Identification of Osteoblast/Osteocyte Factor 45 (OF45), a Bone-specific cDNA Encoding an RGD-containing Protein That Is Highly Expressed in Osteoblasts and Osteocytes*

We describe the cloning and characterization of a novel bone-specific cDNA predicted to encode an extracellular matrix protein. This cDNA was identified by subtractive hybridization based upon its high expression in bone marrow-derived osteoblasts. By Northern blot analysis, we detected a single 2-kilobase mRNA transcript in bone, whereas no expression was detected in other tissues. Immunohistochemistry revealed that the protein was expressed highly in osteocytes within trabecular and cortical bone. RNA and protein expression analysis using in vivo marrow ablation as a model of bone remodeling demonstrated that this gene was expressed only in cells that were embedded within bone matrix in contrast to the earlier expression of known osteoblast markers. The cDNA was predicted to encode a serine/glycine-rich secreted peptide containing numerous potential phosphorylation sites and one RGD sequence motif. The interaction of RGD domain-containing peptides with integrins has been shown previously to regulate bone remodeling by promoting recruitment, attachment, and differentiation of osteoblasts and osteoclasts. Secretion of this RGD-containing protein from osteocytes has the potential to regulate cellular activities within the bone environment and thereby may impact bone homeostasis. We propose the name OF45 (osteoblast/osteocyte factor of 45 kDa) for this novel cDNA.

Bone is a highly dynamic tissue that undergoes continual processes of remodeling and modeling (1). In the growing skeleton, the amount of mineralized bone formed exceeds the amount lost through resorption, whereas in the mature adult lost bone mineral is precisely balanced by an equivalent amount of formation, thereby preserving the integrity of the skeleton. Under certain conditions such as aging, postmenopausal estrogen deficiency, or prolonged steroid treatment, the amount of bone formed is not sufficient to compensate for the quantity lost by resorption. Over time, this imbalance results in reduced bone mass, compromises the structural integrity of the skeleton, and, ultimately, can lead to osteoporotic bone fractures. Osteoporosis and the resulting fractures are causes of significant morbidity and mortality within the aging population.
Bone remodeling is a very complex process of tightly coordinated action by the bone resorbing osteoclasts and the bone forming osteoblasts. Osteoblasts are derived from a mesenchymal cell lineage and are responsible for the formation of new bone matrix in their differentiated state (2). In addition, osteoblasts produce factors that regulate the formation of osteoclasts and osteoclastic bone resorbing activity in response to endocrine signals such as parathyroid hormone and 1,25-dihydroxyvitamin D 3 . It has been postulated that bone loss associated with aging is caused by a defect in the osteoblast cell lineage (3)(4)(5)(6). Either the mesenchymal precursor population is insufficient or has lost the capacity to proliferate and/or differentiate into sufficient numbers of functioning osteoblasts.
Osteoblasts progress through a three-stage process of differentiation: proliferation, matrix maturation, and mineralization (7)(8)(9). During this differentiation process, a well characterized temporal and spatial expression pattern of extracellular bone matrix proteins and other genes occurs (9 -11). The bone matrix is composed primarily of type I collagen, which forms the extracellular structural component for the deposition of hydroxyapatite. Osteoblasts also secrete non-collagenous proteins into the extracellular matrix (12,13). The non-collagenous proteins include proteoglycans, sulfated glycoproteins, highly phosphorylated RGD-motif proteins, and proteins modified to contain Gla amino acid residues. Examples include biglycan, osteonectin, bone sialoprotein, osteopontin, and osteocalcin. The exact functions of many of these proteins have not yet been delineated, although, for most, evidence supports a role in promoting mineralization events (12). An exception is the Gla peptide, osteocalcin, which has been shown to be a negative regulator of bone formation by gene knockout technology (14).
The progression of osteoblast differentiation has been modeled in cell culture using primary calvarial or bone marrow cells. Bone marrow contains pluripotent stem cells of the adipocytic, osteoblastic, fibroblastic, and hematopoetic cell lineages (10,15,16). Rat and human marrow cultures require a differentiation agent such as dexamethasone, whereas mousederived cells can differentiate in the absence of dexamethasone (17)(18)(19). In vitro differentiated osteoblasts display the capacity to secrete noncollagenous proteins into the extracellular matrix in a temporally regulated manner, which may indicate a regulatory function for each of these proteins in the mineralization process (20). Furthermore, differentiated bone marrow cultures have the capacity to facilitate deposition of matrix and formation of hydroxyapatite mineral when grown in the presence of a phosphate source such as ␤-glycerophosphate. These proper-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  ties have made the differentiation of rat bone marrow cells a useful model to investigate the mechanisms of bone remodeling and osteoblast function. In fact, it is ex vivo experiments comparing marrow cultures derived from young versus aged humans that have suggested that aging related osteoporosis may be the result of a diminished capacity of marrow to produced functioning, differentiated osteoblasts (3)(4)(5).
In our efforts to better describe the molecular mechanisms by which osteoblastic cells control bone homeostasis, we have constructed a subtractive cDNA library and identified novel genes that are differentially expressed in rat bone marrow-derived osteoblastic cell cultures versus control cell cultures that were not induced to form osteoblasts. This report is focused on the identification and characterization of a single novel gene that has properties consistent with a bone-specific extracellular matrix protein. The temporal regulation of the mRNA expression specifically in osteoblasts and osteocytes during the differentiation program and the analysis of the predicted amino acid sequence are consistent with a gene having functional significance in bone. We propose the name OF45 (osteoblast/osteocyte factor of 45 kDa) for the protein encoded by this cDNA.

EXPERIMENTAL PROCEDURES
Rat Bone Marrow Cell (RBMC) 1 Culture -Primary RBMC were isolated and differentiated into osteoblastic cells as described previously (21,22). Marrow was isolated from the tibia and femur of 3-month-old Harlan Sprague-Dawley female rats and placed into culture medium (␣-minimal essential medium supplemented with 15% fetal calf serum, 2 mM glutamine, 0.1 mg/ml gentamicin, 100 units/ml penicillin, and 100 g/ml streptomycin; Life Technologies, Inc.). Isolated marrow cells were filtered through 100-m mesh, washed three times with culture medium, and plated into 10-cm tissue culture dishes at a density of 1 ϫ 10 6 cells/cm 2 . Cells were allowed to attach for 4 days undisturbed at 37°C with 5% CO 2 . Cells were grown until 90% confluent, at which time they were treated with 50 g/ml ascorbate and 10 nM dexamethasone in culture medium as indicated. Medium was changed three times per week after the initial attachment period. To assess the ability of these cell cultures to mineralize, cultures were maintained in 50 g/ml ascorbate, 10 nM dexamethasone, 10 mM ␤-glycerophosphate and stained by Von Kossa staining to assess mineral deposition.
Osteoblast Cell Line Culture-UMR-106 cells were obtained from ATCC (American Type Culture Collection, Manassas, VA) and maintained in Dulbecco's modified Eagle's medium/F-12 medium supplemented with 10% heat-inactivated fetal bovine serum, and 50 g/ml gentamycin (Life Technologies, Inc.). Cells were plated in 10-cm tissue culture dishes and grown until 80% confluent. Cells were then switched into control media, 10 nM dexamethasone, or 10 nM dexamethasone and 50 g/ml ascorbic acid. Cells were refed with fresh medium three times per week and were extracted for RNA at either 24 h or 14 days using Trizol reagent as indicated below. ROS 17/2.8 cells were the kind gift of Barbara Kream (Division of Endocrinology and Metabolism, University of Connecticut Health Center, Farmington, CT) and were plated and treated in a manner identical to that for the UMR-106 cell cultures.
Subtractive cDNA Library Construction and Analysis-A subtractive cDNA library was constructed essentially as described previously (23,24). Briefly, 5 g of poly(A) ϩ mRNA from untreated cultures and 5 g from dexamethasone-treated cultures were used to generate cDNA using Superscript II reverse transcriptase following conditions specified by the manufacturer (Life Technologies, Inc.). The PCR product after five rounds of subtractive hybridization was ligated to BamHI-digested/ alkaline phosphatase-treated pBSSKϩ (Stratagene, La Jolla, CA). Colonies were screened by hybridization to round 5 probe as described (24). Plasmid DNA was isolated from hybridizing colonies and sequenced.
cDNA Cloning-A cDNA library was constructed from mRNA isolated from dexamethasone-treated RBMC using the pSPORT1 vector system as described by the manufacturer (Life Technologies, Inc.). The cDNA library was plated, transferred to Hybond N membranes (Amersham Pharmacia Biotech), denatured, cross-linked with a Stratalinker (Stratagene, La Jolla, CA), and hybridized according to standard methods (25). One million colonies were screened by hybridization with three different 32 P-labeled fragments from the subtractive cDNA library. 28 positive clones were identified and sequenced. The longest clone, named 51-35, hybridized to all three probes and was sequenced on both strands to confirm the sequence. The first nucleotide of the longest clone corresponds to base 17 in the sequence shown in Fig. 1. The 5Ј end of the mRNA was determined by RNA ligase-mediated RACE as described (26). The OF45 gene-specific primers were 44A (5Ј-GTTAGGTAGT-GCGTGCTTCG-3Ј) and 44B (5Ј-GGACATCTGTTGGAATTACGC-3Ј). The pGbx-1 RNA linker primers were NRC-1 (5Ј-CCAAGACTCACT-GGGTACTGC-3Ј) and NRC-2 (5Ј-CTAGAGGGGCCTGTTGAACC-3Ј).
RNA Isolation and Northern Blot Analysis-RNA was isolated from cell culture with Trizol reagent (Life Technologies, Inc.) according to a protocol modified for proteoglycan-rich sources (27). RNA was isolated from rat tissues using guanidine isothiocyanate extraction, followed by a cesium chloride gradient (25). Northern blots were performed using Nytran membranes (Schleicher & Schuell) by standard methods (25). Different cDNA probes were used on the identical filters after stripping of the filters overnight in prehybridization buffer at 57°C prior to adding the next radiolabeled cDNA probe.
Each cDNA probe was obtained by reverse transcriptase-PCR from either rat tibia or ROS  Mammalian Expression of Rat OF45 and Western Blot Analysis-A mammalian expression vector for OF45 was created by PCR amplification of the coding region and subcloning into the cytomegalovirus mammalian expression vector pcDNA3 (Invitrogen, Carlsbad, CA). Cell cultures were transiently transfected with pcDNA3 or OF45pcDNA3 using LipofectAMINE Plus according to manufacturer's instructions (Life Technologies, Inc.). Medium was harvested 48 h after transfection. Samples were denatured with reducing sample buffer (Novex, San Diego, CA), separated by electrophoresis on Novex 10% NuPAGE gels, and transferred to polyvinylidene difluoride using a semidry transfer system. Blots were blocked using 5% nonfat dry milk for 1 h, and primary antibody diluted in 0.25% nonfat dry milk was added for 1 h. After washing, goat anti-rabbit peroxidase secondary antibody (Pierce) was diluted into 0.25% nonfat dry milk and applied to the blot for 1 h. After washing, signal was detected using an ECL Plus detection kit (Amersham Pharmacia Biotech, Buckinghamshire, United Kingdom) following the manufacturer's instructions.
Antibody Generation-OF45 protein spanning the entire coding region was expressed in bacteria using pET-28a (Novagen, Madison, WI). This construct created a fusion protein with a His 6 tag and T7 tag at the amino terminus. A glutathione S-transferase fusion protein was created by subcloning the BamHI/EcoRI fragment in the Escherichia coli expression construct pGEX-2TK (Amersham Pharmacia Biotech, Piscataway, NJ). High titer polyclonal antisera was generated by immunization of rabbits by Research Genetics (Huntsville, AL). A polyacrylamide gel slice containing ϳ20 mg of highly purified glutathione S-transferase-OF45 fusion protein was used as the immunogen. This material was obtained by SDS-PAGE purification of glutathione Stransferase-OF45 fusion protein that was solubilized from the inclusion bodies. The His tag OF45 fusion protein expressed in pET28a was purified and coupled to cyanogen bromide-activated Sepharose 4B, as suggested by the manufacturer (Amersham Pharmacia Biotech). Approximately 5 mg of protein/ml of resin was coupled. Anti-OF45 antibody was affinity-purified by standard methods described previously (29).
Immunohistochemistry-Adult rat tibia were fixed in 4% paraformaldehyde and decalcified in 10% EDTA. Bones were embedded into paraffin, and 5-m sections were placed onto Digene silanated slides (Beltsville, MD) and heated overnight at 50°C. Slides were rehydrated, and endogenous peroxidase activity was blocked by incubating sections in 3% H 2 O 2 in methanol. Antigenic sites were exposed by pepsin digestion at 20,000 units in 0.01 M HCl. Sections were blocked in 1% BSA/ PBS and incubated overnight at 4°C in a humidified chamber with primary antibody diluted at 1:500 in 0.1% BSA/PBS or rabbit IgG (Vector Laboratory Inc., Burlingame CA) as a control. Secondary antibody consisting of anti-rabbit IgG Fab fragments linked to horseradish peroxidase (Amersham Pharmacia Biotech) was diluted at 1:100 in 0.1% BSA/PBS and placed onto sections for 1 h at room temperature.
Detection was performed using a diaminobenzidene substrate chromagen system according to the manufacturer's instructions (Vector Laboratory).
Bone Marrow Ablation-Animal experimental protocols were carried out according to Pfizer's animal care approved protocols, and animals were maintained in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Tibial marrow ablation was performed as described by Suva et al. (30). At the indicated times after ablation, the tibia were isolated and split longitudinally. The cellular contents of the marrow cavity were removed by scraping with a spatula, and RNA was isolated as described for tissues.

Isolation of Genes That
Are Induced during Osteoblast Differentiation-Primary rat bone marrow cells were cultured for 4 weeks in the presence of vehicle or 10 nM dexamethasone. Consistent with earlier literature reports (18,21,22), this culture system required dexamethasone treatment to potentiate differentiation into osteoblastic cells as evidenced by induction of osteoblast marker genes (Fig. 4) or the ability to detect mineralization by Von Kossa staining when a phosphate source such as ␤-glycerophosphate was included in the culture medium (data not shown).
A cDNA library enriched for genes induced during osteoblastic differentiation was constructed by subtractive hybridization of cDNA from dexamethasone treated cultures versus cDNA from vehicle-treated cultures. After 5 rounds of subtractive hybridization, clones were analyzed by colony screening using round 5 subtracted cDNA as probe. A summary of the sequence analysis for 153 randomly selected, hybridizing clones is shown in Table I. The sequence data were assembled and compared with known genes in GenBank with the BLAST algorithm (31). An indicator of the efficiency of the subtraction was that there was very little contamination with ribosomal RNA (1/ 153) or with mitochondrial genes (0/153). The identification of genes previously reported to be induced during osteoblast differentiation (10), such as osteopontin, bone sialoprotein, and osteocalcin, gave an initial validation that the subtractive library contained osteoblast marker genes.
In addition to known osteoblast marker genes, sequence analysis revealed a number of unknown genes. The mRNA corresponding to most of the novel sequences could not be detected by Northern blot analysis of 20 g of total RNA. More sensitive methods will be required to confirm their regulation during osteoblast differentiation. In this report, we focus on the characterization of a single novel gene that was strongly represented (17% of analyzed clones) in the subtractive library and the induction of which could be detected by Northern blot analysis of total RNA.
Full-length cDNA Cloning-The full-length cDNA was cloned as described under "Experimental Procedures" and sequenced (Fig. 1). A Kyte-Doolittle hydrophobicity plot demonstrated that the sequence contained a hydrophobic leader sequence followed by a hydrophilic protein ( Fig. 2A). Although the predicted amino acid sequence lacked the NH 2 -terminal positively charged region (N-region) characteristic of a signal sequence for a secreted protein, it does contain an aminoterminal hydrophobic region (H-region) characteristic of a signal sequence as described by von Hiejne (32). Analysis of the amino acid sequence with PSORTII, a program to predict the subcellular localization sites of proteins from their amino acid sequences indicates that this is likely to be an extracellular protein (33,34). The predicted site of cleavage would be between Ala-16 and Ala-17. Assuming the actual cleavage occurs after Ala-16, the secreted protein would be 419 amino acids in length with a calculated molecular weight of 44,744 Da. The predicted amino acid composition of the processed protein is shown in Fig. 2B. The sequence is rich in serine, glycine, and charged residues. It would be a basic protein having a calculated pI of 8.67 with a net charge of ϩ5.13 at neutral pH in the absence of post-translational modification. There are numerous potential phosphorylation sites including nine consensus casein kinase II sites shown in Fig. 1. Phosphorylation sites offer the potential for post-translational modification to increase the acidic character of the protein. Finally, sequence analysis revealed one consensus RGD integrin recognition motif characteristic of some extracellular matrix proteins (Fig. 1).
Transient transfection in Chinese hamster ovary cells confirmed that OF45 can be secreted by mammalian cells (Fig.  2C). The apparent molecular mass by SDS-PAGE was ϳ45 kDa, consistent with that predicted from the cDNA sequence for the cleaved peptide. Recombinant expression in Chinese hamster ovary and UMR cells (data not shown) gave no evidence for additional processing or post-translational modifications that could be detected by SDS-PAGE. However, it is possible that proteolytic or other post-translational processing occurs with endogenously expressed protein in bone.
Comparison of OF45 Sequence to Known Genes-Comparison of the OF45 cDNA sequence to GenBank using the BLAST 1.4 algorithm (31) showed little primary sequence homology to any known genes in the data base. Weak homology was found to the protein DMP1 (also known as AG1) (35). The primary sequence homology is shown in the alignment in Fig.  3. Although the homology at the primary sequence level is very weak (22% identity, 37% similarity, expected probability 3 ϫ 10 6 ) there are general similarities that support the relatedness of these two proteins. OF45 shares properties with DMP1 including the fact that both genes are highly charged, serine-rich, and contain a hydrophobic leader sequence. OF45 is also analogous to DMP1 in that both genes exhibit a highly tissue restricted expression. DMP1 is most highly expressed in mature odontoblasts, the tooth-forming cells analogous to the bone forming osteoblasts that express OF45 (35)(36)(37)(38)(39)(40)(41)(42)(43).
Expression of OF45 in Comparison to Other Osteoblast mRNAs in Differentiating Rat Bone Marrow Cell Cultures-Northern blot analysis was used to detect gene expression changes over 21 days in culture in differentiating RBMC (Fig.  4). As expected, based on earlier literature (18,21,22), RBMC cultured in the absence of dexamethasone did not express mRNA of bone sialoprotein or osteocalcin, although they did express alkaline phosphatase, osteonectin, and osteopontin. Dexamethasone induced differentiation as indicated by increased osteopontin, alkaline phosphatase, osteocalcin, and bone sialoprotein mRNA. OF45 mRNA was first detectable at 6 days and increased throughout the 21 days. OF45 expression thus correlated with the increased differentiation toward the osteoblastic cell type. Temporally, OF45 expression was most similar to the expression of bone sialoprotein and osteocalcin in this culture system. These mRNAs are considered markers of mature osteoblasts, with expression occurring late in the differentiation program of osteoblasts (7,10,21).
Tissue Distribution of OF45 mRNA-OF45 was undetectable in poly(A) ϩ RNA from heart, brain, spleen, lung, liver, skeletal muscle, kidney, and testis by Northern blot analysis (Fig. 5A). Hybridization of the identical filter with an actin probe demonstrated that the filter contained intact mRNA (Fig. 5A). In addition, 20 g of total RNA from pituitary, tibial shaft, uterus, marrow, tibial metaphysis, intestine, aorta, RBMC Ϯ dexamethasone, brown fat, and white fat were probed for OF45 expression (Fig. 5B). OF45 was expressed in tibial shaft and metaphysis as well as dexamethasone-treated RBMC but not detectable in other tissues. Based on these results, expression of OF45 was highly specific to bone.
Expression of OF45 mRNA in Bone Compartments-In order to further characterize the expression of OF45 in bone, total RNA from various bone compartments was subjected to Northern blot analysis (Fig. 6). Expression of OF45 was shown to be highest in tibial shaft, tibial metaphysis, and tibial growth plate with lower expression seen in femoral neck and calvaria. OF45 mRNA was not detectable in bone marrow or periosteum by this method. The expression of OF45 differed from that of other bone markers. Most notably, OF45 mRNA was expressed highly in the tibial midshaft. This sample represented primarily cortical bone. Other osteoblast markers such as alkaline phosphatase, osteopontin, and osteonectin are expressed only at low levels in the cortical bone of tibial shaft. Similarly the osteoclast marker, tartrate-resistant alkaline phosphatase, is detectable only at low levels in the tibial shaft. Even in comparison to osteocalcin and bone sialoprotein, OF45 had a significantly different distribution. OF45 had roughly equal expression in the tibial shaft versus tibial metaphysis and tibial growth plate, whereas osteocalcin and bone sialoprotein were less abundant in the tibial shaft versus the metaphysis or growth plate.
Expression of OF45 mRNA in Osteoblastic Cell Lines-Although RBMC cultures consist predominantly of osteoblastic cells after dexamethasone treatment, they contain multiple cell types. The RNA derived from bone compartments was similarly derived from the multiple cell types present in bone. In order to strengthen the evidence for OF45 expression in osteoblastic cells, we determined the expression pattern of OF45 in the well characterized osteoblastic cell lines, UMR106 and ROS17/2.8 cells. Cells were cultured in the presence of dexamethasone or vehicle for either 1 or 14 days. Total RNA was extracted, and Northern blot analysis was performed (Fig. 7). In UMR106 cells, OF45 mRNA was detectable at a low level on day 1 in the presence of dexamethasone or dexamethasone/ascorbic acid. By day 14, mRNA expression was increased in the presence of dexamethasone and was highest at day 14 in dexamethasone/ ascorbic acid. ROS17/2.8 cells cultured in the presence of dexamethasone or dexamethasone/ascorbic acid showed higher levels of expression at day 1 versus day 14. Clearly, OF45 mRNA was expressed in these osteoblastic cell lines. Further work will be needed to understand the differences in the expression pattern at various times after dexamethasone treatment between UMR106 and ROS 17/2.8 cells. It is possible that the differences in OF45 expression was reflective of the differences in the differentiation character of these two cell lines. The expression of OF45 mRNA in UMR106 cells under various differentiation conditions was compared with other osteoblast markers in UMR cells (Fig. 8). Consistent with Fig. 7, OF45 mRNA was induced by dexamethasone/ascorbic acid and to a lesser extent by dexamethasone alone. Osteocalcin and alkaline phosphatase were also induced by dexamethasone or dexamethasone/ascorbic acid. Although the pattern of OF45 induction resembled that observed for osteocalcin, differences were noted, i.e. osteocalcin was induced over the length of culture even in the presence of ascorbic acid alone. In contrast, OF45 mRNA was not induced after 14 days with only ascorbic acid. Dexamethasone treatment was required for induction of OF45.
Immunohistochemistry-Affinity-purified anti-OF45 poly-clonal antibodies were generated and used to localize OF45 protein expression in normal adult bone tissue by immunohistochemistry. Specific staining was detected in osteocytic cells within both trabecular and cortical surfaces, indicating expression in highly differentiated osteoblasts (Fig. 9). There was no specific staining detected in osteoclasts, chondrocytes or periosteal cells. Additionally, cells of the hematopoetic lineage contained in the marrow compartment were negative for OF45 staining.
Expression of OF45 in Bone Marrow Ablation Model-In order to examine the expression of OF45 relative to previously known bone marker genes, we employed the rat tibial marrow ablation model as an in vivo model of synchronous bone formation and resorption (30,44). Northern blot analysis of the cellular contents that were isolated from the marrow space at various times after ablation are shown in Fig. 10. Standard osteoblast markers, including alkaline phosphatase, bone sialoprotein, collagen, osteonectin, and osteocalcin, were detected as early as days 4 -6. In contrast, OF45 expression was not detectable at significant levels until day 7, peaking at days 9 -10. In fact, OF45 mRNA expression more closely paralleled that of the osteoclastic marker tartrate-resistant alkaline phosphatase.
In contrast to the Northern blot analysis, immunohistochemistry of the marrow-ablated samples detected abundant protein expression within the embedded osteocytic cells even at early times after ablation (Fig. 11). Expression was continually detected throughout the time course. Coupling the Northern analysis with immunohistochemistry, we conclude that OF45 was expressed only after the osteoblastic cells had become embedded within bone matrix and, therefore, could not be detected in the isolated RNA from the marrow space. In striking contrast, known osteoblast markers were readily detected in the contents of the marrow space at early times after ablation. DISCUSSION We have isolated and characterized a novel bone-specific cDNA, named OF45, from a differentiated rat bone marrow cell culture system. OF45 mRNA was present at very low levels in these cultures in the absence of dexamethasone but was induced by the addition of this differentiation agent. The induc- tion continued over time in culture, indicating that expression of OF45 mRNA increased with the increasing differentiation state of the osteoblast lineage. Rat bone marrow cells cultured in the presence of dexamethasone have been shown to be highly enriched in osteoblastic cells. These cultures express numerous markers of osteoblasts including alkaline phosphatase, osteopontin, bone sialoprotein, and osteocalcin. In addition, these cultures contain functional osteoblasts as defined by the capability to facilitate hydroxyapatite mineral formation. However, these cultures are heterogeneous, containing precursors for adipocytic, osteoblastic, fibroblastic, and hematopoetic cell lin-eages. In order to confirm that OF45 mRNA was expressed in the osteoblastic cells rather than another cell population, we utilized two well characterized rat osteoblastic immortalized cell lines, UMR 106 and ROS 17/2.8. UMR 106 cells are thought to reflect the phenotype of a newly formed osteoblast, exhibiting high bone sialoprotein expression and low osteocalcin expression. These cells are able to form apatitic mineral in a confluent state without the formation of bone nodules. ROS 17/2.8 cells are derived from rat osteosarcoma and have been used as a model for a more mature osteoblast phenotype. OF45 mRNA could be detected in both cell lines, confirming that this gene was expressed in osteoblastic cells. Therefore, we believe it is likely that most, if not all, of the expression observed in the marrow cultures was within cells of the osteoblastic lineage.
Interestingly, there were differences in OF45 expression in UMR106 cell versus that observed in ROS17/2.8 cells. In UMR cells, OF45 mRNA was barely detectable under basal culture conditions, but greatly induced when cells were maintained at confluence in the presence of dexamethasone. This parallels closely what was observed in the bone marrow cultures. In contrast to the marrow cultures, UMR106 cells can facilitate hydroxyapatite formation in the absence of dexamethasone. Therefore, the capacity of the cells to facilitate mineral deposition is not sufficient to ensure strong OF45 mRNA expression. The expression in ROS17/2.8 cells was almost the reciprocal of that observed in UMR106 cells. OF45 mRNA expression in ROS17/2.8 cells was detected in cells grown under standard basal medium, was induced after 1 day of dexamethasone/ascorbate treatment, but fell to almost undetectable levels when the cells were maintained at confluence for 14 days. The differences observed between these two cell lines may reflect the inherent differentiation character unique to each line. Further experiments will be required to elaborate the underlying reasons and implications of the differences.
Analysis of the rat tissue distribution demonstrated high OF45 mRNA in bone tissue, as might be expected as this is the FIG. 9. OF45 protein was expressed in osteocytes of adult rat tibia. Sections of adult rat tibia were subjected to immunohistochemical detection for OF45 using an affinity-purified polyclonal antibody. Serial sections were incubated with either the specific antibody or rabbit IgG at the same concentration and detected using a secondary antibody linked to horseradish peroxidase and diaminobenzidene staining. Specific staining can be seen in osteocyte cells in both trabecular and cortical bone. site of differentiated osteoblasts. A survey of tissues revealed that OF45 mRNA was highly restricted to bone even when poly(A) ϩ RNA was used in Northern blot analysis. Of course, based on this analysis, we cannot rule out high levels of OF45 expression in specific rare cell types within other tissues, expression in tissues not examined such as teeth, or possibly expression under very precise physiological conditions. It would not be surprising if other sites of OF45 expression are described, and further work will be needed to determine the functional relevance at all sites. Many other "osteoblast" marker genes, such as osteopontin are widely expressed. Even the most restricted genes have been detected in other tissues under various conditions. For example, osteocalcin, widely regarded as a bone-specific marker, has been detected in platelets (45).
Subfractionation of bone tissue and Northern blot analysis indicated that, although OF45 mRNA was expressed within identical sites as other markers, the relative distribution within the sites differed. Most noticeable was that OF45 was most abundant in the tibial midshaft. This section of bone was composed of primarily cortical bone and contained little contribution from marrow, unattached cells, or trabecular bone. This result suggested expression in the osteoblastic cells that have become embedded in bone matrix and further differentiated into osteocytes. Osteocytes are terminally differentiated osteoblasts, which become encased in the mineralized matrix during bone formation (46). The functional role of the osteocyte re-mains unclear; however, they can communicate through dendritic processes and have been theorized to be the primary sensors of mechanical strain in bone (47)(48)(49). We confirmed the expression within osteocytes at the protein level by immunohistochemistry. The expression of OF45 within these highly specialized cells may imply a functional role in the activity of osteocytes to regulate bone mineral homeostasis.
The difference between the expression of OF45 versus the expression of other osteoblast markers was most dramatically illustrated in the Northern blot analysis of the extracts from the bone marrow ablation experiment (Fig. 10). OF45 was not abundantly detected in the RNA extracted from the cellular contents of the ablated marrow cavity on days 4 -6, whereas all of the other osteoblast markers were abundantly expressed. In fact, OF45 expression more closely resembled that of the osteoclastic marker, tartrate-resistant alkaline phosphatase. Immunohistochemistry demonstrated that OF45 was actually expressed at days 4 -6 but only within the cells embedded in the matrix. From these immunohistochemistry results, we conclude that OF45 expression did occur at early times after ablation when it has been shown that rapid matrix formation was occurring (30,44). However, we did not detect the expression by Northern blot analysis because it was expressed only after the cells had become embedded in the bone matrix and could not be isolated in the cellular extracts from the marrow compartment. Consistent with this conclusion, OF45 was detected at the later times when osteoclast activity resulted in the release of the trapped cells where they could be harvested from the marrow compartment. Therefore, we conclude that, in the ablation model, the expression of OF45 occurred only after the cells became trapped in the matrix. The other osteoblast markers were expressed in this model prior to the osteoblasts becoming trapped in the matrix. The differences in expression between OF45 and other osteoblast markers may indicate a distinct function for OF45 in regulation of bone homeostasis.
Analysis of the predicted amino acid sequence suggests that OF45 shares many features with bone extracellular matrix proteins beyond the already discussed expression pattern. These features include a secreted polypeptide that is highly charged, rich in potential serine phosphorylation sites, and contains a potential integrin recognition RGD sequence. Blast analysis revealed only remote similarity to the protein DMP1 (also called AG1) (35). Originally reported as a dentin-specific protein, recent studies indicate that DMP1 expression is more widespread, appearing in odontoblasts, ameloblasts, cementoblasts, and osteoblasts (36,37). A common feature of these cells types is that all have been implicated in producing mineralizing extracellular matrix (43). It has been proposed that DMP1 plays a role in regulating the bioprocess of matrix mineralization (36,41,43,50,51). OF45 may play a similar role. However, it is important to point out that OF45 is only distantly related to DMP1 and in some characteristics differs markedly. For example, dpm1 is highly acidic whereas OF45 is predicted to be slightly basic in the absence of post-translational modification.
During the remodeling cycle, extracellular matrix proteins produced by osteoblasts and osteocytes such as osteopontin and bone sialoprotein promote cell attachment and activity of osteoclasts (1,12). In the reversal stage, osteoblast-derived extracellular matrix proteins also recruit preosteoblasts and facilitate attachment and differentiation, leading to new bone matrix formation. These interactions are often mediated through interactions of cell surface integrins with an RGD motif. Extracellular matrix proteins have also been implicated in the maturation of bone matrix and directly regulating hydroxyapatite formation through direct binding of calcium ions. Experimental evidence suggests extracellular matrix proteins FIG. 11. OF45 protein was expressed by the embedded cells even at the earliest time points after marrow ablation. Sections of tibia from marrow-ablated rats were subjected to immunohistochemical detection for OF45 using an affinity-purified polyclonal antibody. Serial sections were incubated with either the specific antibody or rabbit IgG at the same concentration and detected using a secondary antibody linked to horseradish peroxidase and diaminobenzidene staining. Sections from days 0, 2, 4, 8, 10, and 16 days after ablation show specific staining in cells embedded in matrix. Sections stained with IgG showed no staining similar to those seen in Fig. 9 (panel 3) (data not shown). can be either inhibitory or stimulatory to mineralization depending on both the peptide and its local concentration. A single protein can function in both cell-matrix interaction and in matrix maturation, as has been suggested for osteocalcin. Gene knockout experiments of the osteocalcin gene have suggested that it functions as an inhibitor of bone formation, perhaps through interaction with a specific receptor (14,52). Further experiments have suggested it also plays a role in matrix maturation (53).
In summary, OF45 is a novel gene expressed specifically in bone. By analogy to other RGD-containing matrix proteins, possible cellular functions include the regulation of cell-matrix interactions with osteoblasts, osteoclasts, and other cell types. In addition, like other noncollagenous bone matrix proteins, its expression in bone tissue could affect matrix maturation and its ability to mineralize properly. We have demonstrated that OF45 expression can differ from that of other osteoblast markers, implying some unique roles. Under specific physiological and pathological conditions, we expect OF45 could impact bone tissue in different ways. As with other matrix proteins, further experiments will be necessary to elaborate the precise molecular and biological functions of the peptide. Given that it is specific to bone, it could be an excellent target for drug discovery efforts attempting to treat bone diseases, either the protein itself or through inhibition or activation of its activity.