A mutation causes MuSK reduced sensitivity to agrin and congenital myasthenia

Abstract

Congenital myasthenic syndromes (CMSs) are a heterogeneous group of genetic disorders affecting neuromuscular transmission. The agrin/muscle-specific kinase (MuSK) pathway is critical for proper development and maintenance of the neuromuscular junction (NMJ). We report here an Iranian patient in whom CMS was diagnosed since he presented with congenital and fluctuating bilateral symmetric ptosis, upward gaze palsy and slowly progressive muscle weakness leading to loss of ambulation. Genetic analysis of the patient revealed a homozygous missense mutation c.2503A>G in the coding sequence of MUSK leading to the p.Met835Val substitution. The mutation was inherited from the two parents who were heterozygous according to the notion of consanguinity. Immunocytochemical and electron microscopy studies of biopsied deltoid muscle showed dramatic changes in pre- and post-synaptic elements of the NMJs. These changes induced a process of denervation/reinnervation in native NMJs and the formation, by an adaptive mechanism, of newly formed and ectopic NMJs. Aberrant axonal outgrowth, decreased nerve terminal ramification and nodal axonal sprouting were also noted. In vivo electroporation of the mutated MuSK in a mouse model showed disorganized NMJs and aberrant axonal growth reproducing a phenotype similar to that observed in the patient's biopsy specimen. In vitro experiments showed that the mutation alters agrin-dependent acetylcholine receptor aggregation, causes a constitutive activation of MuSK and a decrease in its agrin- and Dok-7-dependent phosphorylation.

Figure 1. Identification of the homozygous c.2503A>G in MUSK, hereditary transmission and localization of the mutation in a three-dimensional model of the intracellular part of MuSK.

(A) Sequence chromatograms from a normal individual (control), affected proband IV: 1 and his parents (parents) are shown. (B) Pedigree of the Iranian family. The proband is indicated by an arrow. The expected nucleotide change c.2503A>G transmitted in this consanguineous Iranian family is indicated below the symbols when determined. (C) 3D model of the cytoplasmic kinase domain dimer of human MuSK built by homology modeling based on the structure of rat MuSK (PDB code: 1LUF) superimposed on the structure of the Insulin-like Growth factor 1 (PDB code 3D94), as proposed recently . The model is shown in cartoon representation; also represented in sticks are the catalytic residues (Tyrosines 751, 755 and 756, Aspartate 725 and Arginine 729) and residues whose mutations are associated with myasthenic syndromes A727/M605, V790 and M835). N-terminus, C-terminus and a loop not resolved in the structure are also indicated.

Figure 2. Morphological study of the patient’s biopsied muscle.

(A) Whole-mount preparations stained with α-bungarotoxin for nAChR in red and with an anti-neurofilament (NF) antibody for axons in green. In the normal NMJ, the axonal branch typically ends as a fork innervating a well-defined synaptic structure. In denervated NMJ, axons are absent. In some cases, axonal sprouts induce aggregation of nAChR forming ectopic junctions as a rosary. The histogram shows the classification of the NMJs observed in the patient muscle-biopsy specimen into four categories (expressed as percentage of the 25 NMJs examined). (B) In remodeled NMJs, nodal (arrow) or terminal (asterisk) axonal sprouts reinnervate synaptic gutters. The histogram shows the classification of the sprouting profiles observed into two categories (expressed as a percentage of the 16 NMJs exhibiting sprouts identified amongst the 25 NMJs analyzed). Scale bar = 10 µm and applies to all prints in A and B.

Figure 3. Electron microscopy in control and patient.

(A) Control endplate Patient’s endplates: (B) reinnervated endplate shows decreased and enlarged post-synaptic folds, a vanished basal lamina, and absence of active zones. (C) Ectopic NMJ is characterized by the total absence of subneural folds indicating that they are at an early stage of synaptogenesis. (D) An axon with a smooth endoplasmic reticulum (SER) network, a sign of impaired axonal flow. Zooms show aggregation of unidentified proteins on SER networks. Scale bars represent 500 nm for the four low-magnification prints and 3 µm for both zooms.

Figure 4. Immunocytochemistry of MuSK on cryostat sections.

In the control, nAChR and MuSK are strongly expressed and perfectly colocalized. In the patient, nAChR expression is reduced whereas MuSK expression remains normal and extends slightly in the perisynaptic area. The scale bar represents 10 µm and applies to the six prints.

Figure 5. Quantification of nAChRs in patient’s biopsied muscle.

(A) nAChR expression in MuSK mutated muscle from the patient was decreased by about 50% compared with biopsied control muscle. (B) The nAChR γ-subunit mRNA level, which is synonymous with denervation, was measured in the extrasynaptic zone in patient and was increased by 5-fold compared with the control. The mean value of control AChR (A) or AChR γ-subunit mRNA levels were defined as 1 in arbitrary units. Error bars indicate mean±SD (**, P<0.01, Student t test).

Figure 6. Morphological study of electroporated mice.

(A). Denervated junctions were fragmented and dispersed without terminal axons. Remodeled junctions were partially reinnervated and showed nodal sprouts forming more or less mature ectopic junctions (B). The arrow shows a mature ectopic junction with well-defined borders synonymous with subneural fold formation. The asterisk and the zoom show the first steps of ectopic junction formation characterized by very simplified AChR clusters induced by emergent nodal sprouts or axonal varicosity. Histograms show the distribution of different types of NMJs and sprouting in MV MuSK electroporated mice. The scale bar represents 10 µm and applies to all prints except to the blowup where the scale bar represents 2.5 µm.

Figure 7. Effect of the MV MuSK mutation on AChR aggregation.

(A) MuSK −/− cells failed to aggregate nAChR with or with agrin treatment. We observed some aneural nAChR clusters in both WT and MV MuSK cells. After agrin treatment, nAChR clusters of more than 3 µm length appeared in WT MuSK cells while they remained almost undetectable in MV MuSK cells. Scale bar = 20 µm. (B) Histogram showing the number of nAChR aggregates of more than 3 µm per myotube that crossed the field from 3 independent experiments (20 fields in each Petri dish were randomly counted for each set of experimental conditions). Error bars indicate mean±SD (**, P<0.01, Student t test).

Figure 8. Effect of the mutation on agrin-dependent MuSK phosphorylation.

HEK 293 cells expressing both Lrp4 and MuSK were incubated with recombinant rat agrin. (A) HA-MuSK was immunoprecipitated (IP) with an anti-HA antibody (α HA) and immunoblots (IB) for MuSK (α HA) and phosphorylated MuSK (α pTyr) were performed. A representative experiment out of three realized in duplicate is shown. MuSK amount and phosphorylation were then estimated with ImageJ software. (B) Phosphorylation is expressed as n-fold activation as compared to WT MuSK without Lrp4 and agrin. Error bars indicate mean±SD (n = 3; **, P<0.01, paired t test). Comparison between MV and WT MuSK shows that while WT MuSK needs the presence of both agrin and Lrp4 to be phosphorylated, MV MuSK is already phosphorylated even in the absence of Lrp4 and agrin.

Figure 9. Effect of the use of a dead-kinase MuSK on MV and WT MuSK phosphorylation.

HEK 293 cells expressing both Lrp4 and MuSK were incubated with recombinant rat agrin. HA-MuSK was immunoprecipitated (IP) with an anti-HA antibody (α HA) and immunoblots (IB) for MuSK (α HA) and phosphorylated MuSK (α pTyr) were performed. A representative experiment out of three realized in duplicate is shown. As expected the introduction of the dead-kinase mutation in WT MuSK (WT/KA) abolishes the phosphorylation observed in the presence of both Lrp4 and agrin. The introduction of the kinase-dead mutation together with the MV mutation (MV/KA) abolishes the spontaneous phosphorylation of MV MuSK (MV) seen in the absence of Lrp4 and agrin.

Figure 10. Effects of the mutations on Dok-7-dependent MuSK phosphorylation and its binding to Dok-7.

HA-MuSK (WT, VM, or MV) and/or WT-Dok-7 proteins were exogenously expressed in 293T cells and the whole cell lysates (WCLs) were subjected to immunoprecipitation (IP) of MuSK proteins with anti-HA rabbit polyclonal antibody (α HA pAb) and immunoblotting (IB) with anti-HA rat monoclonal antibody (α HA mAb), anti-Dok-7 antibody (α Dok-7), or anti-phosphotyrosine antibody (α pTyr). (A) Representative data from 3 independent experiments are shown. (B) Ratios of the tyrosine phosphorylation of MuSK to total amount of MuSK were quantified and shown. The mean value of Dok-7-dependent MuSK (WT) phosphorylation levels was defined as 1 in arbitrary units. Error bars indicate mean±SD (n = 3; *, P<0.05; **, P<0.01, paired t test).