All patients fulfilled the 1987 American College of Rheumatology criteria for a diagnosis of RA 30 and had actively inflamed knee joints. RA ST was obtained at arthroscopy, using local anaesthesia as previously described 31. SF was obtained from 49 patients with inflammatory arthritis (RA, n = 29; Spondyloarthropathy (SpA), n = 20) and five OA patients by arthrocentesis and stored at -80°C. Paired serum was obtained in 45 of 49 of these patients and serum from eight healthy volunteer controls was obtained. Fully informed written consent was obtained from each patient and the study was approved by the St. Vincent's University Hospital Ethics and Medical Research Committee.

ST obtained at arthroscopy (RA, n = 11, psoriatic arthritis (PsA), n = 12, OA, n = 3) were snap frozen in liquid nitrogen and homogenised using a Mikro-Dismembrator (B. Braun Biotech International, Allentown, Pennsylvania, USA). Homogenised samples were resuspended in protein lysis buffer and stored at -80°C. Protein concentration was determined by the BCA protein assay (Pierce, Rockford, IL, USA).

IL-17A protein expression was measured in serum, SF and ST lysates by specific ELISA (R&D Systems Europe, Abingdon, Oxon, United Kingdom). The ELISA standard curve range was 15 pg/ml to 1000 pg/ml. The lowest standard was used as the detection limit. IL-17A expression in synovial tissue lysates was corrected for total protein concentration.

Biopsy samples obtained at arthroscopy were embedded in Tissue Tek medium and then snap-frozen and stored in liquid nitrogen until sectioned for analysis. Serial 7 μM microtome sections were mounted on Superfrost slides and fixed in acetone for 10 minutes. The sections were then washed and blocked with 1 × casein for 30 minutes. The sections were then washed and incubated with 2 μg/ml goat polyclonal anti-IL-17A (Santa Cruz, Heidelberg, Germany) or normal immunoglobulin (Ig) G2a control. After incubation for one hour at room temperature and washing, the sections were incubated with biotinylated mouse anti-goat IgG antibody for 30 minutes at room temperature, followed by strepavidin-peroxidase complex (Dako, Glostrup, Denmark) for 30 minutes and 3,3'-diaminobenzidine tetrahydrochloride for five minutes. Nuclear counterstaining was performed using Mayer's haematoxylin, the sections were then dehydrated, and mounted in Dibutyl Phthlate Xylene (DPX). Sections from 11 RA and 3 OA patients and healthy control tissue (n = 1) were analysed for IL-17A synovium expression.

RASFC were obtained by enzymatic digestion of synovial biopsy specimens with 1 mg/ml of type 1 collagenase (Worthington Biochemical, Lakewood, New Jersey, USA) in RPMI (Gibco BRL, Warrington, UK) for four hours at 37°C in humidified air with 5% carbon dioxide. Dissociated cells were plated in RPMI 1640 supplemented with 10% FCS (Gibco BRL, Warrington, UK), 10 ml of 1 mmol/L HEPES (Gibco BRL, Warrington, UK), penicillin (100 units/ml), streptomycin (100 units/ml), and fungizone (0.25 μg/ml) (all from Biosciences, Co. Dublin, Ireland). The cells were incubated and grown to confluence in T75 flasks (about 10 days) at 37°C in humidified air with 5% carbon dioxide before being harvested with trypsin and passaged. RASFs between the fourth and eighth passages were used for experiments.

Normal human articular cartilage was obtained from patients undergoing surgery for traumatic fracture of the femoral neck, each of whom had no history or radiological evidence of arthritis. Chondrocytes were isolated from the tissue by sequential proteolysis 32. Cells were plated in DMEM supplemented with 10% FCS, 10 ml of 1 mmol/L, penicillin (100 units/ml), streptomycin (100 units/ml) and fungizone (0.25 μg/ml). Chondrocytes were used for experiments up until the eighth passage.

Cartilage explant cultures were prepared using 3 mm punch biopsy specimens, thus ensuring that only full-depth cartilage biopsy samples were used. RASFC and cartilage explants were cultured in 96-well plates in serum-free RPMI 1640 supplemented with 10% FCS, penicillin (100 units/ml), and streptomycin (100 units/ml) in the presence of TNF-α (10 ng/ml), OSM (10 ng/ml) and IL-17A (50 ng/ml) alone and in combination. Cartilage explants were cultured over a time course of 15 days, which has been demonstrated previously to be an optimal time-point to examine proteoglycan depletion from cartilage sections 3334. As RASFCs are the primary cells that invade cartilage, RASFC experiments were also performed over a 15-day time course to examine if induction of MMPs in response to IL-17A could be sustained over the same time period. Culture supernatants were harvested every four days, and wells were replenished with fresh medium containing cytokine conditions that were identical to the medium used on day 1. Following the culture period, supernatants were pooled from each time point and stored at -80°C for analysis of proMMP-1, proMMP-13, TIMP-1, MMP-2 and MMP-9. Cartilage and RASFC experiments were performed in duplicate. Cartilage explants were paraffin embedded for immunohistological analysis.

Primary chondrocytes were cultured in 96-well plates in serum-free DMEM supplemented with penicillin (100 units/ml), and streptomycin (100 units/ml) for 24 hours in the presence of TNF-α (10 ng/ml), OSM (10 ng/ml) and IL-17A (50 ng/ml) alone and in combination. Chondrocyte experiments were performed in duplicate.

Following the culture period, human cartilage explants were removed and fixed overnight in 7% formaldehyde in PBS (pH 7.4) and embedded in wax. Five-micrometer sections were stained with H&E and examined microscopically. For analysis of proteoglycans, 5 μm sections were stained with Safranin O-fast green and counterstained with haematoxylin 35.

An ex vivo RA ST explant model was established, as previously described 36. Each ST biopsy section was placed in a 96-well plate in serum-free RPMI supplemented penicillin (100 units/ml), streptomycin (100 units/ml), and fungizone (0.25 μg/ml) for 24 hours at 37°C in air with 5% carbon dioxide. Synovial explants were then stimulated (in triplicate) for 24 hours with TNF-α (10 ng/ml; R&D Systems, Europe, Abingdon, Oxon, United Kingdom) and IL-17A (10 to 20 ng/ml; R&D Systems, Europe, Abingdon, Oxon, United Kingdom). Following incubation for 24 hours, biopsy wet weights are obtained. The conditioned media was aspirated, collected and frozen at -80°C until assayed for proMMP-1, MMP-2, MMP-9 and TIMP-1 by ELISA and zymography. For Humira blockade experiments, each ST biopsy section was placed in a 96-well plate in full DMEM for 48 hours with Humira (10 μg/ml) or IgG control antibody (4 μg/ml). Following the incubation period, biopsy wet weights were obtained. The conditioned media was aspirated, collected and frozen at -80°. Supernatants were assayed for IL-17A by an MSD assay as this has a sensitivity of 0.4 pg/ml.

ProMMP-1, proMMP-13 and TIMP-1 levels were quantified by specific ELISA (R&D Systems Europe, Abingdon, Oxon, United Kingdom). The ELISA minimum detectable doses were 0.021 ng/ml, 7.7 pg/ml, 0.009 ng/ml, 0.08 ng/ml and 4.91 pg/ml. The ELISA standard ranges were 10 ng/ml to 0.156 ng/ml, 5000 pg/ml to 78 pg/ml, 0.002 ng/ml to 0.045 ng/ml, 10 ng/ml to 0.156 ng/ml and 2.14 pg/ml to 10.0 pg/ml, respectively.

Culture supernatants were separated by electrophoresis under nonreducing conditions by SDS-PAGE in 10% polyacrylamide gels copolymerised with 1% gelatin. Gels were washed vigorously twice for 25 minutes in 2.5% Triton X-100 to remove SDS, rinsed for 25 minutes in dH₂O, then incubated overnight in 50 mM Tris/NaCl, pH 7.5, 10 mM CaCl₂ at 37°C. Following overnight incubation gels were rinsed for five minutes in dH₂O before addition of zymography stain (150 ml dH₂O, 75 ml isopropanol, 25 ml acetic acid, 0.6255 g Brilliant Blue R). Gels were visualised using the UVP Bioimaging AutoChemi system (UVP, Cambridge, UK).

A total of 38 patients were recruited from rheumatology outpatient clinics at St. Vincent's University Hospital and were followed up prospectively for one year. All patients had clinically active disease, with 28-joint count Disease Activity Scores (DAS28) of more than 3.2 points despite conventional disease-modifying anti-rheumatic drug therapy, and were offered treatment with biologic agents. Patients who had previously received biologic therapy were excluded from the study. Changes in conventional therapy were permitted during biologic therapy at the discretion of the patient's treating rheumatologist; however, no changes in the disease-modifying anti-rheumatic drug dosage were made during the study. Following approval by the institutional ethics committee at St. Vincent's University Hospital, all patients gave their fully informed written consent prior to inclusion in the study. All 38 patients began biologic therapy after their baseline assessment of disease activity. Patients were evaluated before and at 1, 3 and 12 months after initiation of biologic therapy.

Blood samples were obtained and sera was separated and stored at -80°C until used for biomarker analysis, and samples were available for this study at baseline and three months. Clinical evaluation at each assessment was performed using the DAS28, and the modified Health Assessment Questionnaire 9. The DAS28 response was analysed both by changes in scores from baseline and by response categories according to the European League Against Rheumatism (EULAR) criteria 9. A DAS28 response at three months was defined as a reduction in the DAS28 score of 0.6 points or more and a final DAS28 score of 5.1 points or less. A DAS28 nonresponse was defined as an improvement of less than 0.6 points or a final DAS28 score of more than 5.1 points. Patients achieving clinical remission at six months were identified according to EULAR criteria (DAS28 < 2.6 points) 9. In addition, the patient's global assessment of his or her overall health was recorded at each visit, using a visual analog scale of 0 to 100 mm, where 0 is the best and 100 is the worst score. Serum was assessed for IL-17A, MMP-1, TIMP-1, MMP3, TIMP-4, acute serum-amyloid A (A-SAA), the collagen degradation markers C1, 2C and C2C and the synthesis markers CS846 and CPII at baseline and three months post therapy.

A-SAA protein levels were measured using a sandwich enzyme immunoassay (Biosource, London, UK). Standards ranged from 9.4 to 300 ng/ml. The minimal detectable dose of the assay was 5 ng/ml.

Expression of IL-17A in synovial explant was assayed by MSD assay. The assay standard range was 0.15 pg/ml to 10,000 pg/ml. The lowest limit of detection was 0.4 pg/ml.

The collagen degradation markers C1, 2C and C2C and the synthesis markers CS846 and CPII were measured by competitive immunoassay as per manufacturer's instructions (Ibex, Montreal, Canada). The ELISA minimum detectable doses were 0.03 μg/ml, 10 ng/ml, 20 ng/ml and 50 ng/ml. The ELISA standard ranges were 0.03 μg/ml to 10 μg/ml, 10 ng/ml to 1 μg/ml, 20 ng/ml to 1000 ng/ml and 50 ng/ml to 2000 ng/ml, respectively

Statistical analysis was performed using SPSS 11 for Windows (SPSS, Chicago, IL, USA). Wilcoxon Rank and Mann Whitney U statistical tests were used. P values less than 0.05 were considered significant.

IL-17A levels were detectable in 54% of SF and 23% of serum. IL-17A serum levels were higher in patients with inflammatory arthritis compared with OA patients (24.3 ± 9.6 pg/ml vs. 12.32 ± 12.32 pg/ml). Serum IL-17A levels were significantly higher in patients with inflammatory arthritis compared with healthy controls (P < 0.05; Figure 1a). Furthermore, levels of IL-17A in SF were significantly higher than their matched serum levels (Figure 1b). SF IL-17A levels were also higher in inflammatory arthritis compared with OA (Figure 1c). High levels of IL-17A expression were demonstrated in ST lysates in both RA and PsA, markedly higher than OA ST (Figure 1d). No significant difference was found between levels in RA and PsA tissue. SF IL-17A levels were found to correlate directly with a measure of disease activity – C-reactive protein (CRP) (n = 43, r² = 0.330, P < 0.05) and disease duration (n = 24, r² = 0.470, P < 0.05). IL-17A was expressed in RA sublining, but not OA or healthy control synovium (Figure 1e). IL-17A expression tended to be scattered throughout the sublining; however, in some RA patients we demonstrated an aggregate of IL-17A positive cells (Figure 1e).

RASFC MMP-1 production was significantly increased by IL-17A (P < 0.015; Figure 2a) and for MMP-13 (P < 0.05; Figure 2b). Primary chondrocyte MMP-1 and MMP-13 expression was also significantly up-regulated by IL-17A (P < 0.05; Figure 2b). IL-17A and OSM combined potentiated the effect on MMP-1 and MMP-13 induction in both primary chondrocytes and RASFC compared with baseline and either cytokine alone (P < 0.05; Figures 2b, d). Similar effects were also observed with TNF-α (data not shown). Although no major effect was observed for MMP-2 activity in RASFCs, it was strongly upregulated by the combination of IL-17A/OSM in primary chondrocytes compared with either cytokine alone. (Figure 2e).

IL-17A upregulates MMP-1 production in human cartilage cultures (Figure 3a); however, this did not reach significance. OSM significantly upregulated MMP-1 production in cartilage explants (P < 0.01). In combination IL-17A significantly potentiated the effect of OSM on MMP-1 production compared with basal (P < 0.01) and either cytokine alone (P < 0.05; Figure 3a). No significant effect was seen on TIMP-1 production (Figure 3b). The combination of IL-17A and OSM significantly shifts the MMP-1: TIMP-1 ratio in favor of cartilage degradation compared with baseline and either cytokine alone (P < 0.01; Figure 3d). A similar effect was seen on the MMP-13/TIMP-1 ratio in cartilage (P < 0.05; data not shown). Cartilage MMP-13 production was also upregulated by IL-17A and TNF-α alone, from basal of 8.34 ± 4.39 ng/ml to 14.47 ± 8.50 ng/ml and 12.56 ± 4.6 ng/ml, respectively (Figure 3c); however, this was not significant. The combination of IL-17A and TNFα had a potentiation effect significantly up-regulating MMP-13 production (P < 0.01; Figure 3c). The MMP-13: TIMP-1 ratio was also significantly shifted by the combination of TNF-α and IL-17A compared with basal (P < 0.01) and either cytokine alone (P < 0.05) (Figure 3d). Cartilage incubated with combinations of TNF-α/IL-17A and OSM/IL-17A, demonstrated almost complete proteoglycan depletion as shown by Safranin-O staining whereas the cytokines alone showed only a mild reduction (Figure 3e).

RA ST explants (n = 11) were cultured in the presence of IL-17A (10 ng/ml or 20 ng/ml) or TNF-α (10 ng/ml), nine showed a response to stimulation with both IL-17A and TNF-α. MMP-1 production by ST explants was increased 1.8 and 2.1 fold by IL-17A (10 and 20 ng/ml) and 4.4 fold by TNF-α from basal of 9202.64 ± 8806.31 ng/mg of tissue to 16,711.80 ± 15,535.07 ng/mg of tissue (P < 0.05), 19,510.08 ± 18,261.87 ng/mg of tissue (P = 0.066) and 40,884.073 ± 3,5621.206 ng/mg of tissue (P < 0.01), respectively (Figure 4a). No significant effect was observed for TIMP-1 production following stimulation with both concentrations of IL-17A or TNF-α (Figure 4b). However, a significant shift in the MMP-1: TIMP-1 ratio was demonstrated (P < 0.01). Furthermore increased expression of pro-MMP-9 and both the pro and active forms of MMP-2 were demonstrated in response to IL-17A stimulation (Figure 4d).

IL-17A protein levels were measured in a previously described cohort of patients 9 undergoing biologic therapy by ELISA. Baseline and three month serum samples of 38 patients were assayed. At baseline, nine patients showed detectable levels of IL-17A. Following three months of therapy, two patients whom at baseline showed undetectable levels had detectable levels of IL-17A. These 11 patients were categorised as the IL-17A-positive group. The remaining 27 patients were negative for IL-17A expression at both baseline and three months and were categorised as the IL-17A-negative group, 70% of which had a clinical response to anti-TNF-α therapy at three months.

In the IL-17A-positive group, 82% showed a decrease in IL-17A levels after three months of therapy, while 18% showed an increase. Patients that showed a decrease post therapy were also those patients that had been defined as clinical responders, while those showing an increase post therapy (dotted lines) were non-responders (Figure 5a). All but one of these patients maintained this response to 12 months post therapy. The change in IL-17A levels pre/post biologic therapy strongly correlated with the change in CRP (r² = 0.817, P < 0.01) and the change in A-SAA (r² = 0.627, P < 0.05), markers of the acute phase response. When patients were categorised into those with detectable IL-17A levels and those with no detectable IL-17A levels, a significant difference in serum matrix turnover markers was demonstrated. MMP-1 and MMP-3 levels were significantly reduced in the IL-17A-negative patients (P < 0.05) compared with the IL-17A-positive patients, which showed no significant change (data not shown). The MMP-1/TIMP-1 ratios were higher in the IL-17A-negative patients compared with IL-17A-positive patients but this was not significant (Figure 5b). A significant reduction was demonstrated for the MMP-1/TIMP4, MMP3/TIMP1 and MMP3/TIMP4 ratios in IL-17A-negative patients (all P < 0.05), compared with IL-17A-positive patients where we demonstrated no significant reduction in any of the MMP/TIMP ratio. This demonstrates decreased matrix degradation three months post therapy in those patients with no detectable IL-17A levels (Figure 5b). Although there is no change from baseline MMP/TIMP ratios pre/post anti-TNF-α therapy in the IL-17A-positive patients, the levels of IL-17A were decreased post therapy. This suggests that patients with detectable IL-17A levels may sustain a higher MMP/TIMP ratio than those that are negative; however, other complex regulatory processes may also be involved in regulating matrix turnover, possibly through interactions with IL-17A or independently. Cartilage biomarkers were also measured in the IL-17A-negative and positive patients. There were no significant differences in the serum levels of C1, 2C, C2C and CPII pre/post therapy in either group. However, CS846 a marker of proteoglycan synthesis was significantly higher in the IL-17A-positive patients compared with the IL-17A-negative group at both baseline (262.5 ± 68.1 ng/ml vs. 158 ± 18.4 ng/ml) and three months post therapy (246.2 ± 60 ng/ml vs. 159 ± 14.9 ng/ml; P < 0.05). Finally, IL-17A production was measured in synovial explants following incubation with Humira and IgG control antibody. Levels of IL-17A were too low to be detected, or were at the lower end of the standard curve and were not reliable.