Thrombospondin is synthesized and secreted by human osteoblasts and osteosarcoma cells. A model to study the different effects of thrombospondin in cell adhesion

The endocytic receptor uPARAP is a regulator of extracellular thrombospondin-1

Thrombospondin-1 (TSP-1) is a matricellular protein with a multitude of functions in the pericellular and extracellular environment. We report a novel pathway for the regulation of extracellular TSP-1, governed by the endocytic collagen receptor, uPARAP (urokinase plasminogen activator receptor-associated protein; MRC2 gene product, also designated Endo180, CD280). First, using a novel proteomic approach for unbiased identification of ligands for endocytosis, we identify TSP-1 as a candidate ligand for specific uptake by uPARAP. We then show that uPARAP can efficiently internalize TSP-1 for lysosomal degradation, that this capability is not shared by other, closely related endocytic receptors and that uPARAP serves to regulate the extracellular levels of TSP-1 in vitro. Using wild type and uPARAP null mice, we also demonstrate uPARAP-mediated endocytosis of TSP-1 in dermal fibroblasts in vivo. Unlike other uPARAP ligands, the interaction with TSP-1 is sensitive to heparin and the responsible molecular motifs in uPARAP are overlapping, but not identical with those governing the interaction with collagens. Finally, we show that uPARAP can also mediate the endocytosis of TSP-2, a thrombospondin closely related to TSP-1, but not the more distantly related members of the same protein family, TSP-3, -4 and -5. These findings indicate that the role of uPARAP in ECM remodeling is not limited to the uptake of collagen for degradation but also includes an orchestrator function in the regulation of thrombospondins with numerous downstream effects. This is likely to be an important factor in the physiological and pathological roles of uPARAP in bone biology, fibrosis and cancer. The proteomic data has been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the data set identifier PXD031272.

Thrombospondins in bone remodeling and metastatic bone disease

Thrombospondins (TSPs) are a family of five multimeric matricellular proteins. Through a wide range of interactions, TSPs play pleiotropic roles in embryogenesis and in tissue remodeling in adult physiology as well as in pathological conditions, including cancer development and metastasis. TSPs are active in bone remodeling, the process of bone resorption (osteolysis) and deposition (osteogenesis) that maintains bone homeostasis. TSPs are particularly involved in aberrant bone remodeling, including osteolytic and osteoblastic skeletal cancer metastasis, frequent in advanced cancers such as breast and prostate carcinoma. TSPs are major players in the bone metastasis microenvironment, where they finely tune the cross talk between tumor cells and bone resident cells in the metastatic niche. Each TSP family member has different effects on the differentiation and activity of bone cells-including the bone-degrading osteoclasts and the bone-forming osteoblasts-with different outcomes on the development and growth of osteolytic and osteoblastic metastases. Here, we overview the involvement of TSP family members in the bone tissue microenvironment, focusing on their activity on osteoclasts and osteoblasts in bone remodeling, and present the evidence to date of their roles in bone metastasis establishment and growth.

Thrombospondin-2 deficiency in growing mice alters bone collagen ultrastructure and leads to a brittle bone phenotype

Thrombospondin-2 (TSP2) is a matricellular protein component of the bone extracellular matrix. Long bones of adult TSP2-deficient mice have increased endosteal bone thickness due to expansion of the osteoblast progenitor cell pool, and these cells display deficits in osteoblastic potential. Here, we investigated the effects of TSP2 deficiency on whole bone geometric and mechanical properties in growing 6-wk-old male and female wild-type and TSP2-knockout (KO) mice. Microcomputed tomography and mechanical testing were conducted on femora and L2 vertebrae to assess morphology and whole bone mechanical properties. In a second series of experiments, femoral diaphyses were harvested from wild-type and TSP2-KO mice. Detergent-soluble type I collagen content was determined by Western blot of right femora. Total collagen content was determined by hydroxyproline analysis of left femora. In a third series of experiments, cortical bone was dissected from the anterior and posterior aspects of the femoral middiaphysis and imaged by transmission electron microscopy to visualize collagen fibrils. Microcomputed tomography revealed minimal structural effects of TSP2 deficiency. TSP2 deficiency imparted a brittle phenotype on cortical bone. Femoral tissue mineral density was not affected by TSP2 deficiency. Instead, transmission electron microscopy revealed less intensely stained collagen fibrils with altered morphology in the extracellular matrix assembled by osteoblasts on the anterior surface of TSP2-KO femora. Femoral diaphyseal bone displayed comparable amounts of total collagen, but the TSP2-KO bones had higher levels of detergent-extractable type I collagen. Together, our data suggest that TSP2 is required for optimal collagen fibrillogenesis in bone and thereby contributes to normal skeletal tissue quality.

Thrombospondin-1 regulates bone homeostasis through effects on bone matrix integrity and nitric oxide signaling in osteoclasts

Thrombospondin-1 (TSP1), an endogenous antiangiogenic, is a widely expressed secreted ligand with roles in migration, adhesion, and proliferation and is a target for new therapeutics. While TSP1 is present in the bone matrix and several TSP1 receptors play roles in bone biology, the role of TSP1 in bone remodeling has not been fully elucidated. Bone turnover is characterized by coordinated activity of bone-forming osteoblasts (OB) and bone-resorbing osteoclasts (OC). TSP1-/- mice had increased bone mass and increased cortical bone size and thickness compared to wild type (WT). However, despite increased size, TSP1-/- femurs showed less resistance to bending than expected, indicative of diminished bone quality and a bone material defect. Additionally, we found that TSP1 deficiency resulted in decreased OC activity in vivo and reduced OC differentiation. TSP1 was critical during early osteoclastogenesis, and TSP1 deficiency resulted in a substantial overexpression of inducible nitric oxide synthase (iNOS). Importantly, administration of a NOS inhibitor rescued the OC function defects of TSP1-/- mice in vivo. To investigate the role of bone-derived TSP1 in osteoclastogenesis, we found that WT pre-OCs had defective iNOS expression when cultured on TSP1-/- bone compared to WT bone, suggesting that TSP1 in bone plays a critical role in iNOS signaling during OC development. These data implicate a new role for TSP1 in bone homeostasis with roles in maintaining bone matrix integrity and regulating OC formation. It will be critical to monitor bone health of patients administered TSP1-pathway directed therapeutics in clinical use and under development.

Thrombospondin-1 is a putative target gene of Runx2 and Runx3

Thrombospondin-1 (TSP-1), a matricellular protein widely acclaimed to be involved in the inhibition of angiogenesis and tumorigenesis, is synthesized and secreted by many cell types, including osteoblast and cancer cells. TSP-1 is highly upregulated during early stage of osteogenesis, whereas it inhibits terminal osteoblast differentiation. Expression of TSP-1 is downregulated in cancer cells, and its ectopic expression has been shown to restrain tumor growth. Transcriptional regulation of TSP-1 in osteogenesis and cancer is poorly understood; this prompted us to study its regulation by the two key regulators of the aforementioned processes: Runx2 and Runx3. Through a PCR-based cDNA subtraction technique, we identified and cloned a cDNA fragment for mouse TSP-1, whose expression was dramatically upregulated in response to Runx2 expression in mesenchymal stem cells. Moreover, TSP-1 expression was considerably reduced in the lung of Runx2 knockout mouse. On the other hand, TSP-1 gene expression drastically increased at both the transcriptional and translational levels in response to Runx3 expression in B16-F10 melanoma cells. In line with this, Runx2 and Runx3 bound to the TSP-1 promoter and stimulated its activity. Hence, these results provide first line of evidence that TSP-1 is a transcriptional target gene of Runx2 and Runx3.

Comparative profiling of plasma proteome from breast cancer patients reveals thrombospondin-1 and BRWD3 as serological biomarkers

Breast cancer is the most common cancer in women worldwide. It is necessary to identify biomarkers for early detection, to make accurate prognoses, and to monitor for any recurrence of the cancer. In order to identify potential breast cancer biomarkers, we analyzed the plasma samples of women diagnosed with breast cancer and age-matched normal healthy women by mTRAQ-based stable isotope-labeling mass spectrometry. We identified and quantified 204 proteins including thrombospondin-1 (THBS1) and bromodomain and WD repeat-containing protein 3 (BRWD3) which were increased by more than 5-fold in breast cancer plasma. The plasma levels of the two proteins were evaluated by Western blot assay to confirm for their diagnostic value as serum markers. A 1.8-fold increase in BRWD3 was observed while comparing the plasma levels of breast cancer patients (n = 54) with age-matched normal healthy controls (n = 30), and the area under the receiver operating characteristic curve (AUC) was 0.917. THBS1 was detected in pooled breast cancer plasma at the ratio similar to mTRAQ ratio (> 5-fold). The AUC value for THBS1 was 0.875. The increase of THBS1 was more prominent in estrogen receptor negative and progesterone receptor negative patients than receptor-positive patients. Our results are evidence of the diagnostic value of THBS1 in detecting breast cancer. Based on our findings, we suggest a proteomic method for protein identification and quantification lead to effective biomarker discovery.

The type II collagen N-propeptide, PIIBNP, inhibits cell survival and bone resorption of osteoclasts via integrin-mediated signaling

OBJECTIVE

Type IIB procollagen is characteristic of cartilage, comprising 50% of the extracellular matrix. The NH(2)-propeptide of type IIB collagen, PIIBNP, can kill tumor cells via binding to integrins α(V)β(3) and α(V)β(5). As osteoclasts rely on α(V)β(3) integrins for function in bone erosion, we sought to determine whether PIIBNP could inhibit osteoclast function.

METHODS

We undertook in vitro and in vivo experiments to evaluate both osteoblast and osteoclast functions in the presence of recombinant PIIBNP. Adhesion of osteoclasts to PIIBNP was analyzed by staining of attached cells with crystal violet. PIIBNP-induced cell death was evaluated by counting Trypan Blue stained cells. The mechanism of cell death was evaluated by DNA fragmentation, TUNEL staining and western blotting to detect cleaved caspases. To determine the role of α(V)β(3) integrin, osteoclasts were pretreated with α(V) or β(3) integrin specific siRNA before the treatment with PIIBNP. To explore PIIBNP function in vivo, a lipopolysaccharide-induced mouse calvaria lysis model was employed.

RESULTS

Osteoclasts adhered to PIIBNP via an RGD-mediated mechanism. When osteoclasts were plated on extracellular matrix proteins, PIIBNP induced apoptosis of osteoclasts via caspase 3/8 activation. Osteoblasts and macrophages were not killed. Reduction of α(V) or β(3) integrin levels on osteoclasts by siRNA reduced cell death in a dose-dependent manner. In vivo, PIIBNP could inhibit bone resorption.

CONCLUSION

We conclude that PIIBNP can inhibit osteoclast survival and bone resorption via signal transduction through the α(V)β(3) integrins. Because of this property and the cell specificity, we propose that PIIBNP may play a role in vivo in protecting cartilage from osteoclast invasion and also could be a new therapeutic strategy for decreasing bone loss.

Treatment of hydroxyapatite scaffolds with fibronectin and fetal calf serum increases osteoblast adhesion and proliferation in vitro

BACKGROUND

The use of hydroxyapatite in reconstructive surgery has been hampered by the fact that it is very slowly invaded by host tissues, a process that is critical to graft incorporation. Implant compatibility may be augmented by providing cellular binding sites and by seeding cells before implantation.

METHODS

Bone-forming cells were seeded onto hydroxyapatite disks, and precoated with fibronectin and fetal calf serum or phosphate-buffered saline. Cellular adhesion and proliferation was analyzed in vitro. For in vivo studies, experimental and control hydroxyapatite disks were seeded with green fluorescent protein-expressing cells and implanted into mice.

RESULTS

Fibronectin/fetal calf serum pretreatment improved cell attachment and cell growth significantly in vitro. After 48 hours, experimental disks (n = 5) contained 2.8 times more attached cells than controls (p < 0.001), and after 7 days this difference had increased further (4.2 times) (p < 0.001). In the in vivo part of the study, sections from implants (n = 4) harvested 3 days after implantation demonstrated an average of 122 +/- 50 green fluorescent protein-labeled cells/mm in the fibronectin/fetal calf serum group compared with 85 +/- 21 cells/mm in the phosphate-buffered saline controls. After 10 days, the cells had in general decreased in number in both groups, but the relation in cell density was similar to the first time point (19 +/- 11 versus 12 +/- 11 cells/mm).

CONCLUSION

In vitro attachment and proliferation of bone-forming cells on hydroxyapatite is significantly increased by pretreatment with fibronectin/fetal calf serum, but this difference is less profound and not significant in vivo.

Androgens Repress the Expression of the Angiogenesis Inhibitor Thrombospondin-1 in Normal and Neoplastic Prostate

In order to understand why the angiogenesis inhibitor thrombospondin-1 (TSP1) is often, although not always, associated with prostatic tumors, we have investigated its relationship with the testosterone and the vasculature on which both normal and tumorigenic prostatic epithelia depend. In vivo, androgen withdrawal led to increased TSP1 production and decreased vascularization in the normal rat prostate which was reversed by androgen replacement. Androgen repression of TSP1 production occurred at the transcriptional level and was dependent on the presence of the first intron of the TSP1 gene. In an experimental model of prostate tumorigenesis, TSP1, when delivered by admixed stromal fibroblasts, markedly delayed LNCaP tumor growth and limited tumor vascularization. However, prolonged exposure to TSP1 resulted in the growth of tumors secreting high levels of vascular endothelial growth factor in the bloodstream of tumor-bearing animals and tumor growth was no longer sensitive to TSP1 inhibitory effects. Clinical evidence also suggested that prostate carcinomas are able to adapt to escape the antiangiogenic effects of TSP1. In human androgen–dependent localized prostate carcinomas, TSP1 expression was inversely correlated with blood vessel density. Androgen deprivation in patients with hormone-responsive tumors led to increased TSP1 expression and vascular regression. In contrast, despite a sustained expression in the tumor bed, TSP1 was no longer associated with decreased vascularization in hormone-refractory prostate tumors. Overall, these results suggest that the high in situ TSP1 exposure triggered by androgen deprivation in patients with prostate cancer could lead to early tumor resistance. Such patients could benefit from a combination of androgen deprivation and antiangiogenic therapy in order to minimize the induction of such tumor escape.

Thrombospondin 1: a multifunctional protein implicated in the regulation of tumor growth

Thrombospondins belong to a family of extracellular matrix (ECM) proteins widely found from embryonic to adult tissues. The modular structure of thrombospondins contains a series of peptide sequences implicated in a multiplicity of biological functions. Extracellular matrix undergoes important alterations under proteolysis that occurs in pathological processes like tumorigenesis. An elevated secretion of thrombospondin 1 (TSP1) is often observed in tumors and is sometimes considered as a predictive factor. However, the role of TSP1 in cancer progression remains controversial and must be carefully apprehended. The regulation of cell adhesion, proliferation, apoptosis by TSP1 is examined in the present review and it is clear from the literature and from our investigations that TSP1 presents both stimulatory and inhibitory effects. The exposition of cryptic sites upon conformational changes can partially explain this contradiction. More interestingly, the analysis of TSP1-directed intracellular signaling pathways activated through specific receptors or supramolecular receptors docking systems may be useful to discriminate the precise function of TSP1 in tumor progression. The central role played by TSP1 in the control of matrix-degrading enzyme activation and catabolism reveals attractive tracks of research and highlights the involvement of the lipoprotein receptor-related protein (LRP) receptor in these events. Therefore, TSP1-derived peptides constitute a source of potentially active matrikins which could provide essential tools in cancer therapy.

Primary human marrow stromal cells and Saos-2 osteosarcoma cells use different mechanisms to adhere to hydroxylapatite

One important step in bone formation on hard tissue implants is adhesion of osteoblast precursors to the implant surface. In this study, we used function-blocking antibodies against integrin subunits to characterize the mechanisms used by human marrow stromal cells and Saos-2 osteosarcoma cells to adhere to protein-coated hydroxylapatite (HA). We found that Saos-2 use both alpha5- and alphav-containing integrins, whereas stromal cells use alphav-containing integrins but not alpha5-containing integrins, despite the presence of alpha5-containing integrins on cell surfaces. On the basis of this difference, we examined binding of these cell types to HA coated with fibronectin (FN) or vitronectin (VN), to determine whether these ligands for alpha5 and alphav integrins could enhance the numbers or morphology of cells adhered to them. We also examined the adhesion of cells to HA coated with RGD peptides designed to bind to FN or VN receptors. Morphology and number of adherent stromal cells were markedly enhanced on serum-coated surfaces compared with FN or VN alone, whereas, surprisingly, Saos-2 cells failed to spread on serum-coated HA and displayed superior spreading and stress fiber formation on FN-coated [corrected] HA. Collectively, these results have important implications for the design of protein coatings to enhance the performance of HA implants.

Adhesion of osteosarcoma cells to the 70-kDa core region of thrombospondin-1 is mediated by the alpha 4 beta 1 integrin

Thrombospondin-1 (TSP-1) is an extracellular glycoprotein that is involved in a variety of physiological processes such as tumor cell adhesion, invasion, and metastasis. It has been hypothesized that TSP-1 provides an adhesive matrix for osteosarcoma cells. Here we present data showing that TSP-1 can promote cell substrate adhesion to U2OS and SAOS cells through the alpha 4 beta 1 integrin. The dose-dependent adhesion to TSP-1 was inhibited by anti-integrin antibodies directed against the alpha 4 or beta 1 subunit, but not by control antibodies against other integrins. To localize the potential alpha 4 beta 1-binding site within the TSP-1 molecule, the protein was subjected to limited proteolysis with chymotrypsin in the absence of calcium. The stable 70-kDa core fragment produced under these conditions promoted alpha 4 beta 1-dependent osteosarcoma cell adhesion in a manner similar to that of the intact protein. Moreover adhesion experiments with neutralizing antibodies revealed that the adhesion was totally dependent on the alpha 4 beta 1 interaction. Further blocking experiments with potential inhibitory peptides revealed that the alpha 4 beta 1-mediated adhesion was not influenced by peptides containing the RGD sequence. Attachment to the 70-kDa fragment was strongly inhibited by the CS-1 peptide, which represents the most active recognition domain for alpha 4 beta 1 integrin in fibronectin. The present data provide evidence that TSP-1 contains an alpha 4 beta 1 integrin-binding site within the 70-kDa core region.

The role of thrombospondin-1 in tumor progression

The role of thrombospondin-1 (TSP-1) in tumor progression is both complex and controversial. It is clear from the literature that the function of TSP-1 in malignancy depends on the presence of other factors and the level of TSP-1 expression in the tumor tissue. High levels of TSP-1 secreted by tumors, which were engineered to overexpress TSP-1, inhibit tumor growth, while anti-sense inhibition of TSP-1 production in certain tumors also inhibits growth. Clearly, the presence of other factors in these experimental systems must be important. The role of TSP-1 in angiogenesis also depends on the levels of TSP-1, the presence and level of angiogenic stimulators such as basic fibroblast growth factor (bFGF), and the localization of TSP-1 in the tissue. Matrix-bound TSP-1 promotes capillary tube formation in the rat aorta model of angiogenesis, while TSP-1 inhibits bFGF- induced angiogenesis in the rat cornea model. The inhibitory effect also depends on the proteolytic state of TSP-1 since the amino terminus promotes angiogenesis in the cornea model, while the remaining 140-kDa fragment inhibits bFGF-induced angiogenesis. Both the stimulatory and inhibitory effects of TSP-1 are likely due to upregulation of matrix-degrading enzymes and their inhibitors. These enzymes are critical for maintaining optimal matrix turnover during angiogenesis. These varied TSP-1-dependent mechanisms offer new targets for the development of anti-angiogenic therapeutics for the treatment of a variety of cancers, as well as other pathologies involving inappropriate angiogenesis such as diabetic retinopathy.

Participation of thrombospondin-1 in the activation of latent transforming growth factor-beta in malignant glioma cells

Malignant glioma cells secrete transforming growth factor-beta (TGF-beta) and can activate latent TGF-beta. However, the mechanism of the latent TGF-beta activation has not yet been determined. This study examined whether thrombospondin-1 (TSP-1) secreted by malignant glioma cell lines participates in the activation of latent TGF-beta secreted by the glioma cells. Western blot analysis revealed that TSP-1 was present in both the cell lysates and the culture supernatants of all three malignant glioma cell lines (T98G, A172, and U251). A bioassay for TGF-beta activity revealed that all malignant glioma cell lines used in this study could activate latent TGF-beta by themselves. Latent TGF-beta 1 activation, evaluated by enzyme-linked immunosorbent assay, was inhibited by more than 50% by the addition of neutralizing anti-TSP-1 monoclonal antibody or anti-TSP-1 polyclonal antibody. These results indicate that TSP-1 has a predominant role in the activation of latent TGF-beta in malignant glioma cells.

Correlation of thrombospondin-1 and transforming growth factor-beta expression with malignancy of glioma

The expression of thrombospondin-1 (TSP-1) and its role in gliomas have not been well examined. In the present study TSP-1 expression in a panel of malignant glioma cell lines and the expression of TSP-1 and transforming growth factor (TGF-beta) proteins in low-grade and malignant glioma tissues were investigated. Reverse transcription-polymerase chain reaction analysis showed that nine of nine malignant glioma cell lines expressed TSP-1 mRNA, and seven of nine glioma lines expressed TSP-2 mRNA. Production and secretion of TSP-1 were examined in the T98G glioblastoma cell line by western blot analysis. Total TSP-1 protein content in the supernatant was 10 times higher than that in the cell lysate. Secretion of TSP-1 was examined in these glioma cell lines by western blot analysis. All glioma lines secreted significant levels of TSP-1. Bioassay showed that all tumor lines had the capacity to activate latent TGF-beta. Localization of TSP-1, TGF-beta1, -beta2, and -beta3 was examined immunohistochemically in surgically resected glioma tissues, including 11 glioblastomas, six anaplastic astrocytomas, and eight astrocytomas. Most glioblastomas expressed high levels of both TSP-1 and TGF-beta. Anaplastic astrocytomas expressed moderate levels of TSP-1 and TGF-beta. Most malignant gliomas expressed various levels of TGF-beta1, -beta2, and -beta3. The expression of both proteins, however, was weak in low-grade gliomas. Normal brain tissues around the tumors were negatively or very weakly positively stained for TSP-1 and TGF-beta. These results indicate that most malignant glioma cells express TSP-1 in vitro and in vivo, and the expression of TSP-1 and TGF-beta in vivo correlates with the histologic malignancy of glioma. Overexpression of both TSP-1 and TGF-beta may increase the biologic malignancy of malignant gliomas, through generating the active form of TGF-beta in tumor tissues.

The effect of polymeric chemistry on the expression of bone-related mRNAs and proteins by human bone-derived cells in vitro

This study used human bone-derived cells (HBDC) grown on two defined polymeric substrata to examine the effect of substrata chemistry on the expression of mRNAs and proteins characteristic of the osteoblastic phenotype. The growth profile of cells grown on tissue culture polystyrene (TCP) was exponential whereas for those seeded on polyethyleneterephthalate (PET) there was a pronounced lag period before cellular multiplication. The temporal expression pattern of mRNAs in HBDC cultured on TCP was similar to that of cells on PET. On TCP, the levels of several mRNAs peaked at day 4, as cellular proliferation slowed. In contrast, the induction in mRNA levels in cells grown on PET corresponded to maximum mitotic activity. There appears to be sequential cascade in protein expression in cells grown on TCP with overlapping peaks of thrombospondin (Tsp), osteocalcin (OC) and osteopontin (OP) expression. In contrast, peak intracellular protein expression levels for Tsp, OC and OP did not overlap when cells were grown on PET.

Restricted localization of thrombospondin-2 protein during mouse embryogenesis: a comparison to thrombospondin-1

Thrombospondin-1 and -2 (TSP1 and TSP2) are multifunctional, multimodular extracellular matrix proteins encoded by separate genes. We compared the distributions of TSP1 and TSP2 in mouse embryos (day 10 and later) by immunohistochemistry. TSP1 was detected on day 10 in the heart and intestinal epithelium, on day 11 in megakaryocytes, and on day 14 in the lung. TSP2 was not detected until day 14, with strongest staining in mesenchymal condensation that gives rise to cartilage and bone. The distribution of TSP2 was different from but overlapped with the distribution of TSP1. TSP1 was found in cartilage proper with diminished staining around chondrocytes undergoing differentiation and hypertrophy, whereas TSP2 was restricted to the matrix surrounding chondrocytes of the growth zone cartilage. TSP2 and TSP1 were both expressed in centers of intramembranous ossification that form the skull bones, in reticular dermis, on the apical surface of nasal epithelium, in skeletal muscle, and in the sheath surrounding vibrissae. Areas of exclusive staining for TSP2 included the perichondrium surrounding the cartilage of the nasal cavities, developing bone of the lower mandible, and adrenal gland. The distinct localizations of TSP1 and TSP2 indicate that the two proteins have specific functions during mouse embryogenesis.

Characterization of cell-matrix adhesion requirements for the formation of fascin microspikes

Cell adhesion to thrombospondin-1 (TSP-1) correlates with assembly of cell-substratum contact structures that contain fascin microspikes. In this analysis, cell-matrix requirements for assembly of fascin microspikes were examined in detail. In six cell lines, cell spreading on a TSP-1 substratum correlated with expression of fascin protein and formation of fascin microspikes. Microspikes were not formed by H9c2 cells adherent on fibronectin, vitronectin, collagen IV, or platelet factor 4. However, both fascin microspikes and focal contacts were assembled by cells adherent on laminin-1. Using mixed substrata containing different proportions of TSP-1, and fibronectin, fascin microspike formation by H9c2 and C2C12 cells was found to be reduced on substrata containing 25% fibronectin and abolished on substrata containing 75% fibronectin. Adhesion to intermediate mixtures of TSP-1 and fibronectin resulted in coassembly of fascin microspikes and focal contacts, colocalization of fascin with actin stress fiber bundles and altered distributions of beta 1 integrins, cortical alpha-actinin, and tropomyosin. In cells adherent on 50% TSP-1:50% fibronectin, GRGDSP peptide treatment decreased focal contact assembly and altered cytoskeletal organization but did not inhibit microspike assembly. Treatment with chondroitin sulfate A or p-nitrophenol beta-D-xylopyranoside decreased microspike formation and modified cytoskeletal organization but did not inhibit focal contact formation. In polarized migratory and postmitotic C2C12 cells, fascin microspikes and ruffles were localized at leading edges and TSP matrix deposition was also concentrated in this region. Depletion of matrix TSP by heparin treatment correlated with decreased microspike formation and cell motility. Thus, the balance of adhesive receptors ligated at the cell surface during initial cell-matrix attachment serves to regulate the type of substratum adhesion contact assembled and subsequent cytoskeletal organization. A role for fascin microspikes in cell motile behavior is indicated.

Osteotropic factor-stimulated synthesis of thrombospondin in rat dental pulp cells

The amount of thrombospondin (TSP) mRNA in confluent clonal rat dental pulp cells was increased by transforming growth factor-beta (TGF-beta), 1 alpha-25-dihydroxyvitamin D3 (1,25(OH)2D3), and phorbol 12,13-didecanoate (PDD), with maximal levels 6, 36 and 2 h, respectively, after stimulation. These increases were accompanied by enhanced syntheses of TSP proteins which were found in the different forms in cell layer/matrix fraction (198 and 165 kDa TSP) and the culture medium (180 kDa TSP). These three factors also raised the mRNA level of osteopontin, which is thought to play an important role in mineralization in dentin and bone. The order of effectiveness of these factors was PDD > TGF-beta > 1,25(OH)2D3 for all the stimulations described above. These results suggest that the osteotropic factors enhance TSP synthesis at the pretranslational level and that TSP produced by dental pulp cells participates in formation of reparative dentin.

Phenotypic expression of marrow cells when grown on various substrata

Our aim was to study the role of various extracellular matrices (ECM) on growth and differentiation of marrow stromal cells in vitro. Morphology changes, gene expression, and enzymatic activities were monitored in stromal osteoblastic MBA-15 and adipocytic 14F1.1 cells. These stromal cells were plated on dishes precoated with different substrata, such as matrigel (basement membrane), collagen type I, and endothelial ECM, and compared with cells plated on protein-free dishes. Striking morphological differences were observed when the cells grew on these different substrata. Changes in cell shape and growth also led to differential mRNA expression and enzymatic activities. When MBA-15 cells were plated on collagen, there was a decrease in mRNA for alkaline phosphatase (ALK-P), osteopontin (OP), and osteonectin (ON), and an increase in mRNA for procollagen (I). A differential effect was noted on 14F1.1 cells, the mRNA for ALK-P increased, the expressions of OP and ON lowered, and no expression for procollagen (I) was monitored. MBA-15 cells cultured on matrigel had decreased mRNA for ALK-P and OP, while they had increased ON mRNA expression and remained unchanged for procollagen I. No change in mRNA expression by 14F1.1 cells was monitored when cultured on matrigel. Functional enzymatic activities of ALK-P markedly decreased in MBA-15 cells cultured on various substrata, and increased or were unchanged in 14F1.1 cells. An additional enzyme, neutral endopeptidase (CD10/NEP), altered differentially in both cell types; this enzymatic activity increased or was unchanged when cells were cultured on these matrices. The results indicate a specific role for different ECM on various stromal cell types and their function.

Association of thrombospondin-1 with osteogenic differentiation of retinal pericytes in vitro

Vascular pericytes can differentiate into osteoblast-like cells in vitro, suggesting that these cells may represent a potential source of osteoprogenitor cells in the adult. Pericyte differentiation is associated with a characteristic pattern of nodule formation and mineralisation. Nodules are formed in post-confluent cultures by the retraction of multilayered areas. Crystals of hydroxyapatite are deposited on the extracellular matrix of these nodules which then becomes mineralised. We now demonstrate that thrombospondin-1 (TSP-1) gene expression is modulated during pericyte differentiation in vitro. That is, the relative levels of TSP-1 (protein and mRNA) increased markedly during nodule formation and then decreased when mineralisation of the nodules had taken place. TSP-1 was localised throughout non-mineralised nodules but it was largely excluded from the inner mass of mineralised nodules. The production of a mineralised matrix by vascular pericytes was promoted by the presence of antibodies to TSP-1 in the culture medium and was inhibited by exogenous TSP-1. These effects did not appear to be mediated through the activation of latent TGF-beta, since neither exogenous TGF-beta nor neutralising antibodies to TGF-beta had any effect on the rate or extent of mineralisation seen in the pericyte cultures. Taken together these results suggest that high levels of TSP-1 inhibit pericyte mineralisation, supporting the view that this protein plays a role in pericyte differentiation and bone formation.

Distribution of thrombospondin and integrin alpha V in DCIS, invasive ductal and lobular human breast carcinomas. Analysis by electron microscopy

The ultrastructural distribution of thrombospondin (TSP) and its cell surface receptor, integrin alpha V, was studied in two cases of human breast carcinoma: one of ductal carcinoma in situ (DCIS) with an invasive component, and one of invasive lobular carcinoma. In DCIS, moderate immunolabelling for TSP and integrin alpha V was observed in the rough endoplasmic reticulum and at the plasma membrane of intraductal carcinoma cells. TSP was also associated with extracellular matrix collagen fibrils surrounding in situ carcinoma cells. In the invasive part of this ductal carcinoma, most of the malignant cells were negative for TSP, while integrin alpha V was moderately expressed in these cells. In sharp contrast, typical strands of invasive lobular carcinoma cells in "Indian file" showed moderate TSP immunostaining in the rough endoplasmic reticulum and strong immunoreactivity for TSP at the plasma membrane and in the extracellular matrix. Moderate to strong immunoreactivity for integrin alpha V was also observed in invasive lobular carcinoma cells. Because of the role of TSP during cancer cell invasion, the different expression patterns of TSP in invasive ductal versus lobular carcinoma may well reflect biological differences between these two types of breast carcinoma and could account for the peculiar diffuse invasive behaviour of breast lobular carcinoma cells.

Thrombospondin mediates calcium mobilization in fibroblasts via its Arg-Gly-Asp and carboxyl-terminal domains

Thrombospondin is a matrix glycoprotein found in various cells that can modulate cell attachment, migration, and proliferation. We now show that intact soluble thrombospondin causes a transient [Ca2+]i increase in IMR-90 fibroblasts. This [Ca2+]i increase is mediated partly by the RGD-containing domain of thrombospondin that binds to the integrin alpha v beta 3 as demonstrated by inhibitor studies using anti-alpha v beta 3 antibody and RGD-containing peptides. A non-RGD and non-alpha v beta 3 component of this [Ca2+]i increase is mediated by the carboxyl-terminal domain of thrombospondin through an unidentified receptor on fibroblasts as shown by the antibody to the carboxyl-terminal of thrombospondin, C6.7. In addition, the carboxyl-terminal derived peptide, RFYVVMWK, also triggers [Ca2+]i increase in approximately 35% of fibroblasts. Both EGTA and Ni2+ block the entire [Ca2+]i increase indicating that this is due to an influx of extracellular Ca2+. B6H12, an antibody to the integrin-associated protein, blocks this [Ca2+]i increase by 50%, suggesting that some of the Ca2+ might be entering through an integrin-associated calcium channel. The current findings demonstrate that multiple domains on thrombospondin can trigger signal transduction events by increasing [Ca2+]i through their interactions with different cell receptors.

Bone marrow interface: preferential attachment of an osteoblastic marrow stromal cell line

In this study, we report on the cell adhesion properties of marrow stromal cells to extracellular matrix components such as collagen and noncollagenous proteins. The osteoblastic cells and their non-osteoblastic counter-parts (MBA series) from the marrow stroma differentially recognized a spectrum of extracellular matrix proteins. The osteoblastic cells, MBA-15, preferentially attached to bone matrix proteins, whereas fibroendothelial MBA-2.1 and adipocytic 14F1.1 cells did not. The MBA-15 cells demonstrated a preference in their attachment to fibronectin > mixture of collagens > bone matrix extracts > collagen type I > noncollagenous proteins. Clonal subpopulations derived from the MBA-15 cell line representing various stages along the osteogenic lineage expressed differential attachment preference. MBA-15.4, a less differentiated clonal line, was compared to MBA-15.6, a mature cell line.

Thrombospondin-4 is expressed by early osteogenic tissues in the chick embryo

The thrombospondins are a family of related glycoproteins found in the embryonic extracellular matrix. To date, five members of this family have been identified. Thrombospondin-1 and thrombospondin-2 have similar primary structure, but are expressed in different tissues at different times during development. Thrombospondins-3, -4, and cartilage oligomeric protein belong to a second thrombospondin subgroup in which the carboxyl-half of each molecule is most similar to thrombospondin-1 and -2. Here we report the cloning and sequencing of a novel probe to avian thrombospondin-4. We have used this probe to determine the origins of thrombospondin-4 in the chick embryo by in situ hybridization. Thrombospondin -4 transcripts first appear in the mesenchyme surrounding bone anlage coinciding with the initial stages of osteogenesis. The expression in osteogenic tissues is transient: thrombospondin-4 mRNAs are not seen in the osteoblasts of bone collars in developing long bones. This pattern is distinct from avian thrombospondin-2 which is expressed in perichondrium and embryonic fibrous connective tissues. Our observations indicate that connective tissues are the principal site of thrombospondin-4 expression in the chick. The diverse origins of different thrombospondin gene family members imply distinctive roles for these proteins related to the growth and differentiation of cartilage, tendons, and bone.

Cell-type specific adhesive interactions of skeletal myoblasts with thrombospondin-1

Thrombospondin-1 (TSP-1) is an extracellular matrix glycoprotein that may play important roles in the morphogenesis and repair of skeletal muscle. To begin to explore the role of thrombospondin-1 in this tissue, we have examined the interactions of three rodent skeletal muscle cell lines, C2C12, G8, and H9c2, with platelet TSP-1. The cells secrete thrombospondin and incorporate it into the cell layer in a distribution distinct from that of fibronectin. Myoblasts attach and spread on fibronectin- or thrombospondin-coated substrates with similar time and concentration dependencies. Whereas cells adherent on fibronectin organize actin stress fibers, cells adherent on TSP-1 display prominent membrane ruffles and lamellae that contain radial actin microspikes. Attachment to thrombospondin-1 or the 140-kDa tryptic fragment is mediated by interactions with the type 1 repeats and the carboxy-terminal globular domain. Attachment is not inhibited by heparin, GRGDSP peptide, or VTCG peptide but is inhibited by chondroitin sulphate A. Integrins of the beta 1 or alpha V subgroups do not appear to be involved in myoblast attachment to TSP-1; instead, this process depends in part on cell surface chondroitin sulphate proteoglycans. Whereas the central 70-kDa chymotryptic fragment of TSP-1 does not support myoblast attachment, the carboxy-terminal domain of TSP-1 expressed as a fusion protein in the bacterial expression vector, pGEX, supported myoblast attachment to 30% the level of intact TSP-1. Thrombospondin-4 (TSP-4) is also present in skeletal muscle and a fusion protein containing the carboxy-terminal domain of TSP-4 also supported myoblast adhesion, although this protein was less active on a molar basis than the TSP-1 fusion protein. Thus, the carboxyterminal domain of TSP-1 appears to contain a primary attachment site for myoblasts, and this activity is present in a second member of the thrombospondin family.

Binding and activation of plasminogen on the surface of osteosarcoma cells

Plasmin (Pm) is a broad action serine protease implicated in numerous physiological functions. In bone, Pm may play a role in growth, resorption, metastasis, and the activation of growth factors. The various components of the Pm system are known to bind and function on the cell surface of various cell types, but no pertinent data are available describing membrane-bound Pm or its zymogen, plasminogen (Pg), in either normal or neoplastic bone cells. We report here that Pg binds to the surface of the human osteosarcoma cell line MG-63 and is activated to Pm by endogenous urokinase plasminogen activator (uPA). These conclusions are based on experiments utilizing radiolabeled compounds and a cell surface proteolytic assay measuring amidolytic activity of Pm. 125I-Pg binding to cells was time dependent, saturable, reversible, and specific. Binding was characterized by a relatively low affinity (Kd approximately 0.9 microM) and a high capacity (approximately 7.5 x 10(6) sites/cell). The binding of 125I-Pg was associated with lysine binding sites of the plasminogen molecule. Activation of 125I-Pg to 125I-Pm occurred on the cell surface and was dependent upon cell bound uPA, as determined by inhibitory antibodies. Binding of Pg to MG-63 monolayers represented approximately 80% bound specifically to the cell surface and the remainder to the surrounding extra-cellular matrix. Either co-incubation with uPA or pre-incubation with Pm resulted in increased 125I-Pg binding to osteosarcoma cells. Cell surface Pm proteolytic activity was confirmed by an amidolytic chromogenic assay. Both Pm and Pg bound to cells with Pg being activated by endogenous uPA. Plasmin activated on the cell surface was partially protected from inhibition by alpha 2-antiPm (requiring Pm lysine binding site interaction) but inhibited by aprotinin, (interacting directly with the Pm catalytic site). Resistance of cell bound Pm to alpha 2-antiPm inhibition suggests that cell surface proteolysis can occur in the presence of a soluble Pm inhibitor known to exist in the extracellular space. Based on these results, we speculate that the various bone physiological processes implicating Pm may occur at or near the bone cell surface.

Cellular expression of bone-related proteins during in vitro osteogenesis in rat bone marrow stromal cell cultures

Rat bone marrow stromal cells comprise a heterogeneous mixture of cell lineages including osteoblastic cells. When grown in the presence of ascorbic acid, beta-glycerophosphate and 10(-8) M dexamethasone, osteoprogenitor cells within the population divide and differentiate to form bone nodules (Maniatopoulos et al., 1988, Cell Tissue Res., 254:317-330; Aubin et al., 1990, J. Bone Miner. Res., 5:S81) providing a useful model to investigate temporal and spatial changes in expression of osteoblastic markers. Immunocytochemistry was combined with Northern blotting, enzymatic assay, and radioimmunoassay to analyze the expression of bone-related proteins during the growth and differentiation sequence. By mRNA levels, protein production and/or enzymatic activity, expression of osteocalcin, bone sialoprotein, and alkaline phosphatase increased concomitantly with the development of bone nodules, while osteopontin mRNA levels decreased and those of SPARC/osteonectin did not change significantly. In older cultures with mineralizing nodules, mRNA levels for alkaline phosphatase and bone sialoprotein, but not osteocalcin, declined. Immunolabeling revealed that cells in early cultures stained poorly for SPARC/osteonectin and strongly for thrombospondin. Later, SPARC/osteonectin staining increased in most cells, while thrombospondin staining could be seen in both matrix and in cells, but with marked intercellular variability in intensity. At all time points studied, osteoblasts within bone nodules stained homogeneously for thrombospondin and alkaline phosphatase, and with marked heterogeneity of intensity amongst cells for SPARC/osteonectin and osteocalcin. Labelling with RCC455.4, a monoclonal antibody raised against rat calvaria cells which intensely labels osteoblasts and osteocytes (Turksen et al., 1992, J. Histochem. Cytochem., 40:1339-1352), co-localized with osteocalcin. Alkaline phosphatase activity and the amount of osteocalcin determined by both radioimmunoassay and immunolabelling decreased in very late cultures, a time corresponding to appearance of fully mineralized nodules. These studies indicate that the bone marrow stromal cell system is a useful model to study the temporal and spatial expression of bone-related proteins during osteogenesis and formation, mineralization, and maturation of bone nodules. Further, immunolabelling at the individual cell and single bone nodule level allowed discrimination of marked variability of expression of osteoblast markers during the differentiation sequence.

The growth-supportive effect of thrombospondin (TSP1) and the expression of TSP1 by human MG-63 osteoblastic cells are both inhibited by dexamethasone

Thrombospondin (TSP) is a 450-kDa extracellular matrix glycoprotein which supports the growth of human MG-63 osteoblastic cells [Abbadia et al., FEBS Lett., 329 (1993) 341-346]. In this study, we describe the effect of the glucocorticoid, dexamethasone, on cell proliferation and TSP expression by MG-63 cells. Using a serum-free mitogenesis assay, dexamethasone (25 to 500 nM) caused a dose-dependent decrease in [3H]thymidine incorporation by MG-63 cells in culture, reaching 40% inhibition of cell proliferation at a concentration of 250 nM. Similarly, the stimulatory effect of TSP (500 ng/ml) on proliferation of MG-63 cells was totally abolished in the presence of dexamethasone (250 nM). In situ hybridization indicated that TSP mRNA level in dexamethasone-treated MG-63 cells decreased compared to quiescent cells. As judged by fluorescence-activated cell sorting analysis, dexamethasone treatment of MG-63 cells resulted in a 50 to 70% decrease in TSP cell surface expression compared to quiescent cells. Secretion of TSP in the culture fluid of dexamethasone-treated MG-63 cells also decreased by 40% while, under similar experimental conditions, a 180% increase in alkaline phosphatase activity was observed in dexamethasone-treated cells. Because glucocorticoids induce osteoporosis in vivo and reduce proliferation of osteoblasts in vitro, our results argue for an important role of TSP during bone formation.

Thrombospondin (TSP1) mediates in vitro proliferation of human MG-63 osteoblastic cells induced by alpha-thrombin

Thrombospondin (TSP) is a 450-kDa glycoprotein synthesized and secreted by human MG-63 osteoblastic cells. In this study, we have first studied the effect of alpha-thrombin on TSP expression by human MG-63 cells. In situ hybridization indicated that TSP mRNA level in thrombin-treated MG-63 cells was increased when compared to unstimulated cells. As judged by immunofluorescence, thrombin-treatment of MG-63 cells resulted in increased cell surface expression of TSP when compared to quiescent cells. Because thrombin stimulates proliferation of osteoblastic cells, the involvement of TSP in proliferation of thrombin-stimulated osteoblastic cells was then investigated using a serum-free mitogenesis assay. Both alpha-thrombin (0.01 to 0.15 U/ml) and TSP (5 to 600 ng/ml) caused a dose-dependent increase in [3H]thymidine incorporation by MG-63 cells. Proliferation of osteoblastic cells induced by alpha-thrombin or TSP was specifically and totally inhibited by anti-TSP monoclonal antibodies (3-10 micrograms/ml) or by indomethacin (1 microM), an inhibitor of prostaglandin synthesis. Anti-TSP antibodies which inhibited cell proliferation also inhibit TSP expression to the surface of these cells. Our experiments support the existence of a mechanism whereby TSP bound to the cell surface of thrombin-treated MG-63 cells stimulates secretion of prostaglandins which, in turn, allow cell proliferation to proceed.

Osteoblast-specific factor 2: cloning of a putative bone adhesion protein with homology with the insect protein fasciclin I

A cDNA library prepared from the mouse osteoblastic cell line MC3T3-E1 was screened for the presence of specifically expressed genes by employing a combined subtraction hybridization/differential screening approach. A cDNA was identified and sequenced which encodes a protein designated osteoblast-specific factor 2 (OSF-2) comprising 811 amino acids. OSF-2 has a typical signal sequence, followed by a cysteine-rich domain, a fourfold repeated domain and a C-terminal domain. The protein lacks a typical transmembrane region. The fourfold repeated domain of OSF-2 shows homology with the insect protein fasciclin I. RNA analyses revealed that OSF-2 is expressed in bone and to a lesser extent in lung, but not in other tissues. Mouse OSF-2 cDNA was subsequently used as a probe to clone the human counterpart. Mouse and human OSF-2 show a high amino acid sequence conservation except for the signal sequence and two regions in the C-terminal domain in which 'in-frame' insertions or deletions are observed, implying alternative splicing events. On the basis of the amino acid sequence homology with fasciclin I, we suggest that OSF-2 functions as a homophilic adhesion molecule in bone formation.

The functions of thrombospondin and its involvement in physiology and pathophysiology

The thrombospondin family of molecules is expressed in many different tissues. Its expression is highly regulated by different hormones and cytokines and is developmentally controlled. It can bind to many different cell types, probably via an array of receptors which are similarly regulated. The level of thrombospondins in body fluids and their distribution in tissue change in correlation with various pathological states. It is linked to the growth of primary tumors and to metastasis, to development of the atherosclerotic plaque, to malaria infection and other diseases. The role(s) of thrombospondin(s) are by and large unknown, though specific interaction seem to affect particular cell functions. The wide-spread spatial and temporal regulation, multiple interactions and correlation with major diseases imply important roles in cell function and call for concerted effort to unravel the mystery.

Thrombospondin-platelet interactions. Role of divalent cations, wall shear rate, and platelet membrane glycoproteins

The role of thrombospondin, a multifunctional matrix glycoprotein, in platelet adhesion is controversial: both adhesive and antiadhesive properties have been attributed to this molecule. Because shear flow has a significant influence on platelet adhesion, we have assessed thrombospondin-platelet interactions both under static and flow conditions. The capacity of thrombospondin to support platelet adhesion depended upon its conformation. In a Ca(2+)-depleted conformation, such as in citrated plasma, thrombospondin was nonadhesive or antiadhesive as it inhibited platelet adhesion to fibrinogen, fibronectin, laminin, and von Willebrand factor by 30-70%. In a Ca(2+)-replete conformation, however, thrombospondin effectively supported platelet adhesion. Shear rate influenced this adhesion; percent surface coverage on thrombospondin increased from 5.4 +/- 0.3 at 0 s-1 to 41.5 +/- 6.7 at 1,600 s-1. In contrast to the extensive platelet spreading observed on fibronectin at all shear rates, platelet spreading on thrombospondin occurred only sporadically and at high shear rates. GPIa-IIa, GPIIb-IIIa, GPIV, and the vitronectin receptor, which are all proposed platelet receptors for thrombospondin, were not solely responsible for platelet adhesion to thrombospondin. These results suggest that thrombospondin may play a dual role in adhesive processes in vivo: (a) it may function in conjunction with other adhesive proteins to maintain optimal platelet adhesion at various shear rates; and (b) it may serve as a modulator of cellular adhesive functions under specific microenvironmental conditions.

Regulation of expression of the chondrocytic phenotype in a skeletal cell line (CFK2) in vitro

We have examined in vitro the spontaneous and regulated expression of phenotypic characteristics associated with differentiated chondrocytes in an established skeletal cell line (CFK2) derived from fetal rat calvariae. Extended culture of CFK2 cells resulted in the appearance of glycosaminoglycans and type II collagen in the cell layer in association with the formation of focal nodes of cells. In addition, induction of mRNA-encoding link protein, cartilage-specific proteoglycan core protein, and thrombospondin was observed in the differentiated population (dCFK2 cells). The expression of these mRNAs was present for at least two passages after subculturing the dCFK2 cells. The dCFK2 cells also demonstrated enhanced parathyroid hormone (PTH)-stimulated adenylate cyclase activity. Proliferation of CFK2 cells was stimulated by the peptide regulatory factors EGF and PTH and inhibited by the steroidal agents dexamethasone and retinoic acid. EGF and retinoic acid inhibited the formation of cell foci and glycosaminoglycan deposition and the expression of mRNA-encoding link protein. In contrast, PTH and dexamethasone enhanced the formation of focal cellular nodes and augmented matrix deposition and link protein mRNA expression. These studies therefore show that the CFK2 cell line can serve as a nontransformed model of rat chondrocytic cells in which both induction and regulation of the expression of cartilaginous matrix components can be observed. This line thereby provides a unique renewable source of chondrocytic precursor cells and an excellent in vitro model for evaluating temporal and environmental control of chondrocyte differentiation and cartilage matrix production.

Diverse mechanisms for cell attachment to platelet thrombospondin

Thrombospondin-1 is a component of the extracellular matrix which is thought to play important roles in cell migration and proliferation, during embryogenesis and wound repair. To understand the basis for these activities, we are mapping the regions of the molecule with cell adhesive activity. Here, we use antagonists of specific cell binding sites, adhesion-perturbing thrombospondin monoclonal antibodies and proteolytic fragments of platelet thrombospondin, to investigate the adhesive mechanisms used by G361 melanoma cells, human intestinal smooth muscle cells (HISM), epidermal keratinocytes and MG-63 osteosarcoma cells. When attached to the same preparations of platelet thrombospondin, HISM and MG-63 cells underwent spreading, whereas G361 cells and keratinocytes did not. Attachment of all four cell types involved the carboxyterminal domain. The type 1 repeats and the amino-terminal heparin binding domain were important for stable attachment of G361, HISM and MG-63 cells, but were not involved in keratinocyte attachment. GRGDSP peptide caused near complete inhibition of HISM and MG-63 cell attachment, partially inhibited G361 attachment, but did not inhibit keratinocyte attachment. Attachment of HISM and MG-63 cells involved the alpha v beta 3 integrin. The integrity of the thrombospondin molecule was important for its adhesivity towards G361, HISM, and MG-63 cells, whereas keratinocytes attached to the 140 kDa tryptic fragment as effectively as they did to the intact molecule. These results show that cell attachment to platelet thrombospondin typically involves multiple binding interactions, but the exact profile of interactions is cell type specific. Usage of particular cell-binding sites does not predict whether cells will undergo spreading or not. These data may, in part, explain some of the current controversies surrounding the mechanisms of cell attachment to thrombospondin.

Attachment to extracellular matrix molecules by cells differing in the expression of osteoblastic traits

Two sets of clonal cell populations differing in the expression of osteoblastic traits, the rat osteosarcoma cell lines ROS 17/2.8 and ROS 25/1 and the immortalized fetal rat calvarial cell lines RCT-1 and RCT-3, were compared for their ability to attach to a series of extracellular matrix (ECM) constituents in vitro. Both osteoblastic (ROS 17/2.8, RCT-3) and nonosteoblastic (ROS 25/1, RCT-1) cell lines attached in a time- and concentration-dependent manner to plates coated with fibronectin (FN), osteopontin (OP), type I collagen (Col I), type IV collagen (Col IV), and laminin (LN) but only weakly to osteocalcin (OC) and thrombospondin (TSP). In both systems, the osteoblastic and nonosteoblastic clones attached identically to FN. Both ROS 17/2.8 and ROS 25/1 attached to similar molar amounts of substrate with the same preference order: FN > LN > Col I > or = Col IV. Maximal ROS 17/2.8 attachment to OP was > or = Col I but required approximately 2.5 times more substrate. ROS 25/1 attached less effectively than ROS 17/2.8 to most non-FN substrates. RCT-3 cells attached similarly to ROS 17/2.8 except that the preference order for Col I and LN was reversed and attachment to OP was lower than for ROS 17/2.8 RCT-1 cells attached best to Col I rather than FN, and equaled or surpassed RCT-3 in attachment to other non-FN substrates. Thus in these experimental systems, cells expressing an osteoblast-like phenotype exhibited generally similar ECM attachment properties. Their nonosteoblastic counterparts recognized the same spectrum of ECM constituents but differed from the osteoblastic cells and from each other in the effectiveness of their attachment to substrates other than FN.

Formation of mineralized nodules by bone derived cells in vitro: a model of bone formation?

The identification of the factors which regulate the proliferation and differentiation of cells of the osteoblast lineage remains one of the major challenges in the field of bone cell biology. Although considerable progress has been made in the isolation and culture of cells of the osteoblast lineage from both animal and, more recently, human bone, uncertainties have persisted as to the extent to which these cell populations retain the ability to differentiate into functional osteoblasts in vitro. The formation in vitro of mineralized nodules that exhibit the morphological, ultrastructural and biochemical characteristics of embryonic/woven bone formed in vivo, represents the first evidence that the differentiation of functional osteoblasts can occur in cultures of isolated animal bone-derived cell populations. It is clear, however, that the culture conditions employed at present only permit a small number of cells to differentiate to the extent of being capable of organising their extracellular matrix into a structure that resembles that of bone. Moreover, it has generally been found that the reproducible mineralization of this extracellular matrix requires supplementation of the culture medium with mM concentrations of beta-GP, which raises doubts as to the physiological relevance of this process. The formation of nodules has also been observed in cultures of human bone-derived cells. As found in cultures of animal bone-derived cells, reproducible mineralization of these nodules will occur in the presence of beta-GP. We have shown, however, that in the presence of the long acting ascorbate analogue Asc-2-P, the formation and mineralization of nodules can occur in the absence of beta-GP. The nodules formed in human bone-derived cell cultures have yet to be characterized as rigorously as those formed in cultures of animal bone-derived cells and thus it remains to be shown that they resemble bone formed in vivo.

Rat osteoclasts adhere to a wide range of RGD (Arg-Gly-Asp) peptide-containing proteins, including the bone sialoproteins and fibronectin, via a beta 3 integrin

The ligand binding ability of rat osteoclast adhesion receptors was investigated in an attachment assay using osteoclasts disaggregated from bone. Osteoclasts adhered well to the Arg-Gly-Asp (RGD)-containing proteins osteopontin (bone sialoprotein I) and BSP (bone sialoprotein II), vitronectin, fibrinogen, von Willebrand factor, and fibronectin. Osteoclasts also adhered, but less strongly, to type I collagen. No attachment of osteoclasts was observed to thrombospondin, tenascin, laminin, or a range of non-RGD-containing bone proteins and proteins from other sources. The attachment of osteoclasts to all ligands was abolished in the presence of GRGDSP peptide, indicating the involvement of the RGD cell binding sequence in ligand binding. Attachment of osteoclasts to all substrates, with the exception of type I collagen, was also strongly inhibited by the addition of monoclonal antibody F11 to the beta 3 integrin subunit, indicating that a beta 3 integrin, probably the vitronectin receptor, was involved. Attachment to type I collagen was blocked by EDTA chelation of divalent cations and was not significantly affected by anti-beta 3 or anti-beta 1 antibodies; when taken with the inhibition by RGD peptide, this suggests the involvement of various receptors, possibly including nonintegrin collagen receptors, in the binding of osteoclasts to this protein. These results define the wide range of ligands for extracellular matrix receptors in osteoclasts in vitro. It remains to be established which of these proteins are important in osteoclast adhesion and osteoclastic bone resorption in vivo.

Thrombospondin in ligament, meniscus and intervertebral disc

The presence of thrombospondin in extracts of ligament, meniscus and human nucleus pulposus was demonstrated by Western blot analysis using an anti-human thrombospondin antibody. Metabolic labeling of ligament fibroblast and meniscal fibrochondrocyte cultures followed by immunoprecipitation with anti-human thrombospondin revealed a single band corresponding to the molecular weight of thrombospondin when analyzed by SDS-PAGE and fluorography under reduced and non-reduced conditions. Thrombospondin, therefore, is a matrix constituent of ligament, knee joint meniscus and the nucleus pulposus and is synthesized by ligament fibroblasts and meniscal fibrochondrocytes in vitro.

Osteonectin is an alpha-granule component involved with thrombospondin in platelet aggregation

We previously showed that thrombospondin, a major alpha-granule glycoprotein of human platelets, forms a specific complex with osteonectin, a phosphoglycoprotein originally described in bone that is also present in human platelets. The storage organelles and the function of osteonectin in platelets are still unknown. In this study, using electron microscopy in combination with immunogold staining, the major storage organelle for platelet-secreted proteins, the alpha-granules. Furthermore, osteonectin was qualitatively and quantitatively assessed by studying normal platelets and the platelets from a patient with gray platelet syndrome. Gray platelet syndrome is a rare congenital bleeding disorder characterized by a selective deficiency in morphologically recognizable platelet alpha-granules and in the alpha-granule secretory proteins. Binding of an iodinated antiosteonectin monoclonal antibody to gray platelet proteins transferred to nitrocellulose from SDS-polyacrylamide gels showed no band corresponding to osteonectin compared to control platelets. Using a polyclonal antiosteonectin antibody-based radioimmunoassay, gray platelets contained 0.2 +/- 0.03 ng osteonectin per 10(6) platelets, which is only 20% of the normal platelet content of osteonectin (0.93 +/- 0.16 ng per 10(6) platelets). Study of the localization of osteonectin to the surface of human platelets demonstrated that a radioiodinated antiosteonectin polyclonal antibody bound specifically to thrombin-stimulated platelets but not to resting platelets. Binding was concentration-dependent, saturable (1710 +/- 453 binding sites per platelet, Kd = 1 microM), and inhibited by an excess of cold antiosteonectin polyclonal antibody. No binding was observed on the surface of thrombin-stimulated gray platelets. To gain further insights into the role of osteonectin released from activated platelets, the effect of an antiosteonectin polyclonal antibody was tested on the aggregation of washed platelets. F(ab')2 fragments from the antiosteonectin polyclonal antibody inhibited in a dose-dependent manner the aggregation of collagen-stimulated, washed human platelets without affecting collagen-induced platelet serotonin release. To characterize the mechanism through which antiosteonectin F(ab')2 fragments inhibit platelet aggregation, the expression of endogenous thrombospondin (TSP) on the surface of thrombin-activated platelets was studied using 125I-labeled anti-TSP monoclonal antibody P10. The endogenous surface expression of TSP to thrombin-stimulated platelets was significantly inhibited in the presence of antiosteonectin F(ab')2 fragments (6286 +/- 2065 molecules of P10 per platelet) compared to 11,230 +/- 766 molecules of P10 per platelet in the presence of nonimmune F(ab')2 fragments.(ABSTRACT TRUNCATED AT 400 WORDS)

The concept of osseointegration and bone matrix expression

Osseointegration has been defined as the direct structural and functional connection between ordered, living bone and the surface of a load-carrying implant. To date, this concept has been described by descriptive histological and ultrastructural criteria but not by biochemical means. This review evaluates the basic science work performed on this concept and then applies the concept to the principle of osseous healing. Specific studies are cited where alterations in the healing response are due to clinical management of implant placement and how studies of surface properties may lead to further insights on implant design and prognosis. In addition, a review of bone expression as a function of in vitro stress applications is given. This is followed by an indepth review of the collagens and noncollagenous proteins, described to date, within isolated bone matrix. It is this collagenous matrix (especially type I) that is described as being close to and oriented with a glycoprotein component next to the implant surface. In turn, the large family of noncollagenous proteins are important in mediating bone proliferation, matrix accumulation, orientation, mineralization, and turnover. This section is followed by a discussion of specific growth factors as they may relate to osseous healing around an implant.