Biologic predictors of extension of oligoarticular juvenile idiopathic arthritis as determined from synovial fluid cellular composition and gene expression

Abstract

Objective: To identify biomarkers in the first synovial fluid (SF) aspirate obtained from children with oligoarticular juvenile idiopathic arthritis (JIA), which could be used to identify children whose disease is likely to extend to a more severe phenotype.

Methods: Patients with recent-onset oligoarticular JIA were identified and grouped according to those whose mild disease persisted (persistent disease) or those whose disease would extend from a mild to more severe phenotype (extended-to-be disease) at 1 year after diagnosis. Flow cytometry was used to delineate differences in the mononuclear cell populations between the first blood sample and first SF aspirate from the same patient and between outcome (persistent versus extended-to-be) groups. Proportions of lymphocytes in the joint were modeled on chemotaxis of lymphocytes to CCL5, using Transwell migration assays. Levels of CCL5 in the SF were quantified by enzyme-linked immunosorbent assay. RNA profiles of SF mononuclear cells were compared between groups using the Affymetrix GeneChip hybridization protocol and hierarchical clustering analyses.

Results: Compared with peripheral blood mononuclear cells, SF mononuclear cells displayed an expansion of CD8+ T cells, reduced proportion of B cells, and expansion of CD16- natural killer cells. The lower CD4:CD8 ratio in the SF was recapitulated in vitro by the observed migration of blood T cells in response to CCL5. Synovial CCL5 levels were higher in children whose disease extended to a more severe phenotype. The CD4:CD8 ratio in the SF was significantly lower in patients with extended-to-be oligoarticular JIA (0.57 compared with 0.90 in the persistent disease group, difference 0.33, 95% confidence interval 0.04-0.62; P = 0.009). Gene expression profiling revealed that 344 genes were >1.5-fold differentially expressed between outcome groups (P < 0.05), and these included genes associated with inflammation and macrophage differentiation, which showed increased levels in patients with extended disease at 1 year, and genes associated with immune regulation, which showed increased levels in patients with persistent disease at 1 year.

Conclusion: Analyses of the proportions of synovial lymphocytes, levels of CCL5, and differential gene expression yielded potential biomarkers with which to predict the likelihood of extension of oligoarticular JIA to a more severe disease phenotype.

Figure 1

Differences in mononuclear cell subsets in the synovial fluid compared with peripheral blood of children with oligoarticular juvenile idiopathic arthritis. Flow cytometry was used to enumerate proportions of A, T cells versus B cells from the live cell gate, B, CD4+ and CD8+ T cell subsets from the CD3+ live cell gate, as well as T cell receptor γ/δ–bearing (γδTCR) cells detected with a specific antibody, and C, natural killer cell subsets from the CD56+ live cell gate. Representative results are shown, in flow cytometry plots (left) and quantitatively as the mean ± SD proportion of each cell subset in 14 paired patient samples (right).

Figure 2

Differences in the proportions of synovial fluid mononuclear cell (SFMC) subsets in patients with persistent oligoarticular juvenile idiopathic arthritis (JIA) (n = 21) compared with those with extended-to-be oligoarticular JIA (n = 11), and the utility of these measures as predictors of outcome. A, Proportions of T cell, B cell, monocyte/macrophage (CD14+), and total myeloid cell (CD13+) subsets, derived from flow cytometry, in patients with persistent oligoarticular JIA (light gray bars) compared with patients with extended-to-be disease (dark gray bars). Results are the mean ± SEM percentage relative to that of total live cells. B, CD4:CD8 T cell ratios, derived from flow cytometry, in patients with persistent versus extended-to-be oligoarticular JIA. Circles indicate individual samples; horizontal lines indicate the median. C, Probability of disease extension in a logistic regression model using the proportions of CD4+ and CD19+ lymphocytes to predict the outcome, which was correctly predicted in 27 (84%) of 32 patients. Circles indicate individual patients; horizontal lines indicate the median. D, Receiver operating characteristic curves comparing the CD4:CD8 ratio (area under the curve 0.79) with the CD4+ T cell–B cell model derived from logistic regression (area under the curve 0.90). ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001, between outcome groups, by Mann-Whitney U test.

Figure 3

Measures of proliferation, apoptosis, and migration of T cell subsets in the synovial fluid (SF) of patients with oligoarticular juvenile idiopathic arthritis. A–C, SF (n = 12) (A and B) and peripheral blood (n = 7) (C) CD4+ and CD8+ T cells were compared for A, intracellular expression of Ki-67, B, binding of VAD tripeptide to active caspases, and C, surface expression of CCR5 as determined by flow cytometry gated on live, CD3+ cells. D, The levels of CCL5 (RANTES) in the SF (n = 10 with persistent disease, n = 5 with extended-to-be disease) were measured by enzyme-linked immunosorbent assay. Circles indicate individual patients; horizontal lines indicate the median. E, SF levels of CCL5 were compared with CD8+ T cell levels as a percentage of the total live SF mononuclear cells (SFMC) from 14 patients (r² = 0.13, P = 0.21). F, Migration to CCL5 lowers the CD4:CD8 ratio, as determined in a Transwell assay using nonadherent peripheral blood mononuclear cells from healthy adults. Bars show the mean ± SD results from 3 individual samples. ∗∗ = P < 0.01 by Mann-Whitney U test.

Figure 4

Hierarchical clustering analysis of genes differentially expressed in synovial fluid mononuclear cells of patients with persistent oligoarticular juvenile idiopathic arthritis (JIA) (n = 13) compared with patients with extended-to-be oligoarticular JIA (n = 8). The heat map shows genes (as labeled to the right of the map) that were 2-fold differentially expressed at a significance level of P < 0.05, as determined by Student's t-test with Welch's correction. Each row represents a different probe set, and each column represents an individual patient sample. The normalized expression level for each gene and each patient is indicated by colors, with red, yellow, and blue reflecting expression levels greater than, equal to, or less than the mean value, respectively, for all of the patients. The boxes below the cluster indicate patients whose disease persisted versus those whose disease had extended at 1 year after diagnosis.