PROJECT SUMMARY There is an urgent need to develop non-sputum biomarker-based triage and diagnostic tests for childhood tuberculosis (TB). This is particularly important for young children with HIV infection, who have high TB-related mortality but often cannot produce sputum and have lower sputum bacillary burden. Biomarker discovery for childhood TB requires ultra-sensitive platforms to measure low-abundant Mycobacterium tuberculosis (Mtb) proteins in clinical samples and greater investigation of host proteins, post-translational modifications of proteins, and metabolites that are more likely than upstream RNA expression to reflect the host-pathogen interactions that lead to TB disease. The overall objective of the proposed project is to identify Mtb- and/or host-derived biosignatures in children that can achieve World Health Organization (WHO) target product profile (TPP) accuracy thresholds for a non-sputum biomarker-based triage or diagnostic test for childhood TB. We hypothesize that biosignatures that combine Mtb proteins and host biomarkers with evidence of functional relevance to TB pathogenesis or immunity will have the best diagnostic performance. To assess this hypothesis, we will conduct biomarker discovery and initial clinical validation studies using samples from three well- characterized pediatric TB cohorts in Uganda, The Gambia and South Africa. In Aim 1, we will use an ultra- sensitive electrochemoluminescence (ECL)-based immunoassay to assess the presence of Mtb proteins ESAT- 6, CFP-10, MPT64, MPT32, and Ag85B in a discovery set of banked blood and urine samples from 100 children under 5 years old with confirmed TB and 200 with unlikely TB per NIH consensus definitions (50% HIV prevalence in both groups). In Aim 2, we will use the same discovery set to perform targeted and untargeted mass spectrometry with functional assessment through pathway analysis, in vitro models and in vivo mouse models to identify host proteins, post-translational modifications and metabolites that distinguish children with confirmed versus unlikely TB. In Aim 3, we will use the candidate Mtb and host biomarkers identified in Aims 1 and 2 to derive biosignatures with up to 10 analytes consisting of Mtb proteins only, host biomarkers only, and both Mtb- and host-derived biomarkers. Biosignatures that meet WHO TPP criteria in the discovery set will 1) be evaluated in an independent test set of banked samples from 300 children under 5 years old (100 with confirmed TB, 200 with unlikely TB; 50% HIV prevalence in each group) to verify diagnostic accuracy and establish cut- offs and 2) be evaluated in a prospective cohort of 350 children under 5 years old using the pre-select cut-offs and both microbiological and clinical reference standards. Completion of these aims will result in identification of promising biosignatures that can be further validated in large-scale field studies and translated into point-of-care triage and/or diagnostic tests for childhood TB.

PROJECT NARRATIVE The inability to quickly and accurately diagnose tuberculosis (TB) in children and initiate treatment is a major contributor to the high mortality of childhood TB worldwide. This study will identify novel Mtb- and host-derived biomarkers, and determine the best biomarker panels (i.e., biosignatures) that can meet the World Health Organization-recommended diagnostic accuracy thresholds for a TB triage or diagnostic test. If successful, the results will the first step needed to make progress toward addressing the critical unmet need for non-sputum biomarker-based tests for childhood TB.