Acquired, autoimmune myasthenia gravis (MG), the most common noniatrogenic disorder of neuromuscular transmission (DNMT), remains the favorite child of many neuromuscular clinicians. Arguably, it provides more professional satisfaction than any other neuromuscular disease. This fulfillment is derived in part from the intellectual satisfaction that comes from understanding disease pathogenesis. At the same time, satisfaction is gained by the ability to make meaningful improvements in patient function and quality of life through the application of rational and effective treatment. MG represents one of medicine’s most notable translational successes in bringing basic science to the bedside.

Although descriptions of individuals likely to have been affected by MG can be traced to antiquity, our current understanding of disease mechanism(s) originates from a series of seminal observations and discoveries. Thomas Willis, a seventeenth-century physiologist, is frequently credited for initially describing the clinical syndrome of MG. The concept of MG as a DNMT and the first therapeutic triumph are often credited to British clinician Mary Walker. She described the benefits of cholinesterase inhibitors in the 1930s; her discovery extrapolated from her observations related to the similarities between MG and curare toxicity. In 1960, Simpson first promoted the hypothesis of an autoimmune basis for MG on the basis of his observations of an increased prevalence in young women and in individuals with other autoimmune diseases.¹ In that same decade, support for MG as a DNMT was provided by the in vitro electrophysiological demonstration that miniature end plate potential (EPP) amplitudes in MG were greatly reduced.² In 1973 Daniel Drachman et al. solidified the concept of MG as a postsynaptic DNMT by demonstrating loss of acetylcholine receptors (AChR) in MG patients through α-bungarotoxin labeling techniques.³ In the same year, Patrick and Lindstrom confirmed the autoimmune nature of MG with the development of an experimental MG model in rabbits who became weak when immunized with AChR.⁴ In 1976, the seminal article describing the value of AChR autoantibody testing in the diagnosis of myasthenia was published.⁵ In 2001, Hoch et al. first reported the association between MG and autoantibodies directed against muscle-specific kinase (MuSK).⁶ In 2008, unnamed autoantibodies directed at clustered AChRs were found in low titer in the serum of approximately two-thirds of AChR and MuSK seronegative patients.^7,8 In 2011, patient’s with autoantibodies directed at the lipoprotein receptor protein 4 (LPR4) were identified as a third MG serotype.⁹

Historically, seronegative MG referred to patients lacking AChR autoantibodies. With the discovery of MuSK autoantibodies, the concept of double seronegative MG patients was coined. As our knowledge of LPR4 MG is somewhat limited, and as LPR4 autoantibody tests are not commercially available at the time of writing this, seropositive will be used in this chapter to refer only to AChR and MuSK MG.

The incidence of MG and its serodistribution may vary with geography and/or ethnicity.¹⁰ MG has been estimated to occur in 2–10/10⁶ individuals/year in the Eastern United States and the Netherlands but up to 20/10⁶ people/year in Eastern Spain.^11–13 The prevalence is estimated to be as infrequent as 2 and as frequent as 200/10⁶ individuals.^13–24 The age of onset may also be influenced by geographical and/or ethnic differences. Juvenile onset MG is uncommon in occidental populations but may represent more than half of cases in Asian populations.²⁵ Heritable MG is rare, estimated to occur in approximately 2% of cases although the concordance rate in monozygotic twins is estimated at 40%.^23,26–28 Under the age of 40, AChR MG is almost three times more common in women. Men and women in their 40s however, are affected with equal frequency whereas in older individuals, the prevalence is greater in men at a ratio of 1.5:1.^29,30

Understanding normal and abnormal neuromuscular transmission (NMT) is relevant not only to MG but to other DNMT that may be either acquired or heritable, resulting from genetic, infectious/toxic, or paraneoplastic/autoimmune mechanisms. These mechanisms typically act individually, but be synergistic. For example, a family has been recently reported in which there appears to be synergy between genetic and autoimmune influences. Multiple family members with seropositive MG were found to have mutations in the ecto-NADH oxidase 1 gene (ENOX1), a gene expressed in both thymus and skeletal muscles.³¹ The mechanism by which this mutation, posited to predispose to autoimmunity, might cause MG remains unknown. It is thought to relate to sequence homologies between ENOX1 and the main immunogenic region (MIR) of the alpha-1 subunit of the AChR. This provides a theoretical explanation for a heritable and anatomically targeted autoimmune disorder.³²

This chapter will focus on acquired, autoimmune MG, a postsynaptic disorder representing the prototype of DNMT. Chapter 26 will discuss the less frequently occurring infectious/toxic and genetic (i.e., CMS or congenital myasthenic syndromes) DNMT, both categories which can be conceptualized and categorized as adversely affecting NMT at the presynaptic, synaptic, or postsynaptic levels.

Acquired MG typically evolves subacutely over days to months and reaches its clinical nadir within 2 years.³³ The majority of patients are seropositive and usually have similar phenotypes regardless of the existence or type of autoantibody. Individual cases may have differing clinical features and natural histories, that may allow prediction of an individual serotype and different responsiveness to treatment.¹⁰ Each section of this chapter will describe the general features of the disease in its classical form associated with binding autoantibodies directed toward the AChR, and will then distinguish the differing features of other serotypes when applicable.

The natural history of myasthenia is difficult to predict in an individual patient. Information is, in a large part, derived from historical data accumulated prior to the availability of current therapeutic options.¹³ It is recognized that approximately 15–20% of Caucasian patients with initial signs and symptoms restricted to the oculomotor system will retain this restricted phenotype.³³ Ninety percent of those destined to develop generalized disease will do so within 2 years of symptom onset and a similar number will reach the nadir of their disease within a 7-year period.^13,34,35 It has been suggested, but by no means universally accepted, that immunomodulating treatment may diminish the risk of evolving into generalized disease.^36,37 Spontaneous, remissions unrelated to treatment are estimated to occur in between 10-22% of MG patients.^13,30 These remissions may occur at any time in the course of the disease. Half of the patient’s who achieve spontaneous remission, relapse within 6 months and 90% within a year.^13,30 Regarding therapeutic decisions, however, there is no apparent correlation between the existence and length of remission and maximal disease severity.¹³ Mortality statistics in MG has been undoubtedly altered by therapeutic intervention. At the turn of the twentieth century it was approximated at 70%, reduced to 23–30% by the 1950s, with contemporary estimates in the 1.2–2.2% range.^13,23,38

The diagnosis of myasthenia is clinically established by two phenotypic features, the pattern of weakness and its tendency to fluctuate. There are theories underlying the selected vulnerability of certain muscles in MG that will be addressed in the pathophysiology section. The basis for the characteristic patterns of weakness in MG remains, however, largely speculative.³⁹ It may be related to the distribution of different types of neuromuscular junctions (NMJs) in different muscles.

The fluctuating nature of myasthenic symptoms is related to the dynamic biology of the NMJ.⁴⁰ It is a quality of the disease that may be a dominant feature of the patient’s history or may be overlooked. MG patients may recognize that their symptoms may vary on a minute-to-minute, diurnal, or week-to-week basis.^41–44 For example, a patient may describe normal articulation at the onset of a telephone conversation and may have unintelligible speech 5 minutes later. Patients may observe normal eyelid position upon awakening and then develop ptosis as the day wears on. Fluctuation may not be simply diurnal and patients may have functional hardships one month that seem to improve on their own the next month without apparent explanation. Fluctuations may also occur in response to identifiable variables such as temperature, systemic infection, menses, anxiety, emotional stress, and pregnancy.^18,45–50 Myasthenic visits to emergency rooms are known to increase in frequency in early morning hours in equatorial countries where electricity and therefore air conditioning may not be available at night, making this potentially the warmest period of the day. Variability may also occur not only in the timing but in the pattern of weakness. Alternating ptosis represents the most notable example of this phenomenon. It should be emphasized that diurnal worsening of strength and stamina is not pathognomonic of MG as the weakness of any neuromuscular disease may worsen as the day goes on.

MG may also be suspected by the pattern of weakness. MG should be considered in any patient with painless weakness, particularly when the weakness is multifocal or diffuse in distribution or when weakness of ocular and bulbar muscles predominates. Asymmetry is not uncommon, particularly with ptosis. Most myasthenics will experience ptosis, diplopia, dysphagia, dysphonia, difficulty chewing, or symptoms referable to facial or neck weakness, either in isolation or in combination. Identification of weakness in muscles innervated by anatomically unrelated cranial nerves, for example, concomitant weakness of eyelid opening and eyelid closure represents a common and diagnostically useful example.

Initial symptoms restricted to ptosis or diplopia will be the presenting symptoms of MG patients in 65–85% of cases and 95% will experience oculomotor involvement at some point in their illness.^{33,37,43,51,52} Overt diplopia may be preceded by nonspecific visual blurring when ocular malalignment is minimal and insufficient to produce two distinct images. The presence of other signs and symptoms of myasthenia or the resolution of blurring the covering of either eye aids in the identification of blurring due to ocular malalignment. Other than the levator palpebrae, the medial rectus appears to be most susceptible of the extraocular muscles. Any pattern of ophthalmoparesis may occur, however, potentially mimicking an individual cranial nerve palsy or intranuclear ophthalmoplegia, or at times even producing nystagmus.^53–55

Bulbar symptoms typically refer to disordered speech and swallowing but will be extended here to include weakness of jaw, facial, and neck muscles. Bulbar onset of MG is quite common particularly in our experience in older men, and is estimated to occur in approximately 15% of cases.⁵⁵ Weakness of bulbar muscles is estimated to occur in >40% of MG patients sometime in their illness. As bulbar weakness may be associated with considerable morbidity, we consider bulbar symptoms to be indicative of generalized MG; ocular MG being restricted to ptosis and diplopia. Speech may be adversely affected in a number of ways. The voice may be hypophonic due to vocal cord paresis or expiratory muscle weakness.^56–58 It may have a nasal quality due to palatal insufficiency and nasal air leak. The patient may be dysarthric as a result of weakness of the lips, tongue, or cheeks. Nasal regurgitation of food and liquid, difficulty manipulating food due to tongue weakness, as well as ineffective sniffing, coughing, nose blowing, or throat clearing may be noted.

Facial weakness is common. Lower facial weakness may result in dysarthria or sialorrhea, or in difficulty whistling, inflating balloons, or drinking from a straw. It may also interfere with the accuracy of pulmonary function testing due to poor oral seal. Weakness of upper facial muscles is equally prevalent but less likely to be symptomatic. Occasionally, patients may complain of visual blurring due to lower lid weakness resulting in pooling of tears. Facial weakness can be easily missed if not sought for, particularly if bilateral. Patients who are affected may be unable to bury their eyelashes or maintain eye closure against resistance. They may be unable to whistle, make a kissing noise, or hold air in their distended cheeks against resistance. A “myasthenic snarl,” may occur in which the mid-portion of the upper lip elevates without elevation of the corners of the mouth during an attempted smile.

Symptoms and signs referable to jaw weakness, particularly jaw closing, occur fairly commonly in MG. Although jaw weakness may occur in other neuromuscular diseases, many of the other causes, for example, Guillain–Barré syndrome, are unlikely to have a phenotype readily confused with MG.⁵⁹ Like most neuromuscular diseases, neck flexor weakness is more common and more pronounced than neck extensors weakness in the majority but not in all MG patients. Head drop is not rare however, and may be the presenting symptom. Like other causes of head drop, posterior neck pain relieved by head support may be the most prominent symptom associated with this sign, presumably related to the stretch placed on posterior cervical muscles and ligaments by the weight of the head. Ventilatory insufficiency is a rare presenting symptom of MG but may develop in a significant percentage of patients with untreated or refractory generalized disease.^58,60 Along with dysphagia, it is undoubtedly the predominant basis for mortality in MG unrelated to complications of treatment or immobility.

Limb weakness occurs in 20–30% of affected individuals. Limb weakness preferentially affects proximal muscles.^61,62 One potential but undoubtedly partial explanation is the warmer temperature of the proximal limbs.⁶³ Typically, limb weakness occurs in concert with the signs and symptoms of oculobulbar disease. In approximately 10% of patients with limb involvement, MG may be initially restricted to distal limb muscles, with foot or finger drop being notable presentations that have been reported and that we have seen.^64–69 Again, the pattern may be focal, multifocal, or diffuse.⁴³

MUSK MG

Approximately 7–15% of all MG patients or 40% of seronegative patients with generalized MG are estimated to have MuSK autoantibodies.^{6,10,55,70–76} On average, MuSK MG patients are younger and more severely affected than their AChR MG counterparts although MuSK MG rarely if ever develops in an individual in the first decade of life.^10,77,78 Disease prevalence in MuSK MG is highest in the third and fourth decades of life in non-Asian populations. In some series, males and females are affected equally whereas in other series, females have predominated.^{73,77,79–81}

Most MuSK MG patients will have a phenotypic pattern indistinguishable from AChR MG patients. Nonetheless, there are certain clinical features that suggest an increased or decreased probability of MuSK MG.^70–88 For example, purely ocular disease is a rare MuSK MG phenotype.^{71–73,77,86,87} Patterns that suggest an increase probability of MuSK MG include persistent bulbar symptoms refractory to treatment, prominent neck, facial, and ventilatory muscle weakness and the presence of muscle atrophy in an otherwise typical MG patient, notably in the tongue.^10,72 The latter finding may render the clinical distinction from bulbar amyotrophic lateral sclerosis (ALS) more difficult.^55,77,79,88 As MuSK is known to facilitate reinnervation, it is plausible that the atrophy noted in this serotype may be related to a “denervating” effect of MuSK autoantibodies.⁸⁹ MuSK MG patients in general have more severe disease both at onset and disease nadir, are less likely to achieve complete remission with treatment, and are more likely to experience myasthenic crises than either their AChR or seronegative counterparts.^10,77–79

SERONEGATIVE MG INCLUDING PATIENTS WITH LPR4 AUTOANTIBODIES

Seronegative MG may represent up to 34% of MG patients depending on the study and the ethnic background of the cohort studied.¹⁰ In 2008, a high sensitivity assay discovered IgG1 autoantibodies in two-thirds of these seronegative MG patients.⁷ In 2011, an epitope on the LPR4 protein was found to be the target of IgG1 autoantibodies, presumably the same autoantibodies described 3 years previously.⁹ These patients, historically included in the seronegative group, represent a very small percentage of MG patients.⁸⁹ Studies to date have estimated that LPR4 autoantibodies are found in 3%, 9%, and 50% of seronegative MG patients.^9,25,90 These significant differences may represent differences between ethnicities, differences in methodology, or accuracies relevant to small sample size; the 50% figure originating from a study of only 13 patients.⁹⁰

The weight of current evidence suggests seronegative MG, with or without LPR4 autoantibodies, manifests phenotypic, natural history and response to treatment features similar, if not identical, to AChR MG.^7,10,91 Otherwise seronegative patients possessing low titers of IgG1 autoantibodies directed toward clustered AChRs have been reported to exist in approximately two-thirds of all seronegative AChR/MuSK MG and half of ocular seronegative MG patients, potentially explaining at least in part the high prevalence of conventional seronegativity in this latter population.^7,8,91

The examination of the suspected or established MG patient includes a number of unique or relatively unique features. As ptosis is such a common manifestation of MG, documentation of the baseline upper and lid positions in relationship to the pupil is recommended prior to provocative testing and to distinguish ptosis from squinting, both of which narrow the palpebral fissure.

In MG, pupil function is spared although physiological anisocoria is common enough to provide a potentially confounding feature. In order to unmask or exacerbate ptosis or extraocular muscle weakness, the suspected MG patient is asked to sustain either up or lateral gaze for a minute while limiting blinking as much as possible (Fig. 25-1A–C). Drifting of lid or eyeball position is watched for. Cogan’s lid twitch is another sign thought to be a relatively specific, although not necessarily sensitive, sign in MG assessment. With this maneuver, the patient is first asked to look down and then rapidly saccade to reassume the primary position. Normally the eyelid moves synchronously with the eyeball. A positive sign is defined by the eyelid overshooting the eyeball position leading to a transient scleral exposure and upper lid oscillation.

Another maneuver with potential diagnostic benefit has been referred to as enhanced ptosis.⁹² This maneuver has been described in patients with bilateral asymmetric ptosis but is conceptually of value in apparent unilateral ptosis as well. In this maneuver, the clinician manually elevates the most affected eyelid which may result in the revelation or exacerbation of ptosis on the opposite side in an MG patient. The proposed explanation for this phenomenon is Hering’s law of equal innervation. Theoretically, with manual lid elevation, there is less need for supranuclear stimulation of the levator subnucleus of the oculomotor nerve. As this is a single midline nucleus which innervates both levator palpebrae muscles, lifting the more severely affected lid leads to the need for less supranuclear stimulation of this subnucleus affecting the contralateral as well as ipsilateral levator palpebrae.

The last bedside maneuver relative to the evaluation of ptosis is the icepack test which relies on the recognized physiological enhancement of NMT by cooling. In this maneuver, an ice pack applied to a closed eyelid may result in notable improvement of existing ptosis. As the icepack is potentially noxious, exposure should be limited to a minute or less.

As motor neuron disease is often the major differential diagnostic consideration in patients with painless weakness who do not have ptosis or diplopia, the clinical assessment of the suspected MG patient includes careful observation of muscles for atrophy and fasciculations. Although muscle atrophy may occur in MG patients, particularly in those with MuSK autoantibodies, it is typically notable for its absence. Demonstrating weakness at the bedside that worsens with sustained or repetitive use in limb and trunk muscles is of theoretical, but in our experience, limited value in suspected MG patients. Any cause of neuromuscular weakness may produce reduced strength on repeated effort and normal patients may be reluctant to sustain effort resulting in a false perception of fatigable weakness. A careful assessment of cranial muscle strength is, however, very important in the assessment of suspected or known MG patients. We typically assess the strength of eyelid and lip closure, jaw opening and closing, tongue protrusion, and neck flexion and extension. In our experience, residual evidence of eye closure weakness in patients who otherwise seem to be in remission is a fairly common finding and should not by itself represent a justification to alter treatment. Again, demonstrating weakness in two or more muscles innervated by anatomically unrelated nerves, in the absence of atrophy and fasciculations, pain, sensory, or autonomic symptoms is likely to represent MG. As symptoms suggesting inadequate breathing warrant serious attention, we find bedside or telephone-based estimates of ventilatory muscle strength helpful. We do this by having the patients count out loud as far as one’s vital capacity will allow after full inspiration. The vital capacity in cubic centimeters can be estimated with reasonable accuracy by multiplying this number by one hundred.

MG IN SPECIFIC POPULATIONS

There are special considerations in pregnant women with myasthenia, their newborn children, and children with myasthenia.^93–97 In an extensive review of the literature involving 322 pregnancies in 225 myasthenic mothers, 31% of mothers had no change in their myasthenic symptoms, 28% improved, and 41% deteriorated during the pregnancy.⁹⁸ During the postpartum period, 30% had a disease exacerbation. There is a theoretical risk of transmitting IgG AChR antibodies in breast milk, although most infants have no problem with breastfeeding.

Transient neonatal autoimmune MG develops in approximately 10% of infants born to mothers with MG.^100–112 It has been reported to occur in MuSK MG.¹¹³ Maternal treatment seems to significantly lower the risk of infantile disease.¹¹⁴ Onset is usually within the first 3 days of life. The most notable features include a weak cry, difficulty feeding due to a poor suck, hypotonia, respiratory difficulty, ptosis, and diminished facial expression resulting from facial muscle weakness. The disorder is temporary in most cases, with a mean duration of about 18–20 days. In rare cases, in utero paralysis may lead to a child born with multiple joint contractures.¹¹⁴ Other than recognition and symptomatic treatment, the most important aspect of this disorder is that there appears to be no increased risk of the child developing MG in later life.

Juvenile MG represents a “subclassification” of autoimmune MG.^{96–98,115–118} It is estimated that approximately 10% of acquired (non-neonatal) autoimmune MG cases will occur before 18 years of age in occidental populations, the majority subsequent to puberty.^114–117 This statistic may be inflated, as some reported juvenile, seronegative cases could easily represent congenital myasthenic syndromes (CMS). The clinical features are similar to adult-onset MG, with the majority of patients initially presenting with primarily ocular symptoms.¹¹⁷ Serum AChR antibodies are present in the majority of affected children. The electrophysiologic findings are also identical to the adult form of the disease.¹¹⁸

ASSOCIATED DISORDERS

Effective management of MG requires knowledge of disorders that occur with an increased incidence in patients with MG, particularly thymic abnormalities and other autoimmune diseases which may occur separately or overlap. Thymus gland pathology is the most notorious disease association.^23,71,119 As many as 80% of patients with seropositive MG have thymic hyperplasia while approximately 12% have thymoma.^23,120–122 Hyperplasia is more common in younger patients. Thymomas occur equally in men and women and occur with the greatest frequency in middle-aged and older individuals.²³ Patients with thymoma, on average, have more severe clinical presentations and higher AChR antibody titers than their nonthymomatous counterparts. Thymic abnormalities have been documented to occur in MuSK MG but are thought to occur far less frequently.^{10,23,71,74,78,81,123} Thymic pathology, either hyperplasia or thymoma, occurs in seronegative and presumably in LPR4 MG.²³ In patients with thymoma, slightly more than half will be found to have or will develop MG.²³

Thymomas are not uniquely associated with MG and may coexist with other autoimmune disorders, other autoantibodies, and a host of other neurological and neuromuscular disorders (Fig. 25-2A,B).¹²⁴ Reported associations include granulomatous myositis, myocarditis, Isaacs’ syndrome, rippling muscle disease, limbic and cerebellar encephalitis, and autonomic neuropathy including the syndrome of intestinal pseudoobstruction.^125–130 Eleven patients with a particularly severe phenotype characterized by bulbar involvement, myasthenic crises, thymoma, myocarditis, and prolonged QT electrocardiographic interval have been described associated with Kv1.4 voltage-gated potassium channel in addition to AChR-binding autoantibodies.¹²⁹

Other autoimmune diseases, most notably Hashimoto’s disease, occur with increased frequency in MG in addition to rheumatoid arthritis, systemic lupus erythematosus, Sjögren’s syndrome, red blood cell aplasia, ulcerative colitis, sarcoidosis, Addison’s disease, and hyperparathyroidism.^131–135 Predictably, these disorders may occur with or without thymic abnormalities. Other neurological or neuromuscular disorders reported to coexist with MG include acute and chronic inflammatory demyelinating polyneuropathies, autonomic neuropathy (e.g., intestinal pseudoobstruction) with or without encephalopathy, Lambert–Eaton myasthenic syndrome (LEMS), acquired neuromyotonia or Isaacs’ syndrome, acquired rippling muscle disease and stiff person syndrome,^{125,128,136–153} In addition, approximately 5% of patients with MG also have an inflammatory myopathy.^{137,154–157} Most of these patients also have a thymoma with or without myocarditis. The histopathology often reveals a giant cell or granulomatous myositis. Serum CK levels are usually elevated with concomitant inflammatory myopathy, which would not be expected in MG alone. It is estimated that the coexistence of other autoimmune diseases approximates 30% in AChR or in seronegative MG as opposed to MuSK MG where the prevalence of other autoimmune disease is estimated to approximate 20%.⁷¹

The diagnosis of myasthenia is usually established clinically and supported by a positive response to one or more of the following tests:

• 
  

  serological—autoantibodies against the AChR or MuSK

  
  
• 
  

  pharmacological—response to edrophonium

  
  
• 
  

  electrophysiological—repetitive nerve stimulation (RNS) or single-fiber electromyography (SFEMG)

As has been repeatedly emphasized, MG should be considered in any patient with painless weakness, occurring in a regional, multifocal, diffuse, or even a seemingly focal pattern. The likelihood of MG increases substantially with the objective demonstration of fatigable weakness, particularly in an oculobulbar distribution. Weakness may be asymmetric and occurs in the absence of fasciculations and in most cases, muscle atrophy.

The differential diagnosis of MG includes other disorders in which signs and symptoms reside predominantly if not exclusively within the voluntary motor system (Table 25-1). Considerations include other DNMT such as the CMS, other acquired DNMT such as LEMS or botulism, motor neuron diseases such as spinal muscular atrophy (SMA), ALS, or X-linked spinal bulbar muscular atrophy (SBMA), numerous myopathies, particularly those with a predilection for cranial musculature, or motor neuropathies when the weakness is largely confined to the limbs.

Due to frequently overlapping phenotypes, and potentially similar electrophysiological features and pharmacological response, the most vexing of these considerations are the CMS. CMS should be seriously considered in any child, adolescent, or young adult with apparent seronegative MG. A history suggesting other involved relatives and/or consanguinity increases the probability of CMS. The CMS will be considered in more detail in the subsequent chapter.

In adults with bulbar weakness, considerations other than MG include but are not limited to the progressive bulbar palsy form of ALS, Kennedy disease, LEMS, botulism, and both acquired and hereditary myopathies. Congenital myasthenia cannot be entirely excluded from consideration as the DOK-7 and rapsyn mutations in particular may be associated with a late-onset phenotype and be readily misidentified as seronegative MG.¹⁵⁸ Most causes of multiple cranial neuropathies typically produce sensory in addition to motor symptoms but occasionally these syndromes may be motor predominant. As the presence of ptosis or ophthalmoparesis essentially excludes motor neuron disease or inflammatory myopathy from consideration, careful surveillance for these abnormalities provides valuable differential diagnostic insight. LEMS is usually dominated by symptoms referable to proximal limb muscles with prominent fatigue and symptoms of cholinergic dysautonomia, but ptosis and bulbar symptoms do occur and can make its distinction from MG challenging in some cases.^159,160 Botulism can affect children and adults and as a presynaptic disorder, can produce cholinergic dysautonomia with constipation and enlarged, unreactive pupils. Its acuity and the clinical context in which it occurs are helpful features to help distinguish it from MG. A number of inherited muscle diseases such as oculopharyngeal muscular dystrophy, mitochondrial myopathy, myotonic muscular dystrophy, and rare adult-onset cases of congenital myopathy such as centronuclear myopathy may produce oculobulbar syndromes.

As mentioned above, MG may rarely present with weakness restricted itself to limb or trunk muscles. The pattern of weakness may be predominantly proximal and symmetric or distal and asymmetric. Accordingly, the differential diagnostic considerations are broad and include anterior horn cell diseases, myopathies, neuropathies with motor predominance including some cases of inflammatory demyelinating polyneuropathy, multifocal motor neuropathy in more chronic and focal cases, as well as other forms of DNMT such as LEMS, botulism and acute organophosphate poisoning. Many of these diseases commonly have cranial nerve involvement as well. In children, SMA, congenital myasthenia, botulism, tick paralysis, and various myopathies would be the primary considerations. Poliomyelitis or other enteroviral infections would readily distinguish themselves in most cases but require consideration along with botulism and tick paralysis in any acute–subacute-onset case due to their pure motor characteristics.

PATHOPHYSIOLOGY

Successful NMT is dependent on the anatomical and physiological capabilities of the NMJ to translate and amplify a peripheral nerve action potential (NAP) into a transsynaptic chemical signal mediated by the neurotransmitter acetylcholine (ACh). Subsequently, the NMJ promotes the generation of a postsynaptic EPP, which unlike an action potential, varies in amplitude. If this EPP achieves the necessary magnitude, an action potential will be generated in the corresponding muscle fiber. This single muscle fiber action potential (SMFAP) tranduces the electrical event into a chemical and eventually mechanical event. The SFMAP promotes calcium release into the sarcoplasmic reticulum which is responsible for myofiber contraction and the generation of force.¹⁶¹ The amount of force generated is dependent on the number of motor units activated, and ultimately, the number of muscle fibers in which this sequence of events takes place.

NMT is a process that is highly conserved between species. It has evolved into an extremely efficient system empowered with a substantial safety factor allowing for fail-safe repetitive and sustained muscle contraction in normal individuals. We have obtained considerable knowledge about the complexities of both normal and abnormal NMT. This section will address NMT and DNMT from a very superficial, clinically relevant perspective.

In view of the sophisticated evolution of NMT into a highly efficient system, external influences are more likely to compromise rather than enhance this efficiency. When NMT is compromised, it typically results in a phenotype typically dominated by fatigue and skeletal muscle weakness. As the neuromuscular junction lies beyond the protection of the blood–nerve barrier and is composed of a large number of proteins essential to its optimal function, it is vulnerable to a large number of immune-mediated, toxic, and genetic influences that can adversely affect both its function and structure. DMNT may result from anatomical and/or physiological disruption of one or more of the three components of the NMJ: (1) the presynaptic nerve terminal in which the synthesis, packaging, storage, presynaptic membrane binding, and/or release of ACh and the vesicles (quanta) that contain it take place, (2) the synaptic cleft through which ACh migrates and is eventually metabolized, and/or (3) the postsynaptic muscle membrane where specialized ACh receptors/channels are optimally positioned and organized (Figs. 23-3 and 23-4). Postsynaptic DNMT may result from multiple potential mechanisms including interference with ACh binding, AChR organization, increased AChR turnover, or overt anatomic disruption (Fig. 23-5).^38,55 This section will review the anatomical, biochemical, and physiological aspects of normal and abnormal NMT.

Although DNMT are usually categorized as belonging to one of the three aforementioned anatomical domains, it would be overly simplistic to assume that all NMJs are the same, that each of the three anatomic NMJ domains develops embryologically in isolation, that any domain functions independently of the other two, or that any DNMT results solely from dysfunction of an individual domain. For example, the presynaptic configurations may differ between different NMJs in different muscles with terminal twigs having either an “en plaque” or “en grappe configuration.” The former refers to large, single contacts on each muscle fiber. It is the predominant form in most mammalian muscles. The latter array refers to multiple, smaller contact points on individual fibers. This configuration seems to correlate with the need for tonic muscle contraction and is most prevalent in nonmammalian systems, but exists in humans in extraocular muscles in particular and in the tensor tympani, stapedius, laryngeal muscles, and tongue as well.³⁹ The greater concentration of fetal-type AChR in extraocular muscles represents one hypothesis as to why these muscles appear to be disproportionately susceptible to AChBR autoantibodies and why they are relatively spared in MuSK MG.¹⁶²

In addition, AChR structure may differ between muscles as well. Muscles innervated by terminal twigs with en grappe morphologies are more likely to have fetal-type AChRs (described below) whose physiological properties may differ from their en plaque counterparts.¹ NMT also differs between different fiber types. Type 2 muscle fibers (fast twitch) have greater sensitivity to ACh than their type 1 counterparts translating to a larger safety factor in NMT. This results from larger nerve terminals, a greater average quantal content, an increased number of postsynaptic folds, and a greater density of sodium channels.^39,162 This is teleologically logical, as firing frequencies in fast twitch fibers are much higher than their slow twitch counterparts, translating to greater ACh depletion in the active zone of the presynaptic region, thereby requiring a greater safety factor in type 2 fibers to ensure uniformly successful NMT.¹⁶²

Lastly, there is a significant interdependence on the three anatomical domains of the NMJ that is evident not only during synaptogenesis but in disease. Optimal postsynaptic architecture and function is very dependent on presynaptic influence.⁹⁰ As an example, acquired postsynaptic MuSK MG and the inherited synaptic form (end plate acetylcholinesterase (AChE) deficiency of CMG both have adverse effects on more than one domain of the NMJ anatomy.⁹⁰ AChE deficiency will have presynaptic effects such as reduction in quantal content and in the size of the presynaptic nerve terminal as well as a postsynaptic effect in the generation of an end plate myopathy.¹⁶³

The sequence of events in normal, successful NMT can be conceptualized as beginning with the synthesis and resynthesis of ACh in the presynaptic terminal and can be schematically followed in (Figs. 25-3 and 25-4). This is accomplished primarily by the enzyme choline acetyltransferase (ChaT) that combines acetate and choline after their reuptake into the presynaptic terminal. Synthesized ACh molecules are packaged into vesicles or quanta that exist in three separate zones; a large storage pool, a mobilization pool, and in clusters close to the presynaptic membrane referred to as the active (immediate release) zone.¹⁶⁴ Although there are non–calcium-dependent mechanisms of ACh release, the efficient function of these active zones is dependent on calcium entry into the distal motor nerve terminal. The P/Q type voltage-gated calcium channels (VGCC) integral to this process are distributed along the active zones at vesicle fusion sites on the presynaptic membrane. In response to a NAP, the presynaptic calcium concentration increases to the 100–1000 μM range.^162,165 Vesicle release occurs after a delay of approximately 100 μs following the NAP with the presynaptic calcium concentration dissipating after approximately 200 μs. As calcium channels are not fully activated by a normal NAP, the capacity to increase quantal content exists by other mechanisms not involved in normal NMT.

In mammalian systems, the contents of 50–300 vesicles are typically released in response to a single NAP, referred to in the aggregate as the quantal content.³⁹ There are approximately 200–400 × 10³ individual synaptic vesicles contained in the average nerve terminal. In mammalian NMJs, approximately 20% of these are positioned for immediate release in the active zones.^166,167 Within each vesicle, there are between 5–10 × 10³ molecules of ACh, the number varying somewhat between individual vesicles.^168,169 In at least one form of congenital MG associated with the impaired synthesis of ACh due to a mutation of the ChaT gene, reduction of the number of ACh molecules within a single synaptic vesicle is reduced sufficiently to interfere with NMT.³⁹

The mobilization pool is estimated to contain 300 × 10³ vesicles that can be moved readily to the active zone region.¹⁶⁸ Following vesicle fusion with the presynaptic membrane and exocytosis, ACh resynthesis and repackaging (endocytosis) take place. Experimental data suggest that the rate of resynthesis parallels the rate of ACh release under normal physiological conditions and is capable of increasing to keep pace with neuromuscular activation.¹⁷⁰

The process of ACh synthesis and resynthesis, vesicle packaging, migration, docking, and exocytotic release into the synaptic cleft is dependent on greater than 1000 functional presynaptic proteins (Fig. 25-3). Detailed description of this obviously complex system is incompletely understood and beyond the scope of this chapter. The key proteins underlying vesicular fusion with the presynaptic membrane are referred to as the SNARE protein complex (soluble NSF attachment protein receptor). Key components of the SNARE complex are synaptotagmin and synaptobrevin (bound to the vesicular membrane), and syntaxin-1 and SNAP 25 (bound to the presynaptic plasmalemma). Interaction between synaptobrevin and syntaxin-1 and SNAP 25 prime the binding and fusion process that culminates by the subsequent binding of calcium to synaptotagmin that triggers quantal release into the synaptic space.³⁹ Subsequent to its role in this process, calcium may freely diffuse away from the active sites, be removed from the nerve terminal by a coupled sodium/calcium exchange mechanism, or be sequestered in the smooth endoplasmic reticulum or mitochondria.¹⁷¹ To the best of our knowledge, there are no recognized acquired or heritable disorders related to the proteins discussed in this paragraph.

The quantal content varies in response to each NAP. Under normal circumstances, the quantal content produces an EPP that far exceeds that necessary to produce a postsynaptic muscle fiber action potential ensuring the fail-safe generation of muscle fiber action potentials with repetitive or sustained attempts at voluntary muscle activation. Although largely irrelevant to normal physiology, presynaptic release of ACh can be augmented as mentioned previously. This effect may be either pathological or therapeutic, depending on the context in which it occurs. In disorders of NMT, the EPP may be augmented by pharmacological intervention at the presynaptic terminal that either prolongs depolarization by blocking potassium channels or the duration and effect of calcium. 3,4 diaminopyridine is an example of the former and guanidine is an example of the latter.^39,162,165 In addition, autoantibodies directed at components of the presynaptic potassium channels may produce a pathological condition of muscular hyperactivity (Isaacs’ syndrome) as reviewed in Chapter 10.

Another presynaptic contribution to NMJ development in particular, and to its maintenance and function as well, is the protein agrin. Agrin, synthesized in and released from the presynapse, contributes significantly to postsynaptic differentiation and to the stabilization of end plate receptors¹⁷² This process of end plate differentiation requires interaction with a number of crucial postsynaptic proteins including LPR4, MuSK, downstream of kinase-7 (Dok7), and receptor- aggregating protein at the synapse (rapsyn).⁹⁰ As will be described in this and the subsequent chapter, impaired NMT may result as a consequence of either heritable defects in many of these proteins (agrin, MuSK, rapsyn, Dok7) or in some cases autoantibodies directed against them (MuSK, LPR4).⁹⁰

The morphology of the synaptic space may be subdivided into the major or primary gap (cleft) between the nerve terminal and muscle and multiple secondary clefts formed by the postjunctional folds extending into the postsynaptic region.³⁹ This folding increases the surface area of the postjunctional membrane by 10-fold in comparison to the presynaptic membrane.^173,174 This allows for an increase in the density of AChRs/channels as described below, thereby improving the efficiency of NMT. The synaptic cleft is narrow with an average distance of 50 nm between the presynaptic membrane and the summits of the postsynaptic folds.³⁹ This narrow gap facilitates rapid NMT. Once released into the synapse, the quanta will briefly bind to and interact with ACh receptors or channels that control the ingress of cations into the muscle fibers on which the receptors are located.

ACh binding to the end plate is short-lived, due in part to diffusion away from the receptor and in large part due to catabolism by the enzyme AChE that is anchored to the basal lamina by its collagen tail, the outer layer of the postsynaptic muscle membrane.³⁹ It is encoded by the triple-stranded collagen Q gene (COLQ) which is relevant to one form of CMS that will be described in Chapter 26. Drugs that reversibly inhibit AChE are used both diagnostically (edrophonium) and therapeutically (pyridostigmine) in MG whereas irreversible anticholinesterases (organophosphates) are utilized for their toxic properties as insecticides or in warfare. There are multiple isoforms of AChE.¹⁷⁵ The primary form is AChE-S (synaptic) which is bound to the basal lamina. Other forms, AChE-E (erythropoietic) and AChE-R (read-through) do not play significant roles in ACh catalysis under normal circumstances but compensatory increases in AChR-R in response to chronic treatment with cholinesterase inhibitors may have detrimental clinical effects as described below.¹⁷⁵

ACh catalysis by AChE is one of the fastest enzymatic processes known and occurs at a rate of five ACh molecules per millisecond.¹⁷⁶ Despite this, ACh enjoys a competitive advantage as the density of AChR (15–30 K × m²) is 5–10 times greater than the molecular density of AChE (2–3 × 10³/m²).¹⁷⁷ This allows 50–75% of the quantal content to achieve successful interaction with AChR under normal circumstances.¹⁷⁸

Temperature changes affect NMT in a variety of ways that may have both electrodiagnostic (EDX) and clinical significance.^{44,45,48,49,179,180} These effects will be addressed here as the most notable physiological effect of temperature change on NMT is on AChE. Reduced temperature slows the rate of AChE hydrolysis of ACh, prolonging the duration of EPPs by allowing ion channels to remain open for longer periods of time and enhancing end-plate responsiveness to ACh.^181,182 The net effect of cooling is to enhance NMT. Electrodiagnostically, this may diminish the probability of demonstrating a decremental response to slow repetitive stimulation. Clinically, cooling may improve function, for example, patient recognition that cold liquids are easier to swallow than warm ones.

Cooling has other non-AChE effects on the physiology of NMT. Both the duration and amplitude of the NAP at the presynaptic terminal are increased by cooling.^163–185 This may be a consequence of prolonged calcium channel open time and augmented quantal content.¹⁸⁶ Reducing the affected muscle’s temperature is known to increase the AChR’s open time as well.¹⁸² Finally, a reduction in muscle temperature leads to a lowering of the resting membrane potential, bringing it closer to threshold, allowing a myofiber action potential to be triggered with a smaller EPP. These four factors, and perhaps others as well, serve to improve NMT in response to cooling.

The organization of the postsynaptic membrane provides for efficiency in NMT (Fig. 25-4). The development and maintenance of end plate complexity is related to the proximity of nerve terminals both embryologically and during adult life and to the influence of ACh and agrin produced and released by these nerve terminals. The key anatomic structure on the end plate is the AChR receptor or channel. Each AChR channel is a glycoprotein composed of five subunits that are arranged in a manner similar to barrel staves turned inside out, resulting in a transmembrane structure with a sagittal appearance similar to the cooling tower of a nuclear power plant. In an adult, the channel consists of two α subunits with singular copies of β, δ, and ε subunits.^39,162,187 Embryologically, the AChR has a γ subunit instead of an ε subunit. The transition to an adult configuration occurs at least in part due to the trophic influence of the presynaptic nerve terminal during the innervation process. Fetal AChRs are typically downregulated during adult life but their presence may have two notable clinical influences. As mentioned above, they may persist to some degree in certain adult muscles, particularly those with en grappe nerve presynaptic morphology, and may contribute to the selected vulnerability of ocular and bulbar muscles in acquired MG. Their persistence in some forms of congenital myasthenia may allow for survival in what would otherwise be a lethal condition.

At the NMJ, AChRs preferentially reside in at the apices of the junctional folds of the muscle end plate and span the postsynaptic membrane. These channels are topographically clustered with their density estimated to approach 10,000 molecules per μm².¹⁸⁸ The density of these channels falls to approximately 10 molecules per μm² within a few microns of the end plate.^39,162 The topography of AChR channel distribution both on and within the muscle end plate is essential for optimal NMT. The placement of these channels is established embryologically and maintained during adult life through the contributions of a number of numerous proteins that are essential for the development and maintenance of channel distribution and their optimal function.¹¹⁴ The initiation of this process is through the effects of the presynaptically synthesized and released protein agrin and ACh, both of which are essential to this process. Agrin knockout mice have normal numbers of AChRs but no evidence of clustering. Agrin appears to bind postsynaptically to LRP4 which in turn results in MuSK activation.¹⁸⁹ MuSK is required not only for proper synaptogenesis, but for the stabilization and maintenance of end plates in postnatal life. Knockout MuSK mice embryos fail both to develop AChR clusters or to survive.³⁹ Deletion of MuSK in adult muscle leads to the degradation and complete loss of NMJs. MuSK reacts in turn with Dok-7 to activate certain downstream signaling pathways and with rapsyn. ACh clustering is maintained by rapsyn which directly binds with the cytoskeleton and dystrophin-glycoprotein complex, specifically through α- and β- dystroglycan.^{39,162,190,191} Despite the apparent importance of Dok-7 in this process, defects in this protein do not appear to adversely affect AChR clustering on junctional folds.¹⁶² An additional protein, neuregulin (NRG-1) also appears to play a role in the clustering of AChRs at the NMJ.^39,162

As mentioned, the AChR is a ligand-gated channel that spans the postsynaptic muscle membrane with its long axis oriented perpendicular to this membrane. It has a hydrophilic central pore with the ACh-binding site located on the extracellular surface of each subunit.¹⁹² In the resting state, the central narrow region or waist created by the apex of the convexity of all five subunits meet in opposition, effecting channel closure. Channel opening is dependent on the simultaneous binding of two molecules of ACh with each channel, which then leads to a conformational change, allowing transient opening of the channel pore and, as a result, ion movement. The agonist binding sites for ACh straddle the α/δ or α/ε (α/γ when relevant), identical to the binding site for α bungarotoxin.^114,162 They appear to be distinct from, but close to the MIR which resides on the α subunit and is the locus for AChR autoantibody binding.^114,162 The α/γ binding site appears to have the highest affinity for ACh which could in turn have relevance in regard to the selective vulnerability of certain muscles which have a greater prevalence of this channel type. Although potassium, calcium, and sodium ions are all capable of traversing the channel, sodium conductance is most dynamic due to favorable size, concentration, and electrical gradient considerations.

In response to the random spontaneous release of singular quanta, unrelated to NAPs, individual channels will open and a miniature end plate potential (MEPP) will be created in that muscle fiber. This is presumed to have a trophic influence on muscle but produces no myofiber contraction. MEPPs occur at a frequency of about 0.2–0.03 times per second, resulting in the activation of 1–2 × 10³ AChR channels and the generation of a nonpropagated waveform with a magnitude of 0.5–1 mV.^193,194

The nonpropagating EPP generated by the normal quantal content release typically exceeds 50 mV in amplitude, and in normal individuals, produces a propagating SMFAP in each muscle fiber stimulated.¹⁹⁵ In health there is a “safety factor”, that being an EPP whose magnitude far exceeds that which is required to depolarize the muscle fiber. A typical mammalian resting membrane potential is approximately −80 mV. The threshold for depolarization may be achieved by a change in voltage of only 10–15 mV, thus providing the three- to fourfold safety margin that normally exists in NMT transmission. The magnitude of the EPP is decreased with repetitive stimuli occurring at a frequency of 5 Hz or less, which, at least initially, deplete ACh-containing vesicles in the active zone. Because of the aforementioned safety margin however, this effect has no significance in the normal individual.

The decline in the EPP in response to “slow” repetitive stimulation will not persist indefinitely as the EPP amplitudes begin to increase after the fourth or fifth stimulus attributed to ACh arrival from the mobilization pool. Conversely, and perhaps counterintuitively, the EPP may be augmented substantially by repetitive stimuli at frequencies of 5 Hz or more. This phenomenon of post-tetanic facilitation is attributed, in large part, to enhanced quantal release related to lingering calcium effects within the presynaptic terminal. Again, this phenomenon bears no consequence in the normal individual, as MFAP occurs in each muscle fiber in response to each and every stimulus. This post-tetanic facilitation does not last indefinitely, and the EPP will begin to decline after approximately 1 minute in normal people due to declining ACh availability. This latter phenomenon is referred to as post-tetanic or postexercise exhaustion and can also be utilized as a diagnostic tool with repetitive stimulation studies in patients with suspected DNMT.^196,197 Although post-tetanic exhaustion will not result in NMT failure in normal individuals, it may do so in patients with DNMT whose safety margin for the generation of muscle fiber action potentials is compromised at baseline by disease. In all DNMT, reduction and eventual loss of the EPP safety margin by whatever means is the universal mechanism by which NMT failure and weakness are created.

In addition to the AChR channel, Nav 1.4 sodium channels are also integral to the generation of the muscle fiber action potential.¹⁶⁶ Unlike the AChR, they are clustered at the base rather than the peak of the synaptic folds.^39,175 Their density is 5–10-fold higher in the end plate than in other regions of the sarcolemma and have a greater density in type 2 than in type 1 muscle fibers.¹⁶² Sodium ingress at the NMJ facilitates the EPP generated by ACh channel opening and adds to the safety margin of NMT. Mutation of Nav 1.4 channel gene is a rare form of CMS.

The half-life of an AChR in the junctional membrane is about 8–10 days.¹⁹⁸ Under normal circumstances, there is a normal turnover of AChR which are internalized and degraded. The senescent receptors are internalized by the process of endocytosis and transported to lysosomes for degradation through an intricate network of intracellular tubules. This process is accelerated in disease as described in more detail below, being expedited by cross-linking of channels with AChR autoantibodies.¹⁶² The AChR are not recycled but are replaced by newly synthesized receptors, one reason why DNMT are more treatment responsive than other neuromuscular disorders in which damaged components are not as readily restored, even if the disease process is arrested.

Integral to the pathogenesis of ACHR MG is the reduction of AChR at the end plate as initially demonstrated by Fambrough and Drachman in 1973 through radiolabeled α-bungarotoxin techniques.³ The pathogenic AChBR autoantibodies of MG are of the IgG1 and IgG3 types and bind predominantly to the MIRs of the AChR which exist on the two alpha subunits of the channel.^{32,73,114,199–201} Once bound to the AChR, these antibodies initiate a number of irreversible processes, all of which are directed at the AChR and postsynaptic membrane. The most potent of these appears to be complement-mediated, membrane-attack complex lysis of AChR.^202,203 Resultant end plate changes not only reduce receptor number but have other anatomic consequence of physiological significance that include simplification of the normal corrugated structure of the end plate (Fig. 25-5). This not only reduces the cross-sectional area but widens the synaptic cleft, thus further reducing the probability of ACh/AChR interaction.^173,204,205 Other potential pathophysiological mechanisms include steric hindrance, with autoantibodies physically blocking ACh binding sites on adjacent channels or preventing the conformational change resulting in opening of the ion pore.²⁰⁶

The autoantibodies not only bind to the AChR but also cross-link with other antibodies.^207,208 When the AChRs are cross-linked in this manner, these are reabsorbed by the postsynaptic membrane by a process known as endocytosis. This process takes place under normal circumstances but accelerates up to three times faster in the presence of AChBR autoantibodies. As a result, the normal NMJ AChR half-life of 5–10 days is dramatically reduced and the AChR population diminished.^209–211 As synthesis of new AChRs remains unchanged, there is a significant net reduction of 70–90% AChRs per NMJ.³ Although AChR MG is conceptually a disorder mediated by autoantibodies, there is a significant T-cell-mediated, CD4 lymphocyte component as well.^55,212 T lymphocytes specific to AChRs are found in patients with myasthenia.^213,214 T cell-targeted therapies hold therapeutic promise for MG.^212,215

MG also reduces the number of sodium channels that are clustered at the depths of the end plate folds.¹⁶⁶ This appears to increase the threshold for muscle fiber action potentials.¹⁶² In any event, reduction in the number or function of sodium channels at the muscle end plate will further erode the safety margin of NMT.

MuSK MG

The pathogenesis of MuSK myasthenia is not as well understood as AChR MG but is clearly different. The association between MuSK and MG is based on a number of lines of evidence. There appears to be a strong association between the presence of MuSK autoantibodies and patients with seronegative myasthenic phenotypes. MuSK autoantibodies produce weakness and reduction in MEPP amplitudes in mice and a weakness associated with impaired AChR clustering in rabbits when passively transferred. Children born of MuSK MG mothers may develop transient neonatal MG and respond to plasma exchange (PLEX) treatment.^{74,83,216,217} Lastly, MEPP amplitude reduction has been demonstrated in vitro in an intercostal muscle biopsy specimen acquired from a MuSK MG patient.²¹⁸

Current knowledge suggests that MuSK MG is a disease of disordered AChR distribution as opposed to one of AChR destruction. There appears to be a good clinical pathological correlation between those muscles where clustering is impacted the most and selectively vulnerable muscles such as sternocleidomastoid, diaphragm, and masseter.⁷⁹ Histological studies of MuSK MG patients suggest that unlike AChBR MG, substantial loss of AChR content and IgG/complement does not occur.⁷⁹ Unlike AChBR autoantibodies, MuSK autoantibodies are of the IgG4 subtype and do not activate complement.^72,219 Their primary action in hampering NMT in MuSK MG appears to be fragmentation of the normal AChR clustering at the junctional peaks of the postsynaptic membrane. In addition, MuSK autoantibodies may have synaptic and presynaptic effects including impaired binding of the collagen tail of the AChE molecule and presynaptic reduction in quantal content instead of the usual compensatory increase that occurs in other forms of MG.^178,219 In fact, recent studies suggest that the primary binding site for MuSK autoantibodies is the site of interaction between MuSK and the collagen tail of AChE (ColQ) responsible for the anchoring of AChE on the basal lamina.²²⁰ This may provide an explanation for why MuSK MG patients, as will be subsequently described, uncommonly respond favorably to cholinesterase inhibitors.⁷⁹ Accordingly, MuSK MG may be eventually classified as a synaptic rather than postsynaptic disorder. Like most other postsynaptic proteins, mutation of the MuSK gene may result in rare reported cases of CMG.²²¹

LPR4 MG

The pathogenesis of LPR4 MG is incompletely understood. In support of a pathogenetic role of LPR4 autoantibodies, serum from LPR4 MG patients reduces MEPP amplitudes in mice.⁸ The preponderance of evidence suggests that LPR4 MG is an IgG1-complement–mediated disease similar to AChBR MG.^8,9,25,222 Accordingly, patients with LPR4 MG respond to immunomodulating agents in a manner similar to AChR MG patients although these two autoantibodies rarely, if ever, coexist.²⁵ In contrast, as the normal function of LPR4 involves interaction with MuSK in AChR clustering during synaptogenesis, one would predict a disease mechanism similar to MuSK MG in which complement-mediated end plate destruction does not appear to have a role. Unlike AChR MG, a small percentage of LPR4 MG patients will also harbor MuSK autoantibodies.^9,25

Transient Neonatal MG

The pathophysiology of transient neonatal myasthenia results from the passive transplacental transfer of the mother’s IgG AChR antibodies which bind to the interface of alpha and gamma subunits.¹¹⁴ The reason why this disorder does not happen with greater frequency is unknown.

No discussion of MG pathophysiology would be complete without attempting to provide a cogent explanation for the selective vulnerability of certain muscle groups and the notable asymmetries that may occur in a disease occurring as a consequence of equitable distribution of circulating autoantibodies. Two potential explanations have already been discussed: (1) differing pathophysiological mechanisms with differing autoantibodies and (2) differences in the AChR channel types in different muscles. It has also been hypothesized that the preferential involvement of the external ocular muscles may be related to the elevated temperature of the head compared to limbs. This of course, would provide an inadequate explanation for asymmetry. In the end, adequate pathophysiological explanations for the myasthenic phenotype remain elusive.^223,224

Lastly, the major gap that remains in our understanding of acquired, autoimmune MG pathogenesis is the knowledge of what initiates the disease process.³² As mentioned, Mendelian genetics appears to have a minimal causative role in acquired, autoimmune MG. We agree with those who believe that the future will identify a significant genetic role in disease susceptibility to adverse environmental influence.^23,26,31,32 The role of infectious agents such as the Epstein–Barr virus as one of these potential environmental precipitants remains uncertain.²²⁵ The thymus gland continues to occupy center stage in any MG etiology discussion.^23,226–228 One potential explanation involves the recognition that the thymus contains myoid cells and other types of stem cells that may serve as autoantigens by the expression of AChRs or AChR antigens on their surface.²³ In this same vein, AChR-specific B lymphocytes are found within the thymus gland of MG patients that are capable of generating antibodies to AChR in culture.

The reader is referred to Chapter 2 where the testing modalities used in support of an MG diagnosis are reviewed. In this section, we will briefly review the serological, EDX, pharmacological, and imaging methods available to aid in the diagnosis and management of myasthenic patients. We will focus on their strengths and weaknesses, and the strategies that we employ in their use. The relative sensitivities of different tests used to support the clinical diagnosis of MG have been estimated and are summarized in Table 25-2.^{6,9,10,25,37,55,72–74,77,91,160,229–232}

Diagnostic strategies in MG undoubtedly vary between clinicians and institutions due to individual bias and preference, test availability, and relevant differential diagnostic considerations in individual cases. Even in the most clinically straightforward cases, we believe that diagnostic confirmation should be obtained whenever possible, particularly if immunomodulating treatment or thymectomy is contemplated. The most sensitive test for MG is SFEMG (92–100%) followed by RNS of distal and proximal nerves (0–99%) depending on the muscle tested and whether the disease is limited or generalized in its manifestations.²²⁹ Neither modality is, however, specific for MG. AChR antibody testing is slightly less sensitive (36–94%) depending also on whether the patient has ocular or generalized disease as well as the nature of the assay utilized, but is highly specific.

SEROLOGICAL TESTING

Identification of AChR autoantibodies is the most expeditious means to confirm MG. Although there are different types of AChR autoantibodies as will be subsequently described, the AChR binding antibody is the principal pathogenic antibody tested for. It will be considered synonymous with AChR autoantibodies throughout this chapter unless otherwise specified. Typically, this is the only diagnostic test that we initially order unless there are phenotypic features that would suggest MuSK MG. As mentioned, the sensitivity is estimated at approximately 36–79% in ocular MG, 75–94% in generalized MG, and 66–93% in all MG patients.^{5,37,114,231,233} They are however, highly specific for the MG, being rarely reported as false positives in patients with other autoimmune diseases such as systemic lupus, rheumatoid arthritis, hepatitis, thymoma without MG, inflammatory neuropathy, motor neuron disease, 13% of patients with LEMS, 3% of patients with lung cancer without an apparent neurological disorder and in some asymptomatic relatives of MG patients.¹⁶⁰ We consider the diagnosis to be confirmed if these autoantibodies are present in the appropriate clinical context.

The value of these autoantibodies is for all intents and purposes in establishing the diagnosis initially. Although titers may decline with treatment, in particular following thymectomy, it is generally held that this test cannot reliably determine disease severity, response to treatment, or to predict either remission or relapse.^55,234 Unlike many other tests used in everyday practice, mild elevations in the AChR autoantibody titer are often significant. Conversely, patients without AChR autoantibodies may have a different disease, harbor a different MG autoantibody, have seronegative MG, or on occasion have a false negative result. This latter situation may arise with testing that has been done too early, or in an individual in whom autoantibody formation has been suppressed by immunomodulating treatment or thymectomy.^233,234 For these reasons, testing prior to the initiation of immunomodulating treatment or thymectomy is ideal. As initially seronegative patients may develop autoantibodies over time, repeat testing in a recommended interval of 6 months in the appropriate clinical context may be considered.²³¹

AChR-modulating and AChR-blocking autoantibodies are also commercially available but play a less significant clinical role.^160,235 AChR modulating autoantibodies measure degradation of the AChR in cultured human myotubes.²³² Both of these autoantibodies are most likely to coexist in individuals who are AChR binding autoantibody seropositive and are unlikely to occur in isolation.^126,160,235 In one report, AChR modulating autoantibodies were found in 75% of patients with AChR binding autoantibodies but in only 5% of seronegative patients.²³² It is our practice to order these autoantibodies only in this latter population. High titers of AChR modulating autoantibodies also play a potential role in the detection of thymoma. Seventy three percent of patients afflicted with both thymoma and MG harbor AChR modulating autoantibodies producing a >90% receptor loss.^125,126,142 Although this potentially justifies their use as a screening tool for thymoma detection, we preferentially rely on imaging for this purpose.

AChR-blocking autoantibodies bind to the same site as ACh or α-bungarotoxin, close to but distinct from the MIR on the extracellular domain of the ACh channel.²³² They are found in approximately half of patients with generalized MG, but in only 30% of patients with ocular disease.^160,235 In one study, these autoantibodies were found in 30% of MG patients seropositive for AChR-binding autoantibodies but in no seronegative MG patient. As a result of this insensitivity, we find AChR-blocking autoantibody testing to have limited clinical value.