Influenza Modeling of Correlates of Protection for Optimal Immune Dynamics and Evolution.

PROJECT SUMMARY Despite its high impact worldwide due to its significant disease burden and mortality, we still do not fully understand the molecular basis of host immune responses that lead to optimal protection against Influenza virus disease. Recent evidence with pandemic viruses such as SARS-CoV-2 and the 2009 H1N1 Influenza, has shown that new improved correlates of protection are needed to inform and guide vaccine development beyond our reliance in the induction of neutralizing antibodies. We will leverage and mine all publicly available data and our extensive multi-Omic dataset already available for 280 influenza infected individuals from the CHILE cohort (Training cohort), which include severe and non-severe individuals, with diverse comorbidities and vaccination status. Through the expertise of the investigators of the FLU-CODE consortium, we will conduct complementary longitudinal Omic and immunological readouts of the innate, adaptive and humoral immunity, to develop sparse machine learning models and use deep-learning models to identify host immune response markers mechanistically implicated in immune protection. We hypothesize that the integration and modeling of longitudinal host innate and adaptive immune responses after natural influenza virus infection and vaccination assessed though systems biology, will allow the identification of early immune signatures associated with improved outcome. We will use Omic datasets obtained during acute infection (days 1-14), convalescence (days 15-30) and, through a follow-up longitudinal cohort (Test Cohort) we will analyze samples after recovery (at 3, 6 and 12 months). These will be combined with analyses of the metabolome, and the innate and the B- and CD4+ and CD8+ T-cell responses, as well as humoral effector antibody responses, to identify novel correlates of protection beyond neutralizing antibodies. We will integrate our datasets with publicly available Omic data to enrich our deep-learning models to identify candidates genes and pathways involved protective responses. We will use mouse model systems of severe and non-severe infection, and vaccination/challenged mouse models that recapitulate human disease and immune status, respectively, to validate in vivo the most prominent immunological features identified by our modeling approach. The data generated will be further integrated by our modeling approach, to further fine-tune the model and identify early signature/biomarkers of immune protection. Our systems biology approach will deconvolute the contribution of the different arms of the immune responses to uncover novel of correlates of protection and immune persistence. The datasets generated by the FLU-CODE consortium will provide an unprecedented detailed mechanistical assessment of host immune responses to inform the development of improved vaccines against IV disease.