ObjectivePathogenesis of antiphospholipid syndrome (APS) isn't fully elucidated. We aimed to identify gene signatures characterizing thrombotic primary APS (thrPAPS) and subgroups at high risk for worse outcomes. MethodsWe performed whole blood next-generation RNA-sequencing in 62 patients with thrPAPS and 29 age-/sex-matched healthy controls (HCs), followed by differential gene expression analysis (DGEA) and enrichment analysis. We trained models on transcriptomics data using machine learning. ResultsDGEA of 12.306 genes revealed 34 deregulated genes in thrPAPS versus HCs; 33 were upregulated by at least 2-fold, and 14/33 were type I and II interferon-regulated genes (IRGs) as determined by interferome database. Machine learning applied to deregulated genes returned 79% accuracy to discriminate thrPAPS from HCs, which increased to 82% when only the most informative IRGs were analyzed. Comparison of thrPAPS subgroups versus HCs showed an increased presence of IRGs among upregulated genes in venous thrombosis (21/23, 91%), triple-antiphospholipid antibody (aPL) positive (30/50, 60%), and recurrent thrombosis (19/42, 45%) subgroups. Enrichment analysis of upregulated genes in triple-aPL positive patients revealed terms related to ‘type I interferon signaling pathway’ and ‘innate immune response’. DGEA among thrPAPS subgroups revealed upregulated genes, including IRGs, in patients with venous versus arterial thrombosis (n = 11, 9 IRGs), triple-aPL versus non-triple aPL (n = 10, 9 IRGs), and recurrent versus non-recurrent thrombosis (n = 10, 3 IRGs). ConclusionUpregulated IRGs may better discriminate thrPAPS from HCs than all deregulated genes in peripheral blood. Taken together with DGEA data, IRGs are highly expressed in thrPAPS and high-risk subgroups of triple-aPL and recurrent thrombosis, with potential treatment implications.
Read full abstract