Proteomic profiling of the large vessel vasculitis spectrum identifies shared signatures of innate immune activation and stromal remodelling

Abstract

AbstractTakayasu arteritis (TAK) and giant cell arteritis (GCA) are the primary forms of large vessel vasculitis (LVV) and can result in serious cardiovascular morbidity. Improved understanding of the molecular basis of these diseases is required to develop novel biomarkers and targeted treatments. Moreover, it is unclear whether shared or distinct pathogenic processes underpin the LVV spectrum. To address this, we performed plasma proteomic profiling, quantifying 184 plasma proteins using Olink immunoassays in two independent cohorts totalling 405 individuals. In Cohort 1, comparison of patients with TAK (N=96) and large vessel-GCA (LV-GCA) (N=35) versus healthy controls (HCs) (N=35) revealed 52 and 72 significant differentially abundant proteins, respectively. Correlation with disease activity status identified novel TAK and LV-GCA disease activity markers. Cohort 2 consisted of patients presenting acutely with possible cranial GCA (C-GCA); C-GCA was subsequently confirmed (n=150) or excluded (n=89). 31 proteins were associated with C-GCA. Analyses stratified by temporal artery biopsy results revealed enrichment of the proteomic signal in biopsy-proven GCA, suggesting the presence of distinct endotypes within C-GCA. Cross-disease comparison revealed that active TAK, LV-GCA and biopsy-proven C-GCA had highly similar plasma proteomic profiles. Twenty-six proteomic associations were shared across all three groups including IL6, monocyte/macrophage related proteins (CCL5, CCL7, CSF1), tissue remodelling proteins (VEGFA, TIMP1, TNC) and proteins not previously linked to LVV (TNFSF14, IL7R). We also observed disease-specific associations including increased CXCL9 in LV-GCA and C-GCA but not in TAK and decreases in the extracellular matrix protein COMP in TAK but not in LV-GCA or C-GCA. Evaluation of publicly available transcriptomic data from LV-GCA aortic tissue revealed that 47 of the 112 proteins significantly altered in ≥1 LVV type had significantly altered mRNA expression in LVV aortic tissue. Similarities in LVV proteomic profiles suggest shared pathobiology involving innate immunity, particularly monocyte/macrophages, lymphocyte homeostasis and tissue remodelling processes. Our results highlight a signature of immune-stromal cross talk in LVV and identify potential novel therapeutic targets in this axis (e.g. TNFSF14). The correspondence of plasma signatures to tissue phenotype highlights the potential for non-invasive monitoring of arterial inflammation and injury.