According to our previous study (12), osteoblasts were isolated and extracted by digesting neonatal rats' calvariae using collagenase I (Sigma-Aldrich, St. Louis, MA, USA). Digests 3–6 were neutralized, pooled, and filtered. After we resuspended the single-cell suspensions, the single-cell suspensions was cultured in alpha-Minimal Essential Medium (α-MEM) which supplemented with 12% fetal bovine serum (FBS) and antibiotics. The operation procedures were approved by the Animal Care and Use Committee at our institution and complied with the guidelines of the international regulations.

The Cell Counting kit-8 (CCK-8; Dojindo Laboratories, Kumamoto, Japan) and Cell Apoptosis-DNA ELISA Plus kit (Roche, Palo Alto, CA, USA) was respectively used as quantitative measurement to assess the cell survival and the cell apoptosis. We recorded both the optic density (OD) of each well at 450 nm.

The lactate dehydrogenase (LDH) in the conditional medium was also measured by a two-step enzymatic reaction LDH assay kit (Takara, Tokyo, Japan) to assess the cell apoptosis. The percentage of LDH release=LDH (conditional medium)/LDH (conditional medium + cell lysates).

We used TUNNEL staining to detect the apoptotic osteoblasts by the In Situ Cell Death Detection kit (Roche). The cultured osteoblasts were fixed with freshly 4% paraformaldehyde for 1 h in a six-well plate. Then they were incubated with 3% H₂O₂ and 0.1% Triton X-100 for 10 min. During each step, they were washed three times with phosphate buffered saline (PBS). Finally, cells were stained with DAPI. A fluorescence microscope (Olympus, Tokyo, Japan) was used to obtain the images and assess the apoptotic cells. The TUNEL positive cells were calculated by three different authors.

As described (12), we used six-well plates to culture osteoblasts (3×10⁵ cells/well) and added Dex or/and Sal at different time. In the Sal group and Dex + Sal group, osteoblasts were pretreated with the optimum concentration of Sal (100 nM) for 2 h before the addition of Dex. We used radioimmunoprecipitation assay (RIPA) buffer containing phenylmethane sulfonyl fluoride (PMSF) to extract proteins from the osteoblasts. Then the protein concentration was calculated by the BCA Protein Assay kit (Beyotime Biotechnology, China). 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was adopted to resolve each protein lysate samples (30 µg in average). Then the solution was transferred onto polyvinylidene fluoride (PVDF) membranes. After the membranes were blocked and incubated with the primary antibody (1:1,000) overnight, they were incubated with each secondary antibody (1:3,000) for 2 h. Electrochemiluminescence Plus Reagent (Invitrogen Life Technologies, Carlsbad, CA, USA) was used to make the bands visible. We also used Image Lab 3.0 software (Bio-Rad, Hercules, CA, USA) to assess the band intensity. All the primary antibodies against cleaved-caspase-3, cleaved caspase-9, BAX, Bcl-2, cytochrome C, Akt and phosphor-Akt (Ser473) were purchased from Cell Signaling Technology (Danvers, MA, USA).

We chose 12-weeks-old adult male Sprague-Dawley rats (Animal Center of the Chinese Academy of Sciences, Shanghai, China) to build our animal model. We divided 45 rats equally into three groups (the control group, the model group and the Sal group). According to our previous study (12), The experimental rats in the model group and the Sal group were administered with double doses of lipopolysaccharide (LPS; Escherichia coli oup:B5; Sigma-Aldrich) on day 0 and 1 and received intramuscular injection of 40 mg/kg methlprednisolone (MPS) sodium succinate (Pfizer Pharmaceutical, Beijing, China) on day 3, 4, and 5. The rats in the control group were injected with normal saline (NS) instead. Then, the rats in the Sal group received an intraperitoneal injection of Sal (25 mg/kg/d) for seven consecutive days. The rest of the rats were injected with NS. All the injections were carried out at the same time each day and the time interval was 24 h.

We used 50×35×20 cm plexiglass cages to raise the rats. We also controlled the laboratory conditions (relative humidity of 50±10%, constant temperature of 20±2°C and 12 h light/dark cycle). All the rats were allowed free access to food and water. 8 weeks after the final injection of MPS, all the rats were sacrificed and bilateral femoral heads were collected for histological analysis. Our protocol was carried out in accordance with the National Institutes of Health guide for the care and use of Laboratory animals and was also approved by the Animal Care and Use Committee of Wenzhou Medical University.

Femoral head samples, collected from rat models in vivo, were fixed in paraformaldehyde for 24 h at 4°C. Then we transferred the samples to 10% EDTA solution at 4°C for 2 months. After the samples were dehydrated in ethanol, the samples were embedded in paraffin and cut into 4-µm thick sections in the coronal plane. We used hematoxylin-eosin for tissue section staining. With a microscope (Leica Microsystems, Wetzlar, Germany), we observed the pyknotic nuclei of osteocytes, empty lacunae, bone marrow cell necrosis and bone trabeculae broken. The percentage of empty lacunae was assessed blindly by three authors.

We repeated all the experiments at least three times with consistent outcomes. We used SPSS 21.0 software (IBM SPSS, Armonk, NY, USA) to convert all outcome measurements. The one-way ANOVA followed by Tukey's post hoc test was used to calculate statistical significance, and P<0.05 was considered to indicate a statistically significant difference.

Sal showed a significant protective effect on cell survival assay and against Dex-induced cell death (cell apoptosis assay and LDH release assay) when the concentration reached 10 nM (P<0.05). And the protective effect of Sal was most obvious at 100 nM (P<0.01) (Fig. 1A-C). Furthermore, we used CCK-8 to value the cytotoxic effect of Sal. As shown in Fig. 1D, Sal had no potential cytotoxic effect on the primary murine osteoblasts in Sal alone group, which reflected in the decline of CCK-8 OD, when the concentration was not more than 100 nM (P<0.05).

The TUNEL assay revealed that TUNEL positive cells significantly increased when treated with Dex (P<0.01) (Fig. 2). And the pretreatment of Sal obviously decreased TUNEL positive cells compared with the Dex group (P<0.01) (Fig. 2). What's more, there is no significant difference between the Sal alone group and the control group (P>0.05).

Sal alleviated Dex-induced osteoblasts apoptosis via activating the PI3K/Akt signaling pathway and downregulating the caspase protein expression. To explore the mechanism by which Sal decreased Dex-induced apoptosis, western blot analysis was performed to study changes in the PI3K/Akt signaling pathway. As shown in Fig. 3A, B, Dex significantly decreased p-Akt level in the Dex alone group compared with the control group (P<0.01). Pretreatment with Sal for 2 h significantly increased p-Akt expression in Dex-induced osteoblasts apoptosis (P<0.01). And the expression of p-Akt in the Sal alone group did not show any change compared with the control group (P>0.05). Total Akt levels in each group did not change.

To further investigate the effect of PI3K/Akt signaling in the caspase protein expression, cells were pretreated with Sal for 2 h and then stimulated with Dex for 24 h. Sal significantly downregulated the expression of Dex-induced cleaved caspase-9 and cleaved caspase-3 (P<0.01) (Fig. 3C, D).

We used western blot analysis to explore whether the mitochondrial pathway is involved in the Sal-induced protective effect against Dex-induced osteoblasts apoptosis. The expression levels of BAX, Bcl-2 and cytochrome C were detected. As shown in Fig. 4, compared with the control group, the expression level of pro-apoptotic protein Bax significantly increased (P<0.01) while the expression level of anti-apoptotic protein Bcl-2 was on the contrary in the Dex group after exposure to Dex for 24 h (P<0.01). Pretreatment with Sal for 2 h significantly decreased BAX expression and increased Bcl-2 expression in Dex-induced osteoblasts apoptosis (P<0.01). Moreover, we found cytochrome C increased significantly by Dex (P<0.01), but the expression could be inhibited by Sal (P<0.01). And the expression of BAX, Bcl-2 and cytochrome C in Sal alone group did not show any change compared with the control group.

The Sal group (2/15) significantly decreased the incidence rate of SANFH compared with the model group (13/15) (P<0.01). As shown in Fig. 5, the percentage of empty lacunae increased significantly in the model group (P<0.01). Additionally, Sal reduced the percentage of empty lacunae compared with the model group (P<0.01) while Sal would not increase the percentage of empty lacunae compared with the control group.