Osteoarthritis and Cartilage 18 (2010) 1448e1453 Association between radiographic hand osteoarthritis and RANKL, OPG and inflammatory markers I. Pantsulaia yz, L. Kalichman x, E. Kobyliansky k * y Department of Biomedicine, Institute of Medical Biotechnology, Ministry of Education and Sciences, Georgia z Department of Microbiology and Immunology, Tbilisi State Medical University, Georgia x Department of Physical Therapy, Recanati School for Community Health Professions, Faculty of Health Sciences, Ben-Gurion University of the Negev, Beer Sheva, Israel k Human Population Biology Research Unit, Department of Anatomy and Anthropology, Sackler Faculty of Medicine, Tel Aviv University, Israel a r t i c l e i n f o s u m m a r y Article history: Received 3 February 2010 Accepted 15 June 2010 Objective: The aim of the study was to evaluate the association between prevalence and severity of radiographic hand osteoarthritis (OA) and serum levels of systemic inflammatory markers in a community-based population sample. Design: A cross-sectional observational study was conducted on a population comprised 1452 Chuvashians (763 males, aged 49.23  17.43; and 689 females, aged 50.37  17.47 years). OA was evaluated in 14 joints of each hand using Kellgren and Lawrence (KeL), joint space narrowing (JSN) and osteophyte (OS) scores. Serum levels of systemic inflammatory and osteoclastogenic cytokines were measured by an enzyme-linked immunosorbent assay (ELISA). Statistical analyses included descriptive statistics, correlation analysis and multiple linear regressions. Results: Monocyte chemotactic protein-1 (MCP-1) and osteoprotegerin (OPG) levels were associated with OA traits, but the statistically significant correlations were weak and/or moderate. In particular, the MCP1 inflammation marker showed a statistically significant association with JSN (b ¼ 0.077, P ¼ 0.022) and OS (b ¼ 0.067, P ¼ 0.024) scores, but not with the number of affected joints (KeL  2). OPG was significantly correlated with the scores as to the number of affected joints (b ¼ 0.063, P ¼ 0.035) and OS (b ¼ 0.077, P ¼ 0.028). No significant associations were found between levels of other inflammatory [interleukin (IL)-6, tumor necrosis factor (TNF)-a, IL-17] and osteoclastogenic [receptor activator for nuclear factor k B ligand (RANKL), macrophage colony-stimulating factor (M-CSF)] cytokines and OA characteristics. Conclusions: This study strengthens the premise that OPG might be a valid biomarker of hand OA. Confirmation of these results in larger cohorts of patients will reinforce our theory that the RANKL/OPG pathway is a suitable target for developing novel agents against OA. Ó 2010 Published by Elsevier Ltd on behalf of Osteoarthritis Research Society International. Keywords: Osteoarthritis Hands Inflammation markers Introduction Osteoarthritis (OA) is the most common form of arthritis and can result in substantial disability in the elderly1. Although OA is generally described as a non-inflammatory disease, various studies have indicated that the presence of intraarticular low-grade inflammation contributes to the development and progress of OA2. Recent in-vivo and in-vitro studies have demonstrated that chondrocytes can produce and/or respond to a number of inflammatory * Address correspondence and reprint requests to: Eugene Kobyliansky, Human Population Biology Research Unit, Department of Anatomy and Anthropology, Sackler Faculty of Medicine, Tel Aviv University, Ramat Aviv 69978, Tel Aviv, Israel. Fax: 972-3-6408287. E-mail address: anatom14@post.atu.ac.il (E. Kobyliansky). cytokines and chemokines present in joint tissues and fluids in OA3e7. Interleukin (IL)-1 and tumor necrosis factor (TNF)-a appear to play predominant roles in the initiation and progression of articular cartilage destruction3e5 by decreasing collagen synthesis and increasing the levels of matrix metalloproteinases (MMPs)8, and of other inflammatory mediators, such as nitric oxide, IL-8, IL-6, and prostaglandin E29. Subsequently, nitric oxide influences the chondrocytes, thus promoting cartilage degradation, including inhibition of collagen and proteoglycan synthesis, MMP activation8, and increased susceptibility to other oxidant injury4,10. The role of IL-6 in inflammation differs from that of TNF. IL-6 levels rise in various inflammatory conditions and bone resorption6,7. Serum concentrations of inflammatory markers such as C reactive protein (CRP) and IL-6 are higher among persons with knee 1063-4584/$ e see front matter Ó 2010 Published by Elsevier Ltd on behalf of Osteoarthritis Research Society International. doi:10.1016/j.joca.2010.06.009 I. Pantsulaia et al. / Osteoarthritis and Cartilage 18 (2010) 1448e1453 or hip OA, compared to a population without OA11. Nevertheless, the role of IL-6 in cartilage pathology remains unclear, as this cytokine can induce the production of tissue inhibitors of metalloperoxidase-1 (TIMP), IL-1 receptor antagonists (IL-1Ra), and soluble TNF receptors, all of which inhibit or block MMP or cytokine activity. Among other cytokines, it has been suggested that IL-17 and IL18 play a role in OA pathophysiology, as they share many properties with IL-112. These cytokines also affect an articular tissue, independent of IL-1b. Both seem to be involved in the early phases of the inflammatory process13,14. Moreover, in addition to the inflammatory cytokines, there is ample evidence of involvement of the chemokine system in the pathogenesis of OA15e17. For example, chemokines were detected in OA synovial fluid and were shown to be expressed in synoviocytes and subchondral bone osteoblasts15,17. The investigators also revealed that human chondrocytes express mRNA of several chemokines, including IL-8, monocyte chemotactic protein-1/chemokine (C-C motif) ligand 2 (MCP-1/CCL2) and release these chemokines at various levels15. Increased production of MCP-1 and Regulated upon Activation, Normal T-cell Expressed, and Secreted (RANTES) in OA chondrocytes has also been reported18. On the other hand, synovial inflammation is not the primary cause of OA, but rather a secondary phenomenon related to multiple factors including matrix degradation19,20. Recent studies have confirmed that the subchondral bone is the site of several dynamic morphological and biochemical changes that appear to be part of the OA process19,20. Factors such as osteoprotegerin (OPG) and receptor activator for nuclear factor k B ligand (RANKL), which constitute specific components capable of influencing the bone remodeling process, have also been found to be expressed and modulated in human OA subchondral bone. Moreover, it was recently demonstrated that during longstanding OA and rheumatoid arthritis, the OPG/RANKL ratio in the synovial fluid is far more elevated in OA than in rheumatoid arthritis21. Thus, it is widely accepted that both chemokines, proinflammatory and osteoclastogenic cytokines play important roles in OA pathogenesis. The levels of many of these mediators probably are a more dominant determinant in net cartilage destruction than absolute levels of destructive mediators such as IL-1. Consequently, we hypothesize that the circulating levels of inflammatory and osteoclastogenic cytokines i.e., MCP-1, IL-17, IL-6, TNF-a, OPG, RANKL, and macrophage colony-stimulating factor (M-CSF) might be independently associated with radiographic hand OA and could be used as biomarkers of disease development or severity. We previously reported on plasma levels of cytokines and the influence of genetic and environmental factors on circulating levels of IL-6, TNF-a, MCP-1, RANKL/OPG/M-CSF22e25, but the current study constitutes the first investigation of the correlation of key markers of inflammation and osteoclastogenesis with OA characteristics and cytokine levels. Materials and methods Study sample The study population comprised native Chuvashians residing in a number of small villages in the Republics of Chuvashia and Bashkortostan, Russian Federation. Our aim was to investigate skeletal aging for which the study design (cross-sectional observation study) and data collection methods have been described previously26. The data were gathered during three expeditions undertaken in August/September 1994, May/June 1999, and September 2002 by the Anuchin Research Institute and Museum of Anthropology, Moscow State University (Russia) and the Department of Anatomy and Anthropology, Sackler Faculty of Medicine, 1449 Tel Aviv University (Israel). The same team of investigators collected the data and performed all the measurements during all three expeditions. The collected data included basic socioeconomic parameters, medical history, and standard anthropometric measurements; in addition, plain hand radiographs of both hands and blood samples were obtained from all subjects. The study participants had no chronic or acute infections, metabolic diseases or primary or secondary pathologic amenorrhea, and were not receiving prescription or non-prescription anti-inflammatory drugs, or vitamins, minerals or other dietary supplements on a regular basis. All tested subjects signed an informed consent form. The study was approved by the Ethics Committee of Tel Aviv University, Israel. Cytokine measurements Blood samples were obtained by venepuncture from subjects who had fasted overnight. Cells were removed by centrifugation, and serum aliquots were stored at 80 C until analyzed. Plasma levels of TNF-a and IL-6 were determined using a high-sensitivity commercial enzyme-linked immunosorbent assay (ELISA) with an alkaline phosphatase signal amplification system for TNF-a and IL-6 (Quantikine High-Sensitivity ELISA, Minneapolis, MN, USA). All measurements were carried out in accordance with the manufacturers’ instructions and implemented using a microplate reader. The results were calculated using a four-parameter curve fit and expressed as pg/ml. The intra-assay coefficient of variation was 6%; the inter-assay coefficient of variation was 10%. The minimum detection doses for TNF-a were 0.32 pg/ml and for IL-6, 0.094 pg/ml. MCP-1, IL-17, OPG, RANKL, and M-CSF levels were determined by a sandwich enzyme immunoassay technique using a set of specific antibodies and standards from R&D Systems. The detection limit was 15.6 pg/ml for MCP-1, 40 pg/ml for OPG for IL-17, 20 pg/ml for sRANKL and 9 pg/ml for M-CSF. The inter- and intra-assay variations were less than 4.7% and 10%, respectively. All hormones and biochemical markers were assayed in duplicate. Mean values were used in further analyses. Radiographic assessment of OA Single plain radiograms of both hands were obtained from each study participant in a postero-anterior position, with the X-ray source located 60 cm above the hand, as described in detail by Pavlovsky in 198127. The hands were exposed for 5e10 s at 100e150 mA, without intensifying screens, at 50 kV. The same equipment was used in all three expeditions and X-rays were taken according to the same standardized protocol. Each X-ray was read by an experienced, specially trained researcher. The degree of OA was evaluated in 14 joints of each hand [four distal interphalangeal (DIP), four proximal interphalangeal (PIP), five metacarpophalangeal (MP), and first interphalangeal (IP-1) joints], according to Kellgren and Lawrence’s (KeL) grading criteria, utilizing photographs from the Atlas of Standard Radiographs28. In addition, joint space narrowing (JSN) and osteophyte (OS) formation of the same joints were evaluated using the Atlas of Altman and Gold29. The KeL index for each joint ranged from 0 to 4, with those scored as KeL  2 considered as affected. The JSN and OS indices ranged from 0 to 3, with scores JSN  2 or OS  2 considered as affected. In the present study, we used six indices of radiographic hand OA: (1) three indices of the number of affected joints for each individual (designated Num-KL, Num-JSN and Num-OS) representing severity of hand OA; and (2) three indices of the presence of at least one affected joint (dichotomous indices) (designated DichKL, Dich-JSN and Dich-OS). 1450 I. Pantsulaia et al. / Osteoarthritis and Cartilage 18 (2010) 1448e1453 Table I Descriptive statistics for all studied traits (mean and 95% confidence interval for mean, upper and lower limit) n Men n Women Age (years) Height (m) Weight (kg) BMI (kg/m2) MCP-1 (pg/ml) IL-17 (pg/ml) IL-6 (pg/ml) TNF-a (pg/ml) OPG (pg/ml) 763 763 763 763 479 141 211 212 459 689 689 689 689 413 141 209 209 393 RANKL (pg/ml) 471 M-CSF (pg/ml) 212 49.23 [47.99e50.47] 1.66 [1.66e1.66] 64.60 [63.80e65.40] 23.32 [23.07e23.57] 84.65 [79.16e90.14] 12.97 [9.09e16.85] 2.51 [2.15e2.87] 4.90 [3.91e5.89] 2974.36 [2873.14e 3075.59] 1109.82 [1020.86e 1198.78] 845.31 [791.28e899.34] 50.37 [49.13e51.61] 1.55 [1.55e1.55] 61.06 [60.15e61.97] 25.53 [25.17e25.89] 81.30 [76.34e86.26] 14.25 [10.89e17.61] 2.25 [1.92e2.58] 4.67 [3.70e5.64] 2966.47 [2854.14e 3078.80] 1216.13 [213.89e 2218.37] 938.22 [846.65e 1029.79] 400 209 Reliability of the X-ray readings Two experienced researchers (an orthopedic surgeon and a researcher experienced in radiograph reading) read sets of radiographs and decided on the protocol for evaluating the KeL scores. Ten X-rays were read using this protocol and then re-read separately by two investigators to estimate the intra- and interrater reliability of the readings. All discrepancies were reviewed for systematic errors. This procedure continued for 10 different sets of radiographs until high intra- and inter-rater reliability (k > 0.80) was established. Thereafter, one investigator read all the X-rays, blinded to the patient’s name, sex and age. Before reading each set of radiographs, the investigator re-read five previously read radiographs to “calibrate” his readings to a standard. The intra-rater reliability (k statistics) for KeL scores was 0.84 (P < 0.01), for JSN scores 0.82 (P < 0.01) and for OS scores 0.91 (P < 0.01) based on 20 repeated measurements. Body mass index (BMI) BMI was computed as the ratio of weight (in kilograms) divided by height (in meters) squared. Statistical analysis Preliminary descriptive statistics, including mean, standard deviation (SD), and Pearson correlations were performed using the STATISTICA 8.0 statistical package (Statsoft, Inc, USA). Initial analyses revealed statistically significant deviations from the norm in the data distribution of some cytokines. These data were thus logtransformed prior to further analyses. Thereafter, the correlations between each biochemical index and age, sex, and BMI were calculated. In the female sub-sample, all correlations were first separately evaluated in premenopausal and postmenopausal women. Since no statistical differences were found in the pattern and extent of correlations, all further analyses of female participants were performed without considering menopausal status. For numerical OA indexes, multiple regression models were built with following predictors: age, sex, BMI and biochemical marker concentration (separately for each marker). Results Descriptive statistics and sex differences The main characteristics of the sample population are shown in Table I, presented before log-transformation, in their original units. Average ages of the participants were 49.2 years for the men and 50.4 years for the women. The mean BMI was 23.3 and 25.5 kg/m2 for men and women, respectively. The variation span of the variables for each sex, in the present study was within normal variation range. Testing the distribution of each of the biochemical factors using Fallon et al.’s method30 indicated that individuals with a mean more than four SDs were clearly outliers in the general or the sex-specific samples. These outliers (N ¼ 10) were therefore omitted from the subsequent statistical analysis. There were no significant differences between men and women in the mean concentrations of the studied cytokines (ManneWhitney U test, P > 0.10). For both sexes, the distributions of all measured biochemical indices showed statistically significant deviations from the norm, and were consequently subjected to logarithmic transformation. Statistical data for the other studied biochemical indices (Table I) will not be discussed at present since this was recently performed elsewhere23e26. Age-dependent trends of studied variables The influence of age and BMI on each of the studied variables showed that for both sexes the plasma levels of MCP-1, IL-6, and OPG significantly changed with age P < 0.001 (Table II). The extent of M-CSF correlation with age was greater in women (r ¼ 0.29, P < 0.001) than in men (r ¼ 0.17, P ¼ 0.012), with no significant difference (P > 0.05). OPG was similar in both sexes (r ¼ 0.42e0.43, P < 0.001), while there was no correlation between age and IL-17, RANKL and TNF-a levels for both sexes. There were small but significant regression coefficients between BMI and MCP-1 and OPG (Table II). Interrelationship between OA characteristics and the inflammatory markers An investigation of possible associations between OA characteristics (Num-KL; Dich-KL) and cytokine plasma levels revealed that there was no significant association (P > 0.05) between OA Table II Coefficients of regression between the inflammation biomarkers and age and BMI, according to sex (b with 95% confidence interval, upper and lower limit) Biomarker n Men Age MCP-1 IL-17 IL-6 TNF-a OPG RANKL M-CSF 479 141 211 212 459 471 212 0.108 [0.020e0.196] 0.070 [ 0.229e0.089] 0.253 [0.120e0.386] 0.067 [ 0.202e0.068] 0.473 [0.379e0.567] 0.080 [ 0.172e0.012] 0.277 [0.146e0.408] Significant correlations (P  0.05) are shown in bold type. n Women Age BMI 413 141 209 209 393 400 209 0.212 [0.122e0.302] 0.038 [ 0.205e0.129] 0.210 [0.079e0.341] 0.017 [ 0.152e0.118] 0.556 [0.478e0.634] 0.081 [ 0.174e0.014] 0.195 [0.064e0.326] 0.110 [0.022e0.198] 0.035 [ 0.132e0.202] 0.054 [ 0.079e0.187] 0.021 [ 0.156e0.114] 0.219 [0.125e0.313] 0.034 [ 0.128e0.060] 0.127 [L0.006e0.260] BMI 0.101 [0.017e0.197] 0.071 [ 0.091e0.231] 0.003 [ 0.138e0.132] 0.136 [ 0.271 to 0.001] 0.086 [ 0.008e0.180] 0.025 [ 0.069e0.119] 0.136 [0.001e0.271] 1451 I. Pantsulaia et al. / Osteoarthritis and Cartilage 18 (2010) 1448e1453 Table III Coefficients of regression between the OA characteristics and inflammation biomarkers, age and BMI in males (b with 95% confidence interval, upper and lower limit) Men (n) Num-KL Num-JSN Num-OS Dich-KL Dich-JSN Age (763) BMI (763) MCP-1 (479) IL-17 (141) IL-6 (211) TNF-a (212) OPG (459) RANKL (471) M-CSF (212) 0.648 [0.595e0.701] 0.201 [0.130e0.272] 0.119 [0.027e0.210] 0.078 [ 0.241e0.085] 0.198 [0.064e0.336] 0.099 [ 0.237e0.038] 0.346 [0.256e0.433] 0.066 [ 0.160e0.027] 0.202 [0.067e0.338] 0.196 [0.125e0.267] 0.069 [ 0.003e0.142] 0.135 [0.044e0.226] 0.016 [ 0.147e0.180] 0.111 [ 0.026e0.250] 0.071 [ 0.209e0.067] 0.120 [0.027e0.214] 0.073 [ 0.167e0.021] 0.012 [ 0.126e0.151] 0.531 [0.470e0.592] 0.115 [0.043e0.187] 0.070 [ 0.022e0.161] 0.024 [ 0.138e0.188] 0.103 [ 0.035e0.241] 0.087 [ 0.225e0.051] 0.282 [0.192e0.373] 0.070 [ 0.164e0.024] 0.118 [ 0.019e0.256] 0.727 [0.678e0.776] 0.164 [0.092e0.236] 0.029 [ 0.062e0.121] 0.065 [ 0.228e0.098] 0.208 [0.073e0.344] 0.023 [ 0.115e0.162] 0.356 [0.264e0.441] 0.048 [ 0.142e0.046] 0.262 [0.128e0.395] 0.318 [0.251e0.385] 0.079 [0.006e0.151] 0.041 [ 0.051e0.132] 0.016 [ 0.147e0.180] 0.094 [ 0.044e0.232] 0.006 [ 0.145e0.130] 0.130 [0.037e0.224] 0.089 [ 0.183e0.005] 0.022 [ 0.117e0.161] Dich-OS 0.674 [0.621e0.727] 0.113 [0.041e0.185] 0.013 [ 0.079e0.105] 0.042 [ 0.121e0.206] 0.192 [0.056e0.329] 0.066 [ 0.205e0.071] 0.3235 [0.235e0.413] 0.033 [ 0.128e0.060] 0.162 [0.025e0.299] Significant correlations (P  0.05) are shown in bold type. characteristics and levels of IL-17, TNF, RANKL or M-CSF for both men and women (Tables III and IV). Only MCP-1, IL-6 and OPG were associated with OA traits, however the statistically significant correlations were weak and/or moderate. For testing the effects of age, BMI and cytokines on the degree of OA, multiple regression analysis was performed only for traits significantly correlated with OA. The results showed that only MCP-1 and OPG are significant predictors of the degree of OA, while the other cytokines have no significant effect (Table V). An examination of the association between the studied cytokines showed that there were highly statistically significant correlations between plasma concentrations of RANKL and MCP-1 in both sexes (r ¼ 0.24, P < 0.001). In addition, there was a moderate, but significant, association between MCP-1 and OPG levels in the sub-sample of women (r ¼ 0.19, P < 0.01). Discussion Alterations in OA cartilage are related to changes in chondrocyte metabolism. The course of the destructive process is determined by the balance between anabolic and catabolic mediators and their regulators in the joint and by the local distribution of these mediators in the cartilage3,4. At present, conventional radiography is considered the gold standard in diagnosing OA; however, it cannot detect early stages of OA or monitor OA progression over a short period of time. To overcome these limitations, the search for valid biomarkers has intensified over recent years3,6e10,12,31. The molecular triad, OPG/RANK/RANKL, has been described as a key cytokine system for controlling the differentiation and function of osteoclasts. The balance between the expression of RANKL and OPG determines the extent of osteoclast activity and subsequent bone resorption19,20. Furthermore, in rheumatoid arthritis, the sRANKL/OPG ratio in both mRNA and protein level is increased in synovial fluid/blood and leads to bone destruction observed in this disease20,21. It has also been demonstrated that the ratios of OPG/RANKL and RANK/RANKL in OA chondrocytes are significantly different from those in normal chondrocytes21. Therefore, in the present study, we set out to evaluate the relationship between proinflammmatory and osteoclastogenic cytokines and OA characteristics. We found MCP-1 and OPG to be important predictors of OA development. However, no significant associations between inflammatory markers, osteoclastogenic cytokines (IL-6, TNF-a, IL‑17, RANKL, and M-CSF) and OA characteristics were found. Our results are in partial agreement with Pilichou et al.22 and Toncheva et al.’s32 findings. Pilichou et al. demonstrated that patients with OA of the knee had higher serum levels of OPG and sRANKL and that the local production of OPG strongly correlated with the radiological disease severity of OA19,21. Toncheva et al.32 showed that the sRANKL serum increased in both active and inactive OA patients. However, this elevation was less in active OA, pointing to the possibility that certain mechanisms inhibiting sRANKL are triggered during the active phase32. Consequently, we could conclude that the increase of OPG levels reflect a compensatory response by chondrocytes, macrophages or synovial fibroblasts as to sRANKL elevation and protecting the joint, albeit ineffectively. On the other hand, Vlad et al.33 found no evidence of an association between any of the inflammatory markers and presence of radiographic OA. The results of our study reveal that the levels of the inflammation marker MCP-1 are associated with Num-JSN and Num-OS (Table V). We also showed that MCP-1 levels are significantly correlated (P < 0.01) with RANKL in both sexes and with OPG in women. The association between the osteoclastogenic cytokines and MCP-1 levels in healthy pedigrees is of special interest, since it could shed light on the involvement of MCP-1 in cartilage and bone remodeling. Major strengths of this study: (1) we examined an ethnically, culturally and socio-economically homogeneous population, with phenotype measurements unchanged by the intake of medications or vitamin supplements; (2) the examined individuals were free of acute and chronic immune-related diseases (autoimmune, cancer etc); (3) each individual was simultaneously assessed for all of the examined variables. On the other hand, our study had several limitations. Firstly, levels of markers in and around joints were not measured. We do Table IV Coefficients of regression between the OA characteristics and inflammation biomarkers, age and BMI in females (b with 95% confidence interval, upper and lower limit) Women (n) Num-KL Num-JSN Num-OS Dich-KL Dich-JSN Dich-OS Age (689) BMI (689) MCP-1 (413) IL-17 (141) IL-6 (209) TNF-a (209) OPG (393) RANKL (400) M-CSF (209) 0.650 [0.592e0.709] 0.218 [0.143e0.294] 0.155 [0.063e0.248] 0.041 [ 0.127e0.210] 0.204 [0.072e0.336] 0.012 [ 0.123e0.147] 0.408 [0.321e0.496] 0.030 [ 0.065e0.124] 0.250 [0.119e0.381] 0.169 [0.094e0.245] 0.086 [0.009e0.163] 0.046 [ 0.047e0.140] 0.001 [ 0.168e0.170] 0.152 [0.018e0.285] 0.075 [ 0.060e0.209] 0.116 [0.020e0.211] 0.020 [ 0.115e0.074] 0.099 [ 0.035e0.233] 0.491 [0.424e0.558] 0.074 [ 0.003e0.151] 0.201 [0.110e0.293] 0.028 [ 0.196e0.141] 0.168 [0.035e0.302] 0.032 [ 0.103e0.167] 0.310 [0.219e0.401] 0.012 [ 0.082e0.107] 0.109 [ 0.025e0.243] 0.711 [0.657e0.765] 0.164 [0.088e0.240] 0.107 [0.014e0.200] 0.020 [ 0.189e0.148] 0.109 [ 0.026e0.243] 0.012 [ 0.123e0.147 0.400 [0.312e0.487] 0.041 [ 0.135e0.054] 0.140 [0.006e0.273] 0.257 [0.183e0.331] 0.026 [ 0.051e0.104] 0.056 [ 0.037e0.149] 0.039 [ 0.129e0.208] 0.152 [0.018e0.285] 0.074 [ 0.060e0.209] 0.178 [0.083e0.272] 0.001 [ 0.095e0.094] 0.099 [ 0.035e0.233] 0.596 [0.534e0.658] 0.073 [ 0.004e0.150] 0.147 [0.054e0.239] 0.039 [ 0.208e0.129] 0.133 [ 0.001e0.266] 0.055 [ 0.080e0.190] 0.345 [0.255e0.435] 0.076 [ 0.170e0.018] 0.090 [ 0.044e0.225] Significant correlations (P  0.05) are shown in bold type. 1452 I. Pantsulaia et al. / Osteoarthritis and Cartilage 18 (2010) 1448e1453 Table V Results of multiple regression analyses for Num-KL, Num-JS, Num-OS as dependent variables for single biochemical marker concentration and sex, age and BMI as predictors (b þ SE is given for the biochemical marker variable; R2 is given for the whole model) N Num-KL b P R2 Num-JS b P R2 Num-OS b P R2 MCP-1 IL-17 IL-6 TNF OPG RANKL M-CSF 892 282 420 421 852 871 420 0.029  0.025 0.238 0.439 0.000  0.041 0.990 0.517 0.049  0.036 0.169 0.482 0.011  0.035 0.763 0.481 0.063  0.029 0.035 0.456 0.045  0.025 0.068 0.456 0.068  0.036 0.059 0.484 0.077  0.033 0.022 0.019 0.038  0.057 0.051 0.089 0.085  0.049 0.086 0.048 0.003  0.048 0.937 0.041 0.060  0.040 0.131 0.026 0.017  0.033 0.603 0.025 0.002  0.049 0.972 0.042 0.067  0.029 0.024 0.250 0.011  0.053 0.830 0.211 0.042  0.045 0.395 0.187 0.004  0.044 0.920 0.183 0.077  0.035 0.028 0.246 0.018  0.029 0.528 0.239 0.009  0.045 0.836 0.185 N e number of measured individuals. For all regression models R2 is significant P < 10 bold type. not know whether circulating levels of these markers may correlate well with local levels. However, since osteoarthritic cartilage is the site of increased production of both cytokines and chemokines, including IL-8, monocyte chemoattractant protein-1, and RANTES, IL-615e18, we were unable to reveal the main sources of increased levels of MCP-1 and OPG. We note that Stankovic et al. demonstrated that synovial fluid leukocytes are important sources of chemokines34. Secondly, measurements of the same individuals before and after OA development were not taken, due to the crosssectional design of our study. Thirdly, we evaluated only radiographic OA prevalence and severity. The association between radiographic and symptomatic OA is weak35. In future studies, it will be necessary to evaluate symptomatic hand OA. Fourthly, the studied markers serum levels variations were dependent on genetic factors. In our previous studies22,24 we showed that the additive genetic component was associated with a very substantial variation in MCP-1 (49.25  8.95%), OPG (46.0  9.5%), TNFa (82.3  6.7%) and M-CSF (53.8  10.7%). Consequently, it was not clear that significant relationship between the OA degree and OPG or MCP-1 was due to shared genetic or environmental factors. It is important to note that although the effect of the above-mentioned risk factors, particularly RANKL/OPG and inflammation markers, has been previously evaluated, the simultaneous contribution of all risk factors to OA characteristics has never been assessed. Herein we have begun the assessment process. In summary, this study provides support to the statement that OPG might be a valid biomarker of hand OA. Confirmation of our results in a larger cohort of patients will validate our premise that the RANKL/OPG pathway is a suitable target for the development of novel agents against OA. Author contributions Ia Pantsulaia: study planning, data interpretation, analysis of circulatory inflammatory markers, manuscript preparation; Leonid Kalichman: study planning, X-ray evaluation, data analysis and interpretation, manuscript preparation; Eugene Kobyliansky: study planning and supervising, data analysis and interpretation, manuscript preparation. Conflict of interest None. Acknowledgements We are grateful to Dr Ida Malkin for her help in statistical analysis. This study was supported by a grant from the Israeli 4 ; models with significant biochemical marker influence (P < 0.05) are shown in National Science Foundation e “Academia”, no. 1042-04 and the Hirsch and Genia Wasserman Memorial Fund for medical research. The authors wish to thank Mrs Phyllis Curchack Kornspan for her editorial services. References 1. Zhang Y, Niu J, Kelly-Hayes M, Chaisson CE, Aliabadi P, Felson DT. Prevalence of symptomatic hand osteoarthritis and its impact on functional status among the elderly: the Framingham Study. Am J Epidemiol 2002;156:1021e7. 2. Bonnet CS, Walsh DA. Osteoarthritis, angiogenesis and inflammation. Rheumatology (Oxford) 2005;44:7e16. 3. Goldring MB, Marcu KB. Cartilage homeostasis in health and rheumatic diseases. Arthritis Res Ther 2009;11:224. 4. Goldring MB, Otero M, Tsuchimochi K, Ijiri K, Li Y. Defining the roles of inflammatory and anabolic cytokines in cartilage metabolism. Ann Rheum Dis 2008;67(Suppl 3):iii75e82. 5. Kobayashi M, Squires GR, Mousa A, Tanzer M, Zukor DJ, Antoniou J, et al. Role of interleukin-1 and tumor necrosis factor alpha in matrix degradation of human osteoarthritic cartilage. Arthritis Rheum 2005;52:128e35. 6. Loeser RF. Molecular mechanisms of cartilage destruction: mechanics, inflammatory mediators, and aging collide. Arthritis Rheum 2006;54:1357e60. 7. Hulejová H, Baresová V, Klézl Z, Polanská M, Adam M, Senolt L. Increased level of cytokines and matrix metalloproteinases in osteoarthritic subchondral bone. Cytokine 2007;38:151e6. 8. Burrage PS, Mix KS, Brinckerhoff CE. Matrix metalloproteinases: role in arthritis. Front Biosci 2006;11:529e43. 9. 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. J Immunol 2008;181:5082e8. 10. Fermor B, Christensen SE, Youn I, Cernanec JM, Davies CM, Weinberg JB. Oxygen, nitric oxide and articular cartilage. Eur Cell Mater 2007;11(13):56e65. 11. Pearle AD, Scanzello CR, George S, Mandl LA, DiCarlo EF, Peterson M, et al. Elevated high-sensitivity C-reactive protein levels are associated with local inflammatory findings in patients with osteoarthritis. Osteoarthritis Cartilage 2007;15:516e23. 12. Hussein MR, Fathi NA, El-Din AM, Hassan HI, Abdullah F, Al-Hakeem E, et al. Alterations of the CD4(þ), CD8 (þ) T cell subsets, interleukins-1beta, IL-10, IL-17, tumor necrosis factoralpha and soluble intercellular adhesion molecule-1 in I. Pantsulaia et al. / Osteoarthritis and Cartilage 18 (2010) 1448e1453 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. rheumatoid arthritis and osteoarthritis: preliminary observations. Pathol Oncol Res 2008;14:321e8. Benderdour M, Tardif G, Pelletier JP, Di Battista JA, Reboul P, Ranger P, et al. Interleukin 17 (IL-17) induces collagenase-3 production in human osteoarthritic chondrocytes via AP-1 dependent activation: differential activation of AP-1 members by IL-17 and IL-1beta. J Rheumatol 2002;29:1262e72. Boileau C, Martel-Pelletier J, Moldovan F, Jouzeau JY, Netter P, Manning PT, et al. The in situ up-regulation of chondrocyte interleukin-1-converting enzyme and interleukin-18 levels in experimental osteoarthritis is mediated by nitric oxide. Arthritis Rheum 2002;46:2637e47. Haringman JJ, Smeets TJ, Reinders-Blankert P, Tak PP. Chemokine and chemokine receptor expression in paired peripheral blood mononuclear cells and synovial tissue of patients with rheumatoid arthritis, osteoarthritis, and reactive arthritis. Ann Rheum Dis 2006;65:294e300. Lisignoli G, Toneguzzi S, Grassi F, Piacentini A, Tschon M, Cristino S, et al. Different chemokines are expressed in human arthritic bone biopsies: IFN-gamma and IL-6 differently modulate IL-8, MCP-1 and rantes production by arthritic osteoblasts. Cytokine 2002;20:231e8. Masuko-Hongo K, Sato T, Nishioka K. Chemokines differentially induce matrix metalloproteinase-3 and prostaglandin E2 in human articular chondrocytes. Clin Exp Rheumatol 2005;23:57e62. Alaaeddine N, Olee T, Hashimoto S, Creighton-Achermann L, Lotz M. Production of the chemokine RANTES by articular chondrocytes and role in cartilage degradation. Arthritis Rheum 2001;44:1633e43. Kwan Tat S, Amiable N, Pelletier JP, Boileau C, Lajeunesse D, Duval N, et al. Modulation of OPG, RANK and RANKL by human chondrocytes and their implication during osteoarthritis. Rheumatology (Oxford) 2009;48:1482e90. Tat SK, Pelletier JP, Velasco CR, Padrines M, Martel-Pelletier J. New perspective in osteoarthritis: the OPG and RANKL system as a potential therapeutic target? Keio J Med 2009;58:29e40. Pilichou A, Papassotiriou I, Michalakakou K, Fessatou S, Fandridis E, Papachristou G, et al. High levels of synovial fluid osteoprotegerin (OPG) and increased serum ratio of receptor activator of nuclear factor-kappa B ligand (RANKL) to OPG correlate with disease severity in patients with primary knee osteoarthritis. Clin Biochem 2008;41:746e9. Livshits G, Pantsulaia I, Trofimov S, Kobyliansky E. Genetic influences on the circulating cytokines involved in osteoclastogenesis. J Med Genet 2004;41:e76. 1453 23. Pantsulaia I, Trofimov S, Kobyliansky E, Livshits G. Genetic and environmental influences on IL-6 and TNF-alpha plasma levels in apparently healthy general population. Cytokine 2002;19:138e46. 24. Pantsulaia I, Trofimov S, Kobyliansky E, Livshits G. Contribution of the familial and genetic factors on monocyte chemoattractant protein-1 variation in healthy human pedigrees. Cytokine 2005;32:117e23. 25. Trofimov S, Pantsulaia I, Kobyliansky E, Livshits G. Circulating levels of receptor activator of nuclear factor-kappaB ligand/ osteoprotegerin/macrophage-colony stimulating factor in a presumably healthy human population. Eur J Endocrinol 2004;150:305e11. 26. Malkin I, Kalichman L, Kobyliansky E. Heritability of a skeletal biomarker of biological aging. Biogerontology 2007;8: 627e37. 27. Pavlovsky O. Methods of osseographic studies of the bone. In: Pavlovsky O, Ed. Morpho-physiological Researches in Anthropology. Moscow: MGU Publishing House; 1981:44e61. 28. Kellgren J, Lawrence J. In: Atlas of Standard Radiographs: the Epidemiology of Chronic Rheumatism. Oxford: Blackwell Scientific Publications; 1963;Volume II. 29. Altman RD, Gold GE. Atlas of individual radiographic features in osteoarthritis, revised. Osteoarthritis Cartilage 2007;15 (Suppl A):A1eA56. 30. Fallon A, Spada C. Detection and Accommodation of Outliers in Normally Distributed Datasets, http://www.cee.vt.edu/ewr/ environmental/teach/smprimer/outlier/outlier.html; 2008. 31. Martel-Pelletier J, Boileau C, Pelletier JP, Roughley PJ. Cartilage in normal and osteoarthritis conditions. Best Pract Res Clin Rheumatol 2008;22:351e84. 32. Toncheva A, Remichkova M, Ikonomova K, Dimitrova P, Ivanovska N. Inflammatory response in patients with active and inactive osteoarthritis. Rheumatol Int 2009;29:1197e203. 33. Vlad SC, Neogi T, Benjamin EJ, Larson M, Felson DT. Inflammatory markers and their association with osteoarthritis of the knee. In: American College of Rheumatology Annual Scientific Meeting, Boston, MA, USA. 2007. 34. Stankovic A, Slavic V, Stamenkovic B, Kamenov B, Bojanovic M, Mitrovic DR. Serum and synovial fluid concentrations of CCL2 (MCP-1) chemokine in patients suffering rheumatoid arthritis and osteoarthritis reflect disease activity. Bratisl Lek Listy 2009;110:641e6. 35. Dahaghin S, Bierma-Zeinstra SM, Hazes JM, Koes BW. Clinical burden of radiographic hand osteoarthritis: a systematic appraisal. Arthritis Rheum 2006;55:636e47.