Epstein-Barr Virus in Multiple Sclerosis

Abstract

Epstein-Barr virus (EBV) is a very common herpesvirus that infects more than 90% of the general population. Epidemiologic data indicate that EBV is a requisite risk factor for the development of multiple sclerosis (MS); however, the mechanisms by which EBV contributes to MS pathogenesis are unclear. In this review, we discuss how EBV alters the functions of B cells, its primary cellular reservoir, and the associated dysregulation of anti-EBV immunity in patients with MS. We comprehensively explore the evidence for different potential mechanisms by which EBV may lead to the development of MS, including the so-called driver and hit-and-run models. Finally, we discuss key outstanding scientific questions that must be addressed to advance not only our understanding of the role of EBV in MS pathology but also the development of novel disease therapies.

Introduction

Epstein-Barr virus (EBV), also known as human herpesvirus 4, is a ubiquitous gamma herpesvirus that infects most of the global population. Although primary infection is often asymptomatic, EBV has been linked to a broad range of diseases, including infectious mononucleosis, lymphoproliferative disorders, lymphomas, and certain autoimmune diseases such as multiple sclerosis (MS).¹

History of the EBV-MS Connection

MS is a chronic autoimmune demyelinating disease of the CNS. The concept that MS may be caused by a viral infection has seen a resurgence in recent years, with a particular focus on the association with EBV. The link between MS and EBV was noted as early as the 1980s.^2-4 Numerous studies have confirmed that more than 99% of patients with MS have been infected with EBV, which is significantly higher than the 90–95% seroprevalence in the global population.^5-11 EBV seropositivity is likely 100% in all patients with MS when considering assay sensitivity and increased diagnostic stringency (e.g., exclusion of disorders that mimic MS).^12-14 This strong association has also been corroborated in pediatric MS, where EBV seropositivity exceeds 95%, substantially higher than the seroprevalence in age-matched healthy controls.^13,15-17 EBV seroconversion was found to increase the risk of MS development by 32-fold, with a median of 5 years before MS diagnosis.¹⁸ These findings therefore emphasize the role of EBV as a prerequisite for the development of MS. After EBV seroconversion, the levels of a biomarker for neuroaxonal degeneration increased,^18,19 suggesting that subclinical CNS injury begins years before neurologic symptoms develop.

EBV Overview

EBV is a 172-kb linear double-stranded DNA virus encoding approximately 85 proteins and 50 noncoding RNAs.^20,21 It is typically transmitted by saliva and infects oral epithelial cells during primary infection.²² Tonsillar lymphocytes exposed to EBV enter the bloodstream where EBV remains dormant inside B cells, the principal reservoir of the virus where lifelong latency is established.²³ EBV gp350, gH/gL, or BMRF2 interacts with epithelial cell proteins, including integrin αvβ1, αvβ6, and αvβ8 and type II membrane protein BDLF2.²⁴ EBV subsequently binds to ephrin A2 receptors, leading to gB-mediated fusion. EBV gp350/220 binds to the complement receptor CD21 on the B cell surface, and fusion is triggered by the interaction between the EBV glycoprotein gp42 and MHC II.²²

EBV induces a variety of different transcriptional programs within B cells, which largely differ by the activation state of the B cell^25-29 (Figure 1). EBV infection of naïve B cells initiates B cell activation through a growth program (latency III), involving expression of all latency genes (encoding EBNA1-EBNA3A/B/C and LMP1-LMP2A/B) driven by the master transcriptional regulator EBNA2.³⁰ Activated B cells then progress through a germinal center reaction program in latency II with a more restricted EBV gene expression pattern (EBNA1, LMP1, LMP2A).³¹ In this stage, LMP1 on B cells mimics the function of CD40, leading to NF-κB–mediated signaling and promoting cell survival.^32,33 LMP1 also leads to CD40 ligand (CD40L/CD154) expression on B cells, which normally only express CD40, permitting in-cis CD40 activation on B cells.³⁴ LMP1 signaling permits B cell class-switching but limits affinity maturation.^33,35 CD40 crosslinking downregulates EBNA2, EBNA3A/B/C, and LMP1, permitting evasion of the host immune response.³⁶ LMP2A activates BCR-dependent signaling pathways to promote B cell survival³⁷ and increases IL-10 production to suppress immunity.³⁸ Latency 0 is the most restricted latency stage, in which only the noncoding RNAs EBER1 and EBER2 and the BART microRNAs are expressed, thus preventing detection by the immune system. EBNA1 is switched on transiently in latency I to permit viral episome replication during B cell division.³⁹ Consequently, B cells exist infrequently in latency I outside certain malignancies (e.g., Burkitt lymphoma).³⁹

Figure 1. Overview of EBV Infection and Expression Stages in B Cells.

During primary infection, EBV infects naïve B cells through binding of gp350/220 to CD21, followed by gp42 to MHC II. Newly infected B cells undergo rapid expansion in latency III where all latency-associated genes are expressed. The resulting memory B cells undergo a germinal center–like differentiation process in latency II, where a more restricted latency gene expression program permits B cell survival. Latently infected B cells in healthy convalescent carriers persist in a very quiescent state termed latency 0 and infrequently express EBNA1 in latency I during homeostatic cell division. B cell activation and plasma cell differentiation results in expression of BZLF1 (and/or BRLF1), resulting in activation of the EBV lytic program and virus production. Created in part with BioRender. Sabatino, J. (2024).

BCR crosslinking of latently infected memory B cells leads to plasma cell differentiation and activation of the EBV lytic program.⁴⁰ In this stage, viral immediate-early (IE) genes, early (E) genes, and late (L) genes are expressed in successive order.^29,41 The immediate-early genes BZLF1 and BRLF1 play a critical role in initiating the lytic cycle by acting as transactivator proteins that trigger the expression of early genes.^42,43 The early lytic genes encode essential enzymes necessary for viral DNA replication.⁴³ Amplification of EBV DNA takes place between early and late gene expressions. During late gene expression, structural proteins are produced and assembled into viral particles that encapsulate viral DNA, followed by release of infectious virions.⁴³ Of note, EBV reactivation in B cells does not always lead to successful virus production.⁴⁴

Primary EBV Infection and the Acute Immune Response

During acute infection, the virus undergoes active replication and produces new viral progeny.²³ Most of the primary EBV infections that occur in childhood are asymptomatic.^45,46 Infectious mononucleosis (IM) is a syndrome that is characterized by severe pharyngitis, cervical lymph node enlargement, fever, fatigue, and possible splenomegaly and hepatomegaly and occurs most commonly in adolescents and young adults.¹ IM is triggered by high EBV viral loads in the oral cavity and blood, with up to 50% of memory B cells infected at peak.⁴⁷ EBV is cleared rapidly from the blood but can persist in saliva at high viral loads with retained infectivity for at least 6 months.⁴⁸

In asymptomatic EBV infection in children, age does not seem to influence antiviral antibody titers.⁴⁹ IM-associated acute EBV infection coincides with an overactive immune response manifesting in the generation of IgM antibodies against EBV viral capsid antigen (VCA) as well as a significant expansion of NK cells and CD8⁺ T cells.^46,50 Shortly after the development of early antibodies against EBV VCA, IgG antibodies develop against EBV early antigen (EA) and EBNA2 and then gradually decline.^e1 EBNA1-specific IgG antibodies develop around 3–6 months after infection in parallel with EBNA1-specific CD4⁺ T cells, which persist long term in healthy carriers.^e1-e3

NK cells are rapidly activated in acute IM, with several distinct subsets activated. Interferon-γ (IFNγ)–producing CD56^bright CD16⁻ NKG2A+ NK cells are rapidly expanded in tonsillar and lymphoid tissue, where they control B cell proliferation and transformation.^e4,e5 The proportion of CD56^bright NK cells is typically at their peak levels in young children and declines with age into early adulthood,^e6 suggesting a potential mechanism for IM risk in adolescents. Although highly differentiated cytotoxic CD56^dim CD16⁺ NKG2A+ NK cells are expanded in the blood,^e7 the intermediate CD56^dim CD16⁻ NK cells undergo the greatest expansion during acute EBV infection.^e8

Primary EBV infection is ultimately controlled by T cells, in particular cytotoxic CD8⁺ T cells.^e9 EBV LMP1 expression on the surface of B cells leads to upregulation of MHC I and II and costimulatory molecules, enhancing antigen presentation to T cells.^e10,e11 Conversely, EBNA1 has a large series of Gly-Ala repeats at amino acids 90–325, which interfere with its translation^e12 and antigen presentation on MHC I,^e13,e14 blunting EBNA1-specific CD8⁺ T cell responses.^e15 During acute IM, CD8⁺ T cells undergo a rapid and massive expansion, with up to 50% of those in the blood specific for a single EBV epitope.^e15-e19 EBV-specific CD8⁺ T cell expansion is driven by IL-27 production from EBV-infected B cells.^e20,e21 Most of the CD8⁺ T cells in acute IM are reactive against immediate-early and early lytic antigens, with smaller frequencies reactive to latent antigens.^e9,e19,e22 The frequencies of EBV-specific CD8⁺ T cells are significantly reduced in tonsillar tissue during acute IM.^e19 Healthy EBV-convalescent individuals exhibit a modest reduction in the frequency of EBV-specific T cells with a partial shift toward late lytic antigens,^e7,e9,e23 while maintaining a clonally stable repertoire over many years.^e24 EBV viremia is comparable between individuals with asymptomatic EBV infections and those with acute IM^e22,e25 with similar frequencies of EBV-specific CD8⁺ T cells. However, the former group does not develop the massive CD8⁺ T cell expansion observed in the latter.^e9,e25,e26 The reason for this divergence in bystander CD8⁺ T cell expansion in acute IM remains unknown.

While latently infected B cells in the peripheral blood are found almost exclusively in the CD27⁺ IgD⁻ memory subset,^47,e27,e28 both naïve (IgD+) and memory B cells harbor latent EBV in tonsillar lymphoid tissue.^30,e27 EBV latency is estimated to occur in 1 in 10⁵-10⁶ memory B cells in the peripheral blood of long-term healthy carriers,^47,e27 with 5 to several hundred copies of EBV DNA per infected cell.^e29 Latent EBV can reactivate under certain circumstances, including stress, malignancy, autoimmune disease, and other conditions.^e30 EBV reactivation is believed to be kept in check by T cells, in particular CD8⁺ T cells.^e31,e32 Indeed, the frequencies of EBV-specific CD8⁺ T cells are enriched in tonsillar tissue in long-term healthy carriers and react readily to recently EBV-infected B cells.^e33 This suggests that CD8⁺ T cell reactivity is likely maintained in sites of EBV latency (e.g., B cell–enriched lymphoid tissue) because of recurrent viral reactivation.

Anti-EBV Humoral Immunity in MS

Not only is EBV seropositivity more common in patients with MS, but those who experience symptomatic IM have more than twice the risk of developing MS.^e34-e37 The presence of siblings confers protection against MS development.^e38 This suggests that a delayed EBV infection, which is associated with a higher IM risk, enhances overall MS risk.

Studies investigating antibodies have advanced our understanding of how EBV infection influences MS. Patients with MS have anti-EBNA1 antibody titers several-fold higher than EBV-seropositive healthy controls.^7,e39-e41 Furthermore, higher anti-EBNA1 antibody titers confer higher risk of developing MS,^7,9,10,e42 and titers continue to rise after MS onset.^7,8 Anti-EBV antibodies, in particular to EBNA1, have also been detected in the CSF of patients with MS.^e41,e43-e45 Antibodies against other EBV antigens, including VCA, EA diffuse (EA-D), EBNA2, and EBNA3C, are also increased in MS.^{9,19,e39,e46-e48}

EBV also interacts with a number of MS-associated genetic risk factors.^e49,e50 The HLA-DRB1*15:01 allele, the strongest genetic risk factor of MS,^e50,e51 is associated with significantly increased anti-EBNA1 IgG levels.^e40,e52-e58 HLA-DRB1*15:01–positive MS patients have increased antibody reactivity to specific regions of EBNA1, in particular EBNA1_385-420.^e53,e103 As described above, MHC II facilitates EBV fusion with B cells, and HLA-DRB1*15:01 may confer higher EBV infectivity than other MHC II alleles.^e59,e60 On the contrary, EBV viral load is reduced in patients with MS carrying the MHC I allele HLA-A*02:01,^e40 a known protective factor against MS.^e51 Of interest, HLA-A*02:01 is also associated with reduced likelihood of IM.^e56,e57

EBV is well recognized for its ability to regulate host cell gene expression. A number of EBV genetic variants in different latency genes have been reported to be significantly associated with MS.^e61 EBNA2 occupies transcription factor binding sites in many MS-associated loci,^e62-e65 including those associated with the CD40 pathway.^e66 Several studies have also demonstrated that certain EBNA2 genetic variants are associated with MS. Specific allelic variants of EBV LMP1 have also been shown to interact differently with NK cell and HLA-E–associated immune responses in patients with MS.^e67

In summary, individuals who eventually develop MS exhibit elevated serum antibody titers to specific EBV antigens, which may stem from various underlying factors, including age at infection and host genetic predisposition. In addition, these studies provide potential serologic evidence of impaired immune regulation of the persistent latent EBV carrier state and EBV reactivation. Assuming that EBV viral load is not different in individuals who eventually develop MS, this suggests that patients with MS exhibit distinct immune responses despite similar EBV exposure. Further insight into how EBV infection affects MS onset is therefore critical.

EBV-Specific T Cells in MS

A number of studies have explored EBV-specific T cells in MS. While increased frequencies of EBV-specific CD8⁺ T cells were observed in patients with MS compared with healthy controls in some studies,^e43,e68 other studies have not found a clear difference.^e48,e69,e70 Several studies have also suggested reduced frequencies and/or functionality of EBV-specific CD8⁺ T cells in later stages of MS.^e70-e72 Studies examining the repertoire of the T cell receptor β chain (TCRβ) in the peripheral blood found that patients with MS had a larger fraction of public EBV-specific CD8⁺ TCR sequences compared with non-MS patients.^e73-e75 In a study of monozygotic twins discordant for MS status, twins with MS had increased EBV-specific TCRβ sequences, although the twin pairs shared EBV seropositivity and did not differ in other viral TCRβ specificities.^e73 Changes in the frequencies of certain EBV-specific CD8⁺ T cells during relapse and remission of individual patients with MS have yielded mixed results.^e69,e70,e76

Several studies have reported higher frequencies of EBV-specific T cells in the CSF relative to the blood of patients with MS compared with patients with other neurologic disorders.^e73,e77-e80 It was also recently demonstrated that several highly expanded EBV-specific CD8⁺ T cell clonotypes in the CSF of patients with MS were enriched relative to the blood.^e81 By contrast, CSF enrichment of CD8⁺ T cells specific to other common viral antigens (e.g., influenza A and cytomegalovirus) has not been observed in MS.^e78-e80 There is also evidence of increased frequencies of EBV-specific CD4⁺ T cells in the blood and CSF of patients with MS.^e82-e84 In particular, EBNA1-specific CD4⁺ T cells are increased in a pattern similar to anti-EBNA1 IgG and exhibit a Th1 profile.^e42,e82,e84 It is important to note, however, that EBV-specific T cells, in particular CD8⁺ T cells, can be readily identified from the CSF of patients with MS, as well as patients with other neuroinflammatory conditions.^{e73,e77-e80,e85-e87}

EBV-specific CD8⁺ T cells have also been identified in MS lesions. Using in situ staining with EBV peptide–loaded multimers, EBV-specific CD8⁺ T cells were detected in the perivascular cuffs of active MS lesions and meninges, with frequencies ranging from 0.2% to 2.5%.^e88-e89 By contrast, much lower signals were observed against cytomegalovirus (CMV) and influenza A. Another study detected enhanced reactivity to EBV-transformed B cells in CD8⁺ T cell lines derived from active MS lesions but not from normal appearing white matter,^e86 raising the possibility of focal EBV expression in the CNS.

Other Herpesviruses in MS

CMV is another common virus in the herpes family that infects approximately 40%–60% of the general adult population, resulting in lifelong latent infection. In contrast to EBV, patients with MS are at least 20% more likely to be CMV seronegative compared with age-matched controls.^3,e90,e91 Among patients with MS and healthy controls with similarly high EBNA1 IgG titers, healthy controls are far more likely to be CMV seropositive,^e20 although additional confirmatory studies are needed. This is consistent with the finding that TCRs associated with CMV specificity are reduced in patients with MS.^e75 Increased CMV antibody levels are also associated with reduced MS severity.^e92,e93 The protective effect of CMV in MS is further augmented in the setting of coinfection with EBV and human herpesvirus 6A (HHV-6A),^e91,e94 suggesting interactions between exposures to certain herpesviruses. By contrast, immunity to other human herpesviruses, including herpes simplex 1 and herpes simplex 2 (HSV-1 and HSV-2) and varicella zoster virus, is not different in patients with MS.^e92,e95 The mechanism by which CMV exerts a protective effect in MS is not clear, but host and virus genetic factors may affect HLA-E–mediated NK cell control of pathogenic T cells.^e20

Potential Mechanisms of EBV in MS

We contend that MS cannot occur without previous EBV infection. The most likely explanation for rare patients with MS with negative EBV serology is false-negative antibody testing.^e96 Alternatively, these patients may reflect misdiagnoses of disorders that mimic MS, in particular in patients diagnosed before the discovery of neuromyelitis optica and myelin oligodendrocyte antibody–associated disorder, which do not have strong associations with EBV.^17,e97,e98 It has been posited that MS represents a multistage process in which EBV infection is an early insult, followed by other genetic and/or environmental factors that must align in order for MS to develop in susceptible individuals.^e99,e100 How then does a ubiquitous virus that latently infects B cells lead to CNS autoimmunity in a subset of individuals years after primary infection? Two primary hypotheses have been put forward over the past 4 decades^e100-e102 (Figure 2). The first, known as the “driver hypothesis,” posits that periodic EBV reactivation drives a pathogenic autoimmune response. The second, the “hit-and-run hypothesis,” suggests that primary EBV infection causes immune dysfunction, which then leads to CNS autoimmunity. In this second scenario, EBV latency and reactivation are no longer relevant once MS pathogenesis is underway. In the sections given further, we review the evidence supporting these hypotheses and discuss additional outstanding questions.

Figure 2. Proposed Models of EBV in MS Pathogenesis.

In the “driver” model (left), EBV reactivation within latently infected B cells leads to activation of EBV-specific T cells. EBV expression in the CNS leads to immune activation and CNS pathology. In the “hit-and-run” model (right), EBV-specific B cells and T cells cross-react with CNS autoantigens (i.e., molecular mimicry). Alternatively, bystander CD8⁺ T cells activated during infectious mononucleosis may be reactive against CNS antigens. Created in part with BioRender. Sabatino, J. (2024).

EBV Reactivation and Expression in MS

The lifelong persistence of EBV, its periodic reactivation, and the presence of EBV-specific CD8⁺ T cells in MS lesions suggest that EBV re-expression may be related to MS pathology (Figure 2). Early evidence of differences in EBV latency in MS was observed through higher rates of spontaneous B cell transformation in peripheral blood samples from patients with MS compared with healthy donors.^e103,e104 Subsequent studies have demonstrated higher EBV viral loads and EBV-containing exosomes in the blood of patients with MS compared with healthy individuals.^e40,e105 Another study found that compared with healthy controls, patients with MS had increased lytic EBV gene expression in memory B cells, especially during active disease phases.^e106 Increased EBV DNA was also detected in cervical lymph nodes of patients with MS compared with controls, which correlated with increased EBV viral loads in saliva.^e76

However, not all studies have observed detectable differences in EBV DNA and/or mRNA levels in peripheral lymphocytes between patients with MS and controls.^e107-e111 Furthermore, significant differences in EBV viral loads or transcriptomes have not been reported in the CSF of patients with and without MS.^{e108,e110,e112} While EBV-specific antibodies are present in the CSF, studies have yielded divergent results regarding whether EBV reactivity differs between patients with MS and controls.^e47,e77,e113 One recent study reported an increase in EBNA1 IgG titers in the CSF of patients with MS within the 407–419 region,^e114 similar to the region associated with HLA-DRB1*15:01. In addition, several studies have reported EBNA1-specific oligoclonal bands in the CSF of patients with MS.^e115,e116

A number of studies have demonstrated detectable EBV DNA, EBV protein-expressing exosomes, and/or increased anti-EBV antibody reactivity in patients with actively relapsing MS compared with those with stable MS.^{6,e105,e107,e117} EBNA1 IgG titers were particularly increased in patients with actively relapsing MS^e117 while anti–EA-D did not change.^e118 EBNA1 and VCA IgG titers also positively correlated with T2 lesion numbers in patients with clinically isolated syndrome, whereas antibody titers against other herpesviruses showed no association with MRI outcomes.^e42 Overall, these findings suggest that EBV quiescence is perturbed in MS and may be associated with acute MS inflammation.

It has been posited that MS may reflect a dysregulation of EBV latency II in B cells,^e119 although it remains unclear whether a specific latency or lytic program is dysregulated. These studies therefore highlight the need for more comprehensive analysis of EBV expression in MS, i.e., specific gene products and proteins associated with specific EBV states. The fact that B cells serve as the primary reservoir of EBV and are a key player in MS pathology establishes a strong mechanistic link between EBV and MS. Despite this, key questions remain unanswered. Does EBV reactivation trigger a subsequent inflammatory response? Because EBV expression occurs in a number of different patterns, B cell function may be altered in a way that induces MS inflammation. Alternatively, it is possible that B cell–driven inflammation develops first in MS, leading to subsequent EBV reactivation. In this scenario, changes in EBV reactivation are an epiphenomenon rather than a true driver of pathology. Indeed, this is consistent with evidence of EBV reactivation in other infectious settings, including COVID-19.^e120-e122

Is EBV Present in the CNS in MS?

EBV is present within CNS-homing B cells,^e123 providing a potential gateway for EBV entry into the CNS. EBV reactivation in the CNS has also been reported in cases of fatal MS after natalizumab washout.^e88,e124 Many studies have detected EBV in B cell–enriched regions of the CNS of patients with MS,^{e66,e69,e89,e125-e133} although a number of studies have failed to replicate these findings.^{e86,e134-e189} Several of the negative studies independently analyzed the same MS patient tissue samples as those used in studies reporting positive EBV expression.^e135,e137 These discrepancies could be due to variations in methods, differences in measured EBV products, and differences in sample preparation and quality.^e140 Of the studies that have detected EBV in the CNS, both latent and lytic EBV gene products were reported in the perivascular cuffs of active and chronic lesions and in the meninges. Several of these studies also assessed EBV expression in the CNS of control patients, including those with other neurologic diseases or no neurologic disease—although EBV was detected more frequently in tissue from patients with MS, it was also detected in a sizeable fraction of control brain tissue.^{e125,e127,e129,e133} These studies therefore suggest that while EBV may be present in the CNS of patients with MS, they do not adequately explain how EBV contributes to autoimmune CNS demyelination. Indeed, one must consider the possibility that the presence of EBV in the CNS is a marker of B cell infiltration, and not necessarily indicative of a pathologic role. Of interest, a recent study identified much higher frequencies of EBV-infected B cells in the CNS compared with what has been reported in circulation, with the highest amounts in MS tissue.^e133

Other Reservoirs of EBV

While B cells are considered the primary reservoir of EBV, it is important to recognize that EBV is capable of infecting and persisting within other cell types. Oral epithelial cells are not only sites of initial EBV infection but also reservoirs of latent EBV.^e141,e142 Although generally resistant to EBV infection, T cells can be infected by EBV type II.^e143 EBV has also been reported in thymic tissue,^e144 although the precise cellular reservoir is not clear. NK cells can also be directly infected by EBV.^e145 EBV, along with other herpesviruses, has also been reported in 10%–20% of cervical samples in women.^e146,e147 These findings therefore extend the potential locations of EBV latency and reactivation within the periphery.

It is important to note that there is evidence that EBV may reside within certain cellular populations of the CNS. For instance, EBV can infect endothelial cells,^e148 including human brain microvascular endothelial cells.^e149 Furthermore, EBV is capable of infecting astrocyte lines in vitro^e150 and has been detected in astrocytes and microglia in MS brain tissue.^e127,e133 There is some evidence that EBV can infect neurons directly,^e151 and a recent study revealed the presence of EBNA1 in neurons within MS lesions.^e133 There are no published reports of EBV infection of oligodendrocytes. Collectively, it is clear that EBV can infect a variety of cell types, including certain CNS cells, which may permit EBV entry into the CNS. In addition, the close proximity of EBV-infected cells with neurons, astrocytes, microglia, and oligodendrocytes provides indirect mechanisms by which CNS pathology might occur.^e133

In light of the evidence in favor of the “driver hypothesis,” there is strong interest in developing therapeutic strategies to target EBV itself for the treatment of MS^e152,e153 (Table 1). In several small clinical trials, the antiviral medications acyclovir and valacyclovir showed nonsignificant trends toward reduced relapse rates and MRI activity.^e154-e156 Notably, IFNβ treatment was chosen for clinical testing in MS because of its known antiviral properties, although its specific effects on EBV remain limited.^e157 Adoptive cell therapy involving transfer of EBV-specific T cells for enhanced control of EBV has gained considerable interest as a potential treatment strategy^e119 (Table 1). While an early phase I trial showed some modest effects on progressive MS outcomes,^e158 the more recent phase II EMBOLD trial (ATA188) failed to reach its primary end point of confirmed disability improvement at 12 weeks in progressive MS.^e159 Despite the lack of clear clinical efficacy of EBV-directed therapies, a number of unanswered questions remain, including their effects on EBV reactivation, MS relapse rates, and MRI activity.

Table 1.

Potential EBV-Directed Therapeutic Strategies for MS

Therapeutic strategy	Goal of strategy	Potential limitations
Antiviral treatment	Treatment of MS	Incomplete viral control, tissue penetration
EBV cellular therapy	Treatment of MS	Duration of treatment, HLA matching, identification of appropriate antigens
EBV vaccination	Prevention of MS, treatment of MS	Timing of vaccination, identification of appropriate antigens

Abbreviations: EBV = Epstein-Barr virus; HLA = human leukocyte antigen.

MS Disease-Modifying Therapy Effects on EBV Immunity

Given that memory CD20⁺ B cells are the primary reservoir of latent EBV and the EBV-associated immunologic changes in MS, it is logical to hypothesize that the therapeutic effects of anti-CD20 mAb therapies may be due to removal of the primary EBV niche. In support of this, multiple studies have demonstrated reduced levels of EBNA1 IgG after initiation of anti-CD20 mAb treatment.^e160-e162 Total IgG levels generally decline gradually after long-term B cell depletion treatment,^e163,e164 suggesting that the reduction in EBNA1 IgG is not due to a global reduction in antibody levels. Several studies have also shown a reduction in the reactivity of EBV-specific T cells after anti-CD20 treatment, in particular memory CD8⁺ T cells, whereas other virus-specific T cells remain unchanged.^e165-e167

The reduction in anti-EBV immunity in the setting of strong clinical efficacy after anti-CD20 mAb treatment supports the notion that either EBV reactivation or the adaptive immune response to EBV is directly involved in MS pathology.^e168 Nonetheless, evidence of persistent EBV is demonstrated by the finding that EBV DNA remains in the oropharynx of patients treated with anti-CD20 mAb.^e16 This suggests that anti-CD20 mAb therapies may not completely deplete B cells from lymphoid tissue or that EBV may persist in alternative reservoirs. Anti-CD20 mAbs also do not significantly reduce leptomeningeal inflammation, a site of B cell aggregation in MS,^e160 which may be relevant to ongoing neurodegeneration in progressive MS.

Autologous hematopoietic stem cell transplant (AHSCT) is sometimes used for severe treatment-refractory MS. Viral reactivation, including EBV viremia, is a known feature during post-AHSCT immune reconstitution.^e170,e171 EBV-specific CD8⁺ T cells rapidly expand and persist in AHSCT-treated patients with MS, and the frequency of latently EBV-infected B cells after AHSCT did not correlate with clinical response.^e172 These findings suggest that the presence of latently infected B cells and the expansion of EBV-specific CD8⁺ T cells are alone insufficient to explain MS pathology, suggesting that additional pathologic factors must be involved or that the anti-EBV immune response may be protective in MS.

EBV-Mediated Immune Dysfunction in MS

An alternative model is the “hit-and-run” hypothesis in which primary EBV infection leads to immune dysfunction and subsequent EBV-independent MS pathology (Figure 2). This mechanism is consistent with an immunologic concept known as heterologous immunity, in which an initial immune response (e.g., against virus infection) influences a subsequent immune response.^e173,174 There are a number of plausible ways in which such immune dysfunction could occur. Molecular mimicry, in which foreign antigen-specific B cells and/or T cells are cross-reactive to self-antigens, is one commonly purported mechanism. For instance, primary EBV infection may lead to activation of EBV-specific lymphocytes that cross-react with CNS autoantigens and contribute to MS.

A number of studies have identified EBV antibody cross-reactivity to various self-antigens (Table 2). A region of EBNA1 (386–444) has been identified with homology to several different human proteins. This is the same region of EBNA1 against which CSF IgG levels are increased in MS,^e114 and it is consistent with the finding that EBV latency proteins have a particularly high degree of mimicry potential to self-antigens.^e175 EBNA1 antibody cross-reactivity has been identified with 3 different self-antigens: glial cell adhesion molecule (GlialCAM), αB-crystallin (CRYAB), and anoctamin-2 (ANO2). ANO2 seroconversion occurs after EBNA1 seropositivity and precedes MS onset,¹⁹ and ANO2 antibody reactivity is increased in patients with MS.^e176 EBNA1 cross-reactivity to CRYAB has been detected within 6–12 months of IM and in EBV-seropositive healthy carriers, but not to ANO2 or GlialCAM,^e58 suggesting that cross-reactivity may vary in health and disease. The MS risk allele HLA-DRB1*15:01 further increases the likelihood of autoantibodies cross-reacting with EBNA1.^{e58,e177-e179} These antigens are expressed in neurons, oligodendrocytes, and other glial cells (Table 2) although their expression is not restricted to the CNS. In addition to EBNA1, an MS-specific autoantibody signature identified in a subset of patients with MS recognized a motif found within different self-antigens and EBV BRRF2 and gM^e175,e180 (Table 2).

Table 2.

EBV and Self-Antigen Cross-Reactivity in MS

	EBV antigen	Self-antigen	CNS expression	References
Antibody/B cells	EBNA1	GlialCAM	Oligodendrocytes, microglia	e44,e67,e179,e192,e193
	EBNA1	ANO2	Neurons, microglia	19,e177,e179,e192,e193
	EBNA1	CRYAB	Oligodendrocytes, glial cells	e58,e178,e179,e193,e194
	BRRF2, gM	SRSF4, RIMS2, CLASRP, SRRM3, TRIO	Neurons, glial cells	e180
	BRRF2, BFRF3	Septin-9, DLST	Neurons, glial cells	e195
CD4+ T cells	BALF5	MBP	Oligodendrocytes	e181
	EBV LCL	CRYAB	Oligodendrocytes, glial cells	e183
	EBNA1	Myelin proteins	Oligodendrocytes	e60,e83
	BPLF1, BHRF1	RasGRP2	Neurons	e182
	EBV LCL	Myelin proteins	Oligodendrocytes	e48
	EBNA1	GlialCAM	Oligodendrocytes, microglia	e67
CD8+ T cells	EBV LCL	Myelin proteins	Oligodendrocytes	e48
	EBNA1	GlialCAM	Oligodendrocytes, microglia	e67

Overview of EBV-specific B cells and T cells that are cross-reactive to self-antigens and their CNS expression.

Abbreviations: ANO2 = anoctamin-2; CLASRP = CLK4-associating serine/arginine rich protein; CRYAB = alpha B-crystallin; DLST = dihydrolipoamide S-succinyltransferase; GlialCAM = glial cell adhesion molecule; LCL = lymphoblastoid cell line; MBP = myelin basic protein; RasGRP2 = Ras guanyl-releasing protein 2; RIMS2 = regulating synaptic membrane exocytosis 2; SRRM3 = serine/arginine repetitive matrix 3; SRSF4 = serine and arginine rich splicing factor 4; TRIO = triple functional domain protein.

EBV-specific CD4⁺ and CD8⁺ T cells cross-reactive to a variety of CNS autoantigens have also been reported^{e48,e60,e83,e181,e182} (Table 2). Ras guanyl-releasing protein 2 (RasGRP2) is coexpressed in B cells and CRYAB is present in EBV-infected B cells,^e182,e183 providing a connection between EBV, B cells, T cells, and CNS pathology. While there is clear antibody cross-reactivity between EBNA1_386-444 and GlialCAM, T cell cross-reactivity between these 2 antigens is less definitive^e67 because of wide variation of peptide binding to different MHC I and II alleles. Indeed, CD8⁺ T cell reactivity was similar between the extracellular and intracellular regions of GlialCAM,^e44 suggesting that any potential T cell cross-reactivity extends beyond a narrow region of EBNA1. EBV modulation of B cell antigen-presenting functions, such as through LMP1 expression,^e184 could activate self-reactive T cells in MS. EBV-infected B cells also produce larger amounts of IL-27 to influence T cell activation¹¹⁷, which is dysregulated in MS.^e185,e186 Because antigen-presenting cells can activate numerous effector T cells, a relatively small number of EBV-infected B cells could activate significant numbers of pathogenic T cells. Furthermore, recurrent EBV reactivation could lead to progressively increased activation of self-reactive T cells until a threshold for disease pathology is ultimately crossed.^e187 This scenario could explain the delay between EBV infection and MS onset. While TCR and antibody cross-reactivity to different antigens is a fundamental hallmark of adaptive immunity (and essential for ensuring reactivity to an enormous array of potential foreign antigens), a significant limitation of the molecular mimicry model is the uncertainty of the target autoantigens in MS. Consequently, their clinical relevance is speculative.

In addition to cross-reactivity, bystander activation is another potential mechanism by which EBV could trigger CNS autoimmunity through the “hit-and-run” model^e188 (Figure 2). Primary infection with certain viruses, such as viral hepatitis and influenza, can lead to bystander activation of T cells specific for unrelated antigens.^e189 In the case of MS, the bystander expansion of non–EBV-specific CD8⁺ T cells during IM^e22,e25,e26 could lead to the generation of pathogenic autoreactive T cells, but further investigation is still needed. Ultimately, the “driver” and “hit-and-run” hypotheses help to explain certain features by which EBV may contribute to MS pathology, but each has limitations (Table 3). It is important to note that these potential mechanisms are not mutually exclusive, and it is possible that both could be relevant to MS.

Table 3.

Strengths and Limitations of Potential Mechanisms of EBV in MS

“Driver” hypothesis
Support	Limitations
EBV in B cells and CNS provides link to MS pathology	EBV presence in CNS is controversial
Anti-EBV immunity changes during relapse	Specificity of EBV in MS CNS is not clear
	AHSCT is beneficial in MS despite EBV reactivation and CD8+ T cell expansion
	EBV reactivation may be a nonspecific marker of immune dysfunction

“Hit-and-run” hypothesis
Support	Limitations
Cross-reactivity between EBV and numerous self-antigens expressed in CNS identified significant link with EBNA1	MS-relevant autoantigens remain unconfirmed
IM-associated CD8+ T cell expansion allows for T cell specificity beyond EBV	Cross-reactivity is a normal feature of adaptive immunity and may not be pathologic

Abbreviations: AHSCT = autologous hematopoietic stem cell transplant; EBV = Epstein-Barr virus; IM = infectious mononucleosis.

Where Do We Go From Here?

It is undoubtedly clear that EBV infection must occur before the development of MS. Thus, EBV vaccine–based strategies are being explored for the prevention of MS^e190,e191 (Table 1). While complete prevention of EBV infection (i.e., induction of sterilizing immunity) may not be an easily achievable goal, early vaccination may reduce the likelihood of IM, thereby protecting against the development of MS. A number of EBV vaccine clinical trials are either ongoing or completed using mRNA, recombinant protein, ferritin-nanoparticle, or synthetic peptide strategies, targeting EBV cell entry (gp42/gp220/gH/gL or gp350) for preventing EBV infection (NCT05164094, NCT04645147, NCT05683834, NCT06908096) or EBNA1, LMP2, EBNA3A, or BZLF1 for EBV-related diseases (NCT01094405, NCT01147991, NCT00078494, NCT05714748, NCT05831111, NCT01256853, NCT01800071, NCT06741072). These early-stage studies are focused on safety and EBV outcomes. Some of these efforts aim to augment anti-EBV immunity in patients who have already developed MS. A key caveat of this approach is the unanswered questions regarding the EBV-associated “hit-and-run” model and molecular mimicry in MS. It is conceivable that inclusion of the incorrect antigen(s) in an EBV vaccine could induce a pathogenic autoimmune response precipitating or exacerbating MS, rather than preventing it (even if EBV infection or reactivation is prevented). Alternatively, if EBV induces a pathogenic EBV-independent immune response (e.g., bystander activation), then EBV vaccination after MS onset will be too late to be effective. Similar limitations would apply to antiviral and cellular treatment strategies against EBV (Table 1). Thus, understanding the precise immunologic triggers that occur after EBV infection is paramount.

Although clinically overt MS develops years after primary EBV infection, there is evidence of subclinical neuroinflammation in the interim period.¹⁸ Further research into this prodromal period is needed, acknowledging the challenges in acquiring pre-MS samples. A more comprehensive analysis of EBV expression within B cells is also needed for assessing the potential role of EBV reactivation in the “driver” model of MS. The very low frequency of EBV-infected B cells in nonmalignant conditions makes this challenging, but using sensitive and orthogonal approaches (e.g., quantitative PCR, flow cytometry, and microscopy) and analyzing additional tissues (e.g., lymphoid tissue) may provide further insight. Pairing these findings with changes in B cell function, such as antigen presentation to T cells, will be highly informative. It is also crucial to compare findings in patients with MS with results in those with other autoimmune conditions and healthy controls to understand clinical relevance. Because EBV reactivation may simply be a marker of immune dysregulation rather than a direct contributor to MS pathology (i.e., an epiphenomenon), identifying the antigenic targets of MS-relevant B cells and T cells is clearly needed.^e81,e180

Unraveling the complex interactions between EBV and MS will continue to be a challenge. Nonetheless, the combination of new basic science technologies and various clinical studies has the potential to yield significant new insights and hopefully new therapies. In addition to advancing our understanding of MS, such studies are likely to enhance our understanding of how EBV affects us all, in both health and disease.

Glossary

AHSCT

autologous hematopoietic stem cell transplant

CMV

cytomegalovirus

EBV

Epstein-Barr virus

MS

multiple sclerosis

TCRβ

T cell receptor β chain

VCA

viral capsid antigen

Author Contributions

F. Wahbeh: drafting/revision of the manuscript for content, including medical writing for content. J.J. Sabatino: drafting/revision of the manuscript for content, including medical writing for content.

Study Funding

The authors report no targeted funding.

Disclosure

The authors report no relevant disclosures. Go to Neurology.org/NN for full disclosures.