Spelsberg, Thomas C.

Abstract Because regulation of the differentiation to osteoblasts and adipocytes from a common progenitor in bone marrow stroma is poorly understood, we assessed effects of bone morphogenetic protein-2 (BMP-2) on a conditionally immortalized human marrow stromal cell line, hMS(2–6), which is capable of differentiation to either lineage. BMP-2 did not affect hMS(2–6) cell proliferation but enhanced osteoblast differentiation as assessed by a 1.8-fold increase in expression of OSF2/CBFA1 (a gene involved in commitment to the osteoblast pathway), by increased mRNA expression and protein secretion for alkaline phosphatase (ALP), type I procollagen and osteocalcin (OC) (except for OC protein), and by increased mineralized nodule formation. Transient transfection with Osf2/Cbfa1 antisense oligonucleotide substantially reduced BMP-2–stimulated expression of ALP mRNA and protein. The effects of BMP-2 on adipocyte differentiation varied: expression of peroxisome proliferator-activated receptor γ2 (a gene involved in commitment to the adipocyte pathway) was unchanged, mRNA expression of the early differentiation marker, lipoprotein lipase, was increased, and mRNA and protein levels of the late differentiation marker, leptin, and the formation of cytoplasmic lipid droplets were decreased. Thus, by enhancing osteoblast commitment and by inhibiting late adipocyte maturation, BMP-2 acts to shunt uncommitted marrow stromal precursor cells from the adipocyte to the osteoblast differentiation pathway.

Abstract Although the differentiation of mature osteoblasts has been well studied, there is still a need for a convenient way to study preosteoblast differentiation. Our laboratory has recently described a method for isolating small numbers of authentic osteoblast precursor cells from human bone marrow (Rickard et al., J Bone Miner Res 11:312–324, 1996). Here we describe the conditional immortalization of these cells by retroviral transfection with the amphotrophic vector, pZipSV40tsa58, which encodes for a temperature-sensitive mutant form of the simian virus large T-antigen. At the permissive temperature of 34°C, the cell lines proliferated, but differentiation was arrested, whereas at the restrictive temperature of 39.5°C, proliferation was decreased and differentiation was induced. As assessed by semiquantitative reverse transcriptase PCR after 4 days of culture at 39.5°C, the six cell lines expressed similar mRNA levels both constitutively and in response to dexamethasone (Dex) and 1α,25-dihydroxyvitamin D3 (1,25(OH2)D3) for osteoblast (alkaline phosphatase [ALP], type I collagen [Col I], osteocalcin [OC], and parathyroid hormone receptor [PTH-R] and adipocyte (lipoprotein lipase [LPL]) genes. In the presence of 10−8 M Dex, gene expression for ALP, PTH-R, and LPL increased, but that for OC decreased. Stimulation with 10−8 M 1,25(OH2)D3 increased gene expression for ALP, OC, and Col I. Changes in protein production for ALP, OC, and type I procollagen in response to Dex and 1,25(OH2)D3 were similar to changes in mRNA levels. When cultured at 39.5°C with ascorbate and β-glycerolphosphate for 21 days, mineralization of matrix occurred, whereas culture with Dex plus 1,25(OH2)D3, or rabbit serum led to enhanced formation of cytoplasmic lipid droplets within 6 days. Thus, these cell lines are capable of bipotential differentiation and should serve as an excellent tool to study the molecular mechanisms that regulate and select for osteoblast and adipocyte differentiation in humans.

Background— Rheumatic heart disease is the most common cause of valvular disease in developing countries. Despite the high prevalence of this disease, the cellular mechanisms are not well known. We hypothesized that rheumatic valve calcification is associated with an osteoblast bone formation and neoangiogenesis. Methods and Results— To test this hypothesis, we examined human rheumatic valves replaced at surgery (n=23), normal human valves (n=20) removed at cardiac transplantation, and degenerative mitral valve leaflets removed during surgical valve repair (n=15). Microcomputed tomography was used to assess mineralization fronts to reconstruct the extents of mineralization. Immunohistochemistry was used to localize osteopontin protein, α-actin, osteocalcin, vascular endothelial growth factor, von Willebrand factor, and CD68 (human macrophage). Microcomputed tomography demonstrated complex calcification developing within the heavily calcified rheumatic valves, not in the degenerative mitral valves and control valves. Immunohistochemistry localized osteopontin and osteocalcin to areas of smooth muscle cells within microvessels and proliferating myofibroblasts. Vascular endothelial growth factor was present in areas of inflammation and colocalized with the CD68 stain primarily in the calcified rheumatic valves. Alizarin red, osteopontin, and osteocalcin protein expression was upregulated in the calcified rheumatic valves and was present at low levels in the degenerative mitral valves. Conclusions— These findings support the concept that rheumatic valve calcification is not a random passive process but a regulated, inflammatory cellular process associated with the expression of osteoblast markers and neoangiogenesis.

Background— Calcific aortic stenosis is the third most common cardiovascular disease in the United States. We hypothesized that the mechanism for aortic valve calcification is similar to skeletal bone formation and that this process is mediated by an osteoblast-like phenotype. Methods and Results— To test this hypothesis, we examined calcified human aortic valves replaced at surgery (n=22) and normal human valves (n=20) removed at time of cardiac transplantation. Contact microradiography and micro-computerized tomography were used to assess the 2-dimensional and 3-dimensional extent of mineralization. Mineralization borders were identified with von Kossa and Goldner’s stains. Electron microscopy and energy-dispersive spectroscopy were performed for identification of bone ultrastructure and CaPO 4 composition. To analyze for the osteoblast and bone markers, reverse transcriptase–polymerase chain reaction was performed on calcified versus normal human valves for osteopontin, bone sialoprotein, osteocalcin, alkaline phosphatase, and the osteoblast-specific transcription factor Cbfa1. Microradiography and micro-computerized tomography confirmed the presence of calcification in the valve. Special stains for hydroxyapatite and CaPO 4 were positive in calcification margins. Electron microscopy identified mineralization, whereas energy-dispersive spectroscopy confirmed the presence of elemental CaPO 4 . Reverse transcriptase–polymerase chain reaction revealed increased mRNA levels of osteopontin, bone sialoprotein, osteocalcin, and Cbfa1 in the calcified valves. There was no change in alkaline phosphatase mRNA level but an increase in the protein expression in the diseased valves. Conclusions— These findings support the concept that aortic valve calcification is not a random degenerative process but an active regulated process associated with an osteoblast-like phenotype.

Background — Despite the common occurrence of aortic stenosis, the cellular causes of the disorder are unknown, in part because of the absence of experimental models. We hypothesized that atherosclerosis and early bone matrix expression in the aortic valve occurs secondary to experimental hypercholesterolemia and that treatment with atorvastatin modifies this transformation. Methods and Results — To test this hypothesis, we developed an experimental hypercholesterolemic rabbit model. New Zealand White rabbits (n=48) were studied: group 1 (n=16), normal diet; group 2 (n=16), 1% (wt/wt) cholesterol diet; and group 3 (n=16), 1% (wt/wt) cholesterol diet plus atorvastatin (3 mg/kg per day). The aortic valves were examined with hematoxylin and eosin stain, Masson trichrome, macrophage (RAM 11), proliferation cell nuclear antigen (PCNA), and osteopontin immunostains. Cholesterol and highly sensitive C-reactive protein (hsCRP) serum levels were obtained by standard assays. Computerized morphometry and digital image analysis were performed for quantifying PCNA (% area). Electron microscopy and immunogold labeling were performed for osteopontin. Semiquantitative RT-PCR was performed for the osteoblast bone markers [alkaline phosphatase, osteopontin, and osteoblast lineage-specific transcription factor ( Cbfa-1 )]. There was an increase in cholesterol, hsCRP, PCNA, RAM 11, and osteopontin and osteoblast gene markers (alkaline phosphatase, osteopontin, and Cbfa-1 ) in the cholesterol-fed rabbits compared with control rabbits. All markers except hsCRP were reduced by atorvastatin. Conclusions — These findings of increased macrophages, PCNA levels, and bone matrix proteins in the aortic valve during experimental hypercholesterolemia provide evidence of a proliferative atherosclerosis–like process in the aortic valve associated with the transformation to an osteoblast-like phenotype that is inhibited by atorvastatin.

For many years it has been recognized that sex steroids have profound effects on bone metabolism. The current perception is that estrogen decreases bone resorption and androgen increases bone deposition. To investigate the potential for androgens to directly modulate bone resorption, we have examined avian osteoclast and human and mouse osteoclast-like cells for androgen responsiveness. There was a dose-dependent decrease in resorption activity in response to α-dihydrotestosterone (α-DHT), β-DHT, testosterone, or the synthetic androgen RU1881. This decrease was blocked by cotreatment with the specific androgen antagonist hydroxyflutamide. Further examination of avian osteoclasts revealed that the cells exhibited specific and saturable nuclear binding of tritiated RU1881 and that α-DHT stimulated the activity of the androgen response element as measured by using a chloramphenicol acetyltransferase reporter plasmid. In addition, avian osteoclasts responded to androgen treatment with elevated production and secretion of transforming growth factor β, a well documented response to androgen exposure in other cell systems. Treatment with either α-DHT or β-DHT for 24 hours resulted in a significant dose-dependent decrease in secretion of cathepsin B and tartrate-resistant acid phosphatase. This response to β-DHT, a stereoisomer of α-DHT that is inactive in other androgen receptor-dependent systems, supports the hypothesis that the osteoclast androgen receptor has unusual ligand-binding properties. Taken together, these results confirm the presence of functional androgen receptors in these cells and support the conclusion that osteoclasts are able to respond directly to androgens in vitro and thus are potential androgen target cells in vivo .