Myasthenia Gravis: Historical Perspective and Overview
by David S. Younger, MD; Bradford B. Worrall, MD; and Audrey S. Penn, MD

Reprinted from Neurology, Volume 48, Supplement 5, pgs.S1-S7.

This article is an overview of myasthenia gravis and a background for the other articles that follow in this issue. Other reviews have appeared since 1992.

Pathogenesis
Myasthenia gravis (MG) is probably the best understood autoimmune disorder. None other has captured the attention of so many generations of neurologists and neuroscientists. It is instructive for students of myasthenia gravis to appreciate the historical achievements in the order of their occurrence, and the many individuals who have contributed to the understanding of the disease.

The ancient history of MG is controversial. Possibly the first description of a patient with myasthenia gravis appeared in 1644 in correspondence from colonial Jamestown, Virginia, pertaining to Indian Chief Opechankanough. In 1685, Sir Thomas Willis described a patient with bulbar symptoms that could have been psychogenic. The clinical syndrome was identified by Wilks in 1877, Erb in 1879, and Goldflam in 1893. In 1895, Jolly named the disease “myasthenia gravis pseudoparalytica.”By 1900, Campbell and Bramwell had reported 60 cases. In 1934, the efficacy of physostigmine was shown by Walker. One year later, Dale and colleagues described the chemical nature of neuromuscular transmission at motor end-plates. Harvey and Masland summarized the salient electrophysiologic features of MG in 1941. In the same year, Blalock, and later Keynes, described trans-sternal thymectomy in myasthenia gravis that included as complete a removal of the gland as possible, whether or not a tumor was suspected preoperatively.

In 1960, an autoimmune cause of MG was suggested by Simpson and by Nastuk et al. However, the immunologic basis of MG awaited basic understanding of acetylcholine (ACh) release at motor endplates, as described by Katz and Miledi. Nature provided two gifts that facilitated characterization of the nicotinic ACh receptor (AChR). First, a specific neuromuscular toxin, α-bungarotoxin (BuTx), was isolated from krait snakes. Second, the electric organs of Torpedo served as a rich reservoir of AChRs.

In 1973, Patrick and Lindstrom injected rabbits with AChRs from the electric organ of eels, intending to make anti-receptor antibodies and to see if these antibodies blocked the function of AChRs in intact electric organ cells. The antibodies did block. In addition, the immunized rabbits became paralyzed and died. Experimental autoimmune myasthenia gravis (EAMG) was recognized and resulted from the autoimmune attack against native AChRs. Fambrough and Drachman applied BuTx to motor point biopsies from patients with MG and found a marked reduction in the number of AChRs, averaging 20% of controls. By 1980, Lindstrom, Toyka, Lennon, Engel, and others reproduced the essential clinical and morphologic correlates of human MG in animals by passive transfer of human myasthenic serum, also with AchR-specific monoclonal antibodies.

The past two decades have witnessed spectacular progress in understanding of the microstructure, physiology, and molecular composition of the nicotinic AChR. This, in turn, has been applied to the clinical problem of MG. The receptor is a gated receptor channel and intrinsic membrane glycoprotein of molecular weight 290,000, composed of five subunits, with the stoichiometry Of α₂βδε Distinct but related genes encode the individual subunits, and cDNAs for each have been cloned, showing remarkable homology.

The mechanisms of the end-plate current (EPC) and end-plate potential (EPP) have been, elucidated by noise analysis and patch-clamping or voltage clamping. In response to an incoming nerve action potential, highly localized regions of the nerve terminal release approximately 200 quantal packets, each containing 6-10,000 molecules of ACh. Binding of ACh to specific sites on the α/γ and α/ε subunits of AChR results in the transient openings of the AChR channel that allows a net influx of Na⁺ ions, thus producing the depolarizing potential. The circular arrangement of the five subunits delineates a 2.5-nm channel whose narrowest point is only 0.65 nm in diameter. Each subunit contains four membrane-spanning α-helices termed M1 to M4, with the M2 segment effectively lining the channel.

More quantal packets of ACh are released into the synaptic cleft and more receptor channels are present than are needed to depolarize the muscle fiber to threshold. This creates a “safety factor.” Even when the number of receptors is reduced experimentally, the EPP does not fall below the threshold needed to generate a muscle action potential. The severe reduction in the number of functioning receptors at myasthenic end-plates leads to many EPPs falling below threshold. The steps that link the binding of the ACh molecule to the opening of the ion channel and the relationship of the molecular structure of the AChR to its various physiologic functions and clinical sequelae in MG are still under investigation. These issues are explored further in the article by Kaminski et al. (this issue).

Much has also been learned about the role of thymus-derived T cells and bone-marrow-derived B cells in the pathogenesis of MG since the early reports of altered cellular immunity in human MG and in experimental models. ACh receptor antibodies are a product of plasma cells of the lymphocyte lineage. AChR antibody production starts with activation of T cells in a trimolecular complex composed of major histocompatibility class II molecules, native AChR, and antigen-presenting cells. The initial events in the pathogenesis of myasthenia (loss of self-tolerance) are not understood, but several factors probably contribute, including autoreactive T cells sensitive to native AChR antigens in thymus and blood, usually quiescent and not deleted or anergized.

Diagnosis
The clinical diagnosis of autoimmune acquired MG is made by recognizing a pattern of weakness that has the features of fluctuation and variability over the course of a day or over months or years, leading to perceptible exacerbations and remissions. The distribution of weakness is characteristic, affecting ocular, facial, oropharyngeal, and limb muscles. The diagnosis is confirmed by unequivocal and reproducible improvement after intravenous administration of edrophonium chloride, a rapidly acting anticholinesterase drug. Formal diagnosis is bolstered by finding a decremental. response to repetitive nerve stimulation and AChR antibodies in the serum. Although the accuracy of the clinical diagnosis has never been formally studied, no other disease demonstrates a similar combination of clinical signs, and there are practically no false-positives in edrophonium testing or AChR antibody studies. Practically no other disorder exhibits positivity on edrophonium. testing or AChR antibody studies. Fatigability is rarely the sole finding, but can accompany true weakness and may worsen with repetitive contraction. Selective involvement of limb or respiratory muscles, sparing ocular or oropharyngeal muscles, is rarely if ever encountered. The clinical limits of MG are explored further in the article by Lisak (this issue).

The electrophysiologic evaluation of myasthenia gravis, described later in this volume by Lange, includes repetitive stimulation and single-fiber electromyography. A decremental response of 12 to 15% or more of successive compound muscle action potentials after 3-Hz stimulation and aggravation of the block for several minutes after brief exercise are indicative of MG. Single-fiber electromyography quantitates transmission at individual end-plates while the patient voluntarily activates the muscle under examination. Action potentials are recorded from two muscle fibers in the same motor unit near the single fiber electrode. The variability in the time between the two potentials, which varies among consecutive discharges, is termed jitter, and is calculated as the mean difference between consecutive interpotential intervals. Jitter normally varies from 10 to 50m Blocking occurs when consecutive impulses do not follow. A typical finding in MG is normal jitter in some potential pairs and increased jitter in others. As a rule, 20 potential pairs are studied in each muscle. Up to 85% of patients with generalized and 10% with ocular MG reveal abnormal decrement in a hand or shoulder muscle, and 86% of patients with generalized and 63% of those with ocular disease reveal abnormalities on single-fiber electromyography. With the addition of a second muscle, jitter on. single-fiber electromyography is positive in 99% of patients with generalized MG, making it a more sensitive method of analysis.

Three ACh receptor assays are available for serologic evaluation of myasthenia gravis, as described by Lennon in this issue. They include AChR binding, blocking, and modulating antibody assays. The postulated actions of ACh receptor antibodies include accelerated degradation, endocytosis, and crosslinking of receptors, functional blockade, and complement-mediated lysis of end-plates by the membrane attack complex (leading to flattening and simplification of postsynaptic junctional folds). The binding assay is positive in up to 90% of patients with generalized MG and should be the first line of testing, with a specificity of more than 99%.

Nosology and classification have often been difficult in MG. The term myasthenia has traditionally been used interchangeably for the acquired autoimmune form of the disease, and the term myasthenic has been used for other syndromes of the neuromuscular junction. There are several different clinical classifications of MG, and few institutions endorse the same ones. Early attempts emphasized duration of symptoms because it was believed that the disorder might be progressive. More recent clinical classifications utilize indices of maximal severity and other descriptors of current clinical status. The clinical spectrum, nosology, and suggestions for the etiologic classification of MG and the myasthenic disorders are reviewed by Lisak in a later article.

Approximately 12 to 17% of patients with generalized MG lack demonstrable serum AChR antibodies. This is termed seronegative myasthenia gravis. These patients do not differ clinically from those with elevated titers, and exhibit similar favorable clinical responses to anticholinesterase or immunosuppressant drugs, plasmapheresis, and thymectomy. The pathogenesis of seronegative MG may not differ from that of antibody- positive cases. There are several possible explanations for the failure to detect antibodies with conventional assays. If serum antibody titers are low, the assay may fail to detect antibodies adequately because of binding at endplates. Other possible confounding factors may be low affinity or excessive variability in antibody reactivity to epitopes of the assayed antigen. Alternatively, antibodies may be directed at sites other than the main binding sites, or at sites hidden during extraction of AChR. Antibodies to AChR may not be detected with denervated or immature AChRs that contain γ-chains. A short duration of disease and concomitant immunotherapy before the assay may also contribute to seronegativity. Significantly reduced numbers of AChRs were noted in motor endplate biopsies of patients with seronegative generalized and ocular MG. Passive transfer of sera from seronegative patients to laboratory animals results in a disorder clinically similar to that induced by seropositive sera. Immunoglobulin is usually not bound to extracted AChRs at end-plates of seronegative patients, implying that the disorder might result from a circulating plasmafactor capable of inhibiting AChR function at sites other than the binding site for ACh. These issues are addressed in more detail by Sanders in this issue.

Genetic factors play a pivotal role in the pathogenesis of congenital myasthenic syndromes but not in autoimmune MG. The number of congenital myasthenic syndromes continues to grow. Loss of the safety factor for normal neuromuscular transmission characterizes all of them. The site of the defect may be presynaptic, synaptic, or postsynaptic. Two identified presynaptic disorders are due to defects in ACh resynthesis or to a paucity of synaptic vesicles with reduced quantal release. End-plate acetylcholinesterase deficiency causes a synaptically mediated disorder. The postsynaptic disorders are associated with a kinetic abnormality of AChR with or without AChR deficiency, or with AChR deficiency without a primary kinetic abnormality. Those with AChR deficiency and a kinetic abnormality include a syndrome with a short open time, a slow-channel syndrome associated with a prolonged open time due to delayed channel closure, a slow-channel syndrome due to increased affinity of AChR for ACh causing repeated reopenings during prolonged ACh occupancy, and another syndrome in which the nature of the kinetic abnormality is not elucidated. Those without AChR deficiency include the low-affinity fast-channel syndrome and the high-conductance fast-channel syndrome. Those with AChR deficiency without a primary kinetic abnormality are caused by nonsense mutations in the ε subunit gene. Clues to a congenital myasthenic syndrome include a positive family history, onset in the neonatal period, infancy, or childhood with progression during adolescence or adulthood, lack of significant response to anticholinesterase drugs, and absent serum AChR antibodies.

Investigation of these syndromes requires sophisticated morphologic and electrophysiologic studies of the neuromuscular junction, usually available at only a few medical centers with a specific interest in these disorders. A motor-point muscle biopsy in a suspected patient should be processed for cytochemical localization of acetylcholine sterase and immune deposits at the end-plates. Electron microscopic and cytochemical studies can be used to determine the size and density of synaptic vesicles and the morphology of. nerve terminals, and postsynaptic membranes. Quantitative assessment of AChR binding sites can be performed by employing peroxidase-labeled α-BuTx. In vitro microelectrode studies, including noise analysis and patch-clamp recordings provide information about the kinetic properties of AChR channels. Application of molecular genetics to the detection of mutations in AChR subunit genes has revealed new insights into and valuable correlations with the observed channel abnormalities.

Treatment
Thymomatous MG is discussed by Robert Lovelace in a later article. The following discussion relates primarily to non-thymomatous autoimmune MG. Neurologists must choose the sequence and combination of available anticholinesterase and immunosuppressant medications, thymectomy, plasmapheresis, and intravenous immune globulin (IVIg), as discussed by Massey in this issue. Virtually all patients use pyridostigmine at some time, usually as initial therapy . Optimal dosage is determined by the patient's symptoms, with increases in the dose until undesirable side effects offset maximal clinical benefit. Chronic administration is not known to cause a decline in effectiveness or deleterious effects. However, anticholinesterases do not appreciably change the natural history of the disease, and ultimately, other modalities must be used.

Prednisone and other corticosteroid preparations have been the most widely used immunosuppressive agents in MG. As early as 1935, Simon reported sustained remission in one patient after daily injections of aqueous extracts of the anterior lobe of the pituitary gland. Torda and Woff documented partial remissions in the first five patients treated with adrenocorticotrophic hormone (ACTH). Subsequently, they demonstrated improvement in 10 of 15 additional patients." However, all experienced transient worsening, and one died. The unfavorable experiences of Shy et al., Grob and Harvey, and Millikan and Eaton overshadowed these initial promising results. Enthusiasm waned for almost 20 years until Cape and Utterback, Warmolts and Engel, and Jenkins demonstrated the efficacy of chronic ACTH and chronic oral prednisone therapy. Although corticosteroids exert immunomodulation at various levels of the immune system, their effects on activated T and B cells and on antigen-presenting cells are believed to be the most important in the beneficial response in MG. Long-term administration resulted in eventual improvement in 69 to 80% of patients, but 48% of patients had initial exacerbations and two-thirds had undesirable or serious side effects. Gradually increasing the dose of prednisone averts exacerbations of weakness. Among experts, however, there has been uncertainty as to the optimal regimen. To illustrate, in 1974 the regimen of 25 mg alternating daily, increasing by 12.5 mg every three doses to a maintenance dose of 100 mg, was associated with dramatic or moderate improvement in 11 of the 12 patients thus treated. Twenty years later the same group recommended beginning prednisone with 15 to 20 mg daily, increasing by 5 mg every 2 to 3 days to a maximum of 50 to 60 mg daily, followed by alternate-day dosing.

The use of azathioprine was first reported in 1969. It has since gained worldwide acceptance, with response rates equal to those of prednisone as monotherapy for patients with generalized disease. Azathioprine is appropriate therapy for patients who exhibit poor responsiveness, intolerance, or frequent relapses while receiving corticosteroids, in those deemed unsuitable candidates for thymectomy because of age or co-morbid disease, and in patients with a thymoma . There are, however, three drawbacks to its use. Idiosyncratic side effects occur in about 10% of patients but are mainly gastrointestinal and flu-like and rarely necessitate permanent withdrawal of the medication. Bone-marrow suppression occurs in all patients. There is a long delay in the onset of the therapeutic effect, often delayed by 6 to 18 months. Taking all these factors into account, most clinicians concur with slow advancement of the dose over weeks, from 50 mg/day to maintenance levels of 2 to 3 mg/kg/day. Careful monitoring of liver function, peripheral white blood cell, and platelet counts is necessary.

Cyclosporine is a potent agent that inhibits T cell-dependent antibody responses by reversibly suppressing the clonal expansion of activated helper T cells. Administration of cyclosporine can prevent the expression and induction of EAMG. Cyclosporine was as effective as prednisone and azathioprine in controlled double-blind studies of the treatment of generalized MG. Long-term use is associated with dose-dependent and cumulative nephrotoxicity owing to endothelial vascular injury and intersititial fibrosis, hypertension, and headache.

Plasmapheresis and intravenous immune globulin (IVIg) are administered during myasthenic exacerbations to produce short-term clinical improvement, often within days of commencing treatment. Dalakas discusses the use of IVIg in a later article. Plasmapheresis rapidly lowers AChR antibody titers, which may account for its beneficial effects. IVIg is believed to inhibit specific idiotype-anti-idiotype antibody interactions with downregulation of autoantibody production, inhibition of binding to the AChR, and amelioration of complement-mediated lysis of AChRs. Drawbacks to both include high cost, the need for specialized equipment and staff, potential shifts of body fluids, electrolyte disturbances, and the need for indwelling catheters for vascular access. For IVIg, additional considerations include the small but finite risk for transmissible disease, flu-like syndrome, aseptic meningitis, renal failure, and anaphylaxis in IgA-deficient patients. Selective plasmapheresis using the protein A-Sepharose resin for immunoadsorption of AChR antibodies is not yet available for clinical use.

The earliest trans-sternal procedures were performed for removal of thymic tumors. Thymomatous MG is discussed in the article by Jaretzki (this issue). The beneficial results in non-thymoma patients including the salient abnormal histologic changes and possible contributing factors to the pathogenesis of MG, were appreciated afterwards. In the 1960s, two concepts suggested on intrathymic pathogenesis for MG. Simpson proposed that the thymus was involved in the immunologic pathogenesis of MG. Independently, Van de Velde and Friedman and Van der Geld and Strauss observed crossreactivity of myasthenic serum with thymic myoid determinants.

Although the hypothesis that MG originates in the thymus is hard to prove, several observations support a primary role of the gland in the pathogenesis of the disease in many patients. First, thymic tissue and peripheral blood from myasthenic patients contain an enhanced proportion of autoreactive T cells that are capable of enhancing the production of anti-AChR autoantibodies by MHC-compatible B cells. Second, the thymus contains all of the elements theoretically necessary for activation of AChR-specific autoimmune T cells: antigen-producing thymic myoid cells, auto antigen-presenting interdigitating cells, and immunocompent CD4⁺ T cells. Third, in actively induced or passively transferred EAMG, where the myasthenogenic process is initiated outside the gland, germinal centers are not observed. In addition, total thymectomy suppresses EAMG. Finally, transplantation of myasthenic thymus fragments into mice with severe combined immunodeficiency results in the production of pathogenic mouse antibodies.

The success of early and total thymectomy, and the often-observed fall in antibody titers, . especially in patients with non-involved hyperplastic thymus glands, further strengthened the role of the thymus gland in the primary immunopathogenesis of MG. The technical goal of surgery is complete removal of the thymus gland. Total thymectomy can be difficult because the gland consists of multiple lobes in the neck and mediastinum and small foci usually lie outside the field of classical or extended trans-sternal or transcervical surgical approaches. Trans-sternal, transcervical "maximal" thymectomy should be performed whenever possible to ensure the best chance for complete remission. The surgical management of MG is reviewed later in this volume by Jaretzki, a modern pioneer in optimizing thymectomy surgery.

Forty years ago, Rowland and colleagues focused attention on myasthenic crisis and emerging concepts of myasthenic intensive care to avert death due to crisis. At that time, a patient with MG had a 50:50 chance of surviving a crisis, defined as the need for mechanical ventilatory support. Approximately 16% of all patients experience a crisis, a figure that has not changed appreciably since then. Progressive weakness, oropharyngeal symptoms, refractoriness to anticholinesterase medication, and intercurrent infection precede crisis in most of these patients. As discussed by Mayer in a later article, it is now standard practice to treat severe MG in an intensive care unit because of the ready availability of monitoring to assist in the correct timing of intubation and extubation and the availability of aggressive respiratory and medical therapies to reduce the need for tracheostomy. The overall mortality of crisis has dropped from 50% to 6% in the past four decades. Crisis is a temporary exacerbation, regardless of the proximate cause. The goal is to keep the patient alive until the transient morbidity of viral or bacterial infection, aspiration pneumonitis, surgery, or other complications subsides and responsiveness to anticholinesterase medication returns. In the past, edrophonium was administered to differentiate myasthenic from cholinergic crisis. That issue is now moot because cholinergic crisis is exceedingly rare and withdrawal of anticholinesterase medication is necessary for improvement in both. The underlying immunologic derangements in myasthenic crisis are not well understood, but there is a rapidly fatal antibody-mediated syndrome that bears resemblance to crisis and is associated with inflammation and necrosis of the end-plate region.

The ultimate goal of therapy in MG is a cure or, at the least, prevention or inhibition of the immune response to AChRs. A number of therapies that selectively or specifically interfere with the immune pathogenesis of the disease have been envisioned and may prove useful in the future. Selective immunotherapy inhibits only cells of the immune system, without affecting other cells and without the side effects of generalized immunosuppression. Some examples include cyclosporine A, already described, which inhibits interleukin (IL)-2, and genetically engineered agents that are toxic to IL-2 and thereby kill activated T cells, such as DAB₃₈₉, or that interfere with co-stimulatory signals for T-cell activation, such as. CTLA-4. Specific immunotherapy goes a step further by attempting to inhibit the specific autoimmune response to AChRs. Such strategies are of theoretic interest at present and include the elaboration of AChR-specific suppressor cells, the induction of tolerance to AChR-specific T cells, the inactivation of AChR-specific T cells using targeted APCs, and genetically modified B cells.

Acknowledgments
We thank Lewis P. Rowland, MD, for critical review of the manuscript and for sharing his experience' and wisdom in myasthenia gravis with us over many years. Robert P. Lisak, MD, also reviewed the manuscript and provided helpful comments. We also thank William Wagner, Esq., and Anna El-Quidsi of the Greater New York and National Office of the Myasthenia Gravis Foundation of America, a voluntary organization dedicated to providing patient services, public information, and research support for myasthenia gravis, for their endorsement and support of this supplement issue.

References
A complete list of references can be found in Neurology, Volume 48, Supplement 5, pgs.S5-S6.