Single-cell transcriptomics identifies drivers of local inflammation in multiple sclerosis

Single-cell transcriptomics enables unbiased biological discovery and holds new promise for personalized medicine. However, its potential for understanding human diseases by comparing patient vs. control samples in a clinical setting remains largely unexplored. Here, we applied single-cell RNA-sequencing (scRNA-seq) to rare cerebrospinal fluid (CSF) specimens from well-characterized controls and patients with multiple sclerosis (MS) - a prototypic inflammatory disease of the central nervous system (CNS). We thereby generated and validated the first transcriptional atlas of single CSF leukocytes in health and disease. In MS patients, we found an expansion of natural killer cells and late B cell lineages and based on these insights we developed a score with potential diagnostic relevance. Using this analytical approach, we identified and characterized activated phenotypes of MS-derived CSF leukocytes, including an enrichment in T follicular helper (TFH) cell transcriptional signatures. We validated the expansion of such B cell-helping TFH cells in MS patients and demonstrated that TFH cells exacerbate symptoms in an animal model of MS and promote B cell infiltration of the CNS. TFH-dependent B cell expansion may thus drive local CNS autoimmunity in MS. Our study demonstrates how single-cell transcriptomics can identify novel disease mechanisms in a clinically-relevant case-control study design.