In Vivo Injection of Anti-LGI1 Antibodies into the Rodent M1 Cortex and Hippocampus Is Ineffective in Inducing Seizures

Abstract

Autoimmune encephalitis (AIE) associated with antibodies directed against the leucine-rich glioma inactivated 1 (LGI1) protein is the second most common AIE and is responsible for deleterious neocortical and limbic epileptic seizures. Previous studies demonstrated a pathogenic role of anti-LGI1 antibodies via alterations in the expression and function of Kv1 channels and AMPA receptors. However, the causal link between antibodies and epileptic seizures has never been demonstrated. Here, we attempted to determine the role of human anti-LGI1 autoantibodies in the genesis of seizures by analyzing the impact of their intracerebral injection in rodents. Acute and chronic injections were performed in rats and mice in the hippocampus and primary motor cortex, the two main brain regions affected by the disease. Acute infusion of CSF or serum IgG of anti-LGI1 AIE patients did not lead to the emergence of epileptic activities, as assessed by multisite electrophysiological recordings over a 10 h period after injection. A chronic 14 d injection, coupled with continuous video-EEG monitoring, was not more effective. Overall, these results demonstrate that acute and chronic injections of CSF or purified IgG from LGI1 patients are not able to generate epileptic activity by themselves in the different animal models tested.

Keywords: LGI1; autoimmune encephalitis; electrophysiology; epilepsy; video-EEG.

Figure 1.

Acute injections of CSF and purified serum IgG. A, Schematic of the experimental setup for acute injection and simultaneous electrophysiological recording from the sedated rat. B, Control experiments for the location of the injection sites. Coronal brain slices at the indicated coordinates, showing the hippocampal (top) and M1 cortical (bottom) injection sites, as revealed by Fluoro-Ruby injections (pink), at the exact same volume and coordinates as the control and anti-LGI1 antibody-containing CSF and purified serum injected. Blue labeling is DAPI labeling of cell nuclei. C, ECoG activities and local field potentials recorded before (left) and after (right) an acute injection of purified IgG from LGI1 patients into the M1 cortex. ECoG activity was collected from the left M1 (M1L), right M1 (M1R), and the left APC (APCL). A multichannel electrode was inserted in M1L, 200 µm anterior to the injection site to record LFPs from the different cortical layers. Note the absence of epileptiform activity or seizures on the postinjection recordings. D, Frequency power (mean ± SD) of intracerebral LFP recordings after injection in M1 (top) and hippocampus (bottom), in control (blue) and LGI1 (orange) conditions. Electrophysiological signals were analyzed before injection (left), 2 h after injection (middle), and 5 h after injection (right). The electrode closest to the injection site was selected for the analysis. No significant difference was found between control (M1, n = 8; hippocampus, n = 8) and LGI1 (M1, n = 14; hippocampus, n = 12) experiments (two-tailed Mann–Whitney rank-sum test for each time period, on data binned in frequency bands of 1 Hz width). The statistical power estimated for each frequency bin was on average 0.55 ± 0.15. The frequency content of LFP activities is displayed separately between CSF and serum experiments in the Extended Data Figure 1-1. Stereotaxic coordinates used for the acute experiments are reported in the Extended Data Table 1-1.

Figure 2.

Chronic injections of CSF and purified serum IgG. A, Experimental design. Rats and mice were implanted with an injection cannula in the left hippocampus, in addition to left M1 (M1L) and right M1 (M1R) and bilateral hippocampus electrodes for long-term EEG monitoring of brain activity. The postinjection days of recording are indicated on the schema. B, Representative staining of human IgG on coronal section of a mouse after 7 d of unilateral injection with patient-derived LGI1 serum antibodies to demonstrate the distribution in the hippocampus. Slices were incubated with peroxidase-coupled anti-human IgG followed by DAB staining. C, Typical EEG activity recorded in the different states of vigilance in controls (n = 10 mice and 6 rats) and animals injected with LGI1 antibodies (n = 10 mice; n = 6 rats). Note the absence of epileptiform activity or seizures. Examples of recordings of control and LGI1 awake mice at different time points are shown in Extended Data Figure 2-1. Stereotaxic coordinates used for the chronic injection experiments are reported in Extended Data Table 2-1.