From seropositivity to pathogenicity: causal evidence for autoantibodies in myasthenia gravis

Autoantibody testing has expanded the serological spectrum of myasthenia gravis (MG), but detection of disease-associated immunoreactivity does not by itself establish direct pathogenicity. This Review examines how an MG-associated antibody moves from serological observation to mechanistic causality. We organize the evidence around three questions: whether the antibody signal is reproducible and biologically interpretable, whether circulating antibody can engage the native target and produce neuromuscular-junction dysfunction, and whether the resulting phenotype can be attributed to that antibody specificity. Acetylcholine receptor (AChR) and muscle-specific kinase (MuSK) antibodies provide mature causal benchmarks. Low-density lipoprotein receptor-related protein 4 (LRP4) illustrates established pathogenicity despite assay-dependent detection and an incompletely resolved balance between signaling interference, Fc-dependent effects, and local complement injury. Agrin has substantial biological and functional support, but independent sufficiency of human anti-agrin antibodies remains incompletely resolved. Cortactin provides a contrasting example in which repeated serological detection has not been linked convincingly to native target engagement or direct neuromuscular-junction dysfunction. ColQ, rapsyn, and collagen XIII remain unresolved for different reasons. Rather than assigning a numerical pathogenicity score, we identify the specific causal step that is missing for each antibody and match that uncertainty to the most informative experimental approach. This framework may improve interpretation of expanding antibody panels and help prioritize mechanistic studies that distinguish clinically useful biomarkers from directly pathogenic autoantibodies.