Parathyroid hormone-induced E4BP4/NFIL3 down-regulates transcription in osteoblasts.

Parathyroid hormone (PTH), a major regulator of bone metabolism, activates the PTHR1 receptor on the osteoblast plasma membrane to initiate signaling and induce transcription of primary response genes. Subsequently, primary genes with transcriptional activity regulate expression of downstream PTH targets. We have identified the adenovirus E4 promoter-binding protein/nuclear factor regulated by IL-3 (E4bp4) as a PTH-induced primary gene in osteoblasts. E4BP4 is a basic leucine zipper (bZIP) transcription factor that represses or activates transcription in non-osteoblastic cells. We report here that PTH rapidly and transiently induced E4bp4 mRNA in osteoblastic cells and that this induction did not require protein synthesis. PTH also induced E4BP4 protein synthesis and E4BP4 binding to a consensus but not to a mutant E4BP4 response element (EBPRE). E4BP4 overexpression inhibited an EBPRE-containing promoter-reporter construct, whereas PTH treatment attenuated activity of the same construct in primary mouse osteoblasts. Finally, E4BP4 overexpression inhibited PTH-induced activity of a cyclooxygenase-2 promoter-reporter construct. Our data suggest a role for E4BP4 in attenuation of PTH target gene transcription in osteoblasts.

Parathyroid hormone (PTH 1 ), a major regulator of bone metabolism, affects both osteoblastic and osteoclastic formation and activity (1)(2)(3). However, the target cell of PTH in bone is the osteoblast, which expresses PTHR1, a G protein-coupled heptahelical receptor that also binds PTH-related peptide (4 -7). Ligand-bound PTHR1 initiates signaling cascades that phosphorylate and activate transcription factors that in turn induce transcription of several primary response genes (8 -10). Primary gene products perform many functions and ultimately lead to changes in osteoblastic phenotype. To map early changes in regulation of osteoblastic primary gene expression, we performed representational difference analysis (RDA) of mRNA from control and PTH-treated osteoblastic cells (11). Here we report that PTH rapidly induced expression of the bZIP transcription factor adenovirus E4 promoter-binding protein/nuclear factor regulated by IL-3 (E4BP4).
E4bp4 was isolated by screening a placental cDNA gt11 expression library for binding to the ATF site of the adenovirus E4 promoter (12) or to the CRE/ATF-like sequence of the upstream regulatory region of the human IL-1␤ gene (13). Later, through screening of an expression gt11 cDNA library prepared from phytohemagglutinin-stimulated human T-cell mRNA for binding to regulatory sequences of the IL3 promoter, a nuclear factor named NF-IL3 was identified and found to be identical to E4bp4 (14). The 2-kb cDNA contains an open reading frame of 1386 bases and encodes a 462-amino acid (51.4 kDa) protein (12). E4BP4 is a bZIP transcription factor and acts both as a transcriptional repressor and transcriptional activator. E4BP4 has been hypothesized to be involved in cellular functions such as cell survival versus apoptosis, the antiinflammatory response, and circadian rhythm regulation (15). However, no involvement of E4BP4 expression in the function of bone cells has been reported.
In the present study, we demonstrate that PTH rapidly and transiently induced E4bp4 mRNA in primary mouse osteoblasts and that this induction did not require new protein synthesis. PTH also induced E4BP4 protein synthesis. Furthermore, PTH induced nuclear protein binding to a consensus E4BP4 response element (EBPRE). A specific E4BP4 antibody supershifted the majority of this binding action. E4BP4 overexpression inhibited promoter activity of a reporter luciferase construct containing EBPRE consensus sites in primary osteoblasts. PTH treatment attenuated the promoter activity of the same construct. Finally, E4BP4 overexpression inhibited PTHinduced activity of a cyclooxygenase-2 promoter-reporter construct. Our data suggest that E4BP4 induction might mediate PTH inhibition of target gene transcription.

MATERIALS AND METHODS
Cell Culture-Primary mouse osteoblasts (MOB) were obtained from the calvariae of neonatal mice as previously described (11). MC3T3-E1 cells were plated at 5,000 cells/cm 2 , and MOB cells were plated at 15,000 cells/cm 2 in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 50 g/ml streptomycin. ROS 17/2.8 cells were plated at 5,000 cells/cm 2 and were grown in F-12 medium (Invitrogen, Grand Island, NY) containing 10% fetal bovine serum, 100 units/ml penicillin, and 50 g/ml streptomycin. Cells were grown in a humidified incubator at 37°C and 5% CO 2 . Cells were grown in non-differentiating media (no ascorbic acid or ␤-glycerophosphate) and were treated with agents at the time of confluence.
Mouse Calvariae Organ Culture-Hemicalvariae from 6-to 8-day-old CD-1 mice were dissected and cultured individually in 6-well dishes (Costar, Cambridge, MA) containing 2 ml of BGJ b medium (Invitrogen) with 1 mg/ml bovine serum albumin, 50 g/ml ascorbic acid, 100 units/ml penicillin, and 50 g/ml streptomycin. All animals used in these studies were sacrificed according to protocol approved by the * The research was supported by National Institutes of Health Grant R01-DE13316. 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.
RNA Extraction and Northern Blot Assays-Total RNA was extracted using TRIzol reagent (Invitrogen) according to the manufacturer's protocol. Twelve to fifteen micrograms of total RNA was separated by electrophoresis on a 1% agarose/3.7% formaldehyde gel and transferred to a Gene Screen Plus hybridization filter (PerkinElmer Life Sciences, Boston, MA). Filters were prehybridized at 68°C for 15 min, hybridized at 68°C for 1 h using the QuickHyb hybridization solution (Stratagene, La Jolla, CA) with a 32 P-labeled probe (see below). Filters were washed and were used to expose radiographic film (Kodak, Rochester, NY) at Ϫ80°C. E4bp4 mRNA intensity was measured and corrected for GAPDH expression using a PhosphorImager screen (Amersham Biosciences).
Reverse Transcriptase-PCR-To generate a full-length E4bp4 probe, mRNA from MC3T3-E1 cells treated with 10 nM PTH for 2 h, was reverse-transcribed using the Moloney murine leukemia virus reverse transcriptase (Stratagene). cDNA was amplified by PCR using oligonucleotide primers that span the full-length open reading frame of E4bp4 (sense primer, 5Ј-CCG CCA TGC AGC TGA GAA AAA TGC AG-3Ј; antisense primer, 5Ј-TTA TTA CCT GGA GTC CGA AGC CG-3Ј (16)). The resulting amplification product was purified and subcloned into the pcDNA 3.1 expression vector downstream of the cytomegalovirus promoter (Invitrogen, San Diego, CA). To generate E4bp4 cDNA probe for Northern analysis, the E4bp4 cDNA insert was excised, gel-purified, and labeled with [ 32 P]dCTP by random primer labeling technique.
Total and Nuclear Protein Extraction-MOB cells were treated accordingly and then washed twice with ice-cold 1ϫ phosphate-buffered saline. For total protein extracts, cells were resuspended for 30 min in 1 ml of triple detergent lysis buffer containing 50 mM Tris-Cl, 150 mM NaCl, 0.02% sodium azide, 0.1% SDS, 100 g/ml PMSF, 1 g/ml aprotinin, 1% Nonidet P-40, 0.5% deoxycholate, and 1ϫ protease inhibitors. Lysates were spun at 15,000 ϫ g for 15 min at 4°C, and supernatants were used for Western blot assays. Nuclear proteins were prepared as previously described (17). Cells were resuspended in 400 l of buffer A (10 mM HEPES (pH 7.9), 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA) supplemented with freshly added 1 mM DTT, 0.5 mM PMSF, and 1ϫ protease inhibitor mixture (Roche Applied Science, Indianapolis, IN) for 15 min on ice. Then 25 l of 10% Nonidet P-40 was added, and the cells were vortexed vigorously for 10 s and centrifuged at maximum speed for 30 s in a microcentrifuge at 4°C. Pellets were resuspended in buffer C (20 mM HEPES (pH 7.9), 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA) supplemented with freshly added 1 mM DTT, 0.5 mM PMSF, and 1ϫ protease inhibitor. After incubating for 30 min at 4°C on a rocking platform, nuclear lysates were centrifuged at maximum speed for 10 min in a microcentrifuge at 4°C, and the supernatants containing the nuclear proteins were collected. Protein concentrations were determined by the Lowry method.
Western Blot Analysis-Fifty micrograms total or nuclear proteins were combined with equal volume of 2ϫ loading buffer, boiled for 5 min, separated by SDS-PAGE (5% stacking at 20 mA and 10% resolving gel at 40 mA), and electrotransferred onto a nitrocellulose membrane overnight at 22 V and 4°C. Membranes were stained with 0.1% Ponceau S (w/v) in 5% acetic acid (v/v) to determine equal loading. Then membranes were blocked for 1 h with 5% nonfat milk in TBS-T (0.5% Tween), incubated with a polyclonal E4BP4 antibody (Santa Cruz Biotechnologies, Santa Cruz, CA, catalog number SC-9549) at 1:100 dilution in 5% nonfat milk/TBS-T for 1-2 h at room temperature, washed with TBS-T, and incubated with anti-goat IgG-horseradish peroxidase conjugated secondary antibody (Santa Cruz Biotechnologies) in 5% nonfat milk/TBS-T for 1 h. Following incubation, membranes were washed with TBS-T, and proteins were visualized using ECL reagent (Amersham Biosciences, Piscataway, NJ) according to the manufacturer's instructions and autoradiography.
Statistics-Statistical differences between two groups were detected with Student's t test. Comparisons among groups were performed by analysis of variance, and statistical differences were detected with a Bonferroni post hoc test. All experiments were performed at least three times.

PTH Induced E4bp4 Gene Expression in the Absence of Protein Synthesis in MC3T3-E1
Osteoblastic Cells-Using RDA to identify osteoblastic primary response genes induced by PTH (11), we cloned a DNA fragment that corresponded to E4bp4 (data not shown). To confirm that E4bp4 was a PTH-induced primary response gene, MC3T3-E1 cells were treated with 10 g/l CHX for 30 min followed by 10 nM PTH for 90 min. Total RNA was extracted, and Northern blot analysis was performed using the RDA-isolated cDNA fragment as a probe. E4bp4 was constitutively expressed in CHX-treated cells and was markedly induced by PTH treatment (Fig. 1A).
PTH  primers demonstrated that PTH induced E4BP4 mRNA in both osteoblastic cell lines with a peak at 2 h (Fig. 1B). DNA sequencing confirmed that the PCR product had the previously published E4bp4 sequence (accession number NM_017373, National Institutes of Health data base).
PTH Induced E4bp4 Expression in Primary Mouse Osteoblasts and Calvariae-MC3T3-E1 and ROS 17/2.8 cells are immortalized cell lines that retain a pre-osteoblastic/osteoblastic phenotype. To test if PTH induces E4bp4 gene expression in primary osteoblastic cells and normal bone, MOB cells or calvariae organ cultures were treated with 10 nM PTH for various times, total RNA was extracted, and Northern blot analyses were performed. PTH-induced E4bp4 mRNA expression in both MOB cells and cultured calvariae peaked at 2 h and declined thereafter (Fig. 2).
PTH Induction of E4bp4 mRNA Does Not Require New Protein Synthesis in MOB Cells and Is Regulated at the Transcriptional Level-To verify that E4bp4 is a PTH-induced primary response gene in MOB cells and to investigate if E4bp4 gene transcription is required for the PTH effects, MOB cells were pretreated with 3 g/ml CHX or with 5 g/ml actinomycin D (ActD) for 30 min and then with 10 nM PTH for 0 -6 h (Fig. 3). Pretreatment with CHX enhanced PTH-induced E4bp4 mRNA levels that were sustained for the duration of the experiment. Pretreatment with ActD completely inhibited PTH-induced E4bp4 mRNA expression (Fig. 3).
PTH Induced E4BP4 Protein Synthesis in MOB Cells-To determine if PTH induced E4BP4 protein synthesis, MOB cells were treated with 10 nM PTH for 0 -6 h, and total or nuclear proteins were extracted. Western immunoblot analysis indicated that PTH induced E4BP4 protein expression in MOB cells that peaked between 2 and 4 h and declined thereafter (Fig. 4). Similar results were obtained with both total and nuclear proteins, although nuclear proteins demonstrated a delayed peak and decline (Fig. 4, compare A and B). Recombinant E4BP4 protein was used as a control (lane R, Fig. 4A).
PTH Induced Nuclear Protein Binding to a Consensus E4BP4 Response Element (EBPRE)-EMSAs were performed to test the ability of in vitro synthesized E4BP4 protein to bind consensus or either of two mutant EBPREs. E4BP4 bound with high, moderate, and low affinity to wild type EBPRE, to EBPRE-m1, and to EBPRE-m2, respectively (Fig. 5A). In vitro synthesized luciferase control showed no binding to the DNA probes.
To investigate potential PTH-induced binding to the consensus or mutant EBPREs, EMSAs utilizing nuclear proteins from PTH-treated MOB cells and the EBPRE probes were performed. Untreated cells showed a low binding of nuclear proteins to the consensus EBPRE. Treatment with 10 nM PTH increased this binding that peaked at 4 h and declined by 6 h. Similar to in vitro synthesized E4BP4, the binding of the nuclear extracts was reduced or almost abolished when EBPRE-m1 or EBPRE-m2 probes were used (Fig. 5B).
The PTH-induced binding to the EBPRE probe was completely, partially, or minimally inhibited by 25ϫ and 50ϫ excesses of EBPRE, EBPRE-m1, or EBPRE-m2 cold competitor suggesting specific binding of the nuclear proteins to the wild type probe (Fig. 5C).
E4BP4 Protein Is Part of the PTH-induced Binding to a Consensus EBPRE-Basal E4BP4 mRNA and protein levels were very low, whereas PTH treatment significantly induced E4bp4 gene expression. To test if the PTH-induced binding to the consensus EBPRE was due to an existing or to a newly synthesized protein, such as E4BP4, MOB cells were pretreated with 3 g/ml CHX for 30 min and then treated with 10 nM PTH for 0 -6 h. Nuclear extracts were utilized for EMSAs. In the absence of CHX, PTH induced binding to the EBPRE. However, CHX abolished this PTH effect suggesting that the PTH-induced binding was due to newly synthesized proteins (Fig. 6A).
To confirm that E4BP4 was part of the PTH-induced complexes, we utilized a polyclonal antibody against E4BP4 protein. First we determined that the antibody was able to supershift in vitro synthesized E4BP4 protein. E4BP4 or luciferase proteins were incubated with EBPRE probes and then with E4BP4 antibody or a nonspecific antibody (Nurr1 antibody at the same concentration, Santa Cruz Biotechnologies, catalog number SC-990). The E4BP4 antibody supershifted the entire E4BP4⅐EBPRE binding complex, whereas the nonspecific antibody had no effect (Fig. 6B). Similarly, the E4BP4 antibody supershifted part of the PTH-induced nuclear extract⅐EBPRE complex, whereas the nonspecific antibody had no effect. Interestingly, not all of the PTH-induced complexes were supershifted, in contrast to the in vitro synthesized E4BP4 protein (Fig. 6C). Because the consensus EBPRE is very similar to the consensus C/EBP␤ binding site (18), we tested the possibility that C/EBP␤ is part of the PTH-inducible binding to the EBPRE. Indeed, a C/EBP␤ antibody supershifted part of the PTH-induced nuclear extract⅐EBPRE complex (Fig. 6D).
E4BP4 Inhibited Promoter Activity in Osteoblasts-To study the regulatory role of E4BP4 on osteoblastic gene expression, MOB cells were co-transfected with pcDNA3.1-E4BP4 or empty vector and with pTK-Luc, pTK-EBPRE, pTK-EBPRE-m1, or pTK-EBPRE-m2 reporter constructs. E4BP4 overexpression had no effect on the basal thymidine kinase promoter activity. However, E4BP4 strongly inhibited pTK-EBPRE activity by 92% (Fig. 7A). When mutant EBPREs were inserted upstream of the thymidine kinase promoter, E4BP4 overexpression inhibited promoter activity to a lesser degree (pTK-EBPRE-m1 by 59% and pTK-EBPRE-m2 by 39%), in parallel to the ability of E4BP4 to bind these mutant sites, as described above.
Because E4BP4 overexpression inhibited luciferase activity of the pTK-EBPRE reporter construct, we tested the ability of PTH treatment to affect this construct. MOB cells were transfected with pTK-Luc, pTK-EBPRE, pTK-EBPRE-m1, or pTK-EBPRE-m2 reporter constructs and then treated with PTH for 12 h. PTH treatment caused a statistically significant attenuation of the promoter activity by 25%, compared with the untreated control (Fig. 7B). In contrast, PTH treatment had no statistical effect on the pTK-EBPRE-m1 or pTK-EBPRE-m2 constructs (Fig. 7B).
Finally, we tested the ability of E4BP4 to affect activity of endogenous promoters. Computer analysis revealed four EBPRE-like elements (seven of eight nucleotides) within the proximal 963 bp of the cyclooxygenase-2 (cox-2) promoter (the kind gift of Dr. Harvey Herschman, UCLA). MOB cells were co-transfected with a luciferase reporter that contained 1 kb of the cox-2 promoter and with pcDNA3.1-E4BP4 or empty vector. Four days later, cells were treated with vehicle or 10 nM PTH for 3 h, and luciferase activity was measured. As previously described (19), PTH induced cox-2 promoter activity (Fig. 7C).

FIG. 5. Effect of PTH on nuclear protein binding to wild type and mutant EBPREs.
A, in vitro synthesized E4BP4 or luciferase (control) proteins were incubated with labeled wild type (EBPRE) or mutant (EBPRE-m1 or EBPRE-m2) oligonucleotides, and EMSAs were performed. B, nuclear extracts from MOB cells treated with 10 nM PTH for 0 -6 h were incubated with labeled wild type or mutant EBPRE oligonucleotides. C, nuclear extracts from MOB cells treated with 10 nM PTH for 4 h were incubated with labeled oligonucleotides in the absence and presence of 25ϫ or 50ϫ excess unlabeled wild type or mutant EBPRE oligonucleotides. E4BP4 overexpression inhibited basal luciferase activity, although not at statistically significant levels. More importantly, E4BP4 overexpression significantly inhibited PTH-induced cox-2 promoter activity. DISCUSSION PTH has important effects on bone metabolism (1). Osteoblasts are the direct PTH target in bone, yet the molecular mechanisms that mediate the effects of PTH on osteoblastic gene expression remain largely elusive (9). PTH-triggered osteoblast activity begins with the rapid and transient induction of primary response gene expression (9,10). Subsequently, primary gene products mediate downstream PTH effects (19 -22). Among the primary genes induced by PTH are several transcription factors, including the AP-1 transcription factors c-fos and c-jun (23), the nuclear orphan receptors Nurr1 (11) and Nur77 (24) and the transcriptional repressor ICER (16). Here, we show that PTH induced E4BP4, which appears to be a transcriptional repressor in osteoblasts. To the best of our knowledge, this is the first report of osseous localization and PTH regulation of E4bp4 gene expression.
Similar to other tissues, osteoblastic cell lines and calvariae demonstrated low basal expression of E4bp4 mRNA levels. PTH treatment rapidly and transiently induced E4bp4 gene expression that peaked around 2 h and declined thereafter. In MC3T3-E1 and MOB cells, PTH-induced E4bp4 mRNA levels did not require protein synthesis suggesting that E4bp4 is a primary response gene in PTH-treated osteoblasts. In fact, CHX pretreatment led to sustained E4bp4 mRNA levels following PTH treatment. This is in agreement with the glucocorticoid induction of E4bp4 in mouse fibroblasts, where CHX pretreatment enhances the induction of E4bp4 mRNA levels (28). Interestingly, our findings contrast the observation that CHX abolishes IL-3-induced E4bp4 expression in pro-B lymphocytes (29). The role of E4bp4 as a primary response gene appears to be cell-specific.
E4BP4 contains a domain that is homologous to the DNA binding domain of the bZIP family of transcription factors, and both endogenous and recombinant E4BP4 bind DNA as a dimer at ATF-like sites (12,15). High affinity binding of random sequence oligonucleotides revealed the consensus E4BP4 binding site sequence (EBPRE) to be (G/A)T(G/T)A(C/T)GTAA, which is homologous to the CRE consensus site (TCACGTCA (12)).
The EBPRE sequence that we selected for our experiments conforms to the consensus but precludes CREB or ATF binding (12). Nuclear extracts from PTH-treated MOB cells formed complexes with the EBPRE site that we used. The nuclear protein⅐DNA complexes were completely abolished by pretreatment of the cells with CHX, suggesting that binding was due to newly synthesized, rather than pre-existing, proteins. Part of the PTH-induced nuclear protein⅐EBPRE complexes contained E4BP4 protein, because it supershifted with an E4BP4 antibody but not with a nonspecific antibody.
E4BP4 antibody did not supershift the whole PTH-induced EBPRE-binding complex. Analysis with the TRANSFAC version 4.0 software (18) revealed that the EBPRE site has close homology to the consensus C/EBP␤ binding site. Indeed, a C/EBP␤ antibody supershifted a PTH-inducible EBPRE⅐ nuclear protein complex. C/EBP␤ is induced by PTH in osteoblastic cells (34). 2 Recognition of the same cis-binding element by both E4BP4 and C/EBP␤ points to potentially interesting interactions of these transcription factors for the regulation of target gene expression.
E4BP4 acts as a transcriptional inhibitor in most experimental settings (12,25,(35)(36)(37). However, E4BP4 overexpression transactivates the IL-3 promoter in resting T cells, presumably through binding to a cis element located within Ϫ165/Ϫ128 bp of the IL-3 promoter (14). Why E4BP4 has such distinct effects on transcriptional activity is not clear. Possibly these effects reflect differences in the cell types used with different coactivators/corepressors present. Interestingly, the E4BP4 binding site in the IL-3 promoter is not a canonical EBPRE (14). Thus, the cis-element sequence or the overall promoter context could also influence the effect of E4BP4 on transcription (15).
The E4BP4 transcription repression domain is localized between amino acids 298 and 363. This domain confers transcription repression function in an E4BP4-GAL4 chimeric protein (38). A possible mechanism for the transcriptional inhibition by E4BP4 is the binding of its repression domain to the TBPbinding protein DR1 (39). DR1 is a co-repressor of both basal and induced transcription that interacts with TBP and prevents the association of TFIIB with the TBP⅐TATA complex (40,41).
To test if E4BP4 represses or activates gene transcription in 2 B. E. Kream, University of Connecticut, personal communication.
FIG. 6. E4BP4 protein is part of the PTH-induced binding to a consensus EBPRE. A, MOB cells were pretreated with 3 g/ml CHX for 30 min and then with 10 nM PTH for 0 -6 h. Nuclear proteins were incubated with labeled wild type EBPRE oligonucleotides, and EMSAs were performed. B, in vitro synthesized E4BP4 or luciferase proteins were incubated with labeled wild type EBPRE oligonucleotides and then with E4BP4-specific or nonspecific polyclonal antibodies. C, MOB cells were treated for 0 -6 h with 10 nM PTH. Nuclear extracts were incubated with wild type EBPRE oligonucleotides and then with E4BP4-specific or nonspecific polyclonal antibodies. D, MOB cells were treated for 0 -6 h with 10 nM PTH. Nuclear extracts were incubated with wild type EBPRE oligonucleotides and then with C/EBP␤-specific polyclonal antibodies.
osteoblasts, we overexpressed E4BP4 in resting MOB cells transfected with a reporter construct carrying a basal promoter under the control of three tandem consensus EBPREs. E4BP4 strongly inhibited the chimeric promoter activity (92%), whereas it had no effect on the basal promoter. When mutant EBPREs were used, E4BP4 inhibition closely paralleled its ability to bind to the mutant sites.
When MOB cells, transfected with the reporter construct containing three tandem consensus EBPREs, were treated with PTH, a statistically significant (25%) attenuation of promoter activity was observed. This suggests that PTH not only induces binding of E4BP4 to EBPREs but also that PTH regulates E4BP4-responsive promoters. The inhibition of the same promoter by E4BP4 overexpression was substantially greater than the inhibition by PTH treatment. A possible explanation is that the pcDNA-E4BP4 vector forces much higher E4BP4 expression than the induction by PTH treatment. Another possible explanation is that, in addition to E4BP4, PTH induces transcription activators that compete for EBPRE binding. The resulting promoter regulation will reflect the competition between transcription activators and repressors. Indeed, Fig. 6D shows that, in addition to E4BP4, PTH induces C/EBP␤ binding to EBPRE. PTH inhibits transcription of osteoblastic genes, including both primary (19) and late response genes (20,42,43). The catabolic effects of continuous PTH treatment appear to be associated with decreased expression of many of the genes involved in bone formation and increased expression of genes involved in bone resorption (10). Because E4BP4 appears to be a transcriptional inhibitor in osteoblasts, E4BP4 expression could potentially mediate some of the inhibitory effects of PTH on target gene transcription.
One PTH target gene is cox-2, which is rapidly and transiently induced by PTH through activation of the cAMP-PKA pathway (19). Although induction of cox-2 transcription does not require new protein synthesis, the return to basal transcription does. This suggests that PTH simultaneously induces cox-2 and a transcriptional inhibitor. A possible mode of action for this inhibitor would be competition for promoter binding and transcriptional repression. E4BP4 is an ideal candidate for cox-2 repression, because activation of ATF/CREB would induce transcription, whereas E4BP4 synthesis would compete for binding and would attenuate promoter activity (44). Moreover, computer analysis revealed four EBPRE-like sites (seven out of eight nucleotides) within the cox-2 promoter at positions Ϫ833/Ϫ826, Ϫ683/Ϫ676, Ϫ59/Ϫ52, and Ϫ 34/Ϫ27. Thus, we tested the ability of E4BP4 protein to inhibit basal and, more importantly, PTH-induced cox-2 promoter activity. Indeed, E4BP4 overexpression caused an 80% inhibition of the PTHinduced cox-2 promoter.
The possibility that E4BP4 is part of a tightly regulated system of transcriptional activators and repressors has precedent in the molecular control of circadian rhythms (12,15,25,36). E4bp4 expression follows a circadian pattern in the suprachiasmatic nucleus, the pineal gland, and the liver (37,45). Furthermore, E4bp4 mRNA levels are in opposite phase with the expression of dbp, hlf, and tef genes that also oscillate with a daily rhythm in a number of tissues (45). E4bp4, dbp, hlf, and tef genes are members of the proline-and acidic residue-rich family of proteins and recognize identical promoter elements (15). Thus the involvement of proline-and acidic residue-rich proteins as reciprocal regulators of gene expression controlling circadian rhythms was hypothesized (45).
E4BP4 induction is associated with inhibition of cell apoptosis. E4BP4 overexpression promotes survival of IL-3-dependent pro-B lymphocytes in the absence of IL-3 treatment. In combination with the observation that IL-3 induces E4BP4 expression in these cells, E4BP4 was hypothesized to play a pivotal role in IL-3-mediated pro-B survival (29). However, no gene targets that might mediate the E4BP4 effects on cell survival have been identified. Interestingly, apoptosis inhibition has been proposed as a possible mechanism for the anabolic effects of PTH. PTH inhibits osteoblast and osteocyte apoptosis in normal and osteopenic mice in vivo and dexamethasone-induced osteoblast apoptosis in vitro (46,47). It is conceivable that E4BP4 induction may participate in the PTH-mediated inhibition of apoptosis.
In summary, we demonstrate that PTH induced expression of the transcription factor E4BP4 in osteoblastic cells as a primary response gene. Furthermore, PTH induced binding of nuclear E4BP4 protein to EBPRE and regulated the activity of an EBPRE-containing promoter. E4BP4 inhibited chimeric and endogenous promoters in osteoblasts, suggesting that E4BP4 might mediate the inhibitory effects of PTH on target gene transcription.
Acknowledgment-We thank Dr. Jeanne Nervina for providing insightful comments and reviewing the manuscript.  7. E4BP4 acts as a transcriptional repressor in MOB cells. A, MOB cells were transiently co-transfected with empty or E4BP4 expression vectors and with luciferase-reporter constructs carrying three tandem repeats of wild type (pTK-EBPRE) or mutant (pTK-EBPRE-m1 or pTK-EBPRE-m2) EBPREs. The plot indicates mean Ϯ S.E. luciferase activity, expressed as percentage of control luciferase (*, p Ͻ 0.05). B, MOB cells were transiently transfected with luciferase-reporter constructs carrying three tandem repeats of wild type or mutant EBPREs. Then cells were treated with 10 nM PTH for 12 h. The plot indicates mean Ϯ S.E. luciferase activity, expressed as percentage of control (*, p Ͻ 0.05). C, cells were transiently co-transfected with empty or E4BP4 overexpression vectors and with luciferase-reporter constructs containing 1 kb of the cox-2 promoter. Cells were treated with 10 nM PTH for 3 h. The plot indicates mean Ϯ S.E. luciferase activity expressed as percentage of maximum induction (*, p Ͻ 0.05, significantly different from vehicle in the absence of exogenous E4BP4; ‚, p Ͻ 0.05, significantly different from PTH treated in the absence of exogenous E4BP4).