1  INTRODUCTION
     Osteoporosis is a chronic, progressive disease of the skeleton characterized by bone fragility due to a reduction in bone mass and possibly alteration in bone architecture which leads to a propensity to fracture with minimum trauma^［1］ . Many synthetic agents such as estrogens in hormone replacement therapy, selective estrogen receptor modulators like raloxifen, droloxifen, bisphosphonates and calcitonin have been developed to treat osteoporosis but each one of them is associated with side effects such as hypercalcemia, hypercalciurea, increased risk of endometrial and breast cancer, breast tenderness, menstruation, thromboembolic events, vaginal bleeding and hot flushes ^{［2, 3］} . So active attempts still are being vigorously pursued to find novel inhibitors of bone resorption that minimize bone loss in postmenopausal women, thus decreasing the necessity for drug therapy.
     The Herba Epimedii (Berberidaceae family) is a famous Chinese herbal medicine, made from the dried aerial parts of Epimedium brevicornum Maxim., Epimedium sagittatum (Sieb. et Zucc.) Maxim., Epimedium pubescens Maxim, Epimedium wushanense T.S. Ying. and Epimedium koreanum Nakai ^［4］ . They have commonly been used in the treatment of cardiovascular diseases and other chronic illnesses (bronchitis, infertility, amnesia and asthenia, impotence and senile functional diseases) in China for over 2 000 years ^［5］ . Herba Epimedii is a typical “kidney-tonifying” traditional Chinese medicine. According to traditional Chinese medicine theory, “kidney” controls bone. The “kidney-tonifying” action of traditional Chinese medicine might have relationship with bone formation. Currently, Herba Epimedii is applied in many Chinese formulas for anti-osteoporosis ^［6］ . Pharmacological studies also showed that it had potential function against osteoporosis ^{［7, 8］} . 
     In China, it has also been used to improve and tonify the Yang. However, it is not clear whether Herba Epimedii affects bone formation and what the mechanism is for anti-osteoporosis. The purpose of this study is to evaluate whether the extract of Epimedii sagittatum (ESE) is effective in ameliorating bone loss due to ovariectomy (OVX) and, if so, whether it functions in a manner similar to estrogen.

2   MATERIALS AND METHODS
2.1  Materials and reagents
         Aerial parts of Epimedii sagittatum (Sieb.et Zucc.) (including branches and leaves) were collected in a valley, located in Bozhou, Anhui Province, China, in September 2004. The raw material was identified by Han-chen Zheng, a professor of pharmacognosy in the Second Military Medical University, China. A reference specimen (voucher No. 20040903) has been deposited in the Herbarium of the Second Military Medical University, China. 17β-estradiol was purchased from Shanghai Hualian Pharmaceutical Co. Ltd. The reagent kits for measurement of calcium, inorganic phosphorus and alkaline phosphatase activity in serum were obtained from Fortune Bio-medical Engineering Co. Ltd. (Shanghai, China). Radioimmunoassay (RIA) kits for measurement of estradiol and bone gla protein (BGP) levels were purchased from Atomic Energy Institute of China. Goldner's staining reagents were purchased from Sigma Chemical Co. Ltd. Methyl methacrylate, dibutyl phthalate, benzoyl peroxide and other reagents were domestic analytical grade.
2.2   Preparation of extract
         Three liters of water were added to 0.3 kg fresh whole plant of E. sagittatum (Sieb.et Zucc.), and were decocted for 4 h. After filtration, the aqueous extracts were concentrated in 0.2 g (crude drug)/ml under reduced pressure at 50 ℃, and stored at －20 ℃ until use. The extract yield was about 3%. The extracts (ESE) diluted to 0.1 g (crude drug)/ml in water were administrated orally to rats at a volume of 10 ml/kg body weight. To the control rats, water was administered in a similar way.
2.3   Animals
          Fifty female Sprague-Dawley rats, 12 weeks of age, were purchased from SLAC Laboratory Animal Co. Ltd, Shanghai, China, and acclimated to conditions for one week before the experiment. To develop bone loss in ovariectomized rats, the experimental animals were housed in an air-conditioned room at (23±2) ℃ with 12 h/12 h light-dark illumination cycles at constant temperature (24±0.5) ℃ and humidity (45%-50%). Food and drinking water were supplied ad libitum. At the beginning of the study they were 3 months of age with an average weight of 200 g. The rats were weighed every week during the experiments. All animal experiments were performed according to the ethical guidelines suggested by the Institutional Animal Ethics Committee and Committee for the Purpose of Control and Supervision of Experiments on Animals, Ministry of Health, and Government of China.
2.4  Ovariectomy and administration of Epimedii extract
          The rats were randomly divided into five groups (n=10). Four groups were OVX groups and one group was sham-operated group. Rats in the four OVX groups were treated with vehicle (water), 17β-estradiol (1 mg/kg, once every week, ig) ^［9］ or ESE (0.5, 1 g/kg, daily, ig) for 11 weeks, respectively. Rats in the sham-operated group were treated orally with vehicle. 17β-estradiol tablet was dissolved and adjusted with physiological saline for an appropriate concentration.
          Surgery of animals was done under pentobarbital sodium (50 mg/kg, i.p.) anesthesia. Bilateral ovariectomies were performed from a dorsal approach with a small midline dorsal skin incision. The sham-operated rats were subject to sham surgery exposing, but the ovaries were not removed. Success of ovariectomy was confirmed at necropsy by failure to detect ovarian tissue and by observation of marked atrophy of uterine horns.
          At the end of the treatment, blood samples from rats of all the groups were withdrawn by tail vein method to assess biochemical parameters. The animals were then sacrificed by using ether anesthesia and the femora, tibia, and lumbar vertebra were dissected and stored in a freezer at －80 ℃ until examination for biomechanical testing and to study histopathological changes. Several organs, including the uterus, vagina and pituitary were dissected out, and immediately weighed ^［10］ .
2.5   Evaluation of parameters
2.5.1  Biochemical parameters
          Serum calcium (Ca) and inorganic phosphorus (Pi) concentrations, and serum alkaline phosphatase (ALP) activity were measured on an automatic analyzer (Ciba-Corning 550, USA) by using Beckman’s diagnostic reagent kit for the in vitro determination. The levels of estradiol (E2) and bone gla protein (BGP) were determined by using a specific and sensitive double-antibody RIA kit on a γ-ray counter (CAS-SN 695B, China) ^［11］ .
2.5.2  Bone mineral density (BMD)
          The femurs were cleaned of adhering soft tissues and the total length and width (midshaft) were measured with digital calipers. The bone mineral density (BMD, g/cm2) of the left femur was measured by dual-energy X-ray absorptiometry (LUNAR Co. Ltd., USA) by using the small animal scan mode ^［12］ .
2.5.3   Histomorphometric analysis
          The left proximal tibia metaphyses were opened to expose the marrow cavity by using an isomet low speed saw ( Buechler LTD, USA) and fixed in 10% phosphate buffer formalin for 24 h. They were then dehydrated in ethanol, defatted in xylene and embedded undecalcified in methyl methacrylate. The frontal sections were cut at 4-μm and 10-μm thickness with microtome (Leica RM 2155, Germany). The 4-μm sections were stained with Goldner’s Trichrome staining for static histomorphometric measurements, and the unstained 10-μm sections were used for dynamic histomorphometric analyses. Static measurements included the percentage of trabecular area (%Tb.Ar), trabecular thickness (Tb.Th), trabecular number (Tb.N) and trabecular separation (Tb.Sp). Dynamic measurements included the percentage of labeled perimeter (%L.Pm), mineral apposition rate (MAR), bone formation rate with tissue volume as referent (BFR/TV), bone formation rate with bone volume as referent (BFR/BV) and bone formation rate with bone surface as referent (BFR/BS).
2.6   Statistical analysis
          The data were analyzed by using one way analysis of variance (ANOVA) followed by post hoc Sheffe's Test using SPSS computer software Version 7.0. Level of significance was fixed at 0.05.

3    RESULTS
3.1   Body weight
          The body weight of OVX rats was significantly higher than that of the sham-operated rats (Figure 1). After 11 weeks, the body weights in the sham-operated and OVX rats were (290±13) g (mean±SD, n=10) and (358±14) g (n=10), respectively (Table 1). Increase in the body weight of animals treated with ESE (OVX+ESE) was almost the same as that in OVX+17β-estradiol rats (in 13～14 weeks) and less than that in OVX rats (Figure 1). With the administration of 17β-estradiol to the OVX rats (OVX+17β-estradiol), the increase was almost the same as that in the sham-operated rats.

Figure 1    Changes in body weight in rats
Rats were sham–operated (■), ovariectomized (●), and administered ESE (1 g/kg) (△) and 17β-estradiol (▲). The changes in ESE (0.5 g/kg) group were the same as ESE (1 g/kg) group, the curve of which was not shown. Each point represents the mean±SD (n=10). *P＜0.05 and **P＜0.01, vs ovariectomized group at corresponding times.

3.2   Uterine weight
          As shown in Table 1, the uterine weights of the sham-operated and OVX rats differed significantly (P＜0.01). The weight of the OVX+ESE (0.5 g/kg) group was higher than that of the OVX group nearly by 64%. The difference was statistically significant. The weight of the 17β-estradiol group was similar to that of the sham-operated rats. The weight of the OVX+ESE (0.5 g/kg) group was also almost similar to that of the OVX+ESE (1 g/kg) group.
3.3   Bone mineral density
           As shown in Table 1, the femur bone mineral densities (BMD) of the sham-operated and OVX rats were (0.263±0.011) and (0.248±0.009), respectively. This indicated that the ovariectomy decreased the BMD by 5.7%. The administration of ESE (1 g/kg) to the OVX rats significantly recovered the BMD. The OVX+17β-estradiol rats restored BMD almost completely to the level of the sham-operated rats, whereas the OVX+ESE (0.5 g/kg) group had no effect.
3.4   Serum parameters
           Results of serum parameters of different groups of animals are in Table 2.The results indicate a significant reduction in the serum calcium in OVX rats when compared with the sham-operated, ESE and 17β-estradiol treated animals. The serum phosphorus was decreased in OVX rats. 17β-estradiol can recover the level and ESE has no such effect. Statistically significant (P＜0.05） elevated level of serum ALP was observed in the OVX animals when compared with the sham-operated animals. Treatment with ESE reduced the serum ALP levels (P＜0.05）. Moreover, ovariectomy induced a decrease in serum E2, and treatment with ESE or 17β-estradiol improved E2 level compared with the OVX group. However, there were no significant differences in the serum BGP levels among any groups except OVX+ESE (1 g/kg) compared with OVX rats.
3.5   Histomorphometric examination
           This morphological observation was quantitated by histomorphometric analysis of longitudinal cross sections obtained from the proximal tibiae. A marked bone loss was observed in the OVX rats when compared with the sham-operated controls (Table 3). This bone loss was accompanied with a remarkable decrease in Tb.N and Tb.Th, and increase in Tb.Sp (vs sham-operated, Table 3). The 17β-estradiol could partially prevent the bone loss at the proximal tibia metaphysic (PTM) of OVX rats. In OVX rats treated with ESE, the %Tb.Ar increased (P＜0.05, vs OVX, Table 3）, Tb.Sp decreased (P＜0.05, vs OVX, Table 3）. The dynamic parameters of the bone such as BFR/BS and BFR/BV were increased in rats treated by ESE (Table 4). ESE had no significant influence on parameters such as %L.Pm, MAR and BFR/TV in OVX rats.
     
Table 1   Effects of ESE on body weight, uterine weight and BMD in OVX rats 12 weeks after administration
                                                                                                  (`x±s)

Group	n	Body weight (g)	Uterine weight (mg)	BMD (g/cm2)
Sham-operated	10	290±13**	836±22**	0.263±0.011*
OVX	10	358±14	140±6	0.248±0.009
OVX+17β-estradiol	10	302±18*	813±15**	0.264±0.011*
OVX+ESE (0.5 g/kg)	10	310±10*	224±9*	0.249±0.010
OVX+ESE (1 g/kg)	10	308±11*	230±9*	0.262±0.012*

*P＜0.05, **P＜0.01, vs OVX group.
     
Table 2    Effects of ESE on serum biochemical parameters in ovariectomized rats
                                                                                                                                         (`x±s)

Group	n	Calcium (mmol/L)	Phosphorus (mmol/L)	ALP (U/L)	E2 (ng/L)	BGP ( g/L)
Sham-operated	10	3.43±0.50*	2.32±0.02*	92.13±2.63*	25.1±2.1*	1.29±0.13
OVX	10	2.61±0.21	1.62±0.01	130.02±4.42	14.2±1.5	1.41±0.24
OVX+17β-estradiol	10	3.36±0.37*	2.19±0.02*	94.23±2.92*	21.2±2.0*	1.62±0.36
OVX+ESE (0.5 g/kg)	10	3.31±0.42*	1.67±0.03	112.45±3.00*	16.3±1.8	1.66±0.33
OVX+ESE (1 g/kg)	10	3.29±0.28*	1.65±0.02	113.11±3.51*	31.1±2.7*	1.73±0.21*

*P＜0.05, vs OVX group.
     
Table 3    Effects of ESE on static parameters of proximal tibial cancellous bone histomorphometry in ovariectomized rats
                                                                                                            (`x±s)    

Group	n	%Tb.Ar	Tb.Th (μm)	Tb.N	Tb.Sp (μm)
Sham-operated	10	23.97±2.94**	70.13±8.84*	3.44±0.34*	222.31±22.83**
OVX	10	7.32±1.09	56.57±1.42	1.29±0.19	1 345.50±127.59
OVX+17β-estradiol	10	13.57±2.27*	57.44±4.78	2.36±0.31*	372.54±32.13**
OVX+ESE (0.5 g/kg)	10	8.33±1.26	57.56±4.25	1.33±0.15	901.80±55.71
OVX+ESE (1 g/kg)	10	15.20±2.00*	60.95±0.57	1.54±0.23	825.23±45.56*

*P＜0.05, **P＜0.01, vs OVX group.
     
Table 4    Effects of ESE on dynamic parameters of proximal tibial cancellous bone histomorphometry in ovariectomized rats
                                                                                                                         (`x±s)    

Group	n	%L.Pm	MAR	BFR/BS	BFR/BV	BFR/TV
Sham-operated	10	22.89±5.07	0.69±0.08	15.72±4.08*	141.73±54.06*	32.89±8.50
OVX	10	31.79±5.88	1.19±0.08*	37.43±4.49	402.59±44.86	29.68±6.76
OVX+17β-estradiol	10	24.09±3.73	0.76±0.13*	18.19±3.98	194.25±48.92*	26.60±8.82
OVX+ESE (0.5 g/kg)	10	21.56±3.34	0.92±0.10	55.23±3.35*	523.33±45.23*	26.32±7.23
OVX+ESE (1 g/kg)	10	22.99±2.47	0.88±0.23	51.78±4.75*	517.73±47.71*	26.49±4.62

*P＜0.05, vs OVX group.

4  DISCUSSION
     The present study clearly demonstrates the usefulness and beneficial effects of ESE in the prevention of bone loss induced by ovariectomy. It is well known that estrogen deficiency is an important risk factor in the pathogenesis of osteoporosis. Ovariectomy results in great decrease in uterine weight, bone mineral density, histomorphometric index, and so on. These changes are mostly due to estrogen deficiency.
     Our data on body weight and serum alkaline phosphatase supported the observations of other investigators that both gained body weight and elevated serum alkaline phosphatase due to ovarian hormone deficiency are prevented by estrogen administration^［13］ . The present results showed that both ESE and 17β-estradiol inhibited both OVX-induced gained body weight and elevated serum alkaline phosphatase. It is well known that calcium, phosphorus, E2, and BGP are widely accepted phenotype markers of the bone formation ^［14］ . In this study, treatment with ESE and 17β-estradiol completely corrected the decreased concentrations of calcium, phosphorus, E2, and BGP in serum observed in OVX rats.
     The present data showed that OVX decreased the weight of uterus, when compared with the sham-operated group. Administration of ESE significantly increased the uterine weight in ovariectomized rats. These results suggested that ESE maybe possess of a characteristic of estrogen agonist. It thus further supports our previous observation. It is also suggested that ESE may have a similar effect as 17β-estradiol on bone loss and uterine weight.
     Our findings showed that OVX rats developed bone changes similar to those seen in estrogen deficient osteoporotic women, most markedly a decrease in bone density. ESE may significantly improve the bone density. In the present study, ESE is effective in preventing bone loss and estrogen deficiency for treatment of osteoporosis.
     Histomorphometric examination of the proximal tibiae of ESE groups revealed that the protective action of the extract may be due to an increase in bone formation ^［15］ .
     In conclusion, the results obtained in the present study provide evidences that ESE importantly contributes to the prevention or treatment of the development of bone loss induced by ovariectomy in rats. The studies on the isolation and characterization of the active chemical constituents are in progress.