Neocortical injury–induced status epilepticus

Abstract

Objective:

To characterize neocortical onset status epilepticus (SE) in the C57BL/6J mouse.

Methods:

We induced SE by administering homocysteine 16–18 hours after cobalt (Co) implantation. SE was monitored by video and electroencephalography (EEG). We evaluated brain structure with magnetic resonance imaging (MRI). Neurodegeneration was evaluated 72 hours after SE using Fluoro-Jade C staining.

Results:

Cobalt triggered seizures in a dose-dependent manner (median effective dose, ED₅₀ = 0.78 mg) and the latency to peak seizure frequency shortened with increased dose. Animals developed SE after homocysteine administration. SE began with early intermittent focal seizures, consisting of frontal onset rhythmic spike-wave discharges manifested as focal dystonia with clonus. These focal seizures then evolved into generalized continuous convulsive activity. Behavioral manifestations of SE included tonic stiffening, bilateral limb clonus, and bilateral tonic-clonic movements, which were accompanied by generalized rhythmic spike-wave discharges on EEG. After prolonged seizures, animals became comatose with intermittent bilateral myoclonic seizures or jerks. During this period, EEG showed seizures interspersed with generalized periodic discharges on a suppressed background. MRI obtained when animals were in a coma revealed edema, midline shift in frontal lobe around the Co implantation site, and ventricular effacement. Fluoro-Jade C staining revealed neurodegeneration in the cortex, amygdala, and thalamus.

Significance:

We have developed a mouse model of severe, refractory cortical-onset SE, consisting of convulsions merging into a coma, EEG patterns of cortical seizures, and injury, with evidence of widespread neocortical edema and damage. This model replicates many features of acute seizures and SE resulting from traumatic brain injury, subarachnoid, and lobar hemorrhage.

1 |. INTRODUCTION

Status epilepticus (SE) consists of abnormally prolonged seizures associated with neuronal death, neuronal injury, and altered neuronal networks.¹ There are approximately 126 000 to 195 000 episodes of SE reported every year in the United States. The 30-day all-cause mortality after SE ranges from 10% to 27%.² It has become clear over the past two decades that acute seizures induced by cortical injury associated with trauma, subarachnoid, and lobar hemorrhage, worsen the prognosis of these conditions. These acute neurological insults cause seizures and SE, which then cause secondary injury. There is consensus among neuro-intensivists and neurophysiologists that these injury-induced seizures should be detected and treated.^3–8 However, the mechanism of injury-induced SE and appropriate therapies are unknown.

The pathophysiology underlying acute seizures following cortical injury is poorly understood, in part due to a lack of suitable animal models. An experimental model should recapitulate features of acute seizures or SE following the cortical insult. Seizures often begin with clear clinical manifestations, but as patients become comatose, the seizures become nonconvulsive.^9,10 In these comatose patients, there is often edema and evidence of blood-brain barrier breakdown near the site of injury.¹¹ Electroencephalography (EEG) is necessary to detect these nonconvulsive seizures. On EEG, seizures punctuate periodic discharges and burst suppression patterns in these patients. An experimental model should recapitulate features of acute seizures or SE following the cortical insult.

Status epilepticus occurred in rats with experimental cobalt (Co) focus in the cortex when they were given homocysteine.¹² This study focused on drug therapy in this form of experimental SE. EEG patterns of periodic discharges and burst suppression were present, suggesting cortical injury. We adapted this neocortical injury–induced SE model to mice and characterized behavior, EEG, magnetic resonance imaging (MRI), and neurodegeneration.

2 |. MATERIALS AND METHODS

2.1 |. Animals

All studies were performed according to protocols approved by the Animal Care and Use Committee of the University of Virginia. Adult C57BL/6J mice of either sex (23–27 g, 8–10 weeks old) were used for these studies; four to five mice were kept in a cage, mice had ad libitium access to food and water, and they were maintained on a 12-hour light/dark cycle.

2.2 |. Co-implantation

This animal cohort (1) was stereotactically implanted with increasing doses of Co (0.44–2.11 mg) (Alfa Aesar) orthogonally into the supplementary motor (M2) area (AP, +2.6 mm; ML, +1.8 mm) of the brain under isoflurane anesthesia. Co was implanted at this location for all cohorts. Co wires with increasing length of two different diameters were used (0.44 mg [0.25 mm diameter, 1 mm length]; 0.66 mg [0.25 mm diameter, 1.5 mm length]; 0.88 mg [0.25 mm diameter, 2.0 mm length; 1.76 mg [0.5 mm diameter, 1 mm length]; and 2.11 mg [0.5 mm diameter, 1.2 mm length]).

2.3 |. EEG electrode implantation

During Co implantation, stainless-steel electrodes were also implanted in the ipsilateral frontal (Fi) (to Co) and contralateral frontal (Fc) areas (AP; +1.6 mm, ML, ±1.8 mm and DV +1.0 mm). The bilateral parietal EEG recording electrodes, ipsilateral parietal (Pi), and contralateral parietal (Pc) were implanted at the following coordinates: AP, −2.6 mm; ML, ±1.8 mm; DV +1.0 mm with a cerebellar reference electrode as described earlier.¹³ All electrodes were secured with dental cement. Animals were given ketoprofen (1 mg/kg, Intraperitoneal [i.p.]) to minimize pain and discomfort. Continuous video and EEG monitoring (Grass ARUA LTM64 using Twin software, Grass) was initiated 30 minutes after the surgery and continued for 7 days to understand seizure semiology associated with Co implantation.

2.4 |. Immunohistochemistry and Fluoro-Jade C (FJC) staining following Co implantation

The second animal cohort was used to study the neurodegenerative effect of a single Co dose (2.1 mg) on the brain. Twenty-four hours following Co implantation, animals were transcardially perfused and brains were harvested and processed as described earlier.¹⁴ To prevent any artifact, electrodes were not implanted. Every fourth 40-μm-thick coronal section was processed for NeuN immunohistochemistry (IHC) to observe lesions (anti-NeuN antibodies; 1:500, MAB377, Millipore). The injury was seen qualitatively in images acquired on a fluorescence microscope (Nikon Eclipse Ti-U at 2× magnification, 0.45 NA [Numerical Aperture]).

Every fourth, 40-μm-thick coronal sections in a separate set of serial sections from the same cohort were also processed for FJC-staining (Histo-Chem, Inc) to observe neurodegeneration as described previously.¹⁴ FJC-positive cells in images were counted corresponding to the same stereotactic coordinates shown for lesion using a fluorescence microscope (Nikon Eclipse Ti-U at 20× magnification, 0.45 NA). NIS element acquisition software was used for image acquisition and Imaris version 9.3.1 was used for image processing (Bitplane Imaris, Oxford Instruments). FJC-positive cells were counted manually by two independent observers (one unblinded and one blinded) in the region for each section at 20× magnification in Image J software (Image J, National Institutes of Health [NIH]).

2.5 |. SE induction

This third animal cohort was used to study the development of SE. Homocysteine dose (845 mg/kg, i.p.), as reported previously in rats¹² was injected at different time intervals; 16–18 hours, 1 day, 3 days, 5 days, and 7 days following Co (1.7 mg) implantation. Along with Co implantation, four cortical and a cerebral reference electrode was implanted as described earlier in Sections 2.2 and 2.3.

We increased the Co (2.1 mg) dose to further optimize and characterize the model (Cohort 4). Following Co and EEG electrode implantation, homocysteine (845 mg/kg i.p.) was injected 16–18 hours following Co implantation. SE was defined as intermittent or continuous electrographic seizure activity lasting longer than 5 minutes. Distinct EEG patterns such as spike-wave discharges and burst suppression were identified using the American Clinical Neurophysiology Society (ACNS) guidelines. EEG data were exported to Lab Chart 7 (ADI Instruments), and power spectrograms and power graphs were generated. A referential cortical EEG electrode in the ipsilateral frontal cortex (Fi) was used to generate a power spectrum for the total duration of SE. The analysis consisted of a fast Fourier transform using a cosine-bell data window with a window size of 1024 data points (2.56 seconds). This method resulted in a frequency resolution of 0.375 Hz. A window overlap of 50% was used to help smooth the x-axis of the spectrogram. Power was expressed as mV².

2.6 |. Magnetic resonance imaging

MRI was performed on a 7.0T MR system (Clinscan, Bruker Biospin) using a 4-channel phased-array radiofrequency coil and an MR-compatible physiological monitoring and gating system for mice (SA Instruments, Inc). The maximum gradient strength of the system was 500 mT/m, and the peak slew rate achievable was 6667 mT/m/ms Scanning protocols included standard T2-weighted imaging fast spin-echo sequence with an acquisition time of 20 minutes) with voxel size = 60 × 45 × 500 μm, field of view = 2.3 cm, resolution 512 × 384, and slice thickness = 0.5 mm. Because Co is a ferromagnetic metal, we removed it before the initiation of scans. Images were obtained from animals under continuous isoflurane anesthesia (1%−1.5%), and breathing was monitored continuously (Small Animal Instruments, Inc).

This cohort (5) was Co implanted, and animals were divided randomly into two subsets for MRI; Co control group consisted of Co implanted animals injected with saline (Co) and Co-homocysteine group consisted of animals who had SE and were studied 240 minutes after homocysteine injection. The SE animals were scanned during the coma phase, which was 240 minutes after the injection of homocysteine. MRI scans were obtained in both coronal and horizontal orientation in each animal. In each MRI image, the area of the T2 signal was measured by two independent observers (one unblinded and another blinded) using ImageJ software (ImageJ, NIH). The hemispheres were outlined manually; small, noisy clusters with bright signals outside the brain were removed (ImageJ, NIH); and a threshold was applied. The brain regions with an increased signal were delineated using Franklin and Paxinos (FP) atlas.¹⁵

2.7 |. Fluoro-Jade C staining 72 hours following SE

This animal cohort (6) was used to study neuronal damage associated with SE using FJC-staining. Seventy-two hours after SE, the animals were perfused transcardially, and brains were harvested as described earlier. Every fourth, 40-μm-thick coronal section was processed for FJC-staining, as described earlier.¹⁴ Images of different regions from three coronal sections were taken at 10× magnification using the NIS Elements (Nikon), from the bregma between −1.91 and −2.15 mm. FJC-positive cells were counted manually in each region for each section by outlining the region of interest (ROI) at 10× magnification in Image J software (Image J, NIH). The sections were scanned on a confocal microscope (Nikon Eclipse Ti-U at 10× magnification, 0.45 NA). The excitation laser was 488 nm. Images were tiled as stacks with optical section separation (Z interval) of 10 μm and stitched using NIS Elements software. Imaris 9.3.0 (Bitplane) was used for visualization, and Adobe Photoshop CC was used for cropping the original image and figure display.

2.8 |. Data analysis and statistics

The EEG data and statistical analysis were processed using Graph Pad Prism version 8. Normally distributed data for the T2 signal in MRI is presented as the mean ± standard error of the mean (SEM). The duration of SE and mortality after SE was plotted for survival analysis using the Kaplan-Meier test. The data for the dose-response for Co was fitted to an equation for the sigmoidal curve. The best fit to the equation determines the median effective dose (ED₅₀). Mean ± SEM was used to describe normally distributed data for power graph and FJC-positive cells. Fisher exact test was used to compare the fraction of animals that developed SE following homocysteine injection at different time points. The values were considered significant at P < .05.

3 |. RESULTS

3.1 |. Co caused seizures

The study from cohort 1 demonstrated the effect of increasing Co dose on seizures. All Co doses caused spikes and polyspike-wave discharges, associated with contralateral clonic forelimb jerks, which appeared within 4 hours after the Co implantation. Animals primarily developed two seizure types. (a) Focal motor seizures manifested as focal dystonia with unilateral clonus, which electrographically were associated with rhythmic spike-wave discharges in frontal electrodes (Fi and Fc) (Figure 1B,C). (b) Focal motor seizures with secondary generalization; behaviorally, these seizures manifested as focal dystonia with unilateral clonus and evolving into bilateral forelimb clonus. Electrographic seizure activity became generalized, and spike-wave discharges were present in the frontal and parietal electrodes (Pi and Pc) (Figure 1D,E).

FIGURE 1.

Electroencephalography (EEG) and seizure semiology observed after cobalt (Co) implantation. A, The total power of EEG recorded from the ipsilateral frontal (Fi) electrode during peak seizure frequency was plotted against time. The power is in the range of 0–40 Hz. The scale bar represents power in μV². B and C, EEG traces recorded from four cortical electrodes: ipsilateral frontal (Fi), contralateral frontal (Fc), ipsilateral parietal (Pi), and contralateral parietal (Pc). Traces illustrate that electrographic activity associated with focal motor seizures was manifest as focal dystonia with clonus. Electrographic seizure activity was recorded from frontal electrodes only. D and E, Electrographic seizure activity during seizures that were manifested as focal dystonia with clonus with secondary generalization. F, The number of seizures recorded during 4-hour bins for each of the Co doses. The doses of Co used were 0.44 (n = 8), 0.66 (n = 13), 0.88 (n = 8), 1.76 (n = 12), and 2.11(n = 7) mg. Values are expressed as a mean ± standard error of the mean (SEM). G, The dose-response curve of Co dose and number of seizures. The line represents nonlinear four parametric fits of the data and represents median effective dose (ED₅₀). All data in this figure are from cohort 1

We determined peak seizure frequency for each dose over time by binning them over 4-hour intervals. Lower Co doses (0.44 and 0.66 mg), caused few seizures. However, increasing Co doses (0.88–2.11 mg) increased the number of seizures, and reached a peak at different time points, 32 hours for 0.88 mg, 24 hours for 1.76 mg, and 12 hours for 2.11 mg. The peak seizure frequency did not increase with higher doses (0.88–2.11 mg), but it shifted closer to the time of Co implantation (Figure 1F). The total number of seizures induced in response to increasing doses of Co reached a plateau at 2.11 mg dose, with latency to peak seizure frequency at ~12 hours. The animals implanted with the lowest dose of Co (0.44 mg) experienced 4.0 ± 0.6 seizures (n = 8). This number increased to 6.3 ± 1.1 with 0. 66 mg Co (n = 13), to 10.0 ± 1.8 with 0.88 mg Co (n = 8), to 13.9 ± 2.2 with 1.76 mg Co (n = 12), and to 14.2 ± 2.3 with 2.1 mg Co (n = 7). The dose-response data were fit to the sigmoid curve (4 parameter equation) to determine the ED₅₀ and maximal effect (Figure 1G). Fit confirmed that the seizure number was unlikely to increase with a further dose elevation. However, to confirm, we increased the dose to 2.64 mg, but this dose proved to be lethal, and animals died within 12 hours.

3.2 |. Early pathological effects of Co

Co caused a significant cortical lesion (cohort 2), which was evident in multiple NeuN-stained sections. The lesion was observed in brain sections spanning from bregma AP +2.6 to +1.0, and they encompassed all cortical layers (Figure 2A). Separate serial sections, 160 μm apart, were stained for FJC and NeuN staining. Co also caused neurodegeneration (FJC-positive cells), which diminished as a function of distance from the Co insertion site. FJC-positive cells spanned from bregma +2.6 to +0.5 and were observed primarily in primary and secondary motor, orbital, prelimbic, anterior cingulate, and primary somatosensory areas of the neocortex. Unblinded and blinded analysis by two independent observers showed that there were more FJC-positive cells closer to Co (Figure 2C,D). There were no FJC-positive cells in the contralateral motor cortex or any other part of the brain (n = 5).

FIGURE 2.

Histological effect of cobalt (Co) implantation. A, Serial coronal sections are illustrating Co lesion from anterior to posterior 24 h post-Co (2.1 mg) implantation (arranged from left to right). Images are taken at 2× magnification and scale bar = 300μm. B, Fluoro-Jade C (FJC) staining in the ipsilateral frontal cortex corresponding to the lesion, 24 h after the Co implantation. Please note extensive FJC-positive cells in the sections encompassing the lesion and sparse labeling posterior to the lesion (arranged from left to right) (n = 5). Image magnification was 20× and Scale bar = 300 μm. C and D, The number of FJC-positive cells was counted in representative microscopic fields (20× magnification) in different regions of the coronal brain section encompassing lesion by an (C) unblinded and (D) blinded observer (n = 5). All data in this figure are from cohort 2

3.3 |. Development of SE in Co implanted animals

Walton and Treiman reported that homocysteine (5.5 mmol/kg or 845 mg/kg, i.p.) injected 9.0 ± 2.5 days following Co implantation caused SE in rats.¹² Therefore, in cohort 3, mice received homocysteine (845 mg/kg, i.p.) at different time intervals: at 16–18 hours (n = 5), 1 day (n = 6), 3 days (n = 6), 5 days (n = 5), or 7 days (n = 4) following Co (1.7 mg) implantation to determine the optimal time for inducing SE. A larger fraction of Co-implanted animals developed SE when they received homocysteine after 16–18 hours (2/5) compared to those who received it 1 to 7 days later (0/21), Fisher exact test, P < .05. Mice that received homocysteine 1 day, 3 days, 5 days, and 7 days following Co implantation had intermittent focal seizures. Homocysteine was also injected in steel wire (same dimension as Co) implanted animals, which also caused a single discrete seizure in 50% of animals (n = 4). The seizure latency was ~30 minutes after homocysteine injection and caused no mortality. In summary, Co (1.7 mg) implantation followed by homocysteine injection at 16–18 hours post-implantation caused SE in 40% of animals.

3.4 |. Optimization and characterization of model

To further optimize the model, Co dose was increased to 2.11 mg, followed by homocysteine (845 mg/kg i.p.) injection after 16–18 hours, which caused SE in 7 of 10 animals. Five male and five female mice constituted this cohort, and there were no significant differences between sexes for a fraction of animals developing SE, latency to SE, and duration. Three male animals and four female animals developed SE with a similar latency (13.0 ± 3.7 and 10.6 ± 3.8 minutes, respectively), and it lasted 122.6 ± 17.8 in male and 105.7 ± 2.59 minutes in female animals. Below we present behavior and EEG characteristics observed in this cohort of male and female animals. All further studies to evaluate edema using MRI or neurodegeneration using FJC-staining used this protocol for induction of SE.

3.5 |. Behavior and EEG characteristics of SE

We describe seizures using a behavior seizure scoring scale. (Stage 1) Focal unilateral forepaw clonus: unilateral forepaw clonus, with no tonic stiffening or tonic-clonic movements; (Stage 2) Focal dystonia and clonus: the body of the animal twists or bends sideways in association with unilateral forepaw involuntary repetitive jerking with subtle bilateral jerking; (Stage 3) Tonic stiffening and head-bobbing: whole-body tonic stiffening with head-bobbing occurs but no forelimb and hind-limb movement; (Stage 4) Bilateral clonus with Straub-like tail: bilateral symmetrical forelimb clonus with little or no rearing with Straub-like tail; and (Stage 5) Generalized tonic-clonic seizures with loss of posture or wild jumping or running.

However, to describe the evolution of EEG, we divided SE into three phases: (a) early, intermittent focal seizures; (b) continuous generalized convulsive seizures, and (c) periodic discharges and coma phase.

3.5.1 |. Early, intermittent focal seizures

Status epilepticus started with two to four episodes of focal motor seizures with secondary generalization, observed as focal dystonia with clonus evolving to bilateral clonus (Figure 3B). EEG activity recorded from the ipsilateral frontal electrode consisted of high-frequency discharges primarily in the theta, alpha, and beta frequencies, followed by rhythmic spike-wave discharges in the theta frequency. In contrast, EEG recorded from the contralateral frontal electrode consisted of high-amplitude spike-wave discharges, appearing at the same time as the ipsilateral frontal electrode. Seizures consisting of rhythmic, spike-wave discharges appeared in the ipsilateral parietal and contralateral parietal electrodes with a lag of 4–8 seconds (Figure 3C). However, in-between these seizure episodes, there were intermittent single spikes followed by high-frequency, low-amplitude polyspikes in the frontal electrodes. During this phase, the first seizure after homocysteine injection lasted 25–40 seconds. Subsequent seizures were 60–180 seconds long and eventually they merged into generalized convulsions.

FIGURE 3.

Electroencephalography (EEG) patterns marking early focal intermittent seizures, continuous generalized convulsive seizures, and coma phase during status epilepticus (SE). EEG and associated behavior patterns during cobalt (Co)-homocysteine-induced SE. A, The total power of EEG recorded from the ipsilateral frontal (Fi) electrode was plotted against time. The power is in the range of 0–40 Hz. The scale bar represents power in μV². Time 0 corresponds homocysteine injection. B-E, EEG traces recorded from four cortical electrodes—ipsilateral frontal (Fi), contralateral frontal (Fc), ipsilateral parietal (Pi), and contralateral parietal (Pc)—illustrating activity associated with behaviors corresponding to the early focal intermittent seizure phase. (B) Unilateral forepaw clonus (stage1) and (C) focal dystonia with clonus (stage 2), continuous generalized convulsive seizure phase, that is, (D) tonic stiffening (stage 3) and bilateral clonus (stage 4), and (E) generalized tonic-clonic seizure with loss of posture (stage 5). Please note the changes in EEG recorded from frontal electrode (F-I) EEG patterns observed during the coma and burst suppression phase. F, periodic epileptiform discharges, G, periodic epileptiform discharges with burst patterns (H) burst suppression patterns, and (I) complete suppression of electrographic activity, which marked the end of SE (n = 7). J, Kaplan-Meier curve illustrating the duration of SE. Percentage of animals in SE and 95% confidence intervals plotted against time (n = 7). The total duration of the SE was divided into 10-min slots. The event (end of SE) was plotted slot-wise in a Kaplan-Meier curve. K, The pie chart shows the percentage of time spent by the animals in each stage of behavior observed during SE. The numbers in the wedges represent the average percentage of time spent in the respective phase (n = 5). L, Kaplan-Meier curve illustrating the survival of animals after SE. Percentage of animals survived and 95% confidence intervals were plotted against time (n = 7). The total duration after SE was divided into a 1 h slot. The event (death of the animal) was plotted slot-wise in the Kaplan-Meier curve. All data in this figure are from cohort 4

Power spectrum analysis of EEG recorded from the ipsilateral frontal electrode depicted a sudden burst of high-power activity associated with early, intermittent focal seizures. The first electrographic seizure, which marked the beginning of SE, caused high-power activity in the theta range frequency represented by the red color in the spectrogram and power graphs (Figures 3A and 4B). The end phase of each of these early seizures was characterized by high-power activity in the delta and theta range, as seen with intense red and yellow color. However, overall, during these seizures, the increase in power primarily occurred in the delta, theta, alpha, and beta frequencies (<20 Hz) (Figures 3A and 4A–D). In-between these seizures, power decreased in the <5 Hz range (light blue streaks) (Figure 3A).

FIGURE 4.

Electroencephalography (EEG) power during cobalt (Co)-homocysteine-induced status epilepticus (SE). A plot of EEG power in each of the (A) delta, (B) theta, (C) alpha, (D) beta, and (E) gamma frequencies for 1-min intervals from the injection of homocysteine to the end of SE in animals. All data in this figure are from cohort 4. The power in each frequency range was normalized to its baseline frequency (n = 5). Time 0 represents injection of homocysteine

3.5.2 |. Continuous generalized convulsive seizures

This phase had continuous seizures manifested as tonic stiffening with head-bobbing, bilateral clonus with Straub-like tail, and generalized tonic-clonic seizures with loss of posture. Predominantly, during this phase, animals were head-bobbing with minimal forelimb and hind-limb movements. Some animals stood along the wall of the recording chamber. Occasionally, animals ran around the cage and intermittently jumped vertically. At other times, animals exhibited bilateral clonus with a Straub-like tail or clonus with a loss of posture.

On EEG, continuous spike-wave discharges predominantly in theta, alpha, and beta range frequency (<12 Hz) were present with occasional discrete peaks of activity >20 Hz during this period. Later, the frequency and amplitude of spike-wave discharges diminished (Figure 3F), interrupted by the periods of burst (Figure 3G). Power spectrogram and power graph analysis also revealed continuous high-power activity in the delta, theta, and alpha ranges with bursts of gamma and beta activity (Figure 4). After 15–20 minutes, polyspike-wave discharges appeared, and the activity was predominantly in the delta and theta range in all electrodes (Figure 3A). Later during this phase, high-power bursts disappeared, which marked the last phase of SE.

3.5.3 |. Periodic discharges and a coma

As generalized convulsions abated, motor activity drastically reduced with intermittent head-bobbing, involuntary vertical jumps, and purposeless movements. Electrographically, continuous seizure activity transitioned to generalized periodic discharges interspersed with bursts of polyspike discharges and periods of relatively flat EEG background (Figure 3G). Over time, periodic discharges decreased in amplitude and frequency, but bursts of spike and polyspike-wave activity persisted, which evolved into a burst suppression pattern (Figure 3H). During this phase, electrographic seizures lasting 10–20 seconds occurred between periods of burst suppression. The mean number of these seizures were 3.8 ± 1.0 (n = 7). With time, bursts became less frequent, and interburst intervals of low-amplitude background activity became longer until there was a complete suppression of activity in all the electrodes, which marked the end of SE (Figure 3I). Power spectrogram and power graphs revealed high-power peaks in the theta, alpha, and beta range interspersed with low-power activity (Figure 4). SE ended with low-power activity at all frequencies (Figures 3A and 4).

3.6 |. Latency, duration, behavior, and survival after SE

The latency and duration of SE were 14.14 ± 2.7 and 113.0 ± 7.7 minutes, respectively. The duration of SE was presented as the Kaplan-Meier curve (Figure 3J). The percentage time in each stage observed during SE was stage 1 (28.14%), stage 2 (6.89%), stage 3 (30.71%), stage 4 (11.49%), and stage 5 (22.75%) (Figure 3K). After the SE ended, animals remained immobile and died within 24 hours (n = 7) (Figure 3L).

3.7 |. Altered protocol improved survival

Co implantation was modified to allow its removal in two cohorts (5 and 6) to study cerebral edema and neurodegeneration. Co was soldered to a metallic pin and then inserted in the cortex. Animals received homocysteine (845 mg/kg, i.p.) 16–18 hours later, causing SE, which was observed by behavioral seizures. Four hours after the homocysteine injection and SE, animals were immobile, and they were anesthetized to remove the metallic pin. Animals recovered from anesthesia and survived.

3.8 |. Cerebral injury caused edema

T2-weighted images (cohort 5) obtained from saline-injected Co implanted animals demonstrated an increased signal, very localized to the site of Co implantation. These sites included the lateral orbital cortex, ventral orbital cortex, medial orbital cortex, frontal association area, and secondary motor cortex (Figure 5A,C). However, T2 signal for T240 animals was observed in the primary and secondary motor cortex, frontal association area, primary somatosensory cortex, lateral orbital cortex, ventral orbital cortex, dorsolateral orbital cortex, ventral and dorsal agranular insular cortex, and cingulate cortex area 24a, 24b, 25, and 32 ipsilateral to the Co implantation in comatose animals. T2 signal was also increased in the white matter tracts. The frontal cortex around the Co lesion was swollen, with the ipsilateral lobe crossing the midline (midline shift). The head of the ipsilateral lateral ventricle appeared smaller and slit-like with blurred ventricular margins when compared to that in Co animals (Figure 5B,D). In the horizontal sections, an increased T2 signal was observed up to bregma −2.80 mm, and in the coronal sections, and increased T2 signal was observed from AP = 2.6 to AP = −0.83. The area with increased T2-weighted signal was calculated using Image J for both coronal and horizontal images by two independent observers (one unblinded (TS) and one blinded (SJ)). Both observers found widespread increased T2 signal in T240 animals (shown as a magenta line graph) compared to saline-injected Co animals (shown as a blue line graph) (n = 5 for each Co and T240) (Figure 5E–H).

FIGURE 5.

Spatial distribution of cerebral edema in animals with cobalt (Co) and Co-homocysteine-induced status epilepticus (SE). The spatial distribution of cerebral edema in Co animals (Co) and Co-homocysteine animals developing SE (240 min following homocysteine administration during coma phase) (T240) as assessed in vivo by a 7.0 T magnetic resonance (MR) system. A and B, T2-weighted horizontal brain images displaying region-dependent severity of cerebral edema in animals implanted with (A) Co and (B) Co-homocysteine animals developing SE. C and D, T2-weighted coronal brain images showing region-dependent severity of cerebral edema in animals embedded with (C) Co and (D) Co-homocysteine animals developing SE (n = 5 for each Co and Co-homocysteine SE). E and F, The increased T2 signal measured by an unblinded observer in each slice for both coronal and horizontal orientations. G and H, The increased T2 signal measured by a blinded observer in each slice for both coronal and horizontal directions. Co and T240 animals are shown as a blue and magenta line graph, respectively. All data in this figure represent cohort 5

3.9 |. SE caused widespread neurodegeneration

The study from animal cohort 6 demonstrated that FJC-positive cells were present in the cerebral cortex, thalamus, and amygdala observed by two independent observers (one unblinded and one blinded). In the cortex, there were FJC-positive cells in deep layers of the motor cortex, somatosensory cortex, and auditory cortex. In the thalamus, FJC-positive cells were primarily seen in ventromedial (VM), lateral posterior (LP), lateral dorsal (LD) thalamic nuclei, reticular (RT), reunion (RE) areas, intermediodorsal (IMD), and mediodorsal (MD) thalamic nuclei. In amygdalar nuclei, FJC-positive cells were observed primarily in the lateral amygdala. Hippocampus was devoid of FJC-positive cells (n = 5) (Figure 6).

FIGURE 6.

Fluoro-Jade C (FJC)–positive neurons in several brain regions at 72 h after status epilepticus (SE). A, Widespread neurodegeneration was observed in various parts of the cortex, thalamus, and amygdala 72 h following SE. B and C, The number of FJC-positive cells was counted in representative microscopic fields (10× magnification) in different regions of the coronal brain section by a (B) unblinded (TS), and (C) blinded observer (SJ) (n = 5). All data in this figure represent cohort 6. AUD, auditory cortex; CA1, cornu ammonis 1; CA2, cornu ammonis 2; CA3, cornu ammonis 3 regions of the hippocampus; DG, dentate gyrus; IMD, intermediodorsal nucleus of the thalamus; LA, lateral amygdala; LD, lateral dorsal nucleus of thalamus; LP, lateral posterior nucleus of the thalamus; MD, mediodorsal nucleus of thalamus; MOs, motor area cortex; RE, nucleus of reuniens; RT, reticular nucleus of thalamus; SS, somatosensory cortex; VM, ventral medial nucleus of the thalamus.

4 |. DISCUSSION

In summary, we developed a neocortical injury-induced SE in mice. This model will expand the number of tools to study SE. It will help better understand the mechanisms and potential treatment strategies for this form of SE. It will also allow the use of genetically engineered mouse lines with specific genes added or removed.

Important features of this SE model include (a) cortical onset of seizures near the Co lesion; (b) burst suppression pattern on the EEG toward the end of SE; (c) seizure-associated behaviors including focal dystonia with clonus, tonic stiffening, and coma; (d) substantial cortical injury evident from cerebral edema, midline shift, and ventricular effacement on MRI; and (e) degeneration of cortical, thalamic, and amygdalar neurons but not that of hippocampal neurons.

The cortical tissue damage caused by trauma, hemorrhage, hypoxia, and infection often leads to acute seizures and SE.¹⁶ Previous studies reported that severe traumatic brain injury induced by lateral fluid-percussion injury to rats caused seizures and SE.^17,18 Co implanted in the motor cortex caused focal seizures.^19–21 Co binds to oxygen-binding molecules and causes functional hypoxia, triggers neuronal death, and increases expression of vascular endothelial growth factor.^16,22 Hypoxia triggers seizures in vivo,^23,24 increases neuronal excitability in brain slices, and enhances the amplitude of excitatory postsynaptic currents.^24,25 In addition to these mechanisms, Co implantation in the motor cortex also decreases the density of inhibitory neurons and their terminals, thus creating a seizure focus.²⁰

Homocysteine injection also causes seizures.^26–28 Homocysteine is an agonist for N-methyl-d-aspartate (NMDA) ionotropic glutamate receptors.²⁹ Activation of NMDA receptors allows the entry of Ca²⁺ into neurons and enhances excitability. Therefore, decreased GABAergic inhibition combined with NMDA receptor activation may have increased the excitability and triggered prolonged seizures.

The EEG patterns seen in mice during SE are characteristic of those in patients in whom seizures result from cortical insult.³⁰ Cortical circuits generate burst suppression;³¹ and these patterns arise due to the hyperactivity of cortex followed by refractory periods. The burst suppression pattern is also typically observed in cerebral hypoxia/anoxia following cardiopulmonary arrest.³² Homocysteine also caused burst suppression in young animals²⁶; however, we did not observe burst-suppression pattern when homocysteine was injected in adult animals.

We observed an increased T2-weighted signal in the frontal lobe around the Co lesion, which indicates edema, water accumulation, and blood-brain barrier damage.³³ Similar MRI evidence of edema and blood-brain barrier damage is present in patients with moderate to severe traumatic brain injury and hemorrhage.³⁴ In this model, Co lesion rendered the neocortex vulnerable to blood-brain barrier damage caused by homocysteine.³⁵ Brain edema increases intracranial pressure (ICP), thereby impairing cerebral blood flow, which may have resulted in the death of animals.³⁶ Removing Co from the brain after SE may have reduced ICP, thus improving survival in these animals.

SE-induced neurodegeneration was evident in the cortex, thalamus, and amygdala but not in the hippocampus. Neuronal damage in the cortex could have resulted from seizures, and also due to hypoxia/ischemia resulting from decreased perfusion related to cerebral edema. Burst-suppression pattern on EEG is associated with hypoxic-ischemic injury to the cortex, and in this model, we observed both. In addition, homocysteine activates glutamate receptors and may have contributed to neuronal damage.²⁹

In conclusion, we have developed a mouse model of frontal lobe–onset convulsive SE that evolves into electrographic seizures, and it ended in a comatose state. This model recapitulates electrographic patterns, extensive neocortical edema, and injury observed during human nonconvulsive SE following neocortical injury due to trauma or hemorrhage.

Key points.

• Prolonged self-sustaining, convulsive, and non-convulsive status epilepticus occurred following cortical injury.

• Focal motor seizures, electroencephalography (EEG), and the pattern of neuronal injury suggested that seizures spread throughout the cortex from the injury site.

• There was edema of the frontal lobe and widespread cortical damage.

• This model recapitulates clinical and EEG features of acute seizures following traumatic brain injury, lobar, and subarachnoid hemorrhage.