ABSTRACT
Myasthenia gravis (MG) is a prototypic autoimmune disease that fulfills the strict criteria for an autoantibody-mediated disorder against a known target autoantigen, the AChR at the postsynaptic membrane of the neuromuscular junction (NMJ). First of all, immunization of several animal species with the presumed autoantigen, the AChR, can induce a disease called experimental autoimmune MG (EAMG) that is similar to human MG (1). Second, passive transfer of autoantibodies to experimental animals can induce MG symptoms (2). Third, antibodies can be identified at the presumed pathogeneticallyrelevant site of the disease, the NMJ and AChRs isolated from mysthenic muscle have IgG bound to the receptors (3). Furthermore, the patients often have a family history of autoimmune diseases, they respond to plasmapheresis, and autoantibodies can pass across the placenta from mother to child with the eventual consequence of congenital autoimmune MG or even intrauterine MG resulting in arthrogryposis multiples congenita (4). Knowledge about the molecular structure and structure-function relations of the main target autoantigen of MG, the AChR, is now far advanced (5,6), as the understanding of the structure and function of the NMJ (7). Anti-AChR autoantibodies impair neuromuscular transmission by complement-mediated destruction of the NMJ, by increasing AChR turnover following AChR cross-linking, and by interference with ion channel function (8). Anti-AChR-reactive T-cells of the CD4+ subset are also crucial for the pathogenesis of MG by providing help to anti-AChR antibody-producing B-cells. Given that CD8+ T-cells may be important for the pathogenesis of EAMG (9), a role of CD8+ T-cells with specificity for antigenic structures of the postsynaptic side of the NMJ has recently been considered to play a role in the maintenance of human MG. By direct cytotoxic attack on the NMJ, CD8+ cells may provide continuous release and supply of autoantigen (e.g. AChR) for the ongoing interaction of antigen presenting cells (APCs), autoreactive CD4+ T cells and B cells (10). A genetic predisposition to MG on the MHC level has long been recognized in a majority of patients, but other genetic prerequisites and etiologic mechanisms involved in the activation of anti-AChR T-cells are less well understood, with the exception of certain MG-inducing drugs like D-Penicillamine (11). Likewise, the exact mechanisms by which the heterogeneous thymic pathologies are involved in the pathogenesis of MG are only partially resolved, and what gives rise to the various thymic alterations is unknown (12). In thymic lymphofollicular hyperplasia (thymitis) there is good evidence that the autoimune process occurs completely inside the thymus and might be elicited there: AChRs expressed on thymic myoid cells are presented by APCs to B-cells that produce anti-AChR autoantibodies. By contrast, CD4+, AChR-reactive T cells, which, in turn, provide help to it appears that there is no active autoimmune process going on inside the vast majority of thymomas and in thymic atrophy. In particular, autoantibody production is virtually absent inside these thymic alterations. Instead, there is data from thymoma patients suggesting that abnormal
microenvironments in neoplastic thymuses might result in the generation of insufficiently tolerized T cells. After export to the extratumorous immune system, thymoma-derived T cells might gradually replace the normally tolerant by a more autoimmunity-prone T cell repertoire (13,14). The stimuli that might ultimately trigger the activation of AChRreactive T cells of the abnormal T-cell repertoires in thymoma patients are presently unknown (15). Although MG in patients with thymic atrophy shares some features with thymoma-associated MG (autoantibody profile; poor response to surgery), other features are different (HLA-association, age of onset). Therefore, it is presently unknown whether the generation of abnormal T cells inside atrophic thymuses is also involved in the pathogenesis of atrophy-associated MG or whether its pathogenesis is completely different. Whether still another pathogenesis has to be invoked for some or all cases of so-called seronegative MG is not clear but appears very likely (16,17). In spite of these uncertainties it is now safe to assume that MG is not one disease. Instead MG has to be considered the common symptom shared by a variety of pathogenically different disorders. In support of this notion, surgical as well as immunosuppressive therapies have different effects on the outcome of MG patients as far as MG symptoms are concerned. Antigen-specific immunomodulatory therapies, which are presently being tested in animal models, might overcome the varying degrees of therapeutic success and avoid the side effects of general immunosuppression. For further detailed reviews see also (7,1820). This review does not deal with congenital MG caused by constitutively abnormal presynaptic ACh or vesicle biology, mutations of the endplate asymmetric form of acetylcholine esterase (AChE) or mutations/deficiencies of AChR subunits as reviewed recently (21-24). For a detailed review of the Lambert-Eaton myasthenic syndrome, that is only shortly covered here, see also Lang (25,26), Sherer (27) and Dalmau (28).
