By continuing to browse this site you agree to us using cookies as described in About Cookies
Wiley Online Library is migrating to a new platform powered by Atypon, the leading provider of scholarly publishing platforms. The new Wiley Online Library will be migrated over the weekend of March 17 & 18. You should not experience any issues or loss of access during this time. For more information, please visit our migration page: http://www.wileyactual.com/WOLMigration/
The anti-glucocorticoid potential of BMP-2 in osteoblasts was tested in MC3T3-E1 cells using dexamethasone (1 μM) and rhBMP-2 (10 or 100 ng/ml). rhBMP-2 restored mineralization but not condensation or collagen accumulation. These results demonstrate the potential and limitations of BMPs in counteracting glucocorticoids.
Introduction: Pharmacologic glucocorticoids (GCs) inhibit osteoblast function and induce osteoporosis. Bone morphogenetic proteins (BMPs) stimulate osteoblast differentiation and bone formation. Here we tested the anti-glucocorticoid potential of BMP-2 in cultured osteoblasts.
Materials and Methods: MC3T3-E1 cells were treated with dexamethasone (DEX; 1 μM) and/or recombinant human BMP-2 (rhBMP-2; 10 or 100 ng/ml). Culture progression was characterized by cell cycle profiling, biochemical assays for DNA, alkaline phosphatase (ALP), collagen, and calcium, and by reverse transcriptase-polymerase chain reaction (RT-PCR) of osteoblast phenotypic mRNAs. Mineralization was characterized by Alizarin red and von Kossa staining and by Fourier transform infrared spectroscopy (FTIR) and X-ray diffraction (XRD).
Results: DEX inhibited differentiation-related cell cycle, nodule formation, collagen accumulation, osteocalcin, and BMP-2 gene expression as well as mineralization. Replenishment of GC-inhibited cultures with 10 or 100 ng/ml rhBMP-2 dramatically rescued mineral deposition. The rhBMP-2-rescued mineral was bone-like apatite nearly identical to the mineral of control cultures. The rhBMP-2 rescue was associated with increased mRNA levels for α1(I) collagen, osteocalcin, and Cbfa1 types I and II, as well as ALP activity. In contrast, rhBMP-2 did not rescue the GC-inhibited differentiation-related cell cycle, nodule formation, or collagen accumulation. When administered alone, rhBMP-2 also increased the mRNA levels for α1(I) collagen, osteocalcin, and Cbfa1 types I and II, as well as ALP activity. However, treatment with rhBMP-2 alone inhibited cell cycle progression, nodule formation, and collagen accumulation. Surprisingly, in contrast to its rescue of mineralization in DEX-treated cultures, rhBMP-2 inhibited mineralization in the absence of DEX. In parallel to its bimodal effect on mineralization, rhBMP-2 stimulated endogenous BMP-2 mRNA in the presence of DEX, but inhibited endogenous BMP-2 mRNA in the absence of DEX.
Conclusions: Suppression of BMP-2 gene expression plays a pivotal role in GC inhibition of osteoblast differentiation. However, the inability of rhBMP-2 to rescue the entire osteoblast phenotype suggests BMP-2-independent inhibitory effects of GCs. BMP-2 exerts both positive and negative effects on osteoblasts, possibly depending on the differentiation stage and/or the existing BMP signaling.
The incidence of osteoporosis induced by high dose glucocorticoids (GCs) has been well described; however, the mechanisms of GC action on bone remain uncertain. The overall negative effects of GC on bone include decreased bone formation, increased osteoclastic resorption, impaired calcium metabolism, and decreased production of sex steroids.(1,2) The primary causes of long-term GC-induced bone loss are now thought to be direct inhibitory effects on osteoblasts.(1) These direct effects include inhi-bition of cell cycle progression,(3,4) in particular during a defined commitment stage,(5) induction of apoptosis,(6) and impairment of osteoblast function through transcriptional and/or post-transcriptional inhibition of collagen,(7) Cbfa1,(8) and growth factors, including insulin-like growth factor (IGF)-1 and transforming growth factor (TGF)-β.(1)
Bone morphogenetic proteins (BMPs) are potential antidotes to the detrimental effects of GC on osteoblasts because they promote osteoblast differentiation.(9–16) In primary cultures and osteoblast-like cell lines, BMP-2 has been shown to increase alkaline phosphatase (ALP) activity, parathyroid hormone (PTH) responsiveness, and osteocalcin and Cbfa1 expression levels.(17–20) Furthermore, recombinant human BMP-2 (rhBMP-2) has shown clinical efficacy as a therapy to enhance fracture healing, spinal fusion, and alveolar bone loss.(21–23) The potential anti-GC action of BMP-2 was demonstrated by a recent study in which rhBMP-2 counteracted the GC inhibition of blade-width fracture healing in rabbits.(24) In rat calvarial primary osteoblast cultures, rhBMP-2 overcame the GC inhibition of DNA and collagen synthesis, but this study did not describe the BMP-2 effects on gene expression or mineralization.(25)
In this study, we used MC3T3-E1 osteoblastic cells, in which administration of GC commencing at a defined commitment stage strongly inhibits osteoblast differentiation and calcium deposition.(5) We asked what inhibitory effects of GC are counteracted by BMP-2. We report that administration of rhBMP-2 to GC-inhibited MC3T3-E1 cultures restores several osteoblastic functions, including mineralization. However, other phenomena, such as the differentiation-related cell cycle, the accumulation of collagen, and the development of three-dimensional nodules remain inhibited by GC despite treatment with rhBMP-2. In addition, we describe the surprising inhibition of mineralization when rhBMP-2 is administered alone.
MATERIALS AND METHODS
To maintain the MC3T3-E1 cell line, penicillin/streptomycin and α-minimal essential medium were obtained from Invitrogen Corp. (Carlsbad, CA, USA). Individual lots of FBS, also from Invitrogen, were selected based on their ability to support mineralization. Ascorbic acid, β-glycerophosphate, and dexamethasone were from Sigma Chemical Co. (St Louis, MO, USA). rhBMP-2 was generously provided by Wyeth Research (Cambridge, MA, USA). Normal cell culture dishes were purchased from Corning Incorporated (Corning, NY, USA), and collagen-coated plates were from Becton-Dickinson (Franklin Lakes, NJ, USA). The biochemical assays were conducted using Sircol assay kit (Biocolor Ltd., Newtonabbey, North Ireland) for collagen, Sigma Diagnostics Kit 587 for calcium, Sigma Diagnostics Kit 104-LL for ALP, diaminobenzoic acid (DABA; Sigma Chemical) for DNA, and MicroBCA assay kit (Pierce Biochemical, Rockford, IL, USA) for protein. The histological assays made use of the following reagents: Alizarin red (Sigma), silver nitrate (Sigma), Sirius red (a.k.a. Direct Red 80; Aldrich Chemical Co., Milwaukee, WI, USA), and picric acid (Aldrich Chemical Co.). RNA isolation was accomplished using TRIzol reagent from Invitrogen Corp. The Ambion DNA-free kit (Austin, TX, USA) was used to remove DNA before reverse transcription (RT) with the Thermoscript RT-PCR system (Invitrogen).
A robustly mineralizing subclone of the MC3T3-E1 cell line that has been previously described was used in this study.(5) Cells were plated at 30,000 cells/cm2 in 12-well plates for histological and biochemical assays and in 100-mm plates for cell cycle, Northern blot, RT-polymerase chain reaction (PCR), Fourier transform infrared spectroscopy (FTIR), and X-ray diffraction (XRD) analyses. Cells were maintained in α-modified essential medium with 10% FBS and 1.5% penicillin/streptomycin. Starting at 80% confluency (typically day 3; day 0 = plating day), the culture medium was supplemented with 50 μg/ml ascorbic acid and 10 mM β-glycerophosphate to support differentiation, as well as dexamethasone (1 μM) and/or rhBMP-2 (10 or 100 ng/ml) as indicated.
For Alizarin red staining of calcium, culture wells were washed once in PBS and fixed for 1 h at 4°C in 70% ethyl alcohol. The Alizarin red solution (40 mM, pH 4.2) was filtered through Whatman paper and applied to the fixed wells for 10 minutes at room temperature. Nonspecific staining was removed by several washes in water. For von Kossa staining of phosphate, the cultures were washed once in PBS and fixed with 2% fresh paraformaldehyde for 10 minutes at 4°C. The wells were washed with 0.1 M cacodylic buffer, and then immediately covered with filtered silver nitrate solution (3 mg/100 ml H2O) and exposed to ultraviolet light for 15 minutes. The Sirius red staining for collagen was conducted according to a published procedure.(26) Briefly, cells were washed once in PBS, fixed for 1 h at room temperature in Bouin's fluid, and washed and air dried. The Sirius red dye (1 mg/ml in saturated aqueous picric acid) was added to cover each well and incubated for 1 h under mild shaking.
Biochemical assays for collagen, calcium, DNA, protein, and ALP
To study collagen accumulation, cell layers were collected by scraping in PBS, centrifuged, and hydrolyzed in 0.5 M acetic acid for 18 h at 4°C. The acid extract, which accounted for most of the initial pellet volume, was reacted with Sircol reagent, and the collagen-bound dye was quantitated according to the manufacturer's protocol. Cell layers for the other biochemical assays were collected by scraping in a 10 mM Tris-saline buffer (pH 7.2) containing 0.2% Triton X-100. An aliquot was removed for protein and ALP assays, and the remaining material was incubated in 0.5 M HCl (final concentration) at 70°C for 15 minutes, followed by spectrophotometric calcium assay using Sigma Kit 587 (575 nm) and fluorometric DNA assay(27) using DABA (excitation: 400 ± 15 nm, emission: 485 ± 10 nm). Protein content and ALP activity in the Triton extract were measured spectrophotometrically using the Micro BCA assay kit (562 nm) and Sigma kit 104-LL (410 nm), respectively.
Cell cycle analysis
Cell cycle profiles were obtained as previously described.(5) Briefly, cells were trypsinized and collected by centrifugation (4000 rpm, 4°C, 15 minutes) in PBS. After centrifugation, the cells were resuspended in 100% ethyl alcohol and kept at −20°C. The ethyl alcohol was removed by centrifugation (4000 rpm, 4°C, 15 minutes), and the cells were resuspended in Hank's balanced buffer solution containing 20 mg/ml of propidium iodide and 100 μg/ml of the DNase-free ribonuclease A and incubated for 30 minutes at room temperature. The percentage of cells in each of the S, G1, and G2/M phases of the cell cycle was determined by flow cytometry using the EPICS Profile Analyzer (Coulter Corp., Miami, FL, USA).
Northern blot analysis
RNA was isolated using TRIzol reagent according to the manufacturer's protocol. For Northern analysis, RNA (15 μg) was electrophoretically resolved in a 1% agarose/formaldehyde gel and blotted onto a positively charged nitrocellulose membrane by vertical capillary transfer in 10× SSC buffer. The membrane was prehybridized (5× SSC, 10× Denhardt's, 1% SDS, 50% formamide, 0.1 mg/ml denatured salmon sperm DNA) overnight at 42°C and hybridized in the buffer as detailed above with 70 ng of the denatured radiolabeled osteocalcin probe (107 cpm) at 42°C for 20 h. The probe consisted of a 270-bp fragment of the murine osteocalcin gene 2 (OG2) cDNA labeled with [α-32P]dCTP.(28) To remove nonspecifically bound probe, the membrane was washed four times at room temperature (15 minutes each; 2× SSC, 0.1% SDS), one time at 50°C (30 minutes; 0.5× SSC, 0.1% SDS), and one time at 55°C (30 minutes; 0.1× SSC, 0.1% SDS). Hybridization was visualized by autoradiography at −80°C using Fuji film.
RNA isolated as above was treated with DNase before RT. RT was carried out on 2 μg of denatured RNA using the Thermoscript RT-PCR system according to the manufacturer's protocol. Synthesis of cDNAs was primed with random hexamers. The cDNA was PCR-amplified (MJ Research, Inc., Waltham, MA, USA) in a total volume of 25 μl, which contained 1 μCi (800 Ci/mmol) of [α-32P]dCTP (Perkin Elmer, Boston, MA, USA). For each of the analyzed genes, preliminary experiments were performed with increasing amounts of input cDNA to achieve close-to-linear conditions of signal versus input. Equal cDNA input amounts were determined by 18S cDNA amplification. The primer sequences, annealing temperature, and cycle number for the amplification of each gene are listed in Table 1. The amplified PCR products were resolved by electrophoresis in a 5% polyacrylamide gel. After fixation in 10% acetic acid and drying, the gels were exposed to a storage phosphor screen (Molecular Dynamics, Sunnyvale, CA, USA), and the signal associated with the PCR products was quantitated using the STORM 840 Phosphor Analyzer and Image Quant software (Molecular Dynamics). Relative RNA levels for each gene were calculated based on a standard curve obtained using increasing cDNA input. The resultant values for mRNA were corrected for 18S RNA. The corrected values are expressed in relative units, where control is defined as 1.
Table Table 1. Primers and Conditions for RT-PCR
Mineral content and quality, as well as collagen maturity, were analyzed by FTIR. Cell layers were collected in 50 mM ammonium bicarbonate (pH 8.0) 14 days after plating. The samples were lyophilized and analyzed as potassium bromide pellets on a BioRad FTS 40-A spectrometer (BioRad, Cambridge, MA, USA). The spectral data were baseline corrected and analyzed using GRAMS/386 software (Galactic Industries, Salem, NH, USA) as previously described.(29) Briefly, the mineral-to-matrix ratio was calculated as the ratio between the integrated areas of the phosphate absorbance (900–1200 cm−1) and the protein amide I absorbance (1600–1700 cm−1). The collagen maturity was determined from the areas of the 1660- and 1690-cm−1 bands in the amide I spectral peak.(30) The 1660:1690 ratio corresponds to the ratio of nonreducible to reducible collagen cross-links in bone. Crystallinity, or crystal maturity, was assessed by comparing the ratio of the 1030:1020 sub-bands of the phosphate absorbance spectrum.(31)
The nature of the mineral present was determined by XRD. For crystal size determination, lyophilized tissue culture samples were ground to a uniform size and placed on a graphite sample holder. A Brucker D8 powder diffractometer (Brucker XAS, Madison WI, USA) was used to scan the samples from 20° and 40° (2θ) using Ni-filtered Cu-Kα radiation.(32) The peak at 25.85° (2θ) was used to calculate the particle size in the c axis direction using the Scherrer equation.(32)
The results from each quantitative assay were analyzed to test three effects: (1) the effect of DEX in the absence of BMP-2; (2) the effect of BMP-2 in the absence of DEX; and (3) the effect of BMP-2 in the presence of DEX. The Mann-Whitney test was used to analyze the compiled data from three independent RT-PCR experiments. For all other experiments, means were compared using the Student's t-test. Differences for both the Mann-Whitney and t-tests were considered significant when p ≤ 0.05. Statistical analyses were done using GraphPad InStat version 3.0a for Macintosh (GraphPad, San Diego, CA, USA).
BMP-2 rescues calcium deposition in GC-inhibited cultures
We have previously defined a commitment stage in MC3T3-E1 osteoblast cultures, which occurs during the 2–3 days that follow confluency and is characterized by a unique differentiation-related cell cycle.(5,33) When GCs are administered to MC3T3-E1 cultures commencing at this commitment stage (but not later), both the unique cell cycle and calcium deposition are reversibly inhibited.(5) In this study, we asked whether the osteogenic growth factor BMP-2 would counteract the GC inhibition of osteoblast differentiation. Initially, we tested if rhBMP-2 would (1) rescue calcium deposition in GC-inhibited cultures, (2) promote normal cell condensation and nodule formation in the presence of GC, and (3) restore the differentiation-related cell cycle.
MC3T3-E1 cells were treated with DEX and/or rhBMP-2 commencing just before confluency (day 3) and until the end of the experiment (day 14). As demonstrated by Alizarin red staining, calcium deposition was completely inhibited by DEX but was rescued by co-treatment with rhBMP-2 at both 10 and 100 ng/ml (Fig. 1A). Interestingly, when administered alone, rhBMP-2 inhibited calcium deposition, especially at the 100 ng/ml dose (Fig. 1A).
Cultured osteoblasts typically differentiate in condensed areas and form three-dimensional nodules where mineral is later deposited. Microscopic examination of Alizarin red-stained cultures (Fig. 1A, right panel) and of unstained day 11 and 14 cultures (Fig. 1B) revealed relatively large, well-defined individual cells in the BMP-2-rescued cultures, a morphology more typical of the DEX treatment (Fig. 1B). Additionally, the BMP-2-rescued cultures did not form three-dimensional nodules. Cells treated with 100 ng/ml rhBMP-2 alone had a very well-defined cobblestone appearance, similar to the DEX-treated cultures, and did not form nodules (Fig. 1B). Thus, the anti-GC action of rhBMP-2 on calcium deposition was not associated with rescue of condensation or nodule formation.
BMP-2 has been reported to be either mitogenic(12,25,34) or antimitogenic.(12,35–37) Having previously demonstrated GC inhibition of a differentiation-related cell cycle,(5) we asked how this cell cycle would be affected by BMP-2. Cultures were treated with DEX and/or rhBMP-2 commencing on day 3, and cell cycle profiles were determined on day 6, when cells undergoing commitment still display the postconfluent cell cycle.(5) As shown in Fig. 1C, rhBMP-2 did not rescue the cell cycle, which was inhibited nearly 2-fold by DEX. In fact, the downregulated cell cycle was further suppressed by co-treatment with rhBMP-2, especially at the 100 ng/ml dose, where the percentage of cells in S and G2/M was reduced by 24% compared with DEX alone (p < 0.05). When administered alone, rhBMP-2 also had an antimitogenic effect, reducing the percentage of cells in S and G2/M by 30% at both the 10 and 100 ng/ml concentrations (p < 0.05). Thus, BMP-2 rescued mineralization in the DEX-inhibited cultures (Fig. 1A) in a manner that circumvented nodule formation (Fig. 1B) and the differentiation-related cell cycle (Fig. 1C).
Inhibition of collagen accumulation by GC and BMP-2
Next, we asked how GC and BMP-2 affected collagen accumulation. Cultures were treated with DEX and/or rhBMP-2 starting on day 3 as above, and collagen and DNA were measured from days 4 to 14. Collagen per DNA in the control cultures increased in a linear fashion from day 4 to day 14 (Fig. 2A, top). DEX suppressed collagen/DNA by 75–90% (p < 0.005) at all time points, while inhibiting DNA accumulation by only 20–30% (p < 0.05; Fig. 2A, bottom). Unlike the rescue of mineralization (Fig. 1A), rhBMP-2 did not rescue collagen accumulation in the DEX-inhibited cultures (Fig. 2A, top). Surprisingly, rhBMP-2 alone dose-dependently inhibited collagen accumulation per DNA at all time points, and this inhibition amounted to 70%, with 100 ng/ml rhBMP-2 on day 14 (Fig. 2A, top). The inhibition of collagen accumulation in cultures treated with DEX and/or rhBMP-2 was corroborated in situ by Sirius red histological staining of day 14 cultures (data not shown).
The properties of the extracellular matrix were then evaluated by FTIR of day 14 cultures. Collagen maturity, reflected by the 1660:1690 ratio,(30) was not affected by any of the treatments (data not shown). However, DEX decreased the mineral-to-matrix ratio by 6-fold (Fig. 2B), suggesting that the inhibition of mineral deposition (Fig. 1A and see Fig. 3A) exceeded the inhibition of collagen accumulation (Fig. 2A). Co-treatment of DEX cultures with rhBMP-2 at either 10 or 100 ng/ml restored the mineral-to-matrix ratio to levels greater than those observed in the control cultures (p = 0.0001; Fig. 2B). This restoration is attributable to both the increased mineral (Fig. 1A) and the decreased collagen (Fig. 2A). Interestingly, rhBMP-2 alone also increased the mineral-to-matrix ratio in the absence of DEX (Fig. 2B), suggesting in this case that the inhibition of collagen accumulation (Fig. 2A) exceeded the inhibition of mineral deposition (Fig. 1A; see Fig. 3A).
The inhibition of collagen accumulation by BMP-2 (Fig. 2A) was surprising because BMP-2 has been shown to increase type I collagen mRNA levels.(35,38) Consistent with these reports, RT-PCR analysis showed that rhBMP-2 increased α1(I) mRNA levels in day 6 MC3T3-E1 cultures (Figs. 2C and 2D). In three independent experiments, 10 ng/ml rhBMP-2 significantly increased α1(I) mRNA in the absence of DEX by an average of 3.2-fold compared with control (Fig. 2D). In the presence of DEX, 100 ng/ml rhBMP-2 significantly increased α1(I) mRNA by an average of 12-fold compared with DEX alone (Fig. 2D). These results suggest that the inhibition of extracellular collagen accumulation by rhBMP-2 (Fig. 2A) is post-transcriptional.
Because BMP-2 rescued the GC inhibition of calcium deposition without restoring collagen accumulation, we hypothesized that the GC inhibition of calcium deposition was not dependent on inhibition of collagen accumulation. To address this hypothesis, we tested the ability of DEX to inhibit mineral deposition by MC3T3-E1 cells plated onto type I collagen-coated wells. As shown in Figs. 2E and 2F, DEX inhibited calcium deposition in collagen-coated wells (−93%) to the same extent as in uncoated plastic wells (−90%). In addition, consistent with previous reports,(39) mineralization of the control cultures in the collagen-coated wells was enhanced compared with uncoated plastic wells. However, the presence of collagen did not rescue the deposition of calcium in the DEX-treated cultures (Figs. 2E and 2F), suggesting that the inhibition of mineral deposition was independent of inhibition of collagen accumulation.
Kinetics of BMP-2-induced biochemical markers of osteoblast differentiation in GC-treated cultures
Taken together, the data in Figs. 1 and 2 suggest that mineralization in GC-treated cultures is rescued by BMP-2 without restoration of the differentiation-related cell cycle, nodule formation, or collagen accumulation. The fundamental differences between the normal and rescued cultures prompted us to study the kinetics of calcium deposition and ALP activity in the presence of DEX and/or rhBMP-2. As shown in Fig. 3A, the control cultures started depositing calcium on day 10. At this and each of the following time points, DEX inhibited the calcium content per DNA by 96–98%, far more than the inhibition of DNA accumulation (Fig. 2, bottom). rhBMP-2 rescued the DEX-inhibited calcium deposition, yielding a profile closely resembling that of control cultures. The 10 ng/ml dose of rhBMP-2 reversed the DEX-mediated inhibition of calcium deposition by 22% on day 10 and by 43% on day 14. Similarly, rhBMP-2 at 100 ng/ml reversed the DEX-mediated inhibition by 33% and 54% on days 10 and 14, respectively (Fig. 3A). Consistent with the Alizarin red staining (Fig. 1A), rhBMP-2 treatment in the absence of DEX decreased calcium deposition (Fig. 3A).
ALP activity, a common marker of osteoblast differentiation, was detectable in the control cultures as early as day 4. ALP activity per DNA gradually increased, reaching a maximum level on day 10 (Fig. 3B), which corresponded with the onset of calcium deposition (Fig. 3A). DEX significantly inhibited ALP activity per DNA from day 8 (−44%) through day 14 (−27%). rhBMP-2 counteracted the inhibitory effect of GC at all time points and increased ALP activity to levels that equaled or surpassed those observed in the control cultures (Fig. 3B). Administration of rhBMP-2 in the absence of DEX also stimulated ALP activity (Fig. 3B), which was in sharp contrast to its inhibitory effect on calcium deposition.
Characterization of mineral deposited in GC/BMP-2 co-treated cultures
Despite the similar kinetics of ALP activity and calcium deposition between the control and the GC/BMP-2-rescued cultures, the differences in morphology, cell cycle, and collagen prompted us to further analyze the rescued mineral, first by von Kossa staining, to assess phosphate content and then by XRD and FTIR to assess the nature of the mineral present and its crystal size and perfection (crystallinity).
As demonstrated by von Kossa staining of representative cultures, GC inhibited and rhBMP-2 rescued phosphate deposition (Fig. 4A) similar to their respective effects on calcium (Fig. 1A). In the absence of DEX, rhBMP-2 inhibited the deposition of phosphate (Fig. 4A), resembling the inhibition of calcium accumulation (Fig. 1A).
The XRD patterns for all cultures with mineral were characteristic of bone-like hydroxyapatite (data not shown). The XRD analysis also indicated that the control and rescued apatite crystals were of a similar size (data not shown). Crystallinity of the mineral, assessed by FTIR,(31) was also similar in the control cultures, the DEX/rhBMP-2-co-treated cultures, and the cultures treated with rhBMP-2-alone (Fig. 4B). A slight but significant decrease in crystallinity was noticed in cultures treated with 10 ng/ml rhBMP-2 alone. In the DEX-inhibited cultures, there was not sufficient mineral to analyze crystallinity.
Induction of osteoblast-phenotypic genes by BMP-2 in GC-treated cultures
Although osteocalcin (OC) does not seem to play a direct role in osteoblast differentiation,(40) OC gene expression serves a useful molecular tool because it is tightly and specifically associated with this process.(41) We therefore asked whether the BMP-2-rescued mineralization in GC-treated cultures was associated with increased OC gene expression. MC3T3-E1 cells were treated with DEX, rhBMP-2, or both from days 3 to 6, at which time OC mRNA was evaluated by both Northern blot analysis (Fig. 5A) and RT-PCR (Figs. 5B and 5C). By both methods, DEX dramatically inhibited OC mRNA, and rhBMP-2 restored it to levels higher than control (Fig. 5). As determined by incorporation of radionucleotides during the RT-PCR reaction, DEX inhibited OC mRNA by an average of 6.3-fold, and rhBMP-2 at 10 and 100 ng/ml induced OC mRNA to levels 27- and 62-fold greater, respectively, compared with DEX alone (Fig. 5B). Interestingly, administration of 10 and 100 ng/ml rhBMP-2 in the absence of DEX increased OC mRNA levels by 24- and 128-fold, respectively (Fig. 5C), despite the inhibition of mineralization (Figs. 1A, 3A, and 4A).
Unlike OC, the transcription factor Cbfa1 is both associated with and necessary for osteoblast differentiation.(42–44) Furthermore, an increase in Cbfa1 gene expression in response to BMP-2 precedes induction of bone phenotypic genes in osteoblast precursors.(20,45–47) We therefore asked whether the BMP-2-rescue of mineralization in GC-inhibited MC3T3-E1 cells was accompanied by upregulation of Cbfa1. Figure 6 presents analysis of two major Cbfa1 mRNA isoforms, type I and type II.(48–50) Addition of rhBMP-2 to the DEX-treated cultures increased mRNA levels for both CBFA1 type I and type II. rhBMP-2 at 100 ng/ml significantly increased the type I mRNA by an average of 3.7-fold and the type II mRNA by 2-fold (Fig. 6B). However, rhBMP-2 similarly increased Cbfa1 mRNA levels in the absence of DEX (Fig. 6), where it did not stimulate, but in fact inhibited, calcium and phosphate deposition. In addition, and consistent with previous studies,(51) DEX itself had little effect on Cbfa1 mRNA levels despite its strong inhibition of mineralization. Thus, Cbfa1 mRNA levels cannot explain the opposing effects of GC and BMP-2 on mineralization.
Rescue of mineralization by exogenous BMP-2 occurs under conditions of inhibited endogenous BMP-2 expression
The BMP-2-mediated inhibition of mineral deposition in the absence of DEX versus the opposite effect observed in the presence of DEX (Figs. 1A, 3A, and 4A) was unexpected and intriguing. Some TGF-β family members, such as activin and BMP-4, elicit specific responses in target cells only when present within a strict concentration range.(52–54) Because MC3T3-E1 cells produce BMP-2, which is essential for their differentiation,(55) we speculated that rhBMP-2 inhibited mineralization in the absence of DEX because the existing BMP-2 signal was supplemented beyond the effective range; in the DEX-treated cultures, endogenous BMP-2 could have been downregulated so that addition of exogenous rhBMP-2 re-established an effective BMP signal. To begin addressing this hypothesis, we assessed by RT-PCR the effect of DEX on BMP-2 mRNA levels in MC3T3-E1 cultures. Indeed, as shown in Fig. 7, DEX dramatically inhibited the expression of endogenous BMP-2; the average repression of BMP-2 mRNA in three independent experiments was 100-fold (Fig. 7B). Interestingly, exogenous rhBMP-2 significantly stimulated endogenous BMP-2 gene expression in DEX-treated cultures (by an average of 73- and 99-fold for the 10 and 100 ng/ml doses, respectively) while inhibiting gene expression in the absence of DEX (by an average of 1.8- and 2.6-fold, respectively; Fig. 7). These results further support the notion that rhBMP-2 treatment of control cultures resulted in an excessive, antiproductive signal, whereas the same treatment was productive in the GC-treated BMP-2-deficient cultures.
Bone loss during long-term GC treatment has been attributed to decreased osteoblast number and function.(1,2,6) At the cell and molecular level, GCs have been shown to promote osteoblast apoptosis, impede cell cycle progression, reduce nodule formation, and downregulate the expression of osteoblastic genes such as Cbfa1, α1(I) collagen, ALP, and OC.(1,2,6,), 56 In osteoblast cultures, GCs exhibit both negative and positive effects, depending on the steroid concentration, cell maturity, and species. In this study, we used MC3T3-E1 cells, in which mineralized matrix formation is strongly inhibited by GC.(5,57) It should be noted that in MC3T3-E1 cells, like in rat calvarial osteoblasts,(58) GCs do not promote, but in fact inhibit, apoptosis (data not shown). Thus, our MC3T3-E1 culture model provides an opportunity to study the inhibitory effects of GCs on osteoblast cell cycle and differentiation separate from apoptosis.
Co-treatment with rhBMP-2 reversed many of the inhibitory effects of DEX in MC3T3-E1 cultures. Most notably, rhBMP-2 dramatically restored mineralization as demonstrated by Alizarin red and von Kossa staining and by quantitative analysis of calcium accumulation over time. Furthermore, the rescued mineral had similar crystallinity to that of the control cultures. In addition, rhBMP-2 restored the mineral-to-matrix ratio, which was decreased in the DEX-treated cultures because the steroid inhibited mineral deposition to a greater extent than collagen accumulation.
Administration of rhBMP-2, either to GC-treated or to untreated cultures, increased the mRNA levels for Cbfa1, α1(I) collagen, and OC, and stimulated ALP activity. GCs have been shown to decrease all of these osteoblastic markers.(1,2,7,), 8 However, none of the these BMP-2-induced genes alone seem to have mediated the rescue of mineralization, because (1) rhBMP-2 stimulated all of these genes not only in the presence of DEX, where it stimulated mineralization, but also in the absence of DEX, where it inhibited mineralization; (2) although BMPs stimulated Cbfa1 gene expression in this and other studies,(20,45–47) the downregulation of Cbfa1 by GC is questioned by both the present study and by the report of Prince et al.(51); and (3) the rhBMP-2 stimulation of α1(I) collagen mRNA was not translated into normal levels of extracellular collagen accumulation. The only molecular parameter that consistently paralleled mineralization in our study was the level of endogenous BMP-2 mRNA. However, stimulation of endogenous BMP-2 gene expression cannot explain how rhBMP-2 rescued mineralization. We are currently performing microarray gene expression studies to discover BMP-2 target genes that play a role in rescuing mineralization in GC-treated osteoblasts.
Despite the restoration of mineralization, there were clear differences between the control and the GC-treated/BMP-2-rescued cultures. In the latter, the differentiation-related cell cycle was not restored, nodules were not formed, and collagen accumulation was still inhibited, resulting in increased mineral-to-matrix ratio compared with control cultures. Thus, while considering BMP-2-mimetics in the development of novel therapies for glucocorticoid-induced osteoporosis, attention must be paid to the inhibitory effects of GC that are not restored by BMP-2.
When administered alone, the stimulatory effects of rhBMP-2 on Cbfa1 mRNA, α1(I) collagen mRNA, OC mRNA, and ALP activity were surprisingly accompanied by negative effects on extracellular collagen accumulation and mineralization, the former being more severe. Given the BMP-2-mediated stimulation of α1(I) collagen mRNA, the inhibition of collagen accumulation could occur at the level of translation, post-translational processing, tropocollagen assembly, secretion, or extracellular stability. These collagen results are reminiscent of a recent study on OC, describing increased OC mRNA but decreased OC secretion in BMP-2-treated bone marrow stromal cells.(20) The inhibitory effect of rhBMP-2 on mineralization is intriguing, given the well-known osteoinductive property of BMP-2.(10,12,15,), 16 A possible explanation is that in our study, unlike other reports, rhBMP-2 was administered to relatively differentiated osteoblasts, resulting in the elevation of a pre-existing BMP signal(55) to levels outside a strictly defined effective range.(52–54) The rescue of mineralization in the DEX/rhBMP-2 co-treated cultures shows that under some conditions GCs may be permissive to the osteoinductive action of BMP-2, possibly explaining the frequently reported positive action of GC in cultured osteoblasts and the synergism with BMPs.(19,38) From a clinical perspective, the bimodal action of BMP-2, stimulatory in some circumstances (e.g., high GCs, undifferentiated cells) and inhibitory in others (e.g., low GCs, more differentiated cells), suggest that local BMP therapy may yield the best results when administered for a limited, rather than extended, period of time, because excessive treatment may result in negative effects. Indeed, the inhibition of experimental osteotomy healing by GCs was counteracted by a single rhBMP-2 application.(24) Furthermore, mineralization in GC-inhibited MC3T3-E1 cultures is rescued by as little as 6-h exposure to rhBMP-2 (our preliminary results). However, given the differential response of osteoblasts (and chondrocytes) at various stages of differentiation to BMPs, the significance of our findings to bone remodeling, fracture healing, and GC-induced osteoporosis requires further exploration and validation in vivo.
The present findings on GC/BMP-2 interactions in osteoblasts are important in two distinct ways. First, toward unraveling mechanisms of GC-induced osteoporosis, we demonstrated that inhibition of endogenous BMP-2 gene expression(59,60) is a critical adverse effect of GC in osteoblasts. Although recovery from the GC inhibition was incomplete, the exogenously administered rhBMP-2 remarkably rescued mineralization. In this regard, the critical role of BMP signaling in MC3T3-E1 osteoblasts has been illustrated by the strong inhibition of differentiation markers following Noggin administration.(55) Second, toward better understanding of BMP's osteogenic activity, our study suggests that the actions of BMPs during osteoblast differentiation are tightly monitored, similar to their actions as morphogens during embryonic development. We found that treatment with rhBMP-2 induced mineralization when endogenous BMP-2 was inhibited but not in control cultures where BMP-2 was already expressed. Furthermore, endogenous BMP-2 gene expression was increased when rhBMP-2 was administered to GC-inhibited cells, suggesting a positive feedback loop, whereas the opposite effect was observed in the absence of GC, suggesting a negative feedback loop. Positive BMP-2 autoregulation has been previously suggested to occur at the level of transcription.(61) The details of the BMP-2 positive/negative feedback loops may have important implications in future use of BMP-2 and BMP-2-mimetics in GC-induced osteoporosis and other skeletal disorders.
The authors thank Wyeth Research for generously providing rhBMP-2 for this project and Gerard Karsenty (Baylor College of Medicine) for the OC cDNA. These studies were supported by National Institutes of Health Grants AR47052, AR037661, DE04141, and DE07211 and grants from the Arthritis Foundation.