Inflammation ( # 2015) DOI: 10.1007/s10753-015-0241-y

Alteration of the RANKL/RANK/OPG System in Periprosthetic Osteolysis with Septic Loosening Long Wang,1 Zixun Dai,1 Jie Xie,1 Hao Liao,1 Cheng Lv,1 and Yihe Hu1,2

Abstract—The pathogenesis of periprosthetic osteolysis with septic loosening remains incompletely understood. The purpose of this study was to investigate whether expression of the RANKL/RANK/ OPG system is altered in septic interface membranes (SIMs). Seventeen cases with a SIM, 26 cases with an aseptic interface membrane (AIM), and 12 cases with a normal synovium (NS) were assessed. Scanning and transmission electron microscopy (SEM and TEM, respectively) were used to observe the microscopic morphology of three tissue conditions. Differences in RANKL, RANK, and OPG expression at the mRNA level were assessed by real-time quantitative PCR, and differences at the protein level were assessed by immunohistochemical staining and Western blotting. SEM showed wear debris widely distributed on the AIM surface, and TEM showed Bacillus activity in the SIM. RANKL expression and the RANKL/OPG ratio were significantly increased in SIMs. Imbalance in the RANKL/RANK/OPG system is related to periprosthetic osteolysis with septic loosening but is not the only possible pathogenic mechanism. KEY WORDS: RANKL/RANK/OPG system; periprosthetic osteolysis; septic loosening; aseptic loosening; interface membrane.

INTRODUCTION Total hip arthroplasty (THA) was considered Bthe operation of the century^ [1]; but despite the undoubted benefits of the surgery, premature implant failure remains a considerable clinical problem. Failure may occur due to a number of reasons, and two main reasons are complications known as periprosthetic joint infection (PJI) and aseptic loosening. These complications seriously affect the clinical effects of THA and survival life of prostheses 1

Department of Orthopaedics, Xiangya Hospital, Central South University, No. 87 Xiangya Road, Changsha, Hunan 410008, People’s Republic of China 2 To whom correspondence should be addressed at Department of Orthopaedics, Xiangya Hospital, Central South University, No. 87 Xiangya Road, Changsha, Hunan 410008, People’s Republic of China. E-mail: [email protected] Abbreviations: RANKL, Receptor activator of nuclear factor-κB ligand; RANK, Receptor activator of nuclear factor-κB; OPG, Osteoprotegerin; SIMs, Septic interface membranes; AIM, Aseptic interface membrane; NS, Normal synovium; SEM, Scanning electron microscopy; TEM, Transmission electron microscopy; THA, Total hip arthroplasty; PJI, Periprosthetic joint infection; rTHA, Revision total hip arthroplasty; OCs, Osteoclasts; OCPs, Osteoclasts precursors; MCSF, Macrophage colonystimulating factor; TLRs, Toll-like receptors

[1–4]. PJI is considered a devastating complication of total joint arthroplasty, with an approximate incidence of 1–3 % [5–7], and may contribute to septic loosening. Once the diagnosis of septic loosening is confirmed clinically, treatments with intravenous antibiotics, component resection, temporary antibiotic-containing spacer placement, and a new component reimplantation are needed, which represents a serious burden to the patients, society, economy, and healthcare resources. The pathogenesis of septic loosening and that of aseptic loosening have essential differences; however, both are accompanied by the occurrence of periprosthetic osteolysis. Periprosthetic osteolysis results in significant bone loss around the acetabular and femoral components of the hip joint, which increases the difficulty of revision THA (rTHA) and leads to unsatisfactory surgical outcomes. Therefore, several studies have investigated the course of periprosthetic osteolysis following THA. However, most studies concentrated on osteolysis caused by wear debris/aseptic loosening [8–13]. In the past decade, major advances have been made in our understanding of physiological bone metabolism. The receptor activator of nuclear factor-κB ligand (RANKL)/ RANK/osteoprotegerin (OPG) system is considered to

0360-3997/15/0000-0001/0 # 2015 Springer Science+Business Media New York

Wang, Dai, Xie, Liao, Lv, and Hu play central roles in the development of osteoclasts (OCs) [14, 15]. RANKL is the key cytokine regulator of osteoclast generation and activation. RANKL binds to RANK expressed on the surface of OCs and OC precursors (OCPs) [16] and is necessary for the differentiation of OCPs to mature and functional OCs in the presence of macrophage colony-stimulating factor (MCSF) [17]. OPG is a naturally occurring decoy receptor for RANKL and functions to downregulate osteoclastogenesis by binding RANKL, thus preventing its interaction with RANK [18]. The RANKL/RANK/OPG system is considered an important signal transduction pathway in the formation, differentiation, and activation processes of OCs [14, 19]. Previous studies have suggested that the RANKL/RANK/OPG system is related to periprosthetic osteolysis with aseptic loosening, and several studies have demonstrated this correlation [8–13]. Compared to previous studies on osteolysis related to aseptic loosening, studies of periprosthetic osteolysis with septic loosening are rare, and the corresponding pathogenesis is unknown. Whether the RANKL/RANK/OPG signaling pathway also plays an important role in osteolysis related to septic loosening has yet to be demonstrated. In this study, differences in the mRNA and protein expression of RANKL, RANK, and OPG were assessed in septic interface membrane (SIM), aseptic interface membrane (AIM), and normal synovium (NS) samples in order to investigate the role of the RANKL/RANK/OPG system in the pathogenesis of periprosthetic osteolysis with septic loosening.

MATERIALS AND METHODS Patients and Samples Research involving patient samples was approved by the Ethics Committee of our hospital (equivalent to an Institutional Review Board). Seventeen SIM samples were obtained from patients with PJI and periprosthetic osteolysis who underwent rTHA. PJI after THA was diagnosed according to previously published criteria [2, 20], which are listed in Table 1. Among the 17 patients with a SIM, nine patients were women and eight were men, with a mean age of 68.4 years (range, 54–76 years). Twenty-six AIM samples were obtained from patients with aseptic loosening and periprosthetic osteolysis who underwent rTHA. Among these patients, nine were women and 17 were men, with a mean age of 67.5 years (range, 47– 77 years). Twelve normal synovium samples were

Table 1. Diagnostic Criteria for Periprosthetic Joint Infection after THA 1. Presence of a sinus tract communicating with the prosthesis or articular cavity 2. Isolation of a pathogen from a culture of synovial fluid from the preoperative and intraoperative joint aspirate or periprosthetic tissue from the affected prosthetic joint 3. Visible purulence of a preoperative aspirate or intraoperative periprosthetic tissue 4. Existence of three of the following four criteria: a. Elevated serum erythrocyte sedimentation rate and serum C-reactive protein concentration b. Elevated synovial white blood cell count c. Elevated synovial polymorphonuclear percentage d. Greater than five neutrophils per high-power field in five high-power fields observed via histological analysis of periprosthetic tissue at ×400 magnification. PJI was considered if one of the four criteria above was met

obtained from patients with femoral neck fracture and without periprosthetic osteolysis during the primary THA procedure. Among these patients, three were women and nine were men, with a mean age of 75.5 years (range, 64– 82 years). Patients with a bone tumor, bone tuberculosis, rheumatoid arthritis, or metabolic bone disease were excluded. All patients enrolled in this study provided written informed consent preoperatively. All samples harvested intraoperatively were immediately frozen in liquid nitrogen at −196 °C and stored at this temperature until further analysis. Morphological and Structural Evaluations We studied the morphology of the three different types of tissues using scanning electron microscopy (SEM, Quanta-200, FEI, The Netherlands). Specimens were prepared for SEM viewing by double fixation, dehydration, and gold coating. The structures of tissues were analyzed using transmission electron microscopy (TEM, H-7500, Hitachi, Japan). Specimens were prepared for TEM by double fixation, dehydration, infiltration, embedding, and ultrathin sectioning. The specimen sections were stained by uranyl acetate and lead nitrate for TEM viewing. Immunohistochemistry and Image Analysis All harvested tissues were fixed in 10 % neutral buffered formalin, paraffin embedded, and sectioned at 5 μm. Endogenous peroxidase activity was blocked in 3 % H2O2 for 15 min. Antigen retrieval was achieved via the high-pressure method [21]. First, sections were incubated with the following primary antibodies: anti-RANKL (sc-7628, Santa Cruz Biotechnology, Santa Cruz, CA,

Alteration of the RANKL/RANK/OPG System in Periprosthetic USA, dilution 1:100), anti-RANK (sc-9072 Santa Cruz Biotechnology, dilution 1:100), and anti-OPG (3448-1, Epitomics, Burlingame, CA, USA, dilution 1:200) at 37 °C for 30 min and then at 4 °C overnight. After washing with phosphate-buffered saline (PBS), the slides were incubated with biotinylated secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA, dilution 1:1000) for 30 min at 37 °C. Incubation was stopped in PBS before the DAB-complex was administered. To confirm the specificity of the immunostaining, control straining was performed by replacing the primary antibody with goat serum at the same dilution in PBS. Immunostained sections were characterized quantitatively by digital image analysis using the Image-Pro Plus (version 5.0, Media Cybernetics, Rockville, MD, USA). According to the method introduced by Xavier and Wang [22, 23], five random digital images of each specimen at ×400 magnification were captured using a Leica DM4000B microscope (Leica, Solms, Germany). The measurement parameters included the mean density, summed area, and the integrated optical density (IOD). The optical density was calibrated, and the area of interest was set through: hue (RANKL, 0~64, RANK, 0~56, OPG, 0~64); saturation, 0~255; and intensity, 0~255. Then, each image was converted to a grayscale image, and the IOD value was calculated. Real-Time Quantitative Polymerase Chain Reaction Interface membrane or normal synovium samples (SIM 17 samples, AIM 26 samples, NS 12 samples) stored in liquid nitrogen were cut into pieces and then 100 mg of each sample was ground using a mortar. Liquid nitrogen and 1000 μl RNA TRIzol (Invitrogen Life Technologies, UK) was used to fully grind the tissues. Total RNA was isolated using TRIzol reagent, according to the manufacturer’s specification (Life Technologies, Carlsbad, CA, USA). The yield of RNA was measured photometrically. We used 10 μg mRNA to prepare primary cDNA by using the M-MLV reverse transcriptase kit (Promega, Madison, WI, USA). RNA was analyzed by RT-qPCR using a PCR machine (7500, Applied Biosystems, Foster City, CA, USA) and the SYBR Green RT-PCR kit (Applied Biosystems) according to the manufacturer’s instructions. Conventional PCR was performed to obtain a product of amplification suitable for the construction of standard curves for the real-time PCR procedures. The incorporation of Sybr Green into the PCR products was monitored in real time after each PCR cycle, allowing for the calculation of the threshold cycle or Ct value that represents the PCR cycle number at which exponential growth of PCR products begins. PCR cycle conditions were

as follows: 5 min at 95 °C, 35 to 40 cycles of 15 s at 95 °C, 30 s at 59 °C, and 5 min at 72 °C. Each PCR procedure included a negative control reaction without a template. To exclude residual DNA contamination of the RNA samples, RT-PCR was also performed without reverse transcriptase. For mRNA amplification, the validated primers were designed using Primer 3.0 software and obtained from Sangon (Sangon, Shanghai, China). The sequences of the primers used in RT-qPCR are listed in Table 2. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) housekeeping gene was used as a reference for the relative quantification of each gene of interest, which was expressed as the RQ value, which is the ratio of the expression of the target to the expression of GAPDH.

Western Blot Analysis Tissues stored in liquid nitrogen were cut into pieces, and 100-mg samples were homogenized completely using 400 U of single detergent lysis buffer (containing phenylmethanesulfonyl fluoride). The tissues together with lysates were treated for 30 min on ice and then centrifuged at 12,000 rpm for 20 min at 4 °C. The supernatant was recovered as the total sample lysate, aliquoted, and stored at −80 °C. Equal amounts of protein (50 μg) were separated by 8 % sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and electro-transferred to polyvinylidene difluoride membranes (0.45 μm, Millipore, Bedford, MA, USA). Membranes were blocked with 5 % non-fat milk and 0.1 % Tween 20 in Tris-buffered saline for 1 h at room temperature and then incubated overnight at 4 °C with the same primary antibodies listed above for RANKL (dilution 1:500), RANK (dilution 1:500), and OPG (dilution 1:500). The appropriate peroxidase-conjugated anti-mouse, anti-rabbit, and anti-goat antibodies (Jackson ImmunoResearch Laboratories, dilution 1:10,000) were then incubated for 1 h at 37 °C, followed by the detection of proteins by enhanced chemiluminescence (Amersham Bioscience, Table 2. Sequences of Primers Used for RT-qPCR Gene RANKL RANK OPG GAPDH

Primer sequence Forward Reverse Forward Reverse Forward Reverse Forward Reverse

5′-CACTATTAATGCCACCGAC-3′ 5′-GGGTATGAGAACTTGGGATT-3′ 5′-ATGCGGTTTGCAGTTCTTCTC-3′ 5′-ACTCCTTATCTCCACTTAGG-3′ 5′-GCTTGAAACATAGGAGCTG-3′ 5′-GTTTACTTTGGTGCCAGG-3′ 5′-CAGAACATCATCCCTGCCTCT-3′ 5′-GCTTGACAAAGTGGTCGTTGAG-3′

Wang, Dai, Xie, Liao, Lv, and Hu

Fig. 1. Scanning electron microscopy observation of the three tissue types. a Septic interface membranes (SIM), b aseptic interface membranes (AIM), and c normal synovium (NS). SEM images showed an extensive presence of wear particles and debris in AIM samples (Fig. 1b, arrow). In contrast, only a small amount of wear particles and debris was observed in SIM samples (Fig. 1a), and none were observed in NS samples (Fig. 1c).

Piscataway, NJ, USA). β-Actin was used as a reference protein. For densitometric analyses, blots were scanned and quantified using Image-Pro Plus analysis software (version 5.0). The results were expressed as the percentage of βactin immunoreactivity. Statistical Analysis The statistic significance of differences was determined by one-way analysis of variance (equal variances) or Tamhane’s T2 (unequal variances) in multiple groups. P values less than 0.05 were considered indicative of statistically significant differences. All data were analyzed using SPSS 17.0 software (SPSS Inc., Chicago, IL, USA).

RESULTS Morphological and Structural Observations of the Three Tissue Types SEM images of the three tissue types showed an extensive presence of wear particles and debris in AIM

samples (Fig. 1b). In contrast, only a small amount of wear particles and debris were observed in SIM samples (Fig. 1a) and none were observed in NS samples (Fig. 1c). TEM images of the three tissue types showed that Bacillus infection was present in SIM samples (Fig. 2a). In AIM samples, phagocytes with wear particles and vacuoles as well as nuclear fragmentation were present, and the nuclear membrane had disappeared (Fig. 2b). No Bacillus or wear debris was found in the NS samples (Fig. 2c). RANKL, RANK, and OPG Protein Production According to Immunohistochemical Staining Immunohistochemical staining for RANKL, RANK, and OPG proteins showed differential expression of these three proteins in the three different types of tissues (Fig. 3). AIM samples showed strong staining for RANKL, RANK, and OPG proteins, whereas SIM samples showed weak staining for RANKL, RANK, and OPG proteins. In comparison, NS samples showed equivocal or negative staining for RANKL, RANK, and OPG proteins. The

Fig. 2. Transmission electron microscopy observation of the three tissue types. a Septic interface membranes (SIM), b aseptic interface membranes (AIM), and c normal synovium (NS). TEM images showed that Bacillus infection was present in SIM samples (Fig. 2a, arrow). In AIM samples, phagocytes with wear particles and vacuoles as well as nuclear fragmentation were present, and the nuclear membrane had disappeared (Fig. 2b, arrow). No Bacillus or wear debris was found in the NS samples (Fig. 2c).

Alteration of the RANKL/RANK/OPG System in Periprosthetic

Fig. 3. Immunohistochemical staining of RANKL, RANK, and OPG in the three tissue types. RANKL, RANK, and OPG showed a weak staining in septic interface membranes (SIMs), b strong staining in aseptic interface membranes (AIMs), and c no staining in normal synovium (NS) both at low magnification (×100 in panel a) and high magnification (×400 in panel b). AIM samples showed strong staining for RANKL, RANK, and OPG proteins, whereas SIM samples showed weak staining for RANKL, RANK, and OPG proteins. In comparison, NS samples showed equivocal or negative staining for RANKL, RANK, and OPG proteins.

quantitative data for the IOD values and the corresponding statistical analysis showed that protein expression of RANKL, RANK, and OPG in AIM samples was significantly higher than in SIM and NS samples (P < 0.05).

RANKL and RANK protein expression levels in SIM samples were significantly higher than those in NS samples (P<0.01), but OPG protein expression did not differ significantly between these two tissues (P=0.065; Fig. 4).

Wang, Dai, Xie, Liao, Lv, and Hu

Fig. 4. The IOD values for RANKL, RANK, and OPG determined by Image-Pro Plus according to the staining intensity for each group.

RANKL, RANK, and OPG mRNA Expression According to RT-qPCR The RT-qPCR results revealed that the relative mRNA expression levels of RANKL (21.02 ± 11.37), RANK (10.45±5.01), and OPG (18.23±9.10) were significantly higher in AIM samples than in SIM samples (RANKL 9.37 ± 4.70, RANK 5.23 ± 2.74, OPG 11.81 ±6.27) and in NS samples (RANKL 5.51±2.82, RANK 3.17±2.02, OPG 10.54±4.66; P<0.01). The expression of RANKL mRNA in the SIM group was also higher than that in the NS group (P<0.05), whereas the expression of RANK and OPG mRNA in these two tissues did not differ significantly (P=0.655; Fig. 5). According to the quantitative results, the RANKL/OPG ratios for both the AIM group (1.21 ± 0.44, P < 0.01) and the SIM group (0.88 ±0.49, P<0.05) were higher than that of the NS group

(0.55 ± 0.26). In addition, the RANKL/OPG ratio was higher in the AIM group than in the SIM group (P<0.05; Fig. 6). RANKL, RANK, and OPG Protein Expression According to Western Blot Analysis RANKL, RANK, and OPG protein expressions were assessed by Western blot analysis (Fig. 7). According to the quantitative results, the relative expression levels of RANKL (1.47 ± 0.56) and RANK (0.98 ± 0.42) protein were higher in AIM tissues than in SIM tissues (RANKL 1.07±0.44, RANK 0.68±0.33; P<0.05) and in NS tissues (RANKL 0.66±0.27, RANK 0.54±0.23; P<0.01). The expression of RANKL protein in the SIM group was higher than in the NS group (P<0.05), whereas the expression of RANK protein in these two tissues did not differ

Fig. 5. Quantitative results of RT-qPCR analysis for the expression of RANKL, RANK, and OPG mRNA (RQ value) in SIM, AIM, and NS samples.

Alteration of the RANKL/RANK/OPG System in Periprosthetic

Fig. 6. Comparison of the RANKL/OPG ratios for SIM, AIM, and NS samples according to the quantitative results of RT-qPCR analysis.

significantly (P>0.05). Furthermore, OPG protein expression was not significantly different among the three tissue types (SIM 0.93±0.40, AIM 1.05±0.36, NS 0.84±0.28; P>0.05; Fig. 8). According to the quantitative results, the RANKL/OPG ratios for both the AIM group (1.50±0.71, P<0.05) and the SIM group (1.42±0.83, P<0.05) were higher than that of the NS group (0.88±0.48). However, the RANKL/OPG ratios were not significantly different between the AIM and SIM groups (P>0.05; Fig. 9).

DISCUSSION The RANKL/OPG ratio is considered a key parameter in the regulation of bone resorption and is associated with a variety of bone metabolism abnormalities and bone disorders [24]. In previous studies, wear debris has been

shown to increase the RANKL/OPG ratio in murine calvarial tissues and cultured osteoblasts and fibroblasts [11–13, 25]. Additional studies have shown that RANKL expression is upregulated and the RANKL/OPG ratio is elevated in the interface membrane or synovial fluid of patients with osteolysis related to aseptic loosening, in comparison to levels in NS or synovium samples of patients with osteoarthritis [8, 10, 26]. In good agreement, our results also showed that RANKL expression and the RANKL/OPG ratio were significantly increased in AIM samples from patient with periprosthetic osteolysis, in comparison to those in NS samples. Because imbalanced expression of the members of the RANKL/RANK/OPG system is closely related to periprosthetic osteolysis with aseptic loosening, gene therapy may be a potential therapeutic strategy. Previous studies have supported this theory by showing that RANKL blockade with OPG therapy prevents wear

Fig. 7. Expression of RANKL, RANK, and OPG protein in a septic interface membrane (SIM), b aseptic interface membrane (AIM), and c normal synovium (NS) samples by Western blot.

Wang, Dai, Xie, Liao, Lv, and Hu

Fig. 8. Quantitative results of Western blot analysis for the expression of RANKL, RANK, and OPG mRNA in SIM, AIM, and NS samples.

debris-induced osteolysis in a murine calvarial model [27, 28]. However, the quantitative results for our immunohistochemical and RT-qPCR analyses showed that OPG expression also was increased in AIM samples, which is a different finding from those reported by previous studies [8, 10]. Our results seem contradictory to the role of OPG, which can downregulate osteoclastogenesis by binding RANKL and thereby inhibit bone loss, and this is worthy of our attention. We speculate that the increase in OPG expression is probably a compensatory reaction to the high expression of RANKL in interface membranes. Another possibility is that macrophages activated by wear debris can release a variety of pro-inflammatory cytokines. Highlevel expression of these cytokines can induce osteolysis

by activating the differentiation and maturation of OCs through other signal pathways, which could result in a compensatory reactive increase in local OPG expression. However, the RANKL expression in AIM samples was upregulated more obviously, which led to a significant increase in the RANKL/OPG ratio in interface membranes compared to that in NS. The elevated RANKL/OPG ratio may induce generation, activation, and maturation of OCs as well as promote the occurrence of periprosthetic osteolysis. Most patients with PJI after THA also experience periprosthetic osteolysis. The pathogenesis of osteolysis with PJI was not known until now. Abundant SIMs can be harvested during rTHA procedures and are regarded as the best tissue specimens for diagnosing PJI histologically

Fig. 9. Comparison of the RANKL/OPG ratios for SIM, AIM, and NS samples according to the quantitative results of Western blot analysis.

Alteration of the RANKL/RANK/OPG System in Periprosthetic [29, 30]. However, whether the components of the RANKL/RANK/OPG system are expressed in SIM tissues at all or at differential levels compared to those in AIM and NS samples and whether such imbalanced expression is correlated with osteolysis induced by bacterium remain to be determined. SIM tissues from PJI patients were evaluated via immunohistochemistry, RT-qPCR, and Western blot analysis in this study. The results showed RANKL expression was elevated in SIM samples, whereas OPG expression was not elevated compared to that in NS samples, which led to an increase in the RANKL/OPG ratio in SIM samples. Compared to that in AIM samples, RANKL and RANK expression at both the mRNA and protein levels were lower in SIM samples, and thus, the RANKL/OPG ratio was also significantly lower in the SIM group. Overall, these results suggest that the imbalanced expression of the RANKL/RANK/OPG system components is related to periprosthetic osteolysis with septic loosening, but it is not the only, or even the main, pathogenic mechanism. Recent studies also showed that toll-like receptors (TLRs) are highly expressed in SIM samples, specifically TLR2 and TLR5 [4, 31]. Thus, further research is needed to determine the associations between TLRs and periprosthetic osteolysis. This study has several limitations. First, due to the limited ability to diagnose PJI after THA surgery, a few cases with mild infection were routinely examined and found to be negative or not meeting the diagnostic criteria but might have been diagnosed as aseptic loosening. Therefore, it is questionable whether AIM tissues are completely sterile. In addition, wear debris can induce the imbalanced expression of components of the RANKL/ RANK/OPG system. Although only a few wear particles were contained in SIM samples, we cannot completely exclude the influence on the result of this study. Despite these limitations, all of our experimental results taken together indicate that the imbalanced expression of the RANKL/RANK/OPG system is related to periprosthetic osteolysis with septic loosening. However, it was not the only or even the main pathogenesis.

CONCLUSION In conclusion, RANKL expression and the RANKL/ OPG ratio were significantly increased in SIMs. Imbalance in the RANKL/RANK/OPG system is related to periprosthetic osteolysis with septic loosening, but is not the only possible pathogenic mechanism.

ACKNOWLEDGMENTS This work was funded in part by the Natural Science Foundation of Hunan Province, China (Grant No. 12JJ2055), the Hunan Provincial Innovation Foundation for Postgraduates (Grant No. 2012B091), and the Science and Technology Planning Project of Hunan Province (Grant No. 2011FJ6085).

Competing Interests. The authors declare that they have no competing interests. Authors’ Contributions. LW and YH conceived of the study and drafted the manuscript. LW, ZD, JX, HL, and CL prepared experimental samples. LW, ZD, HL, and CL carried out the experiments. LW and JX performed the statistical analysis. LW, YH, ZD, and JX participated in study design and helped to draft the manuscript. All authors read and approved the final manuscript.

REFERENCES 1. Learmonth, I.D., C. Young, and C. Rorabeck. 2007. The operation of the century: Total hip replacement. Lancet 370(9597): 1508–1519. 2. Parvizi, J., B. Adeli, B. Zmistowski, C. Restrepo, and A.S. Greenwald. 2012. Management of periprosthetic joint infection: the current knowledge: AAOS exhibit selection. Journal of Bone and Joint Surgery (American) 94(14): e104. 3. Goodman, S.B., E. Gibon, and Z. Yao. 2013. The basic science of periprosthetic osteolysis. Instructional Course Lectures 62: 201–206. 4. Pajarinen, J., E. Cenni, L. Savarino, E. Gomez-Barrena, Y. Tamaki, M. Takagi, J. Salo, and Y.T. Konttinen. 2010. Profile of toll-like receptorpositive cells in septic and aseptic loosening of total hip arthroplasty implants. Journal of Biomedical Materials Research. Part A 94(1): 84–92. 5. Maoz G, Phillips M, Bosco J, Slover J, Stachel A, Inneh I, and Iorio R. 2014. The Otto Aufranc award: Modifiable versus nonmodifiable risk factors for infection after hip arthroplasty. Clinical Orthopaedics and Related Research. 6. Meehan, J.P., B. Danielsen, S.H. Kim, A.A. Jamali, and R.H. White. 2014. Younger age is associated with a higher risk of early periprosthetic joint infection and aseptic mechanical failure after total knee arthroplasty. Journal of Bone and Joint Surgery (American) 96(7): 529–535. 7. Tsung JD, Rohrsheim JA, Whitehouse SL, Wilson MJ, and Howell JR. 2014. Management of periprosthetic joint infection after total hip arthroplasty using a custom made articulating spacer (CUMARS); the Exeter experience. The Journal of Arthroplasty. 8. Mandelin, J., T.F. Li, M. Liljestrom, M.E. Kroon, R. Hanemaaijer, S. Santavirta, and Y.T. Konttinen. 2003. Imbalance of RANKL/RANK/ OPG system in interface tissue in loosening of total hip replacement. Journal of Bone and Joint Surgery (British) 85(8): 1196–1201. 9. Chen, D., X. Zhang, Y. Guo, S. Shi, X. Mao, X. Pan, and T. Cheng. 2012. MMP-9 inhibition suppresses wear debris-induced

Wang, Dai, Xie, Liao, Lv, and Hu inflammatory osteolysis through downregulation of RANK/RANKL in a murine osteolysis model. International Journal of Molecular Medicine 30(6): 1417–1423. 10. Kim, K.J., S. Kotake, N. Udagawa, H. Ida, M. Ishii, I. Takei, T. Kubo, and M. Takagi. 2001. Osteoprotegerin inhibits in vitro mouse osteoclast formation induced by joint fluid from failed total hip arthroplasty. Journal of Biomedical Materials Research 58(4): 393–400. 11. Masui, T., S. Sakano, Y. Hasegawa, H. Warashina, and N. Ishiguro. 2005. Expression of inflammatory cytokines, RANKL and OPG induced by titanium, cobalt-chromium and polyethylene particles. Biomaterials 26(14): 1695–1702. 12. Wei, X., X. Zhang, M.J. Zuscik, M.H. Drissi, E.M. Schwarz, and R.J. O’Keefe. 2005. Fibroblasts express RANKL and support osteoclastogenesis in a COX-2-dependent manner after stimulation with titanium particles. Journal of Bone and Mineral Research 20(7): 1136–1148. 13. Pioletti, D.P., and A. Kottelat. 2004. The influence of wear particles in the expression of osteoclastogenesis factors by osteoblasts. Biomaterials 25(27): 5803–5808. 14. Lacey, D.L., E. Timms, H.L. Tan, M.J. Kelley, C.R. Dunstan, T. Burgess, R. Elliott, A. Colombero, G. Elliott, S. Scully, et al. 1998. Osteoprotegerin ligand is a cytokine that regulates osteoclast differentiation and activation. Cell 93(2): 165–176. 15. Yasuda, H., N. Shima, N. Nakagawa, K. Yamaguchi, M. Kinosaki, S. Mochizuki, A. Tomoyasu, K. Yano, M. Goto, A. Murakami, et al. 1998. Osteoclast differentiation factor is a ligand for osteoprotegerin/ osteoclastogenesis-inhibitory factor and is identical to TRANCE/ RANKL. Proceedings of the National Academy of Sciences of the United States of America 95(7): 3597–3602. 16. Hsu, H., D.L. Lacey, C.R. Dunstan, I. Solovyev, A. Colombero, E. Timms, H.L. Tan, G. Elliott, M.J. Kelley, I. Sarosi, et al. 1999. Tumor necrosis factor receptor family member RANK mediates osteoclast differentiation and activation induced by osteoprotegerin ligand. Proceedings of the National Academy of Sciences of the United States of America 96(7): 3540–3545. 17. Quinn, J.M., J. Elliott, M.T. Gillespie, and T.J. Martin. 1998. A combination of osteoclast differentiation factor and macrophagecolony stimulating factor is sufficient for both human and mouse osteoclast formation in vitro. Endocrinology 139(10): 4424–4427. 18. Simonet, W.S., D.L. Lacey, C.R. Dunstan, M. Kelley, M.S. Chang, R. Luthy, H.Q. Nguyen, S. Wooden, L. Bennett, T. Boone, et al. 1997. Osteoprotegerin: A novel secreted protein involved in the regulation of bone density. Cell 89(2): 309–319. 19. Hofbauer, L.C., C.A. Kuhne, and V. Viereck. 2004. The OPG/ RANKL/RANK system in metabolic bone diseases. Journal of Musculoskeletal & Neuronal Interactions 4(3): 268–275. 20. Achermann, Y., M. Vogt, M. Leunig, J. Wust, and A. Trampuz. 2010. Improved diagnosis of periprosthetic joint infection by multiplex PCR

of sonication fluid from removed implants. Journal of Clinical Microbiology 48(4): 1208–1214. 21. Norton, A.J., S. Jordan, and P. Yeomans. 1994. Brief, hightemperature heat denaturation (pressure cooking): A simple and effective method of antigen retrieval for routinely processed tissues. Journal of Pathology 173(4): 371–379. 22. Xavier, L.L., G.G. Viola, A.C. Ferraz, C. Da Cunha, J.M. Deonizio, C.A. Netto, and M. Achaval. 2005. A simple and fast densitometric method for the analysis of tyrosine hydroxylase immunoreactivity in the substantia nigra pars compacta and in the ventral tegmental area. Brain Research. Brain Research Protocols 16(1–3): 58–64. 23. Wang, C.J., Z.G. Zhou, A. Holmqvist, H. Zhang, Y. Li, G. Adell, and X.F. Sun. 2009. Survivin expression quantified by Image-Pro Plus compared with visual assessment. Applied Immunohistochemistry & Molecular Morphology 17(6): 530–535. 24. Hofbauer, L.C., and M. Schoppet. 2004. Clinical implications of the osteoprotegerin/RANKL/RANK system for bone and vascular diseases. JAMA 292(4): 490–495. 25. Xu, J., M.D. Kauther, J. Hartl, and C. Wedemeyer. 2010. Effects of alpha-calcitonin gene-related peptide on osteoprotegerin and receptor activator of nuclear factor-kappaB ligand expression in MG-63 osteoblast-like cells exposed to polyethylene particles. Journal of Orthopaedic Surgery and Research 5: 83. 26. Horiki, M., T. Nakase, A. Myoui, N. Sugano, T. Nishii, T. Tomita, T. Miyaji, and H. Yoshikawa. 2004. Localization of RANKL in osteolytic tissue around a loosened joint prosthesis. Journal of Bone and Mineral Metabolism 22(4): 346–351. 27. Zhang, L., T.H. Jia, A.C. Chong, L. Bai, H. Yu, W. Gong, P.H. Wooley, and S.Y. Yang. 2010. Cell-based osteoprotegerin therapy for debris-induced aseptic prosthetic loosening on a murine model. Gene Therapy 17(10): 1262–1269. 28. Zhang, T., H. Yu, W. Gong, L. Zhang, T. Jia, P.H. Wooley, and S.Y. Yang. 2009. The effect of osteoprotegerin gene modification on wear debris-induced osteolysis in a murine model of knee prosthesis failure. Biomaterials 30(30): 6102–6108. 29. Bori, G., E. Munoz-Mahamud, S. Garcia, C. Mallofre, X. Gallart, J. Bosch, E. Garcia, J. Riba, J. Mensa, and A. Soriano. 2011. Interface membrane is the best sample for histological study to diagnose prosthetic joint infection. Modern Pathology 24(4): 579–584. 30. Tohtz, S.W., M. Muller, L. Morawietz, T. Winkler, and C. Perka. 2010. Validity of frozen sections for analysis of periprosthetic loosening membranes. Clinical Orthopaedics and Related Research 468(3): 762–768. 31. Tamaki, Y., Y. Takakubo, K. Goto, T. Hirayama, K. Sasaki, Y.T. Konttinen, S.B. Goodman, and M. Takagi. 2009. Increased expression of toll-like receptors in aseptic loose periprosthetic tissues and septic synovial membranes around total hip implants. Journal of Rheumatology 36(3): 598–608.