Inhibition of Osteoblast-specific Transcription Factor Cbfa1 by the cAMP Pathway in Osteoblastic Cells

The cAMP pathway, a major intracellular pathway mediating parathyroid hormone signal, regulates osteoblastic function. Parathyroid hormone (through activation of protein kinase A) has also been shown to stimulate ubiquitin/proteasome activity in osteoblasts. Since the osteoblast-specific transcription factor Osf2/Cbfa1 is important for differentiation of osteoblastic cells, we examined the roles of the cAMP and ubiquitin/proteasome pathways in regulation of Cbfa1. In the osteoblastic cell line, MC3T3-E1, continuous treatment with cAMP elevating agents inhibited both osteoblastic differentiation based on alkaline phosphatase assay and DNA binding ability of Cbfa1 based on a gel retardation assay. Cbfa1 inhibition was paralleled by an inhibitory effect of forskolin on Cbfa1-regulated genes. Northern and Western blot analyses suggested that the inhibition of Cbfa1 by forskolin was mainly at the protein level. Pretreatment with proteasome inhibitors prior to forskolin treatment reversed the effect of forskolin. Furthermore, addition of proteasome inhibitors to forskolin-pretreated samples resulted in recovery of Cbfa1 protein levels and accumulation of polyubiquitinated forms of Cbfa1, indicating a role for the proteasome pathway in the degradation of Cbfa1. These results suggest that suppression of osteoblastic function by the cAMP pathway is through proteolytic degradation of Cbfa1 involving a ubiquitin/proteasome-dependent mechanism.

The osteoblast-specific factor-2/core binding factor ␣-1 (Osf2 1 /Cbfa1),hereafter called Cbfa1, is a transcription factor recently shown to be a master regulator of osteoblastic differentiation. Komori et al. (1) and Otto et al. (2) observed that targeted disruption of osteoblast-specific Cbfa1 in vivo blocks skeletogenesis, and that heterozygous mutations of Cbfa1 in both human and mice lead to cleidocranial dysplasia, an autosomal dominant disorder (2)(3)(4). As further evidence of its role in differentiation, Ducy and colleagues (5) showed that forced expression of Cbfa1 induced expression of principal osteoblast-specific genes such as bone sialoprotein, osteopontin, and osteocalcin in non-osteoblastic cells. Furthermore, Ducy et al. (5) and Banerjee et al. (6) demonstrated that disruption of Cbfa1 by antisense oligonucleotides in osteoblast cultures inhibited expression of osteoblastic differentiation markers and formation of mineralized nodules.
The cAMP pathway, the major mediator of PTH signaling in osteoblasts (7), can have either anabolic or catabolic effects (8,9) depending on the cell line used, the stage of differentiation, culture conditions, and treatment duration (10 -12). PTH/ cAMP have been shown to regulate expression of several osteoblastic differentiation genes including Cbfa1-regulated osteoblast-specific genes such as osteopontin (13), type I collagen (14,15), bone sialoprotein (16), and osteocalcin (11). Although sequences conferring cAMP responsiveness have been found in the promoters of osteocalcin (8) and bone sialoprotein (16), it remains largely unknown which transcription factors mediate the cAMP response in these and other osteoblastic differentiation gene promoters.
The 26 S proteasome, a multicatalytic protease complex, is a site of regulatory degradation for most proteins (17,18). It is an ATP-dependent extralysosomal protease that is necessary for viability and essential for regulation of proliferation and differentiation of eukaryotic cells (18). It is ubiquitously distributed throughout the cell, in the nucleus, cytosol, endoplasmic reticulum, and even associated with the cytoskeleton (18). It catalyzes degradation of rate-limiting enzymes (e.g. ornithine decarboxylase), transcriptional regulators (e.g. IB), cell-cycle proteins (e.g. cyclins), abnormal proteins, and turnover of membrane proteins (18). Substrate proteins are tagged with multiple copies of ubiquitin for recognition by the proteasome complex. In osteoblasts, PTH, through the cAMP pathway, has been shown to stimulate 26 S proteasomes (19), which account for 90% of the neutral regulatory proteolysis in eukaryotic cells (20). In yeast, the involvement of proteasomes in protein turnover has been studied using various mutants of the proteasome complex. In mammalian cells, which lack such mutants, its role has been studied mainly using peptide inhibitors (17,18). Although many aspects of the ubiquitin/proteasome pathway have been elucidated, the endogenous substrates it regulates in osteoblastic cells remain largely unknown.
Since it has been shown that cAMP regulates osteoblastic differentiation, and that Cbfa1 is essential for osteoblastic differentiation, we hypothesized that cAMP regulation of osteoblastic differentiation occurs through Cbfa1 activity. In addition, since cAMP also stimulates ubiquitin/proteasome activity in osteoblasts, we examined the role of the ubiquitin/proteasome pathway in regulation of Cbfa1 by cAMP. In the present study, we tested the effect of cAMP elevating agents on Cbfa1 expression and function in MC3T3-E1, a non-transformed osteoblastic cell line established from newborn mouse calvaria which has a capacity to undergo osteoblastic differentiation and mineralization in vitro. Results showed that continuous treatment of cAMP agonists inhibited Cbfa1 by proteolytic degradation involving the ubiquitin/proteasome-dependent mechanism.
Oligonucleotides were annealed and gel purified on 20% native polyacrylamide gels. Purified double-stranded probes were end-labeled with [␥-32 P]ATP by using T4 polynucleotide kinase. Nuclear extracts were prepared as described previously (23). 3-5 g of nuclear extracts were incubated with binding buffer containing 20 mM Tris-HCl (pH 8), 10 mM NaCl, 3 mM EDTA, 0.05% Nonidet P-40, 5 mM dithiothreitol, 5% glycerol, and 1 g of poly(dI-dC)⅐poly(dI-dC) (5) for 15 min, after which labeled probe was added and incubated for additional 15-20 min. The samples were subjected to electrophoresis at room temperature on a chilled 5% polyacrylamide gel in Mini-PROTEAN II apparatus (Bio-Rad) in 0.5% TBE at 400 V for 14 min. For experiments using anti-Cbfa1 antibody, the antibody was added to the reaction mixture for 30 min prior to addition of labeled probe. The samples were subjected to electrophoresis at room temperature on 5% polyacrylamide gel in 0.25% TBE at 160 V for 40 min. In competition experiments, competitor oligos were incubated with nuclear extracts during the first 15 min prior to addition of the labeled probe. Autoradiographs were scanned with AGFA ARCUS II scanner and semi-quantitated with NIH Image software, version 1.49, a public domain program.
Alkaline Phosphatase Activity Assay-Cells were seeded in 24-well plates and treated at subconfluence with vehicle alone (Me 2 SO) or forskolin for the indicated number of days and alkaline phosphatase activity was assayed as described previously (24). The alkaline phosphatase activity was normalized to total protein concentration determined using the Bradford (Bio-Rad) assay. The data were from a representative of two experiments, and shown as the mean Ϯ S.D. of quadruplicate wells. 45 Ca Incorporation Assay and von Kossa Staining-Ca 45 incorporation and von Kossa staining to detect mineralization were performed as described previously (24). The data for 45 Ca incorporation were from a representative of two experiments and shown as the mean Ϯ S.D. of six wells.
RNA Isolation and Northern Blot Analysis-Cells, grown in duplicate 60-mm dishes, were treated at subconfluence with vehicle alone or 25 M forskolin for the indicated period, and total RNA was isolated using an RNA isolation kit (Stratagene, La Jolla, CA). Total RNA (10 g) from duplicate samples was run on 1% agarose/formaldehyde gels and Northern analyses for osteopontin, 28 S rRNA were performed as described previously (25).
RT-PCR-The total RNA (3 g) isolated as described above was reversed-transcribed and amplified by PCR as described previously (25). For semiquantitative RT-PCR, the linear range was first established and amplification cycles were chosen to be within this range.
Western Blot Analysis-Nuclear extracts (5 g) were run on 12% Tris glycine gel (Novex, San Diego, CA) and electrotransferred to nitrocellulose membrane overnight at 4°C. The blots were probed with anti-Cbfa1 antibody at 1:500 dilution for 2 h at room temperature. The cbfa1 protein was detected by enhanced chemiluminescence according to manufacturer's recommendations (ECL, Amersham Pharmacia Biotech).
Immunoprecipitations-15 g of nuclear extracts were diluted to 300 l in a dilution buffer (10 mM Tris HCl, pH 8.0, 150 mM NaCl, 0.1% Triton X-100, 0.025% sodium azide, 0.1% bovine serum albumin), and immunoprecipitated with 4 -6 l of anti-Cbfa1 antibody for 4 -5 h at 4°C. After which time, 50 l of Protein A-Sepharose (CL-4B) (Amersham Pharmacia Biotech) was added and incubated overnight at 4°C. The protein A-antibody-antigen complexes were pelleted and washed twice with the dilution buffer, once with buffer A (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.025% sodium azide), and once with buffer B (50 mM Tris-HCl, pH 6.8). The immunoprecipitated proteins were boiled for 5 min in a sample buffer and separated on a 12% Tris glycine gel. The ubiquitinated Cbfa1 was probed with mouse monoclonal anti-ubiquitin antibody at 1:200 dilution for 2 h at room temperature, and visualized by ECL. The data were from a representative of two separate experiments.

Effect of Forskolin on Osteoblastic Differentiation and Mineralization-
To examine the effect of forskolin on osteoblastic differentiation and mineralization, MC3T3-E1 cells were treated continuously with forskolin and alkaline phosphatase activity, as well as matrix mineralization, were measured over a period of 3, 7, 14, and 21 days. The results showed that alkaline phosphatase activity increased as cells progressed through differentiation in control cells, however, this increase was blocked by forskolin treatment (Fig. 1A). Matrix mineralization, as measured by incorporation of 45 Ca, was also inhibited by forskolin (Fig. 1B). Von Kossa cytochemical stain also showed that forskolin treatment completely inhibited mineralization (data not shown).
Effect of Forskolin on Cbfa1 DNA Binding-Since Cbfa1 has been shown to be an important regulator of osteoblastic differentiation, we examined the effect of the cAMP pathway on Cbfa1 DNA binding via gel retardation assay. Nuclear extracts of MC3T3-E1 cells, treated with 25 M forskolin, 1 mM dibutyryl cAMP (Bt 2 cAMP), or 0.5 M PTH, were incubated with probe containing the Cbfa1 DNA-binding site (5). Results showed that forskolin, Bt 2 -cAMP, and PTH reduced Cbfa1 DNA binding ( Fig. 2A). In contrast, treatment of cells with 1 mM dibutyryl cGMP (Bt 2 -cGMP) or 500 ng/ml epidermal growth factor produced little or no effect (Fig. 2A).
The specificity of Cbfa1 binding to DNA probe was examined using anti-Cbfa1 antibody (22). The addition of anti-Cbfa1 antibody resulted in a supershifted band and a decrease in the faster migrating complex (lanes 1-2 versus lanes 3-4, 1 l of antibody and 5-6, 3 l of antibody; Fig. 2B). The supershifted band was apparently specific to anti-Cbfa1 since an irrelevant antibody (anti-NFB antibody) had no effect (lanes 7-8, Fig.  2B). Cbfa1 DNA binding was also competed by excess unlabeled wild-type probe but not by probe containing mutation in the Cbfa1-binding site (5) (data not shown).
Effect of Forskolin on Expression of Osteoblastic Differentiation Genes-To examine whether forskolin induced changes in Cbfa1 DNA binding correlated with changes in expression of downstream Cbfa1-regulated genes, we determined the effect of forskolin on expression of alkaline phosphatase, bone sialoprotein, osteopontin, and osteocalcin genes. Total RNA was isolated from control and forskolin-treated samples; osteopontin expression was determined by Northern analysis; alkaline phosphatase, bone sialoprotein, and osteocalcin expression were determined by RT-PCR. RNA isolated from 7-day cultured cells was used for analyses of alkaline phosphatase, bone sialoprotein, and osteopontin expression, whereas RNA isolated from 21-day cultured cells was used for analysis of osteocalcin expression. Results showed that forskolin also inhibited the expression of alkaline phosphatase, bone sialoprotein, osteopontin, and osteocalcin genes (Fig. 2C).
Effect of Forskolin on Cbfa1 Expression-To determine whether the inhibition of Cbfa1 DNA binding by forskolin was due to reduced Cbfa1 protein level, nuclear extracts were prepared from cells treated with forskolin for 7 or 14 days and Cbfa1 protein level was determined by Western blot analysis. Results showed that the level of Cbfa1 in control samples, detected as two immunoreactive species of ϳ60and 65-kDa sizes (6), was reduced in both time points of forskolin-treated samples (Fig. 3A). To determine the earliest time required for this effect, Western analysis was repeated with nuclear extracts from cells treated with forskolin for 2-48 h. Results showed that in control samples, Cbfa1 levels increased over this time (compare lanes 1, 3, 5, and 7; Fig. 3A), whereas the Cbfa1 level was reduced in samples treated with forskolin for 48 h (lanes 7 versus 8; Fig. 3A). However, the reduction in Cbfa1 level was not observed prior to 32-h treatment with forskolin (Fig. 3A), indicating that degradation requires a lag time.
We next examined the inhibition of Cbfa1 by forskolin at the mRNA level. Cells were treated with forskolin for the indicated time, total RNA was isolated, and Cbfa1 transcript level was analyzed by Northern analysis using Cbfa1-specific probe (5). Results showed only minor variations in Cbfa1 expression levels with forskolin treatment, except in the samples treated for 7 days, in which the transcript level of Cbfa1 was moderately reduced (Fig. 3B). Nevertheless, these changes do not appear to account for the more dramatic change in Cbfa1 protein level.
Mechanism of Cbfa1 Inhibition by Forskolin-To determine whether the reduction in Cbfa1 protein level in the nucleus was due to its translocation to the cytoplasm, Western analysis was performed using both cytoplasmic and nuclear extracts. Results showed no increase in cytoplasmic Cbfa1 levels with forskolin treatment (data not shown).
Since it has been previously shown that PKA activation in osteoblastic cells stimulates ubiquitination of substrate proteins and proteolytic activity of proteasomes (19), a predominant mechanism of regulatory proteolysis (20), we examined the involvement of the ubiquitin/proteasome pathway in the forskolin-induced reduction of Cbfa1 protein level. Proteasome inhibitors (MG-101, MG-132, or PSI) were added to MC3T3-E1 cells, which had been pretreated with forskolin for 3 days. Nuclear extracts, prepared 6 h after addition of inhibitors, were probed with anti-Cbfa1 antibody. Results showed that, in the absence of inhibitors, forskolin reduced the Cbfa1 level as shown above (lane 2 versus 1; Fig. 4A). However, in the presence of proteasome inhibitors, Cbfa1 was retained at 42-65% of the control level (Fig. 4A), indicating that proteasome inhibitors prevented the forskolin-induced proteolysis of Cbfa1. Similar treatment with proteasome inhibitors for 2 h did not result in the recovery of Cbfa1 (data not shown).
To examine whether inhibition of proteasome prevents forskolin-induced degradation of Cbfa1, cells were pretreated with PSI for 2 h, followed by forskolin treatment for 2 days. Nuclear extracts were probed with anti-Cbfa1 antibody. Results showed that forskolin failed to reduce Cbfa1 levels in the samples pretreated with PSI (compare lanes 2 and 4; Fig. 4B), suggesting a role of proteasomes in proteolytic degradation of Cbfa1 by cAMP.
It has been shown that addition of proteasome inhibitors also results in accumulation of the ubiquitinated form of substrate proteins (19,26,27). Therefore to examine whether Cbfa1 is ubiquitinated in forskolin-treated cells prior to proteasomal degradation, we immunoprecipitated Cbfa1 from the samples that had been treated as in panel A (pretreatment with forskolin for 3 days and subsequently treatment with proteasome inhibitors for 6 h). The immunoprecipitates were separated on a gel and probed with anti-ubiquitin antibody which is reactive to both ubiquitin and ubiquitinated substrate proteins. Results showed an accumulation of the polyubiquitinated Cbfa1 in forskolin-treated cells in the presence of proteasome inhibitors compared with the control samples (no forskolin and no inhibitor) or the samples treated only with forskolin (Fig. 4C). These results suggest that proteasome inhibition leads to accumulation of polyubiquitinated Cbfa1. DISCUSSION These findings indicate that continuous long-term exposure of osteoblastic cells to cAMP elevating agents inhibits DNA binding of the osteoblast-specific transcription factor, Cbfa1, and that this inhibition is through proteolytic degradation involving ubiquitin/proteasome-dependent mechanism. The decreased Cbfa1 DNA binding correlated with inhibition of several downstream Cbfa1-regulated genes including alkaline phosphatase, bone sialoprotein, osteopontin, and osteocalcin (5,28). Forskolin also decreased von Kossa staining and radiolabeled calcium incorporation into the matrix, presumably due to inhibition of alkaline phosphatase and other extracellular matrix synthesis.
Regulation of Cbfa1 has been reported at a variety of levels including transcriptional and post-transcriptional. Ducy et al. (5) showed that 1,25(OH) 2 D 3 inhibits Cbfa1 mRNA levels, whereas bone morphogenetic protein 4/7 increases the mRNA level of Cbfa1 (5,29). In contrast, PTH regulates the collagenase-3 promoter through cooperative interaction of two transcription factors, AP-1 and Cbfa1, without directly affecting their individual binding affinities (30). Our results showed that forskolin regulated Cbfa1 more at the post-transcriptional level than transcriptional level. Similar regulation of Cbfa1 at the protein level has been reported with glucocorticoid treatment in primary osteoblast-enriched cultures (31).
The induction of proteases has been reported in response to PTH stimulation in MC3T3-E1. Murray and colleagues (19,32) previously showed that PTH, through the cAMP pathway, stimulates both calpains (calcium activated papain-like proteases I and II) and 26 S proteasome activities in osteoblasts, and that membrane-permeable protease inhibitors have been shown to attenuate the effects of PTH. They also noted that the 26 S proteasomal pathway, which accounts for more than 90% of the regulatory proteolysis, is 100-to 1000-fold higher than calpain activities (19). This proteolytic mechanism also appears to account for our observed decrease in Cbfa1 activity with forskolin treatment.
The results showed that after 3 days of continuous forskolin treatment (4 days in culture), a 6-h incubation with proteasome inhibitors is sufficient to restore Cbfa1 to ϳ50% of the control level. This recovery time may appear rapid relative to the time required for degradation. However this difference may be explained by 2 considerations. First, the data indicate that Cbfa1 levels increase over this time period in control cells, suggesting that baseline turnover may be different between onset of degradation (32-48 h) versus the time of treatment with proteasome inhibitors (96 h). Second, the data also suggest that a lag time (at least 32 h) is required prior to induction of proteasomal dependent Cbfa1 degradation induced by forskolin. A similar lag time for proteolytic degradation has been reported with the transcription factor YY1, another substrate of the proteasome pathway (33). One possible explanation for this lag time is de novo synthesis of one or more enzymes involved in the ubiquitin/proteasome pathway: in T lymphocytes, induction of the ubiquitin-conjugating enzyme, hUBC9 mRNA, which targets the transcription factor ATF2 for ubiquitin-proteasome pathway, requires a 24 -48-h lag time after stimulation (34). Thus, this lag time for onset of Cbfa1 degradation (presumably required for enzyme synthesis) together with the increasing control levels of Cbfa1 over time may account for this rapid recovery in the presence of proteasome inhibitors.
Preceding degradation, there appears to be a small increase in Cbfa1 protein levels at 24 -32 h in cells treated with forskolin. This transient increase may be due to a stimulatory effect of forskolin on Cbfa1 protein levels during the 32-h lag time prior to onset of degradation by the ubiquitin/proteasome pathway. This is in agreement with reports by other investigators that catabolic versus anabolic effects of PTH/cAMP on osteoblastic differentiation depend on exposure time (8 -10).
Finally, our data indicate that Cbfa1 is a novel endogenous substrate for the proteasome in osteoblastic cells, in this case one that is induced by the cAMP pathway. These results provide new insight into the mechanistic basis of osteoblast inhibition by hormones such as PTH, through the cAMP pathway, namely proteasome-dependent degradation of Cbfa1, the master regulator of osteoblastic differentiation.