Background/Purpose: Type 1 interferons, such as interferon-α, are of increasing interest in autoimmunity due to their pleiotropic effects on the immune system.  Approximately 20-30% of established rheumatoid arthritis (RA) patients have a positive interferon gene signature (IGS) suggesting active type 1 interferon signalling,  While this can predict a poor response to biological therapies, such as rituximab, it has not been shown to correlate with disease activity. Since some treatments, such as steroids, can modify the IGS we wished to examine the prevalence and prognostic value of the IGS in early drug naïve RA.

Methods: 50 DMARD and glucocorticoid naïve early RA patients fulling ACR/EULAR (2010) criteria for RA, 23 established RA patients, 23 SLE patients and 23 healthy controls had whole blood RNA isolated (Tempus, ThermoFisher) and expression of 5 interferon signature genes quantified (MxA, IFI6, OAS1, ISG15, IFI44L) by RT-PCR (Taqman, ThermoFisher) using the Roche Universal probe library. The mean expression of these genes was termed the IGS score and defined as positive if was ≥ 2 standard deviations above that seen in the healthy population. EULAR responses were recorded after 3-6 months on standard DMARD therapy and some early RA patients also had their IGS repeated 1 and 3 months following diagnosis. Statistical analysis included ordinal and nominal logistic regression, multiple regression analysis, Wilcoxon signed rank, Mann-Whitney U and Chi² tests (GraphPad Prism v 5.0, San Diego USA and JMP Statistical Visualization Software v 11, SAS Inc, NC).  Significance was defined when p<0.05.

Results:   Twice as many early RA patients as established RA patients exhibited an IGS (42% vs 21%). There was significantly upregulated expression of interferon signature genes in early RA compared with established RA and were at similar levels to those reached in SLE. We observed a significant fall in early RA IGS expression between baseline and 3 months.  The baseline IGS score significantly associated with DAS-28 at baseline and 6 months. This association was mainly driven by the CRP and SJC components.  The IGS also predicted poorer EULAR response to initial therapies and increased glucocorticoid requirements at both 3 and 6 months (p all <0.05). See table 1 for summary of prognostic findings.

  Table 1: Baseline IGS score significantly predicts a poorer response to initial therapy in a range of clinical outcome measures.

Conclusion: We demonstrate an increased IGS in early RA. This mirrors other autoimmune conditions, such as primary sjogrens syndrome and type 1 diabetes where early exposure to type 1 interferons is key in disease pathogenesis. For the first time in RA we show that the IGS correlates with disease activity and in fact has prognostic value in early disease. This could inform management and therapeutic stratification, particularly with regard to therapies that target type 1 interferons and related pathways.