mechanisms of acetylcholine receptor myasthenia gravis · journal of neurology, neurosurgery,...

Journal of Neurology, Neurosurgery, and Psychiatry, 1980, 43, 601-610

Mechanisms of acetylcholine receptor loss inmyasthenia gravisDANIEL B DRACHMAN, ROBERT N ADAMS, ELIS F STANLEY, ANDALAN PESTRONK

From the Department of Neurology, Johns Hopkins University School of Medicine, Baltimore,Maryland, USA

SUMMARY The fundamental abnormality affecting the neuromuscular junctions of myasthenicpatients is a reduction of available AChRs, due to an autoimmune attack directed against thereceptors. Antibodies to AChR are present in most patients, and there is evidence that they have a

predominant pathogenic role in the disease, aided by complement. The mechanism of antibodyaction involves acceleration of the rate of degradation of AChRs, attributable to cross-linking of thereceptors. In addition, antibodies may block AChRs, and may participate in producing destructivechanges, perhaps in conjunction with complement. The possibility that cell-mediated mechanismsmay play a role in the autoimmune responses of some myasthenic patients remains to be explored.Although the target of the autoimmune attack in myasthenic patients is probably always the acetyl-choline receptors, it is not yet clear which of these immune mechanisms are most important. It islikely that the relative role of each mechanism varies from patient to patient. One of the goals offuture research will be to identify the relative importance of each of these mechanisms in the in-dividual patient, and to tailor specific immunotherapeutic measures to the abnormalities found.

Reduction ofAChRs at neuromuscular junction in MGIn 1964, Elmqvist and his colleagues first made theimportant observation that the amplitude of min-iature endplate potentials (mepps) was diminished inmuscles from patients with MG.3 They concludedthat the mepp abnormality was due to a reductionin the amount ofACh contained in a single quantum.However, there are other theoretically possibleexplanations of decreased mepp amplitudes, in-cluding a reduced number of AChRs or a falsetransmitter. In order to test the receptor hypothesis,we obtained "motor point" biopsies from tenpatients with myasthenia gravis and a group ofunaffected individuals and disease controls. Thestrips of muscle containing endplates were incubatedwith 1251-a-BuTx, to saturate the AChRs, and theexcess a-BuTx was removed by washing. Thebound 125I-a-BuTx was measured by scintillationcounting, and autoradiography. Our findings showeda marked reduction of AChRs at neuromuscularjunctions of patients with MG, averaging 80%less than controls.8 11 1 This finding of reducedAChRs at myasthenic neuromuscular junctions

Address for reprint requests: Dr DB Drachman, Department ofNeurology, Johns Hopkins School of Medicine, 1721 E. Madison St,Baltimore, Maryland 21205, USA.

has subsequently been confirmed by a-BuTxbinding, 13 and by electrophysiological methods. 14We have continued to carry out motor pointbiopsies, and our findings suggest that in manycases measurement of junctional AChRs may be themost sensitive test for MG (Pestronk, Drachman,Josifek, in preparation).

Role ofAChR deficit in myastheniaOur observations raised the question of whetherthe decrease of available AChRs per se couldaccount for the physiological abnormalities in MG,or whether it represented a secondary phenomenon.In order to explore this question further, we producedpharmacological blockade of AChRs in rats bythe use of the a-toxin from the cobra Naja najaatra,15 and compared the resultant changes inneuromuscular transmission with the abnormalitiesoccurring naturally in myasthenia.16 The mosteffective method of administration of cobra toxinproved to be a single intravenous injection (12 to20 tg) followed by a long observation period of upto 14 hours. Evoked potentials were recordedfrom the calf muscles during repetitive stimulationof the sciatic nerve. Initially, the evoked actionpotentials declined in amplitude as the AChR sites

601

by copyright. on M

ay 27, 2021 by guest. Protected

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.43.7.601 on 1 July 1980. Dow

nloaded from

http://jnnp.bmj.com/

602

were blocked by the toxin. Following equilibrationof the toxin in the musculature, the compoundaction potentials returned to normal or near normalamplitudes. Partial blockade of AChRs developed,with a reduced margin of safety of neuromusculartransmission. During this phase, which remainedstable for many hours, the "myasthenic" featureswere readily seen. Repetitive nerve stimulation at3 Hz reproduced the characteristic decrementalpattern of MG. Furthermore, these responses wereexaggerated by minute doses of d-tubocurarine(1/30th of the rat curarising dose), which is typicalof the response in human MG. Treatment of therats with neostigmine (2-5 to 7-5 tg intravenously)produced marked improvement or recovery of thedecremental responses. Finally, post-tetanic re-ponses, thought to be particularly characteristic ofMG17 were also reproduced by the cobra toxinmodel. Both the early improvement (post-tetanicpotentiation) and later exaggeration of the decre-mental responses (postactivation exhaustion) oc-curred in the rats. Thus, from the point of view ofthe electrophysiological features, the cobra toxinmodel faithfully reproduced the abnormalitiesfound in human MG. Since the a-toxin of Najanaja atra used in this study has been shown to blockAChRs specifically,15 our findings strongly supportthe hypothesis that a reduction of availablereceptors per se can account for the defect of neuro-muscular transmission in MG.

Autoimmur,e abnormality in MGThe possibility that MG might be an autoimmunedisease was originally proposed on the basis ofindirect evidence, consisting of a high rate ofthymic abnormalities, the association betweenMG and other presumed autoimmune diseases,7and reduced complement levels in some patients.18A further clue that the autoimmune attack mightbe directed against AChRs came from the findingthat animals immunised with purified AChR inFreund's adjuvant developed a condition analogousto myasthenia.19

Based on the knowledge of a receptor deficit inMG, and the analogy with the experimental animalmodel, the search for antibodies directed againstAChRs in the human disease was soon begun.Antireceptor antibody was identified by severaldifferent methods,20-24 all of which depend ona-BuTx for their specificity. In the most sensitiveradioimmunoassay, which detects antibody thatbinds to AChR labelled with 125I-a-BuTx, elevatedtiters have been found in 80-90% of patients withMG.23 However, the antibody titre corresponds onlyapproximately to the clinical status of the patients.

Daniel B Drachman et al

Passive Transfer ofMGThe question of whether the circulating antibodiesare pathogenic, or merely represent a secondaryresponse to AChR damage caused by some otheragent, is of paramount importance in understandingthe pathogenesis of MG. A well known "experi-ment in nature" suggested that a circulating factor-possibly immunoglobulin-might be pathogenic:approximately one of every six infants born tomyasthenic mothers manifests transient signs ofMG during the first few postnatal weeks.26 In thepast, the results of numerous attempts 26 to transfermyasthenia from humans to experimental animalsor nerve-muscle preparations weie generally nega-tive or were criticised for inherent faults.27 Howeverthe possibility of a circulating factor assumed newimportance in the light of the discovery of anti-bodies to AChR. In most previous studies, theexposure to myasthenic serum had been brief,lasting minutes to hours. We wondered whethermore prolonged exposure to levels of immunoglo-bulin corresponding to those in the patient, mightbe required for the effect to take place. We thereforedesigned an experiment in which physiologic levelsof human myasthenic IgG could be maintained in ahardy strain (B6D2F1) of experimental mice for upto 14 days.28 Immunoglobulin fractions, preparedfrom the sera of myasthenic patients or controls byammonium sulfate precipitation, were injected dailyinto recipient mice. In order to produce toleranceto the foreign serum, we treated the mice with asingle dose of cyclophosphamide (300 mg/Kg), 24hours after the first dose of immunoglobulin,except in experiments lasting less than four days.Some of the experimental mice became clinicallyweak within a few days and showed decrementalresponses to repetitive nerve stimulation. Themost reliable tests were the measure of miniatureendplate potentials and determination of the numberof AChRs per neuromuscular junction in the dia-phragms of the recipient mice. Passive transfer ofimmunoglobulins from 94% of patients resulted indecreased mepp amplitudes, decreased AChRs atneuromuscular junctions, or both.29 Purificationof immunoglobulin by column chromatographyshowed that the active fraction was IgG, while IgMwas without effect. Furthermore, absorption of IgGby staphylococcal protein A completely eliminatedthe myasthenogenic effect of the immunoglobulinpreparation. Thus, the passive transfer experimentclearly demonstrated the pathogenicity of IgG frommyasthenic patients.

Effect of myasthenic IgG on AChRsTheoretically, myasthenic patients' IgG might reducethe number of available AChRs by several possible

by copyright. on M


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


Mechaniisms of acetylcholicne receptor loss in myastheniia gravis

mechanisms:(1) It might alter the turnover of AChRs, either by

increasing the rate of degradation or by decreasingthe rate of synthesis.

(2) It might block the active site of the receptor.(3) It might damage the AChRs, possibly in con-

junction with complement or cellular elements orboth.The evidence now available suggests that ac-

celerated degradation of AChRs, blockade, anddamage may all be involved.

Accelerated degradation of AChRsThe studies of the effects of myasthenic IgG onreceptor degradation were first carried out in atissue culture system, using a modification of themethod of Devreotes and Fambrough.30 TheAChRs are first labelled with 1251-a-BuTx. As theAChR-125I-a-BuTx complexes are endocytosed anddegraded, the 1251 is released into the culture medium,primarily in the form of iodotyrosine. The rate ofdegradation of AChRs can be readily calculatedfrom the rate of release of 1251 in the medium, andis normally approximately 4% per hour for ratskeletal muscle. When IgG from myasthenicpatients was added to the cultures, the AChR de-gradation rate increased up to two-to three-fold,as compared with the cultures treated with controlIgG.31 The antibody-induced acceleration is triggeredby IgG alone without requiring other humoral orcellular components of the immune system. Boththe normal receptor degradation process and theacceleration produced by myasthenic immunoglo-bulin are temperature-dependent, suggesting thatthey involve the active participation of the musclecells. Suggesting that this must be a very commonphenomenon, approximately 75% of myasthenicpatients' serum immunoglobulins produced acceler-ated degradation of AChRs,32. Similar results havealso been reported from other laboratories, usingimmunoglobulin preparations from human myasthe-tic patients 33 or EAMG animals.34

Mechanism of accelerated ACChR degradation (fig 1)(a) Cross linking. We wondered whether the acceler-ated degradation effect might be due to the ability ofeach IgG molecule to bind two molecules of AChR.35lgG molecules are known to be Y-shaped, with twoarms capable of binding to two identical antigenicsites, and therefore cross-linking them. For thisstudy, we prepared pure IgG, divalent F(ab')2,and monovalent Fab fragments from myasthenicsera, by standard purification and enzymaticcleavage methods. When added to muscle cultures,the IgG (fig IA) and divalent F(ab')2 fragments(fig I B) produced equivalent accelerated rates of

IgGG(

A A~

F(ab')2

B

Fab

Anti-IgG { f

D

Anti-cc Butx

Butx

Fig I Diagrammatic representation of cross-linking ofreceptors by the various antibody fragments.M denotes muscle membrane, AChR acetylcholinereceptor, and a-BuTx a-bungarotoxin. Acceleration ofdegradation was observed in the experimental conditionsshown in panels A, B, D, and E, Fab alone failed toaccelerate acetylcholine-receptor degradation.(Reprinted by permission ofNew England Journal ofMedicine 298:1120, 1978).

receptor degradation. By contrast, the monovalentFab fragments failed to accelerate the degradationrate (fig IC), although they bound to AChRs incultured skeletal muscle. This suggested that cross-linking of AChRs might account for the effect of theIgG molecule or its divalent fragment. To test thecross-linking hypothesis further, we carried outadditional experiments:

(1) Monovalent Fab antibody fragments werefirst added to the cultures. In this case, a second"piggyback" antibody, directed against the anti-body fragments was then added (fig ID). The effectof the second antibody was to cross-link the Fabfragments, thus indirectly cross-linking the AChRs.This manoeuver had the same effect in acceleratingAChR degradation as did the original divalentantibodies.

(2) In this experiment, we wished to determinewhether direct contact of an antibody with the AChRwas necessary for accelerated degradation, or

603

by copyright. on M


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from



whether cross-linking per se could produce the sameeffect. To answer this question, we again used a"piggyback" technique, but here a-BuTx ratherthan an antibody fragment was attached to theAChRs. Antibodies against a-BuTx were preparedby immunising rabbits with a-BuTx. When cultureslabelled with 1251-a-BuTx were treated with theantibodies, the degradation rate was again acceleratedtwo-to three-fold (fig 1E). The specificity of thiseffect was demonstrated by control experiments inwhich the binding sites of the antibodies wereblocked by adding a-BuTx, which completelyabolished its effect on degradation. This experimentclearly demonstrated that accelerated degradation isattributable to cross-linking via interposed a-BuTx,without requiring direct contact between antibodyand AChR.(b) Selectivity of degradation. Theoretically, themyasthenic IgG might increase AChR degradationby: (a) accelerating the muscle cell's overall receptordegrading mechanism, or (b) altering only AChRsto which IgG is bound, so that the receptors areselectively degraded at a more rapid rate. Ourresults336 favour the latter possibility.We prepared cultures in which one set of AChRs

was directly exposed to myasthenic patients' IgGfor two hours; a second set of AChRs was allowedto develop immediately after the IgG exposure.Each set of receptors was separately labelled with1251-a-BuTx, and its degradation rate independentlyfollowed.3f (See fig 2 for details.) Our findingsshowed that only the set of AChRs directly exposedto myasthenic IgG (and therefore having bound IgG)was degraded at an accelerated rate two to threetimes normal, while the second set of receptors(without bound IgG) was degraded at the controlrate. This suggested that the binding of IgG alteredthe receptors in some way that caused them to bepreferentially selected for degradation.(c) Mechanism of loss of AChRs. There is consider-able evidence that the normal AChR degradationprocess begins by endocytosis or "internalisation"of AChRs, which are then degraded by lysosomalenzymes. Presumably, the mechanism of accelerateddegradation also involves endocytosis and enzymaticlysis. In both normal and antibody-treated cultures,lowering the temperature to 10°C prevents degrad-ation, indicating that it is an active, energy-dependent process. The role of lysosomal enzymesin this process is demonstrated by the markedeffect of the lysosomal enzyme inhibitors, leupeptinand antipain, in muscle cultures. Both the normaland the antibody-accelerated degradation rates aregreatly slowed by these agents (table l).There is increasing evidence that endocytosis

rather than lysosomal degradation is the rate

IgG(

B

* 125lIQBuTXFig 2 Cultures in set A were first labelled with12iI-a-BuTx then treated for two hours with myasthenicIgG. The labelled receptors have bound IgG. The culturesin set B were first treated with myasthenic JgG. Allexisting AChRs were then blocked with unlabelledat-BuTx. After six hours of incubation to allow forsynthesis and incorporation of additional AChRs, the newAChRs were labelled with 121I-a-BuTx. Thus, thelabelled AChRs do not have bound IgG, although otherreceptors in the same clultures do.The degradation rate of labelled AChRs in set A (withbound IgG) was accelerated while that of labelledAChRs in set B (without bound IgG)was not accelerated.

limiting step in the process of removal of AChRs:(1) As indicated above,3fi two different rates ofdegradation can occur simultaneously in a singlemuscle culture, antibody-bound AChRs being

Table 1 Ejjects ofprotease innibitors on degradationrates ofACh receptors

Control IG Myasthenic IG

No inhibitor 5 1 %±O 35 10O9%±t1-12Leupeptin 100 ug/mi 2 -5 %+0 17 6171% 0 65Antipain 100) ug/mi 2-8%±O00 6-6710-17Effects of protease inhibitors on degradstion rates of AChRs. Setsof 15 cultures were treated with control immunoglobulin or amyasthenic patient's immunoglobulin as indicated. Inhibitors ofIysosomal enzymes were added to sets of 5 cultures each. Meandegradation rates are given as percent of AChRs degraded perhour+5.D. Note that both leupeptin and antipain inhibit degradationin control cultures and in cultures treated with myasthenicimmunoglobulin.

degraded more rapidly than those without boundantibody. This fits the concept of a selective in-crease in endocytosis brought about by the anti-body, but is not consistent with an overall increasein the lysosomal degradation rate. (2) Further,Goldberg et al have reported that treatment ofchick muscle cultures with lysosomal enzyme in-hibitors retards the release of degradation productsinto the medium, but does not slow the rate of lossof AChRs from the surface membrane.37 theabove findings support the concept that the endocy-totic step is rate ilimiting for both normal and anti-body-treated AChRs, while the lysosomal enzymesystem is capable of degrading all the AChRs

604

by copyright. on M


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


Mechanisms of acetylcholine receptor loss in myasthenia gravis

presented to it, unless it is specifically inhibited.Since many patients benefit from treatment with

adrenal corticosteroids, we wondered whether thesehormones might in some way reduce the effect ofantibody on AChRs. The addition of hydrocorti-sone, up to 0 5 mg/ml to IgG-treated muscle cultures,failed to interfere with the accelerated degradationproduced by myasthenic patients' antibody (table 2).(d) How does myasthenic antibody enhance endocy-tosis? Although it is clear that myasthenic antibodymust cross-link AChRs to induce accelerated de-gradation, it is not yet certain how cross-linkingincreases the rate of endocytosis. Autoradiography38and fluorescence microscopy39 suggest that the addi-tion of antibody from myasthenic patients or EAMGanimals causes receptor aggregation. The mostlikely possibility is that the cross-linked AChRaggregates are recognised because of altered size ormobility within the membrane, and preferentiallyselected for accelerated degradation.

Effect of myasthenic immunoglobulin on AChRs atintact neuromuscular junctions.There are important differences between the extra-junctional AChRs of cultured muscle, and AChRs ofneuromuscular junctions, with respect to physicalproperties, pharmacology, and kinetics of turnover.40 These differences have raised questions aboutthe applicability of results of tissue culture experi-ments to the situation at intact neuromuscularjunctions. We therefore studied the effects ofmyasthenic immunoglobulin on the turnover ofAChRs at intact mammalian neuromuscular junc-tions, both in vivo and in vitro.41 Labelling ofAChRs of intact mouse diaphargms was accom-plished by intrathoracic injection of 125I-a-BuTx, andrestriction of the mice in the vertical position toallow the toxin to gravitate to the diaphragms. Themice were then given daily intraperitoneal injectionsof immunoglobulin from individual myasthenicpatients or from a control pool. At intervals of 0 tofour days, the mice were killed, and the 1251-a-BuTxbound to the diaphragms was determined by gammacounting. The mice treated with myasthenic im-munoglobulin showed a loss of radioactivity two tothree times as rapid as in the diaphragms of controlmice. This indicated a more rapid loss of AChRs(fig 3). Similar results have been obtained using serafrom EAMG rats and whole diaphragms invitro.53 54

In order to determine whether the loss of AChRswas due to degradation, diaphragms from micetreated in the same way were maintained in organculture for more than 24 hours. Again, the dia-phragms treated with myasthenic patients' IGshowed a three-fold acceleration of loss of radio-

Table 2 Effect of hydrocortisone on degradation ratesofACh receptors

Concentration of hydrocortisone mg/ml

05 01 001 0

ControlIG 3-76%±0-15 3-25%±0 13 3-75%±0-11 4-01%±0-14

MG 1 I1-59%±0-99 10-62%±0-21 10-87%±0-77 10-87%±0-93MG2 9-09%±0-31 8-83%±1-11 8-71%±0-34 9 04%±034MG3 9 33%±0-32 9-97%±100 8-72%±0-25 9-92%±0±39

Effect of hydrocortisone on degradation rates of AChRs. Degradationrates of AChRs were determined as indicated in the text. Sets of20 cultures were treated with immunoglobulin from three differentmyasthenic patients and a control pool. Hydrocortisone, in theconcentration indicated, was added to sets of five cultures treatedwith each serum immunoglobulin preparation. Mean degradationrates are given as percent of AChRs degraded per hour ± S.D.Note the lack of effect of hydrocortisone on either the normal oraccelerated degradation rates.

activity. The radioactive material released into theculture medium was collected and analyzed bycolumn chromatography. It consisted of a lowmolecular weight substance, which co-migrated withmonoiodotyrosine (fig 4). Our findings thus con-

I

Ecr:U

CL

20,000L

15,000

10,000 _-

0 1 2

AA

3 4

DAYS OF TREATMENT

Fig 3 Effect of in vivo treatment with myasthenic orcontrol IG on the ACCh receptors of mouse diaphragms.Diaphragms were labelled by an intrathoracic injectionof 125I-a-BuTx. The mice were given daily doses of IGfrom pooled serums of control patients (O) or fromtwo different myasthenic patients (0, A). The diaphragmswere removed at intervals after the beginning oftreatment (time zero), and the radioactivity remainingwas counted. Each point represents the number of countsper minute recordedfrom a whole diaphragm of a mousekilled at the time indicated after the first IG treatment.The rate of loss of radioactivity in the diaphragms ofmice treated with myasthenic IG is increased, ascompared to that of the controls.(Reprinted by permission of Science, 200:1285, 1978).

605

0

5,000L

by copyright. on M


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from



Mono- iodo-tyrosineI

200 r

1601-

120Dextranblue

11i80 _

0 0

0

40 _

OL

Fig 4 Chromatogram of culturemedium collectedfrom adiaphragm following an 8-hourculture period. The diaphragm waslabelled in vivo with 122I-a-BuTxand treated with IG from amyasthenic patient. Each pointrepresents the radioactivityrecorded in a 1 25 ml fractioneluted with tris buffer (0 05 M,pH 7 4) from a Sephadex G-15column (1 by 13 cm). More than90% of the radioactivity waseluted at the same position ascarrier monoiodotyrosine(unlabelled). Less than 5 0%appeared in the void volume,marked by the high-molecular-weight dye dextran blue.(Reprinted by permission ofScience, 200:1286, 1978).

I I I I I I

0 5 10 15 20 25 30Fraction number

firmed the degradative nature of the process. Theseobservations indicated that immunoglobulin frommyasthenic patients increases the rate of degradationof AChRs at intact neuromuscular junctions ofmammalian muscles. This process may representan important antibody mediated mechanism inhuman MG.

Blockade of the active site of the AChRThe analogy between MG and curare-like blockadeof neuromuscular tiansmission was first pointedout to Mary Walker by D. Denny-Brown, providingthe background for her trial of physostigmine.Moreover, Simpson's 1960 hypothesis7 suggesteda curare-like antibody capable of blocking AChRs.The first demonstration of anti-AChR antibody inthe serum of myasthenic patients was based on itsability to block a-BuTx binding,20 but only a minority(5 of 15) of myasthenic patients' sera producedsignificant blockade in that report, using solubilisedAChR as the test antigen. In reviewing the liter-ature, it appears that the role of AChR site block-ade in the pathogenesis of MG has been a matterof continuing controversy.4244We have recently developed a method to measure

blockade of the active site of AChRs by myasthenicimmunoglobulin, using the rat skeletal muscle

culture system. The cultures are cooled to 4°C toeliminate degradation, and to minimise possibledissociation of antibody. They are treated overnightin the cold with immunoglobulin prepared frommyasthenic patients' sera, and are then saturatedwith 125I-a-BuTx. The loss of a-BuTx binding sitesin the cultures treated with myasthenic immuno-globulin is attributable to AChR blockade. Ourfindings indicate that serum immunoglobulin fromapproximately 80% of myasthenic patients producedsignificant blockade of 125I-a-BuTx binding. It isnot certain whether this represents binding of theantibody directly at, or near, the active site of theAChR. In the latter instance, steric hindrancecould account for the blocking effect. Finally, itremains to be seen to what extent AChR bloackadecontributes to the clinical manifestations of MG inindividual patients.

Complement mediated mechanism of AChR lossThe effect ofcomplement on AChRs was first demon-strated in the mouse passive transfer model.29 Totest the effect of complement, we depleted mice ofthe C3 component by the use of "cobra venom

factor," a purified fraction derived from the venom

of the Indian cobra Naja naja.45 The C3- depletedmice, and non-depleted controls were treated with

c0

Q.Q

606

by copyright. on M


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


Mechanisms af acetylcholine receptor loss in myasthenia gravis

immunoglobulins from seven different patients. Ineach case, the C3-depleted mice showed a reducedeffect of the immunoglobulin on mepp amplitudesand numbers ofAChR per neuromuscular junctions.To test the effect of the latter part of the comple-ment system, we used pairs of C5-deficient and non-deficient strains of mice. Injection of myasthenicimmunoglobulin produced equivalent changes inmepp amplitudes and AChRs in both strains ofmice. The participation of the early part of thecomplement systems up to and including C3 in thisreaction implies that some effector system inaddition to the binding of antibody to receptormay be involved in the pathogenesis of myasthenia.However, the lack of participation of the latterpart of the complement cascade (C5-C9) in the casestested suggests that cell lysis is not required in thisreaction, since activation of C5 through C9 generallyleads to cytolytic effects.These findings have received further support from

morphological studies demonstrating C3 at neuro-muscular junctions of myasthenic patients46 and bystudies of complement fixation by antibody frommyasthenic patients.47We wondered whether complement might play a

role in the mechanisms of accelerated degradationand blockade described above. To test this, weagain used the rat skeletal muscle tissue culturesystem. Immunoglobulin from each of four differentpatients and a control pool was added to rat skeletalmuscle cultures. In each case, fresh human com-plement, heat-inactivated complement or no com-plement was added to the dishes. At the end of threehours, the cultures were saturated with 125I-a-BuTx,

Table 3 Effect of complement on IG-induced loss ofACh receptors

IG Complemizent Bound 1251 -a BuTxcpm+ S.D.

Control + 35 512±1483Control A 35 324±1421MG I + 28 157±623MGI 1A 30 239±2238MG2 + 26379±1957MG 2 A 27 816±2195MG 3 + 24055±1684MG 3 A 25 048±2763MG 4 + 18 191±1461MG 4 A 20 527 ±558

Sets of five cultures each were treated with immunoglobulin fromeach of four different myasthenic patients or control pool, and withfresh or heat-inactivated human complement (0 3 ml whole serum).At the end of three hours of incubation at 37°C, the cultures weresaturated with 1251-a-BuTx, extracted with Triton X-100 andcounted. Note the loss ofa - BuTx binding sites in all cultures treatedwith myasthenic patients' serum, and the absence of additionaleffect of complement. Each of the myasthenic immunoglobulinsused had previously been tested in vivo in the mouse passive transfermodel, and its effect had been enhanced in vivo by complement.These results. suggest that the effect of complement is not relatedto accelerated degradation or blockade of AChR.

washed, and extracted with Triton X-100. In allcases, the experiments showed that the loss ofAChRs induced by myasthenic immunoglobulinwas not enhanced by the addition of active com-plement (table 3). It thus appears unlikely thatcomplement plays a role in accelerated degradationor blockade induced by myasthenic antibody. Whateffect does complement have on AChRs? Electronmicrographs of neuromuscular junctions frommyasthenic patients show that the postsynapticmembrane sometimes undergoes destructive changes.Fragments of membrane with bound complementappear to be released into the synaptic cleft, oroccasionally undergo phagocytosis by macrophages.46 48 Ultimately, the synaptic folds become short-ened and simplified. It is not yet clear to whatextent the simplification results from complement-induced damage. Alternatively, postsynaptic mem-brane could be lost as a result of endocytosis duringthe process of accelerated degradation of AChRs.

Functional alteration of AChRsThe mechanisms described above all contribute tothe loss or blockade of AChRs. However, thepossibility that antibody might also interferefunctionally with the AChR-ion translocationmechanism has also been considered. Studies ofacetylcholine "noise" have revealed that the opentime of ionic channels was normal at neuromuscularjunctions of myasthenic patients.49 Other possiblemechanism, such as an antibody-induced changein the AChR's affinity for ACh have yet to beexplored.

Cell mediated immunityThe present discussion has focused on humoralfactors effecting a loss of available AChRs inmyasthenia gravis. However, the possibility thatcell-mediated immune responses are also involvedhas been suggested. It is known that lymphocytesfrom myasthenic patients are stimulated by thepresence of purified AChR to undergo blast trans-formation when incubated in the presence of AChRfrom electric eels.50 51 In general, stimulation oflymphocytes in response to a specific antigen indi-cates that the cells have previously been sensitisedto that antigen. It is not yet known, however,whether the responding lymphocytes in MG areT cells or B cells, or whether they are capable ofparticipating in cell-mediated effector responses.Histological studies of skeletal muscles frommyasthenic patients have revealed occasional localcollections of lymphocytes ("lymphorrhages"), sug-gestive of a cell-mediated process.aa However, therarity of these findings at neuromuscular junctionssuggests that direct cell-mediated destruction of

607

by copyright. on M


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


608

AChRs may not be important in most cases of MG.

In 1934 Mary Walker wrote her famous letter to theLancet describing the remarkable effect of physo-stigmine in a patient with myasthenia gravis (MG).1This provided the first evidence directly implicatingan abnormality of the neuromuscular junction inMG. In the debate that followed, it was variouslysuggested that the motor neuron,2 nerve terminal,34neurotransmitter,5 receptor,6 or postsynapticmembrane4 6 was the locus of the defect. Withremarkable perspicacity, Professor Simpson, towhose honor this paper is dedicated, forthrightlysuggested that the acetylcholine "receptor"-at thattime merely a hypothetical construct-was primarilyinvolved, and postulated that it was blocked by anautoantibody.7 With the methods then available,the hypothesis could not be tested. It was not until1973 that we were able to identify the precise natureof the defect in MG as a decrease of availableacetylcholine receptors (AChR) at neuromuscularjunctions.8 In large measure, this was made possibleby the availability of neurotoxins that bind specific-ally to AChRs, including a-Bungarotoxin (a-BuTx)and other elapid toxins.9 During the past sevenyears, our understanding of the pathogenesis of thereceptor defect in MG, and a rational basis fortreatment, have progressed with remarkablerapidity.10 The purpose of this paper is to sum-marise studies that have contributed to the currentconcepts of the mechanism of AChR loss in MG.

We are deeply indebted to the many colleagues whoparticipated in the studies described here, includingD Fambrough, S Satya-Murti, F Slone, K. Toyka,I Kao, D Griffin, J Winkelstein, C Angus, KFischbeck, A Murphy, J Michelson, and G Hoffman.Ms C Barlow provided expert assistance with themanuscript. The original research carried out inthe authors' laboratory was supported in part byNIH grants No 5PONS10920 and No 5RO1HD-04817 and grants from the Musuclar DystrophyAssociation and the Myasthenia Gravis Foundation.

References

1 Walker MB. Treatment of myasthenia gravis withphysostigmine. Lancet 1934; 1:1200-1.

2 Engel WK, Warmolts JR. Myasthenia gravis: A newhypothesis of the pathogenesis and a new form oftreatment. Ann NY Acad Sci 1971; 183:72-87.

3 Elmqvist D, Hofmann WW, Kugelberg J, QuastelDMJ. An electrophysiological investigation ofneuromuscular transmission in myasthenia gravis.JPhysiol 1964; 174:417-34.

4 Engel AG, Santa T. Histometric analysis of theultrastructure of the neuromuscular junction in


myasthenia gravis and in the myasthenic syndrome.Ann NYAcad Sci 1971; 183-46-63.

5 Grob D, Johns RJ, Harvey AM. Studies in neuro-muscular function IV. Stimulating and depressingeffect of acetylcholine and choline on patients withmyasthenia gravis, and their relationships to thedefect in neuromuscular transmission. Bull JohnsHopkins Hosp 1956; 99:153-81.

6 Bergman R, Johns RJ. Ultrastructural alterationsin muscle from patients with myasthenia gravis andEaton-Lambert syndrome. Ann NY Acad Sci 1971;183:88-122.

7 Simpson JA. Myasthenia gravis: A new hypothesis.Scottish Med J 1960; 4:419-36.

8 Fambrough DM, Drachman DB, Satyamurti S.Neuromuscular junction in myasthenia gravis:Decreased acetylcholine receptors. Science 1973;182:293-5.

9 Lee CY. Chemistry and pharmacology of polypeptidetoxins in snake venoms. Ann Rev Pharmacol 1972;12:265-86.

10 Drachman DB. Myasthenia gravis: 2 Parts. N EnglJ Med 1978; 136-142, 186-193.

11 Drachman DB, Fambrough DM, Satyamurti S.Reduced acetylcholine receptors in myastheniagravis. Trans Am Neurol Asso 1973; 98:21-3.

12 Drachman DB, Kao I, Pestronk A, Toyka KV.Myasthenia gravis as a receptor disorder. Ann NYAcad Sci 1976; 274:226-34.

13 Ito Y, Miledi R, Vincent A, Newsom-Davis J.Acetylcholine receptors and electrophysiologicalfunction in myasthenia gravis. Brain 1978; 345-68.

14 Albuquerque EX, Rash JE, Mayer RF, SatterfieldJR. An electrophysiological and morphologicalstudy of the neuromuscular junctions in patientswith myasthenia gravis. Exper Neurol 1976;51:536-63.

15 Chang CC, Lee CY. Electrophysiological study ofneuromuscular blocking action of cobra neurotoxin.Br J Pharmac Chemother 1966; 28:172-81.

16 Satyamurti S, Drachman DB, Slone F. Blockade ofacetylcholine receptors: A model of myastheniagravis. Sciecne 1975; 187:955-7.

17 Desmedt JE. The neuromuscular disorder inmyasthenia gravis. In: JE Desmedt ed, New Develop-ments in EMG and Clinical Neurophysiology. Basle:S. Karger 1973: 241-304.

18 Nastuk WL, Plescia OJ, Osserman KE. Changes inserum complement activity in patients withmyasthenia gravis. Proc Soc Exp Biol 1960; 105:177-84.

19 Patrick J, Lindstrom J. Autoimmune response toacetylcholine receptor. Science 1973; 180:871-2.

20 Almon RR, Andrew CG, Appel SH. Serum globulinin myasthenia gravis: Inhibition of a-bungarotoxinbinding to acetylcholine receptors. Science 1974;186:55-7.

21 Bender AN, Engel WK, Ringel SP, Daniels MP,Vogel Z. Myasthenia gravis: A serum factor block-ing acetylcholine receptors of the human neuro-muscular junction. Lancet 1975; 1:607-8.

22 Appel SH, Almon RR, Levy N. Acetylcholine

by copyright. on M


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


Mechanisms of acetylcholine receptor loss in myasthenia gravis

receptor antibodies in myasthenia gravis. N EnglJ Med 1975; 293:760-1.

23 Lindstrom JM, Seybold ME, Lennon VA, Whittting-ham S, Duane DD. Antibody to acetylcholinereceptor in myasthenia gravis: Prevalence, clinicalcorrelates, and diagnostic value. Neurology 1976;26:1054-9.

24 Mittag T, Komfeld P, Tormay A, Woo C. Detectionof antiacetylcholine receptor factors in serum andthymus from patients with myasthenia gravis. N EnglJ Med 1976; 294:691-4.

25 Namba T, Brown SB, Grob D. Neonatal myastheniagravis: Report of two cases and review of theliterature. Pediatrics 1970; 45:488-504.

26 Bergh NP. Biologic assays in myasthenia gravis forany agents causing a neuromuscular block. ScandJ Clin Invest 1953; 5:Supp 5 5-47.

27 Nastuk WL, Strauss AJL. Further developments inthe search for a neuromuscular blocking agent inthe blood of patients with myasthenia gravis. In:

H.R. Viets, ed. Myasthenia Gravis. Springfield:CC Thomas, 1961: 229-37.

28 Toyka KV, Drachman DB, Pestronk A, Kao I.

Myasthenia gravis: Passive transfer from man tomouse. Science 1975; 190:397-9.

29 Toyka KV, Drachman DB, Griffin DE, Pestronk A,Winkelstein JA, Fischbeck KH, Kao I. Myastheniagravis: Study of humoral immune mechanisms bypassive transfer to mice. N Engl J Med 1977;296:125-31.

30 Devreotes PN, Fambrough DM. Acetylcholinereceptor tumover in membranes of developingmuscle fibers. J Cell Biol 1975; 65:335-58.

31 Kao I, Drachman DB. Myasthenic immuno-globulin accelerates ACh receptor degradation.Science 1977; 196:527-9.

32 Drachman DB, Adams RN, Josifek LF, in prepara-tion.

33 Appel SH, Anwyl R, McAdams MW, Elias S.Accelerated degradation of acetylcholine receptorfrom cultured rat myotubes with myasthenia gravissera and globulins. Proc Nat Acad Sci 1977;74:2130-4.

34 Heinemann S, Bevan S, Kullberg R, Rice J.Modulation of acetylcholine receptor by antibodyagainst the receptor. Proc Nat Acad Sci 1977;74:3090-4.

35 Drachman DB, Angus CW, Adams RN, MichelsonJ, Hoffman GJ. Myasthenic antibodies cross-linkacetylcholine receptors to accelerate degradation.N Eng J Med 1978; 298:1116-22.

36 Drachman DB, Angus CW, Adams RN, Kao I.

Effect of myasthenic patients' immunoglobulin on

acetylcholine receptor turnover: Selectivity ofdegradation process. Proc Nat Acad Sci 1978:75:3422-6.

37 Goldberg AL, DeMartino GN, Libby P. Influenceof Thyroid Hormones and Protease-Inhibitors onProtein Degradation in Skeletal Muscle. In:Aguayo AJ, Karpati G, eds Current Topics inNerve & Muscle Research. Amsterdam: ExcerptaMedica 1979; 53-60.

38 Tarrab-Hazdai R, Yaffe D, Prives Y, Amsterdam A,Fuchs S. Effects of Macrophages and Antibodiesfrom Myasthenic Animals on Muscle Cells inCulture. In: Dau P, ed. Plasmapheresis and theImmunobiology of Myasthenia Gravis, 19:32-40.Boston: Houghton Mifflin, 1979.

39 Lennon VA. Immunofluorescence Analysis ofSurface Acetylcholine Receptors of Muscle: Modu-lation by Auto-Antibodies. In: Jender DJ ed.Cholinergic Mechanisms and PsychopharmacologyNew York: Plenum Press, 1978: 77-92.

40 Brockes JP, Berg DK, Hall ZW. The biochemicalproperties and regulation of acetylcholine re-ceptors in normal and denervated muscle. ColdSpring Harbor Symposium Quant. Biol 1975;40:253-62.

41 Stanley EF, Drachman DB. Effect of myasthenicimmunoglobulin on acetylcholine receptor of intactmammalian neuromuscular junctions. Science 1978;200:1285-7.

42 Albuquerque EX, Lebeda JF, Appel SH, et al.Effects of normal and myasthenic serum factors oninnervated and chronically denervated mammalianmuscles. Ann NY Acad Sci 1976; 274:475-92.

43 Lindstrom J. Immunological studies of acetylcholinereceptors. J Supramolecul2r Structure 1976; 4:389-403.

44 Penn AS. Immunological Features of MyastheniaGravis. In: Aguayo AJ and Karpati G, eds.Topics in Nerve & Muscle Research. Amsterdam:Excerpta Medica, 1975: 123-32.

45 Winkelstein JA, Smith MR, Shin HS. The role ofC3 as an opsonin in the early stages of infectionProc Soc Exp Med 1975; 149:397-401.

46 Engel AG, Lambert EH, Howard FM. Immunecomplexes (IgG and C3) at the motor end-plate inmyasthenia gravis. Mayo Clinic Proc 1977; 52:267-80.

47 Tarrab-Hazdi R, Abramsky 0, Fuchs S. Humoralantibodies to acetylcholine receptor in patients withmyasthenia gravis. Lancet 1975; 11:340-2.

48 Rash JE, Albuquerque EX, Hudson CS, Mayer RF,Satterfield JR. Studies on human myasthenia gravis.Electrophysiological and ultrastructural evidencecompatible with antibody labelling of the acetyl-choline receptor complex. Proc Natl Acad Sci 1976;73:4584-8.

49 Cull-Candy SG, Miledi R, Trautman A. End-platecurrents and acetylcholine noise at normal andmyasthenic human end-plates. J Physiol 1979;287:247-65.

50 Abramsky 0, Aharonov A, Webb C, Fuchs S.Cellular immune response to acetylcholine receptor-rich fraction, in patients with myasthenia gravis.Clin Exp Immunol 1975; 19:11-16.

51 Richman DP, Patrick J, Arnason BGW. Cellularimmunity in myasthenia gravis. N Eng J Med 1976;294:694-8.

52 Russell DS. Histologic changes in the stripedmuscles in myasthenia gravis. J Path Bact 1953;65:279-89.

53 Heinemann S, Merlie J, Lindstrom J. Modulation

609

by copyright. on M


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


610

of acetylcholine receptor in rat diaphragm byanti-receptor sera. Nature 1979; 274:65-8.

54 Reiness CW, Weinberg CB, Hall ZW. Antibody to


acetylcholine receptor increases degradation ofjunctional and extrajunctional receptors in adultmuscle. Nature 1978; 274:68-70.

by copyright. on M


http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


mechanisms of acetylcholine receptor myasthenia gravis · journal of neurology, neurosurgery,...

Documents