Journal of Clinical Immunology, Vol. 7, No. 3, 1987

Special Article

Immunopathogenesis and Treatment of Myasthenia Gravis A R N O L D I. LEVINSON, 1-3 B. ZWEIMAN, l'z and R. P. LISAK z

message is an increase in the amount of the neurotransmitter, acetylcholine (ACh), which is spontaneously released in smaller amounts at rest. The packets of ACh are released at the presynaptic active zones and diffuse across the synaptic cleft to interact with specific nicotinic acetylcholine receptors (AChR) which are integral constituents of the postsynaptic muscle membrane (end plate). The muscle membrane at the end-plate region is highly organized with so-called junctional folds. The shoulders of these folds, which are closest to the nerve, are particularly rich in AChR. The specific interaction of ACh with AChR results in a change in the configuration of AChR which results in local depolarization (end-plate potential; EPP). If the EPP is of sufficient magnitude to depolarize the surrounding muscle membrane beyond the excitation threshold, an action potential is produced which propagates along the muscle surface and penetrates into the muscle T-tubular system, triggering muscular contraction. The change in AChR configuration is accompanied by an influx of Na and an efflux of K which results in the action potential. When ACh is inactivated by acetylcholinesterase, end-plate conductance returns to the normal resting level (1).

Accepted: December 8, 1986 KEY WORDS: Myasthenia; thymus; acetycholinereceptor anti-

body; B ceils; immunoregulation; idiotype; therapy. INTRODUCTION Myasthenia gravis (MG) 4 is a disease of neuromuscular transmission characterized by weakness of striated muscles. The defect in neuromuscular transmission is mediated by autoantibodies which react with nicotinic acetycholine (AChR) receptors at postsynaptic myoneural junctions. In this review, we examine the immunopathogenesis of MG and consider various therapeutic strategies suggested by work in this area as well as by advances in our understanding of immunoregulation and autoimmunity. EVENTS AT THE NEUROMUSCULAR JUNCTION When an action potential arrives at the presynaptic terminal of a motor nerve, the electrical activity is converted to a chemical message. This 1Department of Medicine, Universityof Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104. ZDepartment of Neurology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104. 3Arthritis/Immunology Center, Veterans Administration Medical Center, Philadelphia, Pennsylvania. 4Abbreviations used: MG, myasthenia gravis; AChR, acetycholine receptor; anti-AChR, antibodies to AChR; Ab, antibody, PBMC, peripheral blood mononuclear cells; PWM, pokeweed mitogen; EAMG, experimental myasthenia gravis; SIg, surface immunoglobutin; id, idiotype; EPP, end-plate potential; MIR, main immunogenic region; MAb, monoclonalantibody; ADCC, antibody-dependent cellular cytotoxicity; SLE, systemic lupus erythematosus; IgSC, immunoglobulin-secretingcells; TL, thymic lymphocytes.

PATHOGENIC MECHANISMS IN MG There is general agreement that the weakness and fatigue of neuromuscular transmission in MG are due to a loss of functional AChR with alteration in the postsynaptic membrane at the motor end plate. It seems quite clear that antibodies (ab) to AChR (anti-AChR) are important in the pathogenesis of 187 0271-9142~7/0500-0187505.00/0 © 1987 Plenum Publishing Corporation

188

LEVINSON, ZWEIMAN, AND LISAK

this neuromuscular block (2, 3) (Table I). Immunoglobulin can be visualized at the end plate by immunohistochemical techniques; when eluted, this immunoglobulin has anti-AChR activity. Elevated serum levels of anti-AChR are found in 80-90% of MG patients. Infants born to women with MG will sometimas exhibit a transient decrease in neuromuscular transmission for up to several weeks postpartum, likely due to transplacentally transferred anti-AChR. The administration of IgG from myasthenic patients to mice will induce neuromuscular blockade. Thus, it is unlikely that the anti-AChR found in most MG patients is merely an epiphenomenon or is a secondary response to AChR released from end plates damaged by other mechanisms. There are several postulated mechanisms (Table II) by which the anti-AChR could lead to impaired neuromuscular transmission in MG patients.

Damage at the End Plate Ultramicroscopic studies show marked destructive changes in some end plates, particularly at the shoulders of the folds, where AChR is normally at a greater concentration. There is simplification of the postsynaptic membrane and widening of the adjacent cleft, which contains membrane debris. A role for complement fixation in this reaction is strongly suggested by the presence of both C3 and C9 on the postsynaptic membrane. Serum complement levels are decreased in some MG patients. The anti-AChR of many patients can fix complement in vitro when binding to skeletal muscle and can damage cultured rat myotubes, with a resultant decrease in AChR content (4). Although antibody-determined, complement-mediated destructive events do seem to be important in the pathophysiology of MG, these are likely not the only mechanisms. The rapid clinical improvement in MG following certain therapeutic maneuTable I. Evidence for a Pathogenic Role of Anti-AChR Antibodies in Myasthenia Gravis

1. 2. 3. 4. 5.

In vitro effects of Ig from MG sera on motor end plate MG-like clinical manifestations in rodents injected with MG sera Transient MG clinical picture in some infants born to MG patients Transient decrease in MG clinical signs after plasmapharesis Clinical similarities to experimental MG

Table II. Possible Pathogenic Effects of Anti-AChR Antibodies

at the Motor End Plate 1. 2. 3. 4.

Damage postsynaptic membrane, complement mediated Enhance rate of AChR degradation Decrease synthesis of AChR Block cholinergic binding sites on AChR

vers and the lack of destructive changes in many neuromuscular junctions of symptomatic areas despite prominent local immunoglobulin deposition suggest that a more readily reversible process (discussed below) is likely also involved in the neuromuscular block. Acceleration of AChR Degradation Normally, there is a slow turnover rate of AChR at the motor end plate. In vitro and in vivo studies in animals have shown that anti-AChR can increase the rate of receptor degradation, likely due initially to internalization of AChR. This complement-independent reaction may be a selective alteration of an AChR-Ab complex. Both the F(ab)2 and the F(ab) fragments of anti-AChR Ab bind to the AChR but generally only the former induces increased AChR degradation. Thus, cross-linking of the receptor, reversible after removal of the Ab, appears to be an important (but perhaps not universal) early step in enhanced degradation (3, 4). The effect of MG Ab on the local synthesis of new AChR is controversial (3). Receptor Blockade It is conceivable that anti-AChR block binding of ACh to the cholinergic binding sites on AChR. Such inhibition has been commonly assessed by studying the effects of MG serum on the binding to AChR of the neurotoxin, ~-bungarotoxin. However, most studies show that Ab which block the binding of the bungarotoxin to AChR (i) are found in only a minority of MG sera (ii) induce only partial inhibition, and (iii) may not inhibit binding of cholinergic agonists to AChR (see below). More recent studies suggest that there are subpopulations of Ab directed against such cholinergic binding sites with resultant neurophysiologic effects suggesting a direct block of ligand binding rather than destruction or altered AChR turnover. It is currently felt that such blocking Ab likely play a minor primary role in the pathophysiology of MG (3, 4). However, in some patients such blocking Ab could further diminish

Journal of Clinical Immunology, Vol. 7, No. 3, 1987

189

MYASTHENIA GRAVIS

synaptic function already decreased due to a damaged postsynaptic membrane and/or decreased density of AChR. These events could result in (i) a marked acute clinical deterioration and (ii) a rapid clinical improvement after plasmapharesis before regeneration of AChR and damaged membrane would be expected. The evidence cited above makes a fairly convincing case for an anti-AChR-mediated pathogenesis of neuromuscular dysfunction in MG. However, a disturbing feature has been the consistent lack of correlation between serum anti-AChR titers and disease severity" in untreated patients (2). Also, serum anti-AChR levels may not change significantly even when there is marked clinical improvement following thymectomy. In addition, in a minority of MG patients, no anti-AChR are detectable by the most sensitive available techniques. Several explanations have been postulated for these apparent disparities. (i) AChR contains four gtycoprotein subunits (a2, 13, % 8) bound covalently into one functional molecule of about 250 kDa. There are two cholinergic sites per receptor located on each of the two a subunits. Anti-AChR in sera from different MG patients are both polyclonal and heterogeneous with regard to cross-reaction with AChR of different species, affinity, and cholinergic blocking activity. More than 50% of such Ab bind to one region on the subunit, called the main immunogenic region (MIR), distinct from the cholinergic binding site. Monoclonal Ab (MAb) induced against different AChR epitopes vary considerably in their capacity to compete in binding to AChR not only with sera of different MG patients but also with sera obtained sequentially from the same patient (5). The pathogenic significance of this heterogeneity is unclear. However, to date, generally MAb against only the MIR induce neuromuscular transmission block when injected into rodents (3). (ii) Various subclasses of anti-AChR may be of varying importance in the pathophysiology of MG (3). For example, IgG3-class Ab have a shorter half-life and strong complement fixing characteristics of those anti-AChR with potent in vitro effects on the motor end plate. Serum levels of IgM-class anti-AChR are frequently present in early MG. Although such Ab readily fix complement, it is not known whether they can enter the synaptic cleft and induce damage. No striking difference in the predominant subclass and avidity of serum antiAChR was found in different clinical presentations

Journal of Clinical Immunology, Vol. 7, No. 3, 1987

of generalized MG, although some differences were found in those with MG limited to ocular muscles

(6). (iii) It is conceivable that the lack of correlation between serum anti-AChR levels and disease severity in MG could reflect participation by other immune reactants at the motor end plate. Although some studies of MG patients show weak in vitro blood lymphocyte reactivity to AChR, the local collection of lymphocytes (lymphorrhages) found occasionally in MG are concentrated around necrotic muscle fibers, not at the motor end plate. (iv) Among the other antitissue Ab commonly found in MG sera (discussed elsewhere in this review), particular interest has been directed to Ab reacting with the muscle A band and/or I band found in a majority of MG patients. A recent study (7) has found such Ab to occur predominantly in patients with associated thymoma and anti-AChR, although there was no correlation between antimuscle and anti-AChR titers. However, antimuscle Ab are unlikely pathogenic since they are found in patients with thymoma but not MG; in addition, they do not induce alterations at the motor end plate. Thus, there are several possible reasons why the serum level of anti-AChR Ab by itself does not determine the extent or severity of clinical MG. Further investigation will be required to determine the relative significance of these factors. EXPERIMENTAL MG Investigation and understanding of pathogenetic mechanisms in MG have been enhanced by the development of an experimental MG-like disease (EAMG) (3). Both electrophysiologic and clinical manifestations of MG can be induced in certain rodent strains and monkeys by immunizing with purified AChR, generally obtained from the torpedo electric eel. In EAMG, the density of AChR at the motor end plate is decreased and IgG is found at the postsynaptic membrane. There are increased serum titers of anti-AChR (reacting with both electric eel and homologous AChR), although the titers do not correlate with the clinical severity or extent of AChR loss at the etad plate (8). As in human MG, incubation of these Ab with muscle cultures will decrease the number of AChR and inhibit reactivity to ACh. In the Lewis rat, two expressions of EAMG have been recognized. In the transient acute phase, there is extensive invasion by macro-

190

phages in the postsynaptic membrane with associated tissue destruction, likely involving complement activation and possibly an ADCC reaction. These macrophages may bear cytophilic antiAChR. This is followed weeks later by chronic EAMG with a clinical and pathologic picture resembling human MG, including Ig and complement deposition in the postsynaptic membrane and decreased AChR numbers. Passive EAMG can be induced to varying degree by injecting serum of human MG patients or from EAMG animals into certain rodent strains. There is rapid deposition of IgG and C3 on the junctional folds and a decrease in the density of AChR at the end plates of recipients. This is followed by prominent local invasion by macrophages with intensified tissue damage and then a slow recovery, over weeks, with an increasing density of muscle AChR. Tissue damage requires the availability of early complement components in the recipient, possibly to mediate chemotaxis and/or cytotoxic effects. The disease can be transferred with lymphocytes of actively induced EAMG donors. Thus unlike the situation in human MG, cell-mediated immune responses to AChR may contribute to the pathogenesis of EAMG in some studies (3). Moreover such lymphocytes may damage muscle cultures. One group has described an AChR-specific Tsuppressor cell line which can inhibit development of EAMG (9). Others (10) have induced antiidiotypic antibodies in rabbits injected either with the anti-AChR of EAMG animals or with certain MAb against AChR. Prior injection of these antiidiotypic Ab could inhibit the EAMG induced actively or by passive transfer of an appropriate anti-AChR MAb. However, the specificity of this antiidiotypic effect has been questioned by the provocative finding (1 I) that animals recovered from EAMG induced by passive transfer of one MAb against AChR are resistant to reinduction of EAMG not only by that MAb but by other EAMG-inducing MAb directed against different AChR epitopes. Thus, it is obvious that there are many unresolved questions in EAMG. However, this experimental autoimmune disease should be a very useful model to learn more about MG pathogenesis and immunoregulation of autoimmunity in general. Furthermore, the studies (12) of the MG-like syndrome found in some humans treated with penicillamine should further enhance our understanding of how exogenous agents can trigger this autoimmune response.

LEVINSON, ZWEIMAN, AND LISAK

DEVELOPMENT OF AUTOIMMUNITY IN HUMAN MG Although the serologic hallmark of MG is antiAChR, an increased frequency of a number of other autoantibody specificities is seen in this disorder (Table III). In addressing the etiology of MG, investigators have considered factors which might account not only for the appearance of anti-AChR but also for the expression of a more generalized autoimmune state. A recitation of the various theories of autoimmunity is beyond the scope of this presentation. Suffice it to say that, in MG, investigators have addressed the following areas: immunoregulation, idiotype network perturbations, and molecular mimicry.

Analysis of Immunoregulatory Mechanisms Most investigators have concentrated on the phenotypic and functional properties of T cells since they are the preeminent cellular constituents of the immunoregulatory network. A clear picture has not emerged from phenotypic analyses, likely due to the heterogeneity of the patient populations sampied. Thus, percentages of putative suppressor/cytotoxic (CD8 ÷) cells have been reported to be decreased (13, 14), normal (15, 16), or increased (17). Likewise, the number of putative helper/inducer (CD4 ÷) cells has been reported to be decreased (16), normal (17), or increased (15) in the blood of MG patients. Several factors which may influence T-cell subset numbers were not uniformly controlled in these studies. These include age of patients, activity of disease, treatment regimens, and prior thymectomy. Table III. Serum Autoantibodies in Myasthenia Gravis

Acetylcholine receptor Striated muscle Nuclear~ Mitochondrial Smooth muscle Thyroid Gastric parietal cell Rheumatoid factor Coombs' test Heterophile antibodies False-positive serology Antiplatelet Lupus erythematosus preparation Antitymphocyte Squamous epithelium

70--90% 20-50% 20--40% 4--6% 5-10% 15-40% 10-20% 10--40% 10% 10% 0.5-1% 5-50% b 1-2% 40-90% 8%

aAntinuclear antibodies. bDifferent assays in same study.

Journal of Clinical Immunology, Vol. 7, No. 3, 1987

191

MYASTHENIA GRAVIS

Interestingly, nicotinic AChR has also been examined as a possible phenotypic marker on immunoregulatory T cells. Both thymic cells (18) and peripheral blood mononuclear cells (PBMC) (19) appear to express AChR, and perturbation of these receptors on PBMC augments suppressor activity (20). Patients with childhood MG have both reduced numbers and reduced functional activity of a subpopulation of suppressor T cells (21). This defect was associated with a serum antibody which bound to nicotinic AChR on normal PBMC and caused a reduction of suppressor-cell activity. These observations raised the possibility that antiAChR may react not only with AChR at the myoneural junction but also with receptor on the surface of suppressor T lymphocytes and, thereby, contribute to autoantibody production. Immunoregulatory T-cell functional activity has also been assessed in human MG. Nonspecific T cell-mediated suppression of PWM-stimulated immunoglobulin synthesis or mitogen-induced proliferation is impaired in most (22, 23), but not all (24), studies. Reduced suppressor T-cell activity has been associated with the HLA-B8 haplotype (25), an interesting observation given the increased frequency of this haplotype in this and other autoimmune diseases. Unfortunately, we know less about the regulation of anti-AChR production than we do about the regulation of nonspecific in vitro immune events. The information that is available comes largely from in vitro systems in which lymphoid cells (from either the thymus or blood) are cultured alone or with polyclonal activators such as PWM. A heterogeneous pattern of PBMC responses has been observed. Anti-AChR were detected in unstimulated cultures of PBMC of some, but not all, patients (26-29). The addition of PWM caused either a decrease, an increase, or no change in this spontaneous autoantibody production. In an early study (30), patient thymocytes were shown to augment the anti-AChR response of autologous PBMC even when the latter were unresponsive to PWM. This observation suggested that MG thymus contained a source of potent helper cells or perhaps AChR which it was capable of presenting to the PBMC in a particularly stimulating fashion. (See later discussion of The Thymus in MG.) Two observations have been consistently observed in this in vitro assessment of lymphocytes(i) levels of anti-AChR secreted in vitro correlate with serum anti-AChR titers in the cell donor; and

Journal of Clinical hnmunology, Vol. 7, No. 3, 1987

(ii) PBMC from normal individuals do not produce anti-AChR, whether or not they are stimulated with PWM. Our own studies (31) suggest that the failure of normal PBMC to secrete anti-AChR in vitro is not due to potent suppressor T-cell activity in the CD8 + cell population since anti-AChR were also not observed when CD8 + cells were depleted prior to culture. On the other hand, we did find (32) that anti-AChR secretion by PWM-stimulated PBMC from the MG patients was enhanced when such cultures were initially depleted of CD8 + cells. The latter results indicate that T cells suppressing antiAChR are present in the blood of MG patients and that AChR-specific B cells are sensitive to their influence. Idiotype/Antiidiotye Interactions and Molecular Mimicry

Given the importance of id/anti-id interactions in the regulation of the immune response, evidence for an anti-AChR/anti-(anti-AChR) network has been sought. There is evidence for shared idiotypic specificities on human anti-AChR (33). Furthermore, anti-(anti-AChR) have been reported in MG patients (34, 35) as well as their healthy family members (34). The identification of these molecules as anti-ids was based on the demonstration that (i) they did not inhibit irrelevant antigen-antibody interactions, (ii) they reacted with the (Fab')2 fragment of anti-AChR antibody, and (iii) their reactivity was not inhibited by a large excess of normal polyclonal immunoglobulin. Such anti-ids could be protective by regulating the production of antiAChR or by competing with the action of antiAChR at the receptor level. Alternatively, if such anti-ids were of the internal image variety, as appears to be the case in some subjects, they could be detrimental to the host. The possible role of antiidiotypes in modulating experimental MG is discussed elsewhere in this review. The phenomenon of molecular mimicry implies that determinants expressed on foreign antigens or on the antibodies such foreign antigens elicit are cross-reactive with epitopes on self. For example, the a subunit of torpedo AChR appears to share determinants with membrane constituents of ubiquitous gram-negative bacteria (36). However, studies of MG patients have not yet shown an increased frequency or titer of such antibacterial antibodies. An intriguing possibility is that perturbations of id/anti-id responses engendered by foreign antigens

192

might initiate production of autoantibodies. In this regard, it has been reported that certain anfi-ids generated in the mouse in response to antidextran antibodies can function as anti-AChR (37). Since dextran is present in the cell walls of several bacteria, it is conceivable that the immune response to certain bacteria could initiate a network reaction leading to the production of anti-AChR. Another type of molecular mimicry is suggested by the findings (38) that the immune response to an AChlike ligand elicited not only antibodies to the ligand but also autoantiidiotypic antibodies which mimicked the action of anfi-AChR. Certain of these anti-id ab were capable of inducing an MG-like syndrome in experimental animals. THE THYMUS IN MG Following the discovery of the therapeutic benefits of thymectomy in patients with MG, a pathogenic role for the thymus was intensively sought. Although this issue remains open, several interesting insights into the immunobiology of the thymus in MG patients and normals have been provided. AChR may be expressed on thymic myoid and epithelial cells (39, 40) and perhaps even thymic lymphocytes (41), raising the possibility that the thymus represents a potentially important focus for initiating or perhaps perpetuating the autoimmune response in MG. The striking presence of germinal centers in MG thymus has attracted a great deal of attention since this organ is generally not considered to be a site of B-cell activity. Prominent intrathymic germinal center formation has been described in only two other clinical settings, i.e., SLE and thyrotoxicosis (although less prominent than in MG). Smaller and less numerous follicles may be present in thymus glands of healthy individuals. The cellular constituents of the germinal center are similar to those seen in peripheral lymph nodes (reviewed in Ref. 27). Phenotypic analysis of suspended thymus cells from MG patients has generally shown much lower percentages of surface Ig (SIg)-positive cells, particularly SIgM, than in autologous PBMC (42). This is true even when minced thymus tissue is treated enzymatically to help disperse germinal centers into single-cell suspensions. On the other hand, B cells which lack SIgM but express BI have been observed in these MG thymus-ceU suspensions. This pattern differs from that in the peripheral blood, where almost all B1 + cells also express SIgM.

LEVINSON, ZWEIMAN, AND LISAK

Preliminary studies in our laboratory indicate that some, but not all, thymic BI+/SIgM - cells express SIgG. These phenotypic properties likely relate to the functional activity of thymic B cells described later in this review. The T-cell profile is similar in frequency and distribution in the thymus of MG patients and of age-matched normals (43, 44). The thymocytes in thymoma display the phenotypic markers of normal cortical thymocytes (45, 46). There is an increased percentage of Ia + ceils in suspended thymocytes of about one-third of MG patients (43). These Ia-bearing cells are enriched in the non-T-cell fraction of the thymus and are accounted for in part by B cells but not macrophages. It is still not clear if the Ia + cells in these suspensions include dendritic or epithelial cells. The significance of this finding of increased Ia-cell frequency is currently unclear since the pattern of Ia reactivity is similar in tissue sections of thymus from MG patients and normals, where Ia is seen predominantly on cortical cells with long interdigitating processes and in a confluent pattern in the medulla. However, there are reports of an increased frequency of Ia + interdigitating dendritic cells in sections of MG thymus (47). A pathogenic role of the thymus is also suggested by functional studies of thymic B cells. Myasthenic thymus-cell suspensions demonstrate an increased frequency of spontaneous immunoglobulin-secreting cells (IgSC) compared to control thymus-cell suspensions (obtained from subjects undergoing cardiac surgery), a finding indicative of enhanced in vivo B-cell activation in the myasthenic thymus (42). In a sizable percentage of MG patients, the frequency of IgSC was higher in thymus-cell suspensions than in suspensions of PBM, a pattern not observed in the control group. There was no correlation between the frequency of spontaneous IgSC and the histologic appearance of the thymus. This suggested that accentuated B-cell activation may occur in vivo in MG thymus in the absence of formation of germinal centers. We speculated that this apparent compartmentalized B-cell~ response might reflect (i) the action of a local B-cell activator not found in normal thymus, perhaps altered AChR; and/or (ii) a local immunoregulatory milieu that favors activation of B cells once they are triggered. Interestingly, thymic lymphocytes (TL) from both MG and control subjects differentiated prominently into cells secreting Ig following culture with PWM. Surprisingly, these thymocyte responses fieJournal of Clinical Immunology, VoI. 7, No. 3, 1987

MYASTHENIA GRAVIS

quentty exceeded those in the PWM-cultured autologous PBMC even though the B-cell (SIgM ÷) percentage was uniformly much less in the thymocyte suspensions than in the PBMC at the initiation of culture. These findings suggested that the thymus of most adult humans has the capacity for prominent humoral immune responses. TL secreted predominantly IgG in response to PWM, whereas PBM secreted a mixture of IgM and IgG (48). These quantitative and qualitative features of the Ig secretory responses reflect in part the intrinsic properties of the thymic B cells rather than the immunoregutatory properties of thymic T cells. First, the limited production of IgM is consistent with the paucity of SIgM ÷ B cells in the TL suspensions. Second, thymic T cells supported the production of IgM as well as IgG by autologous PWM-stimulated blood B cells, although the amount of help (measured in functional terms) was less than that provided by blood T cells. Although the studies described above have provided insight into the immunobiology of thymic B cells, they say nothing about the specificity of the antibody they produce. This issue was first addressed by Newsom-Davis et al., who reported that anti-AChR was produced by unstimulated cultures; the addition of PWM to such cultures uniformly inhibited this synthesis (reviewed in Refs. 26 and 30). Their studies showed that TL produced more anti-AChR per cell than did PBMC of the same subject. These authors did not detect anti-AChR synthesis by TL of patients with thymoma, although others have done so. The magnitude of the in vitro anti-AChR response by TL suspensions has been reported to correlate directly with the degree of germinal center hyperplasia, the serum anti-AChR titer, and the duration of disease at the time of thymectomy. No correlation was found between the degree of anti-AChR secretion and the frequency of B cells, as determined by the presence of SIg, or presumptive plasma cells, as defined by the presence of cytoplasmic Ig. In our studies (28) of anti-AChR secretion by TL of MG patients, we found that anti-AChR synthesis by unstimulated TL was frequently greater than that by autologous PBMC, although the PBMC always synthesized more IgG. The addition of PWM had a variable effect on anti-AChR production by thymocytes, i.e., inhibition, enhancement, or no effect. Taken together, the above findings indicate that

Journal of Clinical Immunology, Vol. 7, No. 3, 1987

193

AChR-reactive B cells are important components of the thymic B-cell repertoire in MG. It is not clear, however, that their presence is due solely to local intrathymic events since MG thymocytes also produce antibodies against antigens which are unrelated to AChR and not likely to be involved in the pathogenesis of MG. Newsom-Davis et al. reported that cultured TL of MG patients produced antiinfluenza antibodies (30) and we found that such suspensions secreted anti-tetanus toxoid antibody following in vivo booster immunization with tetanus toxoid (28). We did not find such anti-tetanus toxoid antibody secretion by TL of MG patients not recently injected with tetanus toxoid. Our results suggested that B cells with specificity for antigens other than AChR migrate actively in MG to the thymus and that the B-cell repertoire is fashioned at least in part by recent systemic immune events in the host. THERAPEUTIC IMPLICATIONS

Therapy: Symptomatic The demonstration that immunopathologic mechanisms are responsible for the reduction of available nicotinic receptors for ACh at the postsynaptic membrane of the neuromuscular junction provides an explanation for the symptomatic improvement produced by the administration of drugs which inhibit acetylcholinesterase activity. Although there are fewer AChR, cholinesterase inhibitors produce an increase in available ACh and thus increase the chance for interaction between those available receptors and ACh. An increase in the release of ACh by drugs acting presynaptically (ephedrine and xanthines) would again increase ACh at the junction for interaction with the remaining AChR. Conversely, drugs that inhibit ACh synthesis or release, e.g., aminoglycoside antibiotics, would result in less available neurotransmitter to interact with the already reduced number of receptors. Other agents acting postsynaptically may block or inhibit the interaction of ACh and AChR. In normals, the safety factor is sufficient so that no clinical effect is seen with therapeutic levels of such drugs. Therapy: Immunomodulatory As noted previously, several lines of evidence point to immunopathologic mechanisms in MG. Therefore, it is of importance to consider therapy

194

directed at the immune system in relation to our current understanding of these abnormalities. CORTICOSTEROIDS. Symptomatic control can eventually be maintained in many patients on reasonably modest daily or (when possible) alternateday doses of corticosteroids. However, many investigators have found that MG patients who respond to corticosteroids almost invariably require some corticosteroid or steroid-sparing immunosuppressive agent for very prolonged periods of time, often lifelong (50-52). In many patients, the required dose for even reasonable control of disease (with or without concomitant anticholinesterase drugs) is associated with clinically significant side effects. However, it is more likely that the corticosteroid-responsive patient can eventually be tapered off these drugs or at least be controlled on smaller doses following thymectomy. It must be emphasized that all of these statements reflect clinical impressions since there are no large-scale randomized prospective studies which address any of these questions. The administration of corticosteroids is associated with an eventual decrease in circulating titers of anti-AChR (reviewed in Ref. 52). However, significant improvement commonly occurs before there are demonstrable changes in serum titers, possibly because of (a) a direct effect of corticosteroids on neuromuscular function; (b) a decrease in titer of a particular subset of anti-AChR which is important for the production of disease but is not detected by measuring total anti-AChR binding in the usual assay; or (c) an effect of corticosteroids on one of the immunologic effector mechanisms at the junction. There is evidence in experimental MG to support the first possibility, although there are no studies that directly address this phenomenon in human disease. There is no evidence, even in experimental a systems, for the second or third possibility. Another interesting phenomenon is the worsening of many patients over the first 3-10 days after beginning corticosteroid therapy. This seems unlikely to be on an immunologic basis but seems, rather, to be an effect of these agents on human neuromuscular function. CYTOTOXIC AGENTS. While there are no prospective controlled studies demonstrating a beneficial effect of azathioprine or cyclophophamide, it now seems clear that many patients benefit from these drugs in situations where other therapeutic modalities seem to have been ineffective (53-56). Suppression of anti-AChR antibody synthesis

LEVINSON, ZWEIMAN, AND LISAK

seems to be the mechanism responsible for clinical improvement. Recent studies have suggested that cyclosporine may also be effective (57). PLASMAPHARESIS. Plasmapharesis has had a major impact on the management of patients with myasthenia gravis (58, 59). This expensive but relatively safe therapy is of use in the treatment of myasthenic crisis with rapid deterioration, pre- and postthymectomy, and in some patients who fail to improve with other therapy. It has little use in patients with stable disease of a mild or moderate degree (60). Anti-AChR levels decrease transiently after plasmapharesis in the absence of concomitant immunosuppressive or corticosteroid therapy but these effects and clinical improvement may be quite short-lived. On the other hand, plasmapharesis without immunosuppressive agents may be efficacious in patients with clinical deterioration caused by a clear-cut intercurrent event (60). Of note, clinically employed doses of immune-suppressive therapy do not fully prevent a rebound increase in anti-AChR titers (61). IMMUNOGLOBULIN. Normal immunoglobulin (Ig), administered for many years by a few groups, has recently attracted increasing interest. Intravenous administration of Ig has been reported to result in clinical improvement in some patients, without a change in serum anti-AChR titers noted to date. Controlled studies are now under way. Since pooled normal 7-globulin does not contain antiAChR, it seems unlikely that the therapeutic effect results from displacement or competition with antiAChR at the end plate. It is also unlikely that any beneficial effect of these 7-globulin preparations in MG is due to the capacity of the IgG component to bind to Fc receptors on phagocytic cells, a mechanism thought to underlie the beneficial effect of Iv 7-globulin in some immunologically mediated diseases of red blood cells and platelets. Whether or not pooled 7-globulin contains antiidiotypic antibodies which might down-regulate or compete with endogenously produced anti-AChR remains to be determined. THYMECTOMY. Thymectomy is widely regarded to be of use in patients with MG, especially those who do not have a thymoma (40, 51, 52). In the latter patients, thymectomy is recommended on the basis of the need to remove the intrathoracic tumor. Since the majority of patients who improve does not show a change in anti-AChR titers or show a reduction of titers after improvement is evident, the mechanism of thymectomy-induced improvement is

Journal of Clinical Immunology, Vol. 7, No. 3, 1987

MYASTHENIA GRAVIS

not clear. It is possible that there is a change in a particular antibody population of anti-AChR that cannot be detected in the total binding assays employed in most studies. A consistent effect on other immunologic parameters has not been demonstrated.

Future Therapy Many of the effective therapeutic modalities currently employed in patients with myasthenia gravis were first employed on a somewhat empiric basis without an understanding of the immunopathogenesis of the disease, the nature of defects in immunoregulation, or the role of the thymus in the immune system or thymic abnormalities in MG. Despite our admittedly currently incomplete understanding of the pathogenic events at the end plate, the role of the thymus, and defects of immunoregulation, one would hope that further therapeutic strategies aimed at modifying the immune system will be planned on the basis of our increasing understanding of the pathogenesis of MG. Moreover, the availability of experimental models, such as actively and passively induced EAMG, and the ability to transfer passively the neuromuscular defect with both monoclonal antibodies and IgG of MG patients should allow for screening such therapeutic modalities. The currently available drugs used to immunosuppress patients with MG are broad-based, which, in part, explains their frequently unacceptable side effects. As more is learned about the cellular interactions responsible for the production of anti-AChR antibodies, it may be possible to employ newer agents that allow even more selective inhibition of the autoimmune response, perhaps eventually only the pathologic cellular interactions. No matter how selective such agents may be, unless they are antigen specific, they are unlikely to be ideal. Therefore, one would like to plan therapy designed to inactivate the antigen-specific cells involved in the pathologic response. It might then be possible to inactivate, remove, or even kilt only those antibody-producing cells or helper T cells related to the pathogenically important epitopes by employing synthetic peptides, monoclonal antibodies, or antiidiotypic antibodies. Another therapeutic strategy would entail inhibiting the pathologic effects of the antibody at the end plate. It has been demonstrated that Fab fragments of anti-AChR bind at the end plate, generally

Journal o f Clinical Immunology, Vol. 7, No. 3, 1987

195

do not cause enhanced degradation of AChR, and inhibit EAMG. Through the use of genetic engineering, it would be theoretically possible to obtain human Fab to AChR which could be administered at certain times in patients not responsive to other therapy. It is clear, however, that we need to learn much more about the relevant epitopes of AChR in different patients as well as pathologic events at the neuromuscular junction. Plasmapharesis, as currently performed, is expensive, time-consuming, and nonselective in that it removes all plasma proteins. Even newer machines, which can essentially remove Ig only, remove normal Ig along with the autoantibodies. It has been suggested that plasmapharesis utilizing a column containing bound AChR might be of therapeutic benefit. However, large amounts of purified mammalian AChR (difficult to obtain) would be needed, and there is the concern of possibly activating the complement system during the affinity absorption of anti-AChR on the column. Once the relevant AChR epitopes are identified and their structures characterized, sufficient amounts of protein could be prepared by peptide synthesis. CONCLUDING COMMENTS In this review, we have presented the principles of neuromuscular transmission and discussed how this function is altered by immunopathologic events leading to the clinical manifestation of MG. This disorder appears to be prototypic of a newly recognized group of autoimmune diseases mediated by antibodies directed against tissue receptors. However, the initiating events, the role of the thymus and other immunoregulatory factors, and the quantitative/qualitative aspects of anti-AChR effects in the pathogenesis of MG are not yet clearly defined. Nevertheless, the information obtained has provided a basis for new therapeutic approaches that have assisted in the overall management of MG. With newly emerging techniques to investigate both the human and the experimentally induced disease, one can be optimistic about further advances in our understanding of this fascinating disorder. ACKNOWLEDGMENTS This work was supported by grants from the Muscular Dystrophy Association and the National Institutes of Health (NS 19546).

196

REFERENCES 1. Lisak RP, Barchi RL: Myasthenia gravis. In Physiology Pharmacology and Biochemistry of Neuromuscular Transmission. Philadelphia, WB Saunders, 1982, pp 51-96 2. Lisak RP, Levinson AI, Zweiman B: Autoimmune aspects of myasthenia gravis. Concepts Immunopathol 2:65-101, 1985 3. Ashikawa T, Appel SH: Immtmopathologic events at the endplate in myasthenia gravis. Springer Sem Immunopathol 8:177-t96, 1985 4. Sterz R, Holdfield R, Rajki K, et al.: Effector mechanisms in myasthenia gravis. Muscle Nerve 9:306-312, 1986 5. Whiting PJ, Vincent A, Newsom-Davis J: Myasthenia gravis: Monoclonal anti-human AChR antibodies used to analyze antibody specificities and responses to treatment. Neurology 36:612-617, 1986 6. Vincent A, Newsom-Davis J: Acetylcholine receptor antibody characteristics in myasthenia gravis. I. Patients with generalized myasthenia or disease restricted to ocular muscles. Clin Exp Immunol 49:257-265, 1982 7. Penn AS, Schotland DL, Lamme S: Anti-muscle and antiacetylcholine receptor antibodies in myasthenia gravis. 8. Christadoss P, Lindstrom J, Munro S, Talal N: Muscle acetylcholine receptor loss in murine experimental myasthenia gravis: Correlated with cellular, humoral and clinical responses. J Neuroimmunol 8:29-41, 1985 9. Pachner Ar, Kantor FS: In vitro and in vivo action of acetylcholine receptor educated suppressor T cell lines in murine experimental autoimraune myasthenia gravis. Clin Exp Immunol 56:659-668, 1984 t0. Souroujon MC, Pachner AR, Fuchs S: The treatment of passively transferred experimental myasthenia with antiidiotypic antibodies. Neurology 36:622-625, 1986 11. Corey AL, Richman DR, Shuman CA, Gomez CM, Arnason BGW: Use of monoclonal anti-acetylcholine receptor antibodies to investigate the macrophage inflammation of acute experimental myasthenia gravis. Neurology 35:1455-1460, 1986 12. Vincent A, Newson-Davis J: Acetylcholine receptor antibody characteristics in myasthenia gravis. I1. Patients with penicillamine-induced myasthenia or idiopathic rnyasthenia of recent onset. Clin Exp Immunol 49:266-272, 1982 13; Skolnik PR, Lisak RP, Zweiman B: Monoclonal antibody analysis of blood T cell subsets in myasthenia gravis. Ann Neurol 11:170-176, 1982 14. Berrih S, Gaud C, Bach MA, LeBrigand H, Binet JP, Bach JF: Evaluation of T cell subsets in myasthenia gravis using anti-T ceil monoclonal antibodies. Clin Exp Immunol 45: 1-8, 1981 15. Berrih S, LeBrigand H, Levasseur P, Gaud C, Bach JF: Depletion of helper/inducer T cells after thymectomy in myasthenia patients. Clin Immunol Immunopath 28:272-281, 1983 16. Haynes BF, Harden EA, Olanow CW, Eisenbarth BS, Wechsler AS, Hensley LL, Roses AD: Effect of thymectomy on peripheral lymphocyte subsets in myasthenia gravis: Selective effect on T cells in patients with thymic atrophy. J Immunol t31:773-777, 1983 17. Miller AE, Hudson BS, Tindal RSA: Immune regulation in myasthenia gravis: Evidence for an increased suppressor T cell population. Ann Neurol 12:341-347, 1982

LEVINSON, ZWEIMAN, AND LISAK

18. Engel WK, Trotter JL, McFarlin DE, Mclntosh CL: Thymic epithelial cells contain acetylcholine receptor. Lancet 1:1310-1311, 1977 19. Morrell R: Acetytcholine binding sites on peripheral blood and CSF mononuclear cells from myasthenic patients. Ann NY Acad Sci 377:848-849, 1981 20. Richman DP, Antel JP, Burns JB, Arnason BGW: Nicotinic acetylcholine receptor on human lymphocytes. Ann NY Acad Sci 37:427-435, 1981 21. Shore A, Limatiblul S, Dosch HM, Gelfand EW: Identification of two serum components regulating the expression of T lymphocyte function in childhood myasthenia gravis. N Engl J Med 301:625--629, 1979 22. Kelley RE, Keesy JC, Goymerac V, Larrick SB, Kebo D, Buffkin D: Immunoregulation of total IgG synthesis in myasthenia gravis. Ann NY Acad Sci 377:403-410, 1981 23. Mischak RP, Dau PC: Lymphocyte binding antibodies and suppressor cell activity in myasthenia gravis. Ann Ny Acad Sci 377:436--446, 1981 24. Koethe SM, Cook A, McQuillen DP, McQuillen MP: Clinical and immunologic correlations in myasthenia gravis: Measurement of ConA stimulated suppressor cell activity. NY Acad Sci 377:447-454, 1981 25. Zilko PJ, Dawkins RL, Homes K, Witte CK: Genetic control of suppressor lymphocyte function in myasthenia gravis: Relationship of impaired suppressor function HLAB8/DRW3 and cold reactive lymphocytotoxic antibodies. Clin Immunot 14:222-230, 1979 26. Newsom-Davis J, Willcox NA, Vincent A: Immunological abnormalities in myasthenia gravis. In Neuroimmunology, PO Beehan, F Spreafico (eds). Raven Press, New York, 1984, pp 275-283. 27. Levinson AI, Lisak RP, Zweiman B, Kornstein M: Phenotypic and functional analysis of lymphocytes in myasthenia gravis. Spring Harbor Semin Immunopathol 8:209-233, 1985 28. Lisak RP, Levinson AI, Zweiman B, Kornstein MJ: Antibodies to acetylcholine receptor and tetanus toxoid: In vitro synthesis by thymic lymphocytes. J Immunol 137:1221-1225, 1986 29. Fujii Y, Hashimoto J, Monden Y, Ito T, Nakahara K, Kawashima Y: Specific activation of lymphocytes against acetylcholine receptor in the thymus in myasthenia gravis. J Immunol 136:887-891, 1986 30. Newsom-Davis J, Willcox N, Calder L: Thymus cells in myasthenia gravis selectively enhance production of antiacetylcholine receptor antibody by autologous blood lymphocytes. N Engl J Med 305:1312-1318, 1981 31. Lisak RP, Laramore C, Levinson AI, Zweiman B, Moskovitz AR: Suppressor T cells in myasthenia gravis (MG) and antibodies to acetylcholine receptor. Ann Neurol 19:87-89, 1986 32. Lisak RP, Laramore C~ Levinson AI, Zweiman B, Moskovitz AR, Whitte A: In vitro synthesis of antibodies to acetylcholine receptor (anti AChR ab) by circulating tymphocytes: Role of suppressor T ceils in normal subjects. Neurology 34:802-805, 1984 33. Lefvert AK: Anti-idiotypic antibodies against the receptor antibodies in myasthenia gravis. Scand J Immunol 13:493--497, 1981 34. Lefvert AK, Pirskanen R, Svanborg E: Anti-idiotypic antibodies, acetylcholine receptor antibodies and disturbed

Journal of Clinical Immunology, Vol. 7, No. 3, 1987

MYASTHENIA GRAVIS

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

neuromuscular function in healthy relatives to patients with myasthenia gravis. J Neuroimmunol 9:41-53, 1985 Dwyer DS, Bradley RJ, Urquhart CK, Kearney J: Naturally occurring anti-idiotypic antibodies in myasthenia gravis patients. Nature 30:611-614, 1984 Stephansson K, Diepefink ME, Richman DP, Gomez CM, Marton LS: Sharing of antigenic determinants between the nicotinic acetylchotine receptor and proteins in Escherichia coli, Proteus vulgaris, and Klebsiella pneumoniae. N Engt J Med 312:221-225, 1984 Dwyer DS, Vakil M, Kearney JF: Idiotypic network connectivity and a possible case of myasthenia gravis. J Exp Med i64:1310-1318, t986 Erlanger BF, Cleveland WL, Wassermann NH, Ku HH, Hill BL, Wan KK, Sarahgarajan R, Penn AS: The autoidiotypic route to antireceptor antibodies. Ann NY Acad Sci 475:219-226, 1986 Kao I, Drachman DB: Thymic muscle cells bear acetylcholine receptors: Possible relation to myasthenia gravis. Science 195:74-75, 1977 Werkele H, Ketetsen UP, Zurn AD: Intrathymic pathogenesis of myasthenia gravis: Transient expression of acetylcholine receptors on thymus derived myogenic cells. Eur Immunol 8:579-583, 1978 Fuchs S, Schmidt-Hopfield I, Tridente G, Tarrab-hazdai R: Thymic lymphocytes bear a surface antigen which crossreacts with acetylcholine receptor. Nature 287:162-164, 1980 Levinson AI, Zweiman B, Lisak RP, Moskovitz A: Thymic B cell activation in myasthenia gravis. Neurology 34:462-468, 1984 Lisak RP, Zweiman B, Skolnik P, Levinson AI, Moskovitz A, Guerrero F: Thymic lymphocyte subpopulations in myasthenia gravis. Neurology 33:868-872, 1983 Kornstein MJ, Brooks JJ, Anderson AO, Levinson AJ, Lisak RP, Zweiman B: Immunohistotogy of the thymus in myasthenia gravis. Am J Pathol 117:184-194, 1984 Kornstein MJ, Hoxie JA, Levinson AI, Brooks J: Immunohistology of human thymomas. Arch Pathol Lab Med 109:460-463, 1985 Chilosi M, Ianucci A, Fiore-Donati L, Tridente G, Pampanin M, Pizzolo G, Ritter M, Bofilt M, Hanossy G: Myasthenia gravis: Immunohistotogical heterogeneity in microenvironmental organization of hyperlastic and neoplastic thymuses suggesting different mechanisms of tolerance breakdown. J Neuro~mmunol 11:191-204, 1986 Thomas JA, WiIlcox NA, Newsome-Davis J: Immunohistological studies of the thymus in myasthenia gravis. Correla-

Journal o f Clinical Immunology, Vol. 7, No. 3, 1987

197

48.

49. 50.

51. 52. 53.

54.

55.

56.

57.

58.

59.

60. 61.

tion with clinical state and thymocyte culture responses. Neuroimmunology 3:319-335, 1982 Levinson AI, Zweiman B, Lisak RP: Isotype restriction of thymic B cell immunoglobulin synthesis in myasthenia gravis. Ann NY Acad Sci (in press). Lisak RP: Myasthenia gravis: Mechanisms and management. Hosp Pract 18:101-109, 1983 Johns TR: Treatment of myasthenia gravis: Long-term administration of corticosteroids with remarks on thymectomy. In Adv Neurol 17, RC Griggs, RT Moxleg tli (eds). New York, Raven Press, 1977, pp 99-122 Drachman DB: Myasthenia gravis. N Engl J Med 298: 130-142, 186-193, 1978 Lisak RP, Barchi RL: Myasthenia gravis. In Treatment and Management, Philadelphia, WB Saunders, 1982, pp t84-238 Mertens HG, Balzereit F, Leipert M: The treatment of severe myasthenia gravis with immunosuppressive drugs. Eur Neurol 2:321-339, 1969 Matell G, Bergstrom K, Franksson C, Hammarstrom L, Lefvert AK, Moiler E, Von Reis G, Smith E: Effects of some immunosuppressive procedures on myasthenia gravis. Ann NY Acad Sci 274:608--622, 1976 Hertel G, Merten HG, Reuther P, Richer K: The treatment of myasthenia gravis with azathioprine. In Plasmapharesis and the Immunobiology of Myasthenia Gravis, PC Dau (ed). Boston, Houghton Mifflin, t979, pp 3t5-328 Witt AS, Cornblath DR, Parry G, Lisak RP, Schatz NS: Azathioprine in the treatment of myasthenia gravis. Ann Neurol 15:602--605, 1984 Tindall RSA, Rollins JR, Phillips JT, Greenlee, RG, Belendink G: A double-blind, placebo-controlled trial to determine the efficacy of cyclosporin A on the treatment of myasthenia gravis. Neurology 36 (Suppl): 196-197, 1986 Pinching AJ, Peters DK, Newsom-Davis J: Remission of myasthenia gravis following plasma exchange. Lancet 2:1373-t376, 1976 Dau PC, Lindstrom JM, Cassel CK, Denys EH, Shev EE, Spitler LE: Plasmapharesis and immunosuppressive drug therapy in myasthenia gravis. N Engl J Med 297:1134-1140, 1977 Lisak RP: Plasma exchange in neurologic diseases. Arch Neurol 41:654-657, 1984 Newsom-Davis J, Ward CD, Wilson SG, Pinching AJ, Vincent A: Plasmapharesis: Short and long-term benefits. In Dau PC (ed). Plasmapharesis and the Immunobiology of Myasthenia Gravis, PC Dau (ed). Boston, Houghton Mifflin, 1979, pp 199-208