Twenty-nine adult (7 weeks old) male Sprague-Dawley rats (Orient Animal Cor., Gyeonggi-do, Korea) weighting 280~300 g were used for this study: 14 animals for video-EEG recording and 15 animals for MRI and FDG-PET images. All animals were handled daily for at least 7 days prior to the experiments. The animals had free access to food and water ad libitum and they were maintained on a 12 h light-dark cycle (light on from 08:00~20:00 h) under controlled temperature (20~25℃). All experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of the Ewha Womans University School Medicine.

For EEG recordings, steel cortical screw electrodes were placed over the cortex unilaterally as describe in stereotaxic atlas of Paxino and Watson. Under ketamine anesthesia, rats were chronically implanted with the head mount. Two stainless steel screws were implanted through the head mount into four predrilled holes of the skull, and the device was secured with dental cement. After surgery, each rat was allowed 7 days in an individual transparent barrel for recovery.

Animals were injected with pilocarpine hydrochloride (300~380 mg/kg intraperitoneally, Sigma-Aldrich) after 30 min of methylscopolamine injection to induce SE similar to previous studies [1,2,10,11]. Diazepam was administrated 8.6 mg/kg (Sigma-Aldrich) depending on the seizure severity after 120 min of initial SE. Animal that did not ad SE after pilocarpine injection were additionally injected with pilocarpine (110 mg/kg). The schematic design of this study was summarized in Fig. 1.

For seizure recordings, the rats were placed individually in custom-made 15×40×25 cm Acryl cages, and performed continuous video-EEG recording systems (Twin 7.0, Grass-Telefactor, West Warwick, RI, USA) after the pilocarpine injection where each channel was sampled rate of 200Hz and filtered using digital high-and low-pass filters at cutoff frequencies of 0.1 Hz and 70 Hz, respectively. All animals developed SE after the pilocarpine injection, and developed spontaneous recurrent seizures (SRS) approximately 1~2 weeks after the initial SE. The severity of SE and SRS were rated as class 1 to 5 by modified Racine classification [12], but only class 3 to 5 SRS were counted for the duration and frequency to assess seizure severity. Racine's criteria is as follows: stage 1, stereotype mouth movement, eye blinking and/or mild facial clonus; stage 2, head nodding and/or sever facial clonus; stage 3, myoclonic jerks in the forelimbs; stage 4, clonic convulsions in the forelimbs with rearing; and stage 5, generalized clonic convulsions associated with loss balance. After all SRS had been reviewed by both reviewers independently, a final consensus was reached for each SRS through discussion of those in which their reviews were different. Seizures that occurred within a week was not considered as SRS. Therefore, the time points of the earliest and latest seizures were 8 days and 56 days, respectively. The duration of SRS is the time that SRS lasted, and the frequency of SRS is the number of the occurrence of SRS per day. The criteria for seizures in EEG was defined as distinctive activities from the baseline, often repetitive epileptiform and/or rhythmic discharges evolving for at least 2~3 sec accompanied by behavioral changes on video-monitoring [13,14]. An examples of ictal EEG activity during a seizure was illustrated in Fig. 2A.

All neuroimaging procedures were performed under anesthesia with a mixture of ketamine (80 mg/kg) and xylene (5.2 mg/kg). Structural brain images were studied by 7.0 T MRI scanners (Magnetom 7T, Siemens, Enlargen, Germany). For the MRI, Spin Echo sequence was acquired using the imaging parameters with repetition time (TR)=6000 ms, echo time (TE)=60 ms, voxel size=0.22×0.22×0.4 mm, pixel band width (BW)=80 Hz/pixel, and flip angle (FA)=90°. The total acquisition time (TA) was 14 min and 21 sec.

Each animal was deprived of food for 12 hours, and accumulated [¹⁸F]FDG-6-PO₄ metabolic imaging was performed after administration of [¹⁸F]FDG (500 µCi/100 g body weights) using a Focus 120 Micro PET system (Concorde Microsystems, Knoxville, TN, USA) with 1.18 mm×1.13 mm matrix size, and 1.3 mm full width at half maximum (FWHM) resolution. Both MRI and PET were obtained at 7 time points including the baseline, 3 hours, 1, 3, 7 days, 4 and 8 weeks following the initial insult (Fig. 1). The repeated exposures to MRI and PET would have some changes in the imaging conditions in these animals, which were minimized by the standard procedures as described previously [9].

For MRI and PET coregistration, the brain regions of the animal micro PET image were manually extracted for statistical analysis first. PET images were reconstructed using the ordered subset expectation maximization (OSEM) algorithm. Nominal pixel size was 0.43×0.43×0.81 mm³, and then coregistered to rat brain template using statistical parametrical mapping (SPM8, Wellcome Trust Centre for Neuroimaging, http://www.fil.ion.ucl.ac.uk/spm). Second, it was normalized into the MRI template for accurate anatomical information in stereotaxic space. Third, all of the brain images were normalized the [¹⁸F]FDG rat brain template. The individual normalized images were then smoothed with a 2-mm FWHM isotropic Gaussian kernel. Voxel-based statistical analyses were performed using SPM8 toolbox implanted in MATLAB. Finally, T-value maps were created superimposed on transverse planed of the MRI template to identify voxels of statistical significance [15,16].

Seizure severities were determined by the duration and frequency of SRS for each day after pilocarpine injection were quantified during 8 weeks. The duration of SRS is the time that SRS lasted, and the frequency of SRS is the number of the occurrence of SRS per day. The total duration and frequency of seizures within each phase (acute, latent and chronic phases after SE) were divided by the number of days to calculate average daily seizure duration and frequency.

For statistical analysis, correlation between duration and frequency of class 3 to 5 SRS at chronic phase and changes in MRI and FDG-PET acquired at each time point was examined. The seizure severities estimated by averaged duration and frequency of SRS was calculated by averaging the total durations and frequency of class 3 to 5 SRS during chronic phase, which were examined for correlation with intensity of each voxel from the brain imaging scanned at each time point, using the 3-dimensional mapping of Pearson correlation coefficients, constructed to represent the brain areas that were significantly correlated with seizure duration or frequency (p-value<0.05 with FWE correction for multiple comparison). Co-registration of raw images and statistical images of Pearson correlation was created by MRIcro software.

Continuous video-EEG monitoring data were obtained from all experimental animals after pilocarpine injection. We assessed daily seizure counts and duration during the entire study period (Fig. 2B~E). The initial SE was induced 4.9±2.5 min after pilocarpine injection, and the latent period for the appearance of the first SRS was 8.0±1.3 day after SE. At acute phase, mean seizure count per day were 1.3, 3.1, and 2.8 for class 3, 4, 5 SRS, and mean seizure durations per day were 42.4 sec, 202.1 sec, and 195.9 sec for class 3, 4, 5, respectively. At chronic phase, mean seizure counts per day were 0.4, 0.3, and 0.1 for class 3, 4, 5 SRS, and mean seizure durations per day were 27.4 sec, 26.6 sec, and 14.6 sec for class 3, 4, 5, respectively. Seizures within 3 days were considered to be caused by the initial insult closely related to pilocarpine injection, and seizures that occurred 8 days after the initial insult only were considered as SRS.

We also examined time course of changes in T2 signal intensity on follow-up MR images, from 3 hours to 8 weeks (7 time points), in all pilocarpine-induced epileptic rats following SE. We found that 7T MRI mainly showed increased T₂WI signal intensities in bilateral hippocampi, piriform/entorhinal cortices (Fig. 3A, upper row). The hippocampal T2 signal intensity was slightly increased at 3 hours after the initial insult and became more focal and prominent at 1 day and 3 days after the insult. This T2 signal further increased at 7 days and persisted with atrophic changes throughout the chronic phase at 4 and 8 weeks. Significant T2 increase was seen in the thalamus at 1 day, but resolved by 7 days. In piriform/entorhinal cortices and amygdala, a markedly increased T2 signal appeared at 1 day and 3 days, and significantly decreased at 7 days. T2 signal intensity increased again in piriform cortex and amygdala at the chronic phase (4 and 8 weeks after the initial insult).

Lower rows of Fig. 3A illustrated subtraction FDG-PET glucose metabolism at each time point to the baseline PET image and coregistered with MRI to yield accurate anatomical localization. FDG uptake revealed mostly decreases in bilateral hippocampi, amygdala, and part of subcortical or frontal areas throughout acute, subacute, and chronic stages. Although the main findings of PET subtraction were decreased glucose metabolism, some areas showed focal increase of metabolism; in bilateral entorhinal, cerebellum at acute stages (3 hours, 1 day, respectively), and right hypothalamus at chronic stage (8 weeks).

Neuroimaging changes in MRI and PET at each time point were investigated whether they have significant correlation with duration and frequency of SRS during chronic phase. Fig. 3B showed relationship between MRI/PET changes and seizure severities such as mean daily frequency and/or total duration after SE, namely SRS. Mean daily frequency of SRS was positively correlated MRI T2 signal in bilateral hippocampal areas at subacute stage of 7 days (in Fig. 3B, left, r=0.606, corrected p-value<0.05). Total seizure duration was negatively correlated with FDG-PET glucose metabolism in both hippocampal areas at subacute to chronic stages of 3 days (r=−0.589, corrected p-value<0.05; Fig. 3B, middle) and 7 days (r=−0.764, corrected p-value<0.05; Fig. 3B, right).