturation. FT-IRM analysis on enamel sections revealed changes in the mineral environment of carbonate and phosphate ions and probably in the size of enamel crystals. These modifications contributed to the fragility of enamel in osteogenesis imperfecta. The infrared functional group imaging of a part of implanted biomaterial and the bone ingrowth provided the visualization of chemical modifications occurring in biomaterial implants at 20 μm spatial resolution. The use of FT-IRM, in conjunction with appropriate sampling methods and data analysis should provide further insight into the molecular structure of mineral phases of calcified tissues and help to elucidate mineralization processes, skeletal disorders and properties of the biomaterials used as bone substitute.     

Spatial distribution of tissue level properties in a human femoral cortical bone

The mechanical properties of cortical bone are determined by a combination bone tissue composition, and structure at several hierarchical length scales. In this study the spatial distribution of tissue level properties within a human femoral shaft has been investigated. Cylindrically shaped samples (diameter: 4.4mm, N=56) were prepared from cortical regions along the entire length (20-85% of the total femur length), and around the periphery (anterior, medial, posterior and lateral quadrants). The samples were analyzed using scanning acoustic microscopy (SAM) at 50MHz and synchrotron radiation micro computed tomography (SRμCT). For all samples the average cortical porosity (Ct.Po), tissue elastic coefficients (c(ij)) and the average tissue degree of mineralization (DMB) were determined. The smallest coefficient of variation was observed for DMB (1.8%), followed by BV/TV (5.4%), c(ij) (8.2-45.5%), and Ct.Po (47.5%). Different variations with respect to the anatomical position were found for DMB, Ct.Po and c(ij). These data address the anatomical variations in anisotropic elastic properties and link them to tissue mineralization and porosity, which are important input parameters for numerical multi-scale bone models.     

In vivo fatigue microcracks in human bone: material properties of the surrounding bone matrix

Human bones sustain fatigue damage in the form of in vivo microcracks as a result of the normal everyday loading activities. These microcracks appear to preferentially accumulate in certain regions of bone and most notably in interstitial bone matrix areas. These are remnants of old bone tissue left unremodelled, which show a higher than average mineral content and consequently the occurrence of microcracks has been attributed to the possible brittleness brought about by such hypermineralisation. There is a need, therefore, for information on the in situ bone matrix properties in the vicinity of such in vivo microcracks to elucidate the possible causes of their appearance. The present study examined the elastic, strain rate (viscous) and plastic properties of bone matrix in selectively targeted areas by nanoindentation and in both quasistatic and dynamic mode. The results showed that in vivo crack areas are not as stiff as some well-known extremely mineralised and brittle bone examples (bulla, rostrum); the strain rate effects of crack regions were identical to those of other regions of human bone and agreed well with values collected for human bone in the past at the macroscale; while the plasticity index of the crack regions was also not statistically different from most bone examples (including human at random, bovine, bulla and rostrum) except antler, which showed lower plasticity and thus a greater fraction of elastic recovery in indentation energy. It is difficult, therefore, to explain the susceptibility of these interstitial regions to crack in terms of the mineral content and its after-effects on elasticity, viscosity and plasticity alone, but one need to attribute the cracks to the cumulative loading history of these areas, or raise the suggestion that these areas of bone matrix are in some measure 'aged' or material/quality defective.     

'Universal' microstructural patterns in cortical and trabecular, extracellular and extravascular bone materials: micromechanics-based prediction of anisotropic elasticity

Bone materials are characterized by an astonishing variability and diversity. Still, because of 'architectural constraints' due to once chosen material constituents and their physical interaction, the fundamental hierarchical organization or basic building plans of bone materials remain largely unchanged during biological evolution. Such universal patterns of microstructural organization govern the mechanical interaction of the elementary components of bone (hydroxyapatite, collagen, water; with directly measurable tissue-independent elastic properties), which are here quantified through a multiscale homogenization scheme delivering effective elastic properties of bone materials: at a scale of 10nm, long cylindrical collagen molecules, attached to each other at their ends by approximately 1.5nm long crosslinks and hosting intermolecular water inbetween, form a contiguous matrix called wet collagen. At a scale of several hundred nanometers, wet collagen and mineral crystal agglomerations interpenetrate each other, forming the mineralized fibril. At a scale of 5-10microm, the extracellular solid bone matrix is represented as collagen fibril inclusions embedded in a foam of largely disordered (extrafibrillar) mineral crystals. At a scale above the ultrastructure, where lacunae are embedded in extracellular bone matrix, the extravascular bone material is observed. Model estimates predicted from tissue-specific composition data gained from a multitude of chemical and physical tests agree remarkably well with corresponding acoustic stiffness experiments across a variety of cortical and trabecular, extracellular and extravascular materials. Besides from reconciling the well-documented, seemingly opposed concepts of 'mineral-reinforced collagen matrix' and 'collagen-reinforced mineral matrix' for bone ultrastructure, this approach opens new possibilities in the exploitation of computer tomographic data for nano-to-macro mechanics of bone organs.     

Raman and mechanical properties correlate at whole bone- and tissue-levels in a genetic mouse model

The fracture resistance of bone arises from the composition, orientation, and distribution of the primary constituents at each hierarchical level of organization. Therefore, to establish the relevance of Raman spectroscopy (RS) in identifying differences between strong or tough bone and weak or brittle bone, we investigated whether Raman-derived properties could explain the variance in biomechanical properties at both the whole bone and the tissue-level, and do so independently of traditional measurements of mineralization. We harvested femurs from wild-type mice and mice lacking matrix metalloproteinase 2 because the mutant mice have a known reduction in mineralization. Next, RS quantified compositional properties directly from the intact diaphysis followed by micro-computed tomography to quantify mineralization density (Ct.TMD). Correlations were then tested for significance between these properties and the biomechanical properties as determined by the three-point bending test on the same femurs. Harvested tibia were embedded in plastic, sectioned transversely, and polished in order to acquire average Raman properties per specimen that were then correlated with average nanoindentation properties per specimen. Dividing the ν(1) phosphate by the proline peak intensity provided the strongest correlation between the mineral-to-collagen ratio and the biomechanical properties (whole bone modulus, strength, and post-yield deflection plus nanoindentation modulus). Moreover, the linear combination of ν(1) phosphate/proline and Ct.TMD provided the best explanation of the variance in strength between the genotypes, and it alone was the best explanatory variable for brittleness. Causal relationships between Raman and fracture resistance need to be investigated, but Raman has the potential to assess fracture risk.     

Infrared spectroscopic assessment of the inflammation-mediated osteoporosis (IMO) model applied to rabbit bone

A model of osteoporosis based on induced inflammation (IMO) was applied on rabbit bones. The structural heterogeneity and molecular complexity of bone significantly affect bone mechanical properties. A tool like Fourier transform infrared spectroscopy, able to analyze both the inorganic and organic phase simultaneously, could provide compositional information regarding cortical and trabecular sections under normal and osteoporotic conditions. In this study, we assessed the mineral/matrix ratio, carbonate and phosphate content and labile (i.e., non-apatitic) species contribution to bone mineral and collagen cross-linking patterns. Clear differences were observed between cortical and trabecular bone regarding mineral and carbonate content. Induced inflammation lowers the mineral/matrix ratio and increases the overall carbonate accumulation. Elevated concentrations of labile species were detected in osteoporotic samples, especially in the trabecular sections. Collagen cross-linking patterns were indirectly observed through the 1660/1690 cm^− 1 ratio in the amide I band and a positive correlation was found with the mineralization index. Principal component analysis (PCA) applied to female samples successfully clustered trabecular and osteoporotic cases. The important role played by the phosphate ions was confirmed by corresponding loadings plots. The results suggest that the application of the IMO model to rabbit bones effectively alters bone remodeling and forms an osteoporotic bone matrix with a dissimilar composition compared to the normal one.     

A primary phosphorus‐deficient skeletal phenotype in juvenile Atlantic salmon Salmo salar: the uncoupling of bone formation and mineralization

To understand the effect of low dietary phosphorus (P) intake on the vertebral column of Atlantic salmon Salmo salar, a primary P deficiency was induced in post‐smolts. The dietary P provision was reduced by 50% for a period of 10 weeks under controlled conditions. The animal's skeleton was subsequently analysed by radiology, histological examination, histochemical detection of minerals in bones and scales and chemical mineral analysis. This is the first account of how a primary P deficiency affects the skeleton in S. salar at the cellular and at the micro‐anatomical level. Animals that received the P‐deficient diet displayed known signs of P deficiency including reduced growth and soft, pliable opercula. Bone and scale mineral content decreased by c. 50%. On radiographs, vertebral bodies appear small, undersized and with enlarged intervertebral spaces. Contrary to the X‐ray‐based diagnosis, the histological examination revealed that vertebral bodies had a regular size and regular internal bone structures; intervertebral spaces were not enlarged. Bone matrix formation was continuous and uninterrupted, albeit without traces of mineralization. Likewise, scale growth continues with regular annuli formation, but new scale matrix remains without minerals. The 10 week long experiment generated a homogeneous osteomalacia of vertebral bodies without apparent induction of skeletal malformations. The experiment shows that bone formation and bone mineralization are, to a large degree, independent processes in the fish examined. Therefore, a deficit in mineralization must not be the only cause of the alterations of the vertebral bone structure observed in farmed S. salar. It is discussed how the observed uncoupling of bone formation and mineralization helps to better diagnose, understand and prevent P deficiency‐related malformations in farmed S. salar.     

Tensile behavior of cortical bone: dependence of organic matrix material properties on bone mineral content

A porous composite model is developed to analyze the tensile mechanical properties of cortical bone. The effects of microporosity (volksman's canals, osteocyte lacunae) on the mechanical properties of bone tissue are taken into account. A simple shear lag theory, wherein tensile loads are transferred between overlapped mineral platelets by shearing of the organic matrix, is used to model the reinforcement provided by mineral platelets. It is assumed that the organic matrix is elastic in tension and elastic-perfectly plastic in shear until it fails. When organic matrix shear stresses at the ends of mineral platelets reach their yield values, the stress-strain curve of bone tissue starts to deviate from linear behavior. This is referred as the microscopic yield point. At the point where the stress-strain behavior of bone shows a sharp curvature, the organic phase reaches its shear yield stress value over the entire platelet. This is referred as the macroscopic yield point. It is assumed that after macroscopic yield, mineral platelets cannot contribute to the load bearing capacity of bone and that the mechanical behavior of cortical bone tissue is determined by the organic phase only. Bone fails when the principal stress of the organic matrix is reached. By assuming that mechanical properties of the organic matrix are dependent on bone mineral content below the macroscopic yield point, the model is used to predict the entire tensile mechanical behavior of cortical bone for different mineral contents. It is found that decreased shear yield stresses and organic matrix elastic moduli are required to explain the mechanical behavior of bones with lowered mineral contents. Under these conditions, the predicted values (elastic modulus, 0.002 yield stress and strain, and ultimate stress and strain) are within 15% of experimental data.     

The role of mineral content in determining the micromechanical properties of discrete trabecular bone remodeling packets

In trabecular bone, each remodeling event results in the resorption and/or formation of discrete structural units called ‘packets’. These remodeling packets represent a fundamental level of bone’s structural hierarchy at which to investigate composition and mechanical behaviors. The objective of this study was to apply the complementary techniques of quantitative backscattered electron microscopy (qBSEM) and nanoindentation to investigate inter-relationships between packet mineralization, elastic modulus, contact hardness and plastic deformation resistance. Indentation arrays were performed across nine trabecular spicules from 3 human donors; these spicules were then imaged using qBSEM, and discretized into their composite remodeling packets (127 in total). Packets were classified spatially as peripheral or central, and mean contact hardness, plastic deformation resistance, elastic modulus and calcium content calculated for each. Inter-relationships between measured parameters were analysed using linear regression analyses, and dependence on location assessed using Student’s t-tests. Significant positive correlations were found between all mechanical parameters and calcium content. Elastic modulus and contact hardness were significantly correlated, however elastic modulus and plastic deformation resistance were not. Calcium content, contact hardness and elastic modulus were all significantly higher for central packets than for peripheral, confirming that packet mineral content contributes to micromechanical heterogeneity within individual trabecular spicules. Plastic deformation resistance, however, showed no such regional dependence, indicating that the plastic deformation properties in particular, are determined not only by mineral content, but also by the organic matrix and interactions between these two components.     

Osteocalcin cluster: Implications for functional studies

Osteocalcin is a skeletal member of the family of extracellular mineral binding Gla protein. Osteocalcin is synthesized only by the osteoblast and it is secreted into the bone matrix at the time of bone mineralization. The mineral binding properties of osteocalcin as well as its spatial and temporal pattern of expression suggest that it plays a role during bone mineralization, however until now its biological function is unclear. To understand osteocalcin function during skeletogenesis we mutated the two osteocalcin genes by homologous recombination in embryonic stem (ES) cells. Eight targeted clones were identified by Southern analysis using external probes. One of these clones contributed to the germ line of mouse chimera. Interbreeding of heterozygotes is currently in progress. Mutant mice will be useful to understand osteocalcin function in vivo.     

The many adaptations of bone

Studies concerned with the "adaptations" in bones usually deal with modelling taking place during the individual's lifetime. However, many adaptations are produced over evolutionary time. This survey samples some adaptations of bone that may occur over both length scales, and tries to show whether short- or long-term adaptation is important. (a) Woven and lamellar bone. Woven bone is less mechanically competent than lamellar bone but is frequently found in bones that grow quickly. (b) Stress concentrations in bone. Bone is full of cavities that potentially may act as stress concentrators. Usually these cavities are oriented to minimise their stress-concentrating effect. Furthermore, the "flow" of lamellae round the cavities will still further reduce their stress-concentrating effect, but the elastic anisotropy of bone will, contrarily, tend to enhance it in normal loading situations. (c) Stiffness versus toughness. The mineral content of bone is the main determinant of differences in mechanical properties. Different bones have different mineral contents that optimise the mix of stiffness and toughness needed. (d) Synergy of whole bone architecture and material properties. As bone material properties change during growth the architecture of the whole bone is modified concurrently, to produce an optimum mechanical behaviour of the whole bone. (e) Secondary remodelling. The formation of secondary osteones in general weakens bone. Various suggestions that have been put forward to account for secondary remodelling: enabling mineral homeostasis; removing dead bone; changing the grain of the bone; taking out microcracks. (f) The hollowness of bones. It is shown how the degree of hollowness is adapted to the life of the animal.     

In situ examination of the time-course for secondary mineralization of Haversian bone using synchrotron Fourier transform infrared microspectroscopy.

At the tissue level it is well established that the rate of remodeling is related to the degree of mineralization. However, it is unknown how long it takes for an individual bone structural unit (BSU) to become fully mineralized during secondary mineralization. Using synchrotron Fourier transform infrared microspectroscopy (FTIRM) we examined the time required for newly formed bone matrix to reach a physiological mineralization limit. Twenty-six, four-month old female New Zealand white rabbits were administered up to four different fluorochrome labels at specific time points to evaluate the chemical composition of labeled osteons from the tibial diaphysis that had mineralized for 1, 8, 18, 35, 70, 105, 140, 175, 210, 245, 280, 315, 350, and 385 days. Interstitial bone from 505 day old rabbits was used as a reference value for the physiological limit to which bone mineralizes. Using synchrotron FTIRM, area integrations were carried out on protein (Amide I: 1688-1623 cm(-1)), carbonate (v(2)CO(3)(2-): 905-825 cm(-1)), and phosphate (v(4)PO(4)(3-): 650-500 cm(-1)) IR bands. IR spectral data are presented as ratios of phosphate/protein (overall matrix mineralization) and carbonate/protein. The rate of mineralization of osteonal bone proceeded rapidly between day 1 and 18, reaching 67% of interstitial bone levels. This was followed by a slower, more progressive accumulation of mineral up to day 350. By 350 days the rate of increase plateaued. The ratio of carbonate/protein also increased rapidly during the first 18 days, reaching 73% of interstitial bone levels. The ratio of carbonate/protein plateaued by day 315, reaching levels not significantly different to interstitial bone levels. In conclusion, our data demonstrate that bone accumulates mineral rapidly during the first 18 days (primary mineralization), followed by a more gradual increase in the accumulation of mineral (secondary mineralization) which we found to be completed in 350 days.     

Transforming growth factor-beta inhibition of mineralization by neonatal rat osteoblasts in monolayer and collagen gel culture

Summary The latent form of transforming growth factor-beta (TGF-β) is a component of the extracellular matrix of bone. The active form, when locally injected in vivo, stimulates both inflammation and ectopic bone formation. The present study was undertaken to determine if TGF-β also stimulated mineralization by isolated rat calvarial osteoblasts cultured in collagen gels. Gels were used because they should mimic in vivo conditions better than classical monolayer culture. Compared to cells in monolayers, osteoblasts cultured in collagen gels exhibited slower growth, but higher alkaline phosphatase activity and mineral deposition. Cultured cells also synthesized the osteoblast-specific marker, osteocalcin. The increase in osteocalcin in cell layers was parallel to the increase in mineral deposition. In the presence of TGF-β, neither cell growth nor alkaline phosphatase activity increased. Instead, a small decrease occurred in both parameters when compared to untreated cultures. Accumulation of collagen, the major component of the extracellular matrix where mineralization occurs, was similar in untreated and TGF-β1-treated cultures. However, 8 pM TGF-β1 dramatically suppressed mineral deposition in both types of cultures. Despite TGF-β1 stimulating a fourfold increase in lactic acid, the consequent increase in culture medium acidity did not account for the inhibitory effects of TGF-β1 on mineralization. These results demonstrate that collagen gel culture is an improved technique over conventional monolayer culture for demonstrating differentiated osteoblast function and sensitivity to TGF-β1. TGF-β1, at a concentration that has little effect on cell growth, alkaline phosphatase activity, or collagen accumulation, is a potent inhibitor of mineralization. The mechanism by which TGF-β1 inhibits mineralization remains to be determined.