Address correspondence and reprint requests to Dr. C. Roch at Institut de Physique Biologique (UMR 7004 ULP/CNRS/IFR37), Faculté de Médecine, 4 rue Kirschleger, 67085 Strasbourg Cedex, France. E-mail: email@example.com
Summary: Purpose: Patients with temporal lobe epilepsy (TLE) usually had an initial precipitating injury in early childhood. However, epilepsy does not develop in all children who have undergone an early insult. As in patients, the consequences of the lithium-pilocarpine-induced status epilepticus (SE) are age dependent, and only a subset of 21-day-old rats will develop epilepsy. Thus with magnetic resonance imaging (MRI), we explored the differences in the evolution of lesions in these two populations of rats.
Methods: SE was induced in 21-day-old rats by the injection of lithium and pilocarpine. T2-weighted images and T2 relaxation-time measurements were used for detection of lesions from 6 h to 4 months after SE.
Results: Three populations of rats could be distinguished. The first one had neither MRI anomalies nor modification of the T2 relaxation time, and these rats did not develop epilepsy. In the second one, a hypersignal appeared at the level of the piriform and entorhinal cortices 24 h after SE (increase of 49% of the T2 relaxation time in the piriform cortex) that began to disappear 48–72 h after SE; epilepsy developed in all these animals. The third population of rats showed a more moderate increase of the T2 relaxation time in cortices (14% in the piriform cortex) that could not be seen on T2-weighted images. Epilepsy developed in all these rats. Only in a subpopulation of the 21-day-old rats with epilepsy did hippocampal sclerosis develop.
Conclusions: These results suggest that the injury of the piriform and entorhinal cortices during SE play a critical role for the installation of the epileptic networks and the development of epilepsy.
In temporal lobe epilepsy (TLE), patients typically have an initial precipitating injury (complicated febrile seizures, status epilepticus, encephalitis, head trauma, etc.) in their early childhood. After a silent phase of several years, seizures usually begin toward the end of the first decade of life and commonly respond to appropriate antiepileptic drug (AED) treatment, initially, but patients relapse during adolescence or early adulthood, when they become refractory to medical treatment. However, epilepsy will not develop in all children who have an initial precipitating injury; for example, in the case of complicated febrile seizures, epilepsy will develop in only 3% of the patients (1–4).
The model of TLE induced by pilocarpine alone or associated with lithium is very interesting because it reproduces most clinical, developmental, and neuropathologic features of human TLE (5,6). The consequences of pilocarpine-induced status epilepticus (SE) are age dependent, and the development of the epilepsy is a progressive process (5,7–9). Usually in adult rats, pilocarpine induces SE. This acute seizure phase is followed by a “silent” seizure-free phase of a mean duration of 14–25 days. Thereafter, all animals that underwent SE develop a severe pattern of neuronal damage and exhibit spontaneous recurrent seizures (SRSs). In rats aged 7 to 11 days, lithium-pilocarpine induces SE but does not lead to neuronal damage or SRSs (9). The 21-day-old (P21) rats represent the most interesting age in this model because in only some of the animals does epilepsy develop. Indeed, in a previous study, we showed that SRSs developed in 24%, and in 43%, seizures could be triggered by handling after a mean delay of 74 days in both cases (10–12).
All previous studies give information about the localization of the lesions and structures involved in the circuitry of epilepsy, but they are unable to explain why only a proportion of rats becomes epileptic when SE is induced at P21 because they do not perform a follow-up of the same animals. Studies performed during SE do not permit knowing in which individual will epilepsy develop, and studies performed during the chronic phase do not allow understanding of the critical factors that underlie the occurrence of SRSs in a given animal compared with an animal without epilepsy. Consequently, we used magnetic resonance imaging (MRI; i.e., a noninvasive technique) to follow up the evolution of the events triggered by SE induced by lithium-pilocarpine that may lead to epilepsy in P21 rats. A previous study using MRI in adult rats treated with lithium and pilocarpine showed that the occurrence of early edema in the piriform and entorhinal cortices characterized the initial step leading to the development of epilepsy and that hippocampal sclerosis developed with time and could be both the cause and the consequence of epileptic activity (13). Thus with MRI, we undertook the follow-up of a population of P21 rats to define the differences between those in which epilepsy will develop and those in which it will not.
Animals and treatment
For breeding purposes, adult Sprague–Dawley rats (Janvier Breeding Center, Le Genest-St-Isle, France), one male and two females in each cage, were housed together in mating groups for 5 days. After delivery, litters were reduced to 10 pups for homogeneity (day of birth was considered day 0). All animals were maintained under standard laboratory conditions on a 12/12-h light/dark cycle (lights on at 7:00 a.m.). All experiments were conducted in conformity with the rules of the European Committee Council Direction of November 24, 1986 (86/89/EEC) and French Department of Agriculture (license no. 67-97).
Lithium chloride (3 mEq/kg; Sigma, St. Louis, MO, U.S.A.) was administered intraperitoneally to P21 rats (n = 61), 18 h before the subcutaneous injection of pilocarpine (30 mg/kg; Sigma). The rats received 1 mg/kg methylscopolamine (Sigma) 30 min before the convulsant to reduce the peripheral consequences of pilocarpine administration. To improve survival, the animals received a deep intramuscular injection of 2 mg/kg diazepam (DZP; Valium; Roche, Meylan, France) 2 h after the onset of SE. Control rats (n = 4) also received lithium chloride and methylscopolamine and an equivalent volume of saline instead of pilocarpine.
A video-recording device was used to detect the occurrence of seizures. The animals were observed 8–10 h/day, 5 days a week (8.00 a.m. to 6.00 p.m.).
MRI was performed on a scanner operating at 4.7 T (SMIS). Anesthesia for MRI was induced by an intramuscular injection of 37 mg/kg ketamine (Ketalar; Pfizer, La Jolla, California, U.S.A.) and 5.5 mg/kg xylazine (Rompun; Bayer, Leverkusen, Germany). An MR scan was obtained from all animals 1 day before SE and (a) in the acute phase (6 h, 24 h, and 30 h after SE); (b) at the beginning of the silent phase (2 days and 1, 2, and 3 weeks after SE); and (c) then every 15 days (the last MRI was performed 4 months after SE).
In the 35 rats surviving SE, the brain was scanned from the olfactory bulb to the brainstem by using consecutive 1-mm-thick coronal slices with a T2-weighted, spin-echo fast imaging method sequence (3,800/80) to detect the lesions. A 40-mm field of view, 256 × 256-pixel matrix, 1-mm slice thickness, and two excitations were used as imaging parameters.
In 18 of the 35 rats, an additional spin-echo Carr-Purcell-Meiboom-Gill sequence (2,500/22/8 echos) was performed to measure the T2 relaxation time. The measurements of T2 relaxation time were performed on a 1.5-mm-thick coronal slice taken at the level of the dorsal hippocampus and piriform cortex. The 90-degree pulse was set by maximizing the signal of the profile of the head, using a recovery time long enough to avoid any saturation. Parametric T2 images (T2-calculated images) were created from the analysis of the transverse magnetization decay curve of each pixel. The T2 relaxation times were measured within anatomic boundaries of the piriform cortex by setting the largest possible region of interest. The values are presented as the percentage of the initial T2-relaxation time (i.e., T2 time obtained before the treatment). The T2 relaxation times also were calculated at the level of the parietal cortex to assess the stability of the measures (data not shown).
The four control rats were submitted to the same MRI protocol as the others to control the harmless effect of repeatedly performed MRI examinations (e.g., effects of anesthesia) and a possible age-dependent spontaneous variation of T2 relaxation times.
Behavior during lithium-pilocarpine SE
Control animals receiving lithium and saline did not exhibit any behavioral alteration. Within 5 min after pilocarpine injection, rats developed diarrhea, piloerection, and other signs of cholinergic stimulation. During the following 15 to 20 min, rats exhibited head bobbing, scratching, chewing, and exploratory behavior. Recurrent seizures started 20 to 25 min after pilocarpine administration. These seizures that associated episodes of head and bilateral forelimb myoclonus with rearing and falling progressed to SE at ∼50 min after pilocarpine, as previously described (6,11,12,14).
Of the 61 rats subjected to lithium-pilocarpine, 26 (42.6%) died during the SE, and 35 survived.
Occurrence of seizures
In the rats that survived (n = 35), six (17%) of 35 did not have seizures (no-seizure group), and in 29 (83%) of 35 rats, epilepsy developed after a mean delay of 70.8 ± 24.3 days (mean ± SD). In this epilepsy group, one third of rats (10 of 35) developed SRSs, one third (10 of 35) had seizures that could be triggered by handling or stress, and one third (9 of 35) had both SRSs and seizures triggered by handling or stress (Fig. 1).
MRI and epilepsy
None of the lithium-saline rats showed any change in the T2-weighted signal at any time in the study. The mean value of the T2 relaxation time measured in the piriform cortex was 85.5 ± 2.3 ms (mean ± SD of the values measured in four rats at eight different times).
During the acute phase
In rats that survived SE, a first group of six rats did not develop epilepsy.
• In this group, we did not see any modification of the signal on the T2-weighted image at any time after SE, and the values of the T2 relaxation time calculated in four rats at the level of the piriform cortex remained at all times in the range of the values for control rats (Fig. 2A).
The 29 remaining rats that had epilepsy could be separated in two groups (groups two and three).
• In group two (n = 13), a marked increase of the signal was seen on the T2-weighted image at the level of the piriform and entorhinal cortices (unilateral or bilateral) 24 h after SE. This hypersignal began to disappear as soon as 48 h after SE (Fig. 3). The measurement of the T2-relaxation time calculated in five rats 24 h after SE, showed an increase of 49.3% compared with the initial value. It returned to the range of control values rapidly, over the following 48 h (Fig. 2B)
• In group three (n = 16), no obvious modification of the signal appeared on the T2-weighted images in the cortex 24 h after the SE. However, the measure of the T2 relaxation time in the piriform cortex, evaluated in nine rats, indicated an increase of ∼14.1% compared with the initial value. Moreover, at 24 h after SE, the lowest value of the T2 relaxation time of the present group was always superior to the highest value obtained at the same time in the control rats. T2 relaxation time returned progressively to control values over the following 48 h (Figs. 1 and 2C).
During the silent and the chronic phases
At the beginning of the silent phase (until 25 days), it was not possible to distinguish the different groups either from each other or from the control group. We did not see any modification of the cortical signal on the T2-weighted images (Fig. 1). At the end of the silent phase, after a mean delay of 1 month, a hypersignal began to appear in the dorsal and ventral hippocampus in eight (61.5%) of 13 and five (31.2%) of 16 rats of the second (Fig. 3) and third groups, respectively. The signal hyperintensity in the hippocampus of these rats increased during the chronic phase (Figs. 1 and 3).
The findings of the present study demonstrated that only in P21 rats that had measurable anomalies on T2-weighted scans, visible or not on the images, did epilepsy develop. These abnormalities were present 24 h after the SE induced by lithium-pilocarpine mainly at the level of the piriform cortex and to a lesser extent in the entorhinal cortex. These results show that the piriform and entorhinal cortices play a critical role in the onset of epileptogenesis.
Damage and MRI
MRI is a sensitive in vivo technique permitting the diagnosis of structural abnormalities that underlie seizure disorders. This sensitivity is due to the fact that the signal amplitude is related to the water environment of tissues (i.e., the concentration and mobility of water). In almost all structural changes, including edema, gliosis, and neuronal loss, the concentration of free water is increased in the tissue. This increase in free water is characterized by an increase of the T2 relaxation values. However, a certain threshold in the increase of the T2 relaxation values must be reached for visualizing a hypersignal on T2-weighted images.
In P21 rats, in the pilocarpine or lithium-pilocarpine model of temporal lobe epilepsy, the pattern of neuronal damage is mostly limited to the hilus of the dentate gyrus, entorhinal and piriform cortices, and thalamus (7,9,11,12). Previous studies of our group showed that, in this age group, the extent of damage varied considerably with the animals and ranged from an almost negligible to a quite extended degree during the silent phase at 14 or 60 days after SE (10). However, at this time, the epilepsy outcome is not known. It is possible that the variability observed was linked to the intensity of the initial lesion and to the epilepsy outcome of the rat. Nevertheless, only a longitudinal follow-up of rats as MRI allows it would permit correlation of the extent of damage with the epileptic outcome of the P21 rat. Indeed, most often the lesions are quantified at a single time point, and this measurement necessitates the death of the animal. At the moment, it is not possible to study the evolution of damage from SE until occurrence of SRSs in a single animal with techniques other then MRI. Here we could highlight that the first structure injured after SE is the ventral cortex and mainly the piriform cortex. These structures are strongly activated metabolically during SE (12,15) but in P21 rats, many other structures reach very high metabolic levels during SE. Thus this kind of approach did not permit discrimination of the early role of the ventral cortices in the process of epileptogenesis. However, the constitution of the epileptogenic circuit can be underlined with the measurement of local cerebral metabolic rates for glucose (LCMRglc). During the silent phase, 14 days after the SE, metabolic decreases were recorded mainly in damaged forebrain regions. Conversely at 60 days after the SE (i.e., at the end of the silent phase), LCMRglcs increased in both damaged forebrain and intact brainstem areas. The high metabolic levels in forebrain areas may reflect the genesis of a new circuit underlying the occurrence of spontaneous seizures, whereas metabolic increases in the brainstem (regions involved in the remote control of seizures) might underlie a process of protection against the occurrence of seizures (10,12). During the chronic phase, the interictal glucose metabolism is different whether epilepsy has developed or not. In rats without epilepsy, LCMRglcs of the forebrain were hypometabolic, whereas in epileptic rats, LCMRglcs were normo- to slightly hypermetabolic, indicating an active process taking place in damaged areas of these epileptic brains (11,12). The use of MRI allows us to follow the organization of the circuit. Compared with the early events taking place in ventral cortices, the hippocampus reacts later to the initial injury (i.e., at the end of the silent phase). We could see the hippocampal sclerosis developing progressively on T2-weighted images (characterized by a hypersignal) from the end of the silent phase (∼1 month after SE) until the end of the study (4 months after SE). However, an MRI-detectable hippocampal sclerosis (i.e., macroscopic) does not seem to be essential for the development of epilepsy because we found some rats with seizures that experienced no change on the T2-weighted images at the level of hippocampus.
However, T2-weighted images, per se, are not really sufficient to assess all modifications occurring in neurons. The measurement of the T2 relaxation time was necessary to observe abnormalities in the piriform cortex of rats that developed epilepsy, although they did not show obvious edema on T2-weighted images. Moreover, in the hippocampus of P21 rats, neuronal loss is found mainly in the hilus of the dentate gyrus (10–12,15). The resolution of our imaging procedure may not be sufficient to detect such restricted lesions. Diffusion-weighted imaging would probably be more useful to detect earlier and minimal brain damage, as previous studies on kainate-induced SE have shown (16,17). Indeed, a slight reduction of the apparent diffusion coefficient (ADC) was observed as soon as 12 h in regions like the cerebral cortices and amygdala, whereas no changes could be seen on T2-weighted images before 24 h. Between 24 and 72 h, the ADC decreased and subsequently increased, but at the same time, the T2-weighted signal remained uniformly hyperintense. It was proposed that the initial ADC decrease is associated with neuropil swelling, reflecting a shift of water from extracellular environment to the immobile intracellular region, which permits earlier detection of cell damage. The subsequent increase in ADC results in neuropil fragmentation and the consequent loss of diffusion barriers. T2-weighted signal is mostly due to the fragmentation of the previously swollen neuropil [i.e., an increase of free water in the tissue (18–20)].
Role of the cortex in epileptogenesis
In the present study, we saw that at 24 h after SE, a hypersignal appeared at the level of the piriform and entorhinal cortex. This hypersignal disappeared over the following 48 h, so that, at the latter time, no signal difference could distinguish the rats. However, a measurable increase in the signal in the piriform cortex 24 h after SE is necessary to lead to epilepsy. These results suggest that the cortex plays a key role in the development of epileptogenesis and is predictive of epileptic outcome.
Cortices are often described as being involved in the propagation of seizures. In all models of convulsive SE, the piriform cortex invariably demonstrates major pathology (21), and in the kindling model, animals that underwent complete lesions of the piriform cortex were unable to develop secondarily generalized seizures, demonstrating that the piriform cortex is extremely important for the generalization of kindled seizures (22). The neurons of the piriform cortex have especially reactive properties (23,24), and a number of studies have shown that piriform cortex contains the most susceptible neural circuits of all forebrain regions for electrical (or chemical) induction of limbic seizures (25–29). The injection of bicuculline [antagonist of γ-aminobutyric acid subtype A (GABAA) receptors] into the piriform cortex is able to initiate generalized seizures (30–32). The kainate, AMPA, and muscarinic receptor agonists like carbachol, N-methyl-d-aspartate (NMDA), quisqualate, glutamate, and aspartate were effective in eliciting bilateral clonic seizures after unilateral injection into the anterior part of the piriform cortex (33,34). All these studies showed that the piriform cortex is important for the initiation of seizures and confirm ours (i.e., piriform cortex plays a key role for the development of SE and then for epileptogenesis).
In the pilocarpine models, studies using MRI in adult rats have shown an early MRI signal and lesional process of the cortex (piriform and entorhinal) (13,35). Moreover, the application of NMDA-receptor antagonists into the anterior piriform cortex is able to prevent pilocarpine-induced seizures with doses lower than those required for anticonvulsant actions in other brain areas (36–40). These results indicate that an interaction between the cholinergic and amino acid systems at this level may be responsible for initiation of pilocarpine-induced seizures (31). Likewise, epileptiform activity generated by bicuculline in the piriform cortex propagates and induces secondary excitability changes in the entorhinal cortex and the hippocampus (41–44). The interconnections of the piriform cortex with the entorhinal cortex and the hippocampus are characterized by a direct projection from the deep cells of the piriform cortex to both lateral entorhinal cortex and ventral subiculum (45,46) which comprises the main input and output structures of the hippocampus (47). The lateral entorhinal cortex in turn projects back to the posterior piriform cortex and endopiriform nuclei (46). All these results confirm the key role of the piriform/entorhinal circuit with a secondary spread to the hippocampus in the SRSs induced by lithium-pilocarpine SE.
Development and epileptogenesis
The data of the present study confirm that 21-day-old rats are the most interesting age in this model to study the mechanisms underlying epileptogenesis because in only a subset of the animals does epilepsy develop (7,9–12). In our study, 89% of rats that survived the SE developed epilepsy after a mean duration of 70.8 days. All these rats and only these rats had MRI abnormalities in ventral cortices 24 h after the SE. It appears that, in these rats, the ventral cortices are more sensitive to pilocarpine effects than in rats that will not develop epilepsy. This difference in sensitivity may be linked to the fact that all rats of a same litter most likely do not reach the same degree of maturity at the same time. P21 seems to be a critical age for the appearance of epilepsy because some rats behave like adults, whereas others cannot reach this threshold. Various hypotheses can be advanced to explain this difference. First, the immature rat uses, in addition to glucose, ketone bodies to support its metabolic and biosynthetic needs (for review, see 48). The intake in carbohydrates increases in these animals around P17 when they start eating the carbohydrate-rich lab chow of the dam, and the latter behavior may largely vary between the animals. Moreover, at early ages, glucose is metabolized predominantly by the anaerobic glycolysis pathway because the enzymes responsible for the oxidative glucose breakdown start their rapid maturation phase only after P21 (for review, see 48). Therefore quite large variations may occur in the maturation of brain energy metabolism that may underlie the brain response to SE-induced activation and damage. Second, it is well known that the maturity of the forebrain limbic circuits is reached only toward the end of the third week of postnatal life (49). The concentration of muscarinic receptors at birth corresponds to 10% of adult concentration and increases linearly to 90% by 4 weeks postpartum (50). Because pilocarpine is a muscarinic agonist, it is possible that a minimal concentration of muscarinic receptors is necessary to allow neuronal activity to reach a level high enough during SE to trigger the cellular events that are required for the establishment of epileptogenesis. Again it is highly likely that the maturation of the muscarinic receptor system is not perfectly identical in all rats of a given age.
In conclusion, cortices play a key role in epileptogenesis, and their injury is predictive of epilepsy in the lithium-pilocarpine model of TLE. Nevertheless, the reason that only a subset of P21 rats becomes epileptic remains to be explored.
Acknowledgment: We are very grateful to B. Guignard for skillful technical assistance. This work was supported by a grant from the Hôpitaux Universitaires de Strasbourg, the Université Louis Pasteur, the Institut National de la Santé et de la Recherche Médicale, and the Fondation pour la Recherche Médicale.