LGI1 Autoantibodies Enhance Synaptic Transmission by Presynaptic Kv1 Loss and Increased Action Potential Broadening

Andreas Ritzau-Jost,Felix Gsell,Josefine Sell,Stefan Sachs,Jacqueline Montanaro,Toni Kirmann,Sebastian Maaß,Sarosh R Irani,Christian Werner,Christian Geis,Markus Sauer,Ryuichi Shigemoto,Stefan Hallermann

DOI: https://doi.org/10.1212/NXI.0000000000200284

Abstract:Background and objectives: Autoantibodies against the protein leucine-rich glioma inactivated 1 (LGI1) cause the most common subtype of autoimmune encephalitis with predominant involvement of the limbic system, associated with seizures and memory deficits. LGI1 and its receptor ADAM22 are part of a transsynaptic protein complex that includes several proteins involved in presynaptic neurotransmitter release and postsynaptic glutamate sensing. Autoantibodies against LGI1 increase excitatory synaptic strength, but studies that genetically disrupt the LGI1-ADAM22 complex report a reduction in postsynaptic glutamate receptor-mediated responses. Thus, the mechanisms underlying the increased synaptic strength induced by LGI1 autoantibodies remain elusive, and the contributions of presynaptic molecules to the LGI1-transsynaptic complex remain unclear. We therefore investigated the presynaptic mechanisms that mediate autoantibody-induced synaptic strengthening. Methods: We studied the effects of patient-derived purified polyclonal LGI1 autoantibodies on synaptic structure and function by combining direct patch-clamp recordings from presynaptic boutons and somata of hippocampal neurons with super-resolution light and electron microscopy of hippocampal cultures and brain slices. We also identified the protein domain mediating the presynaptic effect using domain-specific patient-derived monoclonal antibodies. Results: LGI1 autoantibodies dose-dependently increased short-term depression during high-frequency transmission, consistent with increased release probability. The increased neurotransmission was not related to presynaptic calcium channels because presynaptic Cav2.1 channel density, calcium current amplitude, and calcium channel gating were unaffected by LGI1 autoantibodies. By contrast, application of LGI1 autoantibodies homogeneously reduced Kv1.1 and Kv1.2 channel density on the surface of presynaptic boutons. Direct presynaptic patch-clamp recordings revealed that LGI1 autoantibodies cause a pronounced broadening of the presynaptic action potential. Domain-specific effects of LGI1 autoantibodies were analyzed at the neuronal soma. Somatic action potential broadening was induced by polyclonal LGI1 autoantibodies and patient-derived monoclonal autoantibodies targeting the epitempin domain, but not the leucin-rich repeat domain. Discussion: Our results indicate that LGI1 autoantibodies reduce the density of both Kv1.1 and Kv1.2 on presynaptic boutons, without actions on calcium channel density or function, thereby broadening the presynaptic action potential and increasing neurotransmitter release. This study provides a molecular explanation for the neuronal hyperactivity observed in patients with LGI1 autoantibodies.