Background
One complication in treating COVID-19 is that there is substantial heterogeneity in SARS-CoV-2-specific memory immune responses following infection. There also remains a critical need to identify host immune biomarkers predictive of clinical and immunological outcomes in SARS-CoV-2-infected patients. A multiomic COVID-19 study from Stanford University followed a prospective, longitudinal outpatient trial of a type III IFN drug in individuals newly infected with SARS-CoV-2, over the course of 7 months. In this extensive investigation they looked at plasma proteomics with Olink, bulk RNAseq, viral shedding, virus-specific antibody loads and virus-specific T-cell responses in order to characterize early immune responses to infection
Outcome
This approach data identified several early immune signatures, including plasma RIG-I, levels, early IFN signaling (with IFN-responsive cytokine increases in CXCL10, MCP1, MCP-2, and MCP-3) associated with subsequent disease progression, viral shedding, and T-cell responses. Interestingly, several biomarkers for immunological outcomes were also evident both in individuals receiving the Pfizer vaccine and COVID-19 patients. Machine learning models using 2–7 plasma protein markers measured at day zero in the longitudinal analysis could accurately predict disease progression, T cell memory, and the antibody response post-infection. Most strikingly, a 4-protein signature (CXCL10, CXCL11, MCP2 and PRDX3) predicted severe disease progression in patients at the earliest stage of infection with an AUC=0.9.
