All procedures were evaluated and approved by the ethics committee of the Henan University of Traditional Chinese Medicine (Henan, China). A total of 48 male C57BL/6J mice (age, 8 weeks; weight, 20±2 g) were obtained from Slac Laboratory Animal (Shanghai, China; www.slaccas.com) and were allowed to acclimate to the environment for 1 week. The mice were given free access to food and tap water and were caged individually under controlled temperature (23±2°C) and humidity (55±5%) with an artificial 12-h light/dark cycle. The mice were randomly divided into four groups: Vehicle group (n=12) were injected with normal saline; dexamethasone (DXM), where mice were injected intramuscularly with 5 mg/kg body weight DXM three times weekly, for 12 weeks (n=12); low dose icariin GIOP, mice received oral icariin at a dose of 10 mg/kg/day for 12 weeks (Icariin + L, n=12) combined with DXM; high dose icariin GIOP, mice received oral icariin at a dose of 100 mg/kg/day, for 12 weeks combined with DXM (Icariin + H, n=12).

Human embryonic kidney HEK293T cells [cat. no. CRL-11268; American Type Culture Collection (ATCC), Manassas, VA, USA] were maintained in Dulbecco's modified Eagle's medium (DMEM; Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with 10% (v/v) fetal bovine serum (FBS; Sigma-Aldrich; Merck Millipore, Darmstadt, Germany), 100 U/ml penicillin, and 100 mg/ml streptomycin, in a 5% CO₂ atmosphere at 37°C.

Mice were placed in metabolic cages for 24 h to collect and recorded the volume of urine at week 12. Blood were collected after the animals were sacrificed at week 12 by intraperitoneal injection of sodium pentobarbital (2%; 200 mg/kg; cat. no. P3761; Sigma-Aldrich; Merck Millipore). The concentrations of calcium (Ca, absorbance was measured at 490 nm) and creatinine (Cr, absorbance was measured at 630 nm) in serum and urine were measured by standard colorimetric methods, using a microplate reader (BioTek Instruments, Winooski, VT, USA). The urinary Ca level was corrected against the urinary Cr concentration. Serum levels of alkaline phosphatase (ALP, cat. no. P0321; Beyotime Institute of Biotechnology, Haimen, China), tartrate resistant acid phosphatase (TRAP, cat. no. P0332; Beyotime Institute of Biotechnology), osteocalcin (OCN, cat. no. 60–1305; Immutopics, Inc., San Clemente, CA, USA) and deoxypyridinoline (DPD, cat. no. Ek-M20919; EK-Bioscience, Biotechnology Co., Ltd., Shanghai, China) were detected using mouse bioactive ELISA assays and measured using a SpectraMax M5 ELISA plate reader (Molecular Devices, LLC, Sunnyvale, CA, USA).

RAW264.7 cells (cat. no. TIB 71; ATCC) were cultured in α-MEM + 10% FBS (Sigma-Aldrich; Merck Millipore) and were plated in 6-well dishes at 1×10⁵ cells/well and were grown for 5 days with RANKL (20 ng/ml, cat. no. 462-TR; R&D Systems, Minneapolis, MN, USA) induction of osteoclastogenesis, as previously described (23).

Osteoclast (5×10⁴) lysates were prepared in 2 ml of lysis buffer, containing 25 mM HEPES, pH 7.5, 5 mM EDTA and 1 mM EGTA (all from Sigma-Aldrich; Merck Millipore), 5 mM MgCl₂ and 10 mM sucrose (both from Thermo Fisher Scientific, Inc.), 5 mM dithiothreitol (DTT), 1% 3-[-(3-chloramidopropyl) dimethylammonio]-1-propanesulfonic acid (CHAPS), protease inhibitor cocktail (10 µl/ml), and 1 mM PMSF (all from Sigma-Aldrich; Merck Millipore). After 30 min incubation at 0°C, cell lysates were centrifuged at 12,000 × g for 15 min at 4°C, and the supernatants were collected. The activity of caspase-3 was determined using the caspase-3 activity assay kit (cat. no. C1116; Beyotime Institute of Biotechnology), according to the manufacturer's instructions. Release of p-nitroaniline was measured at a wavelength of 405 nm using a SpectraMax M5 ELISA plate reader, according to the manufacturer's instructions.

Quantitative assessment of apoptotic osteoclasts (5×10⁴) was performed by terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling (TUNEL) method, which examines DNA-strand breaks during apoptosis, using an ApoAlert™ DNA Fragmentation Assay kit (BD Biosciences, Franklin Lakes, NJ, USA). Cells were trypsinized, fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100 in 0.1% sodium citrate. The cells were washed with PBS three times and subsequently incubated with the reaction mixture for 60 min at 37°C. Cells were immediately analyzed using a FACScan flow cytometer equipped with the CellQuest version 5.1 (BD Biosciences).

The tibias were collected after the animals were sacrificed and were decalcified in 0.5 M EDTA (pH=8.0) and embedded in paraffin by standard histological procedures. Sections (5 µm) were cut and stained with hematoxylin & eosin and safranin O, and were visualized under a Leica DM 2500 microscope (Leica Microsystems GmbH, Wetzlar, Germany). The trabecular bone microarchitecture of the proximal metaphysis of the tibia was measured using a SkyScan 1076 microtomography scanner (Bruker Corporation, Ettlingen, Germany), with a slice thickness of 22 µm. Bone morphometric parameters, including volumetric bone mineral density (BMD), bone volume over total volume (BV/TV), trabecula number (Tb.N), trabecula thickness (Tb.Th) and trabecula separation (Tb.Sp) were measured, as previously described (24). Femurs were placed in a 3-point bending configuration on an Instron Microtester 5848 system (Instron, Norwood, MA USA), to measure maximum load and stiffness, as previously described (25).

The 6-carboxyfluorescein-conjugated 2′-OMe-oligonucleotides were chemically synthesized and purified by high-performance liquid chromatography (GenePharma Co., Ltd., Shanghai, China). The 2′-OMe-miRNA (miR)-186 mimic was composed of RNA duplexes with the following sequence: 5′-CAAAGAAUUCUCCUUUUGGGCU-3′. The 2′-OMe-miR-1271 inhibitor sequence was: 5′-AGCCCAAAAGGAGAAUUCUUUG-3′, and the 2′-Ome-scramble sequence was: 5′-CUUGGUCAAGUCGCGAUCCGUA-3′. Osteoclasts (1×10⁵) were transfected using Lipofectamine 2000^™ (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA), at a final concentration of 100 nM. The culture medium was changed 24 h post transfection. Cells were harvested for analysis after 48 h.

The 3′-UTR of the cathepsin K gene containing the predicated target site for miR-186 was synthesized by GenePharma. The fragment was inserted into the multiple cloning site in the pMIR-REPORT luciferase miRNA expression reporter vector (Ambion; Thermo Fisher Scientific, Inc.). HEK293T cells (5×10⁴; cat. no. CRL-11268; ATCC) were co-transfected with luciferase reporters containing cathepsin K 3′-UTR and miR-186 mimics using Lipofectamine 2000. The cell lysates were harvested following 48 h transfection, and the luciferase activity was measured using a dual luciferase reporter assay kit (cat. no. RG027; Beyotime Institute of Biotechnology) according to manufacturer's instructions.

The tibias were crushed in liquid nitrogen and RNA extraction was performed using TRIzol^® according to the manufacturer's instructions (Invitrogen; Thermo Fisher Scientific, Inc.). Synthesis of cDNA was performed by reverse transcription of 2 µg total RNA, using Moloney murine leukemia virus reverse transcriptase (Invitrogen; Thermo Fisher Scientific, Inc.), with oligo (dT)₁₅ primers (Fermentas; Thermo Fisher Scientific, Inc., Pittsburgh, PA, USA), according to the manufacturer's instructions. The first strand cDNAs served as the template for the qPCR and were performed using an ABI 7300 thermocycler (Applied Biosytems; Thermo Fisher Scientific, Inc.). GAPDH served as an internal control to determine the relative expression of the target genes. The reaction conditions were set according to the kit protocol. The level of miR-186 was quantified using a mirVana qRT-PCR miRNA Detection kit (Ambion; Thermo Fisher Scientific, Inc.) and SYBR-Green (cat. no. D7260; Beyotime Institute of Biotechnology). Following the cycling reactions were used: 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec, 58°C for 30 sec, and 72°C for 30 sec. U6 served as the internal control. The quantitation cycle (Cq) method (26) was used to determine relative expression levels. The PCR primers (Sangon Biotech, Co., Ltd., Shanghai, China) used in the present study were as follows: forward, 5′-CCAGGGCGTACGGAGGCCATT-3′ and reverse, 5′-GACCAAATTACGGCGTAGCCTC-3′ for ALP; forward, 5′-AGCATAAGGGTCCAAGTCCAA-3′ and reverse, 5′-TACCAAAAGCGGCGTAGTTA-3′ for TRAP; forward, 5′-AGGCGGAGGTCGATGCCCCG-3′ and reverse, 5′-CACGATGATGTCACCCTCGATGT-3′ for cathepsin K, forward, 5′-ACCGTGAGTTGTCCGTAGCATC-3′ and reverse, 5′-CGTAAGGGTCCGATACATCTC-3′ for osteoprotegerin (OPG); forward, 5′-CGTAGGTAGCCGGGTAGGATC-3′ and reverse, 5′-GCGTGGGAGCCTGATGCTCA-3′ for receptor activator of nuclear factor-κB ligand (RANKL); forward, 5′-GGCAGGTGCTTCAGAACTGG-3′ and reverse, 5′-GTGGTGGCAGGTAGGTATGG-3′ for Runt-related transcription factor 2 (Runx2); forward, 5′-TGAGCTGGAACGTCACGTGC-3′ and reverse, 5′-AAGAGGAGGCCAGCCAGACA-3′ for osterix; forward, 5′-CTCAGGGTTTCAGTGGTT-3′ and reverse, 5′-TTTCCACGAGCACCCATC-3′ for collagen type 1, α1 chain (Col1α1); and forward, 5′-ACAGGGGAGGTGATAGCATT-3′ and reverse, 5′-GACCAAAAGCCTTCATACATCTC-3′ for GAPDH.

The tibias were homogenized and extracted in NP-40 buffer, followed by 5–10 min boiling and centrifugation (12,000 × g for 15 min at 4°C) to obtain the supernatant. Samples containing 50 µg protein were separated by 10% SDS-PAGE and transferred onto nitrocellulose membranes (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The membranes were incubated at room temperature for 4 h with primary cathepsin K antibodies (1:500; R&D Systems). The blots were washed 3 times with TBS with Tween-20 (TBST), and incubated with donkey anti-goat immunoglobulin G (IgG; cat. no. OARA01784) and donkey anti-mouse IgG secondary antibodies (cat. no. OARA01848) at room temperature for 2 h, conjugated to IRDye 800CW Infrared Dye (LI-COR Biosciences, Lincoln, NE, USA), diluted 1:10,000. The membranes were washed 3 times with TBST. Blots were visualized using the Odyssey Infrared Imaging System (LI-COR Biosciences) and normalized to β-actin expression, to correct for unequal loading, using a mouse monoclonal anti-β-actin antibody (cat. no. AP0060; Bioworld Technology, Inc., St. Louis Park, MN, USA).

Data are presented as the mean ± standard deviation for each group. All statistical analyses were performed using PRISM version 4.0 (GraphPad Software, Inc., La Jolla, CA, USA). Inter-group differences were analyzed by one-way analysis of variance, followed by a post hoc Tukey's honest significant difference test for multiple comparisons. P<0.05 was considered to indicate a statistically significant difference.

Administration of DXM significantly downregulated the serum Ca level, compared with the control group, and icariin treatment did not recover these levels in GIOP mice (Fig. 1A). In contrast, DXM exposure markedly increased the level of urine Ca; however, high dose icariin-treatment reversed this upregulation in GIOP mice (Fig. 1B). ELISA analysis of bone formation and resorption markers indicated that DXM induced high bone turnover in mice, as evidenced by a significant increase in serum ALP (Fig. 1C), TRAP (Fig. 1D) OCN (Fig. 1E) and urinary DPD/Cr (Fig. 1F) levels. However, mice in the high dose icariin + H group demonstrated decreased serum ALP, TRAP and OCN and urinary DPD/Cr.

BMD decreased by 27.73 and 42.59% at weeks 6 and 12, respectively, following administration of DXM (P<0.001), whereas administration with high-dose icariin (100 mg/kg) significantly restored the DXM-induced reduction in BMD in GIOP mice (Fig. 2A). A gross decrease in bone mass was observed following 12 weeks of DXM administration, which was inhibited when icariin was co-administered (Fig. 2B). Furthermore, structural parameters of the trabecular bone network in the tibial proximal metaphysis were measured. The results demonstrated that DXM-treated mice demonstrated significant decreases in trabecular BV/TV, Tb.N and Tb.Th, and increases in Tb.Sp, compared with the control group (Fig. 2C). Notably, icariin treatment in GIOP mice reversed the DXM-induced bone microarchitectural deteriorations. Bone biomechanical parameters, including maximum load and stiffness, decreased by 61.24 and 45.83% respectively, following DXM treatment. Mice in the icarrin + H group demonstrated significantly restoration of the DXM-induced biomechanical alterations in the tibial proximal metaphysis (Fig. 2D).

ALP and TRAP mRNA were significantly increased in the DXM group, compared with the control group (Fig. 3A and B); however, these were reduced by high-dose icariin co-administration. Furthermore, expression of the bone formation-associated genes Runx2, osterix and Col1α1 were decreased with DXM administration compared with the vehicle group, and this was prevented by high dose icariin treatment in DXM-induced GIOP mice (Fig. 3C). To further investigate the molecular mechanism underlying DXM-induced inhibition of bone formation and promotion of bone resorption, RT-qPCR was used to assess the osteoclastogenesis-associated genes RANKL and OPG, which are produced by the osteoblast. The results indicated that DXM significantly elevated the expression of RANKL, and significantly reduced the RANKL/OPG ratio. However, these alterations in RANKL expression and the RANKL/OPG ratio were reversed by icariin co-administration (Fig. 3D).

Cathepsin K mRNA and protein expression levels were significantly upregulated in DXM-treated mice, compared with the control group (Fig. 4A and B). However, these were reversed when icariin was co-administered in GIOP mice. Furthermore, the tibial levels of miR-186 were suppressed in GIOP mice compared with the control group; however, these levels were increased in icariin-supplemented GIOP mice (Fig. 4C). Notably, Spearman's rank correlation analysis indicated that the expression levels of cathepsin K protein and miR-186 were inversely correlated in DXM and icariin + H group mice (Fig. 4D). To further investigate the effect of cathepsin K on cartilaginous degeneration, safranin O staining was performed to observe the upper epiphyseal cartilage in the proximal tibias. The thickness of cartilage adjacent to the epiphyseal plate was reduced in the proximal tibia of the DXM mice, suggesting that DXM induces a delayed formation of new cartilage on the upper epiphyseal plate. However, the thickness of newly formed cartilage of the GIOP mice was improved by high concentrations of icariin (Fig. 4E).

The target prediction tools TargetScan, PicTar and microRNA.org were used to identify potential target genes of miR-186; the results indicated that miR-186 may regulate cathepsin K (Fig. 5A). To investigate whether miR-186 may directly regulate cathepsin K expression, a luciferase reporter construct containing the 3′-UTR of cathepsin K was used. The wild-type and mutant cathepsin K luciferase expression vectors were transfected alongside an miR-186 mimic into HEK293T, and the level of luciferase enzyme activity was measured. Overexpression of miR-186 suppressed the luciferase activity of the reporter gene; however, a mutation in the miRNA binding site abolished the luciferase activity repression, confirming that the cathepsin K 3′-UTR contains a seed site for miR-186 (Fig. 5B).

Transfection with an miR-186 mimic decreased the levels of cathepsin K in osteoclasts, whereas transfection with miR-186 inhibitor increased these levels. However, cathepsin K protein expression was reduced when icariin was administered alongside the miR-186 inhibitor (Fig. 5C). In addition, apoptotic features induced by transfection with miR-186 mimic or inhibitor were assessed. The results demonstrated that caspase-3 activity was significantly increased in the miR-186 mimic group compared with the control group. In contrast, osteoclasts transfected with an miR-186 inhibitor suppressed caspase-3 activity, and this downregulation of caspase-3 was reversed when icariin was co-administered with the miR-186 inhibitor (Fig. 5D). TUNEL staining indicated that transfection with the miR-186 mimic induced cell apoptosis, whereas the miR-186 inhibitor suppressed cell apoptosis, and icariin treatment reversed the suppression of cell apoptosis in miR-186 inhibitor-treated osteoclasts (Fig. 5E).