Six-to eight-week-old C57BL/6J mice were purchased from Shanghai Silaike Experimental Animals Center. All mice were housed in cages with ventilation filters under natural light and were provided food and water without any restrictions. The temperature in the room was maintained at 23°C. All animal experiments were approved by the Ethics Committee of Lishui Hospital (Zhejiang University, Zhejiang, China) and were conducted according to the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, China).

Azilsartan was obtained from Selleck Chemicals (Selleck Chemicals, Houston, TX, United States). Azilsartan was dissolved in DMSO (Sigma–Aldrich, Sydney, Australia) and stored at −20°C until use. Fetal bovine serum (FBS) and α-MEM were purchased from Gibco (Gaithersburg, MD, United States). Recombinant M-CSF and RANKL were purchased from R&D Systems (Minneapolis, MN, United States). Specific antibodies against c-Fos, CTSK, NFATc1, GAPDH, β-actin and HO-1 were purchased from Abcam (Cambridge, United Kingdom). Antibodies against p-JNK, JNK, p-ERK, ERK, p-P38, P38, p-IκBα, IκBα, p-P65, P65, and Nrf2 were purchased from Cell Signaling Technology (Danvers, MA, United States). Cell Counting Kit-8 (CCK-8) was purchased from Solarbio Science and Technology (Solarbio, Beijing, China).

We isolated bone marrow-derived monocytes/macrophages (BMMs) from bone marrow flushes of mouse lower limb long bones using a previously published protocol (Zhao et al., 2019). Then, BMMs were cultured in complete α-MEM containing 10% FBS, 25 ng/mL M-CSF, and 100 U/ml penicillin-streptomycin. After 4–5 days in culture, the BMMs were collected for experimental purpose.

Cell proliferation and cytotoxicity were assessed using CCK-8 assays according to the manufacturer’s protocol. Briefly, 2×10⁴ BMMs were seeded on the surface of 96-well plates, cultured in α-MEM for 24 h, and then treated with various concentrations of azilsartan. At 48 h and 96 h, the medium was replaced with 10 µL of CCK-8 reagent and 90 µL of complete α-MEM for 2 h in a 37°C incubator. Then, the optical density value of each well was measured at 450 nm.

We tested the effect of azilsartan on osteoclastogenesis and bone resorption function by performing TRAP staining and bone pit formation assays. Briefly, 1 × 10⁵ BMMs were seeded into 24-well plates and cultured with α-MEM containing M-CSF for 48 h. Then, different concentrations (0, 0.25, 0.5, or 1 μM) of azilsartan were added to the complete α-MEM containing RANKL (100 ng/ml) and M-CSF for the indicated periods, and the complete medium was replaced every 2 days. At last, cells were flushed 3 times with phosphate-buffered saline (PBS) and then fixed with 4% paraformaldehyde (PFA) for 20 min at room temperature. Finally, these cells were prepared for TRAP staining. TRAP-positive multinucleated cells with ≥3–5 nuclei were identified as mature osteoclasts.

The bone resorption function of osteoclasts was assessed by bone pit assay. 1 × 10⁵ BMMs were seeded on collagen-coated 6-well plates (Corning, Inc., NY, United States) first and then stimulated with M-CSF (25 ng/ml) and RANKL (100 ng/ml) until small osteoclasts formed. Small osteoclasts were released gently from collagen-coated 6-well plates using the cell dissociation solution (Sigma, MO, United States) and reseeded in equal numbers on Corning hydroxyapatite-coated plates (Corning Inc., NY, United States). Small osteoclasts were cultured in complete α-MEM containing RANKL (100 ng/ml) and different concentrations (0, 0.25, 0.5, or 1 μM) of azilsartan for 2–3 days to observe osteoclast-mediated bone resorption. Finally, the plates were soaked in 5% sodium hypochlorite solution for 2 min followed by washes with purified water until all cells were removed from plate. The bone resorption areas were photographed using Olympus light microscope (Olympus Life Science, Tokyo, Japan). Quantitative analysis was performed by ImageJ software (NIH; Bethesda, MD, United States).

Briefly, BMMs were seeded in 12-well plates (2× 10⁵ cells per well) and cultured in α-MEM containing M-CSF (25 ng/ml), RANKL (100 ng/ml) and different concentrations of azilsartan for 5 days. Total cellular RNA was extracted from osteoclasts using an Ultrapure RNA Kit (CWBIO Inc., Beijing, China) in accordance with the protocol. Next, 1 μg of total mRNA was reverse transcribed to cDNAs using a HiFiScript cDNA synthesis kit (CWBIO Inc., Beijing, China). Real-time quantitative PCR was performed using SYBR Green qPCR Master Mix (Yeasen, Shanghai, China) and an ABI 7500 machine (Thermo Fisher Scientific). Relative gene expression was normalized to the expression of β-actin or GAPDH using the 2^−ΔΔCt method. All primer sequences are listed in Supplementary Table S1.

Small interfering RNAs (siRNAs) that specifically target mouse Nrf2 gene were designed by RiboBio (RiboBio, Guangzhou, China). The siNrf2 sequences are listed in Supplementary Table S1. SiNrf2 were transfected into BMMs using Lipofectamine 3,000 (Invitrogen, CA, United States) in accordance with the manufacturer’s instructions. Briefly, 5× 10⁴ BMMs were seeded in each well of 24-well plates the day before transfection. After 48 h of incubation, BMMs were transfected with 20 nM siRNA. After 6 h, the medium containing transfection reagent was replaced with complete α-MEM containing M-CSF and RANKL (100 ng/ml). Forty-eight hours later, total mRNA was extracted for quantitative PCR analysis; 72 h later, total proteins were collected for Western blot analysis.

Total cellular protein was extracted from osteoclasts using RIPA buffer (CWBIO Inc., Beijing, China) containing PMSF (1%). Ten micrograms of total protein were separated on SDS–PAGE gels and transferred to PVDF membranes (Bio–Rad, Hercules, United States). Then, the membranes were blocked with milk (5%) in TBST for 1 h and incubated with the primary antibody (1:1,000) with shaking overnight at 4°C. After 16 h of incubation, the membranes were rinsed with Tris-Buffered Saline containing Tween-20 (TBST) for 3 times and incubated with a horseradish peroxidase (HRP)-conjugated secondary antibody for 1 h at room temperature. Antibody activities were detected with enhanced ECL hypersensitive chemical luminescence reagents (Yeasen, Shanghai, China). Images were acquired using an Invitrogen iBright 1,500 instrument (Thermo Fisher Scientific) and analyzed using ImageJ software.

Intracellular and intramitochondrial ROS activity were measured using a DCFH-DA probe (Yeasen, Shanghai, China) and a MitoSOX Red assay kit (Yeasen, Shanghai, China). Briefly, BMMs were stimulated with or without RANKL and cultured with different concentrations (0, 0.5, or 1 μM) of azilsartan for 24 h. Then, the probe (5 μM) was added to each well and incubated at 37°C for 20 min. Next, the wells were washed 3 times with cold PBS, and the fluorescence was observed with the fluorescence microscope (Olympus Life Science, Tokyo, Japan). The fluorescence intensity was measured by ImageJ Software.

A mouse ovariectomy (OVX)-induced osteoporosis model was adopted to evaluate the effect of azilsartan in vivo. Ten weeks old mice were randomly divided into 4 groups (n = 6 mice per group): sham group (ovaries were only exteriorized but not resected), vehicle group (bilateral ovariectomy + normal saline gavage), low-dose group (bilateral ovariectomy +1 mg/kg azilsartan gavage) and high-dose group (bilateral ovariectomy + 3 mg/kg azilsartan gavage). The dose of azilsartan used was based on the previous study (Abdelsaid et al., 2014; Sukumaran et al., 2017). All mice were anaesthetized by administering an intraperitoneal injection of pentobarbital (40 mg/kg body weight). Mice with a uniform heartbeat and respiration, relaxed muscles, no movement of limbs and, no touching reaction of whiskers were considered to have reached the state of complete anesthesia. The surgical procedure was performed as previously described (Xu et al., 2021). Azilsartan was administered by oral gavage twice a week for 6 weeks beginning on the seventh postoperative day. All mice were euthanized under anesthesia after 6 weeks of azilsartan intervention. The serum sample was collected from each mouse for liver functional enzyme analysis before the mice were euthanized. Then, all femurs and tibias of mice were harvested for the following experiments.

After fixation with 4% PFA for 48 h, the mouse right tibia and femur were scanned using a SkyScan 1,275 micro-CT (Bruker, Billerica, MA, United States). Data were analyzed using the following conditions: 50 kV, 9 μm resolution and 75 μA. All images were reconstructed with SkyScan CTAn software. The quantitative analysis was performed within a region set at 150 layers below the growth plate. Related parameters, including trabecular thickness (Tb. Th), bone volume per tissue volume (BV/TV), trabecular number (Tb. N), and trabecular separation (Tb. Sp), were recorded and analyzed.

All tibias were decalcified in 10% ethylenediaminetetraacetic acid (EDTA) (Sigma, Australia) at 4°C for 14 days. Tibias were dehydrated, embedded in paraffin blocks, and sectioned using a microtome at a thickness of 5 µm. Then, the bone sections were subjected to hematoxylin and eosin (H&E) and TRAP staining. The number of TRAP-positive osteoclast and osteoclast surfaces per bone surface (Oc.S/BS) was assessed in each sample using ImageJ software.

ROS levels were detected in vivo using a dihydroethidium (DHE) (Yeasen, Shanghai, China) probe. Briefly, fresh left tibias were fixed with 4% PFA at 4°C for 4 h. EDTA was applied for decalcification for 12 h and then replaced with a cryoprotectant solution. All tissues were stored in the cryoprotectant solution for another 24 h. Finally, bone sections were frozen. Sections with a thickness of 5 μm were prepared. Cell nuclei were stained with DAPI. Six random regions per group were quantified using ImageJ software.

In general, all data were recorded from three or more independent experiments and presented as the means ± SD. The results were analyzed using one-way analysis of variance (ANOVA) with Tukey’s post hoc test or Student’s t test with GraphPad Prism 7 software. p < 0.05 was considered statistically significant.

The chemical structure of azilsartan is shown in Figure 1A. We performed CCK-8 cell proliferation and cytotoxicity assays in 96-well plates after BMMs were treated with various concentrations of azilsartan to detect the cytotoxicity of azilsartan toward BMMs. As shown in Figure 1B, azilsartan had no effect on cell viability at concentrations ranging from 0 to 1 μM (48 h and 96 h). Next, TRAP staining results showed the formation of a large number of TRAP-positive multinucleated osteoclasts in the control group (0 μM), while osteoclastogenesis was inhibited in a dose-dependent manner by increasing concentrations of azilsartan (Figures 1C,D).

We treated BMMs with 0.5 μM azilsartan at different times to explore which stage of osteoclastogenesis was affected. Interestingly, osteoclast formation was more significantly inhibited when azilsartan was added at the mid-late stage (Days 3–5), while the inhibitory effect was not apparent at the early stage (Days 1–3) (Figures 1E,F).

We next tested whether azilsartan affected osteoclast bone resorption function. Mature small osteoclasts were reseeded on the surface of Corning hydroxyapatite-coated plates and treated with different concentrations of azilsartan (0, 0.25, 0.5, or 1 μM) for 3 days. As shown in Figure 2A, azilsartan treatment significantly reduced the bone resorption area in a dose-dependent manner compared with that of the untreated control group (Figure 2B).

Several osteoclast markers, including Cathepsin K (CTSK), Tartrate Resistant Acid Phosphatase (TRAP), Dendritic Cell-Specific Transmembrane Protein (DC-STAMP), D2 isoform of vacuolar ATPase Vo domain (ATP6v0d2), NFATc1, and c-Fos are upregulated during RANKL-induced osteoclastogenesis (Zheng et al., 2019; Yang et al., 2021). Our results revealed that azilsartan significantly suppressed the expression of these genes compared with the untreated control group (Figure 2C). Next, we investigated the expression levels of proteins associated with osteoclastogenesis and osteolysis, namely, NFATc1 and CTSK, using Western blot analysis. Consistent with the quantitative PCR results, different concentrations of azilsartan exerted an inhibitory effect on osteoclastogenesis and bone resorption, but were not toxicity effect of azilsartan (Figures 2D,E).

We identified the molecular mechanisms by which azilsartan inhibits osteoclastogenesis by examining the major signaling pathways affecting osteoclastogenesis, including NF-κB and MAPK. Western blot analyses indicated that both NF-κB and MAPK signaling were activated by RANKL stimulation. However, azilsartan treatment reduced p65 phosphorylation and IκBα phosphorylation, accompanied by the inhibition of IκBα degradation at 10–30 min (Figures 3C,D). Regarding the MAPK signaling pathway, the phosphorylation of P38, ERK, and JNK was significantly reduced after azilsartan treatment (Figures 3A,B), indicating a failure to activate MAPK signaling.

Furthermore, the expression of NFATc1/c-Fos, the essential transcriptional regulator involved in RANKL-induced osteoclastogenesis, was significantly suppressed on days 3 and 5 of osteoclastogenesis in BMMs treated with azilsartan (Figures 3E,F). Cathepsin K, which is considered a downstream protein of exercise-induced resorption function, was inhibited by azilsartan treatment (Figures 3E,F). In summary, our results showed that azilsartan suppressed osteoclastogenesis by inhibiting the NF-κB and MAPK pathways.

ROS play a critical role in osteoclast formation (Lean et al., 2003), and previous studies have revealed that angiotensin II (Ang II) and AT1R are associated with intracellular ROS production (Benigni et al., 2010; Chao et al., 2018). Next, we examined the ROS level in osteoclast mitochondria using the MitoSOX Red probe. RANKL stimulation significantly increased the ROS level in mitochondria. However, azilsartan treatment resulted in a dose-dependent decrease in ROS levels in mitochondria (Figures 4A,B). Similarly, ROS levels in the cytoplasm were significantly reduced by azilsartan, as detected with the DCFH-DA fluorescence probe (Figures 4C,D).

Nrf2, a critical protein in the antioxidant stress system, regulates oxidative stress and ROS production by binding to antioxidant response elements (AREs) and promoting the transcription of downstream antioxidant and detoxification enzymes (Sun et al., 2015). In our study, the expression of Nrf2 and HO-1 was increased in the presence of azilsartan (Figures 4E,F). The expression of other antioxidant enzymes in the cellular antioxidant system, including catalase and SOD1, was increased at the mRNA levels after azilsartan treatment (Figure 4G). In summary, azilsartan exhibited potent antioxidant properties in osteoclasts.

Next, we investigated the crosstalk among azilsartan, Nrf2, and osteoclastogenesis. A Nrf2-specific siRNA was transfected into BMMs. The efficiency of transfection was evaluated using Western blot analysis and quantitative PCR analysis, as shown in Figures 5A,B. Compared with the siCtrl group, a larger osteoclast area was observed in the siNrf2 group, and siNrf2 transfection diminished the inhibitory effect of azilsartan on osteoclast formation (Figures 5C,D). In addition, we further harvested the total cellular proteins from the aforementioned groups for Western blot analysis. Nrf2 silencing decreased the expression of antioxidant enzyme HO-1 and increased the expression of crucial osteoclast-associated transcription factors (NFATc1 and c-Fos) in osteoclastogenesis (Figures 5E,F). Surprisingly, azilsartan treatment had little effect on reversing siNrf2-mediated osteoclast formation and protein expression (Figures 5C,E), suggesting that Nrf2 likely acts as downstream of azilsartan.

Furthermore, we detected the levels of critical proteins involved in MAPK signaling to explore the underlying mechanisms by which azilsartan alters NF-κB and MAPK signaling. Nrf2 silencing using siNrf2 reversed the low levels of phosphorylated JNK, phosphorylated P38 and phosphorylated ERK in the azilsartan treatment group (Figures 5G,H). SiNrf2 transfection also reversed the repression of NF-κB signaling, as evidenced by increased levels of phosphorylated P65 and IκBα degradation (Figures 5G,I). Collectively, Nrf2 was identified as a downstream target of azilsartan that regulates MAPK and NF-κB signaling pathways in BMMs.

We established an osteoporosis model in mice to explore the effect of azilsartan in vivo. No mouse death or other adverse events occurred during OVX surgery or azilsartan administration. Compared to the sham-operated group, the vehicle group showed significant bone loss. However, the azilsartan-treated group exhibited increased bone mass (Figure 6A). By measuring and comparing the bone parameters, we observed increases in the trabecular number (Tb. N), trabecular thickness (Tb. Th), and bone volume/tissue volume (BV/TV) and a decrease in trabecular separation (Tb. Sp) in the azilsartan-treated group compared with the vehicle group (Figure 6B). Bone histomorphometry results from H&E staining confirmed that azilsartan treatment prevented estrogen deficiency-induced bone loss (Figure 6C). Significant trabecular destruction and bone loss were observed in the OVX groups. Azilsartan administration exerted a protective effect on bone in the low-dose groups and high-dose groups compared with the vehicle groups.

Likewise, TRAP staining showed that azilsartan treatment reduced the number of osteoclasts on each bone surface compared with that of the vehicle groups. The treatment also reduced the TRAP-positive cells/bone surfaces (Oc.S/BS%) ratio (Figures 6C,D). As azilsartan displayed good antioxidant activity in vitro, we next detected ROS levels using DHE fluorescent probes in frozen tibial sections. ROS levels within the bone marrow microenvironment were significantly increased after OVX surgery; however, azilsartan treatment reversed this trend (Figures 7A,B).

Furthermore, H&E staining of the lung, heart, spleen, liver, and kidney tissues indicated that azilsartan had no organ toxicity at the administered dose (Figure 7C). Blood biochemistry examinations of mouse serum showed no difference between the sham groups and azilsartan-treated groups (Supplementary Table S2; Supplementary Figure S1). Taken together, azilsartan administration ameliorates OVX-induced bone loss, possibly by inhibiting ROS production.