Impaired Estrogen Sensitivity in Bone by Inhibiting Both Estrogen Receptor α and β Pathways

Abstract Although it is well established that estrogen deficiency causes osteoporosis among the postmenopausal women, the involvement of estrogen receptor (ER) in its pathogenesis still remains uncertain. In the present study, we have generated rats harboring a dominant negative ERα, which inhibits the actions of not only ERα but also recently identified ERβ. Contrary to our expectation, the bone mineral density (BMD) of the resulting transgenic female rats was maintained at the same level with that of the wild-type littermates when sham-operated. In addition, ovariectomy-induced bone loss was observed almost equally in both groups. Strikingly, however, the BMD of the transgenic female rats, after ovariectomized, remained decreased even if 17β-estradiol (E2) was administrated, whereas, in contrast, the decrease of littermate BMD was completely prevented by E2. Moreover, bone histomorphometrical analysis of ovariectomized transgenic rats revealed that the higher rates of bone turnover still remained after treatment with E2. These results demonstrate that the prevention from the ovariectomy-induced bone loss by estrogen is mediated by ER pathways and that the maintenance of BMD before ovariectomy might be compensated by other mechanisms distinct from ERα and ERβ pathways.

Postmenopausal osteoporosis is characterized by increased bone resorption due to estrogen deficiency (1,2). The supplement of estrogen markedly prevents the disease (3). Action of estrogen is generally supposed to be mediated by estrogen receptor (ER), 1 a member of the nuclear receptor superfamily (4 -6). However, the precise mechanism of ER responsible for the bone metabolism remains to be understood.
Both classical ER␣ and recently identified ER␤ (7) were differentially expressed during the osteoblast differentiation, implying their functional roles in the bone metabolism (8,9). Several studies have suggested that the estrogen action mediated by ER␣ cause the mineralization of bone-forming cells, the osteoblasts (10 -12). On the other hand, it is reported that the estrogen response element (ERE)-mediated transcription is not necessarily observed in the mineralizing process, suggesting the existence of the estrogen-induced but ER-independent signaling pathways (13).
In contrast to the report of a man homozygous for ER␣ gene mutation who suffered from severe osteoporosis based on the increased bone turnover (14), ER␣ knockout (ER␣KO) mice (15) were shown to decrease in their skeletal mineralization compared with wild-type mice to a small extent (16,17). The unclear phenotypes of ER␣KO mice in the bone may be partly because of the possible compensation of ER␣-mediated cascade by ER␤, as well as the involvement of the so-called nongenomic action of estrogen. These possibilities are also suggested in the vascular walls of ER␣KO mice, which show a response to estrogen similar to that of wild-type mice when they were injured (18,19).
We have recently demonstrated that a dominant negative mutant of human ER␣ constructed by C-terminal truncation, namely ER␣-(1-530), blocked both ER␣ and ER␤ signaling pathways (20). In order to clarify the roles of ERs in exerting the physiological effects of estrogen, we introduced the expression vector encoding the corresponding mutant of rat ER␣, namely rER␣-(1-535), into rats, thereby inhibiting both ER␣ and ER␤ signaling pathways in vivo. Taking advantage of the analytical merits of rat as a species for the physiological and the histomorphometrical studies over that of mice (21,22), we studied effects of estrogen on the bone metabolism using these transgenic rats. As a result, we have established a transgenic animal model with the impaired estrogen sensitivity in the bone, exhibiting the decreased mineralizing response after ovariectomy followed by 17␤-estradiol (E 2 ) administration.

EXPERIMENTAL PROCEDURES
Production of Transgenic Rats-The PCR-amplified products of rat ER␣-(1-535) (cDNA fragment encoding amino acids from 1 to 535) ( were cloned into pCXN2 vector (24) that has chicken ␤-actin promoter to construct pCXN2-ER␣-(1-535). The obtained DNA sequence was compiled and confirmed using DNASIS computer programs (Hitachi Co., Tokyo, Japan). The construction of transgenic rats were described elsewhere (22). Briefly, insert of pCXN2-ER␣-(1-535) digested with BamHI was microinjected into the male pronuclei of fertilized eggs from Wistar rats. Injected eggs were transferred to oviducts of pseudopregnant rats. Three independent founders, each with intact copies of transgene, were identified by Southern blot analysis. Two lines, designated line 3 and line 5, were able to transmit the transgenes to their offspring and were bred out into permanent lines. The copy numbers of the integrated transgenes were 10 and 8, respectively.
Study Protocol-Sixty 6-month-old Wistar rats, weighing approximately 300 g each, were used in this study. The rats were acclimated to the local vivarium conditions (24°C; 12-h light/12-h dark cycle), allowed free access to water and a pelleted commercial diet (PMI Feeds, Inc.) containing 0.60% calcium, 0.40% phosphorus, and 2.2 IU/g vitamin D 3 . All the experimental protocols were approved by the Animal Research Committee of Saitama Medical School. Two weeks prior to the operation, all rats were changed to be fed a pelleted low calcium diet (PMI Feeds, Inc.) containing 0.02% calcium instead of 0.60%.
The rats were sham-operated or bilaterally ovariectomized, and they were received subcutaneous daily injection of E 2 (10 g/kg body weight, Sigma) suspended in corn oil as vehicle or of only vehicle for 4 weeks. Double bone labelings were performed by the injection with demeclocyclin (Sigma) and calcein (Sigma) both at a dose of 15 mg/kg body weight, given intraperitoneally as described (25,26). After 4 weeks of treatment, the rats were sacrificed by cardiac puncture under ether anesthesia. Blood was obtained by cardiac puncture, allowed to clot at 4°C, and then centrifuged at 2000 ϫ g for 10 min. In all experiments, uteri were isolated and wet uterine weights were measured to confirm the effects of treatment.
Northern Blot Analysis-Total RNA was prepared from the tissues and primary osteoblast-like cells using acid guanidinium thiocyanatephenol-chloroform extraction as described previously (27). For each sample, 10 g of total RNA was separated in 1% agarose. Northern blot analysis was performed as described previously (28). The 32 P-labeled, 500-bp, EcoRI-NotI-digested fragment of pCXN2-rER␣-(1-535) was used as a probe. Autoradiography was carried out at Ϫ80°C with an intensifying screen for 1 day.
Cell Transfection and Whole Cell Extract Preparation-Osteoblastlike cells were isolated by three sequential enzymatic digestion as described previously (29). Briefly, 21-day-old rat embryo calvariae were incubated at room temperature for 20 min with gentle shaking of an enzyme solution containing 0.1% collagenase, 0.05% trypsin, and 4 mM EDTA in phosphate-buffered saline. Only the cells released from the fourth to sixth consecutive digests were cultured separately with ␣-modified minimum essential medium supplemented with 10% fetal calf serum and with antibiotics (100 g/ml streptomycin and 100 IU/ml penicillin G), and the second passage of the cells were used for experimental determinations. COS-7 cells were maintained in Dulbecco's modified Eagle's medium without phenol red, supplemented with 10% dextran-coated charcoal-stripped fetal calf serum. The 5 ϫ 10 5 cells in 10-cm Petri dishes were transfected with a total of 15 g of plasmids using DOTAP (Roche Molecular Biochemicals). Cells were harvested 36 h after transfection and whole cell extracts were prepared using 1ϫ Passive lysis buffer (Promega, Madison, WI).
Luciferase Assays-Luciferase assay was performed using Dual Luciferase Reporter Assay System (Promega) as described (30). Briefly, 5 ϫ 10 5 cells were transfected with a total of 15 g of DNA. Two g of ERE-TK-LUC reporter plasmid was co-transfected with indicated amount of receptor expression vectors. All assays were performed in the presence of 2 g of pRL-TK, a Renilla luciferase reporter plasmid, as an internal control (Promega). The total amount of DNA and expression vectors for transfection was adjusted using pGEM3Zf (Promega) and pCXN2, respectively. After a 12-h incubation, the cells were washed with fresh medium and incubated for an additional 24 h in the absence or presence of 10 Ϫ7 M E 2 . Cell extracts were assayed for luciferase activity by Luminoskan (Labsystems Inc., Beverly, MA).
Western Blot Analysis-Twenty g of proteins were subjected to Western blot analysis using the 6F11 mouse monoclonal antibody against ER␣ (Novocastra Laboratories Ltd.) and rabbit polyclonal antibody against ER␤ as described (28,31), using the chemiluminescencebased ECL detection system (Amersham Pharmacia Biotech) according to the manufacturer's instruction.
Bone Size and Bone Mineral Measurements-The left femur from each rat was removed at autopsy and scanned using dual energy x-ray absorptiometry (QDR 1000/W, Aloka, Japan) equipped with Regional High Resolution Scan software, with antero-posterior application of the radiation beams to the specimens using two energy x-ray beams (27 KeV and 53 KeV) with 1-mm beam width. The scan speed was 2 mm/s. The femoral scan images were analyzed, and the value of the BMD of distal one-third region of the femur was determined.
Bone Histomorphometry-Histomorphometry was performed as described (26), using a semiautomatic image analyzing system linked to a light microscope (Cosmozone 1S, Nikon, Tokyo, Japan). In the tibial specimen, the metaphyseal cancellous bone area located within 4.0 mm distal to the growth plate-metaphyseal junction was measured. The parameters of trabecular bone volume (BV/TV, %) and trabecular bone surface (BS, mm) were measured. For indices of bone formation, doubleand single-labeled perimeters (dLS/BS, sLS/BS, %) and the interlabel distance (m) were measured on the trabecular perimeter. The mineralizing surface (MS/BS, %) was calculated as the sum of the doublelabeled surface and half of the single-labeled surface. The mineral apposition rate (MAR, m/day) was calculated as the mean distance between double labels divided by the interval labeling time and multiplied by /4, and the bone formation rate (BFR/BS, m 3 /m 2 /day) was

FIG. 1. Structure of the transgene and its expression pattern.
A, schematic structure of the transgene is presented. The transgene is under control of the chicken ␤-actin promoter with cytomegalovirus immediate-early (CMV-IE) enhancer, and includes the dominant negative ER␣ (rER␣-(1-535)) cDNA, and rabbit ␤-globin polyadenylation site. The box represent gene or cDNA fragments. B, Southern blot analysis of tail DNA from heterozygous offspring of the two founders that were successfully bred. The copy number was indicated for each line of transgenic rats. The EcoRI-NotI-digested 500-bp fragment of the rat ER␣ cDNA was used as a hybridization probe. The 2.9-kilobase pair signal observed in the wild-type and transgenic rats represent the endogenous ER␣. expressed per unit of bone surface. The parameter of trabecular osteoclast surface (OcS/BS, %) was measured.
Serum Parameters-Serum calcium, phosphorus levels were determined with an autoanalyzer (Hitachi 7170). Plasma osteocalcin levels were determined by a two-site sandwich enzyme immunoassay system using anti-rat osteocalcin. Serum E 2 levels were determined by an radioimmunoassay kit (Diagnostic Products Co.).

Construction of the Transgenic Rats-
We introduced the expression vector encoding rER␣-(1-535), the corresponding rat mutant of human ER␣-(1-530), into rats (Fig. 1A). Chicken ␤-actin promoter (32) was adopted for expressing rER␣-(1-535) in various kinds of tissues. Three independent founders were identified by Southern blot analysis. Two lines, designated line 3 and line 5, respectively, were able to transmit the transgenes to their offspring and were bred out into permanent lines. The F1 offspring obtained by the back-cross of the founder rats with normal rats were used for subsequent studies. The fertility and lactation in offspring of these founders were not markedly different from the wild-type rats. The copy numbers of the integrated transgenes, determined by comparison with the endogenous ER␣, were 10 and 8 in line 3 and line 5, respectively (Fig. 1B). RT-PCR analysis detected the rER␣-(1-535) mRNA in various tissues including bone of the F1 transgenic female rat (line 3), as well as the endogenous expression of ER␣ and ER␤ ( Fig. 2A). The similar expression pattern was detected in the F1 transgenic female rat of line 5 (data not shown). Especially in primary osteoblast-like cells of the transgenic embryos, we demonstrated the expression of the rER␣-(1-535) at the mRNA level by Northern blotting (Fig. 2B), as well as the protein level as a 59-kDa band by Western blotting (Fig. 2D). Endogenous ER␣ and ER␤ mRNA was demonstrated in primary osteoblast-like cells of both transgenic rats and their littermates by RT-PCR (Fig. 2C), whose protein could not detected by Western blotting (Fig. 2D).
In Vivo Dominant Negative Activity of rER␣-(1-535)-First, the inhibitory effect of rER␣-(1-535) was confirmed against wild-type ER␣ and ER␤ transactivation in COS-7 cells. In the presence of E 2 , equivalent co-expression ratio of the former leads to nearly the basal level of transactivation by ER␣ and ER␤, respectively (Fig. 3, A and B). Next, when the primary osteoblast-like cells isolated from the transgenic embryo calvariae and those of the littermate were transfected with ER␣ or ER␤ expression vector exogenously, an inhibitory effect of rER␣-(1-535) was also demonstrated (Fig. 3C). Endogenous E 2 -induced transactivation was shown as about 1.8-fold induction in the littermate-derived osteoblast-like cells, but it was lost in the transgenic osteoblast-like cells, indicating the dominant negative effect of transgene against ER␣ and ER␤ transactivation in vivo (Fig. 3D).
Radiographic and Histological Evaluation of Impaired Estrogen Sensitivity in Bone-We then examined the dominant negative effects of the transgene product against both ER␣ and ER␤ signaling pathways with respect to the bone mineral properties in vivo, one of the most important targets of estrogen. The distal femur BMD of the sexually mature 7-month-old female transgenic rats and their littermates were investigated by dual energy x-ray absorptiometry. As shown in Fig. 4A, the BMD of sham-operated rats exhibited no significant difference

FIG. 2. Expression of the rER␣-(1-535) transcript in various rat tissues.
A, RT-PCR analysis with poly(A) ϩ RNA from various rat tissues using specific primers for ER␣-(1-535), endogenous ER␣, and ER␤, respectively. Expression of ER␣-(1-535), ER␣, and ER␤ transcripts (455, 386, and 192 bp, respectively) was analyzed by RT-PCR with 0.1 g of poly(A) ϩ RNA as described. The RT-PCR products were analyzed by agarose gel electrophoresis. B, Northern blot analysis using primary osteoblast-like cells derived from wild-type and transgenic embryos. Total RNA was hybridized with 32 P-labeled, EcoRI-NotI-digested fragment of the rat ER␣ cDNA probe. Each lane contained 10 g of total RNA and g;yceraldehyde-3-phosphate dehydrogenase was used for the internal control. C, RT-PCR analysis from poly(A) ϩ RNA of rat primary osteoblast-like cells using specific primers for endogenous ER␣ and ER␤. Expression of ER␣ and ER␤ transcripts (386 and 192 bp, respectively) was analyzed by RT-PCR with 0.1 g of poly(A) ϩ RNA as described. The RT-PCR products were analyzed by agarose gel electrophoresis. D, determination of rER␣-(1-535) by immunoblot analysis of primary osteoblast-like cells. Twenty g of whole cell extracts were resolved by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose filter, and blots were probed with the ER␣-specific monoclonal antibody (6F11, 1:100) and ER␤-specific polyclonal antibody (1:500), respectively. among transgenic 2 lines and littermate rats. Ovariectomy caused the decrease of BMD in transgenic rats to the same extent as the littermates. The decrease of BMD of the control littermate rats due to ovariectomy was completely prevented in response to E 2 administration. Strikingly however, the BMD of the transgenic rats of both line 3 and line 5 remained decreased even after treatment with E 2 . When intact non-ovariectomized transgenic rats (n ϭ 5) and their littermates (n ϭ 5) were treated with E 2 , the distal femur BMD exhibited no significant difference (data not shown). The representative radiographic analyses of the femurs were compared in Fig. 4B, which supported the dual energy x-ray absorptiometry results that the transgenic femur failed to be prevented from the ovariectomyinduced bone loss in response to E 2 administration. Representative fluorescent micrographs of the calcein-labeled mineralization fronts (green) were also shown in Fig. 4C. Seven-monthold transgenic rats, compared with their littermates, showed the marked decrease in calcified cancellous bone of proximal tibia under the ovariectomizedϩE 2 conditions, whereas no apparent histological difference was observed when they were sham-operated or ovariectomized.
Histomorphometrical Evaluation of the Bone Phenotypes-To investigate the effects of the inhibition of both ER␣ and ER␤ signaling pathways in the bone in detail, the histomorphometrical analysis was performed by double labeling with calcein and tetracycline, markers of the amount of newly formed bone (33). When sham-operated, the cancellous bone volume (BV/TV) of the proximal tibia displayed no significant difference between transgenic rats and their littermates (Fig. 5A). Ovariectomy significantly reduced BV/TV of transgenic rats and their littermates (Fig. 5A). The BV/TV was maintained in control littermates by E 2 supplement; however, it remained decreased in both lines of the transgenic rats even if E 2 was administrated after ovariectomy (Fig. 5A). This finding is in accordance with the BMD changes of the distal femur. The cancellous bone formation rate (BFR/BS) as well as the osteoclast surface as percentages of total bone surface (OcS/BS) were increased after ovariectomy both in transgenic rats and the littermates, indicating a high bone turnover state (Fig. 5, B  and C). Interestingly, only transgenic rats in both lines sus- tained rather high values of BFR/BS and OcS/BS in response to E 2 administration (Fig. 5, B and C), keeping a high turnover state.
Increased Serum E 2 Concentration in Sham-operated Transgenic Rats-The average body and uterine weight revealed no significant difference in any treated groups between transgenic and littermate rats ( Table I). The serum osteocalcin values in the ovariectomized littermate rats were significantly higher than the values in the sham-operated group. However, the significant increase of the serum osteocalcin level did not decline when treated with E 2 in both transgenic lines. In contrast to the sham-operated littermates, the serum level of E 2 was significantly higher (about 1.4-fold) in sham-operated transgenic rats, suggesting impaired negative feedback due to E 2 insensitivity. Ovariectomy decreased and E 2 administration increased the serum E 2 level without any difference in all groups. The serum calcium and phosphorus levels were not significantly changed among any treated groups.

DISCUSSION
In this study, we have established the rats antagonizing both ER␣ and ER␤ pathways in vivo and demonstrated for the first time that the ovariectomy-induced bone loss was prevented by the ER-mediated signaling pathways.
Southern blot analysis of the DNA from each of the transgenic rat line indicated different sites of insertion of the transgene (data not shown). This allows us to conclude that the bone phenotypes observed were consequences of expression of transgene rather than insertional mutagenesis. The rER␣-(1-535) transcript was expressed in various kinds of tissues including bone. Assuming that the expression of the transgene remains rather constant throughout the target tissues, the relative ratio of transgene product to endogenous product should become greater in the tissues where the expression of endogenous ERs is lower. For example, in transgenic primary osteoblast-like cells, the rER␣-(1-535) protein detected by Western blotting is FIG. 4. Bone mineral property of transgenic rats. A, distal femur BMD (g/cm 2 ) measured in 7-month-old sham-operated (sham, bright shaded box), ovariectomized (OVX, dark shaded box), and ovariectomized and E 2 treatment (OVXϩE 2 , black box) rats. Values of measurements are means Ϯ standard deviations (n ϭ 5-6). Significance of difference between groups was evaluated with Student's t test; *, p Ͻ 0.05, significantly different from the sham-operated rats within the each strain; **, p Ͻ 0.05, significantly different from the ovariectomized rats within each strain. B, radiological analysis of the bones of the transgenic and the littermate rats. Representative x-ray of femurs of transgenic rats (line 3) and those of control littermates of 7-month-old are shown. C, histological evaluation of cancellous bone morphology in 7-month-old wild-type and transgenic rats. Longitudinal sections were made through the proximal tibia and fluorescent micrographs of the double-labeled mineralization fronts were shown. Only calcein label is visible (green), and the demeclocyclin label (yellow) is too faint to detect at this magnification. abundant enough to cause the complete inhibition of the E 2dependent transactivation by endogenous ERs. We could not detect the endogenous ER proteins in osteoblast-like cells using anti-ER␣ and anti-ER␤ antibodies by Western blotting (data not shown). Thus, it is suggested that the expression of the endogenous ER proteins confirmed by functional analysis here are under threshold levels of Western blot analysis. 1.4-fold higher serum E 2 level in transgenic rats, although it may not seem so remarkable compared with ER␣KO mice, is statistically significant and reflects the reduced estrogen responsiveness. In the case of human, only a man with a point mutation of ER␣ was reported (14) and his serum E 2 level (119 pg/ml) FIG. 5. Bone remodeling rate and trabecular density changes in the transgenic rats. Bars represent means Ϯ standard deviation. Significance of difference between groups was evaluated with Student's t test. For 7 months, n ϭ 5-6; *, p Ͻ 0.05, significantly different from the sham-operated rats within the each strain; **, p Ͻ 0.05, significantly different from the ovariectomized rats within each strain. A, quantitative representation of the trabecular bone density in the metaphyseal region of the tibia expressed as percentages of the total tissue area (BV%TV). B, measurement of bone-formation rate in the transgenic and the littermate rats. C, osteoclast surface as percentages of bone surface (OcS%BS) in the tibial metaphysis. was remained approximately twice that the normal male serum concentration (50 pg/ml). The absence of reproductive abnormalities including uterine weight in transgenic rat might be partly because the differences in the dose-response curves for ER agonism and antagonism by the dominant negative transgene (34). Further studies are required for clarifying the ERmediated transcriptional inhibition in other cells and tissues.
The bone mineral measurements clearly revealed that the ovariectomy-induced bone loss was prevented through ER-mediated signaling pathway, and the data presented here are consistent with the hypothesis that ER plays a major role in the bone physiology (16). Histomorphometrical analysis supports the mechanism that the prevention from the high bone turnover due to ovariectomy was mediated by the ER cascades. The serum osteocalcin levels of the transgenic rats remained increased after ovariectomy followed by E 2 administration, reflecting the high turnover state where both bone formation and its resorption are accelerated (25,35,36). The reasons why the difference of BMD was not observed between sham-operated transgenic rats and their littermates still remain to be resolved. Taken together with that the bone density in the femurs is not significantly decreased in the ER␣KO (16,37) as well as ER␤KO (38,39) female mice, the compensation of BMD might be achieved by ER-independent mechanisms such as the bioactivity of estrogen-regulated cytokines or by a novel estrogenresponsive pathway distinct from ER␣ and ER␤. This possibility is also supported from our results; the intact transgenic rats and littermates exhibited no difference in the distal femur BMD in the presence of pharmacological (100 nM) E 2 concentration. Alternatively, it may be possible that the dominant negative activity of rER␣-(1-535) is not enough to inhibit ERmediated pathways completely in this condition.
Several studies have focused on the possible involvement of bone resorbing cytokines such as interleukin-1 (IL-1) (40), interleukin-6 (IL-6) (41), and tumor necrosis factor (42) in the stimulation of bone resorption due to estrogen deficiency (43)(44)(45). Estrogen is supposed to suppress the production of these cytokines by osteoblasts and bone marrow stromal cells (44,45). Although substantial lines of evidence regarding the protective effects against the bone loss caused by ovariectomy have been accumulated from the several transgenic animal models such as IL-6 (33), IL-1 receptor type I (IL-1R1) (46) knockout mice, and soluble tumor necrosis factor-␣ receptor type I transgenic mice (35), specific contribution of these cytokines to the bone loss due to estrogen deficiency still remains controversial. Moreover, whether ERs are directly responsible for the regulation of these estrogen-regulated cytokines is completely unknown.
Our results demonstrated that the inhibition of ER␣ and ER␤ still induced bone loss after ovariectomy to the same degree as that of littermates. If we assume that the dominant negative activity is sufficient to inhibit ER pathways as shown in the case of E 2 treatment after ovariectomy, the regulation of the bone-resorbing cytokines by estrogen may be involved in ovariectomy-induced bone loss rather than the classical ER/ ERE-dependent regulation. Alternatively, ovariectomy may cause some changes other than the decline of E 2 concentration in vivo, influencing the crucial effects in the bone metabolism. These possibilities are also supported by the results of the ER␣KO (37,47) and ER␤KO (39) mice showing the bone loss due to ovariectomy.
In the present study, we have created a transgenic rat line with impaired estrogen sensitivity by expressing a dominant negative mutant against both ER␣ and ER␤ in bone and other tissues. Taking advantage of the analytical usefulness of the rat as species, we thus opened a way to examine the physio-logical roles of estrogen through ER-dependent and ER-independent signaling pathways. Moreover, this study provides a mechanism for diseases caused by the dominant negative phenotypes of the mutations of nuclear receptors including ER. In practice, some patients with thyroid hormone resistance are caused by the heterozygous mutation of thyroid hormone receptor to its dominant negative mutant (48,49). The forthcoming analyses of the bone phenotypes of the ER␣/␤ doubleknockout mice and the comparison with our animal model described here will give us further insights into the physiological and pathophysiological actions of estrogen in the bone metabolism.