Decades-long elevation of interferon-alpha drives a Sjogren disease endotype: an interdisciplinary study

BACKGROUNDMechanistic heterogeneity is a major obstacle to the development of effective treatment for Sjogren disease (SjD) and there is a pressing need to stratify SjD according to precision medicine principles. Aberrant activation of the type I interferon (IFN) pathway represents a leading candidate pathway, but a causal role of elevated IFN- in driving a Sjogren disease endotype remains to be established. METHODSWe used ultrasensitive single molecule ELISA, and an oligoprotein interferon signature score (derived from broad capture proteomics), to study the role of IFN- in Sjogren disease. We analysed samples from the UK Primary Sjogren Syndrome Registry (UKPSSR, n=177) and UK Biobank Plasma Proteomics Project (n=47606 without Sjogren, n=257 with Sjogren, including 137 individuals sampled prior to diagnosis) to determine the timecourse and immune endotype associated with elevated IFN-. To address causality we created a new transgenic mouse model of IFN- overexpression to establish whether chronically elevated IFN- drives this immune endotype. FINDINGSOligoprotein interferon signatures can be detected at least 14 years prior to diagnosis of Sjogren disease in the UK Biobank-PPP. IFN- concentrations are elevated in 60% of Sjogren disease patients in the UKPSSR. Individuals with elevated IFN- display a distinct immunological endotype characterised by cytopenias, hypergammaglobulinaemia, multiple autoantibodies and autoimmunity against the Sjogren autoantigen TRIM21/Ro52. To address the key question of causal direction, we created a new mouse model of systemic chronic IFN- elevation, in which Ifn4 is overexpressed by conventional dendritic cells. This model recapitulates key features of the endotype and can be partially reversed by IFNAR1 blockade. INTERPRETATIONElevation of IFN- drives an immune endotype of Sjogren disease, originating over a decade prior to diagnosis. SjD patients with elevated IFN- concentrations are broadly clinically similar to those with normal IFN- concentrations, yet are immunologically distinct. This highlights the mechanistic heterogeneity of SjD and the need for immunological stratification along precision medicine principles, using high resolution biomarkers. As well as demonstrating causal direction, biological modelling shows that chronic IFN- elevation over the lifecourse has the potential to establish persistent immune dysregulation which responds only partially to interferon receptor blockade. These findings provide insights into SjD and other "interferonopathic" rheumatological disorders. Research in contextO_ST_ABSEvidence before this studyC_ST_ABSWe searched MEDLINE for "Sjogrens Syndrome/Disease" and "interferon", including the terms "subsets", "sub-groups", "phenotypes", and "endotypes", filtering by "clinical trial", "stratification", and "immune-mediated inflammatory". We also included major review articles from noted experts. We identified reports of associations between IFN- and SjD, usually using indirect or imprecise measures of IFN-. None of these studies included prediagnostic samples and causal inference was limited. Added value of this studyThis study shows that IFN-, when measured directly using ultrasensitive single molecule ELISA approaches (uniquely optimised to determine healthy control concentrations), is elevated in a subset of people with SjD with a specific immunological endotype. Analysis of prediagnostic proteomics shows this elevation can be detected up to 14 years before diagnosis. We show that IFN- drives this endotype (as opposed to vice versa) by recapitulating key endotype features in a novel and unambiguous experimental mouse model of chronic IFN- elevation. We also show that the pathogenic consequences of IFN- elevation over long periods of time can only be partially reversed using IFNAR blockade. Implications of all the available evidenceThese data have important implications for future research, clinical practice, trial design, and therapeutic development. First, our findings provide clinical evidence, supported by unambiguous preclinical evidence, that decades-long elevated IFN can cause and drive a SjD endotype - and accurately defines the level of heterogeneity. Secondly, we provide biomarkers which may be of use in stratifying clinical trial design and also for early identification of at-risk individuals. Thirdly we provide biological proof of principle that longstanding and potentially undiagnosed elevation of IFN- can establish persistent immune dysregulation which may respond only partially to IFNAR blockade. Together these findings inform precision medicine approaches and future trial design.