Periprosthetic joint infection (PJI) is a potentially devastating complication of joint replacement surgery. Osteocytes comprise 90–95% of all cells in hard bone tissue, are long-lived and are becoming increasingly recognised as a critical cell type in the regulation of bone and systemic physiology. The purpose of this study was to examine role of these cells in PJI pathophysiology and aetiology, with the rationale that their involvement could contribute to the difficulty in detecting and clearing PJI. This study examined the ability of human osteocytes to become infected by Staphylococcus aureus and the responses of both the host cell and pathogen in this scenario.

Several S. aureus (MRSA) strains were tested for their ability to infect human primary osteocyte-like cells in vitro and human bone samples ex vivo. Bone biopsies were retrieved from patients undergoing revision total hip arthroplasty for either aseptic loosening associated with osteolysis, or for PJI. Retrieved bacterial colony number from cell lysates and colony morphology were determined. Gene expression was measured by microarray/bioinformatics analysis and/or real-time RT-PCR.

Exposure to planktonic S. aureus (approx. 100 CFU/cell) resulted in intracellular infection of human osteocyte-like cells. We found no evidence of increased rates of osteocyte cell death in bacteria exposed cultures. Microarray analysis of osteocyte gene expression 24h following exposure revealed more than 1,500 differentially expressed genes (fold-change more than 2, false discovery rate p < 0.01). The gene expression patterns were consistent with a strong innate immune response and altered functionality of the osteocytes. Consistent patterns of host gene expression were observed between experimentally infected osteocyte-like cultures and human bone, and in PJI patient bone samples. Internalised bacteria switched to the quasi-dormant small colony variant (SCV) form over a period of 5d, and the ensuing infection appeared to reach a stable state. S. aureus infection of viable osteocytes was also identified in bone taken from PJI patients.

We have demonstrated [1] that human osteocytes can become infected by S. aureus and respond robustly by producing immune mediators. The bony location of the infected osteocyte may render them refractory to clearance by immune cells, and osteocytes may therefore be an immune-privileged cell type. The phenotypic switch of S. aureus to SCV, a form less sensitive to most antibiotics and one associated with intracellular survival, suggests that infection of osteocytes may contribute to a chronic disease state. The osteocyte may therefore serve as a reservoir of bacteria for reinfection, perhaps explaining the high prevalence of infections that only become apparent after long periods of time or recur following surgical/medical treatment. Our findings also provide a biological rationale for the recognised need for aggressive bone debridement in the surgical management of PJI.

Introduction

Sclerostin has been implicated in mechanotransduction in bone and recent data show a lack of response to loading in the sclerostin transgenic mouse. Sclerostin, the protein product of the SOST gene, is an attractive therapeutic target for low bone mass conditions, including osteoporosis. It is expressed exclusively by mature osteocytes in bone and we have shown that sclerostin targets pre-osteocytes/osteocytes to regulate bone mineralization and osteoclast activity, as well as inducing catabolic gene expression in osteocytes themselves and promoting osteocyte-mediated bone loss (osteocytic osteolysis). The aim of this study was to examine the direct effects of sclerostin on anabolic responses to loading in bone ex vivo.

Osteocytes (OCY) are the end stage differentiation cells of the osteoblast lineage, and are incorporated in the bone matrix during bone formation. In doing so, OCY control the mineralisation of osteoid. OCY form a dense inter-connected network of cell bodies and cell processes throughout the mineralised matrix of bone.

OCY viability depends on interstitial fluid flow along the OCY canaliculi, driven by pulsatile blood flow and loading of the skeleton. Maintenance of the density and viability of OCY are essential for bone health because OCY perform many important functions in bone. Firstly, OCY appear to initiate bone repair of bone microdamage. Secondly, OCY are almost certainly the cells, which initiate new bone formation in response to increased loading of bone. Thirdly, OCY are able to regulate the amount of new bone formation in bone remodelling cycles, at least in part by the production of a molecule called sclerostin (SCL).

Mutations in the SCL gene, or deletion of the SCL gene in transgenic mice, are associated with particularly dense, fracture resistant bones. This information has led to development of anti-SCL antibodies as a potential anabolic therapy for bones. Bone loss in ovariectomised aged rats was shown recently to be reversed by treatment with neutralising SCL antibodies. There is also some data to suggest that these antibodies may promote fracture healing. Reduced OCY viability and/or density have been reported in association with osteoporotic fracture. OCY viability seems to be dependent on skeletal loading, adequate skeletal blood flow and estrogen in females. OCY viability is adversely affected by hypoxia, unloading of the skeleton and pharmacobiology, such as chronic exposure to glucocorticoids. Both micro and macro-fractures result in disruption of the OCY network, as do procedures such as drilling and cutting of bone.

Because of the important roles of OCY in bone, new approaches to bone health may require the identification of agents to protect these cells from harmful influences in disease and ageing.

The aim of this study was to examine the progression of osteolytic lesions following liner exchange surgery and relate this to the size of the lesion prior to surgery, and whether the defect underwent curettage and bone grafting during surgery.

Six patients with well-fixed Harris-Galante-1 acetabular components underwent liner exchange surgery for excessive polyethylene wear and osteolysis. The mean interval from primary arthroplasty to revision was 14 years (range 11–17 years). All patients underwent a CT scan pre-operatively to identify the location and size of the osteolytic lesions and during surgery, accessible lesions were curetted and bone grafted. One patient had recurrent dislocations and the acetabular component was revised one year following liner exchange surgery. The remaining five patients had CT scans taken at a mean of five months (range 3–5 months) and 5 years (range 3.4–8.2 years) following surgery. Osteolytic lesion volume with or without bone grafting was measured.

Of the 19 osteolytic lesions detected pre-operatively, the first post-operative CT scan showed that four lesions were fully bone-grafted, ten lesions were partially bone-grafted and five lesions had no bone grafting during surgery. At a minimum of three years following surgery, all fully bone-grafted lesions remained full of bone- graft. Of the ten partially bone-grafted lesions, the osteolytic non-grafted zone decreased in volume in five lesions and five lesions remained unchanged. Of the five osteolytic lesions with no bone grafting, one lesion increased in volume, one lesion decreased in volume and three lesions remained unchanged. No new lesions were detected in any of the hips.

These preliminary results suggest that liner exchange surgery is effective in treating periacetabular osteolysis. Although bone grafting appears to aid in restoring bone stock, it is not essential in halting the progression of osteolysis, which likely results from the ongoing production of polyethylene particles in the joint.

Sensitive and accurate measures of osteolysis around TKR are needed to enhance clinical management and assist in planning revision surgery. Therefore, our aim was to examine, in a cadaver model of osteolysis around TKR, the sensitivity of detection and the accuracy of measuring osteolysis using Xray, CT and MRI.

Fifty-four simulated osteolytic lesions were created around six cadaver knees implanted with either a cemented or cementless TKR. Twenty-four lesions were created in the femur and thirty in the tibia ranging in size from 0.7 cm3 to 14 cm3. Standard anteroposterior and lateral fluoroscopically guided radiographs, CT and MRI scans with metal reduction protocols were taken of the knees prior to the creation of lesions and at every stage as the lesion sizes were enlarged. The location, number and size of the lesions from images obtained by each method were recorded.

The sensitivity of osteolytic lesion detection was 44% for plain radiographs, 92% for CT and 94% for MRI. On plain radiographs, 54% of lesions in the femur and 37% of lesions in the tibia were detected. None of the six posterior lesions created in the tibia were detected on the AP radiographs; however, three of these six lesions were detected on the lateral radiographs. CT was able to detect lesions of all sizes, except for four lesions in the posterior tibia (mean volume of 1.2 cm3, range 1.06–1.47 cm3). Likewise, MRI was very sensitive in detecting lesions of all sizes, with the exception of three lesions, two of which were in the femur and one was in the medial condyle of the tibia (mean volume of 1.9 cm3, range 1.09–3.14 cm3). Notably, all six posterior tibial lesions, which could not be detected using AP radiographs, were detected by MRI.

This study demonstrates the high sensitivity of both CT and MRI (which uses no ionising radiation) to detect simulated knee osteolysis and can therefore be used to detect and monitor progression of osteolysis around TKR. The study also shows the limitations of plain radiographs to assess osteolysis.

While computed tomography (CT) provides an accurate measure of osteolysis volume, it would be advantageous in general clinical practice if plain radiographs could be used to monitor osteolysis. This study determined the ability of plain radiographs to detect the presence of and determine the progression in size of osteolytic lesions around cementless acetabular components.

Nineteen acetabular components were diagnosed with osteolysis using a high-resolution multi-slice CT scanner with metal artefact suppression. Mean duration since arthroplasty was 14 years (range 10–15 years) at initial CT. Repeat CT scans were undertaken over a five year period to determine osteolysis progression. On anteroposterior pelvis (AP) radiographs and oblique radiographs of the acetabulum seen on the rolled lateral hip view, which were taken at the same time as the CT scans, area of osteolysis was measured manually correcting for magnification.

Osteolysis was detected on the AP radiographs in 8 of 19 hips (42%), on the oblique radiographs in 6 of 19 hips (32%) and on the combined AP and oblique radiographs in 8 of 19 hips (42%). Throughout the study period, osteolysis was detected on 31 of 76 AP radiographs (41%) and 22 of 75 oblique radiographs (29%). Osteolysis was more likely to be detected on plain radiographs if the lesion volume was greater than 10cm3 in size (p=0.005). On CT, osteolysis progressed by more than 1cm3/yr in 10 of 19 hips (55%). In these ten hips, osteolysis progression was detected on AP radiographs in six hips and on oblique radiographs in three hips. No correlation was found between osteolysis progression measured by CT and that measured on AP (r2=0.16, p=0.37) or oblique (r2=0.37, p=0.15) or AP and oblique radiographs (r2=0.34, p=0.17).

Plain radiographs are poor in monitoring progression in size of periacetabular osteolytic lesions. Plain radiographs may detect lesions more than 10cm3 in size, but are unreliable.

Computed tomography (CT) provides a sensitive and accurate measure of periacetabular osteolytic lesion volume, however, there may remain a role for plain radiographs in monitoring osteolysis. This study aimed to compare CT and plain radiographs for determining the progression in size of osteolytic lesions around cementless acetabular components.

A high-resolution multi-slice CT scanner with metal artefact suppression was used to determine the volume and progression of osteolysis around 19 cementless Harris Galante-1 and PCA acetabular components. The mean duration since arthroplasty was 14 years (range 10–15 years) at initial CT. Repeat scans of the hip were undertaken over a five year period to determine the progression in size of osteolytic lesions over time. A second blinded observer manually measured the area of osteolytic lesions off anteroposterior pelvis radiographs and oblique radiographs of the acetabulum that were taken at the same time as the CT scan.

All 19 hips had CT detected osteolysis. Osteolysis was detected on one or both of the anteroposterior pelvis or oblique radiographs from at least one time point in eight of 19 hips (42%). Osteolysis was detected on 31 of 76 anteroposterior pelvis radiographs (41%) and on 22 of 75 oblique radiographs (29%) (p=0.140). Osteolysis was more likely to be detected on plain radiographs if the lesion volume was greater than 10cm3 in size compared to those 5–10cm3 and less than 5cm3 in size (p=0.009). In 10 of 19 hips (55%), CT determined that osteolytic lesions progressed in size by more than 1cm3/yr. The mean volume of osteolysis progression was 3.2cm3/yr (range 1.1–7.5cm3/yr). Progression in size of osteolytic lesions was significantly associated with hips with larger osteolytic lesions at the initial CT (p=0.0004). Radiographic measurements detected progression of osteolytic lesions in 5 of the 10 hips (50%) that progressed. No correlation was found between progression in size of osteolytic lesions as measured by CT and progression in size of osteolytic lesions as measured off the anteroposterior pelvis (r2 = 0.16, p=0.37), oblique (r2=0.37, p=0.15) and combined anteroposterior pelvis and oblique radiographs (r2=0.34, p=0.17).

Periacetabular osteolytic lesions are more likely to be detected on plain radiographs if they are more than 10cm3 in size. Plain radiographs may therefore provide some monitoring value as lesions more than 10cm3 are more likely to be progressive. However, plain radiographs should not be relied upon to monitor the progression of these lesions.

Periprosthetic osteolysis is a serious medium to long-term complication of total hip arthroplasty. Interobserver reliability of detecting osteolysis around cementless ace-tabular components is reported to be poor using plain radiographs. Quantitative computed tomography (CT) provides sensitive and accurate measures of osteolytic lesion volume, however, there may remain a role for plain radiographs in monitoring progression of osteolysis. The aim of this study was to use quantitative CT to monitor the progression of osteolytic lesions around cementless acetabular components and to compare plain radiographs and CT in determining the progression of osteolysis.

A high-resolution multi-slice quantitative CT scanner with metal artefact suppression was used to determine the volume of osteolysis around 18 cementless acetabular components. The mean time since arthroplasty was 14 years (range 10–15 years) at the initial CT. Repeat scans of the hip were undertaken over a five-year period to determine progression of osteolysis with time. A second blinded observer examined anteroposterior and lateral plain radiographs taken at the same time as the CT scans and measured the location and area of osteolytic lesions.

CT measurements determined that in ten of 18 hips (56%), osteolytic lesions progressed by more than 1cm3/yr. Progression in size of osteolytic lesions was significantly associated with hips with larger osteolytic lesions at the initial CT (p=0.0005). The mean volume of osteolysis progression was 4.9cm3/year (range 2.8–7.5cm3/yr) for cases with osteolysis volumes greater than 10cm3 at the initial CT, and 0.7cm3/yr (range 0–2.3cm3/yr) for cases with osteolysis volumes smaller than or equal to 10cm3 at the initial CT (p=0.002). Importantly, the rate of osteolysis progression between CT scans varied greatly in some hips. In contrast, using plain radiograph assessment, progression in the area of osteolytic lesions was only detected in 10% of hips.

In conclusion, quantitative CT provides new insights into the natural history of periacetabular osteolysis. Total osteolysis volume greater than 10cm3 is associated with a high risk of progression and progress, on average, at a greater rate than those less than 10cm3. Plain radiographs, including a lateral view, are an unreliable clinical diagnostic tool to predict substantial progression of periacetabular osteolytic lesions.

Osteoclast Associated Receptor (OSCAR) is a novel member of leucocyte receptor complex (LCR)-encoded family expressed by pre-osteoclasts and mature osteoclasts (OC). Blocking of OSCAR binding to its putative ligand has been shown to inhibit osteoclast formation. To date there is no data available regarding the expression of OSCAR in tissues associated with osteolysis and the objective of this study is to determine if OSCAR is expressed adjacent to focal bone osteolysis near failed implants. A total of 22 samples (10 Peri-implant tissue and 12 OA) were studied. OSCAR antibodies were a gift from R& D Systems Inc. (Minneapolis, MN, USA). The tissues were analysed semi-qualitatively using semi-quantitative scoring (SQA) independently by two observers. Non-parametric Mann Whitney-U test was used to test statistical significance. Dual labelling for OSCAR and CD68 expression was also carried out. Strong expression of OSCAR was seen in the majority of multinucleated cells in peri-implant tissues while OA tissues showed very low levels of OSCAR expression. Dual labelling studies revealed that the cells expressing OSCAR also expressed CD68. There was a significant difference in the expression of OSCAR between peri-implant tissue and OA synovial tissue (p< 0.003). This study shows that OSCAR is expressed at high levels by the numerous CD68 multinucleated cells present is these tissues in peri-implant tissues. These findings and recent reports on the role OSCAR may play in OC formation indicate that OSCAR could be an important mediator of peri-implant osteolysis

Introduction Sheep are being used increasingly for spinal and other skeletal-related research. However, there is still limited information about the molecular pathways of bone remodelling in this species compared to rats or mice. It has been demonstrated in other animal models and in the human that the receptor activator of nuclear factor kappaB ligand (RANKL) and osteoprotegerin (OPG) play major regulatory roles in controlling osteoclast activity and their differentiation. We investigated the expression of RANKL and OPG in trabecular bone of an ovariectomised steroid-treated osteopaenic sheep model.

Methods Trabecular bone from the lumbar spine (LS) and proximal femur (PF) of ten osteopaenic ewes and four normal ewes were collected [1]. Total RNA was isolated and complementary DNA (cDNA) was synthesised. DNA encoding RANKL and OPG were sequenced and ovine specific primers were designed to amplify the cDNA by real time RT-PCR to generate products corresponding to mRNA encoding RANKL and OPG. The results were normalised to 18S RNA.

Results Total OPG expression (in trabecular bone) from the PF region was over two fold higher than the LS (P< 0.0001). The relative expression of OPG in the both LS and PF regions were significantly higher in the treated animals (steroid & oophorectomy) compared to controls (p< 0.05). The relative expression of RANK-L in the PF was significantly higher than in the LS (P< 0.0001). However, the relative RANK-L expression in the treated animals was not significantly different from the control animals in either region. The ratio of RANK-L:OPG in the PF and the LS was not significantly different but it was significantly reduced in the osteopaenic animals.

Discussion Based on this gene expression study and previous histomorphological data, it appears that trabecular bone loss is not due to increased osteoclastic activity but may rather due to lack of osteoblastic activity and function. Higher expression of OPG and RANK-L and greater bone loss compared to LS suggest that the rate of bone turnover is greater in the PF. Further investigation of the molecular pathways of bone loss in this animal model will increase its utility for osteoporosis research.

Introduction and Aims: The usefulness of bone graft substitutes and growth factors to promote bone graft incorporation and prosthesis fixation in hip replacement should be examined in a loaded model, as results from cortical defect models may not apply. This paper reviews the results of femoral impaction grafting using these materials in an ovine hip replacement model.

Method: At cemented hemiarthroplasty, sheep femurs were impacted with allograft bone (control group n=23) or with allograft mixed with: 1) corglaes bioglass (n=12); 2) a synthetic hydroxyapatite (HA) (n=6) or the bone morpohogenetic protein OP-1 (n=6) (study groups) and implanted with a cemented double taper femoral stem. Sheep were sacrificed at between six and 26 weeks. The primary outcome was femoral stem subsidence, as determined more recently by the development of clinical radiostereometric analysis (RSA) in this model. Femoral fixation, as assessed by ex-vivo mechanical testing, and bone graft incorporation, as assessed by histological review and histoquantitation, were also key outcomes.

Results: In the control groups, there was a consistent response with bone graft incorporation by new bone advancing proximal to distally in the femur and advancing from the endocortex towards the cement mantle. Mineralised bone apposition occurred by six weeks and this was preceeded by partial resorption of the graft. Complete graft incorporation, with subsequent remodelling of bone, was evident proximally by 26 weeks. Bone graft incorporation in femurs impacted with a 1:1 allograft: bioglass mix was minimal and there was often partial or complete resorption of the graft with replacement by fibrous tissue, resorption of endocortical bone and instability of the femoral prosthesis. Supplementation of allograft with OP-1 promotes initial graft resorption, thus hastening bone graft incorporation and remodelling but one case of stem subsidence, that may have been associated with early resorption seen in the OP-1 group, reinforces the need for further studies examining dose response. There was excellent incorporation of the allograft and HA, with new woven bone directly apposing the HA surface and integrated into the larger porous spaces of the HA. There was no adverse response to the HA and there was minimal to no subsidence of the stem at the cement-bone interface, as determined by RSA.

Conclusion: This model is extremely valuable for investigating new biological approaches to reconstruction of major bone deficiency at revision hip replacement and demonstrates clear differences between materials used to supplement allograft, with HA and OP-1 giving encouraging results. RSA is an essential outcomes tool for this model.

Introduction Peri-acetabular osteolysis is a serious complication of total hip arthroplasty (THA). The aim of this study was to determine, using quantitative computed tomography (CT), the location, volume and rate of progression of peri-acetabular osteolytic lesions, and to determine the validity of this CT technique with intra-operative measurements.

Methods High-resolution spiral multislice CT scan (Somatom Volume Zoom, Siemens, Munich, Germany), with metal-artefact suppression protocol, was used to measure the volume of osteolytic lesions around 47 cementless THAs in 36 patients (median age 73 years, duration 14 years, range five to 24 years). In vitro validation was undertaken. CT scans were taken from the top of sacroiliac joint down to two centimetres below the end of the prosthesis. Reconstruction images were analysed by two different observers and progression of osteolysis with time was determined. In some patients, subsequently revised, in-vivo CT measurements were compared to intra-operative measurements. The rate of progression of osteolytic lesions was calculated. The technique was optimised and validated by extensive in-vitro studies, using bovine and human pelves.

Results The incidence of lesions located in each site was: the ilium, 65%; around fixation screws, 20%; in the anterior column, nine percent; and in the medial wall, six percent. Some lesions were shown to be relatively quiescent, while others were aggressively osteolytic. Intra and inter-observer error for the CT measurement technique was four percent and 2.8%, respectively. In vitro volumetric measurements of simulated bone defects adjacent to the acetabular component and fixation screw were accurate to within 96% and precise to 98%. In addition, preliminary data obtained intra-operatively indicate the accuracy of CT in identifying the sites of osteolysis.

Conclusions CT is thus a valid and reliable technique for investigating the natural history of osteolysis and the factors that may influence its progression. It will also enable assessment of non-surgical treatments of osteolysis.

In relation to the conduct of this study, one or more of the authors is in receipt of a research grant from a non-commercial source.

There is growing evidence that RANKL (also known as osteoclast differentiation factor), its receptor RANK and its natural inhibitor osteoprotegerin (OPG) are involved in bone loss in a number of pathologies. The aim of this study was to determine if these factors are expressed in a number of bone loss pathologies and what cell types were producing these factors in the tissues using reverse transcription polymerase chain reaction (RT-PCR), in situ hybridisation and immunostaining techniques. Periarticular tissue was obtained from 15 patients undergoing revision of aseptic loose implants. Rheumatoid joint tissue was obtained from the pannus region of 12 patients diagnosed with rheumatoid arthritis undergoing joint replacement or joint fusion. Inflamed gingival tissue from sites near bone erosion were obtained from 11 patients with periodontal disease. 6 normal periodontal and periarticular tissue from 6 osteoarthritic patients was used as controls.

RANK, RANKL, OPG and M-CSF mRNA were expressed in tissues obtained from all the pathologies. Higher ratio’s of RANKL to OPG were observed in the pathological tissues compared to their respective controls. In revision tissues many multinucleated giant cell containing particles expressed RANK mRNA. The pattern of staining of RANK mRNA was markedly different in the rheumatoid and periodontal tissues. Differences were also seen in the pattern of expression for RANKL using both in situ and immunostaining. Overall our results indicate that although similar osteoclastogenic factors are fundamentally involved in these bone loss pathologies, different cell types may be producing and/or responding to these factors. Identifying fundamental mechanisms such as these may indicate that similar treatments, such as using OPG or related compounds, may be used for a diverse range of bone loss diseases.

TRAIL/Apo2L is a member of the tumour necrosis factor (TNF) family of cytokines that induces death of cancer cells but not normal cells. Its potent apoptotic activity is mediated through its cell surface death domain containing receptors, DR4 and DR5. TRAIL binds also to three “decoy” receptors, DcR1, DcR2 and osteoprotegerin (OPG), which lack functional death domains, and do not induce apoptosis. The aim of this study was to investigate the cytotoxic activity of TRAIL as a single agent or in combination with clinically relevant anti-sarcoma drugs on human soft tissue sarcomas that are traditionally resistant to chemotherapy. Human soft tissue sarcomas known to be resistant to chemotherapy were taken at the time of biopsy and cultured to produce a cell line. This cell line was then tested against TRAIL, standard chemotherapeutic agents (including doxorubicin, cis platinum, etoposide, methotrexate and cyclophosphamide) and in combination. When used alone, TRAIL and/or the standard chemotherapeutic agents produced minimal tumour necrosis and this was mirrored in the clinical results. In combination, however, up to 60% necrosis was seen, with doxorubicin the most effective chemotherapeutic agent used. These results indicate that chemotherapy and TRAIL act synergistically to kill sarcoma cells and potentially opens up a new area of cytotoxic treatment for these difficult malignancies.

The cellular and molecular mechanisms that lead to particular trabecular structures in healthy bone and in skeletal disease, such as osteoarthritis (OA), are poorly understood. Osteoclast differentiation factor (ODF) is a newly described regulator of osteoclast formation and function, whose activity appears to be a balance between interaction with its receptor, RANK, and with an antagonist binding protein, osteoprotegerin (OPG). We have examined the relationship between the expression of ODF, RANK and OPG mRNA, and parameters of bone structure and turnover, in human trabecular bone. Intertrochanteric trabecular bone was sampled from patients with primary hip OA (n=13; median age 66 years) and controls taken at autopsy (n=12; median age 68.5 years), processed for histomorphometric analysis and RNA isolated for RT-PCR analysis of ODF, RANK and OPG mRNA expression. The ratios of ODF/OPG and ODF/RANK mRNA are significantly lower in OA (1.78±0.98; 0.59±0.31) compared to the controls (3.41±1.94, p< 0.02; 2.53±1.5, p< 0.001). This suggests that in OA there is less ODF mRNA available per unit RANK mRNA, and that osteoclast formation may be reduced. Furthermore, eroded bone surface (ES/BS[%]) was significantly lower (p< 0.05) in the OA group (6.37±3.17) compared to controls (9.74±4.53). Stong associations were found between the ratio of ODF/OPG mRNA and bone volume (ODF/OPG vs BV/TV[%], r=−0.67; p0.05) and bone turnover (ODF/OPG vs ES/BS, r=0.93; p< 0.001; ODF/OPG vs osteoid surgace (OS/BS[%], r=0.80; p< 0.001) in controls. In contrast to controls, these relationships were not evident in the OA group, suggesting that bone turnover maybe regulated differently in this disease.

We have used a culture system of human peripheral blood mononuclear cells (PBMC)as a source of osteoclast (OC) precursors and murine stromal cells to define the cytokine environment in which human OC form, and to determine the separate contributions of the stromal and haemopoietic elements. We designed a panel of reverse transcription-polymerase chain reaction (RT-PCR) primers that specifically amplify the respective murine or human mRNA species that correspond to cytokines and their receptors previously shown to promote or inhibit OC formation. Murine ST-2 cells and human PBMC were cocultured for up to 21 days in the presence of 1,25(OH) 2vitD3, dexamethasone and human macrophage-colony stimulating factor (M-CSF). OC formation was monitored by the appearance of cells that were positive for tartrate resistant acid phosphatase and able to form resorption lacunae on slices of dentine. We found that the ST-2 cells in these cultures expressed mRNA encoding a repertoire of many of the reported osteoclastogenic factors, as well as the recently described OC differentiation factor (ODF/RANKL). The stromal cells also expressed mRNA encoding osteoprotegerin (OPG), a potent inhibitor of OC formation. We found that agonists and antagonists of OC formation were expressed by both the stromal cells and the PBMC. RANK, the receptor for ODF/RANKL, was expressed only by the PBMC as were IL-1R2 and c-FMS. We identified three features of the cytokine environment that may be a characteristic of normal OC formation. Firstly, the ratio of mouse ODF:OPG mRNA was found to increase during the cocultures, consistent with a key role for ODF in the promotion by stromal cells of OC formation. Secondly, we found that mRNA encoding IL-1 and IL-17, as well as IL-6 and sIL-6R, were coordinately expressed by the PBMC. Thirdly, analysis of the culture medium showed that the PBMC secreted IL-1, IL-6 and TNF-alpha protein only in coculture with ST-2 cells during the first few days of osteoclast development. Similarly, prostaglandin E2, shown to synergise with ODF during OC development, was secreted only in cocultures. Together, these data show OC develop in a complex cytokine environment and suggest that haemopoietic cells provide signals to stromal cells during OC development. Work is in progress to extend these studies to human PBMC interacting with normal human osteoblasts.

Wear particles are thought to be a major factor causing osteolysis that leads to aseptic loosening. The aim of this study was to investigate the role of primary regulators of osteoclast development, RANKL (also known as osteoclast differentiation factor), its receptor RANK and natural inhibitor osteoprotegerin (OPG) in aseptic loosening. Cells were isolated from periprosthetic tissues taken at revision from more than 30 patients and the expression of these mediators in vivo was assessed using semi-quantitative reverse transcription polymerase chain reaction (RT-PCR). These cells were also cultured on dentine to determine their ability to become mature osteoclasts. In situ hybridisation using DIG labelled riboprobes specific for RANK mRNA was used to identify cells likely to become osteoclasts. We were able to compare revison tissues containing several different types of prosthetic wear particles.

RANKL, RANK and OPG mRNA were found in samples of periprosthetic revision tissues. Cells derived from this tissue developed into mature osteoclasts capable of resorbing dentine. Cells that rapidly formed osteoblasts expressed a fifteen fold higher ratio of RANKL:OPG mRNA. In situ hybridisation showed RANK expression by macrophages and giant cells, many of which contained wear particles. Significantly, cells from tissues containing silastic wear particles expressed higher levels of RANKL relative to OPG and more produced large numbers of osteoclasts in vitro. This study shows that different bio materials in a particulate form may differ in their ability to form osteoclasts and that the relative levels of RANKL and OPG are likely to be important in determining if osteolysis will occur. In the future molecules that inhibit RANKL binding, such as OPG, may be considered for therapy of periprosthetic osteolysis.