Urinary metabolomic profiling for characterization, diagnosis and treatment response in paediatric Tuberculous Meningitis

Abstract

The primary hypothesis formulated for this investigative study is: 'Can urinary metabolic profiles of paediatric cases of tuberculous meningitis (TBM) provide sufficient metabolic information to characterise and diagnose TBM, and monitor treatment response, thereafter?' Secondly, if a brain-gut axis exists, then chronic central nervous system (CNS) bacterial infection(s) should theoretically be reflected in the urine. The premise here is that chronic CNS bacterial infection(s) will affect the gut microbiome and that altered metabolism in both the CNS and the gut will release metabolites into the blood that are filtered (kidneys) and excreted in the urine. Upon extensive deliberation and contemplation regarding the elucidation of the hypothesis. Specific objectives became clear for this study: (I) Characterise the urinary metabolic profile of the patient with TBM to comprehensively understand the aetiology, pathogenesis, clinical manifestations, epidemiology, and prognosis of the disease. (II) Identify metabolite profiles that characterise the three stages of severity of TBM and potential biosignatures for early diagnosis of the disease, which may lead to a better outcome. (III) Monitor the patients' responses to treatment and identify possible prognostic markers to predict a successful treatment outcome. The structure of this thesis follows. Chapter 1 is a literature overview of existing knowledge of Mycobacterium tuberculosis (M. tb), with a focus on the brain-gut axis exposed to chronic bacterial Infections in the central nervous system (CNS), underscoring the significance of the brain-gut axis within the realm of chronic bacterial infections in the CNS with concise step-by-step pathological aspects of M. tb in term of CNS infection. A brief description of the existing literature provides a complete overview of chronic bacterial infections of the CNS with a central focus on tuberculous meningitis (TBM), as well as the impact of bacterial infection on the gut microbiome and the potential application of urinary metabolic profiling for the diagnosis of M. tb infections. Infection with M. tb leads to perturbed metabolism in both the CNS and gut and will release metabolites into the blood that are filtered (by the kidneys) and excreted in the urine. Identifying the gap in the literature: no non-invasive (urinary) method of diagnosis of TBM (problem statement), showed that the association between the brain-intestinal connection and bacterial infections of the CNS is relatively scarce in the current literature. The intricate relationships with the gut microbiota are still inadequately elucidated. Metabolomics as a scientific method of choice was selected to fill this gap. But which metabolic markers should we use for such purposes? If a brain-gut axis for CNS infection exists, then metabolic markers of such chronic CNS infection(s) should theoretically be reflected in the urine. From this, the biological research question focuses primarily on individuals with TBM, while also addressing the repercussions of bacterial infection on the gut microbiome. A prognostic metabolic model for TBM within the brain is suggested, using myeloperoxidase (MPO), interferon-gamma (IFN-γ), and vascular endothelial growth factor (VEGF), predicting related downstream metabolic indicators that can be detected in urine. Increased concentrations of MPO and IFN-γ in cerebrospinal fluid (CSF) are associated with TBM. The fundamental principles of biochemistry are used to postulate a prognostic metabolic model for TB and to investigate the potential of urinary metabolic profiling in the diagnosis and surveillance of this chronic bacterial infection. Urinary metabolic profiling could serve as a valuable tool for the detection and monitoring of CNS infections, as well as for the design of new therapeutic approaches for the treatment of TBM according to the prognostic metabolic model, especially in resource-constrained settings where access to advanced diagnostic tools can be limited. Chapter 2 describes the use of an untargeted proton nuclear magnetic resonance (1H-NMR) metabolomics approach to identify characteristic metabolic markers that differentiate severe cases of paediatric TBM from controls, characterizing the urinary metabolic profile of TBM. This is pivotal for better understanding the underlying disease mechanisms, formulating effective prevention and treatment modalities, and facilitating further research. The study compared the metabolic profiles of urine samples collected from the most severe TBM form (stage 3, n=13) to an age-matched non-neurological control group (n=44). The outcome: twenty-nine urinary metabolites were identified as candidates that characterised the advanced TBM. These metabolites were categorized into six dysregulated metabolic pathways: upregulated tryptophan catabolism, perturbation of amino acid metabolism, increased energy production, disrupted gut microbiota metabolism, ketoacidosis, and increased nitrogen excretion. This research offers novel biological insights into urinary metabolite biosignatures that can differentiate paediatric TBM patients from control subjects, serving as potential non-invasive metabolic markers for early differential diagnosis and treatment follow-up. Chapter 3 focuses on the urinary gut metabolites associated with TBM and its treatment, highlighting the potential of identifying novel biomarkers for TBM diagnosis and treatment monitoring. Using a semi targeted liquid chromatography-tandem mass spectrometry (LC-MS/MS) approach to investigate metabolites associated with gut metabolism. Urine samples from a control group (n = 40) were compared against an experimental group (n = 35) with confirmed TBM, which were subdivided into TBM stage 1 (n = 8), stage 2 (n = 11), and stage 3 (n = 16). Seventy-eight selected compounds of microbiome origin were evaluated. These metabolites were also measured at different time points during the 6-month TBM treatment period. Eight unique (statistically significant at time point 1) urinary metabolites associated with altered gut metabolism were identified in the TBM patients when compared to the control group: 2 methylbutyrlglycine, 3-hydroxypropionic acid, 3-methylcrotonylglycine, 4-hydroxyhippuric acid, isobutyrylglycine, phenylacetylglutamine, as well as 5-hydroxyindoleacetic acid (5HIAA), 5 hydroxyhexanoic acid, and methylcitric acid (identified at time points: 3, 4 and 5). These metabolites were linked to alterations in tryptophan metabolism, fatty acid metabolism, and M. tb metabolism, respectively. These metabolites showed a constant elevated level for TBM stage 1 across the entire treatment period. This highlighted the importance of gut metabolism and identifying corresponding microbial metabolites for improved management of TBM patients. The investigation also found that these metabolites are associated with an altered host metabolome, caused by M. tb, and the breakdown of M. tb cell wall components caused by anti-TB medication and/or dysbiosis commonly associated with TBM. Chapter 4 This chapter identifies and describes urinary biomarkers in the urine of patients with TBM using proton nuclear magnetic resonance (1H-NMR) metabolomics, for possible diagnostic applications. This investigation compared urine from 32 patients with TBM (stratified into stages 1, 2 and 3) against that from 39 controls and found four metabolites with good diagnostic potential for severe TBM (stage 2 and 3), but not for mild TBM (stage 1). The comparison between (stage 1) TBM and severe (stages 2 and 3) TBM was unsuccessful. Our multivariate metabolic model could successfully classify severe but not mild TBM. Five significant biological metabolites were identified: 1-methylnicotinamide, 3-hydroxyisovaleric acid, 5-aminolevulinic acid, N-acetylglutamine, and methanol, with the four of these metabolites (excluding methanol: ROC analysis revealed that methanol lacked diagnostic sensitivity) showing good diagnostic potential for severe TBM. These markers have not yet been established as definitive diagnostic tools because validation studies are needed. However, they may serve as candidate biomarkers for TBM diagnosis. Additionally, 14 unknown compounds were identified as important for differentiating TBM patients from controls, but further research is needed to validate these biomarkers and determine their clinical utility. Chapter 5, this investigation covers the semi-targeted approach using gas chromatography-mass spectrometry (1D GC-MS) for the semi-targeted metabolomics analysis of alkanes and other volatile organic compounds (VOCs) in the urine of paediatric patients with TBM. This investigation compared all volatile underivatized compounds present in the urine of 27 confirmed cases of paediatric TBM over a 6month treatment period with 13 control groups. The results showed that four alkanes (pentadecane, 5,7dimethyl-undecane, 4,7-dimethyl-undecane and 2,6-dimethyl-undecane), three alkenes (2,5-dimethyl-2hexene, 4,4-dimethyl-1-pentene and 3-methoxy-1-pentene), and three other VOCs of biological interest were identified as statistically significant. These volatile compounds remained perturbed during the TBM treatment. This new systemic metabolic information on M. tb in the host highlighted the role of alkanes and VOCs in the potential persistence of M.tb in the host. Ultimately, Chapter 6 functions as a comprehensive discourse on the collective outcomes and impacts derived from this investigation, as well as a contemplation of the objectives originally proposed before the study commenced. Subsequently, there is a succinct examination of inquiries that emerged throughout the thesis project, along with recommendations for potential avenues for advancement after this doctoral research. Furthermore, Chapter 6 critically examines the accomplishments of the dissertation within the framework of pertinent biological and clinical dimensions associated with TBM, thereby addressing the specified objectives. The questions culminate with insights regarding the constraints of the current study and potential future directions.