IL-6-driven STAT signalling in circulating CD4+ lymphocytes is a marker for early anticitrullinated peptide antibody-negative rheumatoid arthritis

Abstract

Objectives: A previously identified signal transduction and activator of transcription-3 (STAT3) target-enriched gene signature in circulating CD4+ T cells of patients with early rheumatoid arthritis (RA) was prominent in autoantibody-negative individuals. Here, interleukin (IL)-6-mediated STAT signalling was investigated in circulating lymphocytes of an independent early arthritis patient cohort, seeking further insight into RA pathogenesis and biomarkers of potential clinical utility.

Methods: Constitutive and IL-6-induced expression of phosphorylated STAT1 (pSTAT1) and pSTAT3 was determined in T and B cells using Phosflow cytometric analysis in patients with RA and controls. Contemporaneous levels of serum cytokines were measured by immunoassay. Induced gene expression was measured in cultured CD4+T cells by quantitative real-time PCR.

Results: Among circulating lymphocytes of 187 patients with early arthritis, constitutive pSTAT3 correlated with serum IL-6 levels maximally in CD4+ T cells. Increased constitutive pSTAT3, but not pSTAT1, was observed in circulating CD4+ T cells of patients with early anticitrullinated peptide autoantibody (ACPA)-negative RA compared with disease controls, and these levels decreased alongside markers of disease activity with IL-6R-targeted treatment. Among patients presenting with seronegative undifferentiated arthritis (UA) the ratio of constitutive pSTAT3:pSTAT1 in CD4+ T cells contributed substantially to an algorithm for predicting progression to classifiable RA during a median of 20 months follow-up (area under receiver operator characteristic curve=0.84; p<0.001).

Conclusions: Our findings support a particular role for IL-6-driven CD4+ T cell activation via STAT3 during the induction of RA, particularly as a feature of ACPA-negative disease. CD4+ T cell pSTAT measurements show promise as biomarkers of UA-RA progression and now require independent validation.

Keywords: Cytokines; Early Rheumatoid Arthritis; T Cells.

Figure 1

(A–D) CD4+ T cells isolated from whole blood of eight healthy volunteers were cultured with 50 ng/mL final concentration recombinant interleukin (IL)-6 and an equimolar concentration of soluble IL-6 receptor (sIL-6R). Intracellular phosphorylated signal transduction and activator of transcription-3 (pSTAT3) expression was measured by flow cytometric analysis from one aliquot (A), with induced expression of early rheumatoid arthritis (RA) ‘signature’ STAT3 target genes SOCS3, BCL3 and SBNO2 being measured in another aliquot for each individual (B–D); preliminary work established that peak expression was reached at 15 and 60 min, depicted here for pSTAT and transcripts, respectively. (E) pSTAT3 was measured in circulating CD3+T cells of patients with early RA, established RA counterparts and healthy donors as described; additional clinical information supplied in the online supplementary table S1. (F) Constitutive CD3+ T cell pSTAT3 was measured serially in 11 newly diagnosed patients with RA before (pre) and 2–4 months (median 12 weeks) post initiation of disease-modifying antirheumatic drug therapy. Wilcoxon signed-rank test (A–D and F) or non-parametric analysis of variance with Dunn's posthoc pairwise analyses (E): *p<0.05, **p <0.01 and ***p<0.001. Error bars indicate interquartile ranges. MFI, median fluorescence intensity.

Figure 2

(A) Constitutive phosphorylated signal transduction and activator of transcription-3 (pSTAT3) expression is maximal in CD4+ T cells compared with CD8+ and CD19+ cells in the peripheral blood of patients with early arthritis, and (B) a similar pattern is seen for pSTAT1. (C) Relationship between constitutive pSTAT3 and paired serum interleukin (IL)-6 is strongest in the CD4+ T cell lymphocyte population; significant differences are seen between best-fit line slopes (analyses of covariance (ANCOVA) p<0.001). (D) Comparable strength of association between constitutive pSTAT3 in CD4+ T cells and paired circulating IL-6 levels in patients with early rheumatoid arthritis (RA; black) and a disease control group comprising non-RA inflammatory arthritis (grey); no difference in the slope of the best-fit lines is seen between the comparator groups (ANCOVA). (E) Serum IL-6 does not correlate with basal pSTAT1 in any lymphocyte subset. (F) Surface IL-6R expression is maximal in the CD4+ T lymphocyte subset. ***p<0.001, non-parametric analysis of variance with Dunn's posthoc pairwise analyses; Spearman correlation coefficients (rho) and associated p values are depicted. Correlation plots are also presented individually in the online supplementary figure S1, for further clarity. MFI, median fluorescence intensity.

Figure 3

Longitudinal analysis of total CD4+ phosphorylated signal transduction and activator of transcription-3 (pSTAT3) (A), disease activity score in 28 joints (DAS28) (B) and C reactive protein (CRP; C) in four patients with established rheumatoid arthritis. Measurements were made immediately before, and 1 month after, initial treatments with tocilizumab. Additional clinical information is available in the online supplementary table S4. MFI, median fluorescence intensity.

Figure 4

(A) Representative flow cytometry histograms depicting basal and interleukin (IL)-6-induced phosphorylated signal transduction and activator of transcription-3 (pSTAT3) in CD4+ T cells from exemplar patient with non-rheumatoid arthritis (non-RA) inflammatory arthritis disease control (shaded plots) and anticitrullinated peptide autoantibody (ACPA)-negative RA (non-shaded plots); constitutive pSTAT3 is greater in the patient with ACPA-negative RA (dotted lines depict fluorescence-minus-one control). (B) Constitutive CD4+ T cell pSTAT3 expression differs according to diagnosis, being significantly higher in ACPA-negative RA than in inflammatory/non-inflammatory control groups. (C) The opposite relationship with diagnostic outcome is seen in respect of ex vivo IL-6-induced CD4+ T cell pSTAT3 fold induction. (D) Serum IL-6 concentrations differed according to diagnosis, mirroring CD4+ T cell STAT3 pattern (compare B). (E) Differences in constitutive pSTAT3 between disease phenotypes were not explained by differential CD4+ T cell membrane expression of IL-6R. *p<0.05, **p<0.01 and ***p<0.001 (Dunn's posthoc pairwise analysis following non-parametric analysis of variance); MFI, median fluorescence intensity; OA, osteoarthritis.

Figure 5

Phosphorylated signal transduction and activator of transcription (pSTAT) measurements in circulating CD4+ T cells have potential utility in predicting evolution of rheumatoid arthritis (RA) in 32 Newcastle Early Arthritis Clinic patients presenting with undifferentiated arthritis (UA). (A) Baseline serum interleukin (IL)-6 does not differ according to diagnostic outcome. (B) Constitutive CD4+ T cell pSTAT3 at presentation is highest in patients with UA who progress to RA. (C) This is also true regarding the ratio of constitutive CD4+ T cell pSTAT3:pSTAT1, suggesting a potentially robust biomarker. *p<0.05 and **p<0.01 (Mann–Whitney U test). (D) Receiver operator characteristic curves depicting compared discriminatory utilities with respect to diagnostic outcome. The Leiden prediction rule has modest discriminatory utility (area under the curve (AUC)=0.67; 95% CI 0.47 to 0.87, light grey line) and that of constitutive pSTAT3:pSTAT1 ratio in circulating CD4+ T cells is, by comparison, promising (AUC=0.78; 95% CI 0.60 to 0.96, dark grey line). However, only a composite prediction metric that excludes anticitrullinated peptide autoantibody and instead incorporates a measure of constitutive CD4+ T cell pSTAT3:pSTAT1 demonstrates a predictive utility that is significantly enhanced over that of the Leiden rule alone (AUC=0.84; 95% CI 0.68 to 1.0, black line); †difference between AUCs (Leiden prediction rule vs composite score)=0.17; 95% CI 0.01 to 0.33; p=0.04; see text. IA, inflammatory arthritis; MFI, median fluorescence intensity.