Abstract – Project 1 Human Mtb infection results in a large variety of clinical outcomes, ranging from bacterial eradication, to control and latent infection, to progression and active disease with a range of clinical phenotypes. We recently discovered a blood transcriptional signature that predicts TB risk in Mtb-exposed individuals up to 18 months before they exhibit clinical symptoms, a landmark contribution to the field. Still, the mechanisms that underlie TB disease progression remain poorly understood, in large part because the key immune responses within the human lung cannot be readily monitored. Furthermore, TB is a highly heterogeneous disease in which individuals progress to active disease due to a variety of mechanisms. In this project, we will conduct a comprehensive, multi-scale integration of transcriptomic, cytokine, chemokine and eicosanoid profiles from lung and blood during Mtb infection in order to identify and model molecular mechanisms and pathways that determine the outcome of infection. First, we will use multiple experimental strategies to recapitulate the heterogeneity of human Mtb infection in the mouse. These include a novel “ultra low dose” (ULD) infection model that we have pioneered in which mice are infected with 1-3 bacteria and subsequently exhibit a broad range of outcomes, ranging from immune control to progression. We will also employ mice from the Collaborative Cross project that have demonstrated extreme TB phenotypes and Mtb strains that span a range of pathogenicity. Second, we will interrogate and model the host-Mtb interaction in these mouse models using a variety of systems biology approaches in order to uncover the molecular regulators, pathways, and networks in pulmonary innate and adaptive immune cells. We will test the predicted role of critical regulatory molecules by genetically perturbing them in vivo and examining the impact on control of Mtb infection. We will also apply machine-learning approaches to define multi-omic blood based signatures in mice that predict TB progression. In our preliminary work, we have defined an early blood-based signature that predicts the late-time bacterial burdens in ULD-infected mice. We will correlate this signature with systems-level measurements of immune function in the lung to uncover mechanisms of Mtb control. Third, we will translate these findings to human disease. Through the Africa Health Research Institute, we will leverage a large-scale program that will obtain genomic sequence data as well as associated epidemiological and clinical metadata on 50,000 individuals living in a TB-endemic region. We will conduct a candidate gene genetic association analysis to validate regulatory molecules identified in mice to determine whether mutations in human orthologs are associated with altered risk of TB. In addition, we will use several existing non-human primate and human datasets to refine the blood based multi-omic progression signatures defined in mice and test their ability to predict TB progression in humans.