TBX21+CD21loCXCR3+B cell expansion with an EBV-associated phenotype in clinically isolated syndrome conversion to multiple sclerosis

Epstein-Barr virus (EBV) infection precedes multiple sclerosis (MS) onset and plays an etiologic role in the disease, although the nature of viral involvement is poorly understood. A common early event in the course of MS is Clinically Isolated Syndrome (CIS), a neuroinflammatory demyelinating condition for which 60-80% of cases progress to relapsing-remitting MS (RRMS). To investigate possible viral pathogenesis during early disease stages, we analyzed single-cell gene and surface protein expression of CIS peripheral B cells collected longitudinally during the Immune Tolerance Network (ITN) STAyCIS Trial. CIS B cell profiles were compared to publicly available scRNA-seq data from in vitro EBV infection, other autoimmune disorders, chronic infectious diseases, and healthy controls. Analyses focused on CD19+/CD20+/CD21lo/CD11c+/T-bet+ atypical B cells (ABCs), which are expanded pathogenic effectors in MS and other autoimmune and chronic infectious diseases. CIS PBMCs contained significantly higher frequencies of ABCs than healthy adult controls by scRNA-seq and flow cytometry, establishing ABC compartment expansion as a clinical feature of CIS. An EBV-associated ABC transcriptome, including expression of CXCR3 and immune checkpoint molecules (PD-L1, PD-L2), was enriched in CIS versus controls; however, evidence of direct EBV infection within the ABC niche was infrequently observed. CIS ABCs exhibited significant enrichment of inflammatory cytokine mRNAs including CXCL8 (IL8), IL18, and VEGFA. Several of these cytokines (IL-8, VEGF) were secreted by B cells upon de novo EBV exposure. CIS outcome stratification revealed a rare yet distinctive inflammatory ABC signature significantly underrepresented in long-term non-progressors (LTNP) versus people with primary endpoint RRMS activity. Moreover, CXCR3+ ABCs increased after baseline CIS diagnosis and were significantly enriched in MS vs LTNP outcomes after controlling for age and sex. Thus, we find evidence for ABC expansion, inflammatory responses, and checkpoint that precede MS onset, possibly as a bystander response to EBV infection. Cryopreserved PBMCs from two timepoints for people diagnosed with clinically isolated syndrome (CIS; n = 16 individuals x 2 timepoints = 32 samples) were thawed and prepared as single-cell RNA-seq + CITE-seq libraries using 10x Genomics kits, protocols, and instrumentation. The CITE-seq panel consisted of 18 antibodies with unique antibody-derived tags (ADTs) against the following surface molecules: CD3, CD4, CD8, CD11b, CD11c, CD14, CD16, CD19, CD25, CD27, CD45, CD56, CD64, CD196, CD294, CD366, KLRG1, and TCRVa24-JA18. After ADT labeling, scRNA-seq libraries were prepared with the 10x Genomics Chromium Controller and Single-Cell Gene Expression kit with v3.1 chemistry (5' v1 single index). Cells were normalized to 1000 cells per microliter, viability assessed, and titered to a maximum of 10,000 cells per library. Cells were resuspended in master mix containing reverse transcription reagents and combined with gel beads carrying the sequencing primers, barcodes, unique molecular identifier, and a poly-dT primer for RT. Full length cDNAs were cleaned and assayed to ensure lengths between 200-5000bp. Enzymatic fragmentation of the cDNA was used prior to adapter and sample index ligation; TruSeq read 2 primers were added via End Repair, A-tailing, Adaptor Ligation, and PCR. qPCR will be used to assess P5 and P7 adapter ligation, prior to size assessment of between 400-500bp. We generated 150bp paired end sequencing reads on an Illumina NovaSeq6000 (read 1=26 bp, read 2=91 bp, index read=8 bp) at a target of 50,000 reads/cell for gene expression and 5,000 reads/cell for ADT libraries. Output base calls (.bcl format) from sequencing were used to assemble single-cell RNA and ADT reads. Note: after processing read count data using Seurat pipelines, identified B cell clusters were extracted and subsequently analyzed. Base calls and FASTQ files contain all collected PBMC data. Samples are named by outcome during trial and follow-up period (MS = received MS diagnosis; LTNP = Long-Term Non-Progressor) and collection time (t1, t2). Cell line field includes time details (t1 = baseline CIS diagnosis; t2 = 3, 6, or 12 month follow-up). Genotype field includes demographic information (NHW = non-Hispanic White; HW = Hispanic White; AI = American Indian). Treatment field includes clinical trail arm information (placebo or atorvastatin treatment).