The synovial membrane is not entirely homogeneous, leading to a local variability within a joint. Synovial tissue can be retrieved from patients in a number of ways (e.g. open surgery, blind needle biopsy, and arthroscopy).
To study inter-patient versus intra-patient variability in gene expression in rheumatoid arthritis (RA) synovial tissue.
Patients and methods
Three synovial biopsies were taken from three RA patients (patients 1–3) at joint replacement surgery. All nine biopsies were divided into three, creating nine samples from each joint, and a total of 27 samples.
Two to four synovial biopsies were sampled at arthroscopy from four RA patients (patients 4–7) with an inflamed knee joint. Multiple samples were taken from different sites in all patients.
Handling of synovial tissue and microarray analysis
All samples were snap frozen and stored at -80°C. Following RNA extraction (at 4°C), reverse transcription, cDNA amplification and labelling, each sample was hybridised in duplicate against a reference, on a cDNA array locally produced, representing 18,139 unique genes. After data filtering, genes were defined as differentially expressed (DE) if they had a B score > 0 and a fold change > 2. Several hierarchical clusterings were performed to obtain an overview of the data. DE genes in the samples were used to compare the variation between samples from one biopsy, the variation between biopsies from one joint and the variation between biopsies from different patients. Clustering also allowed comparisons between the sampling techniques.
In the orthopaedic subset, variation between samples within one biopsy (patients 1–3) was between 1.2% and 6.7% of analysed genes. The amount of DE genes between biopsies was between 1.2% and 5.9%. Hierarchical clustering of biopsies showed more similarities within a patient than between patients, except for one biopsy. One biopsy from patient 1 clustered with biopsies from patient 2 because the other two biopsies from patient 1 had higher fat cell content (confirmed microscopically) than any of the other biopsies. This was confirmed by gene expression profiles with activity in the fatty acid and lipid metabolism. When these genes were removed from the analysis, the samples clustered perfectly according to patient. The analysis of the arthroscopic subset (patients 4–7) showed a variation of DE genes between biopsies of 0.3–1.6%, and a perfect matching in cluster analysis with all biopsies from the same patient clustering together, without any further labouring with data.
This study shows that inter-patient variability is greater than intra-patient variability in gene expression, in RA synovial tissue. We found that one biopsy is enough to represent one patient as they clustered nicely together, with one exception, which was related to high content of fat cells in the biopsy. Variability is in the same order of magnitude in samples close to each other as compared with those further apart within the same patient.