To determine, at high resolution, the diversity of cell states found in cerebrospinal fluid (CSF) of patients with acute, non-infectious inflammatory syndromes using single cell-RNA sequencing (RNA-seq). We use this population structure to refine our understanding of the mechanism of multiple sclerosis (MS) onset.

Single cell RNA sequencing facilitates the discovery of rare cell types and context-specific cell states that have not been appreciated either by using bulk profiles of CSF cells, where individual profiles are averaged, or by using cytometric data, where only a limited set of selected markers are measured.

We successfully generated single-cell RNA-seq data from fresh CSF samples obtained from untreated individuals with optic neuritis, new onset relapsing-remitting MS, and tumefactive MS. Each cell has a transcriptome with an average of ~1200 RNA transcripts, which were used to empirically assemble clusters of cells with similar expression patterns. A reference atlas of immune cell gene expression is then used to annotate each cluster and to identify known cell populations.

Clustering analyses identify the expected types of CSF cells, with multiple broad myeloid classes as well as B- and T-cells. T cells predominate as anticipated, and we find substantial heterogeneity among T-cell subtypes, including combinations of genes encoding for surface proteins, transcription factors, and cell-cell signaling factors. We find differential representation of T-cell transcriptomic profiles in different donors, particularly in our case of tumefactive MS. We note the presence of an intriguing novel myeloid cell population with a mixture of monocytic and microglial features.

We identified new cell populations such as a mixed monocyte/microglial cell type whose role needs to be better defined, and it identifies key subtypes of CD4⁺ and CD8⁺ T cells as preferentially targeted by genetic variation that contributes to MS onset, prioritizing their modulation in the development of primary prevention strategies for MS.