This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Quaedackers, M. E. Articles by Van Der Burg, B. Search for Related Content PubMed PubMed Citation Articles by Quaedackers, M. E. Articles by Van Der Burg, B.

4-Hydroxytamoxifen Trans-Represses Nuclear Factor-B Activity in Human Osteoblastic U2-OS Cells through Estrogen Receptor (ER), and Not through ER¹

Hubrecht Laboratory (M.E.Q., C.E.V.D.B., S.W., R.H.M.M.S., P.T.V.D.S., B.V.D.B.), Netherlands Institute for Developmental Biology, Uppsalalaan 8, 3584 CT Utrecht, the Netherlands; and Department of Medical Nutrition (J.-Å.G.), Karolinska Institute, Novum, S-14186 Huddinge, Sweden

Address all correspondence and requests for reprints to: Dr. Bart van der Burg, Hubrecht Laboratory, Netherlands Institute for Developmental Biology, Uppsalalaan 8, 3584 CT Utrecht, The Netherlands. E-mail: bvdb{at}niob.knaw.nl

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
Estrogens are important mediators of bone homeostasis, and postmenopausal estrogen replacement therapy is extensively used to prevent osteoporosis. The biological effects of estrogen are mediated by receptors belonging to the superfamily of steroid/thyroid nuclear receptors, estrogen receptor (ER) and ER. ER, not only trans-activates target genes in a hormone-specific fashion, but it can also neutralize other transcriptional activators, such as nuclear factor (NF)-B, causing repression of their target genes. A major mechanism by which estrogens prevent osteoporosis seems to be repression of transcription of NF-B target genes, such as the osteoclast-activating cytokines interleukin-6 and interleukin-1. To study the capacity of both ERs in repression of NF-B signaling in bone cells, we first carried out transient transfections with ER or ER of the human osteoblastic U2-OS cell line, in which endogenous NF-B was stimulated by tumor necrosis factor . Repression by ER was already observed without 17-estradiol, whereas addition of the ligand increased repression to 90%. ER, however, was able to repress NF-B activity only in the presence of ligand. Because it is known that some antiestrogens can also display tissue-specific agonistic properties, 4-hydroxytamoxifen was tested for its capacity in repressing NF-B activity and was found to be active (albeit less efficient than 17-estradiol) and, interestingly, only with ER. The pure antagonist ICI 164,384 was incapable of repressing through any ER subtypes. Deletion analysis and the use of receptor ER/ER-chimeras showed that the A/B domain, containing activation function-1, is essential for this suppressive action. Next, we developed stable transfectants of the human osteoblastic U2-OS cell line containing ER or ER in combination with an NF-B luciferase reporter construct. In these cell lines, repression of NF-B activity was only mediated through ER and not through ER. These findings offer new insights into the specific role of both ER subtypes in bone homeostasis and could eventually help in developing more specific medical intervention strategies for osteoporosis.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
DURING ADULT LIFE, bone tissue is continuously remodeled, which is needed to replace damaged or aged bone. Bone remodeling is a dynamic process involving bone formation by osteoblasts and bone resorption by osteoclasts. To maintain a constant bone mass, osteoblastic and osteoclastic activities are closely coordinated in a homeostatic system (1, 2). Disturbances of this balance are responsible for a variety of skeletal diseases, e.g. osteoporosis. Osteoporosis is associated with accelerated bone loss caused by increased bone resorption relative to bone formation. Estrogen deficiency is the main cause of osteoporosis in postmenopausal women and is contributing to bone loss in aging men (3). An effective treatment to prevent the development of osteoporosis is estrogen replacement therapy (ERT) (4). The beneficial effect of estrogen as regulator of bone homeostasis is well established; however, the precise molecular mechanisms involved in the action of estrogen in bone are still unclear.

The major action of estrogen on the skeleton in vivo is the inhibition of bone resorption. Bone resorption is inhibited indirectly by suppressing the production of bone-resorbing cytokines in osteoblasts. These cytokines, such as interleukin (IL)-1, IL-6, and tumor necrosis factor (TNF), enhance osteoclast differentiation and activity. Therefore, estrogen deficiency leads to an increase of bone resorbing cytokines in blood and bone marrow; and this, in turn, leads to increased bone resorption (5).

The biological effects of estrogen are mediated by the estrogen receptor (ER), a ligand-dependent transcription factor belonging to the superfamily of steroid/thyroid nuclear receptors. ER exists as two subtypes, ER (6) and ER (7, 8), which share a well-conserved modular structure composed of six functional domains, labeled A–F. The C domain, the DNA-binding domain, is nearly completely conserved between ER and ER (±96% homology); whereas the ligand-binding domain, E domain, is clearly less conserved (±58% homology). The N-terminal A/B region is the least conserved between the two receptors (±20% homology) (7, 8). A separate gene encodes each subtype. Differences in splicing or initiation of translation cause the existence of ER isoforms. For the human ER, two major isoforms are known, one protein of 530 and one of 485 amino acid residues (9).

After ligand binding, the ER undergoes a conformational change, displacing the inhibitory heat-shock protein complex and permitting the receptor to form a dimer. In this form, the ER is capable of binding specific DNA sequences, termed estrogen response elements (EREs), located in the regulatory region of target genes. After DNA binding, the receptor is able to interact with basal transcription factors and/or coregulatory proteins to regulate target gene transcription (10). The ER interacts with the transcriptional complex through two distinct activation functions (AFs), AF-1 in the N-terminal A/B domain and AF-2 in the ligand-binding domain (11).

Recent studies have shown alternative pathways through which ER also participates in regulating target gene transcription. These include pathways in which ER interacts indirectly with nonclassical ERE target gene promoters, e.g. by binding other DNA-bound transcription factors, such as AP-1 and Sp1 (12, 13, 14, 15). In addition, other studies have shown that ER can act as a transcriptional repressor, by inhibiting the activity of transcription factors such as nuclear factor-B (NF-B). NF-B is present, in an inactive form, in the cytoplasm associated with an inhibitory protein, IB. A number of agents, including inflammatory cytokines, cause translocation of NF-B to the nucleus through phosphorylation and subsequent degradation of IB. Because NF-B-binding sites have been identified in the promoters of genes encoding bone-resorbing cytokines, such as IL-6, the hypothesis has been put forward that estrogen is capable of inhibiting cytokine production in osteoblasts by repressing NF-B activity (16, 17, 18, 19). A similar mechanism has been described extensively for the repression of inflammatory responses, involving cytokines, by the glucocorticoid receptor (20). Depending on the cell type, repression may involve direct protein-protein interactions but may also involve steroid-induced stabilization of IB (21, 22).

Besides regulating bone homeostasis, estrogen is an essential regulatory hormone in female and male reproduction systems, in the central nervous system and in the cardiovascular system. Consequently, estrogen deficiency in postmenopausal women can cause undesirable symptoms and other diseases besides osteoporosis. Thus, exposure to ERT has more beneficial effects besides maintaining bone mass, e.g. the prevention of cardiovascular diseases (23), improvement of cognitive functions, and prevention of Alzheimer’s disease (24). Unfortunately, ERT is associated with side effects, including an increased risk for breast and uterine cancer (25). An important alternative to ERT for the prevention of osteoporosis is the use of selective ER modulators (SERMs). SERMs are compounds that bind with high affinity to ER and (depending on the tissue type) display estrogen agonistic or antagonistic activities. Tamoxifen [a triphenylethylene (26)] and raloxifene [a benzothiophene (27)] have been reported to have tissue-specific responses and are thus termed SERMs.

To study the molecular mechanism by which estrogen and SERMs maintain bone homeostasis through the ER, we investigated the effect of several estrogenic compounds on NF-B activity in a human osteoblastic cell type, U2-OS. Because no functional ER was detectable in this cell line, we developed U2-OS clones that express physiological levels of ER or ER. In the present study, we show that both 17-estradiol (E₂) and 4-hydroxytamoxifen (OH-T) selectively trans-repress NF-B activity in osteoblasts through ER and not through ER. These findings provide new insights into the possible molecular mechanism of estrogen action in bone tissue.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Chemicals and reagents
The steroid E₂ was obtained from Sigma (St. Louis, MO). OH-T and ICI 164,384 (ICI) were kind gifts from Dr. A. Wakeling (Zeneca Pharmaceuticals, Macclesfield, UK). Recombinant human TNF was obtained from Roche Molecular Biochemicals (Mannheim, Germany).

Cell culture
Human osteosarcoma osteoblastic U2-OS cells and human embryonal kidney 293 cells were obtained from the American Type Culture Collection (Rockville, MD). Both cell lines were cultured in a 1:1 mixture of DMEM and Ham’s F-12 medium (DF; Life Technologies, Inc., Gaithersburg, MD), buffered with bicarbonate and supplemented with 7.5% FCS (Integro, Linz, Austria). Cells were cultured at 37 C in a 7.5% CO₂ humidified atmosphere.

Plasmids
The expression vector encoding human ER (pSG5-HEGO) was a kind gift of Dr. P. Chambon (Strasbourg, France). Chimeric human ER/ER contains the A/B domain of ER fused with C, D, E, and F domains of ER, whereas chimeric human ER/ER contains the A/B domain of ER fused with C, D, E, and F domains of ER (28) and were kind gifts of Dr. B. S. Katzenellenbogen. Mouse ER (mER) (pMT2MOR) and mutants ER 1–339 and ER 121–599 (29) were kindly provided by Dr. M. G. Parker (London, UK).

The estrogen-responsive reporter plasmid 3xERE-tata-Luc, which contains three copies of a consensus ERE oligonucleotide and a TATA box in front of the luciferase complementary DNA (cDNA), has been described before (30). The luciferase reporter construct 4xNF-B(HIV)tkluc, which contains four copies of a NF-B-binding sequence from the HIV long-terminal repeat (LTR) placed in front of the thymidine kinase promoter coupled to luciferase, was described previously (31). The estrogen-responsive reporter gene plasmid 3xERE-TATA--Galactosidase was a kind gift of Dr. J. G. Lemmen (Utrecht, The Netherlands). This construct contains three copies of a consensus ERE containing oligonucleotide and a TATA box in front of the -galactosidase cDNA in pUC18.

Transient transfection and luciferase assay
U2-OS cells were cultured in 12-well plates and 293 cells in 24-well plates. Both cell lines were cultured in phenol red-free DF medium containing 0.2% BSA, 10 µg/ml transferrin, and 30 nM selenite supplemented with 5% dextran-coated charcoal (DCC) FCS. DCC-FCS was prepared by treatment of FCS with DCC to remove steroids, as described previously (32). Cells were transfected using the calcium-phosphate precipitation method (33). U2-OS cells were transfected with a total amount of 3.33 µg DNA/well, consisting of a mixture of 1.0 µg luciferase reporter plasmid, 1.0 µg PDM-LacZ plasmid, 0.33 µg of the indicated ER expression plasmid, and 1.0 µg pBluescript SK^-. 293 cells were transfected with a total amount of 1.6 µg DNA/well, consisting of a mixture of 0.6 µg luciferase reporter plasmid, 0.6 µg PDM-LacZ plasmid, 0.2 µg of the indicated ER expression plasmid, and 0.2 µg pBluescript SK^-. After 16 h, the medium was refreshed; and, when indicated, antihormones and/or TNF was added to the medium (1:1000). Cells were harvested 24 h later and were assayed for luciferase activity using the Luclite luciferase reporter gene assay kit (Packard Instruments, Meriden, CT), according to the manufacturer’s protocol, in a Topcount liquid scintillation counter (Packard Instruments). Values were corrected for transfection efficiency by measuring -galactosidase activity (34). Luciferase activity in stable transfectants was assayed in a similar fashion. EC₅₀ values of reporter gene induction were determined as described before (30).

Establishment of stable transfectants of U2-OS cells
To obtain stable hER or hER transfectants of U2-OS, cells were grown in 6-well plates, until approximately 50% confluency, before they were transfected with the calcium-phosphate coprecipitation method. A total amount of 10 µg DNA/well was used, consisting of a mixture of 8 µg PSG5 based-expression vector encoding either human ER or ER and 2 µg of a selection plasmid encoding a neomycin-resistance gene. After 16 h, the medium was refreshed; and, 24 h later, cells were trypsinized and replated in the presence of geneticin (G418; 200 µg/ml). After 10 days, surviving colonies were isolated and established as stable cell lines. Stable U2-OS/ER cells were cultured under prolonged G418 selection. Reporter genes [4xNF-B(HIV)tkluc or 3xERE-tata-Luc] were transfected similarly using a selection plasmid encoding a hygromycin-resistance gene. Stable double transfectants were grown under prolonged G418 (200 µg/ml) and hygromycin (50 µg/ml) selection.

RT-PCR
Clones of U2-OS cells, stably transfected with an expression vector encoding hER or hER, were cultured in 100-mm dishes, and total RNA was isolated using the acid-phenol method (35). Five micrograms of RNA was treated with 14 U RQ1 deoxyribonuclease (Promega Corp., Madison, WI) for 30 min at 37 C. One microgram of total RNA was incubated at 65 C for 3 min. After cooling, RNA was incubated for 90 min at 37 C with 200 U Superscript Reverse Transcriptase (Life Technologies, Inc.), Superscript buffer, 100 ng oligo (dT), 1 x 10 mM dithiothreitol, and 500 µM of each deoxynucleotide triphosphate. One fifth of the first-strand product was added to a PCR amplification mixture containing 1x Goldstar reaction buffer, 1.5 mM MgCl2, 200 µM of each deoxynucleotide triphosphate, 0.5 U Goldstar Taq Polymerase (Eurogentec, Seraing, Belgium), 100 ng forward primer, and 100 ng reverse primer. For PCR of hER, forward primer 5'-GACAAGGGAAGTATGGCTATGGA-3' and reverse primer 5'-TTCATCATTCCCACTTCGTAGC-3' were used, corresponding to bp positions 799–822 and 1047–1026, respectively. For PCR of hER, forward primer 5'-TAGTGGT- CCATCGCCAGTTAT-3' and reverse primer 5'-G,GGAGCCACACTTCACCAT-3' were used corresponding to bp positions 125–146 and 518–499, respectively. All samples were positive, using primers for -actin. Mixtures were overlaid with mineral oil, and amplification was carried out for 39 cycles in a Perkin-Elmer Corp. (Wellesley, MA) DNA thermal cycler. Each cycle consisted of 1 min of denaturation at 96 C, 1 min of annealing at 55 C, and 1 min of extension at 72 C. The PCR reaction products were separated on 1.2% agarose gels containing ethidium bromide, to visualize the 393-bp (hER-primers) and 248 bp (hER-primers) PCR product.

Steroid-binding assay
Specific E₂ binding in stably transfected U2-OS/ER cells was measured by performing a saturation ligand-binding experiment. For this, U2-OS cells were grown in 6-well plates, for 2 days, until approximately 80% confluency, in phenol red-free medium containing 5% DF-DCC. Cells were rinsed with PBS, and intact monolayers were incubated with increasing concentrations of ³H-E₂ (Amersham Pharmacia Biotech, Little Chalfont, UK) with or without 200-fold excess unlabeled E₂ in serum and phenol red-free medium. After 1 h, a medium sample was counted in a liquid scintillation counter. Then, cells were rinsed two times with PBS and incubated for 1 h with 0.5 M NaOH. Cell lysates were counted in a liquid scintillation counter and estimated for protein concentration [Bio-Rad Laboratories, Inc. (Philadelphia, PA) protein assay]. The dissociation constant (K_d) and the number of receptor sites (B_max) were obtained using Scatchard analysis (36).

Western blotting analysis
For isolation of whole-cell extracts, U2-OS cells were cultured in 100-mm dishes, treated as described, and harvested in buffer containing 50 mM Tris (pH 7.4), 50 mM NaCl, 0.5% Nonidet P-40, 1 mM dithiothreitol, 1 mM phenylmethylsulfonylfluoride, 1 µg/ml aprotinin, and 1 µg/ml leupeptin at 4 C. Subsequently, cells were centrifuged for 15 min at 4 C, and protein concentration of the supernatant was determined by the Bio-Rad Laboratories, Inc. protein assay according to the manufacturer’s protocol. Twenty-five micrograms of extract was separated on SDS-PAGE gels and transferred to Immobilon (Millipore Corp., Bedford, MA). Blots were immunostained with a polyclonal antibody against IB (catalog no. 06–494, Upstate Biotechnology, Inc., Lake Placid, NY) according to the manufacturer’s protocol. Immunoreactive bands were visualized after incubation with a peroxidase-conjugated second antibody and enhanced chemiluminescence (Amersham Pharmacia Biotech).

Gel shift assay
Gel shift assays were carried out as described before (31). Briefly, nuclear extracts of U2-OS cells were incubated with [³²P]deoxycycidine triphosphate-labeled double-stranded oligonucleotides containing the B site from the HIV LTR and subsequently were run on nondenaturing polyacrylamide gels. Gels were dried and processed for autoradiography. Specificity of binding was assessed by competition with 100-fold molar excess of unlabeled oligonucleotide probe.

Statistical analysis
Data are represented as mean values ± SEM from at least three independent experiments. An unpaired Student’s t test was used to compare differences between mean values of two different treatments. Data for dose response studies were analyzed for statistical significance using one-way ANOVA. When the F test for the ANOVA reached statistical significance, differences between specific mean values were assessed by least-significant-difference (LSD) test (37). Differences of P < 0.05 were accepted as statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
ER and ER repress TNF-induced NF-B activity in osteoblastic and nonosteoblastic cells
To determine whether estrogen is capable of repressing NF-B activity in osteoblasts, we studied the effect of E₂ on TNF-induced NF-B activity in human osteoblastic U2-OS cells. These cells were chosen because they are relatively easy to transfect and showed a superior TNF inducible NF-B activity, compared with ROS 17/2.8, SaOs, and MG63 cells (data not shown). A luciferase reporter construct, 4xNF-B(HIV)tkluc, was used to measure the induction of NF-B activity. This construct contains four copies of a NF-B-binding sequence derived from the HIV LTR placed in front of the thymidine kinase promoter and luciferase. Cells were transiently transfected with this reporter in combination with an expression vector encoding human ER or ER. Cotransfection of ER resulted in repression of the TNF-induced transcriptional activity of NF-B, already in the absence of E₂ (70%; P < 0.001), whereas addition of E₂ resulted in enhanced repression (90%). However, cotransfection of ER without ligand resulted in up-regulation of TNF-induced NF-B activity (P < 0.05), whereas addition of E₂ gave considerable repression (60%) of NF-B activity (Fig. 1A).

View larger version (21K): [in this window] [in a new window]   Figure 1. Repression of TNF-induced NF-B activity by E2 through human ER and human ER in U2-OS and 293 cells. U2-OS (A) and 293 (B) cells were transiently transfected with 4xNF-B(HIV)tkluc reporter plasmid in combination with empty expression vector (control, C) or expression vector encoding ER or ER. Sixteen hours after transfection, cells were treated with vehicle (0.1% ethanol = no treatment; black bars), with 250 U/ml TNF plus vehicle (white bars), or with TNF + 10-8 M E2 (hatched bars). After 24 h, cells were harvested and assayed for luciferase activity. Luciferase values were corrected for transfection efficiency by measuring -galactosidase activity of cotransfected PDM-LacZ construct. Values are represented as the induction of luciferase activity evoked by ER over untreated cells transfected with empty expression vector. Bars, Means ± SEM of four independent experiments; **, P < 0.01; ***, P < 0.001 (by Student’s t test).

These results indicate that both ER and ER can act as transcriptional repressors of TNF-induced NF-B activity on a NF-B reporter construct after E₂ stimulation in human osteoblastic U2-OS cells and nonosteoblastic 293 cells. The observation that ER, and not ER, can act as a transcriptional repressor in the absence of hormone in U2-OS cells, and not in 293 cells, suggests that hormone-independent repression is receptor- and cell type-specific.

OH-T-ligated ER represses NF-B activity in osteoblastic cells
Tamoxifen was initially developed as a drug to treat breast cancer (26) because it acts as an antiestrogen in breast tissue. However, tamoxifen is also known to have beneficial effects on bone density and serum lipids in postmenopausal women by acting as estrogen agonist (38). Because of its tissue-selective responses, tamoxifen can be designated a SERM.

To study the estrogen agonistic effect of tamoxifen in bone cells, we were interested in learning whether OH-T, the main metabolite of tamoxifen, was capable of repressing NF-B activity in U2-OS cells, similar to E₂. For this, U2-OS cells were transiently transfected with 4xNF-B(HIV)tkluc in combination with an expression vector containing ER or ER. Cells were cotreated for 24 h with TNF and increasing concentrations of E₂, OH-T, or ICI. E₂ (10^-8 M) caused about 70% decrease in TNF-induced reporter activity, both through ER and ER (Fig. 2, A and B, respectively). OH-T functioned as estrogen agonist in repressing NF-B activity but only through ER and not through ER (Fig. 2). However, OH-T (10^-8 M) was less effective (40%) in repressing NF-B, compared with E₂ (Fig. 2A). The pure antagonist ICI did not show estrogen agonistic activity. However, a dose-dependent increase in NF-B activity was observed in the presence of ICI in combination with ER and ER (Fig. 2). Remarkably, also OH-T showed this increase with ER (Fig. 2B). Similar experiments were performed in 293 cells. Although, in these cells, E₂ repressed TNF-induced NF-B activity through both receptor subtypes, OH-T and ICI had no effect on NF-B activity (data not shown). These results indicate that OH-T is a selective estrogen agonist in an osteoblastic cell type and selectively trans-represses NF-B activity through ER.

View larger version (16K): [in this window] [in a new window]   Figure 2. Dose response curves of the repression of NF-B activity by different ER ligands through human ER and ER in U2-OS cells. Cells were transiently transfected with 4xNF-B(HIV)tkluc reporter plasmid in combination with expression vector encoding ER (A) or ER (B). Cells were treated with 250 U/ml TNF plus vehicle or in combination with increasing doses (10-11–10-8 M) of E2 (•), OH-T (), or ICI () and were assayed, after 24 h, as described in Fig. 1. The results are expressed relative to luciferase activity caused by TNF in the absence of ER ligand (100%). Values represent the means ± SEM of four independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (by one-way ANOVA and LSD for differences between TNF treatment alone vs. TNF plus ER ligand).

Identification of ER domains involved in the trans-repression of NF-B

View larger version (29K): [in this window] [in a new window]   Figure 3. Requirement of the A/B domain of mER in repressing NF-B activity. A, U2-OS cells were transiently transfected with 4xNF-B(HIV)tkluc reporter construct in combination with empty expression vector or expression vectors encoding mER, mER 121–599, or mER 1–399. Cells were treated with 250 U/ml TNF plus vehicle (black bars) or in combination with 10-8 M E2 (white bars), OH-T (hatched bars), or ICI (double hatched bars) and assayed after 24 h as described in Fig. 1. Values are expressed relative to luciferase activity of cells transfected with empty expression vector treated with TNF (100%). B, U2-OS cells were transiently transfected with 3xERE-tata-Luc reporter construct in combination with empty expression vector or expression vectors encoding mER, mER 121–599, or mER 1–399. Cells were left untreated (black bars) or treated with 10-8 M E2 (white bars), OH-T (hatched bars), or ICI (double hatched bars) and assayed after 24 h as described in Fig. 1. Values are expressed as the induction of luciferase activity evoked by ER and treatment over untreated cells transfected with empty expression vector. Bars, Mean ± SEM of three independent experiments; *, P < 0.05; **, P < 0.01; ***, P < 0.001 (by Student’s t test for cells treated with TNF alone vs. TNF in combination with (anti)estrogen or untreated vs. treated with (anti)estrogen).

Combining the effects of mutant ERs on activity of the NF-Bluc reporter and the EREluc reporter, we can summarize the following. Wild-type ER is already active without ligand activation, both in ERE transactivation and repression of NF-B activity, and this depends on the A/B domain. The natural ligand E₂ can enhance basal activity of the receptor, leading to increased ERE transactivation and stronger NF-B repression, and this is only partly dependent on the A/B domain. The synthetic compound OH-T mimics the estrogen effect, but less effectively, and this phenomenon is completely dependent on the A/B domain.

Comparison between the A/B domain of ER and ER in repressing NF-B activity
To further determine the importance of the A/B domain of ER in NF-B repression, chimeric constructs of hER and hER were used. Chimeric human ER/ER contains the A/B domain of ER fused to C, D, E, and F domains of ER; and chimeric human ER/ER contains the A/B domain of ER fused to C, D, E, and F domains of ER (28). In transient transfection assays, the different receptor constructs of ER were cotransfected with the NF-Bluc reporter in U2-OS cells, and cells were treated with TNF in combination with several ligands (Fig. 4A). Replacement of the A/B region of ER with the A/B region of ER (ER/ER) resulted in a receptor that was at least equally potent as wild-type ER in repressing NF-B without ligand. However, replacement of the A/B region of ER with the A/B region of ER (ER/ER) abolished the ligand-independent repression by ER. Thus, ligand-independent actions of ER and ER were interchanged by switching their A/B domain. However, exchanging the A/B domains did not simply switch the effects caused by OH-T and ICI, suggesting that other factors besides the A/B domain are also involved in their agonist/antagonist action. All receptor constructs were capable of repressing NF-B activity mediated by E₂ because this action is not dependent on the A/B domain, as shown in Fig. 3A.

View larger version (26K): [in this window] [in a new window]   Figure 4. Requirement of the A/B domain of human ER in repressing NF-B activity. A, U2-OS cells were transiently transfected with 4xNF-B(HIV)tkluc reporter construct in combination with empty expression vector or expression vectors encoding hER, hER, hER/, or hER/. Cells were treated with TNF alone (black bars) or in combination with 10-8 M E2 (white bars), OH-T (hatched bars), or ICI (double hatched bars) and assayed after 24 h as described in Fig. 1. Values are expressed relative to luciferase activity of cells transfected with empty expression vector treated with TNF (100%). B, U2-OS cells were transiently transfected with 3xERE-tata-Luc reporter construct in combination with empty expression vector or expression vectors encoding hER, hER, hER/, or hER/. Cells were left untreated (black bars) or treated with 10-8 M E2 (white bars), OH-T (hatched bars), or ICI (double hatched bars) and assayed after 24 h as described in Fig. 1. Values are expressed as the induction of luciferase activity evoked by ER and treatment over untreated cells transfected with empty expression vector. Bars, Mean ± SEM of three independent experiments; *, P < 0.05; **, P < 0.01; ***, P < 0.001 (by Student’s t test for cells treated with TNF alone vs. TNF in combination with (anti)estrogen or untreated vs. treated with (anti)estrogen).

In summary, these findings support the idea that the A/B domain of ER is important for ligand-independent transcriptional activity, both in NF-B repression and ERE induction. Moreover, the A/B domain of ER is important for agonistic activation by OH-T.

Development of U2-OS cells stably expressing ER and reporter genes
A potential drawback of transient transfection assays is that, because of overexpression, physiologically irrelevant responses can be obtained. To verify our observations in a more natural situation, we developed clones of U2-OS that are stably transfected with an expression vector encoding hER or hER. Several ER and ER clones were screened for expression of ER mRNA using RT-PCR analysis. Stable clones with ER and ER were found to express significant amounts of ER and ER mRNA, respectively (data not shown). Transient transfection of the luciferase reporter construct 3xERE-tata-Luc was used to determine whether the expressed ER in the stable clones was capable of transcriptional activation. In all clones, a dose-dependent up-regulation of ERE-luciferase activity was observed after treatment with increasing concentrations of E₂, with EC₅₀ values roughly 10-fold lower in ER-, compared with ER-expressing, clones (data not shown). We selected one representative ER and one ER clone to determine the level of expressed ER protein and the binding affinity of E₂ for ER and ER, respectively. In the ER clone, the level of ER expressed was 238 (±18) fmol/mg protein, with a K_d value of 5 x 10^-11 M, which is in the same range as T47D cells expressing ER endogenously (39). In the ER clone, the level of ER was 296 (±64) fmol/mg, with a K_d value of 3 x 10^-10 M. These results indicate that E₂ has a higher affinity for ER, compared with ER, in this cellular context.

Next, we developed clones of U2-OS stably expressing 3xERE-tata-Luc in addition to either ER or ER. In these lines, half-maximal ERE induction is reached at about 10^-11 M E₂ for ER cells and about 10^-10 M E₂ for ER cells (data not shown), consistent with the differences in receptor-binding activity.

NF-B repression is not mediated by ER in U2-OS cells expressing physiological levels of endogenous ER
We also isolated stably transfected U2-OS lines expressing the 4xNF-B(HIV)tkluc reporter gene in addition to either ER or ER. In these cell lines, the integrity of ERE-mediated transcriptional response was determined by transient transfection of an ERE-tata-LacZ construct. Both ER- (Fig. 5A) and ER-expressing (Fig. 5B) cell lines showed significant (15- to 20-fold) ERE-dependent transcriptional activation. In the ER-expressing line, both E₂ and OH-T effectively trans-repress NF-B reporter gene activity (Fig. 5C). Half-maximal NF-B repression was observed at similar concentrations of ligand, compared with transcriptional activity of the ERE construct (Fig. 5A). Surprisingly, NF-B activity was only slightly (15%) repressed by E₂ in cells stably expressing ER (Fig. 5D). The observation that ER strongly (70%) represses NF-B activity after E₂ addition in transient transfection assays (see Fig. 2), and not in cells stably expressing ER, could be caused by the fact that the receptor is often more highly expressed in cells that are transiently transfected. This was confirmed by the fact that, when the reporter gene was transiently induced in several clones expressing ER or ER only, again only ER-mediated repression of a 4xNF-B(HIV)tkluc reporter gene was observed (data not shown). To further investigate this point, wild-type U2-OS cells were transiently transfected with an increasing amount of expression vector encoding ER or ER, ranging from 0.01–1.33 µg/well, in combination with 4xNF-B(HIV)tkluc. Cells were treated for 24 h with TNF alone or in combination with E₂ (10^-8 M). As shown in Fig. 6, it became clear that E₂-ligated ER could efficiently repress NF-B activity already at much lower amounts of transfected expression vector than ER. At 0.033 µg/well NF-B, activity is repressed by 60% by ER (Fig. 6A), whereas no significant repression is observed with ER (Fig. 6B). At 0.33 µg/well, used in transient transfection assays described above, both receptors efficiently repress NF-B when activated by E₂. In addition, the hormone- independent repression by ER is increased with increasing amounts of transfected receptor, whereas it is not observed with ER. These results indicate that NF-B repression by ER and ER is dependent on the expression level of the receptor. The observation that ER, and not ER, is capable of repressing NF-B activity in stable U2-OS-ER clones suggests that the expression level of ±200–300 fmol/mg protein is at the point where ER (but not ER) already efficiently represses NF-B.

View larger version (26K): [in this window] [in a new window]   Figure 5. Activation of transient ERE-LacZ reporter and repression of stable NF-Bluc reporter in stable U2-OS/ER-NFBluc cells. U2-OS/ER-NFBluc (A) and U2-OS/ER-NFBluc (B) cells were transiently transfected with 3xERE-tata-LacZ reporter construct and treated with vehicle or with increasing doses (10-11–10-8 M) of E2. After 24 h, cells were harvested and assayed for -galactosidase activity. The results are expressed relative to the maximal -galactosidase activity caused by E2 (100%). U2-OS/ER-NFBluc (C) and U2-OS/ER-NFBluc (D) were treated with 250 U/ml TNF plus vehicle or with TNF in combination with increasing doses (10-11–10-8 M) of E2 (•), OH-T (), or ICI (). After 24 h, cells were harvested and assayed for luciferase activity. The results are expressed relative to luciferase activity caused by TNF in absence of ER ligand (100%). Values represent the means ± SEM of three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (by one-way ANOVA and LSD for differences between no-treatment vs. estrogen or TNF treatment alone vs. TNF plus ER ligand).

View larger version (24K): [in this window] [in a new window]   Figure 6. Repression of NF-B activity is dependent on the amount of transfected ER-expression vector. U2-OS cells were transiently transfected with 4xNF-B(HIV)tkluc reporter plasmid in combination with empty expression vector (0 µg point) or increasing amounts (0.01–1.33 µg) of expression vector encoding ER (A) or ER (B). Sixteen hours after transfection, cells were treated with vehicle (0.1% ethanol = no treatment), with 250 U/ml TNF plus vehicle (white bars), or with TNF + 10-8 M E2 (black bars). After 24 h, cells were harvested as described in Fig. 1. The results are expressed relative to luciferase activity of cells transfected with empty expression vector treated with TNF (100%). Values represent the means ± SEM of three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (by Student’s t test for cells transfected with empty expression vector vs. expression vector encoding ER).

Repression of TNF-induced NF-B activity is not caused by induction of IB and is does not lead to decreased NF-B DNA-binding activity

To determine whether estrogen has indirect effects on NF-B activity through modulating IB levels, we analyzed IB expression in protein extracts from U2-OS cells, by Western blotting. Normal U2-OS cells or stable ER- or ER-cells were stimulated for 6 h with various combinations of estrogenic compounds and TNF, as indicated (Fig. 7; A, B, and C, respectively). Stimulation with TNF leads to a decrease in the level of IB protein (compare lanes 1 and 2). Stimulation with E₂, OH-T, or ICI alone did not influence the expression of IB (compare lane 1 with lanes 3, 4, and 5). In addition, the TNF-induced degradation of IB was not influenced by E₂, OH-T, or ICI (compare lane 2 with lanes 6, 7, and 8). There was no difference in the expression pattern between cells without ER (Fig. 7A) or clones stably expressing ER (Fig. 7B) or ER (Fig. 7C). These results indicate that E₂, OH-T, and ICI are not able to modulate IB levels in U2-OS cells. Thus, the observed trans-repression of NF-B activity by E₂ or OH-T cannot be ascribed to changes in IB protein levels.

View larger version (42K): [in this window] [in a new window]   Figure 7. Expression of IB protein in U2-OS cells and binding activity of NF-B to DNA. IB protein expression of wild-type (A), ER-expressing (B), or ER-expressing (C) U2-OS cells. Cells were treated with TNF in the absence or presence of E2, OH-T, or ICI for 6 h. Western blots of whole-cell extracts were immunostained with a polyclonal antibody to IB. D, Wild-type and ER- or ER-expressing cells were treated with TNF or E2 alone or with TNF plus E2. Nuclear extracts were analyzed by electrophoretic mobility shift assay with 32P-labeled probe containing the B site from the HIV LTR.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

 
In the present study, we show that ER, and not ER, is the major ER through which transcription of NF-B-regulated genes is inhibited in osteoblastic cells. ER only repressed NF-B activity at relatively high expression levels and only when activated by a full agonist. In contrast, ER already repressed NF-B activity in the absence of ligand. Both the agonist E₂ and, in different contexts, the antagonist OH-T were capable of enhancing this repressive effect. Our results also show that important domains involved in the regulation of NF-B activity are found in the AB and E regions of the two receptor subtypes. Differences in the nonconserved AB region were found to determine much of the differences observed between repression through ER and ER.

It has been observed before that, in some cell types, glucocorticoids and estrogens (41, 42, 43) lead to enhancement of IB expression, thereby decreasing NF-B activity. This effect is cell type-specific and contributes only in part to glucocorticoid-mediated NF-B repression (21, 22). Although estrogens seem able to induce IB under certain conditions (43), our results clearly show that this mechanism is not involved in NF-B repression in U2-OS cells. Evidence for an alternative model of repression does point to a direct physical interaction between ER and NF-B, involving the ER DBD (18, 44). The mechanism of repression in U2-OS may involve formation of an inactive ER/NF-B complex that seems to be to labile to resist the conditions used in our gel shift assays. Others obtained inconsistent results using the IL-6 promoter in gel shift assays, with respect to the ability of estrogens to repress NF-B-binding activity (44, 45, 46). However, because in some studies a repressive effect on DNA binding has been found, the formation of a labile complex, at present, seems the most attractive model.

Our data suggest that, in addition to the ER DBD (18, 44), the AB and E domains may contain additional sites involved in repression. Interestingly, major functional domains in both the AB and E regions of ERs are the regions essential for transcriptional activation, AF-1 and AF-2, respectively. Our results, showing a high correlation between transcriptional activation and transcriptional repression, are consistent with an additional role of these domains in transcriptional repression. First of all, the role of AF-1 in repression is supported by the fact that OH-T-mediated repression is highly dependent on the presence of the AB domain of ER. In contrast to ER, the AB domain of ER contains a strong and independent AF (28). AF-1 cannot be activated by pure antiestrogens like ICI (11, 47, 48, 49), which also is unable to induce repression. Estrogens are potent AF-2 activators and repress NF-B activity even in the absence of the AB domain of ER.

Strikingly, a dose-dependent activation of NF-B activity by ICI was observed, both through ER and ER. This could involve a decrease in expression of ER protein levels, which has been described to occur upon ICI treatment (50, 51). It could also involve specific ER conformation-dependent effects on NF-B activity, because NF-B activation also occurred through ER without addition of ligand, compared with cells that did not express ER; and also, OH-T dose-dependently increased NF-B activity through this receptor, suggesting that the effect occurs in the absence of repression through ER AF1. This is reminiscent of antagonist-induced activation of AP1 activity, which also occurs in the absence of ER AF1 (52). Our results show that different ligands have very different effects on ER-mediated NF-B repression. Both ER and ER can bind a large number of compounds, and each compound induces distinct conformational changes within the ER structure (53, 54, 55, 56, 57, 58), and this affects the presentation of the receptor surfaces to the transcriptional machinery. This, in turn, may alter the interactions of AF-1 and AF-2 with transcriptional coactivators/repressors and thus affect regulation of gene transcription, not only with respect to transcriptional activation but also repression. Clearly, it is of interest to decipher the molecular determinants involved in ER-mediated NF-B suppression and to find ways to pharmacologically modulate the repressive function of the receptor, which may lead to novel drugs to alter bone physiology.

The fact that ER and ER modulate transcription of NF-B-regulated genes in a different manner suggests that the two ER subtypes have different effects on cytokine production in osteoblasts and, consequently, bone formation. Analysis of the effects on bone in female ER knock-out mice has revealed that disruption of ER results in increased bone resorption (59). In contrast, in adult female ER knock-out mice, the bone mineral content is increased (60). The physiological importance of ER in humans was illustrated when a 28-yr-old male with estrogen resistance was reported. The patient was found to contain a point mutation in his ER gene, and he suffered from increased bone turnover and osteopenia, indicating that ER is important for normal bone remodeling in humans (61).

Interestingly, in osteoblastic SV-HFO cells, ER was expressed at a constant level, whereas ER expression gradually increased concomitantly with differentiation of the cell (62). We would expect, in the light of our current results, that osteoclast development would be inhibited through ER-mediated inhibition of production of cytokines. When the osteoblast reaches the latest phase of development, ER is highly expressed (62), and heterodimerization with ER (63, 64) could lead to diminished repression of NF-B activity and to an increase of osteoclast-activating cytokines. In this speculative model, the coupling of formation and resorption of bone, through the distinct actions of ER and ER, are needed to maintain a constant bone mass. This could mean that diseases like osteoporosis, in which bone homeostasis is disrupted, could be treated by selectively inhibiting or activating one of the receptor subtypes to restore the balance of bone formation and resorption. Ligands that are selective agonists or antagonists of ER or ER may therefore be helpful therapeutic agents.

	Acknowledgments

 
We thank Dr. B. S. Katzenellenbogen, Dr. P. Chambon, and Dr. M. G. Parker for ER cDNAs. We thank Dr. A. Wakeling for providing us with OH-T and ICI. We would also like to thank Drs. J. G. Lemmen and Dr. A.W. Zomer for helpful discussion and assistance with the Scatchard analysis. We are very thankful to Dr. L.G Tertoolen, who excellently advised us about the statistical analysis.

	Footnotes

 
¹ This work was supported by the European Space Agency within the ERISTO project.

Received July 31, 2000.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Erlebacher A, Filvaroff EH, Giltelmann SE, Derynck R 1995 Toward a molecular understanding of skeletal development. Cell 80:371–378[CrossRef][Medline]
2. Reddi AH 1997 Bone morphogenesis and modeling: soluble signals sculpt osteosomes in the solid state. Cell 89:159–161[CrossRef][Medline]
3. Riggs BL, Khosla S, Melton III LJ 1998 A unitary model for involutional osteoporosis: estrogen deficiency causes both type I and type II osteoporosis in postmenopausal women and contributes to bone loss in aging men. J Bone Miner Res 13:763–773[CrossRef][Medline]
4. Eastell R 1998 Treatment of postmenopausal osteoporosis. N Engl J Med 338:736–746[Free Full Text]
5. Spelsberg TC, Subramaniam M, Riggs BL, Khosla S 1999 The actions and interactions of sex steroids and growth factors/cytokines on the skeleton. Mol Endocrinol 13:819–828[Free Full Text]
6. Walter P, Green S, Greene G, Krust A, Bornert JM, Jeltsch JM, Staub A, Jensen E, Scrace G, Waterfield M, Chambon P 1985 Cloning of the human estrogen receptor cDNA. Proc Natl Acad Sci USA 82:7889–7893[Abstract/Free Full Text]
7. Kuiper GJM, Enmark E, Peto-Huikko M, Nilsson S, Gustaffson JA 1996 Cloning of a novel estrogen receptor expressed in rat prostate and ovary. Proc Natl Acad Sci USA 93:5935–5930
8. Mosselman S, Polman J, Dijkema R 1996 ER: identification and characterization of a novel human estrogen receptor. FEBS Lett 392:49–53[CrossRef][Medline]
9. Ogawa S, Inoue S, Watanabe T, Hiroi H, Orimo A, Hosoi T, Ouchi Y, Muramatsu M 1998 The complete primary structure of human estrogen receptor (hER ) and its heterodimerization with ER in vivo and in vitro. Biochem Biophys Res Commun 243:122–126[CrossRef][Medline]
10. Tsai MJ, O’Malley BW 1995 Molecular mechanism of action of steroid/thyroid receptor superfamily members. Annu Rev Biochem 63:451–486[CrossRef][Medline]
11. Tora L, White J, Brou C, Tasset D, Webster N, Scheer E, Chambon P 1989 The human estrogen receptor has two independent nonacidic transcriptional activation functions. Cell 59:477–478[CrossRef][Medline]
12. Sukovich DA, Mukherjee R, Benfield PA 1994 A novel, cell-type-specific mechanism for estrogen receptor-mediated gene activation in the absence of an estrogen-responsive element. Mol Cell Biol 14:7134–7143[Abstract/Free Full Text]
13. Krishnan V, Wang X, Safe S 1994 Estrogen receptor-Sp1 complexes mediate estrogen-induced cathepsin D gene expression in MCF-7 human breast cancer cells. J Biol Chem 269:15912–15917[Abstract/Free Full Text]
14. Webb P, Lopez GN, Uht RM, Kushner PJ 1995 Tamoxifen activation of the estrogen receptor/AP-1 pathway: potential origin for the cell-specific estrogen-like effects of antiestrogens. Mol Endocrinol 9:443–456[Abstract]
15. Elgort MG, Zou A, Marschke KB, Allegretto EA 1996 Estrogen and estrogen receptor antagonists stimulate transcription from the human retinoic acid receptor- 1 promoter via a novel sequence. Mol Endocrinol 10:477–487[Abstract] <