Are antibodies to neuronal surface proteins directly pathogenic?

Neuronal autoantibodies differ fundamentally in whether they can access and directly perturb their targets. The distinction between surface/synaptic and intracellular antigens is clinically meaningful, but pathogenicity remains target- and context-specific.

Clinical and functional evidence

• Surface-directed antibodies have strong direct pathogenic potential. Patient-derived anti-NMDAR IgG acutely reduces NMDAR channel function in primary-neuron electrophysiology; the effect requires intact bivalent IgG cross-linking rather than Fab fragments. This supports a direct antibody effect, not merely an association.
• Functional effects extend beyond NMDAR, although mechanisms vary. AMPAR antibodies reduce synaptic receptor abundance and AMPAR-mediated currents. LGI1 antibodies can disrupt LGI1–ADAM22/ADAM23 interactions, while CASPR2 antibodies can disrupt CASPR2–contactin-2 interactions and are associated with neuronal hyperexcitability.

Mechanistic interpretation

• Surface antibodies bind targets accessible on living neurons and may produce dysfunction through acute receptor modulation, receptor cross-linking and internalization, or disruption of synaptic protein complexes.
• For NMDAR, GluN1 cross-linking causes dose-dependent receptor internalization and reduced synaptic currents. AMPAR loss/internalization likewise reduces AMPAR-mediated currents.
• In contrast, classical intracellular/onconeural antibodies—including Hu and Yo—usually mark an antitumor immune response in which cytotoxic T cells are the principal mediators of neuronal injury, rather than acting as direct pathogenic antibodies.

Clinical implications

• Surface-antibody syndromes are often immunotherapy responsive, consistent with a potentially reversible antibody-mediated process. Intracellular-antibody syndromes generally have poorer immunotherapy responsiveness and should heighten concern for an associated malignancy.
• A positive antibody result alone does not establish causality. Interpretation should match the clinical syndrome and objective inflammatory findings; isolated or unexpected serum results warrant confirmation.

Bottom line

• Antibodies to neuronal surface proteins are substantially more plausibly directly pathogenic than antibodies to intracellular neural proteins, because they can directly alter receptor/channel function, receptor abundance, synaptic organization, or excitability; intracellular antibodies more often signify T-cell-driven, frequently cancer-associated autoimmunity.