Molecular Mimicry as a Mechanism of Autoimmune Disease

Brief History of Molecular Mimicry

Over 30 years ago, molecular mimicry by either a virus [18] or bacteria [23] was hypothesized to initiate and exacerbate an autoimmune response through sequence or structural similarities with self-antigens. Currently, molecular mimicry is the prevailing hypothesis as to how viral antigens initiate and maintain autoimmune responses which lead to specific tissue damage [18]. Initial work by Fujinami, Oldstone, and colleagues identified mouse antibodies to measles virus and herpes simplex virus (HSV-1) obtained from antibody-secreting B cell clones [18]. These antibodies were reactive to both intermediate filaments of normal cells and the proteins of measles virus and HSV-1, thereby demonstrating a relatedness between host and viral antigens [18]. Further work by Fujinami and Oldstone used myelin basic protein (MBP), a nerve sheath protein containing an encephalitogenic T cell epitope in rabbits. The hepatitis B virus polymerase (HBVP) protein was found through computer analysis to share six consecutive amino acids with the encephalitogenic MBP epitope [16], and when rabbits were sensitized with either MBP or HBV peptides, the rabbit’s tissue serum reacted against MBP. Further, rabbits sensitized with the HBVP peptide developed central nervous system (CNS) pathology similar to rabbits sensitized with whole MBP protein or the MBP peptide [16]. Importantly, the rabbits sensitized with HBVP did not contract hepatitis but still developed encephalomyelitis and presented with a similar pathology as MBP-sensitized mice. These experiments were the first experimental demonstration of molecular mimicry, whereby a microbial peptide with similar amino acid sequences to the self-peptide was able to activate autoreactive T cells and subsequently cause specific tissue damage.

Relationship Between Molecular Mimicry and Autoimmune Diseases

Immune cells of the adaptive immune response are specifically activated, but the hallmark of autoimmunity is the dysregulation of the immune system, especially T and B cells recognizing self-antigens as foreign. The ability of T cells to evade central (thymic selection) and peripheral (Tregs) mechanisms of tolerance is evident by the large number of T cell-mediated human autoimmune diseases, such as type-1 diabetes, systemic lupus erythematosus, rheumatoid arthritis, and multiple sclerosis (MS) [24–28]. Molecular mimicry has been implicated in the pathogenesis of many of these autoimmune diseases including MS, spondyloarthropathies, Graves’ disease, and diabetes mellitus [16, 29, 30]. In the case of MS, it has been hypothesized that certain viruses, such as Epstein–Barr virus (EBV), share sequence homology with antigenic structures in the CNS [31].

Activation of an autoimmune response could be enhanced by a variety of other, albeit, non-mutually exclusive non-specific mechanisms including bystander activation and superantigens. The difference between other non-specific mechanisms that initiate autoimmunity and molecular mimicry is that microbial mimics specifically direct the immune response towards a tissue and/or organ. Originally, T cell recognition was postulated to be highly specific and cross-reactivity was thought to be a rare phenomenon. However, the structural requirements for peptide binding by MHC class II molecules that are presented to T cells were found to be based on amino acid properties, and amino acids sharing similar chemical features were able to bind at the same MHC peptide binding groove, thereby demonstrating that binding motifs were degenerate with only a small sequence needed for TCR recognition [32–34]. An illustration of TCR degeneracy was shown by Wucherpfennig and Strominger [35] using chemically related synthetic peptides mimicking the MBP(85–99) epitope that were incubated with human MBP(85–99)-specific T cell clones which were then tested for reactivity. Of the T cell clones that responded to the synthetic peptides, only eight of the 129 synthetic peptides were recognized by the T cell clones, and only one of the synthetic peptides that induced a response was clearly similar to MBP(85–99) [35, 36]. Therefore, these studies clearly demonstrate TCRs binding to a spectrum of specific peptides that is based upon structural relatedness, termed poly-specificity [37]. This flexibility exhibited by TCR binding and the existence of pathogens that share sequence or structural similarities with self-antigens could be one reason why investigators have been unable to conclusively associate a specific virus with autoimmune diseases, such as MS (reviewed in [4]).

Linear sequence matches in amino acid motifs is not the only criteria for mimicry [32]. It has been hypothesized that self-reactive immune cells are primed by molecular mimicry and bystander activation, thereby sensitizing the immune cells and leading to a “fertile field” but no apparent disease. Subsequent environmental insults could induce these sensitized autoreactive cells to cause an autoimmune disease. Work from our laboratory demonstrated that recombinant viruses having molecular mimicry with self-CNS antigens were unable to initiate an autoimmune disease individually [38]. However, infected mice that were subsequently challenged, after viral clearance, with a non-specific immunologic insult developed disease [38]. Further, subsequent experiments showed that conventional inflammatory responses to specific pathogens were able to induce disease in animals primed with a molecular mimic to a CNS antigen [39]. Therefore, not only is the priming of the immune system necessary for an autoimmune disease but the milieu to which the primed immune cells are exposed is an important factor in initiating an autoimmune disease. Animal models of various autoimmune diseases have explored the role of molecular mimicry as a contributing factor (Table 2).

The use of transgenic (tg) mice expressing virus proteins as transgenes in specific organs has been an important model for providing evidence for molecular mimicry. The expression of lymphocytic choriomeningitis virus (LCMV) viral antigens in pancreatic islet cells and the subsequent cross of this tg mouse with a TCR-tg mouse specific for LCMV glycoprotein resulted in an animal that only developed autoimmune disease if virally infected [40, 41]. These results demonstrated that “self”-reactive T cells are present in the periphery and the immune cells appear to remain quiescent until an appropriate signal (viral infection) triggers the T cells to respond.

Dual TCR and How This Impacts Our Interpretation of Molecular Mimicry

There are a variety of non-mutually exclusive factors that lead to a fully activated T cell, such as the quantity of peptide–MHC presented on the surface of antigen-presenting cells and TCR avidity. The interaction between the peptide–MHC and TCR is critical for the initiation of an adaptive immune response and clearance of a pathogen [15]. In order for T cells to reach maturity, the T cell goes through a number of developmental checkpoints leading to somatic recombination of various gene segments. The TCR α- and β-chains are generated by V-D-J recombination, which leads to αβ TCRs expressed on the surface of T cells [42, 43]. Although it was believed that T cell signaling was mediated by a single antigen receptor, recent evidence demonstrates that T cells are capable of expressing functional dual Vα TCRs at a frequency of approximately 30% in humans and 15% in mice; however, an accurate number of dual specific TCRs is lacking due to the limited availability of anti-Vα monoclonal antibodies (mAbs) [44–46]. Interestingly, in contrast to the high frequency of dual expressing Vα T cells, only 1% of humans and 5–7% of mice express two β-chains due to allelic exclusion mechanisms, but the frequencies of dual Vβ TCRs have been found to be higher with age and in TCR-tg mice [47–49]. Expression of multiple TCR Vαs on the surface of a T cell is the result of simultaneous rearrangement of both TCRα loci during thymocyte development [50–52]. Further, TCR Vβ-chains preferentially bind to certain Vα-chains leading to differential expression of chimeric TCRs on the surface of T cells [51, 53, 54].

Due to the heterogeneity of TCRs normally expressed in the periphery of humans and mice, TCR-tg mice have been used to track and determine the fate of T cells expressing dual TCRs. The use of TCR-tg mice has led to the identification of a potential role for dual TCRs in a variety of conditions including graft-versus-host disease, human immunodeficiency virus infection, inflammatory bowel disease, T cell leukemia, T cell lymphoma, and MS [55–61].

The expression of dual TCRs by the same T cell has been proposed to be a potential mechanism for autoimmune disease. Normally, high avidity self-reactive T cells are thymically depleted, but it has been hypothesized that the expression of a self-TCR on a T cell is lower when presented in the context of a second TCR, thereby providing a cover for high avidity self-TCRs from both central and peripheral tolerance. Blichfeldt et al. [62] demonstrated that dual tg-TCRs, which have lower expression of each TCR on the surface of a T cell, needed higher concentrations of peptide, presented by MHC, to induce a similar T cell proliferative response compared to a single receptor T cell.

A potential role of dual TCRs in autoimmunity is in the rescue of autoreactive T cells from thymic selection. For example, the double tg mouse for autoimmune diabetes, in which the mice express a TCR specific for peptide 111–119 of hemagglutinin (HA) (TCR-HA) under the control of the rat insulin promoter and develop spontaneous diabetes and insulitis [63], were used to determine how T cells could escape tolerance mechanisms even if the antigen was ubiquitously expressed [64]. Low expressing TCR-HA co-expressing T cells were more effective at transferring diabetes than TCR-HA high dual TCRs, suggesting that the surface level expression of a dual TCR can be modulated by a second TCR expressed on the same T cell, thus “escape” of autoreactive T cells could be the first step in an autoimmune disease.

The “trigger” of an autoimmune disease could be linked to environmental insults, such as viruses. A T cell co-expressing TCRs specific for a self-antigen and a foreign antigen could potentially allow for autoreactive T cells to be activated if the host is exposed to that foreign antigen. The activation of a subset of T cells could than lead to tolerance being broken and the initiation of an autoimmune disease if these T cells experienced a particular organ or tissue that expressed the self-antigen for the other TCR expressed at the surface of the T cell. In support of a role for dual TCRs in autoimmune diseases, work performed in our laboratory characterized autoreactive CD8⁺ T cells isolated from the spleens of Theiler’s murine encephalomyelitis virus (TMEV)-infected SJL/J mice [65]. In vitro assays testing CD8⁺ T cell killing activity found a population of CD8⁺ T cells that killed uninfected syngeneic cells [65]. Adoptively transferring these TMEV-specific autoreactive CD8⁺ T cells into non-infected SJL/J mice caused CNS pathology [65]. Further support for the importance of the mechanism by which viral infection could induce an autoimmune disease through dual TCR-expressing T cells was performed by Ji et al. [61] using MBP(79–87) TCR-tg mice [66]. Cytometric phenotyping, in vitro CD8⁺ T cell killing assays, and adoptive transfer experiments were used to track the expansion and killing capacity of Vα8Vβ8 MBP(79–87)-specific TCR and Vα8Vβ6-vaccinia virus-specific TCR. Infection of these tg mice with vaccinia virus induced autoimmune disease, thus demonstrating a virus triggering an autoimmune disease through dual TCR expressing T cells [61]. Although several tg TCR β-chains have been described on peripheral T cell [61, 67–70], there is no evidence that co-expression of dual TCRs leads to autoimmunity without the use of TCR-tg mice. As described above, current work in our laboratory has characterized TMEV-specific autoreactive CD8⁺ T cell clones derived from a wild-type animal, and these autoreactive TMEV-specific T cell clones express dual TCRs (manuscript in preparation). Importantly, we were able to induce CNS pathology in naïve SJL/J mice by adoptively transferring the TMEV-specific clones. Although further work is needed in order to identify the self-antigen that activates these CD8⁺ T cells, to our knowledge these results are the first demonstration of an autoimmune disease initiated by a dual expressing TCR characterized in the virus’ natural host.

Taken together, three possible mechanisms could explain how the dual reactivity of the TCR may play a role in autoimmune diseases (manuscript in preparation). The first mechanism is molecular mimicry, whereby the induction of an autoimmune response to self is due to a single TCR recognizing both a virus and a self-antigen. The second mechanism is the expression of dual TCRs on a single T cell, where one TCR is able to recognize a microbial antigen and the other TCR recognizes self. The third mechanism involves a T cell expressing chimeric TCRs generated from either a single Vα combining with two different Vβs or a single Vβ combining with two different Vαs, resulting in a T cell with the potential of expressing two different chimeric TCRs specific for a self-antigen and a foreign antigen.