Defining the architecture of cerebrospinal fluid cellular communities in neuroinflammatory diseases

Abstract

Cerebrospinal fluid (CSF) biomarkers are important for multiple sclerosis (MS) diagnosis. Moreover, absent of autopsy or biopsy tissue, CSF is the most relevant source for studying the immune cells involved in MS pathophysiology. Single-cell RNA sequencing (scRNA-seq) provides new opportunities to advance our understanding of disease-associated changes in CSF immune cells. Here, using scRNA-seq data generated from 58 CSF and 10 PBMC samples, we provide an updated atlas of the immune cells present in human CSF in MS and other neuroinflammatory conditions, including novel lymphoid and myeloid cell clusters. Our atlas can thus serve as a reference for future studies of immune cells in neuroinflammation. Our further characterization of CSF myeloid cells suggests that most CSF microglia-like cells resemble two of the previously-described brain microglia signatures. Additionally, our data from a sex-mismatched bone marrow transplant recipient suggest that CSF microglia-like cells are of peripheral origin. Our comparisons between MS and other neuroinflammatory disorders show a highly-specific increase in plasma cells, along with reductions in the proportion of microglia-like cells in MS CSF. Furthermore, our analyses on MS patients receiving anti-CD20 therapy ocrelizumab suggest that the treatment effects are not limited to B cell depletion, and ocrelizumab appears to reverse some MS-associated T and myeloid changes in CSF. Finally, we utilized our atlas to prioritize (1) CSF cell types expressing genes associated with MS susceptibility, and (2) ligand-receptor gene pairs that are differentially expressed in MS CSF, providing targets for further mechanistic and causal investigations in pathophysiology and treatment of MS.

Integrated single cell transcriptomics of cerebrospinal fluid cells in early Multiple Sclerosis

“…Deeper gene analysis of the latter showed similarities with homeostatic microglia, 36 -a phenotype likely instructed by the CSF microenvironment and not indicating ontogenetically shared origins. 37 Considering these two examples of the integrative data set's advantages, the aim of this study is su ciently met: It supports the previous ndings and und substantiates them with a higher statistical power.…”

High-dimensional investigation of the cerebrospinal fluid to explore and monitor CNS immune responses

“…Next to a reduction of microglia-like cells and the well-known expansion of plasma cells and T cells in the CSF of MS patients, the authors observed a reduction of B and T cells and an increase of myeloid cells of ocrelizumab (anti-CD20) treated progressive MS patients versus therapy-naive relapsing-remitting MS patients [15]. The authors speculate that a CD27 downregulation in plasma cells in ocrelizumab-treated MS patients might mediate the immunomodulatory effect of the therapy [15]. While these studies excel through their transcriptional resolution on a single-cell level, many studies lack methodical validation on the protein levels, although some [11,14] provide flow cytometry verification.…”

“…In fact, CSF contains a monocyte subpopulation with a microglia-like phenotype [12][13][14], which is almost exclusive to the CSF [11]. This CSF microglia-like population was found to originate from the bone marrow in a single bone marrow transplant recipient patient [15], in contrast to resident microglia of the CNS, which derive from the yolk sac [16,17]. Transcriptionally CSF T cells show enhanced expression of transcripts associated with migration (CD99), interaction with antigenpresenting cells (APC) (CD83, CD84) and chemokines (CXCL16, CXCR5), while genes associated with naive cell state (SELL), cytokine response (IL2RG), and integrins (ITGAL, VLA4) are reduced in CSF T cells compared to blood T cells [11].…”

“…Additionally, the deviating findings might be due to the inherent heterogeneity of MS [45] and the small sample sizes. A recent preprint generated a large scRNA-seq CSF dataset of neuroinflammatory, mostly MS, and control patients [15]. Next to a reduction of microglia-like cells and the well-known expansion of plasma cells and T cells in the CSF of MS patients, the authors observed a reduction of B and T cells and an increase of myeloid cells of ocrelizumab (anti-CD20) treated progressive MS patients versus therapy-naive relapsing-remitting MS patients [15].…”

“…A recent preprint generated a large scRNA-seq CSF dataset of neuroinflammatory, mostly MS, and control patients [15]. Next to a reduction of microglia-like cells and the well-known expansion of plasma cells and T cells in the CSF of MS patients, the authors observed a reduction of B and T cells and an increase of myeloid cells of ocrelizumab (anti-CD20) treated progressive MS patients versus therapy-naive relapsing-remitting MS patients [15]. The authors speculate that a CD27 downregulation in plasma cells in ocrelizumab-treated MS patients might mediate the immunomodulatory effect of the therapy [15].…”