Urinary metabolomic profiling for characterization, diagnosis and treatment response in paediatric tuberculous meningitis

ABSTRACT This investigative study explored the central hypothesis: "Can urinary metabolic profiles of paediatric tuberculous meningitis (TBM) provide sufficient metabolic information to characterise and diagnose TBM, and monitor treatment response, thereafter?" It also posited that chronic central nervous system (CNS) bacterial infections would be reflected in urine due to their impact on the gut microbiome and subsequent release of metabolites into the bloodstream. The study aimed to characterise the urinary metabolic profile of TBM paediatric patients, identify biosignatures for early diagnosis and staging of TBM, and monitor treatment response for prognostic markers. Chapter 1 provides a literature review on Mycobacterium tuberculosis (M. tb), focusing on the brain-gut axis in chronic CNS bacterial infections. It highlights the lack of non-invasive diagnostic methods for TBM and the limited understanding of the brain-gut connection in this context. Metabolomics was chosen to address this gap, theorising that metabolic markers of chronic CNS infection would be detectable in urine. A prognostic metabolic model for TBM in the brain is suggested, incorporating TBM markers myeloperoxidase (MPO), interferon-gamma (IFN-γ), and vascular endothelial growth factor (VEGF), with the expectation of detecting related downstream metabolic indicators in urine. This model proposes that urinary metabolic profiling could serve as a valuable diagnostic and monitoring tool, particularly in resource-limited settings. Chapter 2 details an untargeted proton nuclear magnetic resonance (1H-NMR) metabolomics approach to characterise the urinary metabolic profile of severe paediatric TBM. By comparing urine samples between severe TBM (stage 3) patients and controls, 29 urinary metabolites were identified as characteristic of advanced TBM. These metabolites were linked to six dysregulated metabolic pathways: upregulated tryptophan catabolism, perturbed amino acid metabolism, increased energy production, disrupted gut microbiota metabolism, ketoacidosis, and increased nitrogen excretion. This research offers novel insights into potential non-invasive metabolic markers for early differential diagnosis and treatment follow-up. Chapter 3 investigates urinary gut metabolites associated with TBM and its treatment using semi-targeted liquid chromatography-tandem mass spectrometry (LC-MS/MS). Comparing TBM patients (n=35) (stratified into stages 1, 2 and 3) with controls (n=40), and monitoring during a 6-month treatment period, eight unique urinary metabolites associated with altered gut metabolism were identified. These included 2-methylbutyrlglycine, 3-hydroxypropionic acid, 3-methylcrotonylglycine, 4-hydroxyhippuric acid, isobutyrylglycine, phenylacetylglutamine, 5-hydroxyindoleacetic acid, 5-hydroxyhexanoic acid, and methylcitric acid. These metabolites were linked to alterations in tryptophan and fatty acid metabolism and M. tb metabolism. The study found that these metabolites remained elevated in TBM stage 1 throughout treatment, emphasising the importance of gut metabolism and microbial metabolites for TBM management. Chapter 4 explores urinary biomarkers for TBM diagnosis using 1H-NMR metabolomics. Comparing urine from TBM patients (stages 1, 2, and 3) with controls, four metabolites (1-methylnicotinamide, 3-hydroxyisovaleric acid, 5-aminolevulinic acid, and N-acetylglutamine) showed good diagnostic potential for severe TBM (stages 2 and 3), but not mild TBM (stage 1). While promising as candidate biomarkers, further validation studies are required. Additionally, 14 unknown compounds were identified as important for differentiating TBM patients from controls, warranting further research. Chapter 5 focuses on semi-targeted gas chromatography-mass spectrometry (1D GC-MS) analysis of alkanes and other volatile organic compounds (VOCs) in the urine of paediatric TBM patients. Comparing TBM patients with controls over a 6-month treatment period, four alkanes, three alkenes, and three other VOCs were identified as statistically significant and remained perturbed during TBM treatment. This suggests a role for alkanes and VOCs in the potential persistence of M. tb in the host. Chapter 6 provides a comprehensive discussion of the study's collective outcomes, addressing the initial objectives and emerging questions. It critically examines the dissertation's accomplishments within the biological and clinical dimensions of TBM,