TB meningitis is the second most common cause of meningitis in Sub-Saharan Africa, and overall is estimated to account for 1-5% of TB manifestations. This project uses molecular, metagenomic and bioinformatic technologies to better diagnose TB meningitis and better understand its pathogenesis.
Mycobacterium tuberculosis (TB) affects >10 million people worldwide and carries devastating consequences with ~1.3 million deaths in 2016. Meningitis is the most feared complication of TB with 50-60% mortality in HIV+ persons. Unfortunately, current diagnostic assays are insensitive, and the ability to rapidly characterize the antimicrobial resistance (AMR) profile is severely limited, resulting in only 1 in 5 of the 480,000 multi-drug resistant TB patients receiving appropriate therapy in 2015. To address this critical gap, we will engage a novel combination of metagenomic next generation sequencing (mNGS) coupled with a CRISPR/Cas9- based targeted enrichment method to detect TB and AMR genes and longitudinal host transcriptional profiling to better understand TB meningitis pathogenesis. The most recent advance in TB diagnostics is the Xpert MTB/RIF Ultra, an automated nucleic acid amplification test. However, Ultra is only ~70% sensitive for probable/definite TB meningitis, and detects only one resistance target, highlighting the need for new and better diagnostics. We have pioneered cerebrospinal fluid (CSF) mNGS at the University of California San Francisco which is an unbiased tool that, in a single assay, can detect the whole range of neurologic infections. Furthermore, we incorporate a novel CRISPR/Cas9-based gene targeting strategy called FLASH (Finding Low Abundance Sequences by Hybridization) to enhance detection of low abundance TB and AMR gene targets by over 104-fold. Thus, we are well positioned to critically test this proposal's central hypotheses which are that CSF mNGS coupled with FLASH is
1) more sensitive and specific than conventional CSF TB diagnostics,
2) enables simultaneous and comprehensive detection of TB AMR genes, and
3) can be used to identify host gene expression biomarkers that discriminate between patients with variable responses to TB meningitis therapy.
We will enroll a prospective cohort of over ~1,300 adults presenting with suspected CNS infection in Kampala and Mbarara, Uganda and test >600 subjects with suspected TBM. The deep clinical phenotyping of this cohort will critically inform the hypotheses being tested herein. These experiments will generate the most comprehensive molecular assessment of TB meningitis in sub- Saharan Africa to date and specifically inform
1) revisions of the TB meningitis uniform case definition,
2) detection of non-TB neurologic infections masquerading as suspected TB meningitis,
3) empiric antibiotic recommendations based on AMR patterns,
4) targets for the next-generation point-of-care diagnostic assays for TB meningitis, and
5) candidate prognostic CSF transcriptional biomarkers for treatment response.