Deep immune profiling delineates hallmarks of disease heterogeneity in extrapulmonary tuberculosis.

Our understanding of the immune response in tuberculosis (TB) remains incomplete. This applies in particular to extrapulmonary TB (EPTB), a highly heterogeneous disease affecting up to 30% of patients in certain regions. Based on data-driven clustering of blood transcriptomes in an EPTB patient cohort, we define three highly distinct immunotypes. Combining bulk with single-cell RNA-sequencing delineates immunological trajectories characterized by dynamic IFN- and IL-1-mediated signalling in monocytes, alongside hyperactivation of T and NK cells, ultimately resulting in extensive immune dysregulation. Integrative analysis of multi-omics data provides deep insights into different layers of the anti-tuberculous immune response and the identification of immunotypes enabling stratification strategies for personalized host-directed treatments. In addition, our comprehensive approach helps to develop an accurate diagnostic gene expression signature for both EPTB and pulmonary TB highlighting the translational potential of our data.