CCAAT/Enhancer-binding Protein d Activates Insulin-like Growth Factor-I Gene Transcription in Osteoblasts IDENTIFICATION OF A NOVEL CYCLIC AMP SIGNALING PATHWAY IN BONE*

Insulin-like growth factor-I (IGF-I) plays a key role in skeletal growth by stimulating bone cell replication and differentiation. We previously showed that prostaglandin E2 (PGE2) and other cAMP-activating agents enhanced IGF-I gene transcription in cultured primary rat osteoblasts through promoter 1, the major IGF-I promoter, and identified a short segment of the promoter, termed HS3D, that was essential for hormonal regulation of IGF-I gene expression. We now demonstrate that CCAAT/enhancer-binding protein (C/EBP) d is a major component of a PGE2-stimulated DNA-protein complex involving HS3D and find that C/EBPd transactivates IGF-I promoter 1 through this site. Competition gel shift studies first indicated that a core C/EBP half-site (GCAAT) was required for binding of a labeled HS3D oligomer to osteoblast nuclear proteins. Southwestern blotting and UV-cross-linking studies showed that the HS3D probe recognized a ; 35-kDa nuclear protein, and antibody supershift assays indicated that C/EBPd comprised most of the PGE2-activated gel-shifted complex. C/EBPd was detected by Western immunoblotting in osteoblast nuclear extracts after treatment of cells with PGE2. An HS3D oligonucleotide competed effectively with a high affinity C/EBP site from the rat albumin gene for binding to osteoblast nuclear proteins. Cotransfection of osteoblast cell cultures with a C/EBPd expression plasmid enhanced basal and PGE2-activated IGF-I promoter 1-luciferase activity but did not stimulate a reporter gene lacking an HS3D site. By contrast, an expression plasmid for the related protein, C/EBPb, did not alter basal IGF-I gene activity but did increase the response to PGE2. In osteoblasts and in COS-7 cells, C/EBPd, but not C/EBPb, transactivated a reporter gene containing four tandem copies of HS3D fused to a minimal promoter; neither transcription factor stimulated a gene with four copies of an HS3D mutant that was unable to bind osteoblast nuclear proteins. These results identify C/EBPd as a hormonally activated inducer of IGF-I gene transcription in osteoblasts and show that the HS3D element within IGF-I promoter 1 is a high affinity binding site for this protein.

Insulin-like growth factor-I (IGF-I), 1 a highly conserved multifunctional peptide with actions on somatic growth, cell survival, tissue differentiation, and intermediary metabolism (1,2), is produced by many tissues, including bone (3). IGF-I has been shown to be synthesized by bone cells, including osteoblasts, and can act as both a growth and differentiation agent within the skeleton (4 -7). Expression of IGF-I by osteoblasts is regulated by systemic and local factors (8,9). IGF-I synthesis is enhanced by parathyroid hormone and by locally produced prostaglandin E 2 (PGE 2 ) (8,9), two hormones that are involved in coupling the remodeling sequence of bone resorption and new bone formation (10 -15), and is diminished by glucocorticoids (16), hormones that decrease bone formation. IGF-I thus has been proposed to play a central role in skeletal growth and in bone remodeling by acting as a key mediator in hormonal control of these processes.
Both parathyroid hormone and PGE 2 elevate intracellular levels of cAMP in osteoblasts (9,17), and it has been shown that stimulation of cAMP accumulation within osteoblasts enhances production of IGF-I and increases IGF-I mRNA abundance (8,9,17). Previous studies have demonstrated that PGE 2 induces IGF-I gene transcription in primary cultures of rat osteoblasts (Ob cells) by activating promoter 1, the major IGF-I promoter in most tissues (18,19). We have found additionally that stimulation of transiently transfected IGF-I promoter 1-luciferase reporter genes by PGE 2 can be mimicked by co-transfection with the catalytic subunit of cyclic AMP-dependent protein kinase and inhibited by a mutant regulatory subunit that is unable to bind cAMP (20), thus confirming the role of this second messenger in activating IGF-I gene expression. In recent experiments, we mapped a functional cAMP response element (CRE) to the 5Ј-untranslated region of IGF-I exon 1 within a previously footprinted site termed HS3D (21) and showed that this segment of DNA was required for full hormonal responsiveness of IGF-I promoter 1 in osteoblasts (22). We also demonstrated by gel mobility shift assays that PGE 2 treatment induced nuclear protein binding to an HS3D oligonucleotide probe by a mechanism that did not require ongoing protein synthesis (22). Based on these observations and on the absence of a typical CRE DNA sequence within the hormonally respon-sive segment of IGF-I promoter 1 (22), we postulated that the IGF-I gene was regulated by PGE 2 by a potentially novel transcriptional mechanism.
We now have identified the principal cAMP-activated transcription factor in osteoblasts that binds to and transactivates IGF-I promoter 1 via the HS3D site. We show that C/EBP␦, a member of the CCAAT/enhancer-binding protein family, appears in osteoblast nuclear extracts after treatment of cells with PGE 2 , that it interacts with an HS3D oligonucleotide probe with an affinity similar to that of a well characterized C/EBP site, and that it transactivates IGF-I promoter 1 and another neutral promoter only when a native HS3D sequence is present. Our results define C/EBP␦ as a cAMP-regulated activator of IGF-I gene transcription in osteoblasts and show that the HS3D element within IGF-I promoter 1 is a functionally important binding site for this protein.
Experimental Animals-Pregnant female Sprague-Dawley rats were purchased from Charles River Laboratories (Raleigh, NC), and 8-weekold male Sprague-Dawley rats were purchased from Harlan-Sprague-Dawley (Indianapolis, IN). Rats were housed in individual cages under a 12-h light-dark cycle with free access to food and drinking water. All protocols were approved by institutional animal study committees.
Cell Culture-Primary osteoblast-enriched cell cultures (Ob cells) were prepared from the parietal bones of 22-day-old Sprague-Dawley rat fetuses, as described previously (23). Cranial sutures were eliminated during dissection, and the bones were digested with collagenase for five sequential 20-min intervals. Cells from the last three digestions were pooled and then plated at 4800/cm 2 in Dulbecco's modified Eagle's medium containing 20 mM HEPES (pH 7.2), 0.1 mg/ml ascorbic acid, penicillin, and streptomycin (all from Life Technologies, Inc.), and 10% fetal bovine serum (Sigma). COS-7 cells (ATCC CRL-1651) were incubated in antibiotic-free Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum. Cells were typically plated at 1 ϫ 10 5 /60-mm diameter tissue culture plate.
C/EBP␤ expression vectors were constructed from a rat C/EBP␤ cDNA in pHDIL6DBD (26). A segment containing the coding region was subcloned into plasmid Bluescript (Stratagene, La Jolla, CA) to make pBS-C/EBP␤. An EcoRI fragment from pBS-C/EBP␤ was then inserted into EcoRI-digested pSV7d (27) and pcDNA3 (Invitrogen, Carlsbad, CA) in the appropriate orientation to generate pSV7d-C/EBP␤ and pcDNA3-C/EBP␤, respectively. C/EBP␦ expression vectors were constructed as follows. First, the coding region of the intronless rat C/EBP␦ gene was isolated from genomic DNA by nested PCR, using the oligonucleotide primers listed in Table I. The amplified fragment was purified, cloned between BamHI and EcoRI sites of plasmid Bluescript to make pBS-C/EBP␦, and sequenced in its entirety. Appropriate restriction fragments were then subcloned into expression vectors to generate pSV7d-C/EBP␦ and pcDNA3-C/EBP␦. Gene Transfer Experiments-Transfection studies using primary rat osteoblasts were performed as described previously (20,22). IGF-I promoter 1-luciferase fusion genes were co-transfected with C/EBP expression plasmids and with a vector carrying the ␤-galactosidase gene under SV40 promoter control (Promega Corp.) to normalize for transfection efficiency. Cultures at 50% confluent density were rinsed in serum-free medium and exposed to plasmids in the presence of Lipofectin TM (Life Technologies, Inc.) for 3 h. The solution was then replaced with growth medium containing 5% fetal bovine serum, and the cells were incubated for 48 h until reaching confluent density. The cells then were rinsed with serum-free medium and treated for 6 h with vehicle (ethanol diluted 1:1000 or greater in serum-free medium) or 1 M PGE 2 . After incubation, the medium was aspirated, the cultures were rinsed with phosphate-buffered saline and lysed in cell lysis buffer (Promega Corp.), and luciferase activity was measured as described previously (20,22). COS-7 cells were transfected using the calcium phosphate precipitation method, as described (24). Cells were plated at a density of 1 ϫ 10 5 /60-mm dish in complete medium and incubated overnight. Four h after the medium was changed, DNA was added. A total of 1 g of reporter gene was co-transfected with up to 100 ng of expression plasmid (pcDNA3, pcDNA3-C/EBP␤, or pcDNA3-C/EBP␦) and 1 ng of a vector containing the Renilla luciferase gene under control of the cytomegalovirus immediate early enhancer/promoter (pRL-CMV, Promega Corp.) to normalize for transfection efficiency. The medium was changed at 24 h after the addition of DNA, and 24 h later, following aspiration of medium and rinsing in phosphate-buffered saline, the cells were lysed and extracts were assayed, using the Dual-Luciferase Reporter Assay System (Promega Corp.). Light emission was measured with a Monolight 2010 luminometer (Analytical Luminescence, Ann Arbor, MI) by integration over 10 s of the reaction.
Nuclear Protein Extracts-Rat osteoblast nuclear extracts were prepared as described previously (20). Confluent osteoblast cultures were deprived of serum for 20 h. Cells then were rinsed with serum-free medium and incubated with vehicle (ethanol diluted 1:1000 or greater) or 1 M PGE 2 for up to 4 h. Medium was aspirated, and cultures were rinsed twice with phosphate-buffered saline at 4°C. Cells were harvested with a cell scraper and gently pelleted, and the pellets were washed with phosphate-buffered saline. Nuclear extracts were prepared by the method of Lee et al. (28) with minor modifications (20). Cells were lysed in hypotonic buffer (10 mM HEPES, pH 7.4, 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM dithiothreitol) with phosphatase inhibitors (1 mM sodium orthovanadate, 10 mM sodium fluoride, 0.4 mM microcystin CL), protease inhibitors (0.5 mM phenylmethylsulfonyl fluoride, 1 g/ml pepstatin A, 2 g/ml leupeptin, 2 g/ml aprotinin), and 1% Triton X-100. Nuclei were pelleted and were resuspended in hypertonic buffer containing 0.42 M NaCl, 0.2 mM EDTA, 25% glycerol, and the phosphatase and protease inhibitors indicated above. Soluble proteins released by a 30-min incubation at 4°C were collected by centrifugation at 12,000 ϫ g for 5 min, and the supernatant was dialyzed for 2 h against 2000 volumes of buffer (20 mM HEPES, pH 7.4, 100 mM KCl, 0.1 mM EDTA, 0.5 mM dithiothreitol, 1 mM sodium orthovanadate, 20% glycerol), containing the protease inhibitors listed above. Protein concentrations were quantitated using a modified Bradford assay (Bio-Rad).
Rat liver nuclear proteins were extracted according to previously published methods (21). Protein concentrations were measured as indicated above. Nuclear extracts were aliquoted and immediately frozen in liquid nitrogen for storage.
C/EBP␤ and C/EBP␦ proteins were generated using a coupled in vitro

5Ј-GCGGATCCGAGGTGACAGCCCAACTTG
Ϫ45 to Ϫ26 Antisense 5Ј-GGAATTCGGTCGTTCGGAGTCTCTAAG 841 to 822 a Restriction sites used for cloning are underlined. b Relative to translation initiation site (38). transcription-translation system (TNT Coupled Reticulocyte Lysate Systems (Promega)) by following the manufacturer's directions. Plasmid DNA was incubated with amino acid mix minus methionine, [ 35 S]methionine (1000 Ci/mmol, Amersham Corp.), rabbit reticulocyte lysate, and T3 RNA polymerase in reaction buffer at 30°C for 2 h. Proteins were separated by electrophoresis through 10% SDS-PAGE gels, and the dried gels were exposed to x-ray film.
C/EBP proteins were expressed in transiently transfected COS-7 cells. Cells were grown in 150-mm diameter tissue culture dishes and were transfected by calcium phosphate precipitation, as outlined above, using 20 g of expression vectors (pcDNA3, pcDNA3-C/EBP␤, or pcDNA3-C/EBP␦). Twenty-four h later, the medium was changed, and after an additional 24 h, cells were harvested and nuclear extracts were prepared as described above for osteoblast cultures.
DNA-Protein Binding Studies-Gel mobility shift experiments followed previously published methods (22). Radiolabeled double-stranded DNA probes were synthesized by annealing complementary oligonucleotides (Table II), followed by fill-in of single-stranded overhangs with dCTP, dGTP, TTP, and [␣-32 P]dATP (800 Ci/mmol, NEN Life Science Products), using the Klenow fragment of DNA polymerase I. Nuclear protein extracts (5-20 g) were preincubated for 20 min on ice with up to 2 g of poly(dI-dC) with or without unlabeled specific or nonspecific DNA competitor or antibodies in 25 mM HEPES, pH 7.6, 60 mM KCl, 7.5% glycerol, 0.1 mM EDTA, 5 mM dithiothreitol, and 0.025% bovine serum albumin. After the addition of 5 ϫ 10 4 cpm of DNA probe for 30 min on ice, samples were applied to a 4 -20% nondenaturing polyacrylamide gradient gel (Novex, San Diego, CA) that had been pre-electrophoresed for 30 min at 12.5 V/cm at 25°C in 45 mM Tris, 45 mM boric acid, 1 mM EDTA. Electrophoresis proceeded for 2.5 h under identical conditions. The dried gels were exposed to x-ray film at Ϫ80°C with an intensifying screen.
DNA-nuclear protein cross-linking by UV light was performed according to published methods (29). Modified radiolabeled doublestranded DNA probes were synthesized as described above with the substitution of bromodeoxyuridine triphosphate (Sigma) for TTP. Osteoblast nuclear protein extracts (30 g) were preincubated for 30 min on ice as outlined above for gel mobility shift assays. After the addition of 5 ϫ 10 4 cpm of bromodeoxyuridine-modified HS3D probe for 30 min on ice, the reaction mixture was exposed to 300-nm UV light for 1 h at a distance of 5 cm, as described (29). After electrophoresis by SDS-PAGE, dried gels were exposed to x-ray film at Ϫ80°C with intensifying screens.
Southwestern blotting followed a previously described protocol (30, 31). Osteoblast nuclear proteins (30 g) were separated by SDS-PAGE using 15% linear gels, and proteins were transferred to nitrocellulose membranes. Following a denaturation-renaturation step using progressively reduced concentrations of guanidine-HCl in blocking/binding buffer (10 mM HEPES, pH 8.0, 50 mM KCl, 1 mM EDTA, 6.4 mM MgCl 2 , 1 mM dithiothreitol), membranes were incubated in the same buffer for 1 h at 25°C. Blots then were hybridized with 1 ϫ 10 7 cpm of labeled probe in 5 ml of buffer for 1 h at 25°C. After washing with the same buffer, membranes were exposed to x-ray film at Ϫ80°C with an intensifying screen.
Western Blotting-Osteoblast nuclear proteins (20 g) were separated by SDS-PAGE and transferred to nitrocellulose membranes (32). After membranes were blocked with 5% nonfat dry milk and 2% goat serum in 20 mM Tris-Cl, pH 7.6, 137 mM NaCl for 1 h at 25°C, they were incubated with an antibody to C/EBP␦ (c150; diluted 1:1000 in blocking buffer) for 1 h at 25°C. Subsequent steps were performed as described (32). Immunoreactive bands were visualized by enhanced chemiluminescence (ECL Western blotting system, Amersham), followed by exposure to x-ray film.

RESULTS
In previous studies, we identified HS3D as an atypical cAMP response element in the 5Ј-untranslated region of rat IGF-I exon 1 that mediated hormonally stimulated IGF-I gene transcription in primary rat osteoblasts (20,22). By gel mobility shift assay, nuclear protein binding to this element was absent under basal conditions but was induced rapidly in osteoblast cultures exposed to PGE 2 (22). This DNA fragment contained a core octameric sequence, 5Ј-CGCAATCG-3Ј, which was required for both nuclear protein binding and transcriptional activation (22). We thus performed competition gel-mobility shift experiments to define the key nucleotides within the HS3D octameric core that were necessary for binding of PGE 2induced nuclear proteins. A series of double-stranded oligomers 31 bp in length were generated with single nucleotide substitution mutations of purines for purines and pyrimidines for pyrimidines at eight different positions (Fig. 1A). Results with these competitor oligonucleotides indicated that only three mutants, 6B, 6D, and 7A, failed to inhibit binding of the 32 P-HS3D probe when used at a 200-fold molar excess (Fig. 2B), although mutant 6B was partially effective. Therefore, only three nucleotides within the HS3D core, 5Ј-CGCAATCG-3Ј (underlined and in boldface type) appear to be critical for the binding of PGE 2 -stimulated nuclear factors. The octameric HS3D core contains an optimal half-site for transcription factors of the C/EBP family, GCAAT (33), and the nucleotides within the half-site that we identified as essential for DNA-protein interactions with osteoblast nuclear proteins were recently found by others to be absolutely conserved for binding recombinant C/ EBP proteins in a PCR-based binding site selection assay (34).
To begin to characterize the factors interacting with this element, we next performed an additional series of DNA-protein binding experiments. We were able to cross-link osteoblast nuclear proteins by UV light to a double-stranded 32 P-labeled HS3D oligonucleotide containing BrdU and show that labeled proteins migrated at ϳ40 kDa after SDS-PAGE and autoradiography ( Fig. 2A). Binding of the probe could be competed by an excess of unlabeled HS3D but was not inhibited by an equivalent excess of a double-stranded oligonucleotide, M6, that contained a mutation in a 4-nucleotide block within the HS3D core that prevented nuclear protein binding (22). A 32 P-HS3D oligomer also detected a protein band of ϳ35 kDa by a filter binding assay (Southwestern blot) of electrophoretically fractionated osteoblast nuclear proteins (Fig. 2B). The difference in mobility between the bands in Fig. 2, A and B, may be accounted for by the mass of the HS3D oligonucleotide covalently coupled to the protein after UV cross-linking.
Based on the results in Figs. 1 and 2, we next asked if C/EBP proteins were expressed in osteoblasts and then examined their potential presence in DNA-protein complexes. By a ribonuclease protection assay, both C/EBP␤ and C/EBP␦ mRNAs were detected in these cells. By antibody supershift experiments, we determined that an antiserum to C/EBP␦ attenuated the DNAprotein complexes formed with Ob cell nuclear extracts, while two antibodies to C/EBP␤ and an irrelevant antiserum to myogenin were ineffective (Fig. 3). In contrast to the closely spaced doublet formed between osteoblast nuclear proteins and 32 P-HS3D, at least three distinct DNA-protein complexes were seen when hepatic nuclear extracts were used in a parallel experi-  (Fig. 3). Antibodies to C/EBP␤ disrupted the most rapidly migrating of these bands, while antisera to C/EBP␦ or myogenin had no effect. These results support the idea that a protein antigenically related to C/EBP␦ comprises a major part of the HS3D-nuclear protein complex in Ob cells. By contrast, other factors, including C/EBP␤, are able to bind to the same DNA fragment in rat liver, where C/EBP␦ is minimally expressed under basal physiological conditions (35).
The D site within the proximal promoter of the rat albumin gene was one of the first high affinity sequences identified for C/EBP proteins (36). When used in gel mobility shift experiments with osteoblast nuclear proteins, the 32 P-labeled doublestranded D site probe shown in Table II gave rise to a DNAprotein complex of identical mobility as found with the labeled HS3D oligomer (Fig. 4). Cross-competition studies additionally revealed that both sets of protein-DNA complexes could be inhibited equivalently by a 200-fold molar excess of either homologous or heterologous competitor, although neither complex was blocked by the unrelated Oct-1 oligomer. In conjunction with results shown in Fig. 3, these observations indicate that HS3D contains an authentic C/EBP binding site and is capable of interacting with both C/EBP␤ and C/EBP␦ in different tissues.
As shown previously, treatment of primary rat osteoblasts with PGE 2 induces nuclear protein binding to an HS3D probe (22). As indicated by the Western immunoblot in Fig. 5, PGE 2 stimulated the appearance of a protein antigenically related to C/EBP␦ in nuclear extracts of these cells. The ϳ35-kDa immunoreactive protein comigrated with C/EBP␦ generated in a linked in vitro transcription-translation system using the cloned gene as a template. The detected protein also was con-sistent in size with the bands reacting with a 32 P-labeled HS3D oligonucleotide probe after UV-mediated DNA-protein crosslinking and Southwestern blotting shown in Fig. 2.
Next, transient overexpression of C/EBP proteins was tested to see if further stimulation of IGF-I transcription could be obtained in osteoblasts. Co-transfection experiments were performed with IGF-I promoter-luciferase reporter genes and expression plasmids for C/EBP␦ or C/EBP␤. Transfection with C/EBP␦ produced a 5-fold increase in basal luciferase activity compared with no change over control values in cells receiving either the parental (empty) expression plasmid or a C/EBP␤ expression plasmid. Treatment with PGE 2 of similarly transfected cells enhanced the activity of the IGF1711b-luciferase fusion gene in all cases (Fig. 6), in agreement with our previous results (20,22). In addition, forced expression of C/EBP␦ or C/EBP␤ significantly increased the magnitude of the response to PGE 2 treatment. By contrast, co-transfection of each C/EBP expression plasmid or the empty parental vector with IGF1711c, a promoter-reporter gene that lacked the HS3D site (20), did not lead to stimulated luciferase activity in the ab-sence or presence of PGE 2 . C/EBP expression plasmids also had no effect on a promoterless reporter gene, pOLuc. These results demonstrate that C/EBP␦ can transactivate IGF-I promoter 1 in osteoblasts through the HS3D region.
Additional experiments were conducted to determine if a neutral promoter could become responsive to PGE 2 by the addition of the HS3D element and to examine transcriptional activation by forced expression of C/EBP␦ in osteoblasts. Cotransfection with a C/EBP␦ expression plasmid induced activity of a luciferase reporter gene containing four tandem copies of a 19-nucleotide HS3D oligonucleotide cloned 5Ј to a minimal RSV promoter (25) but did not stimulate a fusion gene with four copies of the mutant HS3D oligomer, M6, that was unable to bind to osteoblast nuclear extracts (Fig. 7). Treatment with PGE 2 further increased expression of the wild-type reporter gene. By contrast, co-transfection with either an empty vector (parental) or an expression plasmid for C/EBP␤ did not enhance basal reporter activity; however, forced expression of C/EBP␤ did increase the response to PGE 2 (Fig. 7). These studies show that a multimer of HS3D provides a sufficient target for hormonal activation of gene expression in osteoblasts and indicates that C/EBP␦ and, to a lesser extent, C/EBP␤ induce gene activation through this response element.
To confirm the functional studies, osteoblast cultures were transiently transfected with C/EBP expression plasmids, and nuclear protein extracts were prepared for gel mobility shift experiments. Consistent with the results of promoter activation studies, nuclear extracts from untreated cultures where C/EBP␦ was overexpressed contained enhanced binding activity toward a 32 P-HS3D oligonucleotide probe (Fig. 8). Treatment with PGE 2 further stimulated protein-DNA complex formation. Although forced expression of C/EBP␤ did not alter the gel shift in nuclear extracts from control cultures, it did increase the response to PGE 2 , as was observed in the functional assays.
A nonosteoblast cell line, COS-7, was used to examine C/ EBP-stimulated DNA binding activity toward HS3D in a re-

FIG. 6. C/EBP overexpression enhances IGF-I promoter activation in osteoblasts.
Osteoblast-enriched cultures were transiently transfected with chimeric promoter-reporter genes, IGF1711b/Luc (containing the HS3D element), IGF1711c/Luc (lacking HS3D through deletion), or promoterless p0Luc, along with expression plasmid pSV7d (parental), pSV7d-C/EBP␤, or pSV7d-C/EBP␦, as described under "Experimental Procedures." After treatment with control medium (containing ethanol vehicle) or 1 M PGE 2 for 6 h, cytoplasmic extracts were prepared, and luciferase activity was determined. Results are shown from three independent experiments, where n ϭ 9. The asterisk indicates values from C/EBP co-transfections that are statistically different from data obtained using the parental expression vector (p Ͻ 0.05 for a comparison within similar treatments). constituted cell culture system. Transient transfection of a C/EBP␦ expression plasmid in COS-7 cells led to the appearance of a DNA-protein complex that co-migrated with binding activity detected in nontransfected, PGE 2 -treated primary rat osteoblasts. This complex was disrupted by antisera to C/EBP␦ but not by an antibody to C/EBP␤ (Fig. 9). In a reciprocal experiment, nuclear extracts from COS-7 cells transfected with a C/EBP␤ expression vector also gained DNA binding activity toward a 32 P-HS3D probe. This protein-DNA complex migrated faster on gel electrophoresis than the bands seen using C/ EBP␦-transfected cells and was inhibited only by an antibody to C/EBP␤ (Fig. 9).
To study functional aspects of C/EBP-regulated gene expression in COS-7 cells, reporter genes containing four copies of either the 19-bp natural HS3D site or the mutated sequence M6, cloned 5Ј to a minimal RSV promoter, were co-transfected with C/EBP expression plasmids. Co-transfection of the wildtype 4 ϫ HS3d-RSV-luciferase fusion gene with C/EBP␦ led to a 131 Ϯ 76-fold (mean Ϯ S.E.) increase in enzymatic activity compared with cells co-transfected with the empty parental vector, pcDNA3 (Fig. 10A). Luciferase activity was minimally enhanced in cells receiving the C/EBP␦ plasmid and a reporter gene containing either just the minimal RSV promoter or in addition four copies of the mutated HS3D sequence (Fig. 10A). Thus, C/EBP␦ is necessary and sufficient to mediate transcriptional activation of a gene containing tandem HS3D sites in a nonosteoblast cell.
The stimulatory effect of C/EBP␦ expression also was tested in COS-7 cells using increasing amounts of an expression plasmid and a fixed amount of a reporter gene containing four copies of the wild-type HS3D sequence. As little as 1 ng of a C/EBP␦ plasmid mediated a significant increase in luciferase expression compared with cells transfected with 100 ng of the empty expression vector, and 100 ng of the C/EBP␦ plasmid stimulated a 99 Ϯ 23-fold (mean Ϯ S.E.) rise in reporter gene activity. By contrast, co-transfection with 100 ng of the C/EBP␤ expression vector had no effect (Fig. 10B). Taken together, the results in Fig. 10 demonstrate that HS3D can function as an enhancer for C/EBP␦ in a nonosteoblast cell, and they support the role of C/EBP␦ in regulating IGF-I gene expression in primary rat osteoblasts. DISCUSSION The studies presented in this report identify C/EBP␦ as a PGE 2 -activated transcriptional regulator of the IGF-I gene in primary rat osteoblast cultures and demonstrate that the HS3D element within the 5Ј-untranslated region of exon 1 is a functionally important binding site for this protein. Osteoblast nuclear proteins of appropriate mobility on SDS-PAGE for C/EBP␦ bound to a labeled HS3D oligonucleotide, as determined by Southwestern blotting and UV-mediated protein- Osteoblast-enriched cultures were transiently transfected with recombinant luciferase reporter genes containing four copies of a 19-nucleotide HS3D oligonucleotide cloned 5Ј to a minimal RSV promoter. Construct ϩ4 ϫ HS3D contains wild-type HS3D sequence, while ϩ4 ϫ mut HS3D contains four tandem copies of a mutant HS3D oligomer, M6, that was unable to bind osteoblast nuclear proteins (22). Each promoter-reporter gene was transiently transfected with pSV7d, pSV7d-C/ EBP␤, or pSV7d-C/EBP␦, as described under "Experimental Procedures." After treatment with control medium (containing ethanol vehicle) or 1 M PGE 2 for 6 h, cytoplasmic extracts were prepared, and luciferase activity was measured. Results are shown from three independent experiments, where n ϭ 9. The asterisk indicates values from C/EBP co-transfections that are statistically different from results obtained with the parental expression plasmid (p Ͻ 0.05 for a comparison within similar treatments). DNA cross-linking. Mutational analysis of a C/EBP half-site within HS3D confirmed that it was required for osteoblast nuclear protein binding and for transcriptional activation by PGE 2 and by co-transfected C/EBP␦. Antibodies to C/EBP␦ but not to C/EBP␤ disrupted nuclear protein-HS3D DNA complexes as assessed by gel mobility shift assays. An antibody to C/EBP␦ also recognized an appropriately sized protein by Western immunoblot of osteoblast nuclear extracts only after treatment of cells with PGE 2 . Overexpression of C/EBP␦ in primary rat osteoblast cultures led to elevated basal transcription of an IGF-I promoter-reporter gene containing the HS3D site and to enhanced stimulation after PGE 2 but had no effect on a fusion gene lacking this sequence. Forced expression of C/EBP␤ also potentiated the response to PGE 2 , but to a lesser extent than C/EBP␦. Hormonal responsiveness and stimulation of gene activity by C/EBP␦ additionally were seen with a minimal promoter linked to four copies of wild-type HS3D but not to a mutated element that lacked the ability to bind nuclear proteins. Transcriptional activation of an HS3D-containing heterologous gene by C/EBP␦ but not C/EBP␤ also was observed in COS-7 cells. Taken together, these results show that C/EBP␦ functions as a hormonally stimulated transcriptional activator of the IGF-I gene in osteoblasts and that these actions of C/EBP␦ require the HS3D site located within IGF-I exon 1.
Members of the CREB/ATF family of nuclear proteins are generally responsible for activation of gene transcription by cAMP (37). We previously found that the kinetics of inducible protein-DNA binding at the HS3D site after PGE 2 treatment differed from the constitutive binding seen with an oligonucleotide containing a consensus CRE sequence (22), but we observed that a CRE oligomer could compete with HS3D for nuclear protein binding. Unlike current results with an albumin C/EBP probe, where cross-competition with HS3D was observed, the HS3D oligonucleotide did not displace a labeled CRE probe from nuclear proteins (22). The HS3D oligomer also did not bind to bacterially expressed CREB or ATF-1, even after these proteins were phosphorylated by the catalytic subunit of protein kinase, and overexpression of CREB or ATF-1 in osteoblast cultures did not transactivate IGF-I promoter-reporter genes containing an intact HS3D site. 2 These results are now explained with the identification of C/EBP␦ as the nuclear protein binding to HS3D. Previously, a rat C/EBP␦ cDNA was isolated by expression cloning using a CRE oligonucleotide probe (38), indicating that C/EBP␦, like several other members of the C/EBP family (37), can bind to a CRE with high affinity. Conversely, CREB shows minimal affinity for HS3D, but C/ EBP␦ binds with an affinity similar to that of an albumin D site probe.
In some cell types, C/EBP␤ has been shown to function as a cAMP-activated transcription factor (39). In these cells, C/ EBP␤ becomes phosphorylated by protein kinase A and is translocated to the nucleus, where it binds to target sites on specific genes, such as c-fos (39). Unlike C/EBP␤, C/EBP␦ appears to lack a consensus phosphorylation site for protein kinase A and has been found not to be phosphorylated by the enzyme (38). Thus, the mechanisms of activation of C/EBP␦ by PGE 2 or cAMP remain unclear. However, since the stimulation of IGF-I gene expression and the induction of nuclear protein binding toward HS3D by PGE 2 in osteoblasts does not require concurrent protein synthesis (22), it seems likely that a posttranscriptional mechanism is responsible. Although translocation of C/EBP␦ to the nucleus previously has not been observed after cAMP activation in any cell type, it has been seen in hepatocytes in response to the cytokine, tumor necrosis factor-␣ (40). In preliminary studies, we have observed a disappearance of C/EBP␦ from the cytoplasm of PGE 2 -treated osteoblasts and its rapid accumulation in nuclear extracts. 2 Further experiments are required to define the mechanisms of cyclic AMPdependent protein kinase-dependent activation of C/EBP␦ in these cells.
It is surprising that overexpression of C/EBP␤ failed to stimulate IGF-I promoter activity under basal conditions in osteoblasts, although it enhanced the response to PGE 2 (but to a lesser extent than C/EBP␦). Previously published studies have FIG. 10. C/EBP␦ activates a neutral promoter in COS-7 cells through the HS3D site. Panel A shows results of transient co-transfections with a C/EBP␦ expression plasmid or an empty expression vector (100 ng each) and luciferase reporter genes (1 g each) containing a minimal promoter (RSV promoter, RSV-LUC), or 4 copies of either a wild type or mutant HS3D site cloned 5Ј to the promoter (ϩ4 ϫ WT HS3D and ϩ4 ϫ mut HS3D, respectively). The asterisk indicates a significant enhancement of reporter gene expression (p Ͻ 0.05) for cells transfected with ϩ4 ϫ WT HS3D versus RSRV-LUC or ϩ4 ϫ mut HS3D. Panel B shows effects of co-transfection of the ϩ4 ϫ WT HS3D reporter gene with different quantities of a C/EBP␦ expression plasmid or with 100 ng of either the parental vector, pcDNA3, or a C/EBP␤ expression plasmid. The asterisks indicate significant increases in reporter gene expression for cells co-transfected with a C/EBP␦ expression plasmid versus pcDNA3 (*, p Ͻ 0.05; **, p Ͻ 0.02; ***, p Ͻ 0.01). All experiments in both panels were performed with the addition of 1 ng of a reporter gene, pRL-CMV, that expresses Renilla luciferase, which was used to correct for transfection efficiency. Mean Ϯ S.E. is shown of four experiments performed in duplicate.
indicated that expression of the highly homologous human IGF-I promoter can be induced by this protein after transient co-transfection into Hep3B cells (41). In these published experiments, a C/EBP binding site was mapped to the proximal part of promoter 1 in a location 5Ј to the HS3D site but also within a region of the promoter that is conserved between species (41) and present in both IGF1711b and IGF1711c luciferase plasmids used in Fig. 6. It is possible that osteoblasts have limiting amounts of an essential co-factor for gene activation by C/ EBP␤, that C/EBP␤ requires post-translational activation by cyclic AMP-dependent protein kinase or by another kinase in these cells, or that overexpression of C/EBP␤ in osteoblasts leads to high level production of LIP, a transcriptional inhibitor that represents a truncated form of C/EBP␤ derived by translation from an internal AUG codon (42). It is also surprising that overexpression of C/EBP␤ in COS-7 cells did not activate a promoter with four copies of HS3D, since nuclear proteins from these cells contained predominantly full-length C/EBP␤ 2 that did bind to an HS3D oligonucleotide in gel shift experiments. Again, perhaps a critical co-activator or activation step is absent from these cells.
Hormones that activate cAMP have been shown to stimulate IGF-I gene expression in other tissues, including the liver (43)(44)(45). PGE 2 and dibutyryl cAMP both enhanced IGF-I synthesis in bone marrow macrophages, although both paradoxically produced a decline in steady-state levels of IGF-I mRNA (46). Similarly, human chorionic gonadotropin, a hormone that increases cAMP production, caused a decrease in abundance of IGF-I mRNA and a slight decline in the rate of IGF-I gene transcription in rat Leydig cells (47). It thus appears that the response of the IGF-I gene to cAMP may depend on as yet uncharacterized cell type-specific factors.
In summary, we have demonstrated by multiple criteria that C/EBP␦ is the nuclear factor in osteoblasts that enhances IGF-I gene transcription in response to a signal transduction pathway regulated by activation of cAMP. Our results provide an outline of the critical components of this pathway, from the cell-surface parathyroid hormone and PGE 2 receptors to the genomic HS3D target site for C/EBP␦ on IGF-I promoter 1. Challenges for the future will be to unravel the specific intracellular mechanisms responsible for induction of C/EBP␦ in these cells and to understand how different hormones interactively regulate expression of the IGF-I gene in bone.