Myasthenia Gravis

Acquired myasthenia gravis (MG) is a B cell-mediated T cell-dependent autoimmune disease, characterized by impairment of the neuromuscular junction (NMJ) transmission caused by specific autoantibodies (auto-Abs) against the acetylcholine receptors (AChR) on the post-synaptic membrane of skeletal muscle cells (Engel et al. 1977; Lindstrom et al. 1976; Meriggioli and Sanders 2009). The majority of patients with generalized MG (85%) and 50% with ocular MG develop antibodies against AChR, usually belonging to IgG1 and IgG3 isotypes (Rodgaard et al. 1987); these auto-Abs can be detected by the standard radioimmunoassay method (Lindstrom et al. 1976; Vincent 1994). In approximately 40% of MG patients without anti-AChR antibodies (AChR negative MG), antibodies directed to a muscle-specific tyrosine kinase can be detected (Hoch et al. 2001; Lindstrom 2008). The development of anti-AChR auto-Abs is apparently due to breakdown of self-tolerance in the thymus (Melms et al. 2006; Newsom-Davis et al. 1981), with induction or activation of AChR-specific CD4+ T helper cells and production of pro-inflammatory cytokines, leading to the synthesis of high-affinity antibodies (Hoedemaekers et al. 1997b; Vincent 2002).

Anti-AChR Abs are the most frequent cause of NMJ impairment. Neuromuscular transmission can also be impaired by antibodies against the voltage-gated calcium channels on the presynaptic membrane as occurs in the Lambert-Eaton myasthenic syndrome (Mareska and Gutmann 2004), or in rare occasions by drug-induced breakdown tolerance to the AChR, as observed in penicillamine induced MG (Hill et al. 1999; Penn et al. 1998).

MG generally satisfies the clinical criteria for an antibody-mediated autoimmune disease (Drachman 2003):

  1. 1.

    Autoantibodies are present in patients with the disease. Elevated amounts of anti-AChR Abs are found in 85% of sera from myasthenic patients, but not in sera from individuals without MG, including those with other neurologic or autoimmune diseases (Lindstrom et al. 1976; Vincent 1991);

  2. 2.

    Antibodies interact with the target antigen. IgGs are found localized on segments of the post-synaptic membrane and fragments of degenerating junctional folds in MG patients, whereas no immune complexes are evident in non-myasthenic controls (Engel et al. 1977). Moreover, the majority of AChR-specific antibodies are directed against an extracellular area of the receptor, defined as main immunogenic region (MIR), localized between residues 67 and 76 of the α-subunit of the receptor (Luo et al. 2009; Tzartos et al. 1988);

  3. 3.

    Passive transfer of antigen-specific antibodies in animals reproduces features of the disease. Daily injections into mice of an immunoglobulin fraction of MG patients’ sera reduce amplitudes of miniature end-plate potentials and responses on repetitive nerve stimulation, and decrease the number of AChR at the NMJ (Toyka et al. 1975);

  4. 4.

    Immunization with the specific antigen produces a disease model. Patrick and Lindstrom (1973) first demonstrated in rabbits that repeated injections of AChR, purified from the electric organ of Electrophorus electricus, result in the production of high levels of anti-AChR antibodies; the authors also observed flaccid paralysis in animals and typical MG electromyographs. The immunization with AChR from the electric organ of Electrophorus electricus or Torpedo californica induced experimental autoimmune MG (EAMG) in rats and guinea pigs (Lennon et al. 1975), in association with an autoantibody response to the AChR. Clinical signs of the disease could be also observed in female Lewis rats, injected intradermally with a recombinant protein spanning the extracellular domain (human AChR α-subunit, aa 1–210) (Lennon et al. 1991);

  5. 5.

    Reduction of antibody levels ameliorates the disease. Plasma exchange is a successful therapeutic procedure for the treatment of compromised myasthenic patients. The removal of circulating IgGs, including pathogenic anti-AChR antibodies, is associated with a rapid improvement of clinical symptoms in MG patients (Antozzi et al. 1991; Dau 1981).

Impairment of the neuromuscular transmission by AChR-specific auto-Abs results from three different mechanisms: (a) complement-induced damage of the NMJ, (b) increased degradation of AChR, and (c) direct binding of auto-Abs to the ACh binding site.

The binding of auto-Abs (of the IgG1 and IgG3 subtypes) to the antigen at the NMJ activates the complement cascade, a process triggered by the interaction between the Fc fragment of anti-AChR Abs and the C1 component. The complex activation leads to formation of the membrane attack complex (MAC) and consequently to the focal lysis of the muscle cell at the NMJ (Engel et al. 1977). The destruction of the post-synaptic membrane results in a morphological alteration, in which the typical deep junctional folds are replaced with a relatively flat surface and the density of functional AChRs is reduced (Conti-Fine et al. 2006). An associated phenomenon, due to the loss of the AChR at the NMJ, is a marked reduction in the number of voltage-gated sodium channels in MG patients and passively transferred EAMG rats (Ruff and Lennon 1998), and of AChR-associated proteins utrophin (Slater et al. 1997), and rapsyn (Losen et al. 2005; Martinez–Martinez et al. 2007); for a review see (Gomez et al. 2010). The reduction of functional muscle AChRs might also result from antigenic modulation caused by anti-AChR Abs. Indeed, it has been demonstrated in vitro that the addition of myasthenic IgGs, or the divalent fragment F(ab′)2, to skeletal muscle cultures increases threefold the rate of AChR degradation (Drachman et al. 1978). Hence, the formation of immune-complexes induces an accelerated internalization of AChR by endocytosis, which is not compensated by ex novo synthesis, and increases AChR degradation by lysosomal mechanisms reducing its availability on the post-synaptic membrane (Engel and Fumagalli 1982). Moreover, a subset of the polyclonal anti-AChR repertoire might interact with ACh-binding sites on the receptor and induce a direct impairment of AChR function. Administrations of monoclonal antibodies (mAbs) directed to the α-bungarotoxin binding sites are able to produce an acute paralysis in injected chickens (Gomez and Richman 1983); this evidence suggests that the presence of antibody-binding sites might be a crucial factor influencing the severity of the disease, more than antibody concentration in serum.

Experimental Autoimmune MG

EAMG represents an excellent model to investigate the pathogenic mechanisms underlying the human disease and develop new immunotherapies for its treatment.

The first report on an experimental model of MG dates back to almost 30 years ago, when Patrick and Lindstrom (1973) demonstrated that rabbits immunized with AChR, purified from the Electrophorus electricus electric organ, developed MG-like symptoms. After that pivotal discovery, many animal studies confirmed that an autoimmune response was occurring in MG patients against muscle AChR, and that anti-AChR Abs were responsible for the structural and functional damage of the NMJ.

Myasthenia gravis and its animal models share several features, in particular muscle weakness and fatigability, decremental response after repetitive nerve stimulation, temporary improvement of muscle strength following treatment with anti-cholinesterase drugs and increased curare sensitivity (Christadoss et al. 2000). Moreover, MG and EAMG appear similar in several immunopathological features, such as presence anti-AChR antibodies in serum, deposition of IgGs and C3 complement components at the NMJ, MHC class II-restricted presentation of AChR epitopes and involvement of T helper cells in B-cell antibody production (Christadoss et al. 2000) (Table 1).

Table 1 Similarities and differences between MG and its experimental model
Full size table

The human disease differs from experimental MG for only a few features. Factors inducing MG are unknown and thymic alterations are common in myasthenic patients, suggesting the potential role of the thymus in the pathogenesis of the disease (Meinl et al. 1991). In contrast, animals develop EAMG after AChR-immunization and the auto-sensitization process seems to occur only in draining lymph nodes (Christadoss et al. 2000), apparently without affecting the thymus, as in MG patients (Table 1).

Although EAMG can be induced in a wide variety of animals, among which also monkeys (Tarrab-Hazdai et al. 1975; Toro-Goyco et al. 1986) and rabbits (Eldefrawi 1978; Patrick and Lindstrom 1973), about 65% of experimental models are established in rats and 35% in mice, mainly due to a higher incidence of clinical EAMG signs in the former (Link and Xiao 2001). The lower murine susceptibility to EAMG may be due to a high safety factor of the NMJ, characterized by the release of a large amount of acetylcholine quanta from the nerves (Wood and Slater 2001). In susceptible rats, EAMG is routinely induced by active immunization with torpedo AChR (Link and Xiao 2001), but also a short fragment of the rat (self) AChR (aa 97–116 of the α-subunit) is capable to break immunological tolerance (Baggi et al. 2004). EAMG can also be induced by passive transfer of anti-AChR antibodies (Lindstrom et al. 1976; Tzartos et al. 1987), the simplest protocol for studying the pathogenic effects of auto-Abs in vivo. The accepted general distinction between actively induced and passively transferred models of EAMG implies that the passive transfer models allow the study of the effector phase of the human disease (IgG deposit, complement activation, NMJ destruction), while the active models also include breakdown of AChR tolerance and the immune cell activation mechanisms that might occur in human MG. The next paragraphs will describe these models in detail (Fig. 1).

Fig. 1
figure 1

Schematic representation of EAMG induction factors and treatments. Three different mechanisms to induce experimental MG are shown in the upper left box: active immunization with injection of AChR; passive transfer with administration of anti-AChR antibodies purified from myasthenic animals or MG patients; adoptive EAMG induced with transplantation of PBLs derived from MG patients. In the lower left box, two potential mechanisms to suppress EAMG are represented: cellular therapy, via either DCs or Treg cells, characterized by inhibition of T cell proliferation and pro-inflammatory cytokine production, or pharmacological treatment. Findings in the pathogenetic mechanisms and effective treatments can hopefully be translated from the animal model to the human disease (right box)

Full size image

Active EAMG in Rats

Various inbred rat strains have been tested for the induction of active EAMG via immunization with Torpedo californica AChR (TAChR). While Wistar Furth and Copenhagen rats seem to be resistant to disease induction, maintaining normal weight and muscle strength, Fischer and Wistar Munich rats appear the most susceptible strains showing severe EAMG symptoms and rapid progression of the disease (Biesecker and Koffler 1988). Nevertheless, Lewis rats, which exhibit intermediate susceptibility, represent the rat model most commonly used to induce EAMG, as clinical manifestations in this strain are similar to those of human MG (Biesecker and Koffler 1988).

EAMG in Lewis rats is commonly induced by a single immunization with purified AChR in complete Freund’s adjuvant (CFA), triggering the production of antibodies to foreign AChR, which are able to cross-react with the self-AChR (Lindstrom 1980; Link and Xiao 2001). The disease can also be induced by immunization with a synthetic peptide, corresponding to the region 97–116 of rat AChR α-subunit (R97-116), emulsified in CFA, followed by a second immunization boost of R97-116 in incomplete Freund’s adjuvant (IFA) 30 days after the first immunization. In this induction protocol, the production of auto-Abs and the activation of R97-116-specific T cells derive from the breakdown of self-tolerance (Baggi et al. 2004).

The course of EAMG is evaluated monitoring the loss of body weight and muscular strength of immunized animals. Myasthenic symptoms are assessed after exercise (repetitive paw grips on the cage grid) for 30 s, and are characterized by tremor, hunched posture, muscle weakness and fatigability.

When EAMG is induced in the Lewis rat by injection of Torpedo AChR (40 μg/rat), emulsified in CFA supplemented with additional Mycobacterium tuberculosis H37Ra (0.5 mg/rat) (Aricha et al. 2008), two different disease phases can be clinically distinguished. A transient acute phase of muscle weakness can also occur in EAMG models induced with AChR/CFA supplemented with with Bordetella pertussis; if the additional adjuvant is not used, the acute phase is not observed (Lindstrom 1980). The first one (the acute transient phase) begins approximately 7 days post immunization and is characterized by the synthesis of anti-AChR Abs that induce complement depositions on muscle membrane. Chemotactic signals are released, leading to extensive phagocytic invasion at the NMJ, with destruction of the post-synaptic membrane. The cellular invasion firstly results in a rapid decrease of muscle AChR content, which is followed (after 2–3 days) by an abnormal increase in AChR content, probably due to the formation of extra-junctional AChR (Lindstrom 1980). The second phase (the progressive chronic phase), begins approximately 28 days post immunization (Link and Xiao 2001) and is characterized by production of a larger amount of antibodies and complement deposition at the post-synaptic membrane, which now appears as a nearly flat surface lacking its typical junctional folds. In this phase, there are no phagocytic cells, and there is a drastic reduction in skeletal muscle AChR content reduced to about one-third compared with healthy controls. Importantly, the progressive chronic phase reflects the clinical course of the human disease.

Active EAMG in Mice

Mice would represent the ideal model for the development of the experimental disease due to the availability of transgenic, knockout, and mutant mice as tools to investigate the biological mechanisms at the basis of MG pathogenesis (Berman and Patrick 1980b; Christadoss et al. 2000). Moreover, an almost infinite and a never-ending array of monoclonal antibodies specific for cell markers (cell surface antigens, chemokines, cytokines, grow factors) is available for murine research. Indeed, EAMG has been intensively studied in mice with defined immunological aberrations to better understand genetic factors involved in the disease pathogenesis and investigate their potential modulation and regulation (Table 2).

Table 2 Knockout or transgenic mice in experimental MG studies
Full size table

C57Bl/6, SJL and AKR mice were classified as highly susceptible strains due to the development of myasthenic symptoms, induced by TAChR immunization, in 50–70% of animals; on the contrary, BALB/c and SWR strains appeared poorly susceptible showing low incidence of muscular weakness and flaccid paralysis (Berman and Patrick 1980a).

EAMG in the mouse is commonly induced by two or three immunization boosts with purified AChR in CFA/IFA (Christadoss et al. 2000), does not show a transient acute phase and myasthenic symptoms typically appear 7–14 days after the last injection. Due to the several immunization boosts required to induce this model, it is relatively difficult to define the appropriate window for a preventive/therapeutic treatment in mice EAMG; moreover, clinical features are less severe compared with those observed in the Lewis rat model.

Passive Transfer MG

Experimental MG can be also induced by passive transfer of auto-Abs via two distinct mechanisms. The first one regards daily injections into healthy recipient animals with serum IgG fraction isolated from MG patients (Toyka et al. 1975) or with anti-AChR antibodies purified from AChR-immunized donor animals with chronic EAMG (Lindstrom et al. 1976). Otherwise, the disease can be passively induced via administration of monoclonal antibodies, generally of IgG1 and IgG2a subclasses, directed to the AChR α-subunit and derived from AChR-immunized animals (Tzartos et al. 1987) or from cell line culture supernatants (Piddlesden et al. 1996). Recipient animals show the typical EAMG symptoms 24 h after antibodies transfer, developing the acute phase characterized by phagocytic invasion. Passive transfer MG has also been shown in rhesus monkeys injected with a human monoclonal anti-AChR antibody (IgG1 subtype) isolated from MG thymus (van der Neut Kolfschoten et al. 2007).

Adoptive Transfer EAMG

Finally, EAMG can be induced via transplantation of human tissues or cells in severe combined immunodeficiency (SCID) mice, lacking mature B and T cells and tolerating xenografts (Schönbeck et al. 1992; Wang et al. 1999). This chimeric human-mouse model represents an important tool to investigate the mechanisms involved in the pathogenesis of MG. Published studies show that SCID mice engrafted with thymus tissue fragments derived from MG patients, produce human anti-mouse AChR antibodies 1–2 weeks after transplantation. Moreover, these studies demonstrate that a myasthenic thymus contains all the cellular components required for producing auto-Abs and maintaining or increasing their synthesis for at least 11 weeks after transplantation (Schönbeck et al. 1992). Besides, other published data refer that SCID mice injected with peripheral blood lymphocytes (PBL), derived from MG patients, show the typical signs of the human disease, characterized by circulating anti-AChR antibodies and human IgG deposits at the NMJ. The same myasthenic manifestations are observed in SCID mice AChR-immunized and simultaneously injected with PBL isolated from healthy controls (Martino et al. 1993). Finally, studies on SCID mice treated with PBL from MG patients show the role of CD4+ and CD8+ T cells in the development of MG proving that only CD4+ T cells are necessary for the pathogenesis of the disease (Wang et al. 1999).

EAMG Models as a Tool to Investigate Therapeutic Approaches

The main aims of experimental MG are to understand the pathological mechanisms of the disease and to investigate potential new therapies (Fig. 1). The currently used immunosuppressive antigen-unspecific drugs such as, corticosteroids, mitomycin C, cyclosporine A, azathioprine, linomide and cyclophosphamide, were indeed initially tested in EAMG models, as recently reviewed by Sanders and Evoli (2010) and Gomez et al. (2010); in some occasions, the experimental model was also used to investigate the mechanisms of actions associated with specific drugs used in the treatment of the human disease (Janssen et al. 2008).

Induction of Peripheral Tolerance

Current conventional therapies for MG are not effective in a proportion of patients and immunosuppressive drugs induce numerous side effects; moreover, these drugs mainly suppress lymphocyte activation and proliferation without having much effect on longlived plasma cells that are terminally differentiated cells, which continue producing pathogenic antibodies (Arce et al. 2002; Gomez et al. 2011). Hence, new therapies are necessary to suppress antigen-specific immune cells and reduce the undesired effects usually observed following the inhibition of the whole immune system in MG patients. In order to develop new therapeutic approaches, we must first understand and define the pathogenesis of the disease.

The most supported pathogenetic hypothesis for MG induction is the loss of self-tolerance in the thymus, which induces the production of AChR-specific auto-reactive CD4+ T cell and consequently anti-AChR auto-Abs. The progression of EAMG seems to be associated to an altered balance in the T cell subsets: Th17 cells (and the secreted IL-17) were found significantly increased in EAMG animals. IL-17 is a pleiotropic pro-inflammatory cytokine that enhances T-cell priming and stimulates the production of multiple pro-inflammatory mediators, including IL-1, IL-6, TNF-α and chemokines (Mu et al. 2009). Normally, the immune response is kept under control by a peripheral immune-surveillance system, which eliminates self-reactive T cells that escaped from the selection processes in the thymus. This immune-surveillance is maintained in a steady state by the balance between different CD4+ T cell subsets: breaking that balance leads to failure of immune-surveillance.

The nasal administration to myasthenic rats of human recombinant fragments of the AChR α-subunit, including the whole extracellular domain of AChR (Hα1-210), induces tolerance to the AChR. The treatment prevents the development of EAMG and suppresses the progression of the disease, inhibiting antigen-specific T-cell proliferative responses and reducing the levels of anti-AChR antibodies (Barchan et al. 1999). Moreover, EAMG rats, orally treated during the acute and chronic phase with a human recombinant extracellular domain of the AChR α-subunit (Hα1-205), are similarly tolerized. This treatment results in a shift from Th1 to Th2 response and from IgG2 to IgG1 Abs isotypes, and improvement of the disease (Im et al. 1999). Similar evidence is observed after oral administration of Tα146-162 synthetic peptide, corresponding to the immunodominant epitope of TAChR α-subunit, to mice immunized with TAChR. This treatment modulates the ongoing disease in a dose-dependent manner, reducing T cell proliferative response to both TAChR and Tα146-162 peptide, the production of pathogenic antibodies and the loss of muscle AChR content (Baggi et al. 1999).

Immunomodulatory therapeutic strategies addressing the induction of peripheral T-cell tolerance to AChR should promote the re-establishment of a non-pathological balance among the T helper Th1/Th2/Th17/Treg subsets. The importance of Th1 (CD4+IFNγ+) and Th17 (CD4+IL17+) cells in EAMG induction has been demonstrated (Mu et al. 2009), while the role of the Th2 subset in EAMG is controversial. Deficiency of IL-5 or IL-10 (Th2-type cytokines) makes mice susceptible to EAMG induction; deficiency of IL-4 (also a Th2-type cytokine) does not significantly affect EAMG pathogenesis, and this cytokine is not required for EAMG induction (Balasa et al. 1998).

The pathogenic role of complement-fixing anti-AChR IgGs has been confirmed both in mouse and rat EAMG. However, it should be noted that in rats both Th1 and Th2 cells induce secretion of complement-fixing IgG subclasses. Hence, the restoration of the correct balance between Th1/Th2 subsets in the rat EAMG model might have a different effect compared to the mouse EAMG, because the rat IgG1 also activates complement, in contrast to the mouse IgG1. In human, Th1-induced IgG1, IgG2, and IgG3 subclasses bind and activate complement whereas Th2-induced IgG4 antibodies do not. The role of anti-AChR Th1 and Th2 responses in the pathogenesis of human MG is not clear yet (Milani et al. 2006).

Another therapeutic strategy designed to suppress the antigen-specific response causing the disease involves cellular components participating in the control of peripheral tolerance to the AChR. Dendritic cells (DC) are specialized antigen-presenting cells that recognize and process foreign antigens in the periphery and migrate to lymphoid organs where they expose the processed peptides to naïve T cells. Depending on their maturation and differentiation state, DC acquires tolerogenic rather than immunogenic activity (Banchereau and Steinman 1998). In the absence of inflammation, immature DC seem to control peripheral tolerance by promoting regulatory T-cell differentiation; in contrast, inflammatory conditions provoke morphological and functional changes leading to mature DC able to induce the activation of effector T cells (Roncarolo et al. 2001). Many therapeutic strategies try to modulate DC maturation and differentiation with anti-inflammatory agents rather than growth factors. DC, isolated from spleens of healthy rats and conditioned in vitro with transforming growth factor 1-beta (TGF-β1), can be arrested at their immature differentiation stage. The administration to AChR-immunized rats of TGF-β1-DC reduces the severity of EAMG symptoms (Yarilin et al. 2002). Moreover, healthy animals injected with bone marrow DC, isolated from healthy donors and in vitro pulsed with AChR, and subsequently immunized with the same AChR antigen, do not show clinical signs of EAMG. This observation confirms the role of immature DC in the control of peripheral tolerance (Xiao et al. 2003).

AChR-immunized rats injected intraperitoneally with DC derived from spleens of myasthenic animals and in vitro exposed to IL-10, show amelioration of EAMG clinical symptoms, due to DC ability in modulating T and B cell responses (Duan et al. 2004; Xiao et al. 2006). Curiously, the same effect is not achieved when DC are injected subcutaneously. The failure of this second easier injection route represents a limit of IL-10-DC treatment as a potential immunotherapy for human MG (Xiao et al. 2006).

In addition, treatment with granulocyte–macrophage colony-stimulating factor (GM-CSF) can suppress the development of EAMG manifestations when administered to mice before AChR-immunization. The efficacy of this treatment is due to the activation of specific DC subpopulations and expansion of the regulatory T-cell compartment (Sheng et al. 2006). Finally, recent data demonstrate that the administration of bone marrow DC, RelB-silenced and pulsed with Tα146-162, is able to suppress EAMG progression in mice, by inducing a positive shift in favour of Th2/regulatory T cell responses (Yang et al. 2010).

A third alternative therapeutic approach acts directly on the CD4+CD25+ regulatory T-cell (Treg) compartment. Treg cells arise in the thymus, represent 5–10% of CD4+ T cells in the periphery and constitutively express CD25 molecule (IL-2 receptor α-chain). They exert an essential role in maintenance of peripheral tolerance by suppressing proliferation and cytokine production of CD4+ effector T cells (Sakaguchi 2004). A defect in Treg cell subset is often observed in myasthenic patients: the number of Treg cells is reduced in the peripheral blood (Fattorossi et al. 2005), while their suppressive function, but not their number (Matsui et al. 2010), is altered in the thymus (Balandina et al. 2005). Therefore, the restoration or expansion of the Treg cell compartment can represent an important therapeutic tool for the disease. Induced Tregs can be prepared by ex vivo purification of CD4+ T cells from spleens of healthy rats and in vitro stimulation with anti-CD3 and anti-CD28 antibodies in the presence of TGF-β and IL-2. Published studies show that induced CD4+CD25+ cells, with functional features identical to naturally occurring Treg cells, can suppress clinical signs of EAMG in AChR-immunized rats (Aricha et al. 2008). Other studies show that also naturally occurring Treg cells, purified from spleens of healthy rats, can modulate EAMG progression when administered to AChR-immunized rats (Nessi et al. 2010). The administration of naturally occurring CD4+CD25+ cells is effective in reducing T-cell proliferation in response to the immunizing antigen, decreasing pathogenic Abs and increasing muscle AChR content, when given according to a preventive schedule (starting 7 days post immunization). In contrast, the therapeutic injection of Treg cells (starting 30 days post immunization, at overt clinical symptoms) does not reduce the severity of EAMG (Nessi et al. 2010). Moreover, published data demonstrate that the administration of IL-2/anti-IL-2 mAb complexes inhibits the development of EAMG, mediating the expansion of CD4+CD25+Foxp3+ T cells and the conversion of CD4+CD25− T cells in the periphery (Nessi et al. 2010). The suppression of the disease seems mainly due to the shift of Th1/Th2 ratio in favour of a Th2 phenotype and to the increased production of TGF-β (Liu et al. 2010).

A further candidate to study new potential cell therapies for human MG is represented by bone marrow stromal cells (BMSC), which can modulate the functions of T and B cells, NK and DC. In particular, BMSC inhibit lymphocytes responses to different stimuli by secretion of immunosuppressive factors (Kong et al. 2009a, b). Indeed, stromal cells, derived from healthy rats, induce a strong reduction of disease severity when injected in EAMG rats at the appearance of clinical signs. Such treatment results in the suppression of both T and B cell responses to the immunizing antigen and in modulation of cytokine production, decreasing Th1 and Th17 subsets and increasing Th2 and Treg subpopulations, due to BMSC secretion of immunosuppressive factors, like indoleamine 2,3-dioxygenase as well as TGF-β (Kong et al. 2009a, b).

Pharmacological Immunotherapy

Current pharmacological therapies for the treatment of MG include steroids and immunosuppressants, as long-term drugs, and immunomodulating agents, as short-term therapy (Mantegazza et al. 2011). Azathioprine and cyclophosphamide provided successful results in the treatment of experimental MG in rabbits (Abramsky et al. 1976) and rats (Pestronk et al. 1983), respectively. The synthetic immunomodulant drug, Linomide, has been tested for its effect on EAMG and other experimental autoimmune diseases, showing promising results for the future (Karussis et al. 1994). Other emerging drugs, such as mycophenolate mofetil (MMF) (Janssen et al. 2008), pixantrone (BBR2778; PIX) (Ubiali et al. 2008), and bortezomib (Gomez et al. 2011), show excellent efficacy in suppressing EAMG. MMF is a synthesized pro-drug of mycophenolic acid that inhibits the immune system by preferentially depleting guanosine and deoxyguanosine on both T and B-lymphocyte lines; Bortezomib is a proteasome-inhibitor that depletes auto-Abs-producing plasma cells; PIX is an anti-neoplastic drug, characterized by a reduced cardiotoxicity compared with the structurally-related compound Mitoxantrone. These drugs have been administrated to AChR-immunized rats via different treatment schedules, either preventive (before clinical onset) or therapeutic protocol (at overt clinical symptoms), and the showed effects were on antigen-specific T-cell proliferative responses, on the pathogenic Abs and increasing muscle AChR content. These results offer a new possibility for the treatment of human MG.

Modulation of Complement-Mediated NMJ Destruction

The role of complement components in MG and its experimental models has been intensively studied in the past. The terminal and lytic complement components (C9) are located at the motor end-plate in acquired autoimmune MG: the NMJ is characterized by an inverse relationship between the structural integrity of the junctional folds and the abundance of C9 (Sahashi et al. 1980). Complement breakdown products might play several roles in EAMG development: for instance, C3a and C5a could promote inflammation by recruiting and activating phagocytic cells. Published data demonstrate that anti-C5 antibody treatment ameliorates EAMG weakness and disease course (Zhou et al. 2007); however, C5a-knockout mice are resistant to EAMG induction, showing no direct C5 involvement in EAMG pathogenesis (Qi et al. 2008), other complement factors like C3b and C4b lead to muscle membrane lysis (Morgan et al. 2006). Depleting the complement cascade via treatment with cobra venom factor decreases the formation of anti-AChR Abs/ACh-receptor complexes and ameliorates the acute phase of EAMG in rats (Lennon et al. 1978). Moreover, the effects of some complement components have been analyzed in transgenic models. For instance, following AChR-immunization, C5-deficient mice do not show the same increased incidence of clinical symptoms, significant loss of muscle AChR and disease-induced death as in C5-sufficient control animals (Christadoss 1988).

The role of the MAC has also been investigated in acute passively transferred EAMG in Wistar rats, in which the administration of anti-C6 Fab leads to the inhibition of MAC formation and suppression of clinical and electrophysiological signs of experimental MG (Biesecker and Gomez 1989).

A further study aimed to investigate the role of complement has been performed in C6−/− rats passively transferred with monoclonal anti-AChR antibodies mAb35, an IgG1 immunoglobulin that specifically binds the MIR. Here, C6-deficient rats are resistant to EAMG development and neither C9 nor MAC are observed at the NMJ. Confirming these results, the reconstitution of C6 in EAMG C6−/− rats by injection of human C6 is able to induce the disease with severity and symptoms comparable to that observed in wild-type animals (Chamberlain-Banoub et al. 2006). All these evidences demonstrate the importance of complement activation, in particular MAC formation, in NMJ destruction and generally in the pathogenesis of EAMG. Therefore, depletion of components necessary to MAC assembly could represent an essential element for the treatment of human MG.

An alternative to depleting approaches is represented by blocking the NMJ destruction by pharmacological inhibition of the complement activation pathway. Accordingly, the administration of rEV576, a specific C5 complement component inhibitor identified in the saliva of the tick, is able to reduce the severity of passive transfer MG and the progression of acute experimental MG, reducing C9 deposits at the NMJ (Soltys et al. 2009).

Besides, daily intraperitoneal injections with a soluble recombinant form of human complement receptor 1 reduce weight loss and severity of clinical symptoms in myasthenic Lewis rats, in which EAMG has been induced by passive administration of mAb35 (Piddlesden et al. 1996). Following this line of experimental approach, several studies have been recently performed. Regulatory proteins that inhibit MAC formation (such as MIRL-CD59 which inhibits MAC assembly) (Kaminski et al. 2006) or control the activation of the complement cascade (such as the decay-accelerating factor, DAF or CD55, which inactivates C3 and C5 convertase enzymes) (Lin et al. 2002) can represent crucial players in EAMG development and their depletion might have a role in the pathogenesis of EAMG. Hence, the modulation of these regulatory proteins may be a target for preventing or attenuating the destruction of the NMJ. For instance, Daf1−/− mice passively transferred with McAb-3, an activator of complement belonging to IgG2b isotype, show an increased susceptibility to EAMG compared with normal animals, characterized by muscle weakness, fatigability and C3b deposits at the NMJ already 24 h post the disease induction (Lin et al. 2002). Daf1−/−, CD59a−/− and Daf1−/−CD59a−/− mice intraperitoneally injected with McAb-3 can develop passive transfer MG with different severity: both knockout mice for a single gene become mildly sick, while mice with depletion of both genes show a severe muscle inflammation and loss of AChR (Morgan et al. 2006). Similar findings can be observed after intravenous administration of McAb-3 in mice with depletion of the same genes. Daf1−/−CD59a−/− mice develop a very severe disease leading to death, and Daf1−/− mice appear weaker than CD59a−/− with evident deposits of C9 at the NMJ 48 h post EAMG induction (Kaminski et al. 2006).

These and other experimental approaches underline the fact that complement activation is an essential factor in the destruction process of the post-synaptic membrane. Indeed, IL-12−/− mice provided indirect but strong evidence on the role of complement in EAMG, since they do not produce complement activating IgG2b and IgG2c antibodies, and hence knockout mice are protected from EAMG after immunization with Torpedo AChR (Karachunski et al. 2000).

Increase of NMJ Resistance to Complement-Mediated Lysis

The susceptibility to experimental MG in rats correlates with the age of the animals. Aged rats show a mild disease, characterized by low antibody titer and high resistance of the post-synaptic membrane to the Abs-mediated attack (Hoedemaekers et al. 1997a). These events are accompanied by an increase of rapsyn concentration at the level of the NMJ in aged animals. An increased interaction between rapsyn and AChR may stabilize the receptor molecules, leading to minor AChR loss and muscle weakness in acute EAMG (Losen et al. 2005). Thus, rapsyn and its overexpression could represent a further therapeutical target. However, findings that are more recent show that the increased expression of rapsyn alone is not able to efficiently anchor the AChR to the post-synaptic membrane in chronic EAMG, once the destruction of the NMJ has already occurred (Martinez–Martinez et al. 2007).

Conclusions

Human diseases like MG involve different compartments of the organism: the immune system and the NMJ. The EAMG model allows investigation of both compartments, focusing on the pathogenic mechanisms and the clinical outcome. In vitro models able to fully represent complex pathologies where more tissues and systems are involved are not yet available. Therefore, animal models are necessary to deeply comprehend the pathogenic mechanisms and investigate new therapies. Animal models involving transgenic approaches may not be easily translated to human, considering the difficulties associated with gene therapy, but new drugs and chemicals can be easily tested on experimental models and later on translated to clinical trials.

Although several antigen-unspecific immunosuppressive strategies tested in EAMG were transferred to the clinics (Abramsky et al. 1976; Karussis et al. 1994), other antigen-specific therapeutic approaches, more recently tested and proved effective in EAMG, have not been transferred yet to the human disease. This may be due to two reasons. Firstly, EAMG has a less complex pathogenesis compared to the human disease; secondly, MG is a rare disease and generally new approaches are tested in different experimental models belonging to more common autoimmune disorders. Nevertheless, as we had the opportunity to demonstrate, EAMG has been a valuable tool to prove the efficacy of pixantrone in the experimental model (Ubiali et al. 2008), with results that encourage its investigation in the human disease. This approach is propaedeutic to the investigation of any new immunosuppressive or immunomodulating compound of potential interest.

Always bearing in mind the importance of Russel and Burch (1959) 3R rule for replacement of experimental animal procedures with alternative methods, reduction of the number of used animals and refinement of the animal conditions, we cannot forget that complex diseases such as autoimmune disorders need to be addressed with pre-clinical research in order to obtain therapies that are more efficient. Despite the evident differences between EAMG and MG, primarily the fact that the MG pathology cannot spontaneously arise in mice and rats, the experimental approach remains a very useful tool and an unavoidable method to discover new and more efficient therapies.