Single-cell clues to why tuberculous meningitis is so deadly

Tuberculous meningitis (TBM) is the most severe form of tuberculosis, carrying a fatality rate of 20–50% even among treated individuals. Current supportive care can include corticosteroid therapy, which can increase survival in HIV-negative people with TBM, but clinicians and researchers agree that better antimicrobial and host-directed therapies are needed to improve outcomes. A major barrier to developing those therapies is an incomplete understanding of the immune processes that take place inside the brain and spinal fluid during infection. Single-cell approaches have the power to identify disease-specific cellular profiles, but single-cell RNA-sequencing (scRNA-seq) has been underutilized in cerebral samples in brain infection. To address this gap, a team led by corresponding author Nguyễn Thụy Thương Thương used scRNA-seq to study fresh pretreatment cerebrospinal fluid (CSF) from four TBM patients alongside paired peripheral blood mononuclear cells (PBMCs). Their goal was to map which immune cell types and pathways are present in the CSF at the start of treatment and how these differ from the circulation, providing a clearer view of local immunopathologic pathways in TBM.

The group applied single-cell RNA-sequencing (scRNA-seq) to matched samples of fresh pretreatment cerebrospinal fluid (CSF) and peripheral blood mononuclear cells (PBMCs) from four patients with TBM. Across both tissues they detected 29 cell subtypes, but the relative abundance of those subtypes differed markedly between CSF and blood. Notably, CSF was enriched for highly inflammatory microglia-like macrophages, GZMK -expressing CD8 + T cells, and CD56 bright NK cells. The study reports that the GZMK -expressing CD8 + T cells and CD56 bright NK cells showed features associated with dysfunctional cytotoxicity, suggesting these cells may be present but impaired in their ability to kill infected targets. In addition to these cell-type shifts, the investigators observed that inflammatory signaling pathways were increased across multiple cell types in CSF, while oxidative phosphorylation was decreased in CSF cells compared to PBMCs. These contrasts point to a distinct inflammatory and metabolic state in the central nervous system during TBM.

This work highlights the value of scRNA-seq for exploring CSF immunopathogenesis in TBM patients and creates a resource for future studies of TBM and other brain infections. By identifying CSF-enriched populations such as microglia-like macrophages, GZMK -expressing CD8 + T cells, and CD56 bright NK cells, the study points to specific cell populations that could be investigated as contributors to immune-mediated pathology or as potential targets for host-directed therapies. The finding that inflammatory signaling is up and oxidative phosphorylation is down in CSF cells emphasizes that local immune responses in the brain differ from those in blood and may require different therapeutic approaches. Ultimately, these single-cell profiles could guide research into why some people deteriorate despite treatment and help prioritize interventions—both antimicrobial and immunomodulatory—that specifically address the brain environment in TBM.