Unmixed titanium (Ti) particles (Zimmer Biomet, Warsaw, IN, USA) with an average size of 5 µm were used. Prior to injection, the particles were rinsed in 70% ethanol for 48 h at room temperature, washed twice in PBS, and autoclaved at 180°C for 6 h to remove endotoxins. A commercial detection kit (E-Toxate; Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) was used to test whether the treated wear debris contained endotoxins or not (21).

A total of 45 10-week-old female C57BL/6J mice, each weighing 20±2 g, were used in the present study. All mice were maintained with pressure-controlled ventilation at a constant temperature of 25°C and a relative humidity of 40–70% in a 12/12-h light/dark cycle, and were given lab chow and water ad libitum. The study protocol was approved by the Animal Ethics Committee of Ningxia Medical University (Yinchuan, China).

Animal experiments were performed as previously described (20). In all mice, an intraperitoneal injection of Nembutal (0.6% pentobarbital sodium) was given to induce general anesthesia, and the murine joint prosthesis model was established in the right lower extremities. Under sterile conditions, the tibial plateau was exposed through the medial parapatellar approach and one Ti pin was gently implanted into the proximal tibia, with the pin head being maintained in the same plane as the tibial plateau surface. The skin incision was washed with normal saline containing 100 U/ml penicillin and 100 mg/ml streptomycin, and each layer was separately closed with absorbable sutures (20). Prior to surgically inserting the Ti pin, the mouse tibial canal was injected with 10 µl Ti suspension (4×10⁴ particles of Ti in normal saline). Subsequently, every 2 weeks following surgery, 20 µl Ti particles were injected into the joint capsule at week 2, 4, 6, 8, 10 and 12. Mice were randomly divided into three groups for treatment with SR (S12911-2; PROTELOS^®; Servier, Stoke Poges, UK): Control group (joint prosthesis only), SR625 group (joint prosthesis and SR at a dose of 625 mg/kg/day), and SR1800 group (joint prosthesis and SR at a dose of 1,800 mg/kg/day). A total of 7 days post-surgery, mice were given SR via intragastric gavage. Animals were treated consecutively for 12 weeks and were sacrificed for histological analysis, immunohistochemical (IHC) analysis, Ti prosthesis steadiness examination and micro-computed tomography (µCT) analysis.

Following sacrifice, the tibia containing the Ti pin was removed (20). To expose the Ti pin head, all muscles and tissues around the bone were carefully removed. Each bone was fixed to a special clamp using dental cement, which was designed to align the long axis of the implant with the long axis of the HP-100 Control Electronic Universal Testing Machine (Yueqing Zhejiang Instrument Scientific Co., Ltd). With the position of the mouse limb and the custom fixture controlled, the pin was pulled out of the tibial canal at a rate of 2.0 mm/min. Load data were recorded using automatic software (Edburg version 1.0; Yueqing Instrument Co., Ltd., Yueqing, China).

Following removal of all soft tissues, tibias from four mice per group were fixed in 4% paraformaldehyde, at 4°C for 4 weeks. The fixed shin bones were scanned by µCT (SkyScan 1176; Bruker microCT, Kontich, Belgium) at a resolution of 9 µm. The µCT scans were acquired at a 900-ms exposure time, 45-kW voltage and 550-mA current. Automatic data analysis software (NRecon version 1.1.11; Bruker microCT) was used to reconstruct and acquire images based on the µCT analyses, and to determine the bone volume fraction (BV/TV), trabecular thickness (Tb.Th), trabecular number (Tb.N), bone volume (BV), and specific bone surface (BS/BV) of the shin bone surrounding the Ti pin. All horizontal cutting images were captured at two-fifths of the titanium nail, which was 2 mm from the lower edge of the top hat.

Total RNA was extracted using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA), according to the manufacturer's protocol. The 260/280 absorbance ratio was measured to verify RNA purity (NanoDrop; Thermo Fisher Scientific, Inc., Wilmington, DE, USA). First strand cDNA was synthesized with 1 µg total RNA using the RevertAid First Strand cDNA Synthesis kit (Thermo Fisher Scientific, Inc.). A total of 2 µl cDNA was used for each PCR mixture, containing SYBR^® Premix Ex Taq™ II (Tli RNaseH Plus; Takara Biotechnology Co., Ltd., Dalian, China). The reaction was subjected to a 40-cycle amplification of 95°C for 30 sec, 95°C for 5 sec, and 60°C for 30 sec. The relative mRNA expression of selected genes was normalized to GAPDH and quantified using the 2^−ΔΔCq method (22).

PCR primers used in the present study were: RANKL forward, 5′-TCCTGAGCCTCCATGAAAACG-3′ and reverse, 5′-CCCACACTGTGTTGCAGTTC-3′; OPG forward, 5′-TGAAGTACCGGAGCTGTCCCC-3′ and reverse, 5′-AGGCCATATGTGCTGCAGTTCG-3′; and GAPDH forward, 5′-TTGTCAAGCTCATTGGGCTCATTT-3′ and reverse, 5′-GCCATGTAGGTCCACCCATG-3′.

The tibia was fixed in 4% paraformaldehyde for 24 h at 4°C, and immersed in EDTA solution for decalcification. The samples were dehydrated in a graded series of ethanol followed by xylene, prior to being embedded in paraffin at 60°C. Sections (5 µm) were cut perpendicular to the long axis of the tibia using an RM2235 Rotary Microtome-Basic Instrument (Leica Microsystems, Inc., Buffalo Grove, IL, USA). Sections were stained with hematoxylin and eosin (H&E) for histomorphometric analysis: 0.5% water-soluble Eosin for 5 min at 23°C and Hematoxylin for 3 min at 23°C. IHC staining for OPG and RANKL was implemented to assess the activity of osteoclastogenesis. EDTA was preheated to 60°C in a pressure cooker, then glass slides added to the autoclave for 2 min, then allowed to cool for 20 min. Following washing with PBS, the slides were incubated at 23°C for 10 min with 3% hydrogen peroxide, washed again with PBS and incubated with primary antibodies overnight at 4°C. The primary antibodies used were: Rabbit polyclonal anti-OPG (cat. no. ab183910; 1:300; Abcam, Cambridge, UK); and rabbit polyclonal anti-RANKL (cat. no. ab9957; 1:300; Abcam). To exclude the possibility of nonspecific staining, negative controls were performed with PBS instead of primary antibodies. Then the slides were incubated with secondary antibodies (Enzyme-labeled goat anti-rabbit IgG polymer) part of the PV-9001 kit (Sino Biological, Beijing, China) at 23°C for 40 min. Standardized IHC images were obtained with a microscopic imaging system (DM2000 LED; Leica Microsystems, Inc.), and positive expression was calculated using Image-Pro Plus version 6.0 software (Media Cybernetics Inc., Rockville, MD, USA).

Data are presented as the mean ± standard deviation. Results were analyzed by one-way analysis of variance, among the three groups. The least significant difference post-hoc test was performed for the distinction of means between different groups. P<0.05 was considered to indicate a statistically significant difference. SPSS 19.0 (IBM Corp., Armonk, NY, USA) was used for statistical analysis.

The average pulling load was 1.21±0.61 N for the control group. Compared with the control group, significant increases in pulling force were detected in the SR625 group (8.51±0.52N, P<0.01) and SR1800 group (13.42±0.13N, P<0.01). A significant difference in pulling load was additionally observed between the SR625 and SR1800 groups (P<0.01; Fig. 1).

µCT scanning demonstrated differences in the bone microstructure among the three groups. Osteolysis around the control group pin was most marked (Fig. 2). Tb.Th, Tb.N, BS/BV, BV and BV/TV data were obtained from µCT analysis of the region of interest. Compared with the control group (11.709±0.720%), BV/TV was significantly increased in the SR625 group (13.390±0.628%, P=0.048) and SR1800 group (15.288±0.184%, P=0.002) in a dose-dependent manner (P=0.031). Similarly, compared with the control group (0.109±0.006 mm³), BV was significantly increased in the SR625 group (0.125±0.004 mm³, P=0.048) and SR1800 group (0.142±0.002 mm³, P=0.002) in a dose-dependent manner (P=0.031). Conversely, compared with the control group (75.89±1.82 1/mm), a significant decline in BS/BV was observed in the SR625 group (64.98±1.77 1/mm, P=0.005) and SR1800 group (60.36±1.06 1/mm, P=0.001), although without a dose-dependent effect (P=0.123).

Additionally, compared with the control group (0.041±0.001 mm), Tb.Th was increased in the SR625 group (0.043±0.001 mm, P=0.175) and significantly increased in the SR1800 group (0.047±0.001 mm, P=0.011; Fig. 3D). Furthermore, compared with the control group (2.841±0.218 1/mm), Tb.N was increased in the SR625 group (3.050±0.157 1/mm, P=0.154) and significantly increased in the SR1800 group (3.219±0.027 1/mm, P=0.028; Fig. 3E).

H&E staining indicated areas of bone resorption (Fig. 4). IHC was used to detect the expression of OPG and RANKL in all groups. Fig. 5 illustrates the expression of OPG and RANKL in the bone around prosthesis. Compared with the control group, the expression level of OPG was significantly increased in the SR1800 group (0.422±0.010 vs. 0.770±0.012, respectively; P<0.001; Fig. 6A) and the expression levels of RANKL were significantly decreased (0.723±0.011 vs. 0.221±0.009, respectively; P<0.01; Fig. 6B). Similarly, compared with the control group, expression levels of OPG were significantly increased in the SR625 group (0.422±0.010 vs. 0.590±0.007, respectively; P<0.01; Fig. 6A) and levels of RANKL were significantly decreased (0.723±0.011 vs. 0.400±0.018, respectively; P<0.01; Fig. 6B). In addition, the expression levels of OPG and RANKL were significantly increased and decreased, respectively, to a greater extent in the SR1800 group compared with the SR625 group (P<0.01).

RT-qPCR analysis of OPG and RANKL in periprosthetic tissues demonstrated that, compared with the control group, the expression of OPG was significantly upregulated and the expression of RANKL was significantly decreased in response to treatment with SR (P<0.01). Additionally, this effect was significantly enhanced in the SR1800 group compared with the SR625 group (P<0.05; Fig. 7).