OUP user menu

Increased expression of rapsyn in muscles prevents acetylcholine receptor loss in experimental autoimmune myasthenia gravis

Mario Losen, Maurice H. W. Stassen, Pilar Martínez-Martínez, Barbie M. Machiels, Hans Duimel, Peter Frederik, Henk Veldman, John H. J. Wokke, Frank Spaans, Angela Vincent, Marc H. De Baets
DOI: http://dx.doi.org/10.1093/brain/awh612 2327-2337 First published online: 8 September 2005


Myasthenia gravis is usually caused by autoantibodies to the acetylcholine receptor (AChR). The AChR is clustered and anchored in the postsynaptic membrane of the neuromuscular junction (NMJ) by a cytoplasmic protein called rapsyn. We previously showed that resistance to experimental autoimmune myasthenia gravis (EAMG) in aged rats correlates with increased rapsyn concentration at the NMJ. It is possible, therefore, that endogenous rapsyn expression may be an important determinant of AChR loss and neuromuscular transmission failure in the human disease, and that upregulation of rapsyn expression could be used therapeutically. To examine first a potential therapeutic application of rapsyn upregulation, we induced acute EAMG in young rats by passive transfer of AChR antibody, mAb 35, and used in vivo electroporation to over-express rapsyn unilaterally in one tibialis anterior. We looked at the compound muscle action potentials (CMAPs) in the tibialis anterior, at rapsyn and AChR expression by quantitative radioimmunoassay and immunofluorescence, and at the morphology of the NMJs, comparing the electroporated and untreated muscles, as well as the control and EAMG rats. In control rats, transfected muscle fibres had extrasynaptic rapsyn aggregates, as well as slightly increased rapsyn and AChR concentrations at the NMJ. In EAMG rats, despite deposits of the membrane attack complex, the rapsyn-overexpressing muscles showed no decrement in the CMAPs, no loss of AChR, and the majority had normal postsynaptic folds, whereas endplates of untreated muscles showed typical AChR loss and morphological damage. These data suggest not only that increasing rapsyn expression could be a potential treatment for selected muscles of myasthenia gravis patients, but also lend support to the hypothesis that individual differences in innate rapsyn expression could be a factor in determining disease severity.

  • rapsyn
  • experimental autoimmune myasthenia gravis (EAMG)
  • gene therapy
  • electropermeabilization
  • in vivo electroporation
  • α-BT = α-bungarotoxin
  • AChR = acetylcholine receptor
  • BN rat = Brown Norway rat
  • CMAP = compound muscle action potential
  • EMG = electromyography
  • EAMG = experimental autoimmune myasthenia gravis
  • mAb = monoclonal antibody
  • MAC = membrane attack complex
  • MuSK = muscle-specific kinase
  • NMJ = neuromuscular junction
  • RIA = radioimmunoassay
  • VAChT = vesicular acetylcholine transporter


In myasthenia gravis, the muscle form of the nicotinic acetylcholine receptor (AChR) is the target for pathogenic autoantibodies in eight out of ten patients, and the postsynaptic membrane of the neuromuscular junction (NMJ) is the target organ. AChR antibodies cause loss of functional AChR at the NMJ, either by crosslinking the receptors leading to increased internalization (antigenic modulation), by activation of complement leading to focal lysis and morphological damage to the postsynaptic membrane, or by inhibiting the AChR's ion channel function (reviewed by De Baets and Stassen, 2002). Loss of functional AChR impairs neuromuscular transmission, resulting in skeletal muscle weakness.

The AChR is clustered at high density at the NMJ and rapsyn, a 43 kD a protein, is essential for forming and maintaining AChR clusters. The clustering of these two postsynaptic proteins is initiated by agrin, a neuronal synaptogenic protein which acts via the muscle-specific kinase (MuSK) (Glass et al., 1996). MuSK causes phosphorylation of both MuSK and AChR resulting in the clustering and anchoring of preassembled AChR-rapsyn complexes to the cytoskeleton (Moransard et al., 2003). Mutations in the human rapsyn gene (RAPSN) or its promoter can lead to congenital myasthenic syndromes (Ohno et al., 2002, 2003; Burke et al., 2003; Dunne and Maselli, 2003; Banwell et al., 2004; Muller et al., 2004). Mice deficient in rapsyn die shortly after birth because the postsynaptic specialization of the NMJ has failed to develop (Gautam et al., 1995). Rapsyn metabolically stabilizes the AChR, as shown by cotransfection of rapsyn and AChR expression plasmids (Phillips et al., 1997; Wang et al., 1999), and reduces the antigenic modulation of AChRs induced by the anti-AChR monoclonal antibody (mAb)-35 in transfected fibroblasts (Phillips et al., 1997).

Experimental autoimmune myasthenia gravis (EAMG) is an animal model of the human disease, and can be induced by passive transfer of AChR mAbs, e.g. mAb 35, as well as by active immunization against AChR. Antigenic modulation and complement-mediated focal lysis of the postsynaptic membrane are the main pathogenic mechanisms of EAMG. Passive transfer EAMG in the rat is characterized by a very reproducible course of the disease, with substantial loss of AChR and postsynaptic membrane (Lindstrom et al., 1976a; Engel et al., 1979). The maximum muscle weakness is observed 48–60 h after antibody injection in all animals, but susceptibility to EAMG is strongly dependent on strain, sex and age (Hoedemaekers et al., 1997b, c). Aged Brown Norway (BN) rats, in contrast to young animals, are resistant to the induction of passive or chronic EAMG (Graus et al., 1993) and there is little muscle weakness or loss of AChR. This distinction between young and aged BN rats cannot be explained by differences in immune responses or compensatory mechanisms such as increased expression of AChR or complement modulatory proteins (Hoedemaekers et al., 1997a, b, c). However, rapsyn expression is increased in aged BN rats as compared with young BN rats, suggesting a role for rapsyn in reducing AChR loss at the NMJ (Hoedemaekers et al., 1998).

We hypothesized, therefore, that increasing rapsyn expression in EAMG muscles might protect against antibody-mediated loss of AChRs and prevent clinical evidence of muscle weakness. We induced rapsyn overexpression by in vivo electroporation (electropermeabilization), an efficient, non-viral technique (Mir et al., 1999) that is potentially suitable for in vivo use. We examined the effects on the passive transfer model of EAMG induced by mAb 35.


Construction of expression plasmid

The mouse rapsyn cDNA [a gift from Dr Z. Z. Wang (Wang et al., 1999)], was cloned into the mammalian expression vector pcDNA1.1/Amp (Invitrogen, Breda, The Netherlands) using EcoRI restriction sites, resulting in the plasmid pcDNA-rapsyn. The pcDNA-rapsyn plasmid was prepared for electropermeabilization with the Qiagen Midiprep Kit, according to the manufacturer's manual, and finally dissolved in 0.9% NaCl at a concentration of 2 µg DNA/µl.


Eleven-week-old female Lewis rats were obtained from the Department of Experimental Animal Services, University of Maastricht, The Netherlands, with permission of the Committee on Animal Welfare (DEC), according to Dutch governmental rules.


Rats were anaesthetized by breathing halothane (5% induction, 2% maintenance) in NO2/O2 (70%/30%) supplied over a cylindrical ventilation cap held over the head. The hind legs were shaved. One tibialis anterior was injected with 75 µl pcDNA-rapsyn (2 µg/µl) dissolved in 0.9% NaCl and the other with 75 µl 0.9% NaCl. The volume was injected in aliquots at 5–6 sites equally spread over the muscle; 5 min after the injection, the leg was placed between calliper electrodes (BTX, San Diego, CA) with conductive gel. Eight pulses of 200 V/cm, 20 ms, 1 Hz (reversal of polarity after first 4 pulses) were given (Electro Square Porator ECM 830, BTX) (Mir et al., 1999).

Induction of EAMG

Fourteen days after electropermeabilization with pcDNA-rapsyn, rats were divided into two groups: half of the animals received no further treatment (controls), and the other half was injected intraperitoneally with 20 pmol mAb 35 per 100 g body weight (Hoedemaekers et al., 1997a). After another 2 days, the rats were clinically scored, anaesthetized for electromyography (EMG) and sacrificed. Clinical scoring and the EMG measurements were performed by blinded investigators.

Clinical scoring

The severity of clinical signs of disease in EAMG was scored by measuring muscular weakness. The animals' muscle strength was assessed by their ability to grasp and lift repeatedly a 300 g rack from the table, while suspended manually by the base of the tail for 30 s (Verschuuren et al., 1990; Hoedemaekers et al., 1997b, c). Clinical scoring was based on the presence of tremor, hunched posture, muscle strength and fatigability. Disease severity was expressed as follows: 0, no obvious abnormalities; +, no abnormalities before testing, but reduced strength at the end; ++, clinical signs present before testing, i.e. tremor, head down, hunched posture, weak grip; +++, severe clinical signs present before testing, no grip, moribund (Lennon et al., 1975).


Decrement of compound muscle action potential (CMAP) was measured in the tibialis anterior muscles of five EAMG rats 48 h after mAb 35 transfer. Rats were anaesthetized with 60 mg/kg sodium pentobarbital. For stimulation, two small monopolar needle electrodes were used. The cathode was inserted near the peroneal nerve at the level of the knee and the anode more proximal and lateral at a distance of 3–4 mm. For recording, a third monopolar needle electrode was inserted subcutaneously over the tibialis anterior muscle. A ring electrode distally around the relevant hind leg served as a reference and the animal was grounded by a ring electrode around the tail. Stimulation and recording were performed with the EMG system Viking IV (Nicolet Biomedical Inc., Madison, WI). To detect a decrementing response, series of eight supramaximal stimuli were given at 3 Hz. Stimulus duration was 0.2 ms. The test was considered positive for decrement when both the amplitude and the area of the negative peak of the CMAP showed a decrease of at least 10% (Kimura, 2001). To demonstrate reproducibility, at least three recordings were made of all investigated muscles. During the measurements, skin temperature was kept between 35 and 37°C by means of an infrared heating lamp (DISA, Copenhagen, Denmark). All EMG studies were performed by the same investigator (F.S.).

Immunohistochemical staining

Isolated tibialis anterior muscles of four control and four EAMG animals were frozen in melting isopentane. Cryosections of 10 µm were dried, fixed in acetone at 4°C for 10 min and blocked for 20 min with PBSA (phosphate-buffered saline with 2% bovine serum albumin). Sections were incubated with mouse anti-rapsyn mAb 1234 [1/1000 in PBSA (Bloch and Froehner, 1987); a gift of Prof. S. C. Froehner], with mAb 2A1 against rat C5b-9 [membrane attack complex (MAC)], (1/100 in PBSA; provided by W. G. Couser) or with rabbit anti-vesicular acetylcholine transporter (VAChT, 1 : 500; Phoenix Pharmaceuticals, Belmont, CA) for 45 min and subsequently washed with PBS with 0.05% Triton X-100. Subsequently the sections were incubated for 45 min with Alexa594-conjugated α-bungarotoxin (Alexa594-α-BT, diluted 1 : 300 in PBSA; Molecular Probes, Leiden, The Netherlands) and the corresponding secondary antibody: biotinylated donkey anti-mouse Ig (1 : 200, minimal cross-reaction with rat IgG; Jackson Immunoresearch, West Grove, PA) or Alexa350-conjugated goat anti-rabbit (1 : 200; Jackson Immunoresearch). After washing as described above, the biotinylated antibodies were stained with fluorescein-streptavidin (1 : 300; Jackson Immunoresearch). Coverslips were mounted with 0.2 M Tris, pH 8, with 80% (v/v) glycerol.

Quantitative immunofluorescence analysis

Pictures of muscle sections were taken using a Provis AX70 fluorescent microscope (Olympus, Hamburg, Germany) with a digital camera (U-CMAD-2; Olympus) and the AnalySIS software (Soft Imaging Systems, Münster, Germany). The exposure time was set to a constant value ensuring that no saturation of the pictures occurred. Also all other microscope settings were maintained constant. Sections triple stained for VAChT, rapsyn and AChR were photographed using filters for Alexa350, FITC and Alexa594 fluorescence. A 5-fold reduced concentration of primary antibodies and bungarotoxin did not result in a weaker staining, therefore they did not limit staining intensity. Pictures were analysed using the ImageJ software (version 1.32; http://rsb.info.nih.gov/ij). Endplate areas were identified as regions of VAChT staining and the mean intensity of VAChT, rapsyn and AChR staining was measured in the corresponding area. The ratio of AChR : VAChT and rapsyn : VAChT was calculated for >75 endplates as a relative measure for the synaptic AChR and rapsyn concentration. All sections were stained and processed in parallel to avoid inter-assay variations. Differences between endplates in rapsyn-treated versus control samples were analysed using an unpaired t-test.

Two-photon laser scanning microscopy

High-resolution pictures of endplates were taken using a two-photon laser scanning microscope setup as described by van Zandvoort et al. (2004) with the following modifications. A 60× oil immersion objective with a numerical aperture of 1.0 was used. Further magnification was achieved by optical zoom (3×) of the scan head. Picture stacks with an axial resolution of 0.15 µm and a lateral resolution of 0.13 µm were taken using photomultipliers accepting wavelengths of 520–540 nm and >560 nm. To remove noise, each image was filtered applying the Kalman filtering procedure on two subsequent images. For deconvolution, 3-D reconstruction and projection of the images, the Autodeblur and Imaris software (version 4.0.3, Bitplane AG, Zürich, Switzerland) was used.

Radioimmunoassay for measurement of AChRs

The AChR concentrations of isolated tibialis anterior muscles from five control and five EAMG rats were measured as described previously (Lindstrom et al., 1976b; Verschuuren et al., 1992) with the following modifications. Muscles were minced and homogenized with an Ultra-Turrax (3 times for 30 s at 4°C) in 10 ml Buffer A (PBS, 10 mM EDTA, 10 mM NaN3, 10 mM iodoacetamide and 1 mM PMSF). Homogenate was centrifuged (22 100 g, 30 min) and the resulting pellet was resuspended in 2.5 ml Buffer B (Buffer A with additional 0.5% Triton X-100). Extraction was performed for 1 h at 4°C on a reciprocal shaker. After centrifugation (22 100 g, 30 min), six aliquots of 250 µl were taken from the supernatant and incubated with 125I-α-BT (74 TBq/mmol; Amersham Pharmacia Biotech Benelux, The Netherlands) (12.5 µl/ml) followed by rat polyclonal anti-rat AChR (150 µl/ml). As a negative control, three of the aliquots were supplemented with 1 mM acetylcholine and 1 mM neostigmine bromide. After overnight incubation at 4°C, the immune complexes were precipitated with goat anti-rat Ig (100 µl polyclonal serum) for 4 h and centrifuged at 15 000 g for 5 min. The pellets were washed three times in PBS with 0.5% Triton X-100 and measured in a gamma-counter. Results are calculated in fmol ± standard deviation, and differences between concentrations are presented in percentage ± standard error of the difference of the means. Differences between samples were analysed using a one-sided t-test (paired differences were used for comparison of rapsyn-treated versus control in the same experimental group).

Electron microscopy

Five EAMG rats and two control rats with unilateral rapsyn treatment in the tibialis anterior were anaesthetized with 60 mg/kg sodium pentobarbital and transcardially perfused with Tyrode solution (0.1 M) followed by fixation buffer (2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4). The tibialis anterior muscles were removed, post-fixed for 2 h, and sectioned on a vibratome 1000 at 1 mm. Sections were postfixed with 1% osmium tetroxide in 0.1 M phosphate buffer, pH 7.4, dehydrated through a graded ethanol series and embedded in epoxy resin (Glycid ether 100; Serva, Heidelberg, Germany). Endplates were located in toluidine blue-stained semi-thin sections from the central region of each muscle. Ultra-thin sections from selected areas were contrasted with uranyl acetate and lead citrate and viewed with a Philips CM 100 electron microscope. At least five endplate regions were photographed from each muscle. Pictures were scanned for morphometric analysis using the ImageJ software. Analysed parameters included the size of nerve boutons and the length of the presynaptic and postsynaptic membrane (Engel et al., 1976, 1979). Endplates with infiltrated mononuclear cells could not be analysed, but the number of such endplates was counted.


Effects of rapsyn overexpression on rapsyn and AChR distribution in healthy muscle

To look first at the effects of overexpressing rapsyn in normal rat muscles, the left tibialis anterior muscle was electropermeabilized after pcDNA-rapsyn injection (rapsyn-treated muscle). As a control, the right tibialis anterior muscle was electropermeabilized after injection of saline or the empty vector (untreated muscle). After two weeks, the muscles were isolated and cryosections were double-stained for rapsyn, using a mouse antibody, and for AChR using Alexa594-α-BT. Muscle fibres from untreated or treated muscles had AChRs and rapsyn restricted to the NMJs, with co-localization of AChR and rapsyn (Fig. 1A). Approximately 10% of the fibres of rapsyn-treated muscles had, in addition, extrasynaptic rapsyn aggregates. Most were small aggregates distributed over large parts of the cytoplasm (Fig. 1B and C), but some were large, heavily stained extrasynaptic aggregates (Fig. 1D). AChR did not co-localize with any of the extrasynaptic rapsyn aggregates (Fig. 1D).

Fig. 1

Immunohistochemical staining of control and rapsyn-treated control muscles. Cryosections of muscles were double-stained with Alexa594-conjugated α-bungarotoxin (left) and mouse anti-rapsyn mAb1234 (middle); merge on the right. (A) AChR and rapsyn staining of untreated tibialis anterior muscle of a control rat. (BD) Rapsyn overexpression in control rats induced an increased rapsyn concentration in the endplate and variable amounts of extrasynaptic rapsyn aggregates, which did not co-localize with AChR. Bars are 100 µm; all pictures are taken under the same conditions.

To measure the AChR and rapsyn expression more quantitatively, we performed radioimmunoassay (RIA) of muscle extracts labelled with 125I-α-BT, and quantitative immunofluorescence on the muscle fibres, using an antibody to the VAChT as a standard. The results are shown in Fig. 2A–C (control) and in Fig. 2D. Electroporation of the empty vector did not change the AChR concentration compared with untreated muscles (data not shown). Rapsyn-treated muscles showed significantly more AChR content by RIA (+42 ± 15%, n = 5; P = 0.018, paired t-test) than untreated muscles (Fig. 2A, control). Quantitative analysis of the Alexa594-α-BT staining showed a significantly increased amount of AChR per NMJ (Fig. 2B, control; +21 ± 36%, n = 78, P < 0.001, unpaired t-test). Similarly, the relative amount of rapsyn was also increased (Fig. 2C, control; +44 ± 68%, n = 78, P < 0.001, unpaired t-test) compared with untreated control muscle. When the relative amounts of AChR were plotted against the relative amounts of rapsyn (Fig. 2D), there was a highly linear relationship in the untreated muscles (r = 0.81, P < 0.001), but less so in the rapsyn-treated muscles (r = 0.58, P < 0.001). The ratio of rapsyn to AChR was increased from 1 : 1 up to 2 : 1 in some individual rapsyn-treated fibres. The relationship between rapsyn and AChR in the rapsyn-transfected muscles can be fitted to a saturation curve (Fig. 2D, dashed line; r = 0.79 and P < 0.001). These findings suggest that there was potential for increased expression of rapsyn at the NMJs in normal muscle, but a limitation as to how much the AChR concentration could increase at the same junctions.

Fig. 2

Effect of rapsyn overexpression on AChR and rapsyn level in mAb 35-induced EAMG rats. The AChR concentration of untreated and rapsyn-treated tibialis anterior muscles was measured by using 125I-labelled α-bungarotoxin as shown in A and by quantifying α-BT-Alexa594 staining of endplates as shown in B. In the same endplates rapsyn levels as shown in C were analysed by quantification of immunohistochemical staining with anti-rapsyn mAb 1234. Endplate areas were identified using a polyclonal antibody against the VAChT, and the intensities of α-BT-Alexa594 and rapsyn staining were normalized against the intensity of this protein. The average AChR and rapsyn concentration of the untreated muscle of the control animals was set to 1. The relative amount of AChR versus rapsyn in individual endplates is plotted for the control animals in D and for the EAMG animals in E. Closed symbols, endplates of untreated muscles; open symbols, endplates of rapsyn-treated muscles. In the rapsyn-treated muscles, a large proportion of endplates have an increased amount of rapsyn and AchR.

Effects of rapsyn overexpression on neuromuscular transmission and AChR numbers in EAMG rats

First, EAMG was induced in five rats with mAb 35 two weeks after electropermeabilization of the tibialis anterior muscles. Within 48 h all of these five, and all others used in this study (total number = 20) had clear clinical signs of EAMG (score of ++ in each animal). As expected, the untreated tibialis-anterior muscles showed a decremental response of the CMAP, with a mean amplitude decrement of 19.6 ± 5.7% (n = 5) and a mean area decrement of 18.4 ± 2.5%, confirming a neuromuscular transmission defect. In contrast, the contralateral, rapsyn-treated muscles of the same five animals showed no decrement (−1 ± 1%, n = 5), indicating that rapsyn overexpression prevented the development of the NMJ defect in passive transfer EAMG.

To confirm that the electromyographic differences were associated with changes in AChR expression, the concentration of AChR was first measured in EAMG muscle by RIA (on a different set of five EAMG and five control rats). The results are shown in Fig. 2A where they can be compared with those from muscles of control rats. The AChR concentration in untreated muscles of EAMG rats was reduced by 39 ± 10% compared with the untreated muscles of the control group (Fig. 2A; 11 ± 3 fmol/g compared with 18 ± 2 fmol/g; P = 0.001, unpaired t-test), but rapsyn-treated EAMG muscles had significantly higher AChR concentrations than the contralateral untreated muscles (24 ± 9 fmol/g, compared with 11 ± 3 fmol/g, P = 0.013, paired t-test). Indeed the rapsyn-treated muscles of EAMG rats had similar AChR concentrations to those of the rapsyn-treated control muscles (26 ± 5 gmol/g; P = 0.75, unpaired t-test).

To examine further the changes at the NMJ, we performed immunofluorescence studies as above. The examples shown in Fig. 3A and B can be compared with Fig. 1A and B that were prepared and examined under the same conditions. In EAMG rats, AChR and rapsyn staining of endplates of the untreated tibialis anterior muscle was weak (Fig. 3A), whereas that in the rapsyn-treated contralateral tibialis anterior muscle (Fig. 3B) was similar to that in untreated control muscles (compare Fig. 1A). Interestingly, this effect of rapsyn treatment was independent of visible rapsyn aggregates in the fibres (Fig. 3B). The quantification of immunohistochemical staining of endplates in EAMG animals confirmed a reduction in the relative concentrations of AChR (by 41 ± 17%, n = 100, P < 0.001, unpaired t-test) and rapsyn (by 51 ± 15%, n = 100, P < 0.001, unpaired t-test) at the endplates of the untreated EAMG rats (Fig. 2B and C), and demonstrated that the rapsyn treatment increased the relative AChR and rapsyn concentrations in EAMG muscle to levels similar to those in untreated control muscles (1.02 ± 0.40 of AChR; n = 100, P = 0.75 and 1.12 ± 0.65 of rapsyn; n = 100; P = 0.15) although the wide standard deviations reflect the variability in AChR and rapsyn expression at individual NMJs. There was a highly linear correlation between the AChR concentrations by RIA (Fig. 2A) and by quantitative immunohistochemistry (Fig. 2B) in the four experimental groups (r = 0.93).

Fig. 3

Immunohistochemical analysis of rapsyn expression in the NMJ of EAMG rats using standard fluorescence (A and B; scale bars are 100 µm) and two-photon laser scanning microscopy (C and D; scale bars are 20 µm). (A and C) Endplates of an untreated tibialis anterior muscle 2 days after induction of EAMG. Endplates showed reduced concentration or complete loss of AChR and rapsyn. (B and D) Rapsyn overexpression in the contralateral tibialis anterior muscle of the same EAMG animal. Rapsyn-treated endplates appeared very similar to those of rats without EAMG (Fig. 1).

We used two-photon microscopy, at higher resolution, to examine the NMJs in more detail. A representative endplate in an untreated EAMG muscle (Fig. 3C) shows a simplified postsynaptic membrane with reduced amounts of rapsyn and AChR. In contrast, the rapsyn-transfected fibre shows intense staining of rapsyn and AChR in a normal-appearing NMJ, with less intense rapsyn staining of the surrounding cytoplasm and extrasynaptic membrane (Fig. 3D).

The plot of relative AChR concentration versus relative rapsyn concentration at individual endplates (Fig. 2E, closed symbols) in untreated EAMG muscles shows that the relative values of each are lower in general, as expected from Fig. 2B and C, and there is a highly linear correlation between AChR and rapsyn (r = 0.85, P < 0.001) as also found in untreated control muscles (Fig. 2D). But in sections of rapsyn-treated EAMG muscles (Fig. 2E, open symbols), 64 out of 100 (64%) endplates had a rapsyn concentration that was significantly higher than in the contralateral EAMG muscles (>mean + 2 SDs of untreated EAMG endplates), compared with only 6 out of 110 (5%) of the endplates in the untreated muscle sections. Therefore, ∼59% of endplates (64 − 5%) had increased expression of rapsyn. Similarly, 40 out of 100 endplates (40%) had a significantly increased AChR concentration compared with only 6 out of 110 (5%) in the contralateral untreated EAMG muscles. Since the relationship between rapsyn and AChR is linear (0.82 : 1; P < 0.001; Fig. 2E, open symbols), and different from untreated EAMG muscles, it appears that the increased rapsyn expression has helped to protect ∼35% of the NMJs (40 − 5%) from AChR loss.

Rapsyn overexpression does not prevent complement deposition at the NMJ

In untreated muscles, mAb 35 reduced AChR expression and induced deposition of the MAC at the endplates (Fig. 4A), consistent with a complement-dependent degradation of AChRs. Interestingly, in rapsyn-treated muscles, the MAC was still present, but not associated with decreased AChRs (Fig. 4B).

Fig 4.

Immunohistochemical staining of MAC in the NMJs in EAMG rats. Endplates of untreated and rapsyn-treated muscles as shown in A and B respectively were stained with Alexa594-α-BT (left) and mAb 2A1 against rat C5b-9 (MAC, middle); merge on the right. In untreated muscle, there was a little AchR at the end points but strong MAC staining; whereas in rapsyn-treated muscle both AchR and MAC staining was found.

Rapsyn overexpression prevented the destruction of postsynaptic folds in EAMG rats

Therefore, to see whether the rapsyn treatment was associated with less destruction of the postsynaptic membrane in the treated EAMG muscles (as predicted from Fig. 4D), we performed electron microscopic observations and morphometric analysis on two control and five EAMG rats. Rapsyn treatment did not appear to alter the structure of the NMJ in control muscles (Fig. 5A and B and Table 1) which showed normal postsynaptic folds in 28 out of 28 regions examined. In the untreated muscles of EAMG rats, all endplates observed showed morphological abnormalities. At 28 out of 46 endplates, regions with infiltrating cells were found (macrophages and neutrophils); others had reduced postsynaptic folds or even a complete loss of postsynaptic folding (Fig. 5C). The average length of the postsynaptic membrane was significantly reduced compared with normal endplates (P < 0.0001; Table 1). In the contralateral rapsyn-treated EAMG muscles, only 8 out of 49 regions showed a complete destruction of the postsynaptic folding and three more regions were infiltrated by macrophages; therefore altogether 38 out of 49 endplate regions (78%) appeared normal. Moreover, the average length of the postsynaptic membrane in rapsyn-treated EAMG endplates was reduced only slightly compared with normal endplates (difference not significant, P = 0.09), and the ratio of postsynaptic to presynaptic membrane length was normal (P = 0.85). An example of a well-preserved endplate is shown in Fig. 5D. There was a very significant difference between the postsynaptic membrane length of the rapsyn-treated EAMG endplates and the untreated EAMG endplates (P < 0.0001).

Fig 5.

Electron microscopic examination of the postsynaptic folds of motor endplates. NMJ in an untreated and rapsyn-treated tibialis anterior muscle as shown in A and B respectively. (C) Damaged postsynaptic membrane without folds in an untreated tibialis anterior muscle of an EAMG animal. (D) Normal postsynaptic membrane with folds in rapsyn-treated tibialis anterior of an EAMG animal. Arrows indicate the postsynaptic membranes; asterisks indicate the nerve terminals; scale bars are 1 µm.

View this table:
Table 1

Morphometric analysis of endplates*

Regions analysedNerve bouton area (µm2)Presynaptic membrane length (µm)Postsynaptic membrane length (µm)Membrane length ratio (postsynaptic/presynaptic)
Untreated control466.0 ± 0.48.2 ± 0.641.4 ± 4.05.3 ± 0.4
Rapsyn-treated control287.9 ± 1.28.1 ± 0.939.6 ± 4.45.2 ± 0.5
Untreated EAMG187.3 ± 1.28.3 ± 0.716.8 ± 3.7§1.9 ± 0.3§
Rapsyn-treated EAMG466.2 ± 0.77.1 ± 0.532.9 ± 3.05.2 ± 0.6
  • * Mean ± SE.

  • A region here refers to an area of one or two synaptic boutons.

  • Twenty-eight more regions could not be analysed because of infiltrating mononuclear cells.

  • § Significantly different from untreated control endplates (P < 0.0001) and significantly different from treated EAMG endplates (P < 0.0001); other differences in this table are not significant (P > 0.05).

  • Three more regions could not be analysed because of infiltrating mononuclear cells.


Myasthenia gravis is caused by antibody mediated loss of AChRs at the NMJ and most treatments are aimed at modulating the immune response. However, previous observations in rats suggested that intrinsic differences in expression of rapsyn, the AChR anchoring protein, may modify susceptibility to EAMG induced by immunization with AChR or by passive transfer with antibody to AChR. To prove that rapsyn alone was capable of modifying susceptibility, we transfected rapsyn unilaterally into the tibialis anterior muscles of EAMG rats. The rapsyn-treated muscles showed normal numbers and distribution of AChRs and lack of the morphological changes that typify the experimental and human disease, despite the presence of the MAC. Thus, these results imply that even in the presence of an active immune response against the AChRs, upregulation of rapsyn at the NMJ can substantially protect from antibody-induced AChR loss.

Electropermeabilization of tibialis anterior muscles with pcDNA-rapsyn led to rapsyn overexpression in transfected muscle fibres, ranging from a few large extrasynaptic rapsyn aggregates to small rapsyn aggregates densely distributed throughout the muscle fibre cytoplasm. The amount of AChR in rapsyn-treated normal muscles was increased by 42% after 2 weeks, as measured by RIA of total muscle extracts. Similarly, a 50% increase in AChR was found at the cell surface when rapsyn was coexpressed in COS-7 cells (Marchand et al., 2002). The increased AChR concentration in rapsyn-treated muscles was probably due to the metabolic stabilizing effects of rapsyn (Phillips et al., 1997; Wang et al., 1999; Gervasio and Phillips, 2005) and is unlikely to be attributable to upregulation of AChR subunits because AChR gene expression is rapsyn independent (Graus et al., 1993; Gautam et al., 1995; Hoedemaekers et al., 1998). Electroporation of plasmid DNA alone does not lead to a change of AChR concentration in muscles, neither does it protect against mAb 35-induced AChR loss (unpublished data).

Rapsyn overexpression prevented AChR loss and muscle weakness in EAMG rats, which was confirmed by electrophysiological measurements which showed no decrement of the CMAP. Interestingly, this was not just the result of stabilization of the AChRs themselves, but probably also related to lack of morphological damage since 78% of postsynaptic regions of rapsyn-treated muscles of EAMG rats were not damaged or infiltrated, as shown by EM studies, whereas the endplates of untreated EAMG muscles were all damaged, as expected. Engel et al. (1979) found that the postsynaptic membrane length in EM sections was reduced to 26% by day 5 after passive transfer of polyclonal rat anti-AChR antibodies. Our ultrastructural data indicate that mAb 35 reduces the postsynaptic membrane to 36%, even as early as day 2, and that rapsyn overexpression prevents this antibody-mediated destruction of postsynaptic folds.

AChR loss is also not seen in aged BN rats subjected to passive transfer of EAMG (Graus et al., 1993). These animals, which have altered levels of several postsynaptic proteins, have a 25% increase in rapsyn : AChR ratio (Hoedemaekers et al., 1998). Our results suggest that AChR levels can be protected from antigenic modulation and complement-mediated lysis by overexpressing rapsyn alone. Interestingly, we found much less infiltrating cells at endplates in rapsyn-transfected muscles, which is another parallel to the age-related resistance in EAMG. However, AChR loss in passive transfer EAMG is not primarily determined by infiltrating macrophages (Hoedemaekers et al., 1997a).

One possible explanation for rapsyn-induced EAMG resistance is an increased rapsyn : AChR interaction. In cell culture studies, endogenous rapsyn and AChR are present in equimolar concentrations (Burden et al., 1983; LaRochelle and Froehner, 1986) but ∼50% of rapsyn and AChR are not localized in complexes (Marangi et al., 2001). The 1 : 1 stoichiometry of rapsyn and AChR in cell extracts probably does not reflect the interaction of rapsyn and AChR focally at the endplate (Tsui et al., 1990; Moransard et al., 2003), which is more pronounced and is increased by agrin (Moransard et al., 2003). Rapsyn overexpression could stabilize the pool of unclustered AChR by increasing the level of AChR-rapsyn complexes. These could serve as a pool for cluster formation or maintenance, reducing the susceptibility to antigenic modulation. In rat muscles, we found an average increase of 20% in the rapsyn : AChR ratio, with some individual endplates reaching a 100% increase of rapsyn : AChR ratio, as measured by semiquantitative immunohistochemistry (Fig. 2D). Thus, in the intact muscle, since each of the five AChR subunits can bind rapsyn (Maimone and Merlie, 1993), the increased ratio of rapsyn molecules to AChRs might increase the linkage of the AChRs to the cytoskeleton and reduce its susceptibility to antigenic modulation. Interestingly, the increased rapsyn was also associated with resistance to complement-mediated lysis, indicating that these mechanisms are interdependent in vivo as previously suggested (Lennon et al., 1978; Lennon and Lambert, 1981; Christadoss, 1988).

Another possibility is that EAMG damaged endplates cannot recover from on-going antibody-mediated destruction because rapsyn is not available in sufficient amounts. This seems likely since in passive transfer EAMG the mRNA levels of the AChR subunits are upregulated, while rapsyn expression remains constant (Asher et al., 1993), presumably because rapsyn gene expression in vivo is controlled independent of AChR gene expression (de Kerchove D'Exaerde et al., 2002; Rodova et al., 2004). Therefore, a relative lack of rapsyn could restrict the replacement of other endplate proteins in EAMG; these might include those that link newly produced AChR to the cytoskeleton, perhaps rendering the AChR more susceptible to complement-mediated lysis and antigenic modulation.

The results presented here suggest that even a relatively small upregulation of rapsyn expression in myasthenia gravis patients might be therapeutically helpful. The electroporation of selected muscles in humans is feasible, but probably too complicated using the current technical procedure. However, in the future improved vectors might become available for human gene therapy. Alternatively, future studies might investigate the possibility to increase the expression of rapsyn by modulating synaptic transcription factors such as Kaiso (Rodova et al., 2004). Moreover, one could speculate that differences in endogenous rapsyn expression might be a factor in determining susceptibility of different muscles to myasthenia gravis within an individual, and polymorphic variants might be related to disease severity between individuals. It will be interesting to look for rapsyn expression quantitatively at the human NMJs in different muscles, to determine what controls its level of expression, and to see whether polymorphisms in the promoter or coding regions modify disease severity in myasthenia gravis.


We are very grateful to Dr Christian Fuhrer (Brain Research Institute, University of Zürich, Switzerland) for the critical review of this manuscript, to Dr J. T. Vilquin (Institut National de la Santé et de la Recherche Médicale U-582, Institut de Myologie, Groupe Hospitalier Pitie-Salpetriere, Paris, France) for the helpful discussions concerning electropermeabilization and to Marc van Zandvoort, Wim Engels (Department of Biophysics, Universiteit Maastricht, The Netherlands) and Christoph Schmitz (Department of Psychiatry and Neuropsychology, Universiteit Maastricht, The Netherlands) for their help with two photon confocal microscopy and image deconvolution. We would also like to thank M. van de Waarenburg and H. P. J. Steinbusch for their excellent technical assistance. The BioRad TPLSM was obtained by grant no. 902-16-276 from the Medical Section of the Dutch Scientific Organization (NWO). This work was supported by grants from Prinses Beatrix Fonds and L'Association Française contre les Myopathies and a Marie Curie Fellowship of the European Community program ‘Quality of life and management of living resources’.


View Abstract