EBV, autoimmunity and dual T-cell receptors

EBV

EBV infects 90% of the world’s population, with the majority undergoing seroconversion before the age of 10 years. Primary infection of B cells in the oropharynx leads to a general infection of the circulating blood B-cell pool, normally in the range of 0.1–1%; however, in extreme cases, this can reach up to 10% of all circulating B cells and up to 50% of all circulating memory B cells. In infectious mononucleosis (IM), mononuclear cells can amount to 60% of all white blood cells, resembling acute leukemia. Clinical symptoms disappear with the decline of the activated T cells.

EBV features additional fascinating properties: it stays latent in memory B cells for the lifetime of the human host. The virus is able to immortalize B cells in vitro and drive the growth of B cells in severely immunosuppressed patients with post-transplant lymphoproliferative disease. Its causal association with malignant tumors (e.g., Burkitt’s lymphoma, Hodgkin’s lymphoma, nasopharyngeal carcinoma, gastric carcinoma and other cancers) is well established. Viral latent gene expression underlies a strict epigenetic control that is achieved through covalent histone marks and DNA methylation of CpG motifs within viral promoters. In EBV-associated malignancies, not only the viral genome, but also the host cell genome carries extensive epigenetic alterations, such as an overall genomic hypomethylation and hypermethylation of CpG islands at tumor suppressor gene promoters. The tumor-associated epigenetic profiles seem to reflect virus specificity to some degree [1,2]. Furthermore, a causal association of EBV infection with autoimmune diseases, such as systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), multiple sclerosis (MS) and other less frequent autoimmune diseases, has been suspected or assumed for decades. EBV may trigger autoimmunity through diverse mechanisms, such as generating autoreactive T-cell clones during IM, inducing cross-reactive T-cell clones through antigenic mimicry or through immortalization of pre-existing autoreactive B cells [1,3]. Still, there is no formal proof for the causality of EBV infection in the induction of autoimmune diseases [4]. Host factors, for example dysfunctional regulatory T cells, might predispose both for autoimmunity and for a hyper-reactive, but nonfunctional immune response against EBV. A widespread application of an early childhood EBV vaccine might be able to establish causality in the future. However, even without an EBV vaccine on the market, there is intriguing new data with regard to the association of EBV with autoimmune diseases.

Myasthenia gravis

Myasthenia gravis (MG) has been added to the list of autoimmune diseases with a possible EBV association. MG is the most well-characterized antibody-mediated autoimmune disease [5]. In most cases, the symptoms are due to autoantibodies against parts of the neuromuscular junction, usually against the acetylcholine receptor, and less frequently against the muscle-specific kinase receptor. The thymus is the most likely candidate for the site where MG is initiated and maintained. Acetylcholine receptor proteins are expressed in the thymus, and most MG patients have lymphoid hyperplasia in the thymus, either in diffuse form or resembling the active germinal centers of lymphoid follicles. Removing chronic inflammation through thymectomy is a treatment option that may alleviate symptoms. The molecular mechanisms that induce and sustain the autoimmune reaction, however, are not known. Cavalcante et al. offer a potential answer to the question of how immune reactivity against the neuromuscular junction may be maintained. They regularly found EBV-infected B cells in thymic tissue from MG patients with all MG subtypes, while thymus tissue obtained from non-MG patients in the course of heart or other surgery did not contain EBV-positive B cells [6,7]. Those thymus-infiltrating B cells expressed the EBV latency products EBER1, EBNA1 and LMP1, corresponding to the viral latency class II gene expression program, but also the lytic gene product, BZLF1. Whether EBV-infected infiltrating B cells are in any way causal for MG initiation or perpetuation, or whether impaired antiviral immunity is a consequence of MG disease or immune suppressive MG therapy, remains to be determined at this point. An argument for the former possibility may be provided through the case of a 22-year-old woman who developed MG together with juvenile-onset diabetes immediately after suffering IM [8]. This IM case, however, is not established beyond doubt. IM was diagnosed with a positive monospot assay and positive IgM test, whereas anti-EBV IgG was negative. Since IgM and monospot may react nonspecifically, an immediate follow-up serology with a positive anti-EBV VCA-IgG is required for proving a past IM, but was not presented in this case [8].

Contradictory data to Cavalcante et al. were presented by other groups who did not find EBV-encoded RNAs and proteins in MG thymus tissue at all [9,10]. Since appropriate control experiments were presented, the controversy is unresolved at this point. Are there geographical differences in EBV infection rates of MG thymus? Or is it a technical issue? Whatever the reason for the disagreement, it may be helpful to exchange thymic tissue sections between laboratories [11].

Systemic lupus erythematosus

SLE is a chronic inflammatory multisystem autoimmune disease, which mostly attacks the skin, joints, kidneys and CNS, and can damage any other organ. It is characterized by the presence of IgG autoantibodies against nuclear antigens. Of all infectious agents, seroreactivity to EBV is most closely associated with SLE. Several case reports describe the beginning of a SLE immediately after IM. In the acute phase of IM, the spliceosomal Sm-B´ antigen is recognized by antibodies cross-reactive with EBNA1. Additional cross-reactivities have been described in early IM that resemble SLE serology [12]. A total of 99% of SLE patients have antibodies against EBV, while only 90% of the general population is seropositive. This difference is even more pronounced and highly significant in children and young adults. While 99% of young SLE patients are seropositive, only 70% of the general age cohort reacts to EBV. Anti-EBV antibodies and viral load are elevated in SLE patients. Furthermore, SLE patients have a higher frequency of EBV-reactive CD4⁺ T cells, which secrete IFN-γ, but not CD8⁺ T cells. Therefore, EBV has become a prime suspect as a potential SLE trigger. However, the frequency of CD8⁺ T cells that recognize EBV is not different between SLE patients and healthy individuals, but T cells of SLE patients secrete less IFN-γ and are thus less functional [1,3]. Therefore, the question is again whether EBV is the cause or consequence of SLE disease, or whether common host factors support the development both of SLE and of a dysfunctional immune response to EBV infection. Larsen et al. set out to answer this question. They found that the function of CD8⁺ T cells of SLE patients is impaired upon polyclonal stimulation or upon stimulation with EBV, but not with CMV. T-cell dysfunction correlated with the upregulation of the immune inhibitory receptor PD1 in many CD8⁺ T-cell specificities [13]. Since PD1 deletion alleles correlate with a higher SLE susceptibility in affected families, PD1 upregulation seems to be a protective counter-regulation against inflammation in SLE [14]. Furthermore, Larsen et al. found that the majority of SLE flares are followed by increases of viral serum load, but not vice versa[13]. Thus, they suggest that the increased EBV load in SLE patients is a consequence of B-cell activation that again is due to disease activity. Increased EBV activity may then enter into a positive feedback loop that keeps SLE active. Thus, antiviral drugs might alleviate SLE symptoms. However, more extensive longitudinal studies are required to address the alternative perspectives on EBV and SLE [13].

Rheumatoid arthritis

RA is a very widespread autoimmune disease that affects approximately 2% of the world’s population. Inflammatory T and NK cells infiltrate the articular synovia. The synovia of RA patients is frequently EBV-infected. EBV proteins mimicking MHC class II peptides seem to play a pathogenic role. Anticyclic citrullinated peptide antibodies that react with the Gly-Ala repeat of EBNA1 and with cyclic citrullinated synovial fiber proteins are highly specific and predictive for RA. Anti-EBV antibodies, EBV-infected cell number and the viral genome load are increased in the blood of RA patients, and antiviral treatment occasionally ameliorates RA symptoms. Thus, indirect evidence implicates EBV in the development of RA [1,3].

A severe combined immune deficiency mouse model reconstituted with a human immune system gives further support to a causal role of EBV in inducing RA [15]. Inoculation of those mice with the EBV strain Akata elicited a disease resembling human RA in more than 60% of the infected mice, but not in uninfected controls. Symptoms included erosive arthritis with pannus formation and synovial infiltration of inflammatory cells. Rheumatoid factor, anticyclic citrullinated peptide antibodies and cross-reactive anti-EBV antibodies were not detected, whereas an abundant anti-EBV T-cell response was found. It is interesting to note that, although only few EBV-infected cells could be detected in the synovial membrane of the affected joints, EBV-encoded small RNA (EBER) in situ hybridization revealed numerous EBV-infected cells in the bone marrow near the inflamed joints [16]. Because the nontranslated EBER1 RNA is released from EBV-infected cells and activates Toll-like receptor 3-bearing cells [17], one may speculate that immune activation by EBER1 may contribute to the development of arthritis in a paracrine manner. Thus, EBER-producing EBV-positive cells most probably need not be situated within the synovial membrane to elicit pathological alterations characteristic for RA.

Multiple sclerosis

MS is a common inflammatory demyelinating CNS disease that is mediated by myelin-specific autoreactive T cells. A controversy analogous to surrounding MG exists for MS. Serafini et al. regularly found ectopic follicle-like aggregates of submeningeal EBV-infected B cells, viral RNAs and proteins in sections of MS brain tissue [18,19]. This finding was not confirmed by other groups, who did not find ectopic follicles, EBV DNA or viral gene products in MS lesions [20–23]. Again, the discrepancy may be explained by technical differences and awaits resolution. The isolation of single cells through laser capture microdissection was reported to confirm the earlier finding of EBV-infected B cells in MS brains [24]. Whatever the outcome, there is a clear epidemiological link between EBV infection and MS, which has mostly been explained by molecular mimicry or bystander activation leading to an autoimmune response. MBP and the viral protein EBNA1 share B- and T-cell epitopes that may induce MBP-reactive EBNA1 antibodies, which are frequently found in the cerebrospinal fluid of MS patients. Furthermore, MBP and the viral protein BALF5 share common T-cell epitopes [1,3]. Therefore, while other viruses and microbes, as diverse as Chlamydia pneumoniae, Borrelia burgdorferi, HSV, varicella zoster virus, human CMV, human herpesvirus 6, measles virus, mumps virus, rubella virus, human polyomavirus JC-virus, Torque teno virus and human endogenous retroviruses have been suspected to initiate MS, EBV is clearly the prime candidate [1,4]. Again, there is no formal proof for the causality of EBV infection. Dysfunctional Tregs might predispose both for autoimmunity and for a hyper-reactive, but nonfunctional immune response against EBV. A recent paper by Tzartos et al. supports the association of latent EBV infection with innate immune activation in active MS lesions [26].

Dual T-cell receptors with dual specificities

Dual T-cell receptors (TCRs) may help to clarify the question of EBV causality for autoimmunity in a more general way. In a transgenic mouse model for MS, Ji et al. found that tolerance could be broken through infection with a virus that did not express MBP peptides or peptides with MBP homology [25]. The mechanism for the activation of autoreactive CD8⁺ T cells turned out to be the simultaneous expression of both endogenous and transgenic TCR genes. MBP was recognized by the transgenic surface receptor, while clonal expansion of the autoreactive T-cell clone was induced by engagement of endogenously expressed virus-specific surface receptors [25]. The breaking of immune tolerance through rare CD8⁺ T cells expressing dual TCRs may not only offer a mechanism for how primary EBV infection could sometimes trigger MS, but would also give an explanation for the multitude of pathogens that have been accused of eliciting MS. The very strong and partially nonspecific immune response induced through EBV primary infection may have a greater chance of activating dual-specific T cells than all other suspected bacteria and viruses, but this would not exclude other pathogens. Dual TCRs may, of course, play a role in autoimmune diseases other than MS, and therefore explain the epidemiological association between autoimmunity and EBV in general. Thus, an early childhood EBV vaccine should be useful in curbing not only EBV-associated malignancies in high-risk areas, but also autoimmune diseases. In analogy with EBV-associated malignancies, a comprehensive analysis of EBV-specific epigenetic alterations in immune cells may strengthen the presumed causal relationship between EBV infection and autoimmunity. Furthermore, the epigenetic mechanisms involved in allelic exclusion between the single surface-expressed TCR and all the other normally nonexpressed TCR genes may attract increased attention.