s / Osteoarthritis and Cartilage 21 (2013) S9–S62 S12 The expression of Dkk-1 by chondrocytes was high in the non calcified cartilage layers at baseline and in sham-operated knees, but greatly decreased after partial meniscectomy from 4 weeks. Sclerostin and sFRP-3 were expressed only in calcified cartilage and increased with the development of OA. Conclusion: This in vivo study shows that in OA, the canonical Wnt signaling pathway is mainly activated in subchondral bone, forming osteophytes and synovium. Moreover, each Wnt antagonist has a specific expression pattern in OA cartilage. Further studies are needed to assess whether regulating Wnt activity in joint tissues can impact cartilage loss in OA. 7 IMPAIRED FRACTURE HEALING BY HIGHLY SELECTIVE COX-2 INHIBITION M.P. Janssen y, M.M. Caron y, B. van Rietbergen z, T.J. Welting y, P.J. Emans y. yMaastricht Univ. Med. Ctr., Maastricht, Netherlands; z Technical Univ. Eindhoven, Eindhoven, Netherlands Purpose: Non-steroidal anti-inflammatory drugs (NSAIDs) are widely used for pain relief in osteoarthritis, after trauma and are used early treatment of rheumatoid arthritis. More and more cyclooxygenase-2 selective NSAIDs are used because they induce fewer upper gastrointestinal complications. However, despite the positive applications of these drugs there are also downsides to their use. In addition to the previously described impairment of the osteogenic phase of endochondral ossification, we showed that NSAIDs have adverse effects on the in vitro chondrogenic phase of endochondral ossification as well. We therefore hypothesized that highly selective COX-2 inhibition by celecoxib impairs fracture healing not only in the osteogenic phase but also in the chondrogenic phase in an in vivo rabbit model. Methods: Fractures were simulated by the creation of a ‘critical size defect’ in the left ulnas of 12 skeletally immature SPF New Zealand white rabbits. Half of the rabbits received an oral dose of 10mg/kg celecoxib each day from day 0. The control group was placebo treated. The experiment was ended 25 days after initiation of the fractures. After harvesting the ulnas were scanned using a micro-CT scanner (micro-CT80, ScancoMedical, Bruettisellen Switzerland) and plain X-rays were taken. Micro-CT images of the fracture region were used for Finite Element Analysis (FEA) to determine the consolidation of the fractures. Hereafter the ulnas were embedded in PMMA and 50mm thick sections were cut for histological analysis. To determine the effect of celecoxib on skeletal development, the growth plate region of ulnas from the contralateral legwere decalcified, embedded in paraffin and cut into 5mm thick sections. The sections were haematoxylin stained and the total growth plate thickness as well as thickness of the proliferative and hypertrophic zones were determined. PGE2 concentrations in the plasma (at the day of sacrifice) was determined using a specific ELISA. Results: COX-2 inhibition by celecoxib was confirmed by significantly decreased PGE2 levels in plasma of celecoxib-treated animals as compared to control animals. The ulnar fractures in the rabbits which were treated with celecoxib showed substantially less consolidation than the fractures in the rabbits in the control group. On the plain X-rays bigger gaps were observed in the fracture region than in the control group. These findings were confirmed by our micro-CT analysis. Furthermore the FEA applied on the micro-CT datasets showed a significant decrease in compression stiffness of the fracture-region and a decreasing trend in torsional stiffness, bending stiffness for a moment around its x-axis and its y-axis. Histological sections of the ulnar fractures showed a larger cartilaginous tissue fraction in the fracture regions of the celecoxib-treated rabbits as compared to the control group. Analysis of the ulnar growth plates showed a significant decrease in total growth plate thickness in the celecoxibtreated group. When the proliferative and hypertrophic zones were scored separately, no significant difference was found between the proliferative zones, however the hypertrophic zone was significantly thinner in the celecoxib-treated group as compared to the control group. Conclusions: We have shown that systemic COX-2 inhibition by celecoxib impairs fracture healing in an in vivo rabbit model. In additionwe showed that chondrocyte hypertrophy in the ulnar growth plate is inhibited. These findings suggest that the impaired fracture healing might not be a specific consequence of impaired osteogenic phase, but is at least in part a result of an effect of celecoxib on the chondrogenic phase of endochondral ossification. 8 DIRECT ASSESSMENT OF ARTICULAR CARTILAGE AND UNDERLYING SUBCHONDRAL BONE REVEALS A PROGRESSIVE GENE EXPRESSION CHANGE IN HUMAN OSTEOARTHRITIC KNEES C.-H. Chou y,{, C.-H. Lee z,{, L.-S. Lu x,{, I.-W. Song x,{, H.-P. Chuang x,{, S.-Y. Kuo k,{, J.-Y. Wu x,{, Y.-T. Chen y,x,{, V.B. Kraus y,{, C.-C. Wu k,{, M.-T.M. Lee x,{. yDuke Univ., Durham, NC, USA; z Taipei Med. Univ. Hosp., Taipei, Taiwan; x Inst. of BioMed. Sci., Academia Sinica, Taipei, Taiwan; k TriService Gen. Hosp., Natl. Defense Med. Ctr., Taipei, Taiwan; {Ctr. for Genomic Med., RIKEN, Yokohama, Japan Purpose: To evaluate the interaction of articular cartilage (AC) and subchondral bone (SB) through analysis of osteoarthritis (OA)-related genes of site-matched tissue. Methods: We developed a novel method for isolating site-matched overlying AC and underlying SB from three and four regions of interest respectively from the human knee tibial plateau (n1⁄450). For each site, the severity of cartilage changes of OAwere assessed histologically, and the severity of bone abnormalities were assessed by microcomputed tomography. An RNA isolation procedure was optimized that yielded high quality RNA from site-matched AC and SB tibial regions. Q-PCR analysis was performed to evaluate gene expression of 61 OA-associated genes for correlation with cartilage integrity and bone structure parameters. Results: A total of 27 (44%) genes were coordinately up or down regulated in both tissues. The expression levels of 19 genes were statistically significantly correlatedwith the severity of AC degeneration and changes of SB structure; these included: ADAMTS1, ASPN, BMP6, BMPER, CCL2, CCL8, COL5A1, COL6A3, COL7A1, COL16A1, FRZB, GDF10, MMP3, OGN, OMD, POSTN, PTGES, TNFSF11 and WNT1. Conclusions: These results provide a strategy for identifying targets whose modification may have the potential to ameliorate pathological alterations and progression of disease in both AC and SB simultaneously. In addition, this is thefirst study, to ourknowledge, to overcome themajor difficulties related to isolation ofhighqualityRNA fromsite-matched joint tissues. We expect this method to facilitate advances in our understandingof the coordinatedmolecular responses of thewhole joint organ. Acknowledgments: This study was supported by the Academia Sinica Genomic Medicine Multicenter Study (40-05-GMM), the National Research Program for Genomic Medicine, National Science Council, Taiwan (Translational Resource Center for Genomic Medicine, NSC1012325-B-001 -035 and National Center for Genome Medicine, NSC1012319-B-001-001, NSC101-2320-B-001-020-MY3), NIH/NIA Claude D. Pepper OAIC 5P30 AG028716 and P01 AR50245 (VBK), and an OARSI scholarship (to C-H Chou). 9 RNASE MRP IS A NOVEL REGULATOR OF ENDOCHONDRAL OSSIFICATION M.M. Caron y, M. Steinbusch y, K. Reicherter z, S. Mattijssen x, D.A. Surtel y, L.W. van Rhijn y, G.J. Pruijn x, E. Lausch z, B. Zabel z, T.J. Welting y. yDept. of Orthopaedic Surgery, Maastricht Univ. Med. Ctr., Maastricht, Netherlands; zCtr. for Paediatrics and Adolescent Med., Univ. of Freiburg, Freiburg, Germany; xDept. of Biomolecular Chemistry, Radboud Univ. Nijmegen, Nijmegen, Netherlands Purpose: Mutations in the RMRP gene cause cartilage-hair hypoplasia (CHH). CHH patients present with short stature (dwarfism), sparse hair, immunodeficiency and increased risk for malignancies. CHH is the first identified human disease caused by mutations in a small nucleolar noncoding RNA molecule. RMRP RNA is the RNA subunit of the RNase MRP complex, a macromolecular particle consisting of the RMRP RNA and at least 7 different protein subunits (Rpp20, Rpp25, Rpp30, Rpp38, Rpp40, Pop1 and Pop5). CHH-associated mutations in the RMRP gene are thought to reduce its RNase enzymatic activity or to interfere with its expression, however how this results in the development of CHH is unknown. The most prominent phenotypic hallmark of CHH is dwarfism, suggesting that RNase MRP function is involved in development of the growth plate. We therefore hypothesized that RNaseMRP is involved in chondrogenic differentiation. Methods: To investigate the expression of RNase MRP subunits during chondrogenic differentiation in vivo, spatiotemporal expression of