α-Amino-3-Hydroxy-5-Methyl-4-Isoxazolepropionic Acid Receptor Plasticity Sustains Severe, Fatal Status Epilepticus

Abstract

Objective: Generalized convulsive status epilepticus is associated with high mortality. We tested whether α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor plasticity plays a role in sustaining seizures, seizure generalization, and mortality observed during focal onset status epilepticus. We also determined whether modified AMPA receptors generated during status epilepticus could be targeted with a drug.

Methods: Electrically induced status epilepticus was characterized by electroencephalogram and behavior in GluA1 knockout mice and in transgenic mice with selective knockdown of the GluA1 subunit in hippocampal principal neurons. Excitatory and inhibitory synaptic transmission in CA1 neurons was studied using patch clamp electrophysiology. The dose response of N,N,H,-trimethyl-5-([tricyclo(3.3.1.13,7)dec-1-ylmethyl]amino)-1-pentanaminiumbromide hydrobromide (IEM-1460), a calcium-permeable AMPA receptor antagonist, was determined.

Results: Global removal of the GluA1 subunit did not affect seizure susceptibility; however, it reduced susceptibility to status epilepticus. GluA1 subunit knockout also reduced mortality, severity, and duration of status epilepticus. Absence of the GluA1 subunit prevented enhancement of glutamatergic synaptic transmission associated with status epilepticus; however, γ-aminobutyric acidergic synaptic inhibition was compromised. Selective removal of the GluA1 subunit from hippocampal principal neurons also reduced mortality, severity, and duration of status epilepticus. IEM-1460 rapidly terminated status epilepticus in a dose-dependent manner.

Interpretation: AMPA receptor plasticity mediated by the GluA1 subunit plays a critical role in sustaining and amplifying seizure activity and contributes to mortality. Calcium-permeable AMPA receptors modified during status epilepticus can be inhibited to terminate status epilepticus. ANN NEUROL 2020;87:84-96.

FIGURE 1:

Status epilepticus (SE) was not fatal, was less intense, and had shorter duration in GluA1 knockout (KO) mice. (A) Survival during SE. All 26 GluA1-KO animals survived SE; 9 GluA1 wild-type (WT) animals died during SE, and 22 survived (**p = 0.0022, 1-sided Fisher exact test). (B) Heat maps illustrating the severity of behavior seizure scores in GluA1-WT and GluA1-KO animals during SE. Death is marked in pink. Rows signify individual animals, and 7 animals were randomly chosen to represent each group. The seizures were scored every alternate minute during stimulation and for the first hour of the self-sustained SE seizures. The red and black arrows mark the end of stimulation and the first hour of self-sustaining SE. (C) The median behavioral seizure scores (BSSs) along with 95% CI in the GluA1-WT and GluA1-KO animals illustrated in B, n = 7 each; ****p < 0.0001, 2-tailed Mann–Whitney test. (D, a) Spectrograms illustrating the power of electroencephalograms (EEGs) recorded from the hippocampus (HP) and cortex (CTX) during the initial 120 minutes of the self-sustaining phase of SE from a representative GluA1-WT and GluA1-KO animal. The scale bar represents power in square microvolts. Please note that SE duration was shorter in the GluA1-KO animal, and there was a substantial reduction in the power of EEG by 120 minutes. (D, b) A part of the spectrogram is magnified to illustrate the high-frequency spike-wave discharges marked by streaks of elevated power prominent in the GluA1-WT animal but not in the GluA1-KO animal. These events were generally associated with severe behavioral seizures (4–6). (E) Hippocampal and cortical EEGs illustrating a representative high-frequency spike-wave discharge, marked by red lines, in the GluA1-WT animal. Please note that in the GluA1-KO animals, the amplitude of discharges was reduced in the hippocampal recording during an event, and the event was not prominent in cortical EEGs. (F) The cumulative frequency distribution of the high-frequency (HF) spike-wave discharges in 12 randomly selected GluA1-WT and GluA1-KO animals. The data were binned every 10 minutes. *Approximate p = 0.027, Kolmogorov–Smirnov test. (G) Kaplan–Meier survival curves illustrating the duration of SE, which occurred earlier in the GluA1-KO animals than in GluA1-WT animals, n = 17 GluA1-WT and n = 12 GluA1-KO; *** p = 0.0006 Gehan–Breslow–Wilcoxon test.

FIGURE 2:

Seizure susceptibility in GluA1 knockout (KO) and wild-type (WT) animals was similar. (A) For the pentylenetetrazol (PTZ; 45mg/kg)-evoked seizure, latency was 210 ± 31 seconds in WT animals and 236 ± 32 seconds in KO animals (n = 10 WT and n = 9 KO animals, p = 0.59 unpaired t test). (B) PTZ-evoked seizure duration in the WT and KO animals; WT, 170 ± 35 seconds; KO, 217 ± 61 seconds; p = 0.55, Mann–Whitney test. The median behavioral seizure score (Racine scale) was 5 in WT animals and 4.5 in KO animals, p = 0.54, Mann–Whitney test (data not shown). (C) The intensity of current necessary to evoke a hippocampal seizure (after discharge threshold [ADT]) in WT (range = 40–200μA, median = 120, n = 19) and in KO animals (range = 20–160μA, median = 80, n = 20; p = 0.23, Mann–Whitney test). (D) A representative seizure recorded from an ipsilateral cortical electrode in response to 10 seconds of hippocampal stimulation in the WT and KO animals. The x and y scale bars represent 2 seconds and 1mV, respectively. (E) The duration of afterdischarges evoked in the WT and KO animals. The afterdischarge duration (ADD) in the WT animals was 7.0 to 37 seconds, median = 17 seconds, n = 19 and that in the KO animals was 6.7 to 49 seconds, median = 16.6 seconds, n = 20; p = 0.98, Mann–Whitney test. (F) The number of animals that experienced status epilepticus (SE; self-sustaining seizures lasting >5 minutes) following hippocampal stimulation in the 2 genotypes. SE was induced in 50% of WT (n = 8) and 0% of KO (n = 8) animals after 30 minutes of stimulation (p = 0.038, Fisher exact test) and in 68% of WT (n = 25) and 46% of KO (n = 26) animals after 60 minutes of stimulation (p = 0.098, Fisher exact test).

FIGURE 3:

The enhancement of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor–mediated synaptic transmission during status epilepticus (SE) was blocked, but the diminution in γ-aminobutyric acid receptor–mediated synaptic inhibition during SE occurred in GluA1 knockout (KO) animals. (A) Averaged spontaneous excitatory postsynaptic current (sEPSC) traces recorded from CA1 pyramidal neurons of a representative control GluA1 wild-type (WT) animal (black) and a WT animal in SE (red). (B) Averaged sEPSC traces from a control GluA1-KO animal and in a KO animal in SE. (C) The mean of the median amplitude and the average frequency of sEPSCs recorded from control and SE WT animals. The amplitude in WT controls was 16.4 ± 1.0pA (n = 8 cells, n = 4 animals), and that in WT SE animals was 19.7 ± 0.9pA (n = 8 cells, 5 animals; *p = 0.028, unpaired t test). The frequency in control WT animals was 0.54 ± 0.9Hz, and that in WT SE animals was 0.54 ± 0.07Hz (n is the same as for amplitude measurements; p = 0.98, unpaired t test). (D) The mean of the median amplitude and the average frequency of sEPSCs recorded from control and SE KO animals. The amplitude in KO control and KO SE animals was 16.2 ± 1.0pA (5 cells from 3 animals) and 14.6 ± 0.43pA (8 cells from 5 animals), respectively (p = 0.132, unpaired t test). The frequency in control and SE animals was 0.57 ± 0.05Hz and 0.46 ± 0.09Hz, respectively (n is the same as for amplitude measurements; p = 0.4, unpaired t test). (E) Averaged spontaneous inhibitory postsynaptic current (sIPSC) recorded from the CA1 pyramidal neurons of a control GluA1-WT animal (black) and a WT animal in SE (blue). (F) Averaged sIPSCs recorded from CA1 pyramidal neurons of control and SE GluA1-KO animals. (F) Mean of the median sIPSC amplitude in WT control (49.5 ± 3.3pA, 8 cells from 5 animals) and WT SE animals (36.6 ± 3.3pA, 7 cells from 5 animals; p = 0.017, unpaired t test) and the mean frequency of events in the control and SE animals (1.5 ± 0.25Hz and 1.8 0.33Hz, respectively; p = 0.49, unpaired t test). (G) The mean of the median amplitude and the mean frequency of sIPSCs recorded from GluA1-KO control and SE animals. (H) The mean sIPSC amplitude in the control and SE animals was 48.9 ± 4.3pA (8 cells from 5 animals) and 26.9 ± 2.7pA (6 cells from 5 animals), respectively (**p = 0.0018, unpaired t test). Mean sIPSC frequency in control and SE animals was 1.1 ± 0.12Hz and 0.86 ± 0.19Hz, respectively (p = 0.27, unpaired t test).

FIGURE 4:

Conditional deletion of the GluA1 subunit using adeno-associated virus (AAV)-expressing green fluorescent protein (GFP)-tagged CamKII-Cre. (A) Images illustrating GFP expression in the hippocampus of a GluA1-floxed animal injected with AAV serotype 9 expressing a GFP-tagged Cre recombinase under the control of the CamKII promoter. Immunoreactivity of NeuN (red) illustrates the neurons. Colocalization of GFP fluorescence corresponding to Cre expression with NeuN immunoreactivity labeled transduced neurons. The scale bars correspond to 400μm. (B) A Western blot illustrating the expression of the GluA1 subunit in the hippocampal tissue injected with AAV-expressing GFP-tagged CamKII-Cre. The GluA1 subunit expression in the contralateral hippocampus, which was not injected with the AAV, was used as a control. The expression of β-actin was used as a loading control. The graph shows GluA1 to β-actin expression ratio in the injected and uninjected hippocampi. The GluA1 subunit expression in the virus-injected hippocampus was 45 ± 11% of that in the contralateral hemisphere without virus injection (n = 3, *p = 0.0079, t test).

FIGURE 5:

Conditional deletion of the GluA1 subunit using adeno-associated virus (AAV)-expressing green fluorescent protein (GFP)-tagged CamKII-Cre was protective. (A) A comparison of floxed GluA1 animals injected with AAV9-expressing CamKII-GFP (transgenic [TG] control control) and CamKII-Cre-GFP (CamKII-Cre) that died during status epilepticus (SE); death occurred in 11 of the 22 TG control animals and in 1 of the 9 CamKII-Cre virus animals (p = 0.0496, 1-sided Fisher exact test). (B) The electroencephalographic (EEG) power spectrum shows death in an animal marked by an arrow. High-frequency discharge immediately preceded death. CTX = cortex; HP = hippocampus. (C) Heat maps illustrating behavioral seizure scores in the TG control and CamKII-Cre virus–injected floxed GluA1 animals. Each row represents 1 animal; 7 animals were randomly scored in each group. The time subsequent to the death of the animal is marked in pink and after the end of SE is marked in white. (D) Median and 95% confidence interval of the behavioral seizure score (BSS) in the animals illustrated in A (****p < 0.0001, 2-tailed Mann–Whitney test). (E) Spectrogram illustrating the power of EEGs recorded from the hippocampus and cortex during the first hour of SE from a TG control and CamKII-Cre virus–injected animal. (F) Hippocampal and cortical EEGs during SE from a TG control and a CamKII-Cre virus–injected animal. (G) Cumulative frequency distribution of high-frequency spike-wave discharges recorded from 7 representative TG control and CamKII-Cre virus–injected animals. The data are binned every 10 minutes (*p = 0.0207, Kolmogorov–Smirnov test). (H) A Kaplan–Meyer curve illustrating duration of SE in TG control (n = 14) and CamKII-Cre virus–injected (n = 7) animals (*p = 0.0405 Gehan–Breslow–Wilcoxon test).

FIGURE 6:

Blockade of calcium-permeable α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors with IEM-1460 terminated status epilepticus (SE). (A) Effect of IEM-1460 administered at 15 minutes after the end of hippocampal stimulation on the duration of SE. Log dose of IEM-1460 was plotted against the percentage of animals free of electrographic seizure activity within 60 minutes of the drug administration (n = 11 treated with 10mg/kg, n = 8 treated with 18mg/kg, and n = 12 treated with 30mg/kg IEM-1460). The line represents the best fit of the data to a 4-parametric equation for sigmoidal curve. These studies were performed in a blinded manner; the investigator reading the electroencephalogram (EEG) was blinded to the drug administered to the animals. (B) A heat map illustrating the behavioral seizure scores (BSSs) in C57BL/6 animals in continuous hippocampal stimulation (CHS)-induced SE treated with saline or IEM-1460 (30mg/kg, intraperitoneal) at 15 minutes after the end of hippocampal stimulation. (C) The seizure score and 95% confidence interval from animals depicted in B (****p < 0.0001, 2-tailed Mann–Whitney test). (D) Spectrograms illustrating EEG power during SE recorded from hippocampi of representative saline-treated and IEM-1460–treated animals. The white arrows mark the time of injection of saline or IEM-1460. The EEG power dropped in IEM-1460–treated animal. (E) Cumulative frequency distribution of high-frequency (HF) spike-wave discharges in the saline-treated and IEM-1460–treated animals. The treatments were performed at 15 minutes from the end of CHS (n = 7 each, ****p < 0.0001, Kolmogorov–Smirnov test).