ARTHRITIS & RHEUMATISM Vol. 64, No. 1, January 2012, pp 37–39 DOI 10.1002/art.33330 © 2012, American College of Rheumatology EDITORIAL Changing the Outcome of Osteoarthritis: Still a Challenge for Cyclooxygenase 2 Inhibitors Rik J. U. Lories zymes acting downstream of the COX enzymes in prostaglandin synthesis, were expressed in normal cartilage but appeared to be down-regulated in OA cartilage. In contrast, inducible COX-2 and mPEGS-1 were upregulated in OA cartilage but absent in normal cartilage. Treatment with celecoxib, a specific COX-2 inhibitor currently used in clinical practice for OA symptom relief, did not affect histopathologic severity scores for cartilage damage and osteophyte formation in Fukai and colleagues’ study. Moreover, the outcome in this specific mouse model was not different between wild-type mice and mice in which the Ptgs1 and Ptgs2 genes, coding for COX-1 and COX-2, respectively, were deleted. COX inhibition, and in particular inhibition of COX-2, could exert structural effects on disease progression by interfering with different processes in distinct joint tissues. This includes a direct impact on cartilage or subchondral bone homeostasis, and additional effects on osteophyte formation and on synovial inflammation, dampening the production of cytokines and cartilage matrix–degrading enzymes. In the last few years, there has indeed been accumulating evidence that COX-2 may play a role in the progression of OA. Studies performed in Utrecht, The Netherlands have raised particular interest. In vitro work using OA cartilage explants suggested that not only celecoxib but also aceclofenac, which is considered to preferentially inhibit COX-2 over COX-1, had beneficial effects on proteoglycan synthesis and improved explant proteoglycan content, whereas the nonspecific COX inhibitors indomethacin and naproxen did not (3). However, these results could not be confirmed in the in vivo setting: celecoxib treatment of dogs in the canine groove model of OA did not demonstrate any benefit despite the observations that celecoxib was found in detectable amounts in the synovial fluid and that canine explants benefited from in vitro celecoxib treatment (4). An ex vivo analysis of a 15-week in vivo intervention with celecoxib in patients awaiting knee replacement surgery did suggest that COX-2 inhibition had Making a choice between analgesics such as acetaminophen and nonsteroidal antiinflammatory drugs (NSAIDs) when treating patients with pain caused by osteoarthritis (OA) can be difficult. Current guidelines support the use of analgesics as a first choice, but NSAIDs may be slightly superior for improving pain (1). However, NSAIDs are associated with increased toxicity and side effects. Patients often turn to their physicians only after self-treatment with analgesics has failed. NSAIDs target the cyclooxygenase (COX) enzymes, key enzymes in the synthesis of prostaglandins. Inhibition of COX-1 often results in gastrointestinal adverse effects whereas the cardiovascular safety of COX-2 inhibitors has been heavily debated, in particular for elderly patients who have additional cardiovascular risk factors. Weighing the different options could be either simpler or even more complex if NSAIDs were to be shown to have an effect on the structural progression of disease, and thereby on long-term outcome. The consequences of inhibiting COX enzymes on both cartilage and bone biology have been studied extensively. However, findings indicating potential effects on structural progression of OA have been contradictory, and current concepts remain largely based on in vitro experiments. This is further complicated by the hypothesis that more efficient pain control could result in increased joint use and strain in vivo, thereby accelerating disease progression. In this issue of Arthritis & Rheumatism, Fukai et al report on the lack of chondroprotection by specific COX-2 inhibition in a surgically induced model of OA in mice (2). COX-1, as well as microsomal prostaglandin E synthase 2 (mPGES-2) and cytosolic PGES, two enRik J. U. Lories, MD, PhD: Katholieke Universiteit Leuven, Leuven, Belgium. Address correspondence to Rik J. U. Lories, MD, PhD, Laboratory for Skeletal Development and Joint Disorders, Division of Rheumatology, UZ Leuven, Herestraat 49, B3000 Leuven, Belgium. E-mail: [email protected] Submitted for publication August 22, 2011; accepted August 30, 2011. 37 38 beneficial and specific effects in these patients with end-stage disease (5). Obviously the clinical impact of these findings remains to be further studied in trials with patients in whom inhibition of structural disease progression can still affect the outcome. Other in vivo studies used nuclear magnetic resonance imaging (MRI) of cartilage thickness as an outcome measure. In a cohort of 395 elderly subjects, treatment with selective COX-2 inhibitors appeared to have beneficial effects whereas treatment with nonselective NSAIDs had deleterious effects; however, the number of patients receiving each of these treatments was small (40 and 21, respectively) (6). In contrast, a recent open-label pilot study did not demonstrate a beneficial effect on cartilage, as determined by MRI, after 12 months of celecoxib treatment, compared to historical controls (7). These preliminary data do suggest that there is room for well-defined interventional clinical trials assessing the issue. From the biologic perspective, COX-2 is a highly inducible and regulated enzyme, and is typically found in response to inflammatory signals and cell stress. COX-1, in contrast, is the homeostatic isoform of the enzyme and often disappears when a tissue is challenged. In cartilage, the PGE receptor EP4 is a downstream mediator of the deleterious effects of COX-2 activation and PGE2 synthesis, and could represent a more specific therapeutic target to inhibit disease progression (8). Another notable finding of the study by Fukai et al (2) was the absence of an effect on osteophyte formation. Again, two different explanations can be proposed. Effective treatment of pain with the use of celecoxib in this study may have increased mobility in the treated mice compared to the control group. Exposure to instability may trigger repair-like efforts such as osteophyte formation, in order to stabilize the joint. Since COX-2 inhibition delays fracture healing and is used to prevent heterotopic ossification after prosthesis surgery (9), the lack of effect could also suggest that the driving mechanisms or progenitor cell populations involved are different in the surgically induced OA model than in the above-mentioned situations. This could be of specific clinical interest as systematic treatment with celecoxib has been demonstrated to delay progression of spinal ankylosis in ankylosing spondylitis, a chronic inflammatory disease of the spine and sacroiliac joints (10). Again, inflammation versus instability as the factor inducing new bone formation, as well as differences in target cells (progenitor cells in the entheseal attachment zones versus in the periosteum), can explain some of the LORIES discrepant findings, but direct comparisons in specific models are lacking to date. Taking all of this into account, the report by Fukai and colleagues nicely illustrates both some of the strengths and some of the limitations of using mouse models of human disease in a translational setting for clinicians and basic scientists. Over the last decade new animal models of OA have been developed and optimized. Microsurgical procedures leading to mild or severe instability in the knee joints of mice or rats appear to yield models that resemble the human disease processes much more closely than the chemical or enzymatic models that formerly took center stage. The availability of genetic models with deletion or overexpression of specific genes, at either the organism or the tissue level, has provided the research community an increasing number of tools with which to tackle complex translational research questions. Efforts have also been made to standardize microscopic scoring of disease processes. Nevertheless, when studying genetically modified animals, specific attention should be given to the impact of abnormalities in tissues that may not be the direct focus of the experiment. For instance, many mice in which genes relevant for cartilage biology have been modified may also have a bone phenotype that could affect the biomechanical balance between articular cartilage and subchondral bone. The absence of an effect of celecoxib treatment in the study by Fukai et al and the relevance of this finding to the human disorder could be challenged, as pharmacokinetics and dosages are rarely optimized in rodent models. For example, suboptimal dosing could result in rebound effects, shifting the observations to the opposite direction. The authors tried to partially overcome this issue by also using Ptgs1- and Ptgs2-knockout mice; in those experiments, OA development in the knockout mice did not differ from that in wild-type mice. Of interest, mice with gene deletions for proinflammatory cytokines such as interleukin-1 and tumor necrosis factor ␣ also did not exhibit increased cartilage damage in this model. At first glance the main advantage of rodent models is that they enable straightforward and complete pathologic analysis with direct evaluation of structural damage. However, each of the currently used models has its limitations, and OA progression in all of them is clearly accelerated compared to the slowly developing human disease. Scoring systems for mice may lack detail and discriminatory value, and negative results as described by Fukai et al should always be viewed with caution. In addition, as mentioned above, reduction of EDITORIAL 39 joint pain may lead to increased mobility and thus more strain in the model, an issue that was only partially addressed by Fukai and colleagues. In summary, the findings reported by Fukai et al should stimulate further discussion about the use of NSAIDs, and in particular, specific COX-2 inhibitors, in OA. Emerging well-defined clinical trials and the development of imaging end points are encouraging. Paying close attention to the experience obtained in small animal models is essential and should motivate clinical investigators to carefully define their target populations and protocols. Most importantly, when drawing conclusions from both human and animal studies, investigators should not always aim for broad generalization and should avoid overstating the translatability of the data. The wave of enthusiasm for translational research based on either patient samples or specific animal models to solve clinically relevant questions entails the considerable risk of drifting away from the hard data toward emphasizing potential but unproven effects in patients. A particular effect on structural disease progression may be achieved only in specific patient subsets, putting the emphasis for the future of OA treatment on personalized medicine that takes into account potential adverse effects, the extent and stage of the disease process, and likely also the genetic profile. Indeed, PTGS2 has been identified as a susceptibility gene for knee OA in a genome-wide association study (11). These polymorphisms may have an effect on the tissue in the joint but also on the severity of pain and inflammation or even the efficacy of drugs, highlighting once again the complex impact of COX enzymes on OA. AUTHOR CONTRIBUTIONS Dr. Lories drafted the article, revised it critically for important intellectual content, and approved the final version to be published. REFERENCES 1 Zhang W, Nuki G, Moskowitz RW, Abramson S, Altman RD, Arden NK, et al. OARSI recommendations for the management of hip and knee osteoarthritis: part III: changes in evidence following systematic cumulative update of research published through January 2009. Osteoarthritis Cartilage 2010;18:476–99. 2. Fukai A, Kamekura S, Chikazu D, Nakagawa T, Hirata M, Saito T, et al. Lack of a chondroprotective effect of cyclooxygenase 2 inhibition in a surgically induced model of osteoarthritis in mice. Arthritis Rheum 2012;64:198–203. 3. Mastbergen SC, Jansen NW, Bijlsma JW, Lafeber FP. Differential direct effects of cyclo-oxygenase-1/2 inhibition on proteoglycan turnover of human osteoarthritic cartilage: an in vitro study. Arthritis Res Ther 2006;8:R2. 4. Mastbergen SC, Marijnissen AC, Vianen ME, Zoer B, van Roermund PM, Bijlsma JW, et al. Inhibition of COX-2 by celecoxib in the canine groove model of osteoarthritis. Rheumatology (Oxford) 2006;45:405–13. 5. De Boer TN, Huisman AM, Polak AA, Niehoff AG, van Rinsum AC, Saris D, et al. The chondroprotective effect of selective COX-2 inhibition in osteoarthritis: ex vivo evaluation of human cartilage tissue after in vivo treatment. Osteoarthritis Cartilage 2009;17:482–8. 6. Ding C, Cicuttini F, Jones G. Do NSAIDs affect longitudinal changes in knee cartilage volume and knee cartilage defects in older adults? Am J Med 2009;122:836–42. 7. Raynauld JP, Martel-Pelletier J, Beaulieu A, Bessette L, Morin F, Choquette D, et al. An open-label pilot study evaluating by magnetic resonance imaging the potential for a disease-modifying effect of celecoxib compared to a modelized historical control cohort in the treatment of knee osteoarthritis. Semin Arthritis Rheum 2010;40;185–92. 8. Attur M, Al-Mussawir HE, Patel J, Kitay A, Dave M, Palmer G, et al. Prostaglandin E2 exerts catabolic effects in osteoarthritis cartilage: evidence for signaling via the EP4 receptor [published erratum appears in J Immunol 200915;183:1496]. J Immunol 2008:181;5082–8. 9. O’Connor JP, Lysz T. Celecoxib, NSAIDs and the skeleton [review]. Drugs Today (Barc) 2008;44:693–709. 10. Wanders A, van der Heijde D, Landewe R, Behier JM, Calin A, Olivieri I, et al. Nonsteroidal antiinflammatory drugs reduce radiographic progression in patients with ankylosing spondylitis: a randomized clinical trial. Arthritis Rheum 2005;52:1756–65. 11. Valdes AM, Loughlin J, Timms KM, van Meurs JJ, Southam L, Wilson SG, et al. Genome-wide association scan identifies a prostaglandin-endoperoxide synthase 2 variant involved in risk of knee osteoarthritis. Am J Hum Genet 2008;82:1231–40.