Catabolic and proinflammatory effects of leptin in chondrocytes are regulated by suppressor of cytokine signaling-3

RESEARCH ARTICLE Open Access Catabolic and proinflammatory effects of leptin in chondrocytes are regulated by suppressor of cytokine signaling-3 Anna Koskinen-Kolasa 1 , Katriina Vuolteenaho 1 , Riku Korhonen 1 , Teemu Moilanen 1,2 and Eeva Moilanen 1* Abstract Background: Previous studies provide evidence that adipokine leptin increases production of catabolic and proinflammatory factors in chondrocytes and serves as a link between obesity and osteoarthritis (OA). However, the magnitude of the response to leptin treatment varies greatly between chondrocytes from different donor patients. In the present study, we investigated the regulatory role of suppressor of cytokine signaling-3 (SOCS-3) in the leptin-induced responses in OA cartilage. Methods: Cartilage and synovial fluid samples from 97 patients with OA undergoing knee replacement surgery were collected. Cartilage samples were cultured with leptin (10 μg/ml), and the levels of proinflammatory and catabolic factors in synovial fluid and in the cartilage culture media, and SOCS-3 expression in the cartilage were measured. The role of SOCS-3 in leptin signaling was further studied in H4 murine chondrocytes by downregulating SOCS-3 with siRNA. Results: Leptin-induced expression of matrix metalloproteinases MMP-1, MMP-3, MMP-13, interleukin-6 (IL-6), inducible nitric oxide synthase (iNOS) and cyclooxygenase-2 (COX-2) were higher in the cartilage samples with low SOCS-3 expression. Accordingly, downregulation of SOCS-3 by siRNA in H4 chondrocytes led to enhanced leptin-induced expression of MMP-3, MMP-13, IL-6 and iNOS. Synovial fluid leptin was associated positively, and cartilage SOCS-3 negatively with synovial fluid levels of MMPs in a multivariate model in obese (body mass index (BMI) >30 kg/m 2 )but not in non-obese (BMI <30 kg/m 2 )patients. Conclusions: Our results show, for the first time, that SOCS-3 regulates leptin-induced responses in cartilage, and could thus be a future drug target in the treatment or prevention of OA, especially in obese patients. Keywords: Leptin, Adipokine, SOCS-3, Osteoarthritis, Chondrocytes, Obesity Background Adipokines are cytokine-like hormones produced by adipose tissue and originally discovered to regulate energy metabolism [1, 2]. Their role in inflammation and obesity-related disease, such as type 2 diabetes mellitus and cardiovascular disease, and also in rheumatic disease has attracted increasing interest during the past decade. Leptin was first characterized in 1994 [3] and to date it is probably the most studied adipokine. The circulating levels of leptin are closely associated with the amount of stored body fat and with body mass index (BMI) [4]. Leptin is, however, not only produced by adipose tissue, but also by several other tissues, including cartilage and other joint tissues [5–7]. Interestingly, synovial fluid leptin levels are also correlated with BMI and leptin expression in chondrocytes is increased in obese individuals with OA [5, 6, 8]. The expression of leptin and its functional receptor Ob-Rb is also reported to be increased in chondrocytes in OA, in comparison to healthy chondrocytes [6]. Obesity is a major risk factor for OA [9]. Traditionally obesity has been thought to explain the risk of developing OA due to increased wear-and-tear on weight-bearing joints. However, obesity is also a risk factor for hand OA * Correspondence: [email protected] 1 The Immunopharmacology Research Group, University of Tampere School of Medicine and Tampere University Hospital, Tampere, Finland Full list of author information is available at the end of the article © 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Koskinen-Kolasa et al. Arthritis Research & Therapy (2016) 18:215 DOI 10.1186/s13075-016-1112-0 [10], which points to a systemic factor or factors that mediate the obesity-related impact on cartilage. Leptin, with its strong positive association with body fat stores, fits well in this picture; in fact, increasing evidence supports the role of leptin as a significant factor in the pathogenesis of OA. Leptin has been shown to have direct proinflammatory and catabolic effects on cartilage in experimental settings. We and others have previously shown that leptin enhances production of catabolic enzymes, including matrix metalloproteinase 1 (MMP-1), MMP-2, MMP-3, MMP-9, MMP-13, a disintegrin and metalloproteinase with thrombospondin motifs 4 (ADAMTS-4) and ADAMTS-5 and proinflammatory mediators, such as nitric oxide (NO), interleukin 6 (IL-6), IL-1β,IL-8and prostaglandin E 2 (PGE 2 ) in chondrocytes, synoviocytes and in cartilage [6, 11–19]. These findings suggest that leptin is not only a bystander of cartilage breakdown, but an active detrimental factor in the pathogenesis of OA. According to our experience, cartilage from different donor patients respond to leptin treatment in a quite versatile manner: some of the samples produce large amounts of catabolic/proinflammatory mediators like MMPs, IL-6 and NO following leptin treatment, while in some samples leptin-induced changes in the production of these factors are very small. Similar wide variation in the response to leptin is also supported by other studies [17]. A study by Pallu et al. showed that primary chondrocytes received from obese patients with OA respond to smaller amounts of leptin to enhance MMP-13 production than chondrocytes obtained from non-obese patients [17], suggesting that obese individuals might be more susceptible to the harmful effects of leptin on cartilage. However, the mechanisms regulating leptin responsiveness in chondrocytes remain unknown. Suppressor of cytokine signaling 3 (SOCS-3) belongs to SOCS proteins, which are intracellular molecules that have an important function of limiting excessive inflammatory activation of the innate and adaptive immune system [20]. In inflammatory cells SOCS-3 expression is induced by type I and type II cytokine receptors via the JAK-STAT pathway. SOCS-3 binds to the gp130 subunit of those receptors and inhibits the JAK-STAT pathway, thus forming a negative feedback loop to limit cytokine actions [21]. Interestingly, SOCS-3 is also involved in regulating leptin responsiveness in the central nervous system (CNS) [22]. The metabolic function of leptin is to serve as a sensor of body fat stores for the CNS. Elevation of blood leptin due to calorie intake, whether short-term or long-term, in a lean person normally suppresses food intake, whereas decreased leptin levels due to fasting or loss of adipose tissue lead to increased food intake [23]. In obesity however, elevated leptin does not lead to the expected responses in weight control. This is thought to be due to disturbed leptin signaling, also called leptin resistance. Elevated SOCS-3 expression in the CNS is proposed to be the primary mechanism that causes leptin resistance and subsequent failure in controlling food intake in obesity [22]. Consistently, leptindeficient mice develop severe obesity [24], whereas SOCS-3 conditional knockout mice are resistant to diet-induced obesity [25]. In humans, genetic leptin deficiency also causes severe obesity, though leptin and leptinreceptor-related mutations are extremely rare [26]. SOCS-3 is also expressed in cartilage [27–29], and we reported previously that its expression is lower in cartilage from obese patients with OA than from non-obese patients [8]. That led us to hypothesize that SOCS-3 could be a significant mechanism behind the variable leptin responsiveness in cartilage samples from different donor patients. We addressed the hypothesis by investigating SOCS-3 expression and leptin responsiveness in cartilage samples obtained from 97 patients with OA. In addition, the role of SOCS-3 expression in leptin signaling was studied by downregulating SOCS-3 by siRNA in chondrocyte cultures. Methods Cartilage and cell cultures Cartilage and synovial fluid (SF) samples were collected from 97 patients with OA who were undergoing knee replacement surgery. All patients fulfilled the American College of Rheumatology classification criteria for OA [30]. Cartilage samples were processed for tissue culture as previously described [15]. Cartilage pieces were incubated for 42 hours with or without leptin (10 μg/ml). The concentration of leptin used was chosen based on our previous studies and on existing literature [15, 17–19]. Recombinant human leptin was purchased from R&D Systems Europe Ltd, Abindgon, UK. Synovial fluid (SF) samples from the corresponding patients were also collected at the beginning of the arthroplasty. The SF samples were centrifuged at 4000 g at 4 °C and supernatants were collected and kept at −70 °C until assayed. The immortalized murine H4 chondrocyte cell line [31], developed in the Laboratory of Experimental Rheumatology, University Medical Center, Nijmegen, The Netherlands, was used in the siRNA experiments. The chondrocytes were cultured at 37 °C in humidified 5 % carbon dioxide atmosphere in Dulbecco’s modified Eagle’s medium (DMEM) with L-glutamine and Ham’sF-12 medium (1:1) supplemented with 5 % fetal bovine serum (all obtained from Lonza Group Ltd, Basel, Switzerland). Immunoassays and nitrite measurements Concentrations of MMP-1, MMP-3, MMP-13 and IL-6 were determined by immunoassays with commercial reagents according to the protocol provided by the manufacturer (human total MMP-1, human total MMP-3, Koskinen-Kolasa et al. Arthritis Research & Therapy (2016) 18:215 Page 2 of 13 human total MMP-13, mouse total MMP-3 and mouse IL6 ELISA kits were from R&D Systems; human IL-6 ELISA kit was from Sanquin, Amsterdam, The Netherlands; MMP-1 in SF was determined by Multiplex bead array, Fluorokine® Human MMP Multi Analyte Profiling Base Kit, purchased from R&D systems). Nitrite, stable metabolite of nitric oxide (NO), was measured in the culture media by the Griess reaction [32]. The cartilage culture media samples were filtered through Amicon Ultra 10-K filters (from Millipore, Cork, Ireland) at 14,000 g prior to the Griess analysis in order to remove large proteins that might interfere with the Griess analysis. RNA isolation and quantitative reverse transcription/ polymerase chain reaction Culture medium was removed at the indicated time points and total RNA of H4 chondrocytes was extracted with GenElute™Mammalian Total RNA Miniprep kit (Sigma-Aldrich, St Louis, MO, USA). Total RNA was treated with DNAse (Fermentas UAB, Vilnius, Lithuania) and reverse-transcribed to cDNA using TaqMan Reverse Transcription reagents and random hexamers (Applied Biosystems, Foster City, CA, USA). cDNA obtained from the RT reaction was diluted 1:20 with RNAse-free water and subjected to quantitative PCR using TaqMan Universal PCR Master Mix and the ABI Prism 7000 Sequence detection system (Applied Biosystems). Primers and probes for SOCS-3, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), iNOS, IL-6 and MMP-13 were obtained from Metabion International AG (Martinsried, Germany). The primer and probe sequences and concentrations (Table 1) were optimized according to the manufacturer’s instructions in TaqMan Universal PCR Master Mix Protocol part number 4304449 revision C. The expression of mouse MMP-3 mRNA was measured using TagMan Gene Expression Assay (Mm00440295_m1, Applied Biosystems). PCR reaction parameters were as follows: incubation at 50 °C for 2 minutes, incubation at 95 °C for 10 minutes, and thereafter 40 cycles of denaturation at 95 °C for 15 s and annealing and extension at 60 °C for 1 minute. Each experimental reaction was performed in duplicate. The relative mRNA levels of SOCS-3, GAPDH, iNOS, IL-6 and MMP-13 were quantified using the standard curve method as described in Applied Biosystems User Bulletin number 2. To calculate the relative expression of MMP-3 mRNA, the 2 (−ΔΔCT) method [33] was used. According to the method, the cycle threshold (C T ) values for MMP-3 mRNA expression in each sample were normalized to the C T values of GAPDH mRNA in the same sample. Western blot Preparation of cell lysates, SDS-polyacrylamide gel electrophoresis and western blot analysis were carried out as previously described [15]. Mouse monoclonal SOCS-3 antibody (sc-51699), rabbit polyclonal iNOS antibodies (sc-651 and sc-650), goat polyclonal cyclooxygenase-2 (COX-2) antibody (sc-1745) and rabbit polyclonal β-actin antibody (sc-1615R), and secondary horseradish peroxidase (HRP)-conjugated goat anti-mouse (sc-2005), goat antirabbit (sc-2004) and donkey anti-goat (sc-2020) antibodies were all from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Rabbit polyclonal MMP-13 antibody (ab39012) was from Abcam (Cambridge, MA, USA). Leptin-induced iNOS and COX-2 expression was determined by running the control and leptin-induced samples side by side and the result is given as fold of change in the β-actinnormalized densitometry value of the leptin-induced versus the control sample. Downregulation of SOCS-3 expression by siRNA H4 murine chondrocytes were seeded at 1 × 10 5 cells/well in 24-well plates. Cells were incubated for 24 hours and transfected with SOCS-3 siRNA or with non-targeting control siRNA. On-Target SMART pool SOCS-3-specific siRNA (targeting sequences of GGCUAGGAGACUCGC CUUA, GGACCAAGAACCUACGCAU, CUAAUGAAA CCUCGCAGAU and GAAGGGAGGCAGAUCAACA) and siGENOME Non-Targeting siRNA were used at 100 nM to transfect the cells using DharmaFECT 1. All transfection reagents were from Thermo Scientific Dharmacon (Lafayette, CO, USA) and transfection was carried out according to the manufacturer’s protocol. The experiments were started 48 hours after the transfection by adding leptin (10 μg/ml) (mouse recombinant leptin from R&D systems) in fresh culture medium. Statistical analysis The chi-square test, unpaired ttest and Mann–Whitney test (where appropriate) were used to analyze differences between subgroups of the patients. The Wilcoxon test was used to calculate the significance of leptin-induced effects in the cartilage culture. To analyze the differences in leptin responsiveness in relation to SOCS-3 expression, the samples on each western blot gel were divided to two equal sized groups (low SOCS-3 or high SOCS-3) according to SOCS-3 expression. Median leptin responses, measured as change in the production of MMP-1, MMP-3, MMP-13, IL-6 and NO in the leptin-treated versus control sample, and as fold of change in the expression of iNOS and COX-2, were compared between the low SOCS-3 and the high SOCS-3 groups. Possible intergel differences in SOCS-3 expression were controlled by analysis of variance (ANOVA) in which the leptin response variable (e.g., leptin-induced change in production of MMP-1) was set as a dependent variable, western blot gel (1 to 8) as a grouping variable and SOCS-3 Koskinen-Kolasa et al. Arthritis Research & Therapy (2016) 18:215 Page 3 of 13 expression as a continuous variable as a covariate. Associations were further tested by adjusting for BMI and age. Correlation between the factors of interest in SF were determined by Pearson’s correlation analysis. The associations between MMPs or IL-6 and leptin in SF, and SOCS-3 expression in cartilage were further analyzed by ANOVA modeling, by including the variable of interest (SF MMP-1, MMP-3 or IL-6) as a dependent variable, leptin in SF and SOCS-3 expression in the cartilage as covariates and gel number as a grouping factor. The analysis was done in BMI subgroups (obese, BMI >30 kg/m 2 ;non-obese,BMI <30 kg/m 2 ). Natural logarithms were formed of the leptin response values, SOCS-3 expression levels and SF levels of the measured variables where appropriate in order to have normally distributed variables for the ANOVA modeling and for the correlation analyses. ThedatawereanalyzedbyIBMSPSSStatistics19(IBM Corporation, NY, USA) and Graph-Pad InStat version 3.00 software (GraphPad Software Inc., San Diego, CA, USA). The results of the siRNA experiments are presented as means (SEM). The statistical significance of these data was calculated by two-way ANOVA with Bonferroni multiple comparisons post-test using Graph-Pad Prism 5 for Windows version 5.04 (GraphPad Software Inc.). Differences were considered statistically significant at p< 0.05. Results Leptin-induced production of proinflammatory and catabolic factors in osteoarthritic cartilage in relation to clinical factors and SOCS-3 expression Patient characteristics and leptin responses in the cultured cartilage across the whole study population and in the obese (BMI >30 kg/m 2 ) and non-obese subgroups are presented in Table 2. Leptin significantly enhanced the expression of MMP-1, MMP-3, MMP-13, IL-6, iNOS and COX-2 and NO production in OA cartilage ex vivo (Fig. 1). However, there was considerable variation in these responses between the samples from different donor patients (Table 2). There were no statistically significant differences in the leptin responses between obese and non-obese patients (Table 2), and neither did the leptin responses correlate with age, sex or radiographic scaling of OA. When the patients were divided into subgroups according to SOCS-3 expression in the cartilage, leptin-induced changes in the expression/production of MMP-1, MMP-3, MMP-13, IL-6, NO, iNOS and COX-2 in the cartilage were significantly greater in the samples with low SOCS-3 expression than in the samples with high SOCS-3 expression (Fig. 2). This suggests that the level of SOCS-3 expression determinates the magnitude of leptin-induced inflammatory responses. The results remained statistically significant (p< 0.05) for the responses in the expression of MMP-3, MMP-13, IL-6, NO, iNOS and COX-2, and almost significant for response in the expression of MMP-1 (p= 0.10) in the ANOVA modeling after controlling for intergel variation, BMI and age. Synovial fluid levels of MMPs and IL-6 in relation to SF leptin and SOCS-3 expression in cartilage from patients with OA SF samples were obtained from 90 of the 97 patients. Obese patients had significantly higher SF leptin than non-obese patients, while SF MMP-1 and MMP-3 did not significantly differ between obese and non-obese Table 1 Primer and probe sequences for quantitative RT-PCR Gene Oligunucleotide Sequence Conc. (nM) Forward primer GCATGGCCTTCCGTGTTC 300 Mouse GAPDH Reverse primer GATGTCATCATACTTGGCAGGTTT 300 Probe TCGTGGATCTGACGTGCCGCC 150 Forward primer GCGGGCACCTTTCTTATCC 300 Mouse SOCS-3 Reverse primer AAGCTGCCCCCCTCACA 300 Probe CTCGGACCAGCGCCACTTCTTCA 150 Forward primer CCTGGTACGGGCATTGCT 300 Mouse iNOS Reverse primer GCTCATGCGGCCTCCTT 300 Probe CAGCAGCGGCTCCATGACTCCC 150 Forward primer TCGGAGGCTTAATTACACATGTTC 900 Mouse IL-6 Reverse primer CAAGTGCATCATCGTTGTTCATAC 300 Probe CAGAATTGCCATTGCACAACTCTTTTCTCA 200 Forward primer TTGTGTTTGCAGAGCACTACTTGA 900 Mouse MMP-13 Reverse primer AACTGTGGAGGTCACTGTAGACTTCTT 900 Probe CATCCTGCGACTCTTGCGGGAATC 250 SOCS-3 suppressor of cytokine signaling-3, iNOS inducible nitric oxide synthase, IL-6 interleukin-6, MMP-13 matrix metalloproteinase-13, Conc. concentration Koskinen-Kolasa et al. Arthritis Research & Therapy (2016) 18:215 Page 4 of 13 patients (Table 2). Leptin correlated positively with MMP1 and with MMP-3 in SF from obese but not from nonobese patients (Fig. 3). In ANOVA modeling, leptin concentrations in SF and SOCS-3 expression in cartilage significantly explained levels of SF MMP-1 and MMP-3 in the obese but not in the non-obese group (Table 3) pointing to obesity-related association of leptin and SOCS-3 in OA pathophysiology. In addition, SF IL-6 levels were explained by SOCS-3 in the obese but not in the non-obese group, while leptin did not significantly explain SF IL-6 levels in either of the BMI subgroups (Table 3). SOCS-3 modulates leptin responses in chondrocytes In order to investigate further the role of SOCS-3 in the regulation of leptin-induced responses in chondrocytes, we used siRNA to downregulate SOCS-3 in the H4 chondrocyte cell line. H4 chondrocytes expressed SOCS-3 mRNA at relatively high levels and it was reduced by approximately 80 % in the SOCS-3-siRNA-treated cells when compared to the cells transfected with control siRNA. Leptin had a clear effect on inducing MMP-3, MMP-13, IL-6 and iNOS expression in the SOCS-3deficient cells, whereas in the control siRNA-treated cells leptin did not have any statistically significant effect on the production of these factors (Fig. 4), confirming that SOCS-3 negatively regulates leptin-induced proinflammatory responses in chondrocytes. Discussion Leptin has been shown to have detrimental effects on cartilage metabolism in several studies [6, 11–19]. However, considerable variation in leptin responsiveness between cartilage/chondrocytes from different patients has been observed. Our present results indicate that a significant mechanism behind the differential leptin responsiveness could be SOCS-3. SOCS-3 is a known negative regulator of inflammatory signals [20]. Its role in controlling the effects of leptin in chondrocytes has not been previously investigated, but it has been reported to regulate the responses of leptin in the CNS [22]. In the present study we show, for the first time, that SOCS-3 regulates the proinflammatory and catabolic effects of leptin in chondrocytes. This was demonstrated as greater leptin responsiveness in cartilage explants with low SOCS-3 expression in comparison to lower leptin responsiveness in the explants with high SOCS-3 expression. The causality of this association was illustrated by downregulation of SOCS-3 by siRNA in the chondrocyte cell line, which led to increased leptininduced expression of proinflammatory and catabolic genes. In addition, SF leptin levels were shown to be positively associated, and cartilage SOCS-3 expression negatively associated with SF MMP levels in obese, but not in non-obese patients with OA. This points to dysregulation of the leptin-SOCS-3 axis, especially in obese Table 2 Patient characteristics and leptin responses in cartilage cultures in the whole study population and compared across body mass index subgroups All Non-obese, BMI <30 kg/m 2 Obese, BMI >30 kg/m 2 n=97 n=49 n=48 P Gender (female/male) a 60/37 26/23 34/14 0.072 Body mass index (kg/m 2 ) b 30.9 (6.1) 26.2 (2.4) 35.7 (4.6) <0.001 Age (years) b 69.8 (10.0) 72.8 (9.7) 66.8 (9.4) 0.003 Synovial fluid leptin (ng/ml) c, d 12.8 (17.8) 7.6 (11.0) 21.5 (26.7) <0.001 Synovial fluid IL-6 (pg/ml) c, d 118.9 (196.0) 126.8 (204.3) 114.0 (280.2) 0.784 Synovial fluid MMP-1 (ng/ml) c, d 14.4 (25.7) 10.4 (16.6) 18.1 (27.3) 0.325 Synovial fluid MMP-3 (ng/ml) c, d 649.5 (929.6) 591.6 (571.0) 764.9 (1159.0) 0.106 Leptin response in cartilage MMP-1 (change pg/mg cartilage) c 145.8 (247.1) 123.2 (253.6) 150.0 (258.3) 0.773 MMP-3 (change ng/mg cartilage) c 5.2 (8.8) 6.0 (8.6) 4.9 (10.2) 0.920 MMP-13 (change pg/mg cartilage) c 5.8 (13.6) 6.3 (11.8) 5.4 (15.9) 0.983 IL-6 (change pg/mg cartilage) c 123.2 (310.2) 114.6 (295.1) 130.9 (312.0) 0.740 NO (change pmol/mg cartilage) c 44.5 (133.4) 31.0 (125.8) 52.2 (140.0) 0.359 iNOS (fold of increase) c, e 11.7 (160.6) 5.4 (143.1) 15.2 (209.6) 0.501 COX-2 (fold of increase) c, e 6.9 (18.4) 7.1 (15.0) 6.4 (22.6) 0.748 a Values are numbers of female/male subjects; pvalue was calculated for comparison between non-obese and obese subjects using the chi-square test. b Values are mean (SD); pvalues were calculated for comparison between non-obese and obese subjects using the unpaired ttest. c Values are median (IQR); pvalues were calculated for comparison between non-obese and obese subjects using the Mann–Whitney test. d Synovial fluid sample was obtained from 90 patients. e Numbers of patients (non-obese/obese) in the analysis were 26/31 for inducible nitric oxide synthase (iNOS) and 25/29 for cyclooxygenase-2 (COX-2). MMP matrix metalloproteinase, IL interleukin, NO nitric oxide Koskinen-Kolasa et al. Arthritis Research & Therapy (2016) 18:215 Page 5 of 13 individuals, and to a possible obesity-related pathogenic mechanism in OA. In the present study we observed a positive association between leptin levels and matrix metalloproteinases in SF that was only present in the obese patients with OA. However, obesity did not explain the differential leptin responsiveness in the cartilage culture experiments, unlike in the study by Pallu et al. where greater leptin Fig. 1 Effect of leptin on the production of matrix metalloproteinase-1 (MMP-1)(a), MMP-3 (b), MMP-13 (c), interleukin-6 (IL-6)(d), nitric oxide (NO) (e) and on the expression of inducible nitric oxide synthase (iNOS)(f) and cyclooxygenase-2 (COX-2)(g) in cartilage from patients with osteoarthritis (OA). Cartilage samples from 97 patients with OA were cultured with and without leptin (10 μg/ml) for 42 hours. Concentrations of MMP-1, MMP-3, MMP-13 and IL-6 were measured by ELISA; NO production was determined as its metabolite nitrite by the Griess reaction and iNOS and COX-2 proteins by western blotting. The circles represent the medians. The whiskers represent 95 % confidence interval of the median. Statistical significance was calculated using the Wilcoxon test; ***p< 0.001 Koskinen-Kolasa et al. Arthritis Research & Therapy (2016) 18:215 Page 6 of 13 Fig. 2 Leptin-induced production/expression of matrix metalloproteinase-1 (MMP-1)(a), MMP-3 (b), MMP-13 (c), interleukin-6 (IL-6)(d), nitric oxide (NO)(e), inducible nitric oxide synthase (iNOS)(f) and cyclooxygenase-2 (COX-2)(g) in cartilage from patients with osteoarthritis (OA) in subgroups stratified by suppressor of cytokine signaling-3 (SOCS-3) expression in the non-treated cartilage. Human osteoarthritic cartilage was cultured with leptin (10 μg/ml) for 42 hours. Concentrations of MMP-1, MMP-3, MMP-13 and IL-6 were measured by ELISA, NO was determined as its metabolite nitrite by the Griess reaction and iNOS and COX-2 proteins were analyzed by western blotting. The circles represent the median change in the leptin-induced effects. The whiskers represent the 95 % confidence interval of the median. Numbers of patients from whom the cartilage samples were collected are indicated. Statistical significance was calculated using the Mann–Whitney test; *p< 0.05, **p< 0.01 Koskinen-Kolasa et al. Arthritis Research & Therapy (2016) 18:215 Page 7 of 13 Fig. 3 Correlation between leptin and matrix metalloproteinase-1 (MMP-1) and MMP-3 in non-obese (a) and obese (b) patients with osteoarthritis. Leptin and MMPs were measured in synovial fluid (SF) by immunoassay. Natural logarithms (LN) were formed of the SF levels of leptin and MMPs in order to have normally distributed variables for the Pearson correlation analysis. Correlation coefficients (r) and pvalues are indicated. Samples were collected from 90 patients (non-obese, BMI <30 kg/m 2 ,n= 44; obese, BMI >30 kg/m 2 ,n= 46) Table 3 Associations between interleukin-6 (IL-6), matrix metalloproteinase-1 (MMP-1), MMP-3 and leptin in synovial fluid and suppressor of cytokine signaling-3 (SOCS-3) expression in cartilage from non-obese and obese patients with osteoarthritis Non-obese, BMI <30 kg/m 2 Obese, BMI >30 kg/m 2 Dependent variable Covariates R 2 adjusted PR 2 adjusted P LN (SF MMP-1) 0.15 0.30 LN SOCS-3 0.818 0.007 LN (SF leptin) 0.884 0.023 LN (SF MMP-3) 0.03 0.27 LN SOCS-3 0.608 0.004 LN (SF leptin) 0.733 0.015 LN (SF IL-6) −0.05 0.20 LN SOCS-3 0.945 0.003 LN (SF leptin) 0.808 0.466 Pvalues are calculated for covariates in analysis of variance modeling. The model is controlled for intergel variation in SOCS-3 expression levels. Analysis was performed in body mass index (BMI) subgroups. Natural logarithms (LN) were formed where appropriate. SF synovial fluid Koskinen-Kolasa et al. Arthritis Research & Therapy (2016) 18:215 Page 8 of 13 Fig. 4 The effect of silencing of suppressor of cytokine signaling-3 (SOCS-3) by siRNA on leptin-induced expression of matrix metalloproteinase-3 (MMP-3)(a,e), MMP-13 (b,f), interleukin-6 (IL-6)(c,g) and inducible nitric oxide synthase (iNOS)(d,h) in H4 murine chondrocytes. The cells were transfected with SOCS-3 siRNA or non-targeting siRNA (siNEG) and treated with leptin (10 μg/ml) for 4 (c,d), 8 (a,b,h)or24(e-g) hours. mRNA expression (a-d) was determined by quantitative RT-PCR, the levels of MMP-3 (e) and IL-6 (g) in the culture media supernatants by ELISA, and MMP-13 (f) and iNOS (h) expression in the chondrocyte lysates by western blotting. Results are expressed as means ± SEM; n=6(a-eand g) and n=3(f,h). MMP-3 protein level in siNEG and in non-treated SOCS-3 siRNA samples was below the detection limit and is set as half of the lowest standard. Representative bands of the western blots are shown. Statistical analysis was carried out by two-way analysis of variance with Bonferroni multiple comparisons post hoc test; *p< 0.05, **p< 0.01, ***p< 0.001. n.s. not significant Koskinen-Kolasa et al. Arthritis Research & Therapy (2016) 18:215 Page 9 of 13