Summary
An observational study of people with relapsing-remitting multiple sclerosis found Epstein–Barr virus-associated immune signals and expansion of atypical B cells weeks to months before clinical relapse. The findings identify a possible peripheral precursor of relapse, but the connection to inflammation in the brain and spinal cord remains unresolved.
A study in Nature Medicine has identified a reproducible pattern of Epstein–Barr virus (EBV)-associated immune activation in the blood before relapse in people with relapsing-remitting multiple sclerosis (MS). The changes appeared up to three months before clinical symptoms and were concentrated in B cells and monocytes, with the strongest signals involving CD11c-positive atypical B-cell populations.
The observational study used single-cell RNA sequencing, bulk transcriptomics, flow cytometry and targeted viral RNA testing across longitudinal blood samples from patients enrolled in the CLIMB MS cohort. Its findings support a model in which episodic EBV activity may help prime a genetically susceptible immune system before relapse.
A blood signature before symptoms
The researchers analysed samples from 114 people with MS and 21 healthy controls. The initial single-cell analysis included 15 people with MS and 21 controls, producing a map of 281,799 high-quality peripheral blood cells across remission, pre-relapse and relapse timepoints. Relapses were confirmed by two neurologists and required a contrast-enhancing lesion on MRI; remission samples came from periods of clinical and radiographic stability.
Compared with remission, the pre-relapse samples showed broad transcriptional changes across immune-cell populations. Monocytes, activated conventional dendritic cells and cytotoxic γδ T cells were among the most perturbed populations, while several B-cell subsets showed both strong transcriptional changes and enrichment for MS genetic-risk programs.
The most prominent B-cell pattern involved atypical or ABC-like cells. These are antigen-experienced B cells characterised here by surface markers including CD11c; the study used the term “ABC-like” because the flow-cytometry panel did not measure the intracellular transcription factors T-bet and ZEB2 needed to confirm a canonical ABC identity.
In 23 paired patients, ABC-like B cells increased before relapse in 21 cases, with a median 1.6-fold increase compared with remission. The cells also showed interferon, antigen-presentation and viral-defence signatures. The study found that ABC and activated memory B-cell states combined three features: pre-relapse transcriptional disruption, enrichment for MS genetic risk and overlap with gene-regulatory programs associated with the EBV protein EBNA-2.
Evidence linking the pattern to EBV activity
Several measurements pointed to EBV-associated activity in the pre-relapse window. Flow cytometry detected more B cells carrying the EBV surface protein gp350 before relapse than during remission. The model-predicted frequency rose from 0.186% of CD19-positive B cells during remission to 0.263% before relapse. In samples collected within 90 days of relapse onset, the increase was stronger.
The gp350-positive cells were concentrated in activated memory and ABC-like B-cell groups. They remained a small fraction of circulating B cells, but their frequency increased alongside the broader ABC-like population.
The researchers also sorted CD19-positive B cells from 30 patients for targeted reverse-transcription quantitative PCR; 25 paired sample sets passed quality control. EBV LMP-1 transcripts were higher before relapse than during remission, with an odds ratio of 1.50 and a 95% confidence interval of 1.26–1.93. LMP-1 is associated with EBV latency and lytic activity and can influence B-cell survival, antigen presentation and inflammatory signalling.
An independent bulk RNA-sequencing analysis of 130 samples from 70 patients found stronger host-gene responses associated with LMP-1 in samples collected 0–90 days before relapse. The pattern was enriched for early-lytic and latent EBV-associated programs, while the study found no significant enrichment for true late-lytic modules that depend on viral DNA replication. The authors interpret this combination as consistent mainly with abortive EBV reactivation, although a minority of cells may complete the viral cycle or retain late-lytic protein.
A possible pathway to relapse
The study proposes that EBV activity in latently infected memory B cells could activate B-cell survival and inflammatory pathways, favour expansion of MS-risk-enriched ABC-like cells and stimulate bystander monocytes through interferons and other cytokines. These peripheral changes could then increase the likelihood of immune-cell recruitment into the central nervous system.
That sequence remains a working model rather than a demonstrated mechanism. The study measured peripheral blood, so the link between the pre-relapse immune state and the formation of brain or spinal-cord lesions still needs to be mapped. Because the design was observational, EBV activity could be an initiating event, an amplifier of another immune trigger or a response to early disease activity that precedes detectable clinical symptoms.
The findings identify ABC-like cell abundance, gp350 positivity and LMP-1-related transcriptional responses as candidate biomarkers for future testing alongside MRI and serum neurofilament light chain. Prospective studies will need to determine their sensitivity, specificity and added value before they can support relapse prediction or treatment decisions. The results also point to EBV-responsive B-cell signalling, including the LMP-1-associated non-canonical NF-κB pathway, as a possible target for mechanistic research.