Osteoporosis International

, Volume 19, Issue 6, pp 793–800

Comparative effects of 17β-estradiol, raloxifene and genistein on bone 3D microarchitecture and volumetric bone mineral density in the ovariectomized mice


  • A. Cano
    • Department of Pediatrics, Obstetrics and GynecologyUniversity of Valencia
  • S. Dapía
    • Trabeculae, Empresa de Base Tecnológica S.L.Parque Tecnológico de Galicia
  • I. Noguera
    • Research Unit, Faculty of MedicineUniversity of Valencia
  • B. Pineda
    • Research FoundationHospital Clínico Universitario
  • C. Hermenegildo
    • Research FoundationHospital Clínico Universitario
    • Department of PhysiologyUniversity of Valencia
  • R. del Val
    • Research FoundationHospital Clínico Universitario
  • J. R. Caeiro
    • Trauma and Orthopedic Surgery ServiceComplexo Hospitalario Universitario
    • Research FoundationHospital Clínico Universitario
    • Department of GeneticsUniversity of Valencia
Original Article

DOI: 10.1007/s00198-007-0498-6

Cite this article as:
Cano, A., Dapía, S., Noguera, I. et al. Osteoporos Int (2008) 19: 793. doi:10.1007/s00198-007-0498-6



This study assessed the effect of estradiol, raloxifene and genistein on the preservation of bone 3D-microarchitecture and volumetric bone mineral density (vBMD) in the ovariectomized mouse model. Our results indicated that raloxifene was more effective in preserving bone ovariectomized-induced changes, the advantage being concentrated in both bone microarchitecture and vBMD.


This study assessed the effect of different estrogen receptor (ER) agonists on the preservation of bone 3D-microarchitecture and volumetric bone mineral density (vBMD) in the ovariectomized (OVX) mouse model.


Twelve-week-old female C57BL/6 mice were randomly assigned to one of five groups: (1) SHAM-operated + vehicle; (2) OVX + vehicle; (3) OVX + 17β-estradiol (5 μg/kg); (4) OVX + raloxifene (1 mg/kg); (5) OVX + genistein (25 mg/kg), during 4-weeks. Bone microarchitecture and trabecular, cortical and total vBMD of distal femur were imaged by ex vivo microcomputed tomography (micro-CT).


Ovariectomy produced a global deterioration involving both trabecular and cortical 3D-microarchitecture and vBMD. Raloxifene maintained both microarchitecture and vBMD, whereas estradiol prevented deterioration of some microstructural parameters, such as trabecular thickness (Tb.Th), trabecular bone pattern factor (Tb.Pf), and cortical periosteal perimeter (Ct.Pe.Pm), but did not completely block the loss in vBMD. Mice treated with genistein exhibited the less favourable profile in both vBMD and microstructural parameters preserving only cross-sectional bone area (B.Ar) and Ct.Pe.Pm in cortical bone.


Our data indicate that, at the selected doses, raloxifene was more effective in preserving bone OVX-induced changes than either estradiol or genistein, the advantage being concentrated in both bone microarchitecture and vBMD.


GenisteinOsteoporosisRaloxifeneThree-dimensional microarchitecture17β-estradiol


Osteoporosis is a disease characterized by the progressive loss of bone mass, deterioration of cortical and trabecular architecture, and a decrease in mineralization of bone tissue. These changes, among others, may compromise bone strength leading to an increased risk of fracture. The hypogonadic state associated with menopause is the main determinant of postmenopausal osteoporosis, the most frequent form of the disease [1, 2]. Indeed, estrogen deficiency increases bone remodeling within a context where bone resorption prevails over bone formation [3].

Hormone therapy (HT) prevents bone loss and reduces fracture risk [2, 3]. However, recent clinical studies [4, 5] have shown an unfavourable balance for HT when administered to non-symptomatic women. Advances in the knowledge on the molecular details of estrogen receptor (ER) action have made clear that there is a wide array of molecules exhibiting a whole range of agonistic/antagonistic profile on the ER [6]. The recent denomination of selective estrogen receptor modulators (SERMs) reflects this concept. Moreover, an additional source of ER agonists is represented by phytoestrogens, a generic denomination including plant-derived compounds with the ability to bind ER as well.

Raloxifene is a SERM approved for prevention and treatment of osteoporosis in postmenopausal women. Interestingly, together with this agonistic effect on bone, raloxifene displays antagonistic effects on the breast [7, 8]. Genistein (GEN), one main component of the isoflavone family of phytoestrogens, has been shown to have a preferential affinity for the β isoform of the ER [9, 10]. There are data suggesting a beneficial effect of GEN on the cardiovascular tree, on the control of menopausal symptoms, as well as lowering the RANKL/OPG ratio [11, 12].

The aim of the present work was to compare the effects of three different agonists of the ER, estradiol, raloxifene and genistein, on two parameters crucial for bone quality, 3D microarchitecture and volumetric bone mineral density (vBMD) in an animal model of accelerated bone loss by using ovariectomized young female mice. Loss of trabecular connectivity resulting in a shift from plate-like to rod-like trabeculae is a well-established characteristic of osteoporosis [13]. Given that the nature of trabecular architecture is anisotropic, the study used microcomputed tomography (micro-CT), a technique capable of analyzing 3D images. Ovariectomized mice were used as a model to perform the experiments.

Materials and methods

Mice and treatments

Twelve-week-old female C57BL/6 mice (Charles River Laboratories, Barcelona, Spain) were housed in an environmentally controlled laboratory upon arrival and acclimatized for 4 days. The animals were either dorsal ovariectomized or falsely operated (SHAM) under general anesthesia by using 150 mg/kg of ketamine (Merial, Lyon, France) and 5 mg/kg acepromazine (Calmo Neosan, Pfizer, NY). Animals were kept at 21°C with 12:12-h light-dark cycle and were allowed free access to a pelleted standard mouse laboratory diet containing 0.88% calcium, 0.59% phosphorus and 900 IU/kg vitamin D3 (Panlab SL, Barcelona, Spain) and tap water. All procedures for consideration of animal welfare were reviewed and approved by the ethical committee of our institution.

The mice were divided into five groups: SHAM-operated (SHAM, n = 6), ovariectomized without treatment (OVX, n = 6), and ovariectomized mice treated with 17β-estradiol (E2, n = 8), raloxifene (RAL, n = 8) or genistein (GEN, n = 8), all compounds were supplied by Sigma-Aldrich Chemical (St. Louis, MO). Mice were subcutaneously injected 5-times per week with 100 μl of vehicle (10% DMSO, 90% sesame oil; Sigma) in the SHAM and OVX groups, and with 17β-estradiol (0.1 μg/mouse/day = 5 μg/kg/day), raloxifene (20 μg/mouse/day = 1 mg/kg/day) or genistein (0.5 mg/mouse/day = 25 mg/kg/day), dissolved in vehicle for the remaining groups from the third day after surgery. The administered doses suppose equipotent dose according to the estrogenic capacity of each agonist [14, 15] and were chosen based on doses frequently used in the literature [1620].

Mice were sacrificed 4 weeks after beginning the treatment by cardiac puncture exsanguination under halothane anesthesia (Fluothane; Zeneca, Macclesfield, UK), and the uterus and right femur were aseptically removed. The uterus was weighed to confirm the success of ovariectomy and the femur was cleaned of adherent soft tissues and deposited in a tube with 10% formalin.

Three-dimensional trabecular and cortical microarchitecture analysis by micro-CT

The distal region of femora was analyzed without further sample preparation by micro-CT, (SkyScan 1172), imaged with an X-ray tube voltage of 50 kV and current of 200 μA, with a 0.5-mm aluminium filter. The scanning angular rotation was 185°, and the angular increment was 0.45°. The voxel size was 6.5 μm isotropically. Data sets were reconstructed using a modified Feldkamp algorithm [21] and segmented into binary images (8-bit BMP images) using adaptive local thresholding. For analysis of the microarchitectural properties of trabecular and cortical bone regions, femora specimens were evaluated within a conforming volume of interest (VOI). In the case of the trabecular bone region, a VOI was selected starting at a distance of 1.40 mm from the growth plate (GP), extending a further longitudinal distance of 2.70 mm in the proximal direction (202 image slices analyzed, cortical bone excluded). For the cortical bone region, a VOI included in a proximal distance of 2.75 to 3.40 mm from the GP (100 images analyzed) was selected.

Both trabecular and cortical bone regions were obtained by free drawing regions of interest, and analyzed using the commercial software provided with the equipment (SkyScan™ CT-analyzer software, version 1.6.0). Morphometric indices of trabecular bone region were determined from the microtomographic data sets (integrated over a VOI) using direct 3D morphometry. Total volume of VOI (tissue volume TV, mm3) and trabecular bone volume (BV; mm3) were calculated based on the hexahedral marching cubes volume model of the VOI. Trabecular bone volume (BV/TV; %) was directly calculated. Trabecular thickness (Tb.Th; mm), trabecular separation (Tb.Sp; mm), and trabecular number (Tb.N; 1/mm) were measured directly on 3D images using methods previously described [22, 23]. Measurements of trabecular thickness were calibrated by scanning and analyzing three aluminium foils with thicknesses of 50, 125 and 250 μm. The non-metric indices, structure model index (SMI), and trabecular bone pattern factor (Tb.Pf; 1/mm) were also calculated using the direct 3D model. The SMI parameter indicates the relative prevalence of rods and plates in a 3D structure [24]. DA (degree of anisotropy) represents trabecular anisotropy, defined as the ratio between the maximal and minimal radius of the mean intercept length (MIL) [25]. The trabecular bone pattern factor (Tb.Pf; 1/mm) measures the relative convexity or concavity of the total bone surface [26].

Morphometric variables of cortical bone region were measured in 2D from individual 2D cross-sectional images. We assessed the cortical periosteal perimeter (Ct.Pe.Pm; mm), cortical endosteal perimeter (Ct.En.Pm; mm), mean polar moment of inertia (MMI; mm4), which indicates the resistance to rotation of a cross-section about a chosen axis being the rotational analog of mass for linear motion, cross-sectional thickness (Cs.Th; mm), and mean of total cross-sectional bone area (B.Ar; mm2). The coefficient variation (%CV) values for all these measurements were <5%.

Volumetric bone mineral density (vBMD) measurements using micro-CT imaging

Volumetric trabecular, cortical and total bone mineral density (vBMDtrab, vBMDc and vBMDt, respectively) were determined by micro-CT scanning. vBMDc and vBMDtrab were calculated in the conforming VOI described above for cortical and trabecular region, respectively. vBMDt was calculated in the trabecular region described above (cortical bone included). Tomographic scans were performed ex vivo on excised femur as described above. Every pixel of the reconstructed 8-bit BMP images has a colour or grey value between 0 and 255. Grey value 255 was assumed to be white (void space), whereas 0 is black or the densest part of the image. To express grey values as mineral content, hydroxyapatite phantom rods of 2 mm of diameter immersed in pure water with known vBMD (0.25 g/cm3 and 0.75 g/cm3) were employed for calibration.

Statistical analysis

Untransformed data were analyzed for normality by the Kolmogorov–Smirnov test. If the data were normally distributed, a two-tailed t test was performed to compare the SHAM group with the OVX/vehicle group. A one-way (treatment) ANOVA was performed to ascertain the treatment effect difference between OVX/vehicle or SHAM and the three different treatments (E2, RAL and GEN) groups at an α = 0.05 level. If the ANOVA test was significant, Dunnett’s test (when the variances were assumed to be equal) or Dunnett’s T3 test (when the variances were assumed to be unequal) was applied to perform post hoc pairwise comparisons at α = 0.05 level. Levene test was used to test the homogeneity of variance for each dependent variable across all level combinations of the between-subject factors. If the data were not normally distributed, a nonparametric Kruskal–Wallis was performed to determine intergroup differences, if any. Results are presented as means ± SEM. A P value <0.05 of was considered statistically significant. The statistical analysis was carried out using the Statistical Package for Social Sciences (SPSS Inc., Chicago, IL), v. 14.0 for Windows.


Effect of ovariectomy and ER agonists administration on trabecular bone microarquitecture

Figure 1 shows representative micro-CT images of the analyzed region of femora. Differences in trabecular architecture between the groups of mice are rather apparent. Analysis of data indicates that, as compared with the SHAM group, ovariectomy decreased Tb.N and Tb.Th by nearly 23% (P < 0.01) and 11% (P < 0.001), respectively, and increased Tb.Sp by 9% (P < 0.05, Fig. 2a,c,e). Additional effects of ovariectomy were a 25% decrease in trabecular BV/TV (P < 0.01, Fig. 2b), and 11% increase in Tb.Pf (P < 0.01), suggesting a more disconnected trabecular structure (Fig. 2d). There were no significant differences with respect to SMI (Fig. 2f), a variable that quantifies the relative prevalence of rods or plates in a 3D structure such as trabecular bone.
Fig. 1

3D trabecular microarchitectural images of distal mouse femora by means of micro-CT. a) SHAM group. b) OVX group at 4 weeks postsurgery. c) E2 group after 4 weeks of estradiol treatment d) RAL group after 4 weeks of raloxifene treatment. e) GEN group after 4 weeks of genistein treatment

Fig. 2

Effects of ovariectomy and treatment on trabecular microarchitectural parameters. Values are expressed as the mean ± SEM. a: P < 0.05 vs. SHAM; b: P < 0.01 vs. SHAM; c: P < 0.001 vs. SHAM; x: P < 0.05 vs. OVX; y: P < 0.01 vs. OVX; z: P < 0.0001 vs. OVX

Also in Fig. 2, data were similar for the RAL and the SHAM groups except for SMI, which was slightly lower for RAL (P < 0.05). However, RAL-treated mice behaved differently from OVX animals in all parameters, clearly demonstrating that this compound inhibited every change induced by ovariectomy in trabecular bone. The effect was more mixed for E2 and GEN (Fig. 2). Estradiol prevented the changes induced by ovariectomy in both Tb.Th (P < 0.0001) and Tb.Pf (P < 0.05), but not in Tb.N, Tb.Sp, or BV/TV. GEN, in turn, was unable to correct any change induced by ovariectomy.

Effect of ovariectomy and ER agonist administration on cortical bone microarchitecture

As compared with the SHAM group (Table 1), ovariectomy induced changes in Ct.Pe.Pm (13.1% increase, P < 0.01), MMI (13.4% decrease, P < 0.01) and B.Ar (8.2% decrease, P < 0.05). RAL significantly normalized all the three parameters, while E2 only normalized the Ct.Pe.Pm (P < 0.01), although a tendency to prevent the B.Ar deterioration (P = 0.06) was observed. GEN neutralized the changes in Ct.Pe.Pm (P < 0.05) and B.Ar (P < 0.0001) (Table 1).
Table 1

Effect of ovariectomy and ER agonist administration on cortical bone microstructure of different groups of mice







Ct.En.Pm (mm)

4.95 ± 0.27

4.93 ± 0.17

5.04 ± 0.16

5.02 ± 0.18

5.30 ± 0.09

Ct.Pe.Pm (mm)

6.05 ± 0.21

6.84 ± 0.36 b

6.06 ± 0.07 y

6.06 ± 0.08 y

6.28 ± 0.08 x

MMI (mm4)

0.57 ± 0.02

0.50 ± 0.01 b

0.49 ± 0.02 b

0.55 ± 0.01 x

0.53 ± 0.02

Cs.Th (mm)

0.13 ± 0.01

0.12 ± 0.01

0.13 ± 0.01

0.14 ± 0.01

0.12 ± 0.01

B.Ar (mm2)

0.76 ± 0.02

0.70 ± 0.01 a

0.73 ± 0.01

0.80 ± 0.01 z

0.78 ± 0.01 z

Ct.En.Pm: cortical endosteal perimeter; Ct.Pe.Pm: cortical periosteal perimeter; MMI: mean polar moment of inertia; Cs.Th: cross-sectional thickness; B.Ar: total cross-sectional bone area.

a: P < 0.05 vs. SHAM, b: P < 0.01 vs. SHAM

x: P < 0.05 vs. OVX, y: P < 0.01 vs. OVX, z: P < 0.0001 vs. OVX

Effect of ovariectomy and ER agonist administration on vBMD

Ovariectomy resulted in a reduction of 34% (P < 0.01) and 4% (P < 0.01) in vBMDtrab and vBMDc, respectively, an effect that represents a vBMDt reduction of approximately 32% (P < 0.001) (Fig. 3). The different ER agonists prevented, with differing effectiveness, trabecular bone mass loss. Thus, RAL entirely blocked trabecular bone loss after 4 weeks of treatment (P < 0.0001). Mice treated with E2 and GEN showed vBMDtrab that was intermediate between SHAM and OVX animals, an effect implying that they inhibited trabecular bone mass loss only partially (Fig. 3). No treatment inhibited bone loss at the cortical level, although the E2 group nearly reached a favorable difference in relation to the OVX group (P = 0.09).
Fig. 3

Volumetric trabecular (vBMDtrab; a), cortical (vBMDc; b) and total (vBMDt; c) bone mineral density in SHAM, OVX, and OVX treated groups. vBMD was determined by micro-CT scanning as described in Materials and Methods section. Values are expressed as the mean ± SEM. a: P < 0.05 vs. SHAM; b: P < 0.01 vs. SHAM; c: P < 0.001 vs. SHAM; d: P < 0.0001 vs. SHAM; x: P < 0.01 vs. OVX; and y: P < 0.0001 vs. OVX

In terms of vBMDt values, RAL, and not GEN, effectively neutralized the loss of bone mass associated with OVX. E2 exhibited an intermediate effect, since vBMDt was 21.4% higher than in the OVX group, but 14.6% lower than in the SHAM operated animals.

As expected, strong correlations between vBMDt values and almost all trabecular bone microarchitectural parameters were found. Thus, vBMDt correlated positively with BV/TV (r = 0.62, P < 0.001), Tb.Th (r = 0.60, P < 0.001), and Tb.N (r = 0.59, P < 0.001) and negatively with Tb.Pf (r = 0.59, P < 0.001) and SMI (r = 0.46, P < 0.01).


In the present study it was observed that estrogen deficiency, as induced by ovariectomy, deteriorated both bone mineral density (mainly vBMDtrab) and bone quality, as suggested by the important changes in almost all microstructural parameters assayed. Regarding trabecular bone, ovariectomy induced a significant reduction in trabecular thickness (Tb.Th), trabecular number (Tb.N), and trabecular bone volume (BV/TV), as well as an increase in trabecular separation (Tb.Sp), a series of changes that result in a decreased connectivity among trabeculae (increase in Tb.Pf value). At the cortical level, ovariectomy induced an increment of the periosteal perimeter (Ct.Pe.Pm) and a diminution in the polar moment of inertia (MMI) and cortical surface (B.Ar). These results confirmed that 4-week-old ovariectomized mice presented both a very deteriorated microarchitecture as well as a large reduction in trabecular mass (35%). Therefore, this mouse model of adult young mice seems suitable for investigating phenomena underlying accelerated bone loss and osteoporosis.

To our best knowledge, this is the first report on microstructural comparative effects on bone, as assessed by micro-CT, of estradiol, raloxifene and genistein, administered at equipotent dose to ovariectomized mice. The data confirmed that raloxifene was the most effective treatment, preventing the deterioration of both vBMD as well as trabecular and cortical microarchitectural parameters. Moreover, this treatment not only neutralized the loss in vBMDtrab, but, in agreement with published data in rat and in women [7, 16], it further increased this parameter above the levels observed for the control group.

The data add substantial information to the sparse knowledge about the effect of raloxifene on microstructural properties of bone. It has been proposed that raloxifene improves the mechanical properties of bone in ways that do not involve increases in BMD or BV/TV [27]. The present results confirmed that raloxifene improved the deterioration induced by ovariectomy in BV/TV, Tb.N, Tb.Sp, and Tb.Th in trabecular bone. Furthermore, raloxifene increased the connectivity between trabeculae and diminished the SMI. This last effect is particularly important because SMI, which is characterized by a transition from plate-like to rod-like architecture, shows particular importance in trabecular bone affected by osteoporosis [24]. Lower SMI values in the RAL group suggest a more plate-like structure, which is in accordance with improved bone strength. This is the first time that the effect of raloxifene on SMI has been reported. The effects of ovariectomy were less dramatic in cortical bone microstructure but, again, raloxifene neutralized the structural deterioration induced by estrogen deficiency.

Estradiol exhibited a more mixed effect. It only partially avoided vBMD loss, but was the treatment that performed best on vBMDc. This profile is in agreement with recently published data [28, 29]. Regarding microarchitectural parameters, estradiol inhibited loss in trabecular thickness (Tb.Th) and in connectivity (Tb.Pf). We do not have yet explanation for this behaviour, since a significantly improvement in number and in trabecular separation would be expected before of a inhibition of connectivity loss. Additional studies will be needed to clarify this behaviour.

Given the differences observed for raloxifene and estradiol, it might be argued that the dose of estradiol was not equipotent to that of raloxifene. However, the raloxifene and estradiol dosages used in this work would correspond to 60 mg and 0.3 mg, respectively, for a woman of 60 kg of weight, levels that have been demonstrated to have significant bone protection effects in humans [30, 31]. Additionally, it has been described that similar doses of estradiol to those used in the current work inhibit BMD loss induced by ovariectomy, although in those studies, bone mineral content was calculated per area unit (measured by dual energy X-ray absorptiometry) instead of per volume unit [1720]. To further support the significance of the present data, it has been reported that micro-CT can detect differences in mineral content that are comparable or superior to the gold standard defined by ash content values [32]. It seems conclusive, therefore, that raloxifene was more effective than E2 in this ovariectomized mice model. Indeed, it has been described that estradiol, administered to ovariectomized rats at similar doses to those of this study, was unable to improve some proximal tibia microstructural parameters, such as BV/TV, connectivity density, SMI, and Tb.Sp [29].

A previous study demonstrated that identical doses of estradiol as those used in the present one prevented both uterine atrophy and thymic hypertrophy, as well as the increase in B-cells in BM of ovariectomized mice [33]. Moreover, in another study, the same doses reestablished the changes induced by ovariectomy in uterus, thymus, and cortical BMD levels, but were are not able to exceed 20–30% of the response achieved with the maximum dose of estradiol in trabecular BMD in eight-week-old mice [28]. Interestingly, it has been described that mouse age is determinant in this type of study [20]. Thus, administration of 5 μg/kg/day estradiol to 6-month-old mice completely prevented the loss of trabecular bone without stimulative effects on the uterus, while the same dose injected to 3-month-old mice (identical age to that of our mice) did not avoid the bone loss, and resulted in uterine hypertrophy [20], which agrees with our published data [33].

Genistein was the treatment with the lowest antiresorptive power in our study. It was unable to limit vBMD loss or the microstructural deterioration induced by ovariectomy in trabecular bone, although it inhibited the changes in periosteal perimeter (Ct.Pe.Pm) and cortical surface (B.Ar) in cortical bone. It is unlikely that the poor effect of genistein in our study was due to insufficient dosage. That used in our work is normal for this type of study [19, 34, 35], and it would be equivalent to 1500 mg for a woman of 60 kg of weight, 20–25 times superior to the dose generally prescribed. It has been described that genistein dosages around this range are toxic for reproductive tissue [3638]. Our results agree with some previously published studies of the minimal effect of genistein or other isoflavones on bone microarchitectural parameters [29, 39], albeit some evidence in favor of protection against ovariectomy-induced bone loss has been detected [34, 35, 40].

To give an answer to these questions we used C57BL/6 mice as a model, a mouse strain largely used for investigating bone effects of ovariectomy. Moreover, we have used twelve-week-old C57BL/6 mice that, with regard to skeleton, are virtually mature [41]. A certain degree of modeling and remodeling activities, however, can still subsist at this age in the skeleton. This specific metabolic status might have influenced the pattern of response observed for each ER agonist tested in our study. In fact, the interaction of ER agonists with either the own receptor and transcription factors generates a wide array of responses from pure agonism to complete antagonism [42]. We cannot discard variable contributions of each compound through pathways affected by the age of the animal or even of the phase of skeleton growth.

In conclusion, at doses that were chosen to achieve equivalent agonistic power for each of the three compounds, raloxifene has been shown to be a more effective treatment than either estradiol or genistein in preserving bone vBMD and microstructure in the adult young ovariectomized mice. Since the present work was a comparative study, it limits the interference of confounding variables. One cannot discard, however, that the use of alternative routes for administrating the treatment, mouse strain, or age of the mice, may have influenced the results.


The authors are indebted to Mrs. Rosa Aliaga and Mrs. Elvira Calap for their excellent technical assistance. This work was supported by the Generalitat Valenciana (Grant GV05/141) and by a research contract (to M.A.G.-P.) from Consellería de Sanitat de Valencia and Fondo de Investigación Sanitaria (FIS).

Copyright information

© International Osteoporosis Foundation and National Osteoporosis Foundation 2007