Proteome-wide autoantibody screening and holistic autoantigenomic analysis unveil COVID-19 signature of autoantibody landscape

This study presents “aUToAntiBody Comprehensive Database (UT-ABCD)”, a comprehensive catalog of autoantibody profiles in 284 human individuals. The subjects include patients diagnosed with Coronavirus disease 2019 (COVID-19; n = 73), systemic sclerosis (SSc; n = 32), systemic lupus erythematosus (SLE; n = 60), anti-neutrophil cytoplasmic antibody-associated vasculitis (AAV; n = 29), atopic dermatitis (AD; n = 26), as well as healthy controls (HC; n = 64). Our investigation employs proteome-wide autoantibody screening (PWAbS) that utilizes 13,352 autoantigens displayed on wet protein arrays (WPAs). Our WPAs display human proteins synthesized in vitro utilizing a wheat germ cell-free system, maintained in a hydrated state. Our findings demonstrated significant increases in the number of IgG autoantibodies in COVID-19, SSc, SLE, AAV, and AD compared to HCs, whereas the sum of IgG autoantibody levels was significantly higher in COVID-19, SSc, and SLE. Employing machine learning, we distinguished COVID-19 cases with high accuracy based on autoantibody profiles, notably identifying antibodies against proteins encoded by BCORP1 and KAT2A as highly specific to COVID-19 (specificity: 87% and 97%, respectively). Our research highlights the effectiveness of integrating PWAbS and autoantigenomics in exploring immune responses in COVID-19 and other diseases. Although external validation and functional assays were not performed, the identified reactivities represent serological biomarkers and immune signatures associated with COVID-19 and provide a comprehensive resource for future biomarker discovery.