Subjects

Synovial fluid was taken from the TMJs of 30 patients, 24 bilaterally and six unilaterally for a total of 54 joints. Twenty patients (35 TMJs) were children with JIA, five patients (nine TMJs) were adults with previous JIA (AJIA) continuing into adulthood, and five patients (10 TMJs) had RA. Clinical findings in the JIA patient cohort have been published previously.

Clinical examinations were performed before synovial fluid collection and at two follow-up appointments (3 and 8 months). The results of these examinations have been described previously. They included recordings of maximum incisal opening (MIO), pain incisal opening (PIO), lateral excursion, and pain upon palpation of muscles and the TMJs, as well as subjectively reported pain and function of the jaw (visual analogue scale, VAS). Bacteria were detected in synovial fluid in 19 patients. Six patients (12 TMJs; three JIA, one AJIA, and two RA) were excluded from the study due to possible contamination.

This prospective clinical trial of arthrocentesis of the TMJ in patients with JIA, AJIA, and RA was approved by the regional medical ethics committee.

Sample collection

Samples were collected from the TMJ using a push and pull technique, as previously used by Alstergren et al. and in a recently published study. Ultrasound-guided sampling was performed in an operation field after prior disinfection of the skin in an area of 3 cm surrounding the penetration site; this was done using five washes of chlorhexidine–ethanol 5 mg/ml solution (Klorhexidinsprit 5 mg/ml; Fresenius Kabi, Nydalen, Norway) in accordance with the recommendations of Oslo University Hospital. Skin swabs were taken from both the disinfected skin and non-disinfected skin in the puncture area prior to sampling, using a sterile Whatman OmniSwab (GE Healthcare, USA). All samples were frozen immediately at −80 °C.

DNA isolation, PCR amplification, and sequencing

The extraction of genomic DNA was performed using a MasterPure Complete DNA Purification Kit (Epicentre Biotechnologies, Madison, WI, USA), based on the manufacturer’s extraction protocol for fluid samples.

PCR amplification was performed using 16S rRNA gene primers flanking the V3 and V5 region: E334F 5′-CCAGACTCCTACGGGAGGCAGC-3′ and E939R 5′-CTTGTGCGGGCCCCCGTCAATTC-3′. The PCR amplification steps included a denaturation step at 96 °C for 2 min and 32 amplification cycles: 96 °C for 30 s, 61 °C for 40 s, 72 °C for 30 s.

Amplicons were gel-purified (GenElute, NA1111; Sigma Aldrich) and re-sequenced with fusion primers, Roche 454 FLX adaptor primers, 8-base bar code in front of E334F, reverse adaptor primer (B) fusion with E939, and AMPure bead purification (Agencourt AMPureXP beads, A63880; Beckman Coulter, Inc.) of fusion PCR products, and were quality check by Agilent Bioanalyzer high sensitivity chip (Agilent High Sensitivity DNA, catalogue No. 5067-4627; Agilent Technologies, Inc.).

Amplicons were sequenced using a GS-FLX 454 sequencer (454 Life Sciences, Roche) at the Norwegian Sequencing Centre, Centre for Ecological and Evolutionary Synthesis, University of Oslo, Norway. Sequences (reads) were edited using the Ribosomal Database Project (RDP) pipeline and mothur software. Chimera UCHIME and reference 16S rRNA ( http://www.arb-silva.de ) were used before trimmed FASTA sequences were blasted using the Bioportal at the University of Oslo. MEGAN (Centre for Bioinformatics ZBIT, Tübingen University, Germany) was used to visualize the clustering and taxonomy of the samples.

Statistical analysis

Statistical analyses of the correlation between clinical data, immunological data, and bacteria detection were performed using Sigma Plot 12.0 for Windows (Systat Software, Inc., San Jose, CA, USA). Parametric and non-parametric statistics ( t -test and Wilcoxon test) were performed for patients with either one or two TMJ synovial fluid samples positive for bacteria detection, and to compare patients with and without the presence of bacteria in the TMJ synovial fluid.