MuSK myasthenia gravis monoclonal antibodies: Valency dictates pathogenicity

Abstract

Objective: To isolate and characterize muscle-specific kinase (MuSK) monoclonal antibodies from patients with MuSK myasthenia gravis (MG) on a genetic and functional level.

Methods: We generated recombinant MuSK antibodies from patient-derived clonal MuSK-specific B cells and produced monovalent Fab fragments from them. Both the antibodies and Fab fragments were tested for their effects on neural agrin-induced MuSK phosphorylation and acetylcholine receptor (AChR) clustering in myotube cultures.

Results: The isolated MuSK monoclonal antibody sequences included IgG1, IgG3, and IgG4 that had undergone high levels of affinity maturation, consistent with antigenic selection. We confirmed their specificity for the MuSK Ig-like 1 domain and binding to neuromuscular junctions. Monovalent MuSK Fab, mimicking functionally monovalent MuSK MG patient Fab-arm exchanged serum IgG4, abolished agrin-induced MuSK phosphorylation and AChR clustering. Surprisingly, bivalent monospecific MuSK antibodies instead activated MuSK phosphorylation and partially induced AChR clustering, independent of agrin.

Conclusions: Patient-derived MuSK antibodies can act either as MuSK agonist or MuSK antagonist, depending on the number of MuSK binding sites. Functional monovalency, induced by Fab-arm exchange in patient serum, makes MuSK IgG4 antibodies pathogenic.

Figure 1. Patient-derived recombinant MuSK antibodies bind mouse NMJs

Control biotin antibodies did not stain the NMJ. Scale bar is 25 μm. MuSK = muscle-specific kinase; NMJ = neuromuscular junction.

Figure 2. Patient-derived recombinant MuSK antibodies can activate or inhibit MuSK phosphorylation and AChR clustering depending on the antibody valency

Bivalent monospecific recombinant MuSK antibodies (clones 11-3F6 and 13-3B5) activated MuSK phosphorylation in the presence and absence of agrin (A). Activation of MuSK phosphorylation was dose dependent (B). Clone 13-3B5 was slightly more potent compared with 11-3F6. A biotin control antibody did not affect (agrin dependent) MuSK phosphorylation. Monovalent Fab fragments generated from these recombinant MuSK monoclonals inhibited MuSK phosphorylation (C). Agrin-dependent AChR clustering was unaffected when exposed to a biotin control antibody or Fab fragments thereof. Bivalent monospecific recombinant IgG4 MuSK antibodies and Fab fragments inhibited the number of agrin-dependent AChR clustering significantly compared with the biotin control antibody (for IgG4: p < 0.001, for Fab fragments: p < 0.001, one-way analysis of variance (ANOVA) test, Dunnett corrected). Fab fragments reduced AChR clusters to the level of purified patient IgG4 and the “no agrin” condition. Bivalent monospecific antibodies significantly increased AChR clustering compared with Fab fragment–mediated inhibition that seems independent from agrin (D & E; 11-3F6: p < 0.001, 13-3B5: p = 0.03, biotin antibody: p = 0.30, unpaired t test, no agrin conditions 11-3F6: p = 0.03, 13-3B5: p = 0.20 one-way ANOVA Dunnett corrected). Data represent mean ± SEM. Scale bar represents 50 μm. AChR = acetylcholine receptor; MuSK = muscle-specific kinase.

Figure 3. Graphical depiction of the conclusions of this study