morphological effects of myasthenia gravis patient sera on human muscle cells

ABSTRACT: Myasthenia gravis (MG) is caused primarily by autoantibod-ies against the nicotinic acetylcholine receptor (AChR), but autoantibodies toother muscle proteins may be present. Many of these proteins have struc-tural or signalling functions, the disruption of which may affect muscle cellmorphology or viability. In order to investigate the role of such autoantibod-ies in MG, we examined the effect of MG patient sera, of different autoan-tibody composition and obtained at different stages of disease severity, onprimary human muscle cells. Sera from MG patients induced changes in cellmorphology from typical elongated cells to an irregular phenotype, causedthe formation of inclusion bodies and intracellular vesicles, and led to adisordered arrangement of actin microfilaments. Sera from the most se-verely affected patients also induced cell death, which did not occur viaclassic apoptosis. The effects were not complement-mediated and weredose- and time-dependent. As the effects observed in the cell culture systemcorrelated with disease severity, a greater understanding of the individualfactors responsible for these effects may improve our understanding of MGpathogenesis and be of value in the assessment of disease in individualpatients.

Muscle Nerve 33: 93–103, 2006

MORPHOLOGICAL EFFECTS OF MYASTHENIA GRAVISPATIENT SERA ON HUMAN MUSCLE CELLS

STEVEN PAUL LUCKMAN, PhD, GEIR OLVE SKEIE, MD, PhD, GEIR HELGELAND,

and NILS ERIK GILHUS, MD, PhD

Institute of Clinical Medicine, Section for Neurology, University of Bergen and HaukelandUniversity Hospital, N5021 Bergen, Norway

Accepted 18 August 2005

Myasthenia gravis (MG), an autoimmune diseasecharacterized by an increased fatigability of skeletalmuscle, has a prevalence of approximately 1 in10,000.33 Ocular muscle involvement is typical ini-tially, with approximately 50% of patients progress-ing to develop weakness of facial, oropharyngeal,and limb muscles. Myasthenic crisis (respiratory fail-ure that requires hospitalization and mechanicalventilation) is a life-threatening complication thatoccurs in 15%–20% of MG patients.15 Autoantibod-ies against the acetylcholine receptor (AChR), which

is present on the postsynaptic membrane of theneuromuscular junction, are present in 85%–90% ofMG patients with generalized disease and in 70% ofpatients with ocular MG.19,20 A thymoma is presentin 15% of patients with MG.35 AChR antibody-medi-ated impairment of neuromuscular transmission oc-curs mainly by complement-mediated muscle dam-age at the postsynaptic membrane, direct blockadeof ligand–receptor interaction, and an increaseddegradation of AChR.17,19,20 Antibody-dependentcell-mediated cytotoxicity (ADCC) has also been re-ported as a mechanism by which autoantibodies maycause muscle damage in MG.37 The AChR antibodytiter does not correlate with disease severity withinthe MG patient population as a whole, due to theheterogeneity of AChR antibody specificity and thepresence of additional antimuscle autoantibodies.AChR antibody concentration, however, does corre-late longitudinally with disease severity within thesame AChR antibody–positive MG patient.

In addition to anti-AChR antibodies, autoanti-bodies to various intracellular muscle proteins havealso been detected in MG patient sera. Antimuscleautoantibodies include antibodies against the ryan-

Abbreviations: AChR, acetylcholine receptor; ADCC, antibody-dependentcellular cytotoxicity; BSA, bovine serum albumin; CA, citric acid antigen; DAPI,4,6-diamidino-2-phenylindone; DMEM, Dulbecco’s Modified Eagle medium;DMSO, dimethylsulfoxide; FCS, fetal calf serum; FITC, fluorescein isothiocya-nate; Hepes, N-(2-hydroxyethyl)piperazine-N�-ethansulfonic acid; IgG, immu-noglobulin G; MG, myasthenia gravis; MuSK, muscle-specific kinase; MW,molecular weight; MWCO, molecular weight cut-off; PBS, phosphate-buff-ered saline; RyR, ryanodine receptor; SLE, systemic lupus erythematosus;SNMG, seronegative MGKey words: acetylcholine receptor; autoantibody; cytotoxicity; morphology;muscle cell culture; myasthenia gravisCorrespondence to: S. P. Luckman; e-mail: [email protected]

© 2005 Wiley Periodicals, Inc.Published online 14 October 2005 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mus.20443

Effects of Myasthenic Sera MUSCLE & NERVE January 2006 93

odine receptor (RyR), a Ca2�-release channel of thesarcoplasmic reticulum of striated muscle,23,24 thegiant filamentous muscle protein titin,2 citric acidantigen (CA),1 myosin, alpha actinin and actin,28,36

rapsyn (a membrane-associated cytoplasmic proteinessential for AChR clustering),3 the transmembranemuscle-specific kinase (MuSK), involved in AChRclustering,10,13 and the kinase scaffold proteingravin.12,26 These autoantibodies occur in variouscombinations both in conjunction with AChR auto-antibodies or in seronegative MG (SNMG; MG with-out detectable levels of AChR autoantibodies22). Therelevance of non-AChR autoantibodies to the devel-opment and progression of MG is uncertain, and it isparticularly controversial whether antibodies againstintracellular targets are capable of entering intactmuscle cells.

Histologically, light-microscopic findings in mus-cle of MG patients include muscle fiber necrosis andevidence of denervation.9,11,30 Some patients also dis-play inflammatory changes such as lymphocyte clus-tering (lymphorrhage), which can be associated withmuscle fiber necrosis.9,11 It is unclear whether lym-phocyte invasion of muscle is the cause of, or aresponse to, necrosis.21 Lymphorrhages do not ap-pear, however, to colocalize specifically with endplates and may represent a general immune distur-bance in MG.25 Muscle fiber atrophy is also a com-mon light-microscopic finding.9,11 Apoptotic celldeath is rarely observed in muscle fibers, with detect-able apoptotic cells likely being infiltrating mononu-clear cells rather than multinucleated myotubes.29 Atthe electron-microscopic level, a widening of theprimary cleft and a loss of the secondary clefts withinthe end plate is often observed.21 This is accompa-nied by antibody and complement localization at thepostsynaptic membrane, indicating that the changesto the neuromuscular junction are mediated, at leastin part, by a complement-dependent mechanism.However, pharmacological and electrophysiologicalabnormalities have been observed in the end platesof MG patients without obvious degenerativechanges to the end plate, indicating additionalpathogenic mechanisms are also involved.21

In vivo studies of MG have indicated potentialdisease mediators and possible mechanisms of ac-tion, such as AChR antibody-mediated impairmentof neuromuscular transmission through comple-ment-mediated muscle membrane damage. How-ever, it is not possible to definitively establish thepathogenically relevant factors and mechanisms insuch a complex system, in which several autoanti-bodies, cytokines, and potential effector cells (e.g.,cells-mediating ADCC) are present. For this reason,

we have established a primary human muscle cellculture system to examine disease mediators andmechanisms. Since many of the target proteins arestructural in nature, we examined the effect of MGpatient sera displaying different antibody composi-tion or derived from patients with different diseaseseverity on muscle cell morphology and cytoskeletalorganization. In order to examine complement-me-diated effects on the cell cultures, heat-inactivatedsera were also examined. MG sera induced cell deathwas also studied in these cultures.

METHODS

Tissue culture reagents were obtained from Biowhit-taker (Verviers, Belgium) All other reagents wereobtained from Sigma Chemical Company (Stein-heim, Germany) unless otherwise stated.

Patients and Sera. Muscle biopsies (frontalis mus-cle, approximately 150 mg) were obtained from pa-tients during cranial surgery at our institution. Theselected patients displayed no muscle or autoim-mune disease. Little variation was observed in theeffects of MG sera between muscle cultures obtainedfrom different donors. However, for consistency, thedata presented here were obtained using culturesfrom a single donor. Sera were collected from MGpatients and control patients and stored as frozenaliquots at �20°C. Sera were selected from 13 pa-tients with MG of varying severity and with a varietyof autoimmune antibodies, including 1 patient withSNMG (MuSK antibody-negative). Where sera wereobtained during plasmapheresis, the first dischargedserum filtrate obtained was used in the studies. Serawere examined for the presence of titin and RyRautoantibodies by Western blotting using standardprotocols. Patient sera are summarized in Table 1.Sera from patients with systemic lupus erythemato-sus (SLE; 6 patients), Sjogren’s syndrome (2 pa-tients), and normal healthy donors (5 patients ran-domly selected from a biobank of over 100 donors)were also examined. Serum obtained during plasma-pheresis from a patient with small cell lung carci-noma and paraneoplastic limbic encephalitis/pe-ripheral neuropathy with anti-Hu autoantibodies wasexamined as a control for the plasmapheresis pro-cess. Normal pooled sera (sera pooled from 100healthy donors, designated np100) were also exam-ined in addition to unpooled sera from the 5 indi-vidual donors. Complement inactivation was per-formed by heating to 56°C for 30 min. Sera werediluted to the appropriate concentration in differen-tiation medium (as detailed later) prior to filter

94 Effects of Myasthenic Sera MUSCLE & NERVE January 2006

sterilization using 0.2-�m syringe filters (Acrodisk,Gelman Sciences, Ann Arbor, Michigan). Separationof sera into fractions of molecular weight (MW)�100 kDa and �100 kDa was performed usingVivaspin 2 100,000 molecular weight cut-off(MWCO) centrifugal concentrators (Sigma). Sera (2ml) were spun for 30 min at 5000� g. The concen-trate (�100 kDa) was then resuspended at one tenththe original concentration in differentiation me-dium (20 ml), and the filtrate diluted 1 to 10 indifferentiation medium, prior to filter sterilization.

All experiments were performed blind with re-spect to the sera being tested.

Muscle Cell Culture Muscle cells were obtained ac-cording to the methods of Thompson et al.34 withslight modification. Biopsies were finely diced andthe cells dissociated by treatment with 0.25% (wt/vol) trypsin, 0.1% (wt/vol) type IV collagenase, and0.1% (wt/vol) bovine serum albumin (BSA). Incu-bation was performed for 1 h at 37°C, with continu-ous agitation. Undigested tissue debris was removedby centrifugation (300� g for 30 s) and the cellsuspension pelleted by centrifugation at 300� g for 5min. Cells were then washed once with Dulbecco’sModified Eagle medium (DMEM), pelleted as previ-ously described, and resuspended in growth medium[DMEM containing 25 mM N-(2-hydroxyeth-yl)piperazine-N�-ethansulfonic acid (Hepes) and 4.5g/L glucose, supplemented with 20% fetal calf se-rum (FCS), 2 nM insulin, 0.5 mg/ml BSA, 2 mMglutamine, 25 �g/ml gentamycin and 0.25 �g/mlfungizone]. The resultant cell suspension contains aheterogeneous cell population. Differential adhe-sion was employed to purify the muscle cells (myo-blasts). The cell suspension was plated on uncoated25-cm2 flasks for 1 h at 37°C to allow contaminatingfibroblasts to adhere. The remaining cell suspensionwas highly enriched in satellite cells, which readilyadhered to flasks coated with 1% gelatin. Cells ad-hered for 24 h were washed once with growth me-

dium to remove remaining nonadherent cells, and10 ml of fresh growth medium was added. This wasremoved and replaced with fresh growth mediumevery 48 h. Once cells had achieved 50% confluency,a second differential adhesion was performed tofurther purify the myoblasts.

Myoblasts were grown until 60%–70% confluent,at which time the cells were trypsinized (0.25% wt/vol trypsin for 10 min), counted, and plated on75-cm2 flasks coated with 1% gelatin in phosphate-buffered saline (PBS) at a density of 5 � 105 cells/flask. Cells were subcultured 10 times, at which pointthey were considered senescent. Cell stocks werefrozen in liquid nitrogen [freezing medium compris-ing 70% DMEM, 20% FCS, and 10% dimethylsulfox-ide (DMSO)]. Stored cells remained viable underthese conditions for at least 2 years. All experimentsemployed 8-well chamber slides (Lab-Tek, NalgeNunc Int, Pittsburgh, Pennsylvania) coated with 1%gelatin in PBS. Cells were plated at a density of 1 �104 cells/well. Cultures were then allowed to grow to60%–70% confluency prior to addition of differen-tiation medium, as subsequently described.

Once the desired level of confluency wasachieved, growth medium was replaced with differ-entiation medium (DMEM containing 25 mM Hepesand 4.5g/L glucose, supplemented with 2% horseserum, 2 mM glutamine, and 25 �g/ml gentamycin)and the cells incubated for a further 3 days prior toaddition of patient or control sera.

Immunostaining for titin (a key structural com-ponent of myotubes) using anti-titin antibody T11 (agift from Dr. Seigfried Labeit, EMBL Heildelberg,Germany) and standard protocols, confirmed theidentity of the cells obtained in primary culture asmyoblasts. In addition, the morphology and fusionof the cells proved that an authentic muscle cellpreparation was obtained.

Effects of MG Sera on Muscle Cell Cultures. Cellswere subcultured into 8-well chamber slides and

Table 1. Summary of the MG sera examined and the morphological changes induced.

Patient sera code a b c d e f g h i j k l m

Antibody compositionAChR � � � � � � � � � � � � �Titin � � � � � � � � � � � � �RyR � � � � � � � � � � � � �Morphological change �� 0

Sera were tested for the presence of AChR, titin, and RyR autoantibodies. The morphological changes induced by these sera were graded as described in thetext. Sera a, b, and c represent MG crisis patients, the sera being obtained during plasmapheresis. Serum c was seronegative (AChR autoantibody negative).Sera j and k represent pre-thymectomy and post-thymectomy samples from a single patient (patient j/k). At the time of sera collection, patients a, b, c, f, g, i,and j/k were all receiving pyridostigmine; patients d–m were receiving prednisone; and patients g, j/k, and m were receiving azathioprine.


differentiated into myotubes as described above.After 3 days, the differentiation medium was aspi-rated and replaced with MG patient or normalcontrol serum (diluted 1:10 in differentiation me-dium), or differentiation medium only, and incu-bated for a further 3 days. For time course studies,cells were incubated with sera for 1– 4 days,whereas for concentration studies, sera were di-luted 1:10 –1:100 in differentiation medium andincubated for 3 days. Heparin control (20 IU/mlin medium) was included for plasmapheresis pa-tient sera. Cells were then washed once with PBSand fixed with 4% formaldehyde for 10 min. Cellcultures were stained with crystal violet to aid vi-sualization by light microscopy. Crystal violet(0.056% solution), 200 �l/well, was added for 3min, followed by two PBS washes. PBS (300 �l) wasthen added to each well, the cells visualized (un-mounted) by light microscopy using an invertedmicroscope (Leitz DMIL, Leica, Wetzler, Ger-many), and representative digital images taken.

To examine any induction of apoptosis, 4,6-dia-midino-2-phenylindone (DAPI) staining for charac-teristic changes in nuclear morphology was utilized.DAPI staining allows the visualization of chromatincondensation and the formation of apoptotic bodiesthat occurs during the later stages of classic (type I)apoptosis. Cells were treated for 3 days, fixed with4% formaldehyde (for 10 min), and stained for 15min with 1 �g/ml DAPI in PBS. Fixed cells were alsocounterstained with fluorescein isothiocyanate(FITC)-conjugated phalloidin (2 �g/ml FITC-phal-loidin in PBS for 1 h) to observe changes to thepolymerized actin cytoskeleton specifically andchanges in cell morphology in general. Slides werethen washed twice with PBS, mounted (Immu-mount, Thermo Shandon, Naperville, Illinois) andvisualized by fluorescence microscopy. Representa-tive images were taken using a Leica DMRXA fluo-rescence microscope (Leitz DMIL, Leica, Wetzler,Germany).

Quantification. Since differentiated myotubes con-tain numerous nuclei, it is difficult to accuratelyquantify cell numbers in these cultures. For thisreason, “cell number” has been defined as the num-ber of nuclei, as assessed by DAPI staining for nu-clear morphology. Cells were cultured and DAPIstaining was performed as described previously. Fiverandom fields were selected and the total number ofnuclei per field were counted. The number of apo-ptotic and dividing nuclei were also assessed, basedon typical morphological criteria: normal nucleiwere defined as those displaying a typical rounded

morphology, apoptotic nuclei as those displayingapoptotic body formation (obvious nuclear disinte-gration), and dividing nuclei as those showingclearly visible chromatin condensation and chromo-somal structures.

The potency for induction of morphologicalchange was also assessed, as defined by the percent-age of cells displaying inclusion body formation. Cul-tures were graded as follows: 0 (less than 5% of cellswith inclusion bodies, similar to control), � (5%–25%), �� (26%–50%), �� (51%–75%), and�� (76%–100%). MG disease severity wasgraded by whether crisis (severe weakness with respi-ratory muscle involvement requiring plasmaphere-sis) had occurred.

Figures. All illustrations presented here utilizedsera diluted 1:10 with differentiation medium and anincubation time of 3 days unless otherwise stated. Allcontrols shown here utilized normal pool serumnp100.

RESULTS

Two of the 13 sera examined (sera a and b, MG crisispatients) elicited a dramatic change in cell shapefrom the typical elongated morphology observed incontrol cultures to a more contracted, irregular,morphology (Figs. 1 and 2) after incubation for 3days. In addition, intense, granular crystal violetstaining was observed in the perinuclear region ofcells treated with these sera, indicating the forma-tion of inclusion bodies, and vesicles were also ob-served in this region (Fig. 1). These effects involved�95% of the cells. The factors responsible for themorphological changes are of MW �100kDa, a sizerange that includes immunoglobulins but not cyto-kines or drugs that may be present in the plasma-pheresis filtrate (Fig. 1). Serum c (seronegative MGcrisis patient) also caused morphological changes,but inclusion body formation occurred to a lesserdegree than with sera a and b (Fig. 2E). Sera d-ldemonstrated intermediate morphological effects,and serum m showed no obvious difference fromcontrol cultures (Fig. 2). The morphologicalchanges correlated with disease severity (crisis vs.noncrisis patients). Sera from patients with mildersymptoms (noncrisis) induced little or no morpho-logical change, whereas sera of more severely af-fected patients, obtained during plasmapheresis forMG crisis, caused more extensive morphologicalchanges (Table 1).

Sera a and b caused a marked reduction in cellnumber (by approximately 50% and �90%, respec-


tively, compared to controls; Figs. 2A and Fig. 3).Sera from patients with less severe disease (noncrisispatients), and serum c caused little or no reductionin cell number over a 3-day incubation period. Mus-cle cell cultures incubated for 3 days in the presenceof 1:10 diluted sera from patients with SLE andSjogren’s syndrome displayed elongated muscle cellmorphology similar to that observed in culturestreated with normal control sera or differentiationmedium only, and little or no evidence of inclusion-body formation. Normal donor sera (when testedeither as individual serum or as pooled sera np100)did not cause morphological changes or induce celldeath (Figs. 1 and 2), and displayed morphologysimilar to untreated (differentiation medium only)controls. Serum obtained during plasmapheresis of apatient with paraneoplastic limbic encephalitis/pe-ripheral neuropathy also failed to elicit the morpho-

logical changes observed with MG patient sera, indi-cating the observed effects are not an artifact due tothe plasmapheresis process. No correlations betweenthe observed morphological effects and medicationat the time of serum collection were observed (Table1).

The effects of MG patient sera were time depen-dent, with inclusion-body formation observed after24 h, followed by retraction of the cell membrane,formation of discrete multicellular clusters of cells,and a progressive loss of cell number over 4 days ofincubation compared to control cultures (Fig. 3,serum b). MG sera also induced morphologicalchanges and caused cell death in a dose-dependentmanner, with 1:10 diluted serum inducing the great-est effects while 1:100 diluted serum caused a smallerchange in morphology after 3 days of treatment (Fig.2A and B).

FIGURE 1. Representative crystal violet staining of cells treated with control serum and MG serum a, demonstrating cellular retraction,inclusion body formation (dark condensed material), and vesicle formation following treatment with MG serum. (A and B) Control, low,and high power. (C–E) MG serum a, low to high power; V indicates cytoplasmic vesicles. (F) �100 kDa MW fraction, serum a. (G) �100kDa MW fraction, serum a. Bars, 100 �m.


Membrane blebbing, a potential indicator of ap-optotic cell death, was observed in MG serum–treated cells (Fig. 4, serum b). Figure 5 shows typicalDAPI staining of the muscle cell cultures treatedwith control serum and with patient serum a. Nor-mal, apoptotic, and dividing cells were detected inboth cultures (Fig. 5A and B), but greater than 99%of cells exhibited the more normal, rounded nuclearmorphology (Fig. 5C and D). No differences in theproportions of apoptotic and dividing cells were ob-served between the two sera. DAPI staining of nor-mal patient serum- and differentiation medium-

treated controls appeared slightly different from thatobserved following treatment with MG patient se-rum. Control cultures showing a bright, punctatestaining pattern, whereas MG patient serum-treatedcells demonstrate a more diffuse staining patternwith some areas showing little or no staining (Fig. 5E–G).

The morphological changes induced by MG serawere accompanied by a rearranged, more disorga-nized actin cytoskeleton. Upon treatment with themore potent MG sera (sera a and b), actin microfila-ments ran in multiple directions, as compared to the

FIGURE 2. Representative crystal violet staining of cultures treated with 7 of the 13 MG sera examined and control serum, demonstratingthe range of morphological effect observed. (A) Serum a; (B) serum a, 1:100 dilution; (C) serum a, heat-inactivated; (D) serum b; (E)serum c; (F) serum f; (G) serum i, (H) serum k; (I) serum m; (J) normal control. Bar, 100 �m.


more parallel alignment observed in control cultures(Fig. 6). Numerous small vesicle-like structures stain-ing both for DNA and polymerized actin were ob-served in the cytoplasm of MG sera–treated cells. Inaddition, occasional nuclei displaying morphologi-cal abnormalities and staining strongly for polymer-

ized actin were observed in MG serum–treated cul-tures (Fig. 6). The morphology of these nuclei wasnot typical of apoptosis. In contrast, sera c–m in-duced changes to the actin cytoskeleton to a lesserdegree, whereas control sera did not affect the actincytoskeleton.

Heat inactivation of MG sera did not prevent themorphological changes and cell death (Fig. 2A andC), indicating that such changes occur via a comple-ment-independent mechanism. Heparin, a possiblecontaminant of sera obtained by plasmapheresis,which has antiproliferative and proapoptotic effectson various cell types,16,18 did not cause the morpho-logical changes observed following treatment withMG sera and caused only a slight decrease in overallcell number.

DISCUSSION

We have demonstrated a direct cytotoxic effect ofMG patient sera on human muscle cells, using aprimary muscle cell culture system. Muscle cells ex-posed to sera from severely affected patients showeda dramatic change in cell morphology that included

FIGURE 3. Time course for induction of morphological changes by control and MG serum. (A) Control, incubation for 1 day; (B) control,incubation for 4 days; (C–F) MG serum b, incubation for 1–4 days, respectively. Cells were stained with crystal violet. Bar, 100 �m.

FIGURE 4. Membrane blebbing typical of apoptosis followingtreatment of cells with MG serum b. Arrows indicate membraneblebs. Cells were stained with crystal violet. Bar, 100 �m.


cell retraction and clustering with rearrangement ofthe actin cytoskeleton, led to the formation of intra-cellular vesicles and inclusion bodies, and inducedcell death. Sera from less severely affected patientscaused morphological changes and induced celldeath to a lesser degree or had no effect. Theseeffects were time and dose dependent and were notcomplement mediated.

The cell death induced by MG patient sera wasnot via the classic apoptotic pathway, as assessed bycharacteristic changes in nuclear morphology. Sincethe observed morphological changes were not com-plement mediated, complement-mediated cell-mem-brane damage leading to necrotic cell death can beexcluded as a potential mechanism. The nuclei ofMG sera–treated cells showed abnormalities, with amore diffuse staining and the presence of dark (ves-icle-like) areas. The significance of these changes,and of the colocalization of polymerized actin andDNA in cytoplasmic vesicles and morphologicallyabnormal nuclei, remains to be established. Suchobservations might indicate that nuclear degrada-

tion is occurring in the absence of typical apoptoticnuclear morphology.

The precise mode of cell death occurring inthese cultures upon exposure to MG patient seraremains to be elucidated. Apoptosis (or type I pro-grammed cell death) is a rare occurrence in myotu-bules, and it is generally considered that cells inmuscle that display typical apoptotic characteristicsare invading mononuclear cells (often seen at sitesof inflammation).29 However, few studies have beenperformed on apoptosis in skeletal muscle (as re-viewed by Primeau et al.32). It is therefore possiblethat multinucleated myotubes undergo a form of celldeath that, although controlled (unlike necrosis),does not exhibit the classic signs of apoptosis ob-served in mononuclear cells. A possible alternativemechanism is autophagic cell death (or type II celldeath), which differs from apoptosis in the followingways: (1) Apoptosis does not involve the formation ofautophagic vacuoles in the early stages, whereas au-tophagic (type II) cell death display prominent au-tophagic vacuole formation. (2) Apoptosis involves

FIGURE 5. Nuclear morphology of cells treated with control serum (A, C, and F), MG serum a (B, D, and G), or differentiation mediumonly (E), as assessed by DAPI staining. (A and B) Examples of normal, dividing, and apoptotic nuclei in cells treated with control serum(A) and MG serum (B). (C and D) Representative images of cells treated with control serum (C) and MG serum (D). (E–G) High-powerimages of individual nuclei treated with differentiation medium only, control serum, or MG serum respectively. Arrows indicate thepresence of nonstaining “vesicle-like” areas present only in MG-treated cells. Bars, 25 �m (A and B), 100 �m (C and D), 10 �m (E–G).


cytoplasmic and nuclear condensation, accompa-nied by cleavage of actin and other cytoskeletal pro-teins, whereas condensation is not a predominantfeature of autophagy, and the cytoskeleton is largelypreserved.7 Since treatment with MG sera inducedthe formation of vesicles/vacuoles, the cytoskeletonappears largely intact, although rearranged, and nu-clear condensation does not appear to occur upontreatment with MG sera, it is possible that MG serainduce autophagic cell death in the muscle cell cul-tures. It should also be noted that excessive autoph-agy has been observed in other muscle disorderssuch as the autophagic vacuolar myopathies.27

It is unlikely that the morphological changes inthe cells that we observed in vitro would occur invivo, since surrounding healthy cells are likely toconstrain affected cells to a shape similar tohealthy myotubes until the cell has become suffi-ciently degraded. Despite this, disorganization ofthe actin cytoskeleton, and potential functionaldisruption through a possible binding of autoan-tibodies against intracellular targets such as titinand RyR, are likely to render affected myotubes

unable to contract normally. The functional impli-cations of the presence of inclusion bodies andvesicles within the cell are unclear. It can be ar-gued that their presence indicates disruptionwithin the cell that may compromise normal func-tion. The presence of inclusion bodies in MG-treated muscle cells is also noteworthy since inclu-sion bodies have also been observed in two othermuscle diseases characterized by muscle weaknessand atrophy, namely sporadic inclusion-body my-ositis and hereditary inclusion-body myopathies.5

The observed effects appear to correlate withdisease severity, with sera from patients requiringplasmapheresis due to MG crisis eliciting thegreatest morphological change and reduction incell number compared to patients with more sta-ble disease. However, the observed effects did notappear to correlate with the presence of individualautoantibodies, such as titin or the RyR, as sum-marized in Table 1. Furthermore, the role ofAChR autoantibodies in the observed effects re-mains unclear, since SNMG serum also causedmorphological changes. Although the observed ef-

FIGURE 6. Arrangement of the polymerized actin cytoskeleton (FITC phalloidin staining) and DNA (DAPI staining) in cells treated withcontrol serum (A and B) and MG serum (sera a; C and D). The n indicates the position of nuclei displaying the normal roundedmorphology, while n* indicates a possible degrading nucleus, staining strongly for both DNA and polymerized actin. Arrows indicateexamples of colocalization of polymerized actin and DNA in small cytoplasmic vesicle-like structures. Bar, 50 �m.


fects do not correlate with the presence of specificautoantibodies, it should be noted that titin andryanodine receptor antibody presence was as-sessed by Western blotting, and so the absolutelevels of these antibodies cannot be correlated tothe observed effects. Purification and testing ofthe autoantibodies will give further insight intotheir individual roles. We are currently investigat-ing the effects of the purified immunoglobulin G(IgG) and non-IgG fractions of MG patient sera inthe muscle cell culture system. Since the factorsresponsible for the morphological changes ob-served are present in the �100 kDa MW fraction ofthe sera, a role of cytokines or medications presentin the patients serum can be excluded as a cause ofthe observed effects. Screening of sera from alarger MG population will be required to fullyinvestigate potential correlations between the au-toantibodies present, disease severity, and the abil-ity to induce morphological changes and celldeath in this culture system.

The relevance of antimuscle autoantibodies suchas anti-titin and anti-RyR in the development andprogression of MG is unclear, due to the intracellu-lar location of the target antigen. In order to interactwith the antigen, circulating autoantibodies mustcross the intact plasma membrane and be free tointeract with the target, which has yet to be proven.However, cell penetration and binding to intracellu-lar antigens has previously been reported for otherautoantibodies,4,8 and circumstantial evidence existsfor internalization of RyR antibodies in vivo.14 Weare currently investigating antibody internalizationand the role of autoantibodies such as anti-RyR andanti-titin in the muscle cell culture system.

Since the muscle cell culture system is not inner-vated, it is unlikely that AChR clustering occurs andthis may lead to less efficient complement-mediatedcell damage. It has previously been reported that ratspinal cord explants are capable of forming neuro-muscular junctions (i.e., innervating) human pri-mary muscle cell cultures.6 It will be of interest tostudy the effects of MG sera in such cultures. Sincecomplement does not play a role in the cytotoxicityobserved, as demonstrated by heat inactivation ofsera, the precise role of AChR autoantibodies, andthe mechanism by which cytotoxicity occurs, remainsto be elucidated.

Part of this material was presented at the 12th InternationalCongress of Immunology and 4th Annual Conference of FOCIS,Montreal, Canada, July 2004, and at the 7th International Con-gress of Neuroimmunology, Venice, Italy, September 2004. Thiswork was supported by EU grant QL61-CT-2001-01918 and theNorwegian Association for Muscle Disease. We wish to acknowl-

edge the staff of Haukeland University Hospital who aided us inthe collection of patient samples.

REFERENCES

1. Aarli JA, Flood PR, Gilhus NE, Hofstad H, Matre R. Non-receptor striated muscle antibodies in sera from patients withmyasthenia gravis. Monogr Allergy 1988;25:516–521.

2. Aarli JA, Stefansson K, Marton LSG, Wollmann RL. Patientswith myasthenia gravis and thymoma have in their sera IgGautoantibodies against titin. Clin Exp Immunol 1990;82:284–288.

3. Agius MA, Zhu S, Kirvan CA, Schafer AL, Lin MY, FaircloughRH, et al. Rapsyn antibodies in myasthenia gravis. Ann NYAcad Sci 1998;841:516–521.

4. Alarcon-Segovia D, Ruiz-Arguelles A, Llorente L. Brokendogma, penetration of autoantibodies into living cells. Immu-nol Today 1996;17:163–164.

5. Askanas V, Engel WK. Unfolding story of inclusion-body my-ositis and myopathies: role of misfolded proteins, amyloid-beta, cholesterol, and aging. J Child Neurol 2003;18:185–190.

6. Braun S, Croizat B, Langrange MC, Poindron P, Warter JM.Degeneration of cocultures of spinal muscular atrophy mus-cle cells and rat spinal cord explants is not due to secretedfactors and cannot be prevented by neurotrophins. MuscleNerve 1997;20:953–960.

7. Bursch W, Ellinger A, Gerner CH, Fruhwein U, Schulte-Her-mann R. Programmed cell death (PCD): apoptosis, autoph-agic PCD, or others? Ann NY Acad Sci 2000;926:1–12.

8. Deng J-S, Ballou B, Hofmeister JK. Internalization of anti-nucleolin antibody into viable HEp-2 cells. Mol Biol Rep1996;23:191–195.

9. Engel WK, McFarlin DE. Discussion of Fenichel GM. Musclelesions in myasthenia gravis. Ann NY Acad Sci 1966;135:68–78.

10. Evoli A, Tonali PA, Padua L, Monaco ML, Scuderi F, BatocchiAP, et al. Clinical correlates with anti-MuSK antibodies ingeneralized seronegative myasthenia gravis. Brain 2003;126:2304–2311.

11. Fenichel GM. Muscle lesions in myasthenia gravis. Ann NYAcad Sci 1966;135:60–67.

12. Gordon T, Grove B, Loftus JC, O’Toole T, McMillan R, Lind-strom J, et al. Molecular cloning and preliminary character-ization of a novel cytoplasmic antigen recognized by myasthe-nia gravis sera. J Clin Invest 1992;90:992–999.

13. Hoch W, McConville J, Helms S, Newsom-Davis J, Melms A,Vincent A. Auto-antibodies to the receptor tyrosine kinaseMuSK in patients with myasthenia gravis without acetylcho-line receptor antibodies. Nat Med 2001;7:365–368.

14. Iwasa K, Komai K, Takamori M. Spontaneous thymoma rat asa model for myasthenic weakness caused by anti-ryanodinereceptor antibodies. Muscle Nerve 1998;21:1655–1660.

15. Kirmani JF, Yahia AM, Qureshi AI. Myasthenic Crisis. CurrTreat Options Neurol 2004;6:3–15.

16. Leskinen MJ, Kovanen PT, Lindstedt KA. Regulation ofsmooth muscle cell growth, function and death in vitro byactivated mast cells—a potential mechanism for the weaken-ing and rupture of atherosclerotic plaques. Biochem Pharma-col 2003;66:1493–1498.

17. Li Z, Forester N, Vincent A. Modulation of acetylcholinereceptor function in TE671 cells by non-AChR ligands: possi-ble relevance to seronegative myasthenia gravis. J Neuroim-munol 1996;64:179–183.

18. Liang LH, Ye KH, Zhang HW, Luo YR, Ren XD, Xiong AH, etal. Effect of heparin on apoptosis in human nasopharyngealcarcinoma CNE2 cells. Cell Res 2001;11:311–315.

19. Lindstrøm JM, Seybold ME, Lennon VA, Whittingham S,Duane DD. Antibody to acetylcholine receptor in myastheniagravis. Prevalence, clinical correlates and diagnostic value.Neurology 1976;26:1054–1059.


20. Lindstrøm J. Immunobiology of myasthenia gravis, experi-mental autoimmune myasthenia gravis, and Lambert-Eatonsyndrome. Annu Rev Immunol 1985;3:109–131.

21. Lisak RP, Barchi RL. Myasthenia gravis. Philadelphia: WBSaunders; 1982. p 97–99.

22. Mossman S, Vincent A, Newsom-Davis J. Myasthenia graviswithout acetylcholine receptor antibody: a distinct diseaseentity. Lancet 1986;1:116–119.

23. Mygland Å, Tysnes OB, Aarli JA, Flood PR, Gilhus NE. Myas-thenia gravis patients with a thymoma have antibodies againsta high molecular weight protein in sarcoplasmic reticulum.J Neuroimmunol 1992;37:1–7.

24. Mygland Å, Tysnes OB, Matre R, Volpe P, Aarli JA, Gilhus NE.Ryanodine receptor autoantibodies in myasthenia gravis pa-tients with a thymoma. Ann Neurol 1992;32:589–591.

25. Nakano S, Engel AG. Myasthenia gravis: quantitative immu-nocytochemical analysis of inflammatory cells and detectionof complement membrane attack complex at the end-plate in30 patients. Neurology 1993;43:1167–1172.

26. Nauert JB, Klauck TM Langeberg LK, Scott JD. Gravin, anautoantigen recognized by serum from myasthenia gravis pa-tients, is a kinase scaffold protein. Curr Biol 1997;7:52–62.

27. Nishino I. Autophagic vacuolar myopathies. Curr NeurolNeurosci Rep 2003;3:64–69.

28. Ohta M, Ohta K, Itoh N, Kurobe M, Hayashi K, Nishitani H.Anti-skeletal muscle antibodies in the sera from myasthenicpatients with thymoma: identification of anti-myosin, actomy-osin, actin and alpha–actinin antibodies by a solid phaseradioimmunoassay and Western blotting analysis. Clin ChimActa 1990;187:255–264.

29. Olive M, Martinez-Matos JA, Montero J, Ferrer I. Apoptosis isnot the mechanism of cell death of muscle fibers in humanmuscular dystrophies and inflammatory myopathies. MuscleNerve 1997;20:1328–1330.

30. Oosterhuis H, Bethlem J. Neurogenic muscle involvement inmyasthenia gravis. J Neurol Neurosurg Psychiatry 1973;36:244–254.

31. Oosterhuis HJGH, Limburg PC, Hummel-Tappel E. Anti-ace-tylcholine receptor antibodies in myasthenia gravis. Part 2.Clinical and serological follow-up of individual patients.J Neurol Sci 1983;58:371–385.

32. Primeau AJ, Adhihetty PJ, Hood DA. Apoptosis in heart andskeletal muscle. Can J Appl Physiol 2002; 27:349–395.

33. Storm-Mathisen A. Epidemiology of myasthenia gravis in Nor-way. Acta Neurol Scand 1984;70:274–284.

34. Thompson DB, Pratley R, Ossowski V. Human primary myo-blast cell cultures from non-diabetic insulin resistant subjectsretain defects in insulin action. J Clin Invest 1996;98:2346–2350.

35. Willcox N. Myasthenia gravis. Curr Opin Immunol 1993;5:910–917.

36. Williams CL, Lennon VA. Thymic B-lymphocyte clones frompatients with myasthenia gravis secrete monoclonal autoanti-bodies reacting with myosin, alpha-actinin or actin. J Exp Med1986;164:1043–1059.

37. Xu B-Y, Pirskanen, R, Lefvert, AK. Antibody dependent cell-mediated cytotoxicity—an additional mechanism in humanautoimmune myasthenis gravis. J Neuroimmunol 1999;99:183–188.


morphological effects of myasthenia gravis patient sera on human muscle cells

Documents