B-Cell Depletion Therapy Reshapes Antiviral Immunity in Multiple Sclerosis, Study Finds

Multiple sclerosis has long been treated as a disease of errant T cells attacking the nervous system, but the most effective modern therapies work by eliminating an entirely different cell type: the B cell. Antibodies that target the CD20 protein strip the blood of these immune cells and, in doing so, dramatically reduce relapses, new brain lesions and disability progression. Yet B cells are also important players in antiviral defense, and their sustained removal has raised an uncomfortable question for clinicians. Does wiping out B cells leave patients broadly defenseless against viruses, or does the therapy act with far more surgical precision than anyone assumed? A new cross-sectional study published in the Journal of Neurology by Silvia D’Orso of the Santa Lucia Foundation in Rome and colleagues suggests the answer is strikingly specific: B-cell depletion selectively contracts immunity to Epstein–Barr virus, the pathogen most strongly linked to multiple sclerosis, while leaving defenses against other viruses largely intact.

The research team recruited adults with relapsing multiple sclerosis at two neurological centers in Rome and divided them into two cohorts. The first, an immunophenotyping cohort of 45 untreated patients and 25 patients treated with the anti-CD20 antibody ocrelizumab, underwent high-parameter flow cytometry that resolved more than 40 distinct immune cell populations across every major blood lineage. The second, an antiviral cohort of 24 untreated patients, 13 ocrelizumab-treated patients and 19 healthy donors, was tested for T-cell responses against three viruses: Epstein–Barr virus (EBV), cytomegalovirus (CMV) and SARS-CoV-2. The researchers also measured cytokine and chemokine release in culture supernatants and quantified serum neurofilament light chain, or sNfL, an ultrasensitive blood marker of neuroaxonal injury, using single-molecule array technology.

The immunophenotyping results confirmed what was expected and revealed what was not. As anticipated, ocrelizumab-treated patients showed profoundly reduced total B cells and switched-memory B cells, the long-lived reservoir in which EBV establishes latency. Antibody-secreting cells, which lack CD20, were spared in absolute terms. But the therapy’s fingerprints extended well beyond the B-cell compartment. T follicular helper cells, the CD4-positive helpers that shepherd B cells through germinal center reactions, and mucosal-associated invariant T cells, an innate-like population involved in early antimicrobial and antiviral responses, both showed significantly reduced absolute counts in treated patients. The loss of T follicular helper cells is consistent with the disruption of the cognate B-cell–T-cell interactions required for their maturation, while the contraction of MAIT cells may reflect the loss of B-cell-derived survival signals such as interleukin-23, which was also reduced in treated patients.

Innate immune function told an equally nuanced story. When whole blood was stimulated with R848, a synthetic agonist of toll-like receptors 7 and 8 that mimics viral RNA, monocytes from untreated patients stood out for their heightened inflammatory tone. These cells produced excess interleukin-1 beta and the chemokine MCP-3, both at rest and after stimulation, compared with monocytes from healthy donors. At the same time, plasmacytoid dendritic cells, the body’s principal source of type I interferons, were functionally impaired in untreated patients: their interferon-alpha response to stimulation was markedly blunted. Remarkably, in ocrelizumab-treated patients this pDC defect disappeared, with interferon-alpha responses returning to levels indistinguishable from healthy donors, while monocyte interleukin-1 beta was modestly attenuated. The combination of overactive monocytes and hyporesponsive interferon production in untreated disease, the authors suggest, may fuel chronic cellular activity without the negative feedback needed to restrain it.

The centerpiece of the study, however, was the virus-specific T-cell analysis. Using an activation-induced marker assay, in which virus-specific T cells are identified by their rapid upregulation of CD40L and CD69 on CD4 cells or CD137 and CD69 on CD8 cells after exposure to viral peptide pools, the team quantified memory T-cell responses in seropositive donors. EBV seroprevalence was near-universal, ranging from 95 percent in healthy donors to 100 percent in treated patients, confirming that the comparison was biologically meaningful. The result was unambiguous: absolute numbers of EBV-specific CD4-positive T cells and their stimulation index were significantly lower in ocrelizumab-treated patients than in both untreated patients and healthy donors. Responses to CMV and SARS-CoV-2, by contrast, were comparable across groups in both the CD4 and CD8 compartments.

The selectivity ran deeper than cell counts. Detailed phenotyping of the EBV-specific CD4 memory compartment showed significant decreases across central memory and effector memory subsets in treated patients, along with broad downregulation of activation markers including CD25, CD137 and ICOS. Intriguingly, both EBV-specific CD4 and CD8 cells from treated patients expressed less CCR6, a chemokine receptor that guides T cells across the choroid plexus into the central nervous system during early neuroinflammation. CCR6-positive T cells enriched in Th17-like and GM-CSF-secreting subsets are known to accumulate in the cerebrospinal fluid of people with multiple sclerosis and are considered pathogenic. Their reduction suggests that B-cell depletion may not merely remove EBV-infected B cells but also strip the residual EBV-specific T-cell pool of its neuroinvasive potential, a finding with direct implications for how the therapy works.

The authors interpret these results through the lens of EBV biology. Memory B cells are the principal site of viral latency, so their sustained depletion substantially reduces the EBV antigenic burden. With less virus to control, the large EBV-specific CD4 T-cell pool is no longer maintained, and it contracts. This mechanism aligns with recent independent work showing that anti-CD20 therapy reduces EBV-specific CD4 responses several-fold while leaving CMV, varicella-zoster and bacterial antigen responses largely untouched, and that it abolishes salivary EBV shedding. In this view, part of the clinical benefit of ocrelizumab derives not from generic immunosuppression but from dismantling a specific EBV-driven immunopathological circuit. Equally important for patient care, the preservation of CMV- and SARS-CoV-2-specific T cells, together with normalized pDC interferon responses, indicates that cellular antiviral defense is not globally eroded, providing a mechanistic counterpart to the clinical observation that infection risk under these drugs stems mainly from low antibody levels rather than lost T-cell immunity.

The study also delivered a candidate biomarker. Serum neurofilament light chain was significantly elevated in untreated patients and trended lower after treatment, consistent with reduced neuroaxonal damage under therapy. When the researchers stratified patients by sNfL levels, those above the cohort median showed higher concentrations of the chemokines MCP-3, MIP-1 alpha and MIP-1 beta, and MCP-3 was the only mediator that correlated positively with sNfL in untreated patients, both at baseline and after EBV peptide stimulation. This association was not detectable after ocrelizumab treatment, although a formal interaction test could not confirm that the correlation itself differed between treated and untreated groups. The authors are appropriately cautious, noting that no correction for multiple hypothesis testing was applied and that all soluble-mediator comparisons are exploratory. Still, MCP-3 has historical credentials in the field, having first been described in post-mortem multiple sclerosis lesions in 1998, and the new data position it as a candidate blood marker associated with ongoing neuroaxonal injury.

The study’s limitations frame how far its conclusions can travel. Its cross-sectional design precludes causal inference, the treated antiviral arm included only 13 patients, and all measurements came from peripheral blood rather than the central nervous system. Participants were recruited in Rome and predominantly reflect an Italian relapsing MS population, all treated patients received ocrelizumab specifically, and HLA typing and detailed vaccination histories were not available, leaving HLA-dependent differences in EBV responsiveness impossible to exclude. Even so, the central message carries weight for a therapy used by hundreds of thousands of patients worldwide. Anti-CD20 treatment is not a blunt instrument that dismantles antiviral immunity wholesale. It reshapes the immune system in a patterned way, depleting the B-cell reservoir that sustains EBV and, in doing so, selectively relaxing the T-cell response to that virus, while interferon production and immunity to CMV and SARS-CoV-2 remain substantially preserved. For a disease in which EBV infection raises the risk of onset roughly 32-fold, that pattern of selectivity looks less like a side effect and more like a glimpse of the therapy’s actual mechanism at work.

Subject of Research: Immune effects of anti-CD20 B-cell depletion therapy on antiviral T-cell responses in multiple sclerosis

Article Title: Antiviral and inflammatory immune signatures in anti-CD20-treated multiple sclerosis: a cross-sectional study

Article References: Antiviral and inflammatory immune signatures in anti-CD20-treated multiple sclerosis: a cross-sectional study. (n.d.). https://doi.org/10.1007/s00415-026-14202-4

Image Credits: AI Generated

DOI: 10.1007/s00415-026-14202-4

Keywords: multiple sclerosis, ocrelizumab, anti-CD20, Epstein-Barr virus, B-cell depletion, T cells, neurofilament light chain, MCP-3, neuroimmunology, plasmacytoid dendritic cells, immunophenotyping, antiviral immunity