Tibolone (Org OD14) is a tissue-specific steroid with estrogenic effects on the bone and vagina but not endometrium or breast and has been shown to prevent ovariectomy-induced bone loss in young and old rats. We evaluated the effect of long-term tibolone treatment on bone parameters in mature ovariectomized (OVX) rats. Six-month-old rats were allotted to one of six groups (n = 8). Sham-operated and control OVX groups received vehicle, whereas other groups (all OVX) received tibolone (125, 250, or 500 μg/day orally) or 17α-ethinylestradiol (EE; 24 μg/day orally) for 16 months. Treatment with tibolone prevented ovariectomy-induced bone loss in peripheral (femur and tibia) and axial (L1-L2 and L4) skeleton. In peripheral skeleton, tibolone and EE prevented loss of bone mass and quality to a similar extent. Tibolone dose-dependently inhibited trabecular bone volume loss in L1-L2 and tibia, and at 500 μg/day it inhibited 88% of L1-L2 and 55% of tibial volume loss (p ≤ 0.05 in each case). Tibolone, 500 μg, resulted in 10% greater cortical strength of femur (p ≤ 0.05) and 60% greater compressive strength of L4 (p ≤ 0.05) compared with vehicle-treated OVX rats. Tibolone and EE inhibited bone resorption and turnover, assessed by urinary deoxypyridinoline/creatinine and plasma osteocalcin, respectively. We conclude that 16 months of tibolone treatment prevents ovariectomy-induced deterioration of axial and peripheral skeleton and preserves cortical and trabecular bone strength by reducing bone resorption.
OVARIAN HORMONE deficiency in rats accelerates bone remodeling as it does in humans. This is evidenced by the increased number of trabecular bone osteoclasts and osteoblasts, the increase in the plasma bone formation markers osteocalcin and alkaline phosphatase, and by the urinary bone resorption markers hydroxyproline and deoxypyridinoline.(1) In ovariectomized (OVX) rats, trabecular bone loss from the peripheral skeleton (metaphyseal distal femur and proximal tibia) and axial skeleton (lumbar vertebrae) has been shown using histological as well as noninvasive densitometrical techniques.2-7) In the rat, ovariectomy-induced bone loss can be prevented by treatment with the female sex hormone 17β-estradiol.(2, 3) Because of the similarities between rats and humans in their skeletal responses to estrogen deficiency, the mature OVX rat is considered a good animal model for studying early postmenopausal bone loss.(8, 9)
Tibolone [Org OD14; (7α,17α)-17-hydroxy-7-methyl-19-norpregn-5(10)-en-20-yn-3-one] is structurally related to the 19-nortestosterone-derived progestogens norethisterone and norethynodrel. It exhibits combined estrogenic, androgenic, and progestogenic activity in various in vivo bioassays.(10, 11) Tibolone behaves as a tissue-specific steroid in postmenopausal women, having estrogenic effects on bone and vagina without stimulating the endometrium or breast. Histomorphometrical and densitometrical techniques have shown that tibolone is effective in preventing ovariectomy-induced trabecular bone loss in the axial and peripheral skeleton of young mature and senescent rats.12-14) Tibolone was shown to preserve bone mass like estrogens through an inhibition of ovariectomy-increased bone resorption. However, if the intensity of bone remodeling is reduced too much, the biomechanical competence of the bone material can be affected on long-term treatment.
The objective of this study was to evaluate the effects of a long-term (16-month) treatment with tibolone on biomechanical properties of cortical and trabecular bone in OVX rats. Additionally, the effects on bone mass in the axial and peripheral skeleton and on bone turnover were examined. Orally dosed 17α-ethinylestradiol (EE) was used as a reference compound for estrogenic activity on bone. Because of species differences in the bone-remodeling cycles, a 16-month treatment in the rat can be considered equivalent to a 4-year treatment period in humans.(15, 16)
MATERIALS AND METHODS
Tibolone or EE were suspended in 0.5% wt/vol gelatin and 5% wt/vol mannitol for oral administration. Drug suspensions were prepared by the Department of Pharmaceutics, N.V. Organon, Oss, The Netherlands. All chemicals were of analytical grade.
Mature 3-month-old female Wistar rats (strain Hsd/Cpd: Wu) were supplied by Harlan, the Central Institute for Breeding of Laboratory Animals (Zeist, The Netherlands). The animals were housed under standard conditions (14-h light/10-h near dark in an air-conditioned room maintained at 21 ± 2°C) and had free access to food and water during the preexperimental period. The experiment started when the animals were 6 months old. At this time, body weight was 268 ± 18 g (mean ± SD). During the experiment the animals were housed individually, had free access to tap water, and were pair-fed with a maximum of 16 g pelleted food per day. The diet contained 1.06% calcium, 0.8% phosphorous, 2000 IU/kg vitamin D3, 4.8% fat, and 22% protein (Hope Farms, Linschoten, The Netherlands). All animal procedures were approved by the Animal Ethics Committee.
Forty-eight animals were allocated into one of six groups of eight, according to a randomized block design using body weight as selection parameter. At the start of the experiment, five groups were OVX bilaterally under diethyl ether anesthesia (OVX groups) and the remaining group (intact) was sham-operated. From the day of surgery onward, the animals of the intact group and of the OVX control group were treated with vehicle (0.5% wt/vol gelatin/5% wt/vol mannitol, 0.25 ml twice daily orally). The other four OVX groups received tibolone (2 × 62.5 μg, 2 × 125 μg and 2 × 250 μg/day, orally) or EE (2 μg/day × 12 μg/day orally) for 16 months. Drug/vehicle administration was by gavage (0.25 ml bolus) directly into the stomach.
After 16 months of treatment, the animals were starved for 16 h (but with access to demineralized water) while housed in metabolic cages for urine collection. The animals were then anesthetized with diethyl ether and an autopsy was performed. Blood was collected from the abdominal aorta and heparinized plasma was prepared by centrifugation (3000g for 10 minutes at 4°C) and stored in aliquots at −20°C.
Biochemical bone turnover and liver parameters
Urine was analyzed for deoxypyridinoline (ELISA Pyrilinks-D; Metra Biosystems, Inc., Palo Alto, CA, USA), calcium, phosphate, and creatinine. These concentrations were used to determine the molar ratios of deoxypyridinoline, calcium, and phosphate to creatinine. Plasma alkaline phosphatase activity, calcium, phosphate, and glutamate-pyruvate transaminase (GPT) were determined by standard methods using an automated analyzer. Additionally, plasma osteocalcin was determined by radioimmunoassay.(17)
At autopsy, the right femur was dissected out and stored in saline before bone strength measurement in a three-point bending test on the same day. The left femur was also dissected out and bone mineral density (BMD) was measured as described in the following section. For histomorphometric measurements, the right tibia and the lumbar vertebrae L1 and L2 were dissected out and placed immediately in Burckhardt's fixative for 18-24 h before being transferred to 100% ethanol for ≥48 h. The vertebral body of lumbar L4 was adapted for the compression test as described in the following section.
Bone mineral densitometry
Trabecular BMD of the femur was measured with a peripheral quantitative computed tomography (pQCT) machine (XCT 960A; Stratec, Birkenfeld, Germany). The machine was adapted for measurement in small animals and was calibrated with a standard of hydroxyapatite embedded in acrylic plastic. Two 360° scans were taken, with a standard thickness of 1 mm and a resolution of 0.148 mm × 0.148 mm. Trabecular BMD in the metaphysis was measured from a scan taken 5 mm from the distal end of the femur. Cortical BMD was determined from the second scan, which was taken in the diaphysis, 14 mm from the distal end of the bone. Intra-assay variations were 2.2% and 0.5% for trabecular and cortical measurements, respectively; the interassay variations were 2.6% and 0.6%, respectively. This scan also was used to determine geometrical parameters such as cortical thickness, total bone area, and outer and inner diameter. Similarly, BMD and total bone area of the vertebral body of lumbar L4 were measured at the center of the planoparallel section.
The right tibia and lumbar vertebrae L1-L2 were cut sagittally with a low-speed diamond saw (Buehler Isomet, Chicago, IL, USA). One-half was dehydrated with ethanol and embedded in methylmethacrylate. Histological sections (5 μm) were cut and stained according to von Kossa's method.(18) Trabecular bone volume was calculated as bone volume divided by total volume (BV/TV) in the secondary spongiosa of the lumbar vertebrae and in the proximal metaphysis of the tibia. Bone volume and total volume were measured interactively with an image-analysis system (Context Vision, Linköping, Sweden).
Three-point bending test on the right femur
Three-point bending tests were performed using a tensile testing machine (Zwick model 1445; Zwick GmbH and Co., Ulm, Germany). The load cell was a Hottinger Baldwin 5 kN cell with a resolution of 0.1N and a compliance of 0.6 μm/N. Crosshead displacement was measured with a Heidenhain MT101K extensometer (Dr. Johannes Heidenhain GmbH, Traunreut, Germany) (resolution, 0.2 μm). Data were recorded and analyzed using Zwick software version 5.02. The bending test used 4-mm-diameter rollers separated by 15 mm. The supporting rollers were fixed on the crosshead of the tensile tester and the loading roller was attached to the load cell. The right femur was cleaned of surrounding tissue and positioned freely in the tensile testing machine with the anterior aspect down such that the loading roller acted on the middle of the femur. Load was applied by moving the crosshead with the supporting rollers upward. A preload of 0.5N was reached with a speed of 0.1 mm/minute. The actual testing was performed at 0.2 mm/minute at 23°C and 50% relative humidity.
When a load is applied to the femur midshaft, it bends yielding a load-deformation curve. Initially, the femur displacement (d) increases elastically (i.e., linearly) with applied load (F). The maximum slope of this part of the curve (ΔF/Δd) represents the structural rigidity of the bone and is termed the stiffness. When the applied force is increased further, a yield point is reached after which bone deforms nonelastically (plastically) and linearity is lost. After maximum load is attained, bending increases farther while load decreases until fracture occurs. All force and displacement data were recorded for later evaluation, which took place after determination of the inner and outer width and the inner and outer height of the bone at the point of fracture. A video camera and image-analyzing system were used for the measurements of both sides of the fractured femur, assuming an elliptical shape of the femur shaft. Corresponding measurements for the two sides were averaged.
The actual effect of treatment on biomechanical competence can only be fully evaluated if the structural biomechanical parameters (i.e., maximum load and stiffness) are corrected for changes in geometric properties of the femur midshaft, yielding material biomechanical parameters, that is, maximum stress and Young's modulus. Therefore, in addition to the biomechanical characterization and diameter measurements, other geometric properties of the femur were calculated, such as the tested BV between the supporting rollers, second moment of inertia of the cross-section, and the wall-to-lumen ratio of the femur shaft at the fracture point. The calculation of these parameters was based on formulas already described.(19) Young's modulus (which represents the intrinsic stiffness of the bone material) was determined from the initial, linear part of the stress-strain curve. The formulas for stress (σ) and Young's modulus (E) are
where F is load, L is distance between supporting rollers, D is outer diameter height, Ix is second moment of inertia of the cross-section in relation to the horizontal axis, and d is displacement.
Preparation of vertebral body of lumbar L4 for compression test
The body of the isolated vertebra L4 was separated from the spinous processes by a longitudinal cut through the spine with a low-speed diamond saw. Subsequently, the vertebral body was fixed in the holder of a diamond precision parallel saw (Exakt Apparatebau; Otto Herrmann, Norderstedt, Germany) and sawed 1-2 mm underneath the end plates to obtain specimens with planoparallel ends and a total length of approximately 4 mm. To isolate the vertebral body from the remaining spinous cross-elements, the excised body was fixed in another holder and two 3-mm cuts were made to remove the pedicles.
Compression test on vertebral body of lumbar L4
Compression tests were performed using the same tensile testing machine and load cell as in the three-point bending test, and the crosshead displacement was measured with a Heidenhain MT12B extensometer (resolution 0.2 μm). The upper of two parallel, 20-mm diameter plates was attached to the load cell and the lower plate was fixed on the crosshead. The planoparallel L4 section was placed on the lower plate and the load was applied by moving the crosshead upward. A preload of 10N was reached with a speed of 1 mm/minute; the actual testing was performed at 2 mm/minute. Maximum load (Fmax) and corresponding displacement (dmax) and compression stress (σmax) were calculated. The stiffness (F/d) and Young's modulus were determined from the initial, linear part of the load-displacement curve (between 30% and 80% of maximal load). Formulas used for the compression test were
where H is original gauge length (height of vertebra) and A is cross-sectional area of vertebra (other parameters as mentioned previously).
Data analysis involved determination of geometric mean, SEM, and two-way analysis of variance (ANOVA) on log-transformed data by the Statistical Department of N.V. Organon using SAS software (SAS Institute, Inc., Cary, NC, USA). For each of the groups, the difference from the OVX vehicle-treated group was calculated. Results were considered statistically significant if p ≤ 0.05.
With pQCT, true bone mineral densities (mg/cm3) are determined and a discrimination between trabecular BMD and cortical BMD can be made. The effect of long-term estrogen deficiency on trabecular BMD was measured in the distal part of femur, which contains a large number of trabeculae, and on cortical BMD in the medial part of the femur, which consists of cortical bone only. The induction of estrogen deficiency caused by ovariectomy caused a statistically significant 75% decrease in trabecular BMD in the distal femur after 16 months (Fig. 1A). This decrease, as measured by pQCT in the inner part of the distal femur, is presumed to be caused by solely the loss of metaphyseal trabecular bone. Trabecular BMD was significantly higher in the OVX animals treated with tibolone 250 μg/day or 500 μg/day. EE treatment prevented loss of trabecular BMD to a similar extent as the highest dose of tibolone (Fig. 1A).
In the axial skeleton (lumbar vertebra L4), a 50% decrease in trabecular BMD was recorded 16 months after ovariectomy (Fig. 1B). This decrease was prevented completely by treatment with tibolone 250 μg/day or 500 μg/day. The 16-month treatment with EE only partially prevented the loss of trabecular BMD in vertebra L4 (not significantly different from the vehicle-treated group).
After 16 months of estrogen deficiency, cortical BMD of the medial (diaphyseal) part of the femur was slightly (3%) but significantly reduced (Table 1). Treatment with the two highest doses of tibolone partially prevented this decrease, although only the 250-μg/day dose reached statistical significance. However, EE treatment was not effective in preventing this decrease in cortical BMD in the medial part of the femur. The cortical BMD data are in line with the cortical thickness values (Table 1).
Table Table 1.. Effects of 16-Month Treatment with Tibolone or EE on BMD and Bone Biomechanical Parameters of OVX Rats
Histomorphometric measurements confirmed that trabecular bone volume (BV/TV) in the peripheral skeleton and axial skeleton was significantly reduced after ovariectomy in the vehicle-treated group. Decreases were 88% in the proximal tibia (peripheral skeleton; Fig. 2A) and 55% in lumbar vertebrae L1-L2 (axial skeleton; Fig. 2B). Trabecular bone volume was higher in the tibolone-treated groups compared with the vehicle-treated group in both tibia and lumbar vertebrae L1-L2. These effects appeared to be dose dependent. EE treatment had a similar effect as the highest dose of tibolone in preventing loss of trabecular bone volume in both tissues.
Biochemical bone turnover and liver parameters
Table 2 shows the effect of the 16-month tibolone or EE treatment on biochemical bone turnover and liver parameters. Sixteen months after ovariectomy, urinary deoxypyridinoline/creatinine ratio, a bone resorption marker, and plasma osteocalcin and alkaline phosphatase activity, both bone formation markers, were significantly higher in the OVX animals compared with the sham-operated rats. All three doses of tibolone significantly reduced the levels of osteocalcin and deoxypyridinoline/creatinine, indicating that bone turnover was reduced. Again, EE treatment had a similar effect as the highest dose of tibolone in reducing bone turnover. The alkaline phosphatase level was not significantly changed by either compound. Compared with the sham-operated controls, the level of GPT was significantly increased in the OVX group, and tibolone, 125-500 μg/day, or EE significantly reduced this level. There were no significant effects of tibolone or EE on plasma calcium, plasma phosphate, or urinary calcium/creatinine and phosphate/creatinine ratios (Table 2).
Table Table 2.. Effects of 16-Month Treatment with Tibolone or EE on Biochemical Parameters in Plasma and Urine of OVX Rats
Biomechanical quality of femoral cortical bone
The biomechanical quality of cortical bone was examined after 16 months' treatment with tibolone or EE using an ex vivo three-point bending test (Fig. 3A; Table 1). After 16 months of estrogen deficiency, maximum load and stiffness showed a tendency to decrease, although only load reached statistical significance (Table 1). Neither tibolone nor EE treatment had significant effects on these directly measured parameters. Estrogen deficiency for 16 months significantly increased the inner diameter of the diaphyseal midshaft of the femur. Treatment with the highest dose of tibolone inhibited this OVX-induced increase (Table 1).
Wall-to-lumen ratio at the point of fracture, which is an indicator of the relative intensity of periosteal bone apposition and endosteal bone resorption, was significantly decreased after ovariectomy (Table 1). Tibolone 250 μg/day or 500 μg/day tended to inhibit this decrease, although the effect was not statistically significant. EE treatment did not affect wall-to-lumen ratio at the point of fracture. Tibolone did not significantly alter femur length.
The total bone breaking force and corresponding structural biomechanical parameters in a three-point bending test are influenced greatly by bone size and geometry. Correcting maximum load and stiffness for changes in geometry yields maximum stress and Young's modulus, respectively. Ovariectomy resulted in a 20% decrease in maximum stress in femur, as compared with intact animals (p ≤ 0.05; Fig. 3A). Compared with the vehicle-treated OVX group, 16 months of treatment with 500 μg/day tibolone significantly increased maximum stress by 10%. With EE treatment, a similar increase was found, although statistical significance was not reached. Comparing the effect of EE treatment on maximum load (Table 1) and maximum stress (Fig. 3A) illustrates the importance of correcting biomechanical parameters for geometry changes, in modeling species like the rat. There was no dose-dependent effect of tibolone on Young's modulus in OVX rats (Table 1).
Biomechanical competence of the vertebral body
The vertebral body of lumbar L4 was subjected to ex vivo biomechanical testing, and load and stiffness values were normalized for areas of the midsection of the specimens using the formulas already described, yielding maximum stress and Young's modulus (Fig. 3B and Table 1, respectively). The midsectional areas of the vertebrae were not significantly different between the groups (not shown). OVX-induced long-term estrogen deficiency decreased maximum stress by 40% and Young's modulus by 55% in lumbar vertebra L4. Tibolone prevented the decreases in maximum stress and Young's modulus, and these effects appeared to be dose dependent. For both measures, 24 μg/day EE was only slightly more effective than the lowest dose of tibolone, and the effects were nonsignificant.
Retrospective analysis revealed a significant positive correlation between trabecular BMD in the midsection of lumbar L4 and maximum stress (r = 0.68; p ≤ 0.0001).
Ovariectomy in the rat causes a loss of trabecular bone from different sites of the skeleton, both peripheral and axial.3, 7-9, 20-21) This loss is caused by an increase in bone resorption, which is accompanied by an increase in bone formation. In contrast to intact rats in which bone formation equals bone resorption, the balance between bone resorption and bone formation is disturbed in OVX rats, and thus more bone is resorbed by the osteoclasts on the trabecular bone surface than is replaced by the osteoblasts.(3, 21) The remodeling activity in healthy bone is essential to retain bone quality and produce bone that can adapt appropriately to mechanical stimuli.
We have shown previously that tibolone is effective in preventing trabecular bone loss from the peripheral and axial skeleton of young and old OVX rats through a reduction in bone turnover (i.e., a reduction in resorption).12-14, 22) The objective of this study was to evaluate the effect of long-term tibolone treatment in OVX rats on the biomechanical competence of femoral cortical and vertebral trabecular bone, and also its effects on bone mass and bone turnover. EE was included in the study as a reference compound for the effect of estrogenic activity on bone modeling and remodeling. The dose of EE was chosen based on its estrogenic activity, which tibolone displays in a hormonal assay in rats (Allen-Doisy test).(10) Tibolone displayed approximately the same activity as a 10- to 20-fold lower oral dose of EE (by weight) in this in vivo test for estrogenic activity.(10, 11) In this study the estrogenic activities of the two highest doses of tibolone and the EE dose were roughly comparable.
As expected, the ovariectomy resulted in significant decreases in trabecular BMD (in the femur and in lumbar vertebra L4; Fig. 1) and in trabecular bone volume (in the tibia and in lumbar vertebrae L1-L2; Fig. 2) after 16 months. This loss of bone mass was accompanied by a significant increase in skeletal remodeling, as was evidenced by the enhanced levels of the bone turnover markers osteocalcin, alkaline phosphatase, and deoxypyridinoline (Table 2). Treatment with tibolone prevented the decreases in trabecular bone mass at least as effectively as EE in the peripheral and axial skeleton. These effects were reflected by the decreases in plasma osteocalcin levels and the urinary deoxypyridinoline/creatinine ratio (Table 2). Both markers were lower in tibolone-treated animals than in sham-operated animals, indicating a reduction in bone turnover.
These effects of tibolone are in line with data obtained after 4 months and 8 months of tibolone treatment in senescent, 20-month-old OVX rats(12) and in 12-month-old OVX rats.(22) The changes in bone turnover markers suggest that the mechanism of action of tibolone on bone may be similar to that of EE, which also inhibits bone resorption.(1, 3, 21, 23, 24) Although such inhibition generally would be considered beneficial, biomechanical competence of bone may be decreased if bone remodeling is inhibited excessively. In addition, biomechanical bone quality may be diminished if bone mineralization is decreased; this factor cannot be detected by bone mass measurements. A 25-day treatment with pamidronate (14 mg/kg per day), for instance, has been shown to decrease intrinsic diaphyseal bone strength in rats.(25)
Structural biomechanical properties are influenced by the biomechanical quality of the bone tissue (material biomechanical properties) and by the bone architecture or geometric properties. Performing three-point bending tests on rat femora, Ferretti et al.(26) found a significant positive correlation between second moment of inertia, which is a measure of the distribution of bone mass, and diaphyseal strength (r = 0.71) or stiffness (r = 0.68). However, ovariectomy, estrogen treatment, and the normal maturation process can each affect the geometric properties of the femur and thus affect bone strength. The practical effect of treatments on bone material biomechanical competence therefore only can be evaluated fully when structural biomechanical parameters (such as maximum load and stiffness) are corrected for the changes in geometric properties. These calculations yield the material biomechanical parameters maximum stress and Young's modulus. It was found that treatment with tibolone increased both maximum stress and Young's modulus of femoral cortical bone (Fig. 3A; Table 1). EE treatment produced effects of similar magnitude as seen with the highest dose of tibolone. These data suggest that tibolone treatment, like EE, improved cortical bone quality in these mature estrogen-deficient rats. At the start of the experiment the rats were 6 months old, skeletally mature, and were pair-fed for 16 months. This may explain why only limited changes in femur geometry and length were found after such a prolonged period of estrogen deficiency.
To examine the effect of long-term tibolone treatment on the quality of trabecular bone, a compression test was performed on a planoparallel cylinder of the L4 vertebral body. Estrogen deficiency for 16 months significantly decreased maximum stress and Young's modulus. Tibolone, 250-500 μg/day, prevented these decreases, yielding values that were significantly higher than in the vehicle-treated animals (Fig. 3B).
It is concluded that treatment of rats with tibolone for 16 months prevents the OVX-induced deterioration of trabecular bone in axial and peripheral skeleton, thereby maintaining biomechanical quality of trabecular bone. Additionally, biomechanical quality of cortical bone is improved by tibolone treatment compared with controls. In view of the similar effects of tibolone and EE on bone metabolism markers and bone parameters, tibolone is assumed to have an estrogen-like action on bone in that it reduces bone resorption and bone turnover. In light of tibolone's estrogenic actions on bone, it is speculated that tibolone may have positive effects on biomechanical bone quality in postmenopausal women.
This study was supported by N.V. Organon, Oss, The Netherlands.