High levels of circulating interferons type I, type II and type III associate with distinct clinical features of active systemic lupus erythematosus

Abstract

Background and aim: Interferons (IFNs) are considered to be key molecules in the pathogenesis of systemic lupus erythematosus (SLE). We measured levels of type I, II and III IFNs in a large cohort of patients with systemic lupus erythematosus (SLE) and controls and explored associations among high levels of different IFN types and distinct SLE features.

Methods: Four hundred ninety-seven well-characterized SLE patients and 322 population controls were included. Disease activity was assessed by SLE Disease Activity Index (SLEDAI) and Systemic Lupus Activity Measure (SLAM). Functional type I IFN activity was estimated by a WISH reporter cell assay. Levels of IFN-γ were estimated by MSD 30-plex assay. IFN-α and IFN-λ1 were measured by ELISA. Values above the third quartile of patients' measurements were defined as high. Associations among high IFN results and SLE features were investigated by nominal regression analysis.

Results: All IFN measurements were higher in SLE patients than in controls. High type I IFN activity correlated with levels of IFN-γ and IFN-α and associated with active SLE in most domains: weight loss, fatigue, fever, rash, lymphadenopathy, arthritis, nephritis and haematological manifestations. Specific SLE subsets were linked to the upregulation of different subtypes of circulating IFNs: high IFN-γ to arthritis, nephritis and anti-Ro60 antibodies and high IFN-α to mucocutaneous engagement and anti-Ro52 and anti-La antibodies. Isolated high IFN-λ1 was coupled to anti-nucleosome antibodies and less severe SLE.

Conclusions: High functional type I IFN activity captures active SLE in most domains, but more distinct patterns of organ involvement are associated with profiles of circulating IFNs. High IFN-γ as well as high functional type I IFN activity is a characteristic of severe SLE with nephritis and arthritis, while elevated levels of IFN-α associate with active mucocutaneous inflammation and a more benign cardiovascular profile. IFN-λ1 in isolation is associated with milder disease. Our findings suggest that IFNs contribute to the heterogeneity of clinical manifestations in SLE, and measuring circulating IFNs could assist in designing clinical trials with therapies targeting IFN pathways.

Keywords: Autoantibodies; Disease activity; Interferons; SLE.

Fig. 1

Type I IFN activity and levels of IFN-α, IFN-γ and IFN-λ1 in SLE patients and population controls. Type I IFN activity in vitro (a) and IFN-γ levels (b), IFN-α (c) and IFN-λ1 (d) were all higher in SLE patients then population controls (Mann-Whitney U test). The dashed boxes indicate individuals with high levels (> 75th percentile of patient measures) of each investigated IFN

Fig. 2

The distribution of the proportions of patients with high levels of each IFN measurement. The Venn diagram depicts how patient groups with high levels of different IFNs distribute and overlap in the cohort (a), and among those with active SLE as scored by SLAM (b) and SLEDAI (c). Only patients in whom all four measurements were available are included in the analysis (n = 248). The number outside diagram indicates the number of patients who had none of the IFNs expressed at a high level (> 75th percentile of patient measures)

Fig. 3

Distribution of IFN-high subsets among SLE patients with different characteristics. The figure illustrates how subsets of different IFN types distribute among patients with active SLE manifestations (a), past manifestations and events (b), positivity for autoantibodies (c) and laboratory parameters (d) (presented in %) as assessed at inclusion. Abbreviations: LN lupus nephritis, NPSLE neuropsychiatric SLE (*at inclusion, NPSLE was classified by 1982 ACR criteria, seizures or psychosis), aCL anti-cardiolipin, B2GP1 β2glycoprotein-I, LA lupus anticoagulant, ESR erythrocyte sedimentation rate, WBC white blood cells, PLT platelets. Only patients in whom all four measurements were available are included in the analysis (n = 248)