Healthy adult New Zealand rabbits (weight, 2.6–3.2 kg) were obtained from the Experimental Animal Center of Zhongnan Hospital of Wuhan University (Wuhan, China). All the experimental protocols used in this study were approved by the Animal Care and Use Committee of our Hospital.

The SANFH rabbit model was prepared according to previously described methods (1,18), with minor modifications. Briefly, 48 healthy New Zealand rabbits were randomly divided into 3 groups (n=16/group); the normal (no treatment), control (vehicle only) and treatment groups (SANFH model). Rabbits of the control and treatment groups were injected with horse serum (10 ml/kg; HyClone Laboratories, Inc., Logan, UT, USA) through an ear vein. After 2 weeks, 6 ml/kg horse serum was similarly injected once a day for 2 days, followed by an injection of 20 mg/kg methylprednisolone (Pharmacia and Upjohn Company, Puurs, Belgium) into the abdomen of rabbits in the treatment group twice a week for 2 weeks. Following the methylprednisolone injection, 200,000 units of penicillin was injected into the buttock of each rabbit. In the normal and control groups, an equal amount of saline was injected into the buttock muscle. Three animals from each group were examined using magnetic resonance imaging (MRI) and histopathological analysis 2 weeks after the hormone injection. Additionally, the blood and femur head samples of 3 animals from each group were collected 0 (prior to hormone injection), 2, 4 and 8 weeks after the hormone injection.

An orthogonal head coil was placed on the anesthetized rabbit, with its center located on the hip joint, and the fast spin echo (SE) was used. T₂-weighted imaging (T₂W_I; TR/TE, 2500/74 ms), T₁-weighted imaging (T₁W_I; TR/TE, 420/20 ms) and T₂W_I fat-suppression sequence (FS-T₂W_I) were performed twice at the coronal position.

Following the sacrifice of each rabbit, both femoral heads, including the metaphyses and thigh-bones, were removed. The tissue sections were dissected and fixed in 10% formalin, embedded in paraffin, cut into 5-μm-thick sections, de-paraffinized, rehydrated using standard techniques and stained using hematoxylin and eosin (H&E). The histopathological changes were evaluated using a microscope (Olympus, Tokyo, Japan).

The automatic biochemistry analyzer (Advia 1650; Siemens Medical Solutions, Erlangen, Germany) was used to determine the total cholesterol and triglyceride contents in the blood samples of SANFH rabbits.

Total proteins of the femoral head tissue were extracted and equal amounts of proteins (75 μg) were separated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and blotted on polyvinylidene difluoride (PVDF) membranes. These were probed with anti-HSDB1 and anti-HSDB2 rabbit polyclonal IgG (Wuhan Boster Bio-Engineering Co., Ltd., Wuhan, China) and subsequently with goat anti-rabbit/HRP IgG (Wuhan Boster Bio-Engineering Co., Ltd.), and detected using chemiluminescence. To determine the protein loading, antibodies against β-actin were used.

All the experiments were conducted in triplicate (at least) and the data are presented as the mean ± standard deviation (SD). The data were evaluated using one-way ANOVA, followed by Dunnett’s multiple comparisons test between different groups. The statistical significance of differences was analyzed using SPSS software (SPSS for Windows 15.0; SPSS Inc., Chicago, IL, USA) and P<0.05 was considered to indicate a statistically significant difference.

The femoral heads of rabbits in the normal and control groups were symmetrical. In FS-T₂W_I, low fat level signals were observed in the cortex of the femoral head in the normal and control groups. By contrast, MRI in the treatment group revealed a larger articular cavity of the femoral head in rabbits of the treatment group. The high-level FS-T₂W_I signal at the metaphyses indicated that edema was present in the bone marrow in the treatment group (Fig. 1).

As shown in Fig. 2, changes in the periosteum, cartilage, trabeculae and hematopoietic organization were observed. The periosteum of the femoral heads in the control and normal groups were smooth, and cartilage cells were arranged in an organized manner. The trabeculae were intact and their arrangement was regular, compact and full. The bone cells in the trabeculae were clearly visible with few empty bone lacunae. There were abundant medullary hematopoietic and small fat cells with a normal morphology. By contrast, the periosteum of the femoral heads in the treatment group was incomplete with partially shed cartilage cells. There were a few thin trabeculae with a disordered texture and a number of trabeculae were broken into fragments. A few spindle-shaped osteoblasts were distributed along the trabeculae. The medullary hematopoietic areas were poorly organized, with fewer cells, a sparse capillary network and partially obstructed blood vessels (Fig. 2).

The total cholesterol and triglyceride levels in the blood of rabbits in the treatment group were significantly higher compared with those of the control and normal groups (P<0.01; Fig. 3). Furthermore, the total cholesterol and triglyceride levels in the blood of rabbits in the treatment group gradually increased following injection until the end of the observation period.

The expression levels of 11β-HSD1 and 11β-HSD2 proteins were determined using western blotting (Fig. 4). The expression levels of 11β-HSD1 and 11β-HSD2 proteins in the control and normal groups maintained a stable level. By contrast, the expression levels of 11β-HSD1 protein in the treatment group increased 2 weeks after the injection, whereas the expression levels of 11β-HSD2 protein in the treatment group decreased 2 weeks after the injection. Additionally, the expression levels of 11β-HSD1 protein in the treatment group were significantly higher compared with those of the control and normal groups at each time-point after the injection (P<0.001). However, the expression levels of 11β-HSD2 protein in the treatment group were significantly lower compared with those of the control and normal groups at each time point after the injection (P<0.001). The relative protein level was normalized to the intensity of β-actin, as determined by western blotting.

Steroids have been used in clinical practice to inhibit inflammation, allergy and immune responses in numerous diseases. In 2003, steroids were widely used to treat severe acute respiratory syndrome (19). However, serious adverse reactions may be induced by steroid use, with SANFH being one of the most common. According to the results of previous studies, an increasing number of ANFH cases caused by steriod use have been reported and SANFH often occurs at an age of 30–50 years. Without treatment, hip joint dysfunction occurs in SANFH patients and this disease is considered to be an irreversible process (20,21). Numerous therapeutic methods have previously been used to treat SANFH, including hip arthroplasty, prosthesis and reconstructive surgery of the femoral head, in addition to treatments without surgery; however, these treatments were not particularly effective (22–24). Furthermore, the pathogenesis of SANFH remains unclear. Thus, investigation into the pathogenesis of SANFH and novel therapeutic strategies is urgently required.

In the present study, a rabbit model of SANFH was established to explore potential molecular mechanisms underlying SANFH development. MRI and histopathological analyses were used to evaluate the SANFH rabbit model and these determined that the model had been successfully established and was able to be used for further investigation in this study.

Metabolic disorder has been hypothesized to be the mechanism that underlies SANFH development. Following high hormone intake, the serum concentration levels of lipids, including total cholesterol and triglycerides, are increased, leading to hyperlipidemia. This may cause fat embolism to be induced in the peripheral vessels, leading to intravascular coagulation (3,7,25). Subsequently, fibrin platelets are formed and thrombopoiesis occurs, which induces osteonecrosis via microcirculation disturbance in the femoral head (26–28). Results of the present study demonstrated that the total cholesterol and triglyceride levels in the blood of rabbits treated with methylprednisolone were significantly increased compared with those of the control and normal groups; these results provide evidence in support of the metabolic disorder theory.

According to the results of previous studies, glucocorticoids are able to affect fat metabolism and endotheliocyte function, and may be associated with the necrosis of the femoral head (17). Additionally, glucocorticoids cause serious intramedullary fatty infiltration and induce a diminished blood flow. Complications, including intravascular coagulation, microcirculation disturbance and vascular embolization, are induced by fat cell and endotheliocyte dysfunction, which subsequently lead to ANFH. 11β-HSD1 and 11β-HSD2 are the two key enzymes that catalyze the conversion of inert 11-keto forms (cortisone) to active cortisol, or vice versa, thus regulating the access of glucocorticoids to steroid receptors (13). Results of the present study showed that the expression levels of 11β-HSD1 protein were increased, while the expression levels of 11β-HSD2 protein were decreased following injection. Therefore, these results indicate that 11β-HSDs are important in the development of SANFH and that 11β-HSDs are potentially important targets for preventing the development of ANFH in steroid-treated patients.