DeutschClear Cookie - decide language by browser settings
Schubart, Anna (2002): Aspects of the Immunobiology of Myelin Oligodendrocyte Glycoprotein (MOG)-induced Experimental Autoimmune Encephalomyelitis (EAE). Dissertation, LMU München: Faculty of Biology



This study investigated the immunobiology of MOG-induced EAE in the DA rat, an animal model, which reproduces the immunopathology of the type II MS lesion (Lucchinetti et al., 2000). A newly established immunisation protocol results in a highly synchronised biphasic form of EAE, which mimics the disease course of secondary progressive MS, albeit in a strongly abbreviated time course (Figure 3.1.1). This study demonstrates that MOG-specific autoantibodies are responsible for initiating clinical relapse and driving disease progression. On the background of mild, sub-clinical inflammatory activity in the CNS, pathogenic antibodies enter the CNS and mediate demyelination, a process that in turn amplifies the local inflammatory response (Figure 3.1.14 A). It should however be noted that lethal clinical relapses may also occur in the absence of a pathogenic antibody response if an inflammatory lesion develops in a region of the CNS that is particularly sensitive to damage, or where it may perturb vital functions, such as the brain stem. Although antibodies have been shown to amplify the severity of ongoing clinical EAE (Schluesener et al., 1987; Linington et al., 1988; Lassmann et al., 1988), firm evidence for a role in driving relapse and disease progression was missing. This study has now established this principal, which in all probability is relevant to our understanding of the pathogenesis of severe, steroid non-responsive relapses in MS patients. However, this model of EAE is an artificial system, in which the role of antibody is only apparent because of the different kinetics of MOG-specific T and B cell responses. In MS we still have to answer two crucial questions, namely the identity of the autoantigens targeted by the demyelinating antibody response, and the factors that may trigger this response. MOG is the only myelin protein known to initiate a demyelinating antibody response in EAE, and MOG-induced EAE has provided a valuable tool to identify the role of pathogenic autoantibodies in immune mediated demyelination. However, there is a major discrepancy between the proportion of MS patients with pathogenic MOG-specific antibodies in their circulation (5%; Haase et al., 2000) and the frequency of patients with pathological changes suggestive of antibody-mediated pathomechanisms (>50%; Lucchinetti et al., 2000). This discrepancy may in part be accounted for by the absorption of the pathogenic antibodies into the CNS, which will lead to a dramatic reduction of the antibody titre in the periphery, as demonstrated in section of this study. On the other hand, it is unlikely that MOG is the only target autoantigen, which is exposed on the myelin surface and can therefore initiate a demyelinating autoantibody response. The identification of potential targets is a prerequisite to develop diagnostic kits to identify those patients with pathogenic autoantibody responses and then provide an appropriate therapy such as plasma exchange, or immuno-absorption. As demonstrated in this study, DNA vaccination using a plasmid encoding a myelin antigen is one approach to generate high titre autoantibody responses directed against the native protein. The pathogenicity of this antibody response can then be assayed in the same animal by inducing EAE. This method circumvents problems such as purity, yield and denaturation, all of which complicate any study using antigens isolated from the CNS or generated using recombinant technologies. Coupling this approach to a proteomics based analysis of the myelin membrane and reverse genomics to identify candidate gene products provides the means to map out those protein antigens that can be targeted by a demyelinating autoantibody response. The feasibility of this concept is currently being tested in the rat using PLP and MAG as myelin components that may in certain circumstances provoke a pathogenic autoantibody response. Such an analysis will, however, not detect pathogenic antibody responses to glycolipid antigens, which are major target autoantigens in a number of diseases affecting the peripheral nervous system such as Guillain Barré syndrome (GBS). In GBS a pathogenic antibody response to gangliosides appears to be triggered by infections with particular serotypes of Campylobacter jejuni (Fredman, 1998; Willison and O´Hanlon, 1999). In the majority of patients these antibody responses are an acute phenomenon and disappear as the patients recover (Hahn, 1998). It is conceivable that a similar mechanism is responsible for the initiation of severe relapses in some MS patients, if an infection triggers a cross-reactive antibody response to a surface glycolipid epitope. This would induce an episode of acute CNS demyelination that would not be immediately responsive to immunosuppressive therapy, as tissue damage and amplification of the local inflammatory response would be driven by the pre- existing antibody response. Analysis of the autoantibody responses in MS should therefore be extended to examine lipid as well as protein autoantigens. Such studies should also not be restricted to myelin, but also address the question of responses to other structures such as the axon and oligodendrocyte progenitor cells. Such autoantibody responses are however only conditionally pathogenic, in other words their pathogenic potential is only expressed if they can enter the CNS across the blood brain barrier (BBB)(Litzenburger et al., 1998; Bourquin et al., 2000). In EAE the inflammatory insult to the CNS is responsible for the disruption of BBB function and the entry of antibody into the nervous system. MS is characterised by repeated episodes of CNS inflammation but what initiates and maintains this response is unclear. The observation, that DA rats develop a similar, although eventually self-limiting response in the CNS after immunisation with MOG-peptide in CFA provides a model to investigate the immuno-regulatory deficit(s) responsible for chronic CNS inflammation. The disease model is very reproducible with >90% of animals relapsing after peptide immunisation as opposed to <40% after immunisation with MBP in IFA (Lorentzen et al., 1995). This will make it feasible to use genetic methods, such as disease induction in congenic and intra-MHC congenic rat strains and whole genome screens in F2 backcrosses, to identify genetic loci responsible for this defect. The understanding of the mechanisms involved may help to identify new targets for therapeutic strategies concerning chronic inflammatory diseases like MS and rheumatoid arthritis (RA). The second part of this thesis investigated a very different aspect of the autoimmune response to MOG, the consequences of immunological cross-reactivity with BTN, a major component of the milk fat globule membrane. The demonstration of cross-reactive T cell responses between MOG and a dietary antigen opens a new perspective for the aetiology of MS, since former investigations of environmental influences were concentrated on molecular mimicry with microbial peptides (Bray et al., 1983; Wucherpfennig et al., 1995; Challoner et al., 1995; Gautam et al., 1998; Ufret-Vicency et al., 1998; Burgoon et al., 1999). Epidemiological studies identified a link between milk consumption and other dietary factors and MS (Butcher, 1986; Malosse et al., 1992; Lauer, 1997), but the identity of the mechanistic basis in the immune system was unknown. It would be naïve to imagine that milk in the diet would per se induce an auto-aggressive response to MOG and thereby trigger MS. Indeed, disease induction is now thought to involve the chance interactions of several environmental factors on a susceptible genotype. In the case of the cross-reactive pair of antigens BTN/MOG, BTN in the diet would normally induce oral tolerance to the cross-reactive epitope, but this may be broken either by early post-natal exposure to bovine milk products (Miller et al., 1994), gastro-intestinal infections (Hornquist and Lycke, 1993; Weiner 1997) or a combination of both (as discussed in Introduction 3.2). However, would this combination of effects be sufficient to induce an inflammatory autoimmune mediated response in the CNS? Any prediction would at this time be premature. Analysis in the DA rat revealed that the cross-reactive repertoire in this model is complex and involves multiple clonal expansions. Moreover, the sequence of the BTN peptide is not identical to the corresponding MOG sequence and the cross-reactive BTN peptide acts as an APL. In view of the degeneracy of TCR - peptide/MHC recognition, the BTN peptide may initiate a range of responses ranging from superagonistic (Vergelli et al., 1997) to antagonistic (DeMagistris et al., 1992) in the different T cell clones. The identification of TCR -chains used by the cross-reactive T cells is a first step towards generating a transgenic animal model. This may allow us to examine the immunopathological consequences of cross-reactivity involving a dietary antigen and the impact of manipulating the gastro-intestinal flora on the immune response. Whether or not this is relevant for the aetiology of MS is uncertain. Certainly T cell cross- reactivity between the two proteins is very limited and as yet was only demonstrated in the context of the RT1av1 rat MHC haplotype (Stefferl et al., 2000). No cross-reactive T cell response was detected in LEW and BN rats (Stefferl et al., 2000), as well as in SJL/J, C57/BL6, DBA.1 and CBA.1 mice (Schubart and Wissing, unpublished results). In addition, despite the presence of regions with a high level of sequence identity between the two proteins it was not possible to induce a significant cross-reactive antibody response in the DA rat (section 3.2.7). Why is this? In the course of this study it became apparent that BTN is a member of a family of structurally related gene products that are termed the BTN-, or extended B7- gene family. The N-terminal IgV-like domain of all these proteins exhibits a high degree of amino acid sequence identity with both MOG and BTN (Henry et al., 1999; Stammers et al., 2000; Rhodes et al., 2001; see discussion of chapter 3.2), and members of this family are expressed in a variety of organs. It is possible that during B and T cell maturation in the bone marrow and thymus, cross- reactive peptides derived from these proteins will eliminate many clones that would otherwise cross-react with MOG reactive T and B cells, and only those cells which escape this network of tolerogenic stimuli enter the periphery. As the sequences of the BTN-gene family members in the rat are currently unknown (with the exception of MOG), the identification of the extracellular domains of rat BTN provides the first opportunity to test this hypothesis. The oral consumption of milk during suckling should induce rat-BTN-specific suppressor T cells, characterised by low proliferative responsiveness and the secretion of IL-10 and TGF- (Weiner et al., 1994). The identification and characterisation (epitope specificity) of these suppressor T cell responses in the different rat strains (LEW, BN and DA) may provide an explanation why MOG-specific T cell responses are only poorly encephalitogenic in Lewis rats and might help to elucidate the mechanisms of the development of tolerance in the newborn animals.