Biglycan Modulates Osteoblast Differentiation and Matrix Mineralization


  • The authors have no conflict of interest


MC3T3-E1 cell-derived clones expressing higher (S) or lower (AS) levels of biglycan were generated and characterized. The processes of cell differentiation and matrix mineralization were accelerated in S but delayed in AS, indicating that BGN modulates osteoblastic cell differentiation.

Introduction: Biglycan (BGN), a member of the small leucine-rich proteoglycan family, is one of the major proteoglycans found in bone and has been implicated in bone formation. In this study, the effects of over- or underexpression of BGN on osteoblastic cell phenotypes and matrix mineralization were studied.

Materials and Methods: MC3T3-E1 cells were transfected with vectors containing the BGN cDNA in a sense or antisense orientation to generate clones expressing higher (S clones) or lower (AS clones) levels of BGN. MC3T3-E1 cells and those transfected with an empty vector (EV) were used as controls. The levels of BGN synthesized by these clones were evaluated by Western blot analysis. Cell growth was analyzed by cell counting and cell differentiation by the gene expression patterns of several osteoblastic markers using quantitative real-time PCR. The abilities of these clones to form mineralized matrices were evaluated by in vitro and in vivo mineralization assays. Furthermore, the clones were treated with BMP-4 and their responsiveness was assessed.

Results: The cell growth in these clones was unaffected; however, osteoblast differentiation was significantly accelerated in S clones and suppressed in AS clones. The in vitro matrix mineralization in S clones was significantly enhanced but severely impaired in AS clones. When transplanted into immunodeficient mice, S clone transplants exhibited larger areas of lamellar bonelike matrices, whereas only minute amounts of woven bone-type structure was found in AS transplants. The response to BMP-4 was higher in S clones but poorer in AS clones compared with that of controls.

Conclusions: BGN modulates osteoblast differentiation, possibly by regulating BMP signaling, and consequently matrix mineralization.


BIOMINERALIZATION IS A process by which mineral is deposited within or around the extracellular matrices in various organisms. In the collagen-based mineralization system (i.e., bones, dentin, and cementum), collagen regulates the spatial aspect of mineralization by defining the space for mineral deposition and growth. (1, 2) The temporal aspect of this process, however, is likely regulated by a number of noncollagenous matrix proteins (NCPs) including members of the small leucine-rich proteoglycans (SLRPs) family. (3–5)

Biglycan (BGN), a member of class I molecules in the SLRP family, (6, 7) was originally identified in the mineral-associated compartment in bone. (8) It has been reported that BGN is highly expressed in the areas of growth of skeletal tissues and on the cell surface of both chondroblasts and osteoblasts, (9) suggesting its involvement in mineralization. (10) It likely binds to type I collagen(11) and possibly acts as a nucleator for apatite formation in a gelatin gel system. (12)

Recently, mice with targeted disruption of the Bgn gene have been generated. (13) The mice revealed an age-dependent osteoporosis-like phenotype characterized by a reduced growth rate of bone and decreased bone mass. The phenotype is possibly caused by a decrease in the number of bone marrow stromal cells (BMSCs; the osteogenic precursors) and an increase in apoptosis of these cells. (14) In addition, when genes for Bgn and decorin (Dcn), its structurally related SLRP member, were both disrupted, (15) a more severe bone phenotype was seen compared with those of the single BGN-deficient mice. These findings indicated that the effects of BGN/DCN double deficiency are synergistic. Because bone is not significantly affected in the absence of DCN, the results suggested that these two SLRP members may partially rescue or compensate each other in single deficient mice. (10, 15)

To gain more insight into the potential functions of BGN in bone biology, MC3T3-E1-derived stable cell clones expressing higher or lower levels of BGN were generated, and their phenotypes were partially characterized.


Cell culture

The mouse MC3T3-E1 cell line subclone 26(16, 17) (referred to as MC3T3-E1 cells hereafter) were maintained in α-MEM (Gibco) supplemented with 10% FBS, 100 units/ml penicillin, and 100 μg/ml streptomycin. The medium was changed twice a week.

Construction of cDNA and generation of stably transfected cell clones

Total RNA was extracted from MC3T3-E1 cells using Trizol reagent (Invitrogen) according to the manufacturer's protocol. The first-strand cDNA was synthesized by Ominiscript RT kit (Qiagen), and the cDNA containing the coding sequence of mouse biglycan (GenBank accession X53928) was isolated by RT-PCR using Hotstar Taq polymerase (Qiagen). The sequences of the specific primers were as follows: forward primer, 5′-ACAGGTACCATGTGTCCCCTGTGGCTACTC-3′ (positions 235–256) and reverse primer, 5′-CTGCTCGAGCTTCTTATAATTTCCAAATTG-3′ (positions 1322–1341). The PCR products were ligated into the pcDNA3.1/V5-His-TOPO mammalian expression vector (Invitrogen) and sequenced at the UNC Sequencing Facility (University of North Carolina, Chapel Hill, NC, USA). After the sequence of the insert was verified, MC3T3-E1 cells were transfected with the construct containing BGN cDNA in either sense (S) or antisense (AS) orientation using FuGENE6 transfection reagent (Roche Diagnostics) and maintained in a medium containing 400 μg/ml of G418 (Gibco) for up to 4 weeks. Stable cell clones (S or AS clones) were isolated by cloning rings and further cultured under the same conditions. MC3T3-E1 cells were also transfected with an empty pcDNA3.1/V5-His A vector (Invitrogen), maintained, isolated in the same manner as described above, and used as a control (EV).

Western blot analysis

Several stable clones and controls (MC3T3-E1 cells and EV) were seeded on 10-cm dishes at a density of 1 × 106 cells/dish and cultured for 2 weeks. The cell/matrix layers were washed with PBS and extracted with 6 M guanidine-HCl, pH 7.4. The extracts were exhaustively dialyzed against distilled water and lyophilized. The dry weights of the extracts were all comparable among the clones and controls. Protein concentration in each sample was measured by a DC Protein assay kit (Bio-Rad). Equal amounts of protein in each sample were digested with chondroitinase ABC (Seikagaku, Tokyo, Japan) at 37°C for 3 h, dissolved in SDS sample buffer under reducing conditions, separated by 4–12% gradient SDS-PAGE, and transferred onto a polyvinylidene fluoride (PVDF) membrane (Immobilon-P; Millipore). The membrane was blocked with Tris-buffered saline/0.1% Tween 20 (TBST) containing 5% nonfat dry milk and incubated with polyclonal antibody against mouse BGN core protein (LF-159)(18) (generous gift from Dr LW Fisher, National Institute of Dental and Craniofacial Research, Bethesda, MD, USA). The immunoreactivities were visualized by alkaline phosphatase conjugate substrate kit (Bio-Rad) and analyzed using Scion Image software (Scion Corp., Frederick, MD, USA). The levels of DCN in the same extracts were also assessed using polyclonal antibody against DCN core protein (LF-113)(18) as described above. The BGN and DCN levels in cultured medium were also analyzed. Cultured medium pooled during the first week of culture was dialyzed exhaustively against distilled water at 4°C and lyophilized. After aliquots were reconstituted in distilled water, protein concentrations were determined using a DC protein assay kit, and equal amounts of protein in each sample were treated with chondroitinase ABC and subjected to Western blot analyses as described above.

Cell growth

The clones and control cells were plated on a 24-well plate in triplicate at a density of 2 × 104 cells/well and cultured for up to 21 days. At days 2, 5, 7, 10, 14, and 21, cell numbers were determined by cell counting and plotted on a logarithmic scale.

In a separate set of experiments, cells were plated in triplicate at a density of 5 × 104 cells/well in a 96-well plate, cultured for up to 8 days, and subjected to MTS cell proliferation assay (CellTier 96; Promega) at days 2, 5, and 8, according to the manufacturer's protocol. The amounts of formazan compound produced by metabolically active cells were measured by absorbance at 490 nm.

Quantitative real-time PCR

To determine the mRNA expression patterns of osteoblastic markers in S clones, AS clones, and control cells, cells were plated at a cell density of 2 × 105 cells/35-mm dish. After reaching confluence, the medium was replaced with the one containing 50 μg/ml ascorbic acid and 2 mM β-glycerophosphate (mineralization medium). Total RNA was extracted as described above at days 7 and 14. Two micrograms of the total RNA extract was used for RT using the Omniscript RT kit (Qiagen). Quantitative real-time PCR was performed using the sequence specific primers and probes for core binding factor 1/runt-related transcription factor 2 (Cbfa1/Runx2; ABI assay number Mm00501578_m1), Osterix (Osx; Mm00504574_m1), type I collagen α 2 chain (Col1A2; Mm00483888_m1), and bone sialoprotein (BSP; Mm0000492555_m1) as markers of osteoblast differentiation and analyzed by the ABI Prism 7000 Sequence detection system (Applied Biosystems, Foster City, CA, USA). All analyses were done in triplicate, according to the manufacturer's protocol, and to confirm reproducibility, two independent experiments were performed. The mRNA expression levels relative to GAPDH (ABI assay 4308313) were determined, and the fold changes were calculated using the values of MC3T3-E1 cells at each time-point as a calibrator by means of 2−ΔΔCT method. (19)

In vitro mineralization assay

The clones and control cells were seeded at a density of 2 × 105 cells/35-mm dish, cultured in the mineralization medium as described above, and maintained for up to 28 days. To evaluate the mineralized nodule formation in vitro, at the end of each week, cell/matrix layers were washed with PBS, fixed with 100% methanol, and stained with 1% Alizarin red S (Sigma). (17) At day 21 of culture, the calcium contents were quantified by measuring the amount of Alizarin red S bound to mineralized nodules in the cultures. (20) Briefly, after staining with Alizarin red S, cultures were rinsed extensively with water, and extracted with 10% (w/v) cetylpyridinium chloride (CPC) in 10 mM sodium phosphate, pH 7.0, for 15 minutes at room temperature. The dye concentrations in the extracts were determined by absorbance at 562 nm and calculated based on the Alizarin red S standard curve in the same solution. Values were normalized based on the cell numbers determined by cell counting. To confirm the reproducibility of the results, three independent experiments were performed.

In vivo mineralization assay and histomorphometry

The protocol for animal experiments was approved by Institution Animal Care and Use Committee at the University of North Carolina at Chapel Hill.

To further confirm the abilities of the cells to differentiate and form mineralized matrices in vivo, MC3T3-E1 cells, EV, two S clones, and two AS clones (S1 and S2 that synthesized the highest and second highest levels of BGN, respectively, and AS1 and AS3 that synthesized the lowest and one of the lower levels of BGN, respectively, see Fig. 1A) were transplanted into immunodeficient mice. (21–23) Briefly, for a single transplant, 2 × 106 cells were mixed with 40 mg hydroxyapaptite/tricalcium phosphate (HA/TCP) ceramic powder (Zimmer, Warsaw, IN, USA). Trypsin-released cells and the ceramic carrier were incubated at 37°C for 90 minutes with slow rotation (25 rpm). The pelleted HA/TCP powder with adherent cells were mixed with 15 μl each of mouse fibrinogen (3.3 mg/ml solution in PBS) and mouse thrombin (25 U/ml in 2% CaCl2, both from Sigma) and allowed to form a gelatinous solid to stabilize the transplant. The resulting fibrin clot with ceramic powder and attached cells were transplanted subcutaneously into 8- to 12-week-old female beige mice (NIH-bg-nu-xidBR; Charles River Laboratories, Wilmington, MA, USA). Mice were anesthetized by intraperitoneal injection with a combination of ketamine (Fort Dodge Animal Health, Fort Dodge, IA, USA) at 140 mg/kg body weight and xylazine (The Butler Company, Columbus, OH, USA) at 7 mg/kg body weight. Four subcutaneous pockets per mouse were created by blunt dissection on the dorsal surface of each mouse and a single transplant was placed in each pocket. The transplants were harvested 7 weeks after transplantation, fixed in 10% buffered formalin (Sigma), demineralized in 0.5 M EDTA (Fisher), cut in half, and processed for histological examination (H&E staining). Two independent experiments were performed, and at least five representative sections from each transplant were evaluated and photographed under a light and polarized light microscope (Nikon FXA) at x10 magnification. The bone matrix-like area (B), HA/TCP area (HA), and total area (T) were measured by Scion image software and calculated as percentage of B/T and B/T-HA.

Figure FIG. 1..

(A) BGN and (B) DCN synthesis in stable clones and controls. (Top) Cell/matrix extracts at week 2 and (bottom) cultured medium pooled during the first week of culture were analyzed. Those fractions were digested with chondroitinase ABC and analyzed for (A) BGN and (B) DCN by Western blot analysis. All three S clones, referred to as S1, S2, and S3 according to their expression levels (S1 > S2 > S3), possessed higher levels of BGN compared with that of controls, MC3T3-E1 cells (MC), or an empty vector-transfected clone (EV). All three AS clones synthesized lower levels of BGN compared with that of controls. AS1 clone exhibited the lowest level of BGN, and AS2 and AS3 were comparable. The levels of DCN in S and AS clones were all comparable in cell/matrix layer among clones and controls, but slightly increased in AS and decreased in S clones in cultured medium.20

Responsiveness to BMP-4

S1, S2, AS1, AS3, and controls were plated on 35-mm dishes at a density of 2 × 105 cells/dish. On the following day, cells were incubated in the presence or absence of 50 ng/ml BMP-4 (R&D systems) and cultured for 4 days. Total RNA was collected, and the expression of osteoblastic markers was analyzed by real-time PCR as described above.

In another set of experiments, cells were cultured in the same manner and incubated for 48 h in the presence or absence of 50 ng/ml BMP-4 in serum-free medium and lysed in lysis buffer (20 mM Tris, 0.5 mM MgCl2, and 0.1% Triton X), and alkaline phosphatase (ALP) activity was measured using 10 mM p-nitrophenol phosphate (PNP; Sigma) as a substrate. After incubation for 30 minutes, the reaction was terminated with 3N NaOH to a final concentration of 0.5N NaOH, and the production of PNP was measured by absorbance at 405 nm. The activity was calculated based on the standard PNP solutions and expressed as micromolar PNP produced per minute per milligram of total protein. (24)

Statistical analyses

All statistical analyses were performed using Sigma Stat software. Data are expressed as means ± SD. Statistical differences were determined by one-way ANOVA followed by a Tukey-Kramer multiple comparison test.


Generation of cell clones expressing higher (S) and lower (AS) levels of BGN

The levels of BGN present in cell/matrix layers (top panel) and cultured medium (bottom panel) evaluated by Western blot analyses are shown in Fig. 1A. The level of BGN in EV was comparable with that of MC3T3-E1 cells. All three S clones examined, designated S1, S2, and S3 according to their expression levels, revealed significantly higher levels (3.48-, 2.5-, and 2.25-fold in S1, S2, and S3, respectively, based on the analysis by Scion image software). In all AS clones, the levels of BGN synthesized were significantly lower, AS1 being the lowest level (0.17-fold compared with that of MC). AS2 and AS3 showed similar suppression levels (0.43- and 0.47-fold, respectively; Fig. 1A). In cultured medium, the levels of BGN in the respective cell groups were also higher in S clones and lower in AS clones in comparison with controls; thus, they correlated well with those observed in cell/matrix layer. Western blot analyses without chondroitinase ABC digestion were also performed, and a broad immunoreactive band corresponding to the BGN parental molecule as a proteoglycan form (∼90-140 kDa), was observed in all clones examined (data not shown). These results verified that BGN synthesized by these clones is a proteoglycan composed of a core protein and glycosaminoglycan chains and is identical to the endogenous BGN synthesized by MC3T3-E1 cells. Thus, S and AS clones were successfully established. The levels of DCN present in cell/matrix layer were essentially identical among the clones and control cells (Fig. 1B, top panel); however, DCN levels in cultured media (Fig. 1B, bottom panel) tended to be slightly higher in AS (10∼20%) and lower in S clones (∼10%).

Cell growth

Cell growth of each cell type is depicted in Fig. 2. At all time-points examined, there were no significant differences in the cell growth among S clones, AS clones, and controls. The results were also confirmed by MTS cell proliferation assay performed at different time-points (days 2, 5, and 8; data not shown).

Figure FIG. 2..

Cell growth of S and AS clones. Cell number of each group is shown in a logarithmic scale at each time-point indicated. Note that cell growth was comparable among all cell types.20

Expression of osteoblastic markers

The effects of higher or lower levels of BGN on osteoblast differentiation were studied by quantifying the mRNA expression levels of several osteoblastic markers. The results are presented (Fig. 3) as the fold changes relative to the respective values of MC3T3-E1 cells at the first time-points. The expression levels of all markers tested at all time-points were virtually identical between MC3T3-E1 cells and EV. However, the expression levels of Cbfa1/Runx2 and Osx, transcription factors essential for osteoblast differentiation, (25, 26) and Col1A2 and BSP were markedly increased in S clones at day 7 and remained upregulated at day 14 compared with those of controls. In AS clones, on the contrary, the expression levels of all these markers were significantly suppressed at day 7. Although the expression levels of Cbfa1/Runx2 and Col1A2 in AS clones at day 14 became comparable with the controls, those of Osx and BSP still remained markedly suppressed (Fig. 3). Two independent sets of experiments were performed, and the results were essentially identical.

Figure FIG. 3..

mRNA expression profiles of osteoblastic markers. The mRNA expression levels of CBFA1, OSX, COL1A2, and BSP in S clones, AS clones, and controls (MC and EV) are shown as fold changes ± SD relative to the corresponding gene expression levels of MC3T3-E1 cells at each time-point examined. At all time-points analyzed, the expression levels of all osteoblastic markers were significantly increased in S clones. At day 7, the expression of all osteoblastic markers in AS clones were suppressed and at day 14; those of OSX and BSP remained markedly suppressed in all AS clones.ap < 0.005 vs. controls.20

In vitro mineralization assay

Because the results of mRNA expression of osteoblastic markers indicated that osteoblast differentiation was significantly affected by the level of BGN expression, we further studied the effects of BGN on matrix mineralization in vitro. The results are shown in Fig. 4A. In the controls, the initial mineralized nodule formation was observed at day 21, and the extent of mineralized nodules was increased thereafter. In S clones, the onset of mineralization was significantly accelerated, and the extent of mineralization was also markedly increased compared with the controls. The timing and extent of mineralization were correlated with the levels of BGN synthesized by these clones (i.e., mineralized nodules were formed earlier and to a greater extent in S1 followed by S2 and S3). In all AS clones, however, no mineralized nodules were observed even at day 28, whereas mineralization in the controls was progressed noticeably at this time-point. Measurement of calcium-bound Alizarin red S dye showed higher calcium accumulation in S clones and lower calcium accumulation in AS clones compared with those of controls (p < 0.005; Fig. 4B).

Figure FIG. 4..

In vitro mineralization assay. (A) Mineralized nodules formed by controls (MC and EV), S clones, and AS clones were stained with Alizarin red S. Representative dishes of each cell type at different time-points are shown. The timing and extent of matrix mineralization in S clones were accelerated and increased compared with those of controls. In AS clones, mineralized nodules were not observed in any clone, even at day 28 when mineralized nodules were readily observed in controls. (B) Quantitative analysis of calcium in clones and controls. The values represent concentration of extracted Alizarin red S dye normalized to cell numbers at day 21. Note that the calcium contents in S clones were significantly higher and lower in AS clones compared with those of controls.ap < 0.005 vs. controls.20

In vivo mineralization assay

The capabilities of the S and AS clones (S1 and S2 for the former and AS1 and AS3 for the latter) and the controls to form mineralized matrices were further assessed by an in vivo transplantation/mineralization assay (Fig. 5).

Figure FIG. 5..

In vivo mineralized matrix formation. (A) Paraffin-embedded cross-sections of representative S1, S2, AS1, AS3, and controls transplants harvested at 7 weeks of transplantation and stained with H&E are shown. Bonelike matrices (*) can be observed on the surface of HA/TCP carrier (HA) including osteocytes (arrowheads). The bonelike matrices in transplants from S clones appeared to be denser with compact, lamellar-like structure compared with those of controls and such structure is almost absent from AS clones. Scale bar = 100 μm. (B) The structure of bonelike matrices in S clones was further enhanced when observed under a polarized light microscope, whereas it was almost absent in AS clones. Scale bar = 50 μm.20

At week 7, the formation of bonelike matrices was observed in all transplants from S clones and controls with similar spatial arrangement (Fig. 5A). Highly stained bonelike matrices were formed on the surface and surroundings of HA/TCP carrier, and osteocyte-like cells were embedded in the matrices. Mineralized matrices appeared to be denser and have a discernible direction in the S clone transplants (Fig. 5A), which resembled lamellar bone. Although bonelike matrices were also seen in AS clone transplants, they were significantly less than those of S clones or controls and revealed as an immature/woven bone-type structure. They were discretely formed only on the surface of the HA/TCP carrier. The distinct lamellar and compacted appearance of collagen matrices in S clone transplants was further enhanced when observed under a polarized light (Fig. 5B), whereas in AS clone transplants, these structures were essentially absent. The areas of bonelike matrices formed by S clones, as shown by percentage (mean ± SD) of bonelike matrix area/total area (B/T) or bonelike matrix area/total area-carrier area (B/T-HA), were significantly greater than those of controls and AS clones (p < 0.005; Table 1). The areas of the bonelike matrices formed in the AS clone transplants were significantly smaller (p < 0.005) than those of controls (Table 1).

Table Table 1.. Histomorphometric Analysis on the Areas of Bonelike Matrices Formed by Control Cells and Clones
original image

Responsiveness of S and AS clones to BMP-4

Because BMP-4 is a potent osteogenic inducer, (27) we sought to determine the responsiveness of the clones to BMP-4 treatment. S1, S2, AS1, and AS3, together with controls, were treated with BMP-4 and analyzed for mRNA expression of Cbfa1/Runx2, Osx, BSP, and ALP activity. On treatment, expression of Cbfa1/Runx2, Osx, and BSP was significantly increased and exhibited higher levels in S1 and S2 compared with those of controls. Although the expression levels of these markers in AS1 and AS3 also increased in response to BMP-4, they remained significantly lower than those of controls (Fig. 6A). Before BMP-4 treatment, the basal levels of ALP activity were significantly higher in S clones and lower in AS clones compared with those of controls. When treated, both S clones exhibited increased levels of ALP activity, and they were significantly higher than controls treated in the same manner (Fig. 6B). However, in AS clones (AS1 and AS3), ALP activity did not significantly change by the treatment and remained at low levels.

Figure FIG. 6..

Response to BMP-4 treatment. (A) mRNA expression levels of CBFA1, OSX, and BSP with or without BMP-4 treatment. In S clones, when treated with BMP-4, the expression levels were significantly higher than those of controls. Although the treatment increased the expression levels of osteoblastic markers in AS clones, they remained significantly lower than those of S clones or controls. (B) ALP activity induced by BMP-4 in S and AS clones. The levels of activity in S clones were significantly higher than those of controls, whereas they were significantly lower in AS clones compared with those of controls.ap < 0.005 vs. controls.20


Despite the fact that BGN was discovered >20 years ago and has been linked to bone formation, (8) the roles of BGN in this process are still not well understood. In this study, by establishing and characterizing MC3T3-E1 cell-derived clones expressing higher or lower levels of BGN, we showed that BGN modulates osteoblast differentiation and subsequent matrix mineralization without affecting the cell growth.

Osteoblast differentiation has been characterized by the sequential mRNA expression of osteoblast-associated molecules. (28, 29) In S clones, all osteoblastic markers examined (i.e., two of the key transcription factors for osteoblast differentiation {Cbfa1/Runx2 and Osx}, (26, 30) an early stage marker {Col1A2}, and an intermediate/late stage marker {BSP}), were significantly upregulated, indicating that osteoblast differentiation was accelerated in these clones. On the other hand, the expression of these osteoblastic markers was downregulated in all AS clones at day 7, indicating that osteoblast differentiation was delayed. At day 14, however, two of those markers (i.e., Cbfa1/Runx2 and Col1A2) showed expression levels comparable with those of controls, whereas the other two (i.e., Osx and BSP) still remained significantly suppressed. The reason for this differential regulation is not clear at this point; however, it may reflect the complexity of signaling cascades in osteoblasts involving multiple feedback mechanisms induced by various transcription factors. (31, 32) In any case, these results strongly suggest a role of BGN in modulating osteoblast differentiation.

The in vitro mineralization assay revealed an enhanced ability of all S clones to form mineralized nodules in all S clones, whereas this ability was impaired in all AS clones. The upregulation of genes encoding bone matrix proteins essential for mineralization (i.e., type I collagen and BSP) seen in S clones and their downregulation in AS clones may well explain this observation. BSP has long been implicated in the initiation of biomineralization because its expression is correlated with the onset of matrix mineralization(33) and it is capable of binding collagen to promote hydroxyapatite formation in vitro. (34–36) Therefore, the marked differences among clones in regard to the timing of mineralized nodule formation may in part be related to the BSP expression level of the respective clones.

The results obtained from an in vitro mineralization assay were further substantiated by in vivo transplantation studies. The areas of bonelike structure in transplants are significantly larger in S and smaller in AS clones in comparison with those of controls. Previous studies have shown that the presence of bonelike structures formed in the transplants is of donor origin, (22, 23) indicating that the cells still retain their osteogenic activity on in vivo transplantation. Therefore, bonelike matrices seen in the transplants in this study are most likely formed by the respective donor cells. In the case of AS clones, whereas these clones failed to form mineralized nodules in vitro during the culture period undertaken, when transplanted into immunodeficient mice, small amounts of bonelike matrices were observed (Fig. 5; Table 1). This is likely because of the fact that the experiments were conducted in completely different settings (in vitro versus in vivo) and conditions (e.g., the period of culture versus transplantation). Concerning the difference in morphology of bonelike matrices in AS in comparison with those of controls and S clones (woven versus lamellar), there are several possible explanations. First, insufficient levels of BGN led to abnormal collagen matrix organization. However, considering relatively low binding affinity of BGN to collagen(11) and moderate changes in collagen fibrillogenesis in BGN-deficient mice, (15) this may not be the direct cause. Second, they were formed by host-derived cells rather than donor cells. Although this is not likely, (22, 23) because recent reports have shown that osteogenic/stromal cells may support hematopoiesis establishment on transplantation, (37) this possibility cannot be completely ruled out. Third, this may be the result of delayed/impaired osteoblast differentiation of AS clones resulting in immature/altered bone formation. Because the gene expression pattern of AS clones (Fig. 3) indicated that the differentiation process in these cells was delayed/impaired, this is likely the case. Since an in vitro study suggested that BGN may act as a nucleator for apatite formation, (12) the lack of BGN may also cause delayed/impaired mineralization in vivo. Nonetheless, these in vivo results are consistent with in vitro mineralization assay.

A potential compensation by DCN for the absence of BGN has been suggested, (10, 38) although the data from the same animal model are somewhat inconclusive. (13, 14, 38) DCN in the cell/matrix layer did not change among different clones and controls. However, in cultured medium, slight increases and decreases in DCN were seen in AS and S clones, respectively. Although the slight changes do not seem to rescue the phenotypes seen in the clones, a partial compensation mechanism may occur between the two SLRP members as suggested. (38)

The increased expression levels of Cbfa1/Runx2, Osx and BSP, and ALP activity in S clones in response to BMP-4 indicate that BGN modulates BMP-4-induced osteoblast differentiation. The poor response of AS clones to BMP-4 also supports this notion. These results are in good agreement with those reported by Chen et al. (38) Because Cbfa1/Runx2 has been suggested to be one of the target genes regulated by BMPs, (39) it is possible that BGN affected BMP-induced signaling, thus modulating genes encoding Cbfa1/Runx2 and its downstream molecules. In the case of BSP, transcriptional regulation of this gene is still poorly understood. Nevertheless, our study suggested the potential role of BGN as a modulator of BSP gene expression.

During the preparation of this manuscript, it was reported that BGN binds BMP-4 and chordin and antagonized BMP-4 activity in dorsoventral patterning of xenopus by forming ternary complexes with chordin and twisted gastrulation. (40) Interestingly, the same group has noted a potential positive effect of BGN on BMP-4 activity in a certain cell culture system. Likely, a modulatory function of BGN for BMP-4 activity depends on several factors such as cell types, the presence/absence of other BMP/BGN-binding molecules, and local concentrations of BGN. Further studies are warranted to elucidate the modulatory mechanisms of BGN in BMP-induced osteoblast differentiation.

In conclusion, this study shows potential roles of BGN in modulating osteoblast differentiation and subsequent matrix mineralization. These findings further underscore the importance of extracellular matrix molecules not only as a scaffold for mineral deposition but also as key regulators of cellular functions.


This study was supported by NIH Grant NIDCR-DE 10489 and NASA Grant NAG2-1596.