Identification of transcriptomic and proteomic biomarkers for TB and evaluation of point-of-care tests for diagnosis and treatment response

Introduction: Tuberculosis (TB) remains a major global health challenge, partly driven by the limitations of current diagnostic and treatment monitoring tools, and an incomplete understanding of immune correlates of protection. Standard diagnostic tools, culture and molecular tests are often inaccessible in resource-constrained settings and rely on quality sputum specimens, which are difficult to obtain from children, people living with HIV, and individuals with extrapulmonary TB. Furthermore, gaps in understanding of the immune mechanisms that confer protection limit the development and evaluation of effective vaccines. This thesis addresses two key gaps: (1) the need for non-sputum-based diagnostic tools, and (2) the identification of immune correlates of early protective immunity. Methods: This research was conducted in two complementary arms. In the first, the diagnostic performance of two fingerstick blood-based point-of-care tests were evaluated: the Cepheid MTB-HR assay, which measures three host genes (GBP5, DUSP3 and KLF2), and the multi-biomarker test (MBT), which measures three host proteins (SAA1, CRP, and IP-10). These tests were assessed in adolescents and adults presenting with symptoms suggestive of active pulmonary TB. In the second arm, discovery-based approaches were used to identify host biomarkers relevant to TB diagnosis, treatment response, and immune protection. Targeted transcriptomic profiling was performed on PAXgene blood samples from microbiologically confirmed TB patients during treatment, while cytokine biosignatures were analyzed in serum from adolescent and adult patients with presumptive TB. In addition, single-cell transcriptomic analysis was conducted on cryopreserved peripheral blood mononuclear cells isolated from highly exposed adolescent and adult household contacts of TB index cases who were QuantiFERON (QFT) negative at baseline and either remained negative at 6 months or converted to QFT positive. Samples were analysed at baseline, prior to any signs of Mycobacterium tuberculosis infection. Results: The MTB-HR test showed high diagnostic accuracy, with an AUC of 0.95 (95% CI 0.92-0.97), and specificity of 0.85 (95% CI 0.76-0.95) at a fixed sensitivity of 90%, meeting the minimal TPP for a triage test. Similarly, the MBT test yielded an AUC of 0.93 (95% CI: 0.87–0.98), with a specificity of 0.73 (95% CI: 0.58 – 0.90), also meeting the minimal TPP targets for a triage test. Both tests demonstrated robustness across varying bacillary loads and co-infections, and potential for treatment monitoring. Targeted RNA transcriptomic profiling identified a 7-gene signature that distinguished patients with good versus poor treatment outcomes, achieving AUCs of 0.91 (95% CI: 0.85–0.99), 0.98 (95% CI: 0.96–0.99), and 1.0 (95% CI: 0.99–1.00) at baseline, month 2, and month 6, respectively. Additionally, a 6-gene signature at month 2 differentiated fast from slow responders, with AUCs of 0.49 (95% CI: 0.33–0.64), 0.58 (95% CI: 0.07–1.00), and 0.93 (95% CI: 0.78–1.00) at diagnosis, week 2, and month 2, respectively. Serum multiplexed cytokine analysis identified a biosignature of MIG, IL-2Ra, and M-CSF with an AUC of 0.93 (95% CI: 0.89–0.96), and specificity of 82% (95% CI: 74–90) at a fixed sensitivity of 90%, meeting the target product profile for a non-sputum-based screening test. All three biomarkers declined significantly during treatment, supporting their potential use in treatment monitoring. Single-cell transcriptomic analysis revealed no major differences in the abundance of immune cell subsets between resisters and converters at baseline. However, distinct transcriptional programs were observed: resisters showed enrichment in inflammatory pathways, particularly TNF signaling via NF-κB and enhanced immune suggesting heightened immune activation and responsiveness, while converters exhibited upregulation of type I and II interferon pathways. The enrichment of cytotoxic and inflammatory signatures in resisters indicates potential immune-mediated protective mechanisms. Conclusions: This study highlights the diagnostic potential of non-sputum-based tests for TB diagnosis and treatment monitoring and reveals novel immune pathways associated with natural resistance to TB. Together, these findings support the development of improved diagnostics, host-directed therapies, and more effective vaccines for TB prevention and control.