Stability and degradation of fibroblast growth factor 23 (FGF23): the effect of time and temperature and assay type

There is growing need for a reliable assay for measuring fibroblast growth factor 23 (FGF23), a regulator of phosphorus and vitamin D. In this work, we analyze and compare the performance of three available assays, including the effect of temperature and time. This knowledge will allow for better understanding of FGF23 in the future.

INTRODUCTION

Intact and C-terminal FGF23 (iFGF23 and cFGF23) concentrations are important in the diagnosis of hypo- and hyperphosphatemic diseases. The effects of temperature, storage, and specimen handling on FGF23 levels are not well known. We investigated the effects of various factors on plasma and serum measurement of FGF23 using three different assays.

METHODS

Serum and plasma FGF23 were measured using three commercially available ELISA assays-two measuring iFGF23 and one measuring cFGF23. Samples from subjects with known FGF23 disorders were stored at 4, 22, and 37 °C and analyzed at different intervals up to 48 hours (h). A subset of samples underwent repeated freeze-thaw cycles, and samples frozen at -80 °C for up to 60 months were reanalyzed. The effect of adding a furin convertase inhibitor on FGF23 degradation was investigated using samples stored at 37 °C for 48 h. Intact FGF23 levels were measured from plasma samples of four different groups to test the correlation of the two assays.

RESULTS

Plasma FGF23 levels were stable when stored at 4 and 22 °C for 48 h. Both plasma and serum FGF23 levels demonstrated relative stability after five freeze-thaw cycles. Long-term storage at -80 °C for 40 months induced some variability in FGF23 levels. The addition of a furin inhibitor did not affect FGF23 degradation. Intact FGF23 levels showed good correlation only at the upper limit of the assay range when comparing the two assays.

CONCLUSIONS

Sample type, handling, and choice of assay are factors that affect FGF23 levels and should be considered when measuring this hormone.

Serum levels of matrix extracellular phosphoglycoprotein (MEPE) in normal humans correlate with serum phosphorus, parathyroid hormone and bone mineral density

Matrix extracellular phosphoglycoprotein (MEPE), a member of the Small Integrin Binding Ligand N-linked Glycoprotein (SIBLING) family, is primarily expressed in normal bone and has been proposed as a phosphaturic factor because of high expression and secretion in oncogenic hypophosphatemic osteomalacia tumors. In order to begin to address the role of MEPE in normal human physiology, we developed a competitive ELISA to measure serum levels of MEPE. The ELISA was used to characterize the distribution pattern in a population consisting of 114 normal adult subjects. The mean value of MEPE was 476 +/- 247 ng/ml and levels decreased significantly with increasing age. MEPE levels were also significantly correlated with serum phosphorus and parathyroid hormone (PTH). In addition, MEPE levels correlated significantly with measures of bone mineral density in the femoral neck and total hip in a subset of 50 elderly subjects. The results are consistent with MEPE being involved in phosphate and bone metabolism in a normal population.

Elevated serum bone sialoprotein and osteopontin in colon, breast, prostate, and lung cancer

PURPOSE

Histological studies have shown that the two sialoproteins, bone sialoprotein (BSP) and osteopontin (OPN), are induced in multiple types of cancer. We have recently found that these proteins are bound in serum to complement factor H and that the complex must be disrupted to generate free protein to measure their total levels. We hypothesized that measuring total BSP and OPN levels would provide informative markers for the detection of cancer.

EXPERIMENTAL DESIGN

As a proof of concept study, serum from patients with diagnosed breast, colon, lung, or prostate cancer (n = 20 for each type) as well as normal serum (n = 77) were analyzed using competitive ELISAs developed for BSP and OPN. Sensitivity, specificity, as well as positive and negative predictive values were determined for each sialoprotein and cancer type. The relationship between sensitivity and specificity was profiled by receiver operating characteristic curves.

RESULTS AND CONCLUSIONS

Determined values for serum BSP in ng/ml were 285 +/- 19 for prostate, 373 +/- 19 for colon, 318 +/- 18 for breast, 155 +/- 11 for lung cancer sera, and 154 +/- 13 for normal sera. Values of OPN in ng/ml were 653 +/- 39 for prostate, 449 +/- 22 for colon, 814 +/- 53 for breast, 724 +/- 33 for lung, and 439 +/- 30 for normal sera. The assays provide a high degree of sensitivity and specificity that enables the detection of colon, breast, prostate, and lung cancer.

Flexible structures of SIBLING proteins, bone sialoprotein, and osteopontin

Bone sialoprotein (BSP) and osteopontin (OPN) are two members of the SIBLING (Small Integrin-Binding LIgand, N-linked Glycoprotein) family of genetically related proteins that are clustered on human chromosome 4. We present evidence that this entire family is the result of duplication and subsequent divergent evolution of a single ancient gene. The solution structures of these two post-translationally modified recombinant proteins were solved by one dimensional proton NMR and transverse relaxation times. The polypeptide backbones of both free BSP and OPN rapidly sample an ensemble of conformations consistent with them both being completely unstructured in solution. This flexibility appears to enable these relatively small glycoproteins to rapidly associate with a number of different binding partners including other proteins as well as the mineral phase of bones and teeth. These proteins often function by bridging two proteins of fixed structures into a biologically active complex.

Factor H binding to bone sialoprotein and osteopontin enables tumor cell evasion of complement-mediated attack

Metastatic cancer cells, like trophoblasts of the developing placenta, are invasive and must escape immune surveillance to survive. Complement has long been thought to play a significant role in the tumor surveillance mechanism. Bone sialoprotein (BSP) and osteopontin (OPN, ETA-1) are expressed by trophoblasts and are strongly up-regulated by many tumors. Indeed, BSP has been shown to be a positive indicator of the invasive potential of some tumors. In this report, we show that BSP and OPN form rapid and tight complexes with complement Factor H. Besides its key role in regulating complement-mediated cell lysis, Factor H also appears to play a role when "hijacked" by invading organisms in enabling cellular evasion of complement. We have investigated whether BSP and OPN may play a similar role in tumor cell complement evasion by testing to see whether these glycoproteins could promote tumor cell survival. Recombinant OPN and BSP can protect murine erythroleukemia cells from attack by human complement as well as human MCF-7 breast cancer cells and U-266 myeloma cells from attack by guinea pig complement. The mechanism of this gain of function by tumor cell expression of BSP or OPN has been defined using specific peptides and antibodies to block BSP and OPN protective activity. The expression of BSP and OPN in tumor cells provides a selective advantage for survival via initial binding to alpha(V)beta(3) integrin (both) or CD44 (OPN) on the cell surface, followed by sequestration of Factor H to the cell surface and inhibition of complement-mediated cell lysis.

Bone sialoprotein mediates human endothelial cell attachment and migration and promotes angiogenesis

Bone sialoprotein (BSP) is a secreted glycoprotein primarily found in sites of biomineralization. Recently, we demonstrated that BSP is strongly upregulated in osteotropic cancers and particularly those that exhibit microcalcifications. BSP contains an Arg-Gly-Asp (RGD) motif found in other adhesive molecules that interact with cellular integrins. In bone, BSP has been shown to mediate the attachment of osteoblasts and osteoclasts via alpha(v)beta(3) integrin receptors. Ligands for alpha(v)beta(3) integrin are considered to play a central role during angiogenesis. Therefore, we used human umbilical vein endothelial cells (HUVECs) to study the potential role of BSP in angiogenesis. We found that purified eukaryotic recombinant human BSP (rhBSP) is able to promote both adhesion and chemotactic migration of HUVECs in a dose-dependent manner. These interactions involve HUVEC alpha(v)beta(3) integrin receptors and the RGD domain of BSP. Indeed, HUVECs attach to a recombinant BSP fragment containing the RGD domain, whereas this response is not observed with the same fragment in which RGD has been mutated to Lys-Ala-Glu (KAE). A cyclic RGD BSP peptide inhibits both adhesion and migration of HUVECs to rhBSP. Moreover, anti-alpha(v)beta(3) but not anti-alpha(v)beta(5) monoclonal antibodies also prevent BSP-mediated adhesion and migration of HUVECs. We observed that both rhBSP and the RGD BSP recombinant fragment stimulated ongoing angiogenesis on the chorioallantoic chick membrane assay. BSP angiogenic activity was inhibited by anti-alpha(v)beta(3) antibody, and the KAE BSP fragment was inactive. Our findings represent the first report implicating BSP in angiogenesis. BSP could play a critical role in angiogenesis associated with bone formation and with tumor growth and metastatic dissemination.

Interleukin-13 modulates collagen homeostasis in human skin and keloid fibroblasts

Interleukin (IL)-13 has been implicated in the pathogenesis of various diseases characterized by fibrosis. We describe the effects of IL-13 on collagen homeostasis from normal (NF) and keloid (KF) fibroblasts and compare these effects with those of IL-4 and transforming growth factor (TGF)-beta(1). Total collagen generation was up-regulated in NF after 48 h of stimulation by IL-13; in KF, IL-13 stimulated a more rapid collagen response. The kinetics and magnitude of collagen generation induced by IL-13 were equivalent to those induced by similar concentrations of IL-4 and TGF-beta(1). Collagen type I production paralleled total collagen generation from both NF and KF; however, IL-4-induced collagen type I and total collagen production from KF was more transient than that induced by either IL-13 or TGF-beta(1). Procollagen 1alpha1 gene expression was induced in KF by stimulation with IL-13 for 24 h. Moreover, IL-13 was unique among these three cytokines in its ability to induce gene expression for procollagen 3alpha1. Finally, IL-13 inhibited IL-1beta-induced matrix metalloproteinase (MMP)-1 and MMP-3 production and enhanced tissue inhibitor of metalloproteinase (TIMP)-1 generation from NF; although similar effects were observed with IL-4, TGF-beta(1) transiently enhanced MMP-1 and MMP-3 generation without effecting TIMP-1. In KF, IL-13 and IL-4 inhibited MMP-3, whereas TGF-beta(1) enhanced MMP-3; TIMP-1 was unaffected by any of the three cytokines. These data demonstrate both the profibrotic effects of IL-13 on collagen homeostasis and the potential differential regulation of collagen homeostasis in fibroblast subtypes by IL-13.

Receptor tyrosine kinase expression in human bone marrow stromal cells

Bone marrow stromal cells (BMSCs) are a heterogeneous population of cells derived from colony-forming units-fibroblastic (CFU-Fs). These cells reside in the bone marrow cavity and are capable of differentiating into several cell phenotypes including osteoblasts, chondroblasts, hematopoiesis-supporting stromal cells, and adipocytes. However, the factors that regulate the proliferation and differentiation of the BMSC population are for the most part unknown. Since many members of the receptor tyrosine kinase (RTK) family have been shown to participate in growth control of various mesenchymal cell populations, in this study we examined the expression and function of RTKs in the BMSC population. Degenerate oligonucleotides corresponding to two conserved catalytic domains of the RTK family and RT-PCR were used initially to determine which RTKs are expressed in the human BMSC (hBMSC) system. After subcloning the amplification product generated from mRNA of a multicolony-derived hBMSC strain, PDGF receptor (beta), EGF receptor, FGF receptor 1, and Axl were identified by DNA sequencing of 26 bacterial colonies. Furthermore, PDGF and EGF were found to enhance BMSC growth in a dose-dependent manner and to induce tyrosine phosphorylation of intracellular molecules, including the PDGF and EGF receptors themselves, demonstrating the functionality of these receptors. On the other hand, bFGF was found to have little effect on proliferation or tyrosine phosphorylation. Since single colony-derived hBMSC strains are known to vary from one colony to another in colony habit (growth rate and colony structure) and the ability to form bone in vivo, the expression levels of these RTKs were determined in 18 hBMSC clonal strains by semiquantitative RT-PCR and were found to vary from one clonal strain to another. While not absolutely predictive of the osteogenic capacity of individual clonal strains, on average, relatively high levels of PDGF-receptor were found in bone-forming strains, while on average, nonbone-forming strains had relatively high levels of EGF-receptor. Taken together, these results indicate that RTKs play a role in the control of hBMSC proliferation, and that the differential pattern of RTK expression may be useful in correlating the biochemical properties of individual clonal strains with their ability to produce bone in vivo.

Age-related changes in human oestrogen receptor alpha function and levels in osteoblasts

Oestrogen receptors (ERs) are present in human osteoblasts and mediate anti-resorptive effects on bone. Human osteoblast-like cells derived from different aged healthy female donors not on hormone replacement therapy were utilized under well-defined conditions in vitro to investigate ER function and levels. Treatment with 0.1 nM oestradiol-17beta of cell strains derived from eight young women (less than 50 years of age) increased hydroxyproline levels significantly [an average (2.2+/-0.1 S.E.M.)-fold increase], whereas cells derived from nine older women (more than 50 years of age) were not significantly affected. Similarly, cell strains, derived from younger women, transfected with a consensus oestrogen-responsive element linked to chloramphenicol acetyltransferase exhibited a greater response to oestrogen than strains derived from older women. When basal ERalpha levels were measured by enzyme immunoassay and normalized on a per cell basis, osteoblast-like strains derived from younger women (n=24) had a mean value of 2.54+/-0.16 fmol of ERalpha per 10(6) cells. In contrast, strains derived from older women (n=20) had a mean value of 5.44+/-0.48 fmol of ERalpha per 10(6) cells. An age-related increase in ERalpha number was also observed in human skin-derived fibroblasts and directly in dermal biopsies from women not on hormone replacement therapy. The results demonstrate ligand concentration-dependent ERalpha induction and indicate a loss of receptor regulation and diminution of ligand-receptor signal transduction with increasing donor age.

Immortalization and characterization of bone marrow stromal fibroblasts from a patient with a loss of function mutation in the estrogen receptor-alpha gene

A male patient with abnormal postpubertal bone elongation was shown earlier to have a mutation in both alleles of the estrogen receptor, resulting in a nonfunctional gene. Marrow stromal fibroblasts (MSFs) derived from this patient were called HERKOs (human estrogen receptor knock outs), and in order to obtain continuous HERKO cell lines, they were immortalized using a recombinant adenovirus-origin-minus SV40 virus. MSFs are unique cells because they support hematopoesis and contain a mixed population of precursor cells for bone, cartilage, and fat. Three established cell lines (HERKO2, HERKO4, and HERKO7) were characterized and compared with the heterogeneous population of nonimmortalized HERKOs for their osteogenic potential. We performed Northern analysis of matrix genes implicated in bone development and metabolism and an in vivo bone formation assay by transplanting the cells subcutaneously into immunodeficient mice. All three HERKO lines expressed high amounts of collagen 1A1, osteopontin, osteonectin, fibronectin, decorin, biglycan, and alkaline phosphatase. Except for osteopontin, expression of these genes was slightly lower compared with nonimmortalized HERKOs. In the in vivo bone formation assay, the heterogeneous population of nonimmortalized HERKOs formed bone with high efficiency, while the HERKO lines induced a high-density, bone-like matrix. Finally, all HERKO cell types secreted high levels of insulin-like growth factor I and interleukin-6 into the culture medium relative to cells of normal human subjects. In summary, these lines of HERKO cells retain several of the phenotypic traits of MSFs after immortalization, including matrix and cytokine production, and provide a valuable source of a unique human material for future studies involving estrogen action in bone and bone marrow metabolism.

Age-related changes in effects of insulin-like growth factor I on human osteoblast-like cells

The role of insulin-like growth factor I (IGF-I) in extracellular matrix metabolism was studied in both proliferating and confluent human osteoblast-like cultures derived from donors of different ages. In proliferating cultures, recombinant human (rh)IGF-I was found to increase the incorporation of [3H]thymidine in a dose- and age-dependent manner. To study cell proliferation dynamically, continuous growth curves with and without rhIGF-I were modelled by a modified logistic function. Increasing doses of rhIGF-I decreased the lag time and maximal growth rates, whereas plateau values decreased only at the highest dose (100 ng/ml). In post-proliferative cell strains, rhIGF-I (0.1-100 ng/ml) increased levels of type I collagen, biglycan and decorin, and to a smaller extent fibronectin and thrombospondin, whereas it decreased the levels of hyaluronan and a versican-like proteoglycan when protein and proteoglycan metabolism were followed by steady-state radiolabelling with [3H]proline, [3H]glucosamine or [35S]sulphate. These responses to rhIGF-I were found to be age-dependent, with osteoblast-like cells derived from younger patients being more responsive to rhIGF-I. When extracellular matrix turnover was analysed by pulse-chase experiments, rhIGF-I had no effect. The steady-state levels of collagen, decorin, hyaluronan and a versican-like proteoglycan for bone cells treated with rhIGF-I on day 7 in culture were equivalent to levels of these matrix components in untreated osteoblasts grown for 14 days. These results are consistent with rhIGF-I's altering cellular proliferative capacity and matrix synthesis, causing a change in the osteoblast differentiated state.

2-Sulfotrehalose, a novel osmolyte in haloalkaliphilic archaea

A novel 1-->1 alpha-linked glucose disaccharide with sulfate at C-2 of one of the glucose moieties, 1-(2-O-sulfo-alpha-D-glucopyranosyl)-alpha-D-glycopyranose, was found to be the major organic solute accumulated by a Natronococcus sp. and several Natronobacterium species. The concentration of this novel disaccharide, termed sulfotrehalose, increased with increasing concentrations of external NaCl, behavior consistent with its identity as an osmolyte. A variety of noncharged disaccharides (trehalose, sucrose, cellobiose, and maltose) were added to the growth medium to see if they could suppress synthesis and accumulation of sulfotrehalose. Sucrose was the most effective in suppressing biosynthesis and accumulation of sulfotrehalose, with levels as low as 0.1 mM being able to significantly replace the novel charged osmolyte. Other common osmolytes (glycine betaine, glutamate, and proline) were not accumulated or used for osmotic balance in place of the sulfotrehalose by the halophilic archaeons.

Long-term extracellular matrix metabolism by cultured human osteogenesis imperfecta osteoblasts

Osteopenia due to deficient extracellular matrix synthesis is a hallmark of osteogenesis imperfecta (OI), Previous studies carried out within 72 h of osteoblast subculture, at an early stage of matrix synthesis, indicated that for osteoblasts derived from human OI patients the total amounts of collagen, osteonectin, and three proteoglycans were significantly reduced, while total amounts of thrombospondin, fibronectin, and matrix hyaluronan were elevated compared with age-matched controls. The current study was undertaken to follow OI osteoblast matrix metabolism as that matrix is synthesized, deposited, and matured. Steady-state metabolic radiolabeling was used to follow the metabolism of collagen, hyaluronan, and total proteoglycan by OI and normal osteoblasts for up to 5 weeks. Trabecular osteoblasts from non-OI controls showed an increase in total and matrix-associated collagen synthesis during the first and second week, reaching steady-state levels by week 4. In contrast, cultured OI osteoblasts did not increase either the total (medium + matrix-associated) or matrix-associated collagen during the entire 5-week period. Proteoglycan synthesis exhibited a pattern similar to that for collagen. OI-derived proteoglycans differed from controls in that levels in OI cultures did not reflect the normal time-dependent increase in total proteoglycan and proteoglycan matrix deposition. OI osteoblast hyaluronan synthesis was increased when compared with age-matched controls during 4 weeks of culture. In contrast, the ratios of calcium to phosphorus solublized from control and the OI extracellular matrix were not statistically different. Thus, with respect to the synthesis of collagen, proteoglycans, and hyaluronan, OI osteoblasts fail to parallel controls in depositing and elaborating extracellular matrix during 35 days in culture.

Cell proliferation of human fibroblasts and osteoblasts in osteogenesis imperfecta: influence of age

Clinical studies indicate that as a group, osteogenesis imperfecta (OI) subjects are shorter than age- and sex-matched controls. Not only somatic growth, but also cellular growth appears to be impaired, and these may be related to defects in extracellular matrix common to this disorder. We have investigated the growth characteristics of dermal fibroblasts and trabecular osteoblasts isolated from patients with OI and control subjects of various ages. Cell growth curves and cell doubling times were determined by measuring cell number using crystal violet dye binding. Growth curves were modeled by a modified logistic function, the three parameters of which are markers for biologically relevant growth parameters: the plateau value or upper asymptote, which reflects the maximum cell density upon confluence; the maximal growth rate (microM); and the lag time. Both normal human fibroblasts and osteoblasts showed an age-dependent decrease in microM. Normal fibroblasts exhibited no age-dependence to their upper asymptote or lag time. Fibroblasts derived from patients with OI did not have significantly different upper asymptote values microM, or lag times when compared with normal fibroblasts. Normal osteoblasts had a decrease in upper asymptote, decrease in microM, but a relatively constant lag time with increasing age. In contrast, OI osteoblast microM was decreased relative to that of normal subjects. For osteoblasts from OI patients, decreased microM appeared unrelated to the age of the subject, whereas OI fibroblasts did exhibit an age-dependent decrease in microM. The percentage of collagenase-digestible protein (a measure of collagen synthesis) produced by normal human fibroblasts correlated well with microM. Treating normal human osteoblasts with the proline analogue 3,4-dehydroproline, which destabilizes collagen triple helix formation and alters collagen synthesis, secretion, and turnover, also decreased microM. A dose response to varying concentrations of 3,4-dehydroproline was observed for normal human bone cell microM. These data suggest a link between type I collagen synthesis and cellular proliferation.

Extracellular matrix stoichiometry in osteoblasts from patients with osteogenesis imperfecta

In previous work, we compared the steady-state levels of specific matrix components in human bone cells derived from patients with osteogenesis imperfecta (OI) to those of age-matched controls. A remarkable finding was the observation that there was a reduction not only in the total levels of collagen, but also in osteonectin and three proteoglycans (a large chondroitin sulfate proteoglycan, biglycan, and decorin). This pattern was observed in patients with and without detectable collagen defects. More recent analysis of extracellular matrix composition have yielded that, compared with age-matched controls, bone cells from OI patients produced higher steady-state levels of fibronectin and thrombospondin. The percentage of these two proteins incorporated into the cell layer pool was also higher in OI than in age-matched controls. In addition, the steady-state levels of hyaluronan and a heparan sulfate proteoglycan were analyzed in both OI and age-matched controls. Although the total (medium + cell layer) steady-state levels of hyaluronan were reduced by 1/3, the percentage of the hyaluronan in the cell layer pool of patients with OI increased between 100-250% of age-matched control. Thus the matrix elaborated by human OI bone cells is not only quantitatively different but also qualitatively distinct from that of age-matched controls. Not only have specific bone cell matrix components (collagen, osteonectin, the large chondroitin sulfate proteoglycan, biglycan, and decorin) been found to be present in reduced levels in OI bone cells, but some matrix components (thrombospondin, fibronectin, and hyaluronan) have also been found to be present in elevated levels in the matrix of OI cells.(ABSTRACT TRUNCATED AT 250 WORDS)

The X-chromosomal human biglycan gene BGN is subject to X inactivation but is transcribed like an X-Y homologous gene

We report the mRNA and protein expression levels of human biglycan (BGN) in patients with different numbers of sex chromosomes. BGN maps to the distal long arm of the X chromosome, band Xq28, near the second pseudoautosomal region. BGN expression levels are reduced in 45,X Turner patients and increased in patients with additional sex chromosomes. This is suggestive of a pseudoautosomal gene or a gene that escapes X inactivation and that has an active Y chromosomal copy. However, we also provide evidence from hybrid cell lines that BGN is subject to X inactivation and that there is no homolog on the Y chromosome. This evidence excludes an escape from X inactivation. Moreover, additional Y chromosomes increase BGN expression levels, despite the absence of a Y chromosomal BGN gene. Therefore, another explanation has to be invoked. The "pseudoautosomal expression" of BGN may be attributed to a gene or genes that escape X inactivation and that regulate the transcriptional activity of BGN. This is the first report concerning an X chromosomal gene that does not show the conventional correlation between gene dosage and expression rate known from other X chromosomal genes.

The bone cell biology of osteogenesis imperfecta

Osteogenesis Imperfecta (OI) has been defined as a heritable connective tissue disorder with variable severity of clinical expression. OI is a type I collagen based disease. Consequently, much research has focused on identifying specific mutations in the pro-alpha (I) genes. Our interest in OI lies in the metabolism of the non-collagenous proteins (NCPs) of the bone matrix. Although type I collagen is the most abundant protein in bone extracellular matrix, it is the NCPs which bind to, modify and have the potential to regulate that collagen matrix. Our approach has been to determine the levels of the NCPs for both OI and age-matched controls. Most recently, we have utilized an in vitro human osteoblast system to study normal and OI NCP metabolism (Fedarko et al. J. Bone Min. Res. 7, 921-930, 1992). It is our hypothesis that the altered stoichiometry of collagen and NCPs is, in part, responsible for the phenotypic variation of the disease.

OIM and related animal models of osteogenesis imperfecta

Osteogenesis imperfecta (OI) is characterized by fragile bones, skeletal deformity, and growth retardation. This heritable disorder of connective tissue is the result of mutations affecting the COL1A1 and COL1A2 genes of type I collagen. Progress in OI research has been limited because of dependence on human fibroblast and osteoblast specimens and the absence of a naturally occurring animal model for this genetic disorder. Recent technology in molecular biology has led to the development of transgenic models of OI based on site directed mutagenesis of type I collagen genes. OIM is a naturally occurring model which incorporates both the phenotypic and biochemical defects of moderate to severe osteogenesis imperfecta. This powerful tool permits the development of models based on different type I collagen mutations. The collagen type I mutation in OIM is a C propeptide deletion which impairs the production of normal pro-alpha2(I). Tissues in OIM contain only [pro-alpha1(I)]3 homotrimer. Thus, although several animal models are now available for research in osteogenesis imperfecta few are viable or fully mimic human disease disorders. OIM duplicates the phenotype and biochemistry of human disease and has a normal life span.

Isolation and purification of proteoglycans

Purification of a protein typically involves development of a quantitative assay to track protein integrity (e.g. enzyme activity) during subsequent isolation steps. The generalized procedure involves choosing the source of the protein, defining extraction conditions, developing bulk purification methods followed by refined, more selective methods. The purification of proteoglycans is often complicated by a) limited source quantities, b) necessity of chaotrophic solvents for efficient extraction, c) their large molecular size and d) lack of defined functions to enable purity (i.e. activity, conformation) to be assessed. Because the usual goal of proteoglycan purification is physical characterization (intact molecular weight, core protein and glycosaminoglycan class and size), the problems of a suitable assay and/or native conformation are avoided. The 'assay' for tracking proteoglycan isolation typically utilizes uronic acid content or radiolabel incorporation as a marker. Once extracted from their cellular/extracellular environment, proteoglycans can be isolated by density gradient centrifugation and/or column chromatography techniques. Recent advances in the composition of chromatographic supports have enabled the application of ion-exchange, gel permeation, hydrophobic interaction and affinity chromatography resins using efficient high-pressure liquid chromatography to proteoglycan purification.

Structure and molecular regulation of bone matrix proteins

The organic matrix of bone contains several protein families, including collagens, proteoglycans, and glycoproteins, all of which may be extensively modified by posttranslational events, such as phosphorylation and sulfation. Many of the glycoproteins contain Arg-Gly-Asp (RGD), the integrin-binding sequence, within their structure, whereas other constituent proteins contain gamma-carboxyglutamic acid. The deposition of bone matrix by cells in the osteoblastic lineage is regulated by extrinsic factors, such as systemic and local growth factors and physical forces, and factors that are intrinsic to the cell, such as position in the cell cycle, maturational stage, and developmental age of the donor. Recent studies of several bone matrix gene promoters have identified cis- and trans-acting elements that are responsible for gene activity, although the precise sequence of regulatory events is not known. Development of in vitro assays, coupled with studies of the appearance of these proteins during development in vivo, provides insight into the functions of these proteins during the various stages of bone metabolism. Potential roles for these proteins include proliferation and maturation of stem cells, formation of matrix scaffolding elaborated by bone-forming cells, modeling, and remodeling. Changes in the functional properties of the extracellular matrix may be involved in a variety of disease processes, including osteoporosis and oral bone loss.

Isolation and purification of proteoglycans

Purification of a protein typically involves development of a quantitative assay to track protein integrity (e.g. enzyme activity) during subsequent isolation steps. The generalized procedure involves choosing the source of the protein, defining extraction conditions, developing bulk purification methods followed by refined, more selective methods. The purification of proteoglycans is often complicated by a) limited source quantities, b) necessity of chaotropic solvents for efficient extraction, c) their large molecular size and d) lack of defined functions to enable purity (i.e. activity, conformation) to be assessed. Because the usual goal of proteoglycan purification is physical characterization (intact molecular weight, core protein and glycosaminoglycan class and size), the problems of a suitable assay and/or native conformation are avoided. The 'assay' for tracking proteoglycan isolation typically utilizes uronic acid content or radiolabel incorporation as a marker. Once extracted from their cellular/extracellular environment, proteoglycans can be isolated by density gradient centrifugation and/or column chromatography techniques. Recent advances in the composition of chromatographic supports have enabled the application of ion-exchange, gel permeation, hydrophobic interaction and affinity chromatography resins using efficient high-pressure liquid chromatography to proteoglycan purification.

Extracellular matrix formation by osteoblasts from patients with osteogenesis imperfecta

Extracellular matrix proteins synthesized by bone cells isolated from 16 patients with different forms of osteogenesis imperfecta (OI) were analyzed in vitro. Specific components of the extracellular matrix by OI and age-matched cultures were investigated by steady-state radiolabeling followed by quantitation of label into specific proteins and comparison of OI cultures to those of age-matched controls. The in vitro proliferation of OI bone cells was found to be lower than that of control cells. In seven patients, abnormalities of the alpha 1(I) and/or alpha 2(I) chains of type I collagen were detected by gel electrophoresis. In two of these patients, the mutations in the COLIA1 and COLIA2 genes have been previously identified. Although the amount of total protein synthesized by the cells in culture was the same for OI bone cells and age-matched control cells, OI bone cells showed a significantly reduced synthesis of not only collagen but also other bone matrix glycoproteins. The synthesis of osteonectin (SPARC/BM40) and three proteoglycans [a large chondroitin sulfate proteoglycan, biglycan (PGI), and decorin (PGII)] was found to be decreased in OI cells. The reduction was most pronounced at the developmental age at which these macromolecules reach maximal levels during normal development.

Age-related changes in hyaluronan, proteoglycan, collagen, and osteonectin synthesis by human bone cells

Human bone cells grown in culture, representative of a preosteoblastic stage of maturation, produce an extracellular matrix composed of collagen, several noncollagenous glycoproteins, hyaluronan, and four distinct proteoglycans (PGs). The influence of donor age on the levels of expression of these molecules in vitro has not been well characterized. In this study, human bone cells derived from sources ranging from fetal to 60-year-old donors were grown in culture, radiolabeled for 24 h, and the amount of incorporation of [35S]sulfate into PGs, [3H]glucosamine into hyaluronan, [3H]leucine/proline into osteonectin, and [3H]proline into collagen was determined. Cell proliferation was most rapid in fetal-derived bone cells and decreased with increasing age. Total protein and PG synthesis also decreased with increasing age, falling to 1/3 and 1/4, respectively, of fetal levels after age 30. A large chondroitin sulfate PG (Mr approximately 600,000 Da) was the major fetal PG and its levels were highly correlated with cellular proliferation. [3H]Collagen and [35S]decorin levels increased with the increasing age of the donor, reached a maximum in puberty-derived cells, and decreased to 1/3 maximal levels after age 20. The heparan sulfate PG (Mr approximately 400,000 Da) exhibited steady-state levels regardless of donor age. [3H]Osteonectin and [35S]biglycan levels were high in fetal-derived cells and in cells derived from pubescent donors. The percentage of collagen and four proteoglycans associated with the cell layer pool changed with donor age. All fetal-derived PG core proteins possessed more N- and O-linked oligosaccharides than newborn or adult derived PGs.

High-performance liquid chromatographic separation of hyaluronan and four proteoglycans produced by human bone cell cultures

Four proteoglycans and hyaluronan synthesized by cultured human bone cells were isolated using a two-step high-performance liquid chromatography system involving desalting and buffer exchange with a TSK-GEL HW 40(S) column followed by ion-exchange separation on a Nucleogen 4000-10 DEAE column. The desalting of 4 M guanidinium HCl extracts by a TSK-GEL HW 40(S) column equilibrated in a formamide:KH2PO4 buffer produces greater than 95% recoveries, enables quantitation of label incorporation and requires only 40 min to complete. The Nucleogen 4000-10 DEAE column utilizes the same buffer system and requires only 100 min for the resolution of four distinct types of proteoglycans. The formamide:KH2PO4 buffer system is compatible with a previously developed polyacrylamide gel system for the electrophoretic profiling of proteoglycans. After separation by charge density, proteoglycans were further resolved by size distribution using a calibrated TSK-GEL HW 75(F) column which also enabled the estimation of the apparent Mr of hyaluronan produced by the bone cells. The same TSK-GEL HW 40(S) resin is used to exchange pooled proteoglycans into buffers for analyzing enzyme digests of glycosaminoglycan chains and core proteins. The technique has been applied to the analysis of biosynthetically labeled proteoglycans produced in culture by fetal and adult human bone cells. A distinct pattern of proteoglycan size and secretion for both cell types could be shown using this method. The method of analysis is useful for high yield and rapid screening of various cell types for both biosynthetic rate studies and analysis of patterns of proteoglycan synthesis.

Temporal regulation of hyaluronan and proteoglycan metabolism by human bone cells in vitro

Osteoblasts elaborate a dynamic extracellular matrix that is constructed and mineralized as bone is formed. This matrix is primarily composed of collagen, along with noncollagenous proteins which include glycoproteins and proteoglycans. After various times in culture, human bone cells were labeled with [35S]sulfate, [3H] leucine/proline, or [3H]glucosamine and the metabolism of hyaluronan and four distinct species of proteoglycans (PGs) was assayed in the medium, cell layer, and intracellular pools. These cells produce hyaluronan (Mr approximately 1,400,000; a chondroitin sulfate PG (CSPG), Mr approximately 600,000; a heparan sulfate PG (HSPG), Mr approximately 400,000; and two dermatan sulfate PGs with Mr approximately 270,000 (biglycan, PG I) and Mr approximately 135,000 (decorin, PG II) that distribute between the medium and cell layer. Two days following subculture, 12 h [35S]sulfate steady-state labeling yielded a composition of 24, 27, 31, and 18% for total CSPG, HSPG, biglycan, and decorin, respectively. While HSPG and decorin levels and distribution between medium and cell layer remained relatively constant during steady-state labeling at different times in culture, CSPG and biglycan levels increased dramatically at late stages of growth, and their distribution changed throughout culture. These results were independent of cell density, media depletion, and labeling pool effects. In contrast, hyaluronan synthesis was uncoupled from PG synthesis and apparently density-dependent. Pulse chase labeling at different stages of culture showed that the CSPG and decorin behaved as secretory PGs. Both HSPG and biglycan underwent catabolism, with HSPG possessing a t1/2 of 8 h and biglycan a t1/2 of 4 h. While the rate of HSPG turnover did not appreciably change between early and late culture, that of biglycan decreased. The mRNA for decorin was constant, while that of biglycan changed during culture. These results suggest that each PG possesses a distinct pattern of cellular and temporal distribution that may reflect specific stages in matrix formation and maturation.

Human bone cell enzyme expression and cellular heterogeneity: correlation of alkaline phosphatase enzyme activity with cell cycle

Alkaline phosphatase, long implicated in biomineralization, is a feature of the osteoblast phenotype. Yet in cultured bone cells, only a fraction stain positive histochemically. To determine whether osteoblast enzyme expression reflects cellular heterogeneity with respect to cell cycle distribution or length of time in culture, the activities of alkaline phosphatase, tartrate-resistant and -sensitive acid phosphatases, and non-specific esterases were assayed kinetically and histochemically. In asynchronous subconfluent cultures, less than 15% of the cells stained positive and assayed activity was 0.04 IU/10(6) cells/cm2. After 1 week, the percent of alkaline phosphatase positive-staining cells increased 5-fold, while activity increased 10-fold. Non-specific esterases and tartrate-sensitive acid phosphatase were constitutive throughout time in culture, whereas tartrate-resistant acid phosphatase activity appeared after 2 weeks. Cell cycle analysis of human bone cells revealed a growth fraction of 80%, an S phase of 8.5 h, G2 + 1/2 M of 4 h, and a G1 of 25-30 h. In synchronous cultures induced by a thymidine-aphidicolin protocol, alkaline phosphatase activity dropped precipitously at M phase and returned during G1. A majority of the alkaline phosphatase activity lost from the cell surface at mitosis was recovered in the medium. Tartrate-sensitive acid phosphatase and non-specific esterase levels were relatively stable throughout the cell cycle, while tartrate-resistant acid phosphatase activity was not assayable at the density used in synchronous cultures. From these data, variations in alkaline phosphatase activity appear to reflect the distribution of cells throughout the cell cycle.

Control of cell division in hepatoma cells by exogenous heparan sulfate proteoglycan

The effects of cell surface heparan sulfate proteoglycan (HSPG) prepared from log and confluent monolayers of a rat hepatoma cell line on hepatoma cell growth were studied. When HSPG isolated from confluent cells was added exogenously to log phase cells, it was internalized and free heparan sulfate (HS) chains appeared transiently in the nucleus. Concurrently, the growth of the treated cells was inhibited, but the cells resumed logarithmic growth as the level of nuclear HS fell, and the cells grew to confluence and became contact inhibited. When HSPG prepared from log-phase hepatoma cells was added exogenously to log phase cells, it was internalized but very little of the internalized HS appeared in the nucleus, and there was no change in the rate of cell growth. However, when the rate of cell growth was reduced by culture of the cells in serum- and insulin-deficient medium, HSPG prepared from log-phase cells stimulated the growth rate of these slow-growing cells. The cell cycle dependency of HSPG uptake and growth inhibition was studied in cultures synchronized by a thymidine/aphidicolin double block. When [35SO4]HSPG from confluent cells was added to synchronized cells just as they were released from the second block, a portion of the [35SO4]HSPG was internalized and [35SO4]HS appeared in the nucleus. However, at mitosis the [35SO4]HS disappeared almost completely from all of the cellular pools, and after mitosis, more of the [35SO4]HSPG was taken up and [35SO4]HS reappeared in the nucleus and remained in the nucleus until the cells divided again. When cultures were released from the aphidicolin block, both control and HSPG-treated cells progressed through the S, the G2, and the M phases of the cell cycle. However, the length of the G1 phase of the cycle was increased in the HSPG-treated cells. The treated cultures then progressed through the second S, G2, and M phases. Thus, the inhibition of cell division occurred in the G1 phase of the cell cycle, prior to the G1/S boundary. Addition of the HSPG to the synchronized cultures just after the first mitosis resulted in an immediate arrest of the cell cycle in G1.(ABSTRACT TRUNCATED AT 400 WORDS)

Analysis of the proteoglycans synthesized by human bone cells in vitro

Proteoglycans were isolated by ion-exchange chromatography from the extracted cell layer and culture medium of human bone cell cultures following incubation in the presence of [35S]sulfate and [3H]leucine. On sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), the synthesized proteoglycans consisted of at least five polydisperse species having median apparent Mr = 600,000, 400,000, 270,000, 135,000 and 40,000. When chromatographed further on octyl-Sepharose CL-4B, the proteoglycans of the cell layer resolved into three peaks. The unbound fraction (peak cell layer-I) contained a 40,000 species consisting of a single glycosaminoglycan chain with or without peptide. Peak cell layer-II contained three sulfated species on electrophoresis: a 600,000 species uniformly distributed across the peak, a 135,000 species enriched in the ascending limb (similar to bone PG-I as described previously), and a 270,000 species (similar to bone PG-I) enriched in the descending limb. Peak cell layer-III, eluting at 0.2% Triton X-100, was highly enriched in a 400,000 proteoglycan component. When media proteoglycans were chromatographed on octyl-Sepharose, two labeled peaks were found. Peak medium-I (unbound) contained a species exhibiting electrophoretic mobility similar to that of the 400,000 species present in peak cell layer-III. Peak II of the culture medium (medium-II) was apparently identical to that of peak cell layer-II, containing the 600,000, 270,000 and 135,000 species. No appreciable 40,000 species was observed in the medium. Treatment of the 600,000 species with either chondroitinase ABC or ACII generated a core protein preparation with bands of 390,000 and 340,000 on SDS gels. Neither the intact nor the chondroitinase ABC-treated 600,000 species was immunoprecipitated by a purified, polyclonal antiserum raised against the core protein of the large chondroitin sulfate proteoglycan of human articular cartilage. Treatment of the 270,000 and 135,000 proteoglycans with chondroitinase ABC, but not ACII, generated a core protein preparation with bands of 52,000 and 49,000 on SDS gels, indicating that they were dermatan sulfate-containing species. The 400,000 species contained both heparan sulfate and chondroitin sulfate, in approximately a 3:1 labeling ratio. This species changed in electrophoretic mobility following treatment with chondroitinase ABC, heparatinase, or both enzymes in combination, which suggested that it may be a hybrid proteoglycan (i.e. both types of glycosaminoglycan chain on the same core protein).(ABSTRACT TRUNCATED AT 400 WORDS)

Involvement of phosphatidylinositol and insulin in the coordinate regulation of proteoheparan sulfate metabolism and hepatocyte growth

A rat hepatocyte cell line was cultured in Higuchi's medium with fetal calf serum and insulin and labeled with 35SO2/4-. The cells were treated with a number of ligands to displace the heparan 35SO4 proteoglycan (HSPG) from the pericellular matrix. Maximum release was obtained with D-mannose-6-PO4 (50 mM), D-glucose-6-PO4 (50 mM), myo-inositol-2-PO4 (2-5 mM), myo-inositol hexaphosphate (2-5 mM), and DL-myo-inositol-1-PO4 (1-2 mM). D-myo-Inositol-1,3,4-(PO4)3 (1 mM) and L-myo-inositol-1-PO4 (2 mM) were intermediate in their ability to release the cell surface HSPG, whereas heparin (2 mg/ml), yeast phosphomannan (4 mg/ml), D-xylose-1-PO4 (50 mM), D-glucose-6-SO4 (50 mM), and myo-inositol hexasulfate (5 mM) were ineffective. When 35SO2/4- was added to cell cultures, the total cell surface HSPG increased linearly, but the percentage of the total cell surface [35SO4]HSPG that was released by myo-inositol-PO4 increased with time during the labeling period, reaching a maximum of 65% after 5 h. When cells were labeled for 12 h without insulin in the medium, the maximum amount of cell surface HSPG that was released by myo-inositol-PO4 was reduced to 30%. However, when cells labeled in the absence of insulin were treated with phosphatidylinositol-specific phospholipase C and then myo-inositol-PO4, the release of the cell surface [35SO4]HSPG was increased to 73%. When the [35SO4]HSPG that was released from the cell surface by treatment with myo-inositol-PO4 was added to cultures of unlabeled hepatocytes, it was taken up very rapidly and a portion of the internalized HSPG was converted to free heparan SO4 chains which appeared in the nucleus. Uptake was Ca2+- and Mg2+-independent. The amount of [35SO4]HSPG taken up was markedly reduced when the myo-inositol-PO4-releasable [35SO4]HSPG was pretreated with trypsin, thermolysin, alkaline borohydride, or alkaline phosphatase. When the cells were grown in inositol-deficient medium or in the presence of myo-inositol-PO4, the amount of heparan SO4 found in the nucleus was markedly reduced, and the cells no longer exhibited contact inhibition. These effects of myo-inositol deficiency on the growth and nuclear heparan SO4 were accentuated by addition of LiCl to the cultures to prevent phosphatidylinositol synthesis from the endogenous myo-inositol-PO4.(ABSTRACT TRUNCATED AT 400 WORDS)

Transport of heparan sulfate into the nuclei of hepatocytes

Monolayer cultures of a rat hepatocyte cell line shown previously to accumulate a nuclear pool of free heparan sulfate chains that are enriched in sulfated glucuronic acid (GlcA) residues (Fedarko, N.S., and Conrad, H.E., (1986) J. Cell Biol. 587-599) were incubated with 35SO4(2-), and the rate of appearance of heparan [35S]sulfate in the nuclei was measured. Heparan [35S]sulfate began to accumulate in the nuclei 2 h after the administration of 35SO4(2-) to the cells and reached a steady state level after 20 h. Heparan [35S]sulfate was lost from the nuclei of prelabeled cells with a t1/2 of 8 h. Chloroquine did not inhibit the transport of heparan sulfate into the nucleus, but increased the t1/2 for the exit of heparan sulfate from the nucleus to 20 h and led to a doubling of the steady state level of nuclear heparan sulfate. Heparan [35S]sulfate which was obtained from the medium or from the cell matrix of a labeled culture and which contained only low levels of GlcA-2-SO4 residues was incubated with cultures of unlabeled cells, and the uptake of the exogenous heparan [35S]sulfate was studied. At 37 degrees C the cells took up proteoheparan [35S]sulfate and transported about 10% of the internalized heparan [35S]sulfate into the nucleus, where it appeared as free chains. The heparan [35S]sulfate isolated from the nucleus was enriched in GlcA-2-SO4 residues, whereas the heparan [35S]sulfate remaining in the rest of the intracellular pool showed a corresponding depletion in GlcA-2-SO4 residues. At 16 degrees C, where endocytosed materials do not enter the lysosomes, the cells also transported exogenous proteoheparan [35S]sulfate to the nucleus with similar processing. Thus, the metabolism of exogenous heparan sulfate by hepatocytes follows the same pathway observed in continuously labeled cells and does not involve lysosomal processing of the internalized heparan sulfate.

A unique heparan sulfate in the nuclei of hepatocytes: structural changes with the growth state of the cells

Growing and confluent cultures of a rat hepatocyte cell line were labeled with 35SO4(2-) and the heparan sulfate in the culture medium, the pericellular matrix, the nucleus, the nuclear outer membrane, and the remaining cytoplasmic pool was purified by DEAE-cellulose chromatography. The heparan sulfate in all pools from the confluent cells was bound more strongly on the DEAE-cellulose column than the corresponding pools from the growing cells. Gel filtration of each pool before and after beta-elimination showed that the heparan sulfate from the nuclear and nuclear membrane pools was composed of primarily free chains, whereas the heparan sulfate in all of the other pools was a mixture of proteoglycans and free chains. The heparan sulfate in each pool was cleaved with nitrous acid to obtain mixtures of di- and tetrasaccharides. Analysis of these mixtures showed that the structural features of the heparan sulfates in each pool were different and were altered significantly when the growing cells became confluent. The nuclear-plus-nuclear membrane pools represented 6.5% and 5.4% of the total cell-associated heparan sulfate in the growing cells and the confluent cells, respectively. The structural features of the heparan sulfate in the two nuclear pools were very similar to each other, but were markedly different from those of the heparan sulfate from the other pools or from any previously described heparan sulfate or heparin. The most unusual aspect of these structures was the high content of beta-D-glucuronosyl(2-SO4)----D-glucosamine-N,O-(SO4)2 disaccharide units in these sequences. The mode of biosynthesis and delivery of these unusual sequences to the nucleus and the potential significance of these observations are discussed.