Altered EBV specific immune control in multiple sclerosis

Since the 1980s it is known that immune responses to the Epstein-Barr virus (EBV) are elevated in multiple sclerosis (MS) patients. Recent seroepidemiologial data have shown that this alteration after primary EBV infection identifies individuals with a more than 30-fold increased risk to develop MS. The mechanisms by which EBV infection might erode tolerance for the central nervous system (CNS) in these individuals, years prior to clinical MS onset, remain unclear. In this review I will discuss altered frequencies of EBV life cycle stages and their tissue distribution, EBV with CNS autoantigen cross-reactive immune responses and loss of immune control for autoreactive B and T cells as possible mechanisms. This discussion is intended to stimulate future studies into these mechanisms with the aim to identify candidates for interventions that might correct EBV specific immune control and/or resulting cross-reactivities with CNS autoantigens in MS patients and thereby ameliorate disease activity.


Introduction
Epstein-Barr virus (EBV) infection was originally discovered in one of its associated B cell lymphomas, namely Burkitt's lymphoma (Epstein et al., 1964a;Epstein et al., 1964b).Indeed, EBV remains best known for its tumorigenic potential, transforming human B cells in vitro and being associated with epithelial and lymphocytic malignancies, such as nasopharyngeal carcinoma and Hodgkin's lymphoma (Münz, 2019;Nilsson et al., 1971).EBV associated malignancies are thought to contribute with 1-2% to all cancers in humans with annually around 300′000 new cases (de Martel et al., 2020;Wong et al., 2022).In addition, and possibly due to the potent immune stimulatory activity of EBV transformed B cells (Leung et al., 2013) EBV infection is associated with immune pathologies, such as infectious mononucleosis (IM), the primary EBV infection that drives a massive CD8 + T cell lymphocytosis in around 10% of initial virus encounters (Dunmire et al., 2018).During IM viral loads peak 4 to 6 weeks after EBV encounter, are controlled in peripheral blood after one additional month but often allow elevated viral shedding into the saliva for more than half a year (Fafi-Kremer et al., 2005), suggesting more rapid systemic immune control than at mucosal surfaces.In some individuals, however, the high setpoint of viral loads persists during chronic active EBV (CAEBV) allowing the infection to spread to T and natural killer (NK) cells in addition to B cells, and developing life-threatening hyperinflammation often culminating in hemophagocytic lymphohistiocytosis (HLH) (Collins et al., 2021;Fournier et al., 2020).Therefore, it was often speculated that this propensity of EBV infection to cause hyperinflammation might contribute to autoimmune diseases, in some of which systemic EBV viral loads also stay elevated (Münz et al., 2009).Such a connection to autoimmunity was further strengthened by recent findings that EBV infection with seroconversion initiates a prodromal phase of neuroaxonal damage in individuals that then develop the central nervous system (CNS) autoimmune disease multiple sclerosis (MS), while alteration in the antibody responses was not detected for around 200 other viral pathogens during MS disease progression (Bjornevik et al., 2022;Bjornevik et al., 2023).This prodromal phase can last for years and therefore it becomes important to understand how EBV infection and its immune control might be altered in this subset of EBV infected individuals that progress to MS (Fig. 1).Thus, it is important to understand how EBV infection and its immune control looks in healthy virus carriers and is altered in MS patients.

Composition and tissue distribution of EBV infection in MS patients
EBV is thought to be primarily transmitted via saliva exchange (Damania et al., 2022;Farrell, 2019;Münz, 2019).This transmission is incredibly efficient resulting in more than 95% of the human adult population getting persistently infected with this virus.This happens within the first month in Sub-Saharan African countries, while around one third of the populations in Europe and the US are only infected in the second decade of life (Dunmire et al., 2018;Piriou et al., 2012).The delayed primary infection is more frequently associated with IM.From saliva, EBV most likely enters oropharyngeal secondary lymphoid tissues, such as tonsils, via transcytosis across the mucosal epithelium (Tugizov et al., 2003;Tugizov et al., 2013).In secondary lymphoid tissues EBV finds its main host cell, the human B cell, that it enters after attachment of EBV gp350 to complement receptors 1 or 2 (CD21 or CD35) with EBV gp42 attachment to MHC class II as a co-receptor (Connolly et al., 2021).The gH/gL/gB fusion machinery of EBV then allows viral envelope fusion at the cell membrane or after uptake into endosomes.EBV's capsid is then transported along microtubules to nuclear pores for injection of the linear non-methylated double-stranded viral DNA of around 172 kb into the nucleus.Within the nucleus the viral DNA circularizes into an episome that gets successively methylated and chromatinized during the next three weeks and only then lytic virus replication with infectious virion production can start (Woellmer et al., 2012).During this time only latent EBV gene products are expressed.It was found that naïve B cells of healthy virus carriers express all 8 latent EBV proteins, namely six nuclear antigens (EBNA1, EBNA2, EBNA3A, EBNA3B, EBNA3C and EBNA-LP) and two latent membrane proteins (LMP1 and LMP2), two small RNAs (EBERs) and more than 40 miRNAs of the BHRF1 and BART clusters (Babcock et al., 2000;Skalsky and Cullen, 2015).This latency III program is also found in B cells that are transformed by EBV infection into immortalized lymphoblastoid cell lines (LCLs) in vitro and in a subset of diffuse large B cell lymphomas (DLBCL).In germinal center B cells of healthy virus carriers only three latent EBV proteins are expressed, namely EBNA1, LMP1 and LMP2 (Babcock et al., 2000).The respective EBV latency IIa program is also found in Hodgkin's lymphoma of which around 40-50% is EBV associated (Shannon-Lowe and Rickinson, 2019).EBV is thought to drive B cells by activation and proliferation with the latency III program into the germinal center reaction (Babcock et al., 2000).EBV latency II then rescues them by mimicking B cell receptor and follicular helper T cell induced CD40 signaling with LMP2 and LMP1, respectively, from cell death in the germinal center reaction (Babcock et al., 2000).This allows infected cells to access the memory B cell pool, in which EBV shuts down all latent EBV protein expression to persist (Babcock et al., 1998).During homeostatic proliferation of these EBV infected memory B cells EBNA1 expression is transiently switched on because it allows for the replication of viral episomes and attaches them to condensed chromosomes during cell division for viral DNA distribution into daughter cells (Hochberg et al., 2004).These programs are called latency 0 (no viral proteins) and I (EBNA1 only), and lytic EBV replication can be initiated by the immediate early transcription factor BZLF1 from both of these (Rowe et al., 1992), most likely after B cell receptor engagement on infected memory B cells and their differentiation into plasma cells that stimulates BZLF1 expression (Laichalk and Thorley-Lawson, 2005;Reusch et al., 2015).The resulting infectious viral particles might be able to infect in an EphA2 and integrin dependent fashion polarized mucosal epithelia from the basolateral side with gH/gL and BMRF2 as the viral envelope proteins mediating this infection (Chen et al., 2018;Tugizov et al., 2003;Zhang et al., 2018).This might allow for another round of lytic replication in epithelial cells before shedding into saliva for transmission.This shedding is only observed at three weeks after EBV infection when latent infection has already allowed systemic dissemination of EBV (Dunmire et al., 2015).This life cycle restricts EBV primarily to blood circulation, secondary lymphoid organs and mucosal surfaces.
However, under certain circumstances EBV infection can spread to the CNS and might contribute to immunopathology and autoimmunity in this tissue.Along these lines it has been shown that during immune suppression by human immunodeficiency virus (HIV) infection, EBV associated CNS lymphomas can emerge, when systemic immune control of EBV is not yet deficient (Roschewski and Phelan, 2021).Furthermore, immune checkpoint blockade by antibody mediated inhibition of PD1 that increases EBV viral loads in mice with reconstituted human immune system components (humanized mice) (Chatterjee et al., 2019;Volk et al., 2020) leads to neurological symptoms in a subset of treated patients (Martins et al., 2019).In some of these patients homing of EBV infected B cells to the CNS was reported (Johnson et al., 2022).Moreover, several percent of encephalitis in children might be associated with EBV (Doja et al., 2006), sometimes following IM but possibly only requiring high viral loads that are often observed in children without IM caused by CD8 + T cell lymphocytosis (Abbott et al., 2017;Jayasooriya et al., 2015;Silins et al., 2001).Indeed during conditions of high levels of systemic EBV infection, EBV infected B cells might have the intrinsic ability to home to the CNS.This capacity was also enriched by passaging Fig. 1.Altered EBV specific immune control during prodromal, relapsing-remitting and progressive MS.Timeline of multiple sclerosis (MS) progression in most patients with a prodromal (asymptomatic) phase of 5-10 years, a relapsing-remitting disease (RR-MS) for 10-15 years developing into a secondary progressive course (SP-MS).In susceptible individuals EBV infection is thought to initiate prodromal MS.During this time period elevated EBNA1 specific antibodies can be detected in peripheral blood.At the time of first clinical manifestation, some studies suggest an elevated EBV specific CD8 + T cell response in peripheral blood, and these CD8 + T cells can be found in cerebrospinal fluid as an indication of central nervous system (CNS) infiltration during RR-MS.In peripheral blood at least some EBV specific CD8 + T cells might be functionally impaired.Finally, in post-mortem brains of MS patients an enrichment of T-bet + B cells and submeningeal tertiary lymphoid structures (TLS) of B with T cell aggregates were found.This graph was in part generated with Servier Medical Art (https://smart.servier.com).

C. Münz
of an EBV infected B cell line in immune compromised mice always retrieving the brain homing subpopulation (Soldan et al., 2021).This treatment led to epigenetic alterations with EBNA1 up-regulation.Thus, diminished but not completely defective immune control of EBV might allow EBV infected B cells to home to the brain.
Such an alteration in the distribution of EBV infected B cells might also occur during MS.Controversial data reported EBV infected B cells in mostly meningeal tertiary lymphoid structures of post-mortem brain tissue of MS patients (Moreno et al., 2018;Serafini et al., 2007).Nontranslated viral RNAs (EBERs), latent EBV gene products (EBNA2 and LMP1) and lytic viral proteins (BZLF1 and BFRF1) were detected.Oligoclonal bands in the cerebrospinal fluid of MS patients, an established hallmark of this autoimmune disease (Zipp et al., 2019), indeed point towards a substantial oligoclonal plasma cell differentiation of B cells in the CNS that might support lytic EBV replication.Along these lines, LCLs that have been generated by spontaneous outgrowth of EBV infected B cells from MS patients seem to reactivate more readily lytic replication, even so the endogenous EBV strains do not seem to dramatically differ from sequences found in healthy virus carriers (Soldan et al., 2023).To what extent potentially brain homing B cells with elevated lytic EBV replication overlap with CXCR3 + CD11c + T-bet + B cells that have been found to be enriched in post-mortem brain tissue of MS patients (van Langelaar et al., 2019), remains unclear.However, this B cell subset also correlates with EBV viral loads during viral reactivation after transplantation (van Langelaar et al., 2021) and constitutes a subset of EBV transformed B cells in LCLs (SoRelle et al., 2022).Thus, although overall viral titers for EBV do not seem to dramatically differ between MS patients and healthy EBV carriers (Lünemann et al., 2006) the composition of EBV infection with respect to tissue distribution and lytic versus latent gene expression might be associated with MS pathology.

Altered immune control of EBV specific immune responses in MS patients
As outlined above all healthy virus carriers harbor premalignant stages of EBV infection with the same viral infection programs that are also found in EBV associated malignancies (Münz, 2019).The transition into overt lymphomas is prevented by immune control, as becomes apparent during immune suppression, e.g. during HIV co-infection or after transplantation (Carbone et al., 2022;Toner and Bollard, 2022).This immune control is primarily based on cytotoxic lymphocytes, such as natural killer (NK) cells and CD8 + T cells, because individuals with defects in these immune cell populations due to individual gene mutations that lead to primary immunodeficiencies, frequently develop EBV associated pathologies (Damania and Münz, 2019;Fournier and Latour, 2021).These mutations affect the generation of these cytotoxic lymphocytes, e.g. by loss of MCM4 for NK cell deficiency (Gineau et al., 2012), their T cell receptor (TCR) signaling, e.g. by loss of IL-2 inducible T cell kinase (ITK) (Huck et al., 2009), their co-stimulation, e.g. by loss of CD27 and its ligand CD70 (Ghosh et al., 2020), their capacity to respond to cytokines, e.g.IL-27Rα deficiency (Martin et al., 2024), their cytotoxic function, e.g.loss of perforin (Katano et al., 2004), and their ability to expand and survive during EBV infection, e.g.due to loss of cytidine triphosphate synthetase 1 (CTPS1) (Martin et al., 2014).Interestingly, interferons and antibodies do not seem to be required for the immune control of EBV (Latour and Fischer, 2019).The capacity of some of the cytotoxic lymphocyte populations that are affected by these mutations to control EBV infection and associated pathologies has been demonstrated in vitro, in patients and in humanized mice.
In vitro CD4 + and CD8 + T cell clones, isolated from healthy EBV carriers and specific for latent and lytic EBV antigens, have been shown to kill autologous EBV transformed B cell lines (Taylor et al., 2015).Moreover, T cell inhibition by cyclosporin A allows EBV transformed B cells to grow out from EBV infected PBMCs of EBV seropositive donors (Bickham et al., 2003).Adoptive transfer of EBV specific T cell lines has also shown clinical efficacy, efficiently eliminating virus associated posttransplant lymphoproliferations (Vittayawacharin et al., 2024).EBNA1 or LMP1 and 2 specific T cell preparations have been used to treat EBV associated malignancies (Bonifacius et al., 2023;Icheva et al., 2013;McLaughlin et al., 2018).In humanized mice, CD4 + and CD8 + T cell depletion by antibodies leads to increased viral loads and lymphoma formation after infection (McHugh et al., 2020;Murer et al., 2019;Strowig et al., 2009;Yajima et al., 2009).Similar to the results in the regression assay and consistent with increased EBV associated lymphoma formation in patients under immune suppression, T cell inhibition by FK506 treatment also leads to increased viral loads and tumor formation in EBV infected humanized mice (Caduff et al., 2020).Furthermore, antibody mediated blocking of 2B4 or CD27 costimulation on CD8 + T cells compromises EBV specific immune control in humanized mice (Chijioke et al., 2015;Deng et al., 2021).Especially, lytic EBV antigen specific CD8 + T cell responses require CD27 for their more pronounced expansion (Deng et al., 2021).Lytic EBV infection is also controlled by NK cells in humanized mice (Chijioke et al., 2013).Early differentiated CD56 dim NKG2A + KIR − NK cells are particularly efficient in targeting lytic EBV replication and get expanded during IM (Azzi et al., 2014).Thus, cytotoxic lymphocytes are required and sufficient for EBV specific immune control but contribute to immune pathology if to strongly activated and expanded, as occurs during IM.
Indeed, in MS patients EBV specific immune control also seems to be altered.Already in the 1980s it was noted that EBV specific antibodies, directed against viral capsid antigen (VCA) or EBNAs, were elevated in MS patients (Bray et al., 1983;Larsen et al., 1985;Sumaya et al., 1980).Especially EBNA1 specific antibodies are 2-fold elevated in MS patients (Lünemann et al., 2008c) and this extends also to pediatric MS cases (Lünemann et al., 2008b).Their elevated titers are associated with a nearly 4-fold increased risk to develop MS (Olsson et al., 2017).These increased EBNA1 specific antibody levels can already be observed in the C. Münz prodromal disease phase prior to first clinical symptoms (Fig. 1) (Ascherio et al., 2001;Bjornevik et al., 2022;Levin et al., 2005) and their magnitude at first clinical presentation correlates with progression of MS (Kuhle et al., 2015;Lünemann et al., 2010).Furthermore, they seem to increase during relapses in relapsing remitting MS which constitutes the first clinical phase in most patients prior to continuous disability progression (Farrell et al., 2009).While interesting and a potential biomarker for disease activity, the function of these elevated EBNA1 antibody titers remains unclear.However, they seem to be detectable in individuals carrying MS genetic risk factors, primarily a particular MHC class II molecule, HLA-DRB1*1501, that increases MS risk 3-fold (Olsson et al., 2017;Tengvall et al., 2019).In these individuals they could be driven by elevated EBNA1 antigen reservoirs and indeed in humanized mice EBNA1 specific antibody responses seem to correlate with viral loads (Zdimerova et al., 2021).Along these lines elevated EBNA1 specific antibodies could indicate a poorly controlled EBNA1 expressing cell population and defective EBV specific immune control.HLA-DRB1*1501 positive immune compartments are indeed less efficient in the immune control of EBV infection in humanized mice despite expanding T cells to higher levels (Zdimerova et al., 2021).This is reminiscent of IM in which primary infection is also poorly controlled initially despite massive T cell expansion (Dunmire et al., 2018).IM increases the risk to develop MS 2-fold and in HLA-DRB*1501 positive individuals even 7-fold (Olsson et al., 2017).However, HLA-DRB1*1501 has not been found associated with IM (Tian et al., 2017).In addition to a consistent EBNA1 specific antibody response that however only develops after three months of EBV infection, EBNA1 is recognized by CD4 + T cells in nearly every donor (Long et al., 2005;Münz et al., 2000) while consistent CD8 + T cell responses are mainly observed in HLA-B*3501 carrying individuals (Blake et al., 1997).EBNA1 specific CD4 + T cell responses are also elevated in MS patients (Lünemann et al., 2006;Lünemann et al., 2008c).While during relapsing remitting MS no other EBV specific CD4 + or CD8 + T cell responses were found to be elevated in peripheral blood (Lünemann et al., 2006, Lünemann et al., 2008c), at first clinical manifestation elevated CD8 + T cell responses were detected that then decreased and in part even become functionally impaired (Jilek et al., 2012;Jilek et al., 2008).Some of these and especially CD8 + T cells against lytic EBV antigens might migrate to the CNS (Fig. 1) because they can be accumulated in blood upon natalizumab treatment that blocks lymphocyte homing to the brain (Angelini et al., 2013).Accordingly, TCR sequences of expanded T cell clones that recognize autologous LCLs and lytic EBV antigens were found in cerebrospinal fluid and brain parenchyma (Gottlieb et al., 2024;Holmoy and Vartdal, 2004;Lossius et al., 2014;van Nierop et al., 2016;van Nierop et al., 2017;Wuest et al., 2014).On the basis of TCR sequences a broader repertoire of EBV specific and also lytic antigen reactive CD8 + T cells was also predicted in the CNS of MS patients (Schneider-Hohendorf et al., 2022).Some of these might acquire a tissue residency phenotype in lymphocytic aggregates in the CNS (Serafini et al., 2023), and such tertiary lymphoid structures (TLS) correlate with disease activity in secondary progressive MS (Magliozzi et al., 2007).Therefore, a less well controlled reservoir of EBNA1 expressing cells might drive elevated antibody and CD4 + T cell responses against this viral antigen.While EBV specific CD8 + T cells, including EBV lytic antigen specific T cells, might be expanded at first clinical manifestation of MS in response to this poorly controlled EBV reservoir, they do not seem to stay elevated in circulation and some of these seem to accumulate in the CNS during MS continuation and possibly take residence in the brain.

Cross-reactivity of EBV specific immune responses to MS autoantigens
Some of the EBV specific CD4 + T cell responses were found to crossreact with myelin antigens (Fig. 2).The first example was a TCR that recognized amino acids (aa) 84 to 102 of the myelin basic protein (MBP) on the MS associated HLA-DRB1*1501 allele and also the EBV DNA polymerase BALF5 peptide aa627 to 641 (Wucherpfennig and Strominger, 1995).CD4 + T cells with this specificitiy were later also isolated from the CSF (Holmoy et al., 2004).Furthermore, EBNA1 specific HLA-DRB1*1501 restricted CD4 + T cells were queried for their crossreactivity with peptides of MBP, proteolipid protein (PLP), myelin oligodendrocyte glycoprotein (MOG), and cyclic nucleotide phosphodiesterase (CNPD) (Lünemann et al., 2008a).EBNA1 specific CD4 + T cell cross-reacted more frequently with these myelin antigens than with proinsulin, an autoantigen in diabetes.This occurred for EBNA1 specific HLA-DRB1*1501 restricted CD4 + T cells of both MS patients and healthy virus carriers.The myelin cross-reactive T cells produced more frequently IL-2 than the only EBNA1 specific clones.IL-2 hyperresponsive transitional naïve CD4 + T cells were also the only lymphocyte population that was significantly increased in peripheral blood of MS patients compared to their healthy twins (Ingelfinger et al., 2022).EBNA1 specific CD4 + T cells of both healthy controls and MS patients were recently found to also react to the myelin autoantigen GlialCAM (Vietzen et al., 2023a).An additional cross-reactive CD4 + T cell specificity was identified for HLA-DRB1*1501 restricted peptides from the lytic EBV antigens BMRF1 and BPLF1 cross-reacting with the CNS antigen RasGRP2 (Wang et al., 2020).Finally, HLA-DRB1*1501 restricted CD4 + T cells that are primed during EBV infection of humanized mice and recognize LCLs cross-reacted with MBP peptides (Zdimerova et al., 2021).Thus, cross-reactive CD4 + T cell specificities have been defined for EBV and myelin antigens, but it remains unclear if these are enriched in the CNS of MS patients, correlate with disease activity or are even causative for autoimmunity.
In addition to T cell cross-reactivity antibodies have been reported to cross-react to EBV and autoantigens (Fig. 2).Some of the EBNA1 specific antibodies that are elevated in MS patients cross-react with GlialCAM, anoctamin 2 or α-crystallin B (Lanz et al., 2022;Tengvall et al., 2019;Thomas et al., 2023).Antibody responses to the aa380 to 440 in the EBNA1 sequence mediate this cross-reactivity.Anoctamin 2 antibody responses synergize with elevated EBNA1 specific antibodies, HLA-DRB1*1501 expression and absence of HLA-A*0201 to increase MS risk nearly 25-fold (Tengvall et al., 2019).Similarly a-crystallin B antibodies synergize with high EBNA1 antibodies to a 9-fold increased MS risk (Thomas et al., 2023).Antibodies that cross-react with EBNA1 and GlialCAM were found in around 20% of MS patients and immunization with the cross-reacting aa386-405 peptide of EBNA1 exacerbated CNS autoimmunity in mice (Lanz et al., 2022).In addition to EBNA1, crossreactive antibodies have been detected in MS patients that recognize the lytic EBV antigens BFRF3 and BRRF2 and the autoantigens septin 9 and mitochondrial proteins, respectively (Lindsey, 2017).Since memory B cell depletion by CD20 targeting antibodies that do not affect antibody producing plasma cells or their production of oligoclonal bands in the CNS, is strongly disease modifying in MS (Merino-Vico et al., 2023) cross-reactive antibody specificities might not reflect pathogenic antibodies but be indicative of memory B cells with cross-reactive B cell receptors.These might have been primed during EBV infection and might efficiently take up and process myelin antigens to stimulate autoimmune T cell responses.

Immune control of autoreactive immune responses in MS
Pathogenic cross-reactivities might be differently immune controlled in healthy EBV carriers and MS patients.Along these lines it was recently shown that GlialCAM specific B and T cells are preferentially recognized by NKG2D + and NKG2C + NK cells (Vietzen et al., 2023a).This immune control was linked to genomic variations in EBV and human cytomegalovirus (HCMV).NKG2C + NK cells expand in carriers of persistent HCMV infection (Guma et al., 2004), and HCMV infection has been found to confer a mild protective effect against development of MS (Bjornevik et al., 2022;Olsson et al., 2017).Indeed NKG2C + NK cells are in part expanded due to stabilization of the NKG2C ligand HLA-E on the surface of infected cells by HCMV derived peptides (Hammer et al., 2018).Interestingly, it was found that healthy subjects with high GlialCAM specific antibody and T cell reactivity primarily carried HCMV strains that are efficiently able to stabilize HLA-E while HCMV strains that are compromised in this activity due to sequence variation in their UL40 gene are enriched in MS patients (Vietzen et al., 2023a).The resulting expansion of NKG2C + NK cells could protect these individuals from MS.In the absence of these protective HCMV induced NKG2C + NK cells in MS patients it was found that EBV viral strains are enriched in MS patients that carry LMP1 variants that efficiently stabilize HLA-E for inhibition of NKG2A + NK cells (Vietzen et al., 2023b).
A second protective NK cell population was identified as NKG2D + cells and healthy individuals with high GlialCAM specific T and B cell reactivity carried homozygously more frequently the high cytotoxicity mediating allele of NKG2D (Vietzen et al., 2023a).HLA-E expression on GlialCAM specific EBV infected B cells seems to be also used by EBV specific CD8 + T cells to target such autoreactive cells.HLA-E restricted CD8 + T cells specific for the lytic EBV antigen BZLF1 have previously been suggested as a protective immune cell population against EBV associated pathologies (Vietzen et al., 2023b).In MS patients compared to healthy controls, less HLA-E restricted BZLF1 specific CD8 + T cells were detected and the HLA-E*0103 haplotype that preferentially leads to the development of these EBV specific CD8 + T cells was also underrepresented in MS patients (Vietzen et al., 2023a).Interestingly, in individuals with high EBNA1 antibody levels the combination of HLA-E stabilizing LMP1 variants and the HLA-E*0101 haplotype that does not allow for the efficient development of HLA-E restricted BZLF1 specific CD8 + T cells increased MS risk more than 200-fold (Vietzen et al., 2023a).Thus, particular NK and CD8 + T cell responses might restrict auto-antigen specific B and T cell responses that are generated by EBV infection.The respective immune control might be deficient in MS patients and influenced by EBV strain variation as well as HCMV coinfection.These findings should be recapitulated in larger MS patient cohorts.Furthermore, why both healthy individuals and MS patients, despite the above described NK and CD8 + T cell differences, had similar levels of GlialCAM specific antibody responses, needs to be investigated further.

Conclusions
Recent seroepidemiological data suggest that elevated immune responses against EBV can distinguish individuals that will go on to develop MS after primary infection (Bjornevik et al., 2022).These immune responses stay elevated during disease progression and correlate with disease activity (Farrell et al., 2009).These elevated EBV specific immune responses might drive EBV specific CD8 + T cells into the CNS (Gottlieb et al., 2024;Schneider-Hohendorf et al., 2022), generate crossreactive T and B cell populations that recognize myelin autoantigens (Lanz et al., 2022;Tengvall et al., 2019;Thomas et al., 2023;Wang et al., 2020) and compromise immune control of autoreactive B and T cell responses (Vietzen et al., 2023a).The challenge now becomes to identify the Achilles heel of these altered EBV specific immune responses that can be therapeutically targeted to ameliorate MS.

Outlook
Along these lines it will be important to characterize if the altered EBV specific immune response is maintained by a reservoir of EBV infected B cells that could be specifically targeted by anti-viral therapies (Messick et al., 2019), EBV specific T cells (Ioannides et al., 2021) or antibody mediated depletion of EBV induced B cell subsets (Bogers et al., 2023).CD20 targeted B cell depleting therapies that are so successful in MS might target this reservoir but compromise also B cell immunity to many other pathogens (Merino-Vico et al., 2023).Therefore, any refinement of these therapies could significantly improve treatment of MS.

Declaration of competing interest
None.

Fig. 2 .
Fig. 2. Hypothetical model for cross-reactive and diminished autoimmunity controlling immune responses during MS.Cross-reactive antibody responses for the nuclear antigen 1 of EBV (EBNA1) and the myelin autoantigens GlialCAM, anoctamin 2 and α-crystallin B have been reported in MS patients.For some of these, cross-reactive CD4 + T cell responses have also been observed (EBNA1 and BALF5 with myelin basic protein (MBP) and GlialCAM, as well as BMRF1 and BPLF1 with RasGRP2).These might be stimulated by viral and autoantigen presentation on MHC class II of the respective B cells for stimulation of the T cell receptor (TCR) of these cross-reactive T cells.Cross-reactive antibody production and T cell responses can be targeted by NKG2C + NK cells via inflammation mediated HLA-E up-regulation being recognized by CD94/ NKG2C and by NKG2D + NK cells via stress-induced NKG2DL up-regulation.EBV infection also regulates HLA-E mediated immune surveillance.Viral latent membrane protein 1 (LMP1) up-regulates HLA-E to inhibit NK cells via CD94/NKG2A.During lytic EBV replication BZLF1 peptides can also stabilize HLA-E for recognition by HLA-E restricted BZLF1 specific CD8 + T cells.This graph was in part generated with Servier Medical Art (https://smart.servier.com).