SEARCH

SEARCH BY CITATION

Keywords:

  • Seizures;
  • Status epilepticus;
  • Epilepsy;
  • Development;
  • Neuronal injury;
  • Synaptic plasticity

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Summary:  Purpose: To use animal models of variable seizure induction in rats at different developmental stages to determine contributing factors for spontaneous seizures resulting from status epilepticus (SE) early in life.

Methods: Two models of SE with distinct modes of seizure induction, lithium–pilocarpine (LiPC) and perforant path stimulation (PPS), were used at different ages. Multiple methods of determining neurodegeneration during an acute period and plastic changes in those monitored during the chronic phase were used.

Results: Different modes of seizure induction lead to varying types and extents of damage, dependent on the age of the animals at the time of insult. LiPC resulted in injury to animals as young as 2 weeks and became widespread in animals 3 weeks old, whereas widespread damage after PPS was not seen until P35. Rats at an age with widespread damage in response to seizures also showed extensive immediate-early gene activation and often developed spontaneous seizures and features of hippocampal plasticity seen in the epileptic brain.

Conclusions: SE early in life results in multiple consequences to the developing brain. These changes, coexisting in the nonepileptic brain, can overlap in a maladaptive combination to result in the diseased state of epilepsy. The consequence of early seizures in immature animals is a function of both the developmental stage and the method of seizure induction.

The consequences of seizures in the developing brain continue to be a dynamic topic of discussion (1) and debate (2,3). This issue is not readily answered by human studies. Historically, febrile convulsions have been cited as the classic example of the benign nature of seizures in children with no previously demonstrated brain pathology (4,5). However, some studies report adverse sequelae after febrile status epilepticus (SE) (6,7). Additionally, recent studies suggest that prolonged seizures in children, including complex febrile seizure, may result in brain injury that becomes discernible with new, noninvasive imaging techniques (8–11).

Over the years, numerous investigators have attempted to provide further insight into this controversy using the inherent advantage of animal research to control factors that are problematic in the clinical situation. Nevertheless, the general concept of age-specific seizure-induced vulnerability has been further complicated by the multiple ages examined in diverse studies, and the method used to produce seizures or SE. For example, opinions of the resistance of the developing brain to seizures were bolstered by early work showing lack of histologic damage or mossy fiber sprouting (MFS) in response to kainic acid (KA)–induced SE in 2-week-old animals (12–15). Conversely, corticotropin-releasing hormone (CRH) is a potent convulsant that induces limbic seizures at picomolar doses in infant rats (P10–P13) (8). CRH-induced SE causes neuronal damage in the CA3b subfield of the hippocampus and mossy fiber reorganization in these rats (17,18). These effects are not demonstrable in mature rats. Investigations in our laboratory and others with the lithium–pilocarpine (LiPC) model have shown that young animals can undergo region-specific brain damage after SE (19,20) and that those patterns can vary with age (21).

Further, SE-induced damage may have been underestimated in the past. Many studies evaluated only cresyl violet sections to count remaining neurons in an animal with a chronic disorder, rather than label acutely damaged neurons with high sensitivity. In addition, markers of apoptosis were not customarily examined as they are today. DNA fragmentation in response to SE has been shown in subpopulations of damaged hippocampal (21) and extrahippocampal neurons (22). Meanwhile, some damaged hippocampal neurons may have been replaced by increased neurogenesis reported in the mature as well as in developing brain (23–27).

Even in the absence of discernible injury, many changes have been described in recent years as a consequence of early-life seizure. For instance, in a model of hyperthermic seizures in rat pups, postnatal age 10–12 (P10–P12) designed to mimic febrile convulsions in children, even seizures lasting ∼20 min resulted in lasting changes in presynaptic, γ-aminobutyric acid (GABA)ergic inhibitory transmission (28). Moreover, such relatively brief hyperthermia-induced seizures have been shown to reduce thresholds to chemical convulsants in vivo and electrical stimulation in vitro, indicating persistent enhancement of limbic excitability that may facilitate the development of epilepsy (29). Seizures induced by kainic acid in P15 pups enhanced the damage sustained by these animals when rechallenged by KA at P45 (30). Holmes et al. (31) examined the effect of recurrent, brief, flurothyl-induced seizures in neonatal rats. These rats demonstrated impaired learning and decreased activity levels. They also displayed MFS in area CA3 and the supragranular region. When studied as adults, these animals had a decreased threshold to pentylenetetrazol (PTZ)-induced seizures. The authors concluded that recurrent brief seizures during the neonatal period have long-term detrimental effects on behavior, seizure susceptibility, and brain development. In another article, these authors reported MFS in the CA3 region after brief PTZ-induced seizures in P1 and P10 rats (32). Not surprisingly, in a review in Neuron, the authors concluded that seizures were not so benign to the immature brain after all (33). More recently these authors showed that when neonatal rats are subjected to flurothyl seizures, and the hippocampal slices are studied subsequently, there are differences in the intrinsic membrane properties of CA1 pyramidal cells (34). CA1 pyramidal cells from rats treated with flurothyl in the neonatal period displayed a marked reduction in the spike frequency adaptation and after hyperpolarizing potential after a spike train compared with controls.

We present some of our findings of SE-induced changes in the immature brain and attempt to delineate multiple factors that together may ultimately result in the chronic epileptic state in these animals.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Seizure induction

Wistar rat pups (Simonson Labs, Gilroy, CA, U.S.A.) at 1, 2, 3, 4, 5, and 9–12 weeks old were used in these studies. The committees on Animal Research at VAGLAHS and UCLA approved the protocols. Some rats were given LiPC SE (3 mEq/kg lithium chloride, i.p., followed 16–20 h later with pilocarpine, s.c.). Perforant-path stimulation (PPS) was used to induce SE in rats 3 weeks and older (8 h in the awake state for pups or 30 min in adults, with 10-s, 20-Hz trains delivered every minute on a background of 2-Hz continuous stimulation via a bipolar stimulating electrode in the entorhinal cortex and recorded via recording electrode in the dentate granule cell layer of the hippocampus).

s-NSE elevation after SE

Twenty-four hours after LiPC SE, the rats were overdosed with methoxyflurane anesthesia, and a cardiac blood sample was drawn for the estimation of serum neuron-specific enolase (s-NSE). Approximately 1 ml was allowed to coagulate for 20 min at room temperature and then centrifuged at 8,000 rpm. The serum was collected and stored at −70°C until the s-NSE assay was run using a radioimmunoassay (RIA) kit (Pharmacia Inc., Kalamazoo, MI, U.S.A.).

Histologic examination using hematoxylin and eosin (H&E) and TUNEL

Rats were perfused with phosphate-buffered saline (PBS) followed by 4% paraformaldehyde. The brains were dehydrated in graded alcohol solutions, cleared in xylene, embedded in paraffin, and cut at 8-μm sections. For H&E, sections were rehydrated in water and stained with modified concentrations of eosin to enhance the detection of damaged cells when viewed under fluorescent light. For TUNEL staining, paraffin-embedded sections were labeled with a commercially available in situ labeling kit (Oncor, Gaithersburg, MD, U.S.A.) and lightly counterstained with 1% methyl green (Sigma, St. Louis, MO, U.S.A.).

Electron microscopy

Some animals were perfused with PBS followed by 3% glutaraldehyde (Electron Microscopy Sciences, Ft. Washington, PA, U.S.A.) and 1% paraformaldehyde (Fisher) in 0.1 M phosphate buffer (PB), pH 7.4. Brains were removed and postfixed for 1 day and then cut on a vibratome into 100-μm sections. The sections were washed with PB, stained for 1 h in 2% OsO4 (Electron Microscopy Sciences), and en bloc stained in 2% aqueous uranyl acetate (Mallinckrodt, Mundelein, IL, U.S.A.) for 30 min. The sections were then dehydrated in ascending graded solutions of alcohol for 10 min each, placed in 1:1 100% EtOH/propylene oxide for 5 min, followed by two incubations in propylene oxide for 5 min. The sections were then incubated in 1:1 and 1:3 propylene oxide/Durcupan (Fluka, Buchs, Switzerland) for 1 h each, followed by Durcupan overnight. The sections were placed between strips of Aclar plastic and cured for 24 h at 60°C, after which they were cut on an ultramicrotome. They were placed on single-hole grids, covered by butvar, stained with 6% uranyl acetate and lead citrate, and examined with a Philips Electron Microscope 201C.

Immunohistochemistry for c-Jun or substance P (SP)

Two hours after the beginning of PPS or the onset of SE in pilocarpine-treated pups, rats were deeply anesthetized and perfused transcardially with PBS followed by 4% paraformaldehyde. They were postfixed in the same solution for 2 h and placed in Tris PBS/30% sucrose (pH 7.4, 4°C) for 24 h, and then serially sectioned at 40-μm-thick coronal sections on a cryotome; blocked in Tris-PBS containing 3% normal goat serum (NGS) and 0. 3% triton X-100 for 1 h, incubated in rabbit anti c-Jun serum (Santa Cruz) at a concentration of 1:4,000 in 3% NGS and 0. 3% triton X-100 overnight at room temperature; washed with Tris-PBS 3 × 10 min each, incubated in biotinylated goat antirabbit immunoglobulin G (IgG; Vector) at 1:200 in 1% NGS and 0. 3% triton X-100 for 1 h at room temperature; washed with Tris-PBS for 3 × 10 min each, and then incubated in avidin/biotin/peroxidase complex (Vicastain; Vector Laboratories) for 1 h. To identify the immunoreaction product, the horseradish peroxidase was visualized with diaminobenzidine and glucose oxidase, with nickel intensification. For SP immunohistochemistry, rats were perfused, and brain sections were prepared as described earlier 24 h after SE, incubated in rabbit anti-SP antiserum (1:8,000; Peninsula Labs), and visualized in the same manner.

Measuring paired-pulse inhibition (PPI) and monitoring for chronic spontaneous seizures

Animals that underwent SE were subjected to EEG monitoring ≥2 months after the initial experiments and were implanted as described earlier. The depth of both electrodes was optimized by maximizing the amplitude of the population spike evoked from the dentate gyrus by stimulation of the perforant path. After another week-long recovery, the animals were examined for PPI (30 V, 0.1 Hz, 40-ms interstimulus interval) under ketamine anesthesia. Five consecutive waveforms were recorded and averaged for quantifying evoked stimuli. PPI is expressed as an inhibition score calculated as 100 minus the percentage ratio of the second population spike to the first. After a week-long recovery, the animals were monitored for 24 h/day using a video camera and EEG software (the Monitor 81 program from Stellate Software) configured for automatic detection of seizures. If seizures were not detected after an initial 24-h recording, the animals were monitored for an additional 48 h. The EEGs were analyzed off-line with the same software.

Mossy fiber sprouting using Timm's staining

After monitoring for spontaneous seizures and ≥4 months after the initial SE, rats were anesthetized with pentobarbital and perfused transcardially with an aqueous solution of 250 ml of 0.1% (wt/vol) sodium sulfide, followed by 250 ml of 4% paraformaldehyde solution. Coronal 30-μm sections were developed in the dark for 30 min in a 6:3:1 mixture of gum arabic (20%, wt/vol), hydroquinone (5.6%, wt/vol), citric acid–sodium citrate buffer with 1.5 ml of a silver nitrate solution (17%). Images were captured on a frame grabber and converted to gray scale for analysis using a tracing function in Image Pro Plus software version 2.0 (Silver Spring, MD, U.S.A.) to determine the gray value difference (GVD) between the inner and outer molecular layers on a 0–255 scale.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

s-NSE elevation after SE

The s-NSE concentrations from control animals and those undergoing LiPC SE are shown in Table 1. The lack of change in s-NSE in the 1-week-old rat corresponded to the absence of any discernible cell injury at that age in response to LiPC SE. Elevation was seen beginning at age 2 weeks and increased proportionately thereafter. However, the absolute value peaked at week 3 of life. The very young animals showed a higher s-NSE level even without SE, perhaps reflecting apoptosis during development, which is maximal during the first postnatal week in the rat.

Table 1.  Serum neuron-specific enolase (ng/ml) in rats subjected to LiPC SE
Age (wk)ControlLi-PC SE
  • LiPC, lithium–pilocarpine; SE, status epilepticus.

  • a

     p < 0.005 compared with 2-, 3-, or 4-week-old animals.

  • b

     p < 0.001 compared with controls.

117.1 ± 1.0a14.8 ± 0.6
211.5 ± 0.518.9 ± 0.8b
312.1 ± 0.835.8 ± 2.1b
49.3 ± 0.534.9 ± 1.7b
Adult (12–16 wk)5.4 ± 0.4a30.4 ± 1.3b

Characterization of neuronal injury

Evaluation of the type of cell injury resulting from LiPC SE showed that both apoptotic and necrotic features could be discerned in the CA1 region of P14 rat pups. At this age, the CA3 and hilar neurons seemed to be spared. By P21, the hilar and CA3 injury resulting from LiPC SE was comparable to that seen in fully mature rats (Fig. 1A). There was no evidence of apoptosis in the CA1 region in these animals, but apoptotic morphology could be discerned along the hilar border of the granule cells sing several morphologic techniques that included nick-end labeling of fragmented DNA and electron microscopy (EM; Fig. 1B).

image

Figure 1. Immediate effects of status epilepticus in P21 rats. Lithium–pilocarpine (LiPC) treatment results in dentate hilus and granule cell injury, viewed as brightly eosinophilic cells under fluorescent light (A). Electron microscopy of granule cells near the hilar border shows nuclei with apoptotic features (arrows) (B). c-Jun immunoreactivity 2 h after perforant-path stimulation (C) shows little labeling in the amygdala and piriform cortex, but is present after LiPC (D).

Download figure to PowerPoint

SP expression

SP, an 11-amino acid neuropeptide, positively modulates both pre- and postsynaptic effects of glutamate in the central nervous system (CNS). Therefore, we tested the age-dependent seizure vulnerability of excitotoxic injury in the hippocampus and found that it correlates with SP expression. Maximal expression was seen in the CA1 region of 2-week-old rats and decreased thereafter. However, in the dentate granule cell and CA3 regions, its expression in response to SE was not discernible in the 2-week-old pups but was maximal in the 3-week-olds.

c-Jun immunocytochemistry

A similar pattern of hippocampal immunoreactivity, with a higher signal in the dentate granule cells than the pyramidal cell layer, was seen in 3-week-old rats that were subjected to LiPC or in 3- and 5-week-olds with PPS SE. Animals that underwent LiPC SE at P21 or PPS at P35 showed higher immunoreactivity in extrahippocampal structures including the amygdala and piriform cortex than did those that experienced PPS at P21 (Fig. 1B and C). Even though only light staining for immunoreactivity was seen in the substantia nigra of the P21 animals after LiPC SE and of the P35 animals after PPS, none was discernible in the P21 animals after 2 h of PPS. The P21 animals also showed very weak immunoreactivity in the temporal neocortex and thalamus after 2 h of PPS compared with those that underwent LiPC SE or those that experienced PPS at P35. The cingulate cortex and hypothalamus reflected the same pattern of immunoreactivity, with strong signals in the P21 pups subjected to LiPC SE or P35 animals subjected to PPS, but very light staining in the P21 pups after 2 h of PPS.

Paired-pulse inhibition (PPI) in the dentate gyrus

All groups of animals, given either LiPC or PPS, and their respective controls, showed similar PPI scores (80–100) before the administration of ketamine. Recordings were made under ketamine anesthesia after the rats had developed chronic spontaneous seizures or 3 months after the initial PPS or LiPC SE experiments. In control animals (P21, P35, and adult animals that were implanted but not subjected to PPS), no PPI was discernible at 40-ms interstimulus interval (ISI). The group subjected to 8 h of PPS at P21 showed mild paired-pulse facilitation, as evidenced by a negative PPI score (−27; Fig. 2A). However, rats that had undergone LiPC SE at that age (P21) demonstrated clearly discernible dentate inhibition with a PPI score of 57 (p < 0.05 compared with PPS at P21; Fig. 2B). The rats that had undergone 8 h of PPS at P35 and those that had been subjected to 30 min of PPS as adults had inhibition scores of 70 and 90, respectively (p < 0.05 compared with PPS at P21) 2–4 months after the initial PPS (Fig. 2E).

image

Figure 2. Long-term effects of status epilepticus in P21 rats. Perforant-path inhibition (PPI) is not seen after perforant-path stimulation (PPS) in 3-week-old rats (A), but is strong several months after lithium–pilocarpine (LiPC) (B). Mossy fiber sprouting (MFS) is not seen after PPS (A) and is visible after LiPC (B). E: Results of PPI, MFS, and incidence of spontaneous seizures in several age groups with different treatments. * p < 0.05.

Download figure to PowerPoint

Chronic spontaneous seizures

Systematic observation for 24-h observation periods revealed spontaneous seizures in only one (11%) of nine rats that underwent PPS at age 21 days. However, all the surviving rats that underwent PPS for 30 min as adults or PPS for 8 h at 35 days (six of six) and eight (73%) of 11 surviving rats that underwent LiPC SE at 21 days demonstrated epochs of spontaneous seizure activity with high-frequency, high-amplitude EEG discharges.

Mossy fiber sprouting

Hippocampal plasticity as evidenced by sprouting of mossy fibers, the axons of dentate granule cells, is shown in the Timm-stained sections. A sham-stimulated rat at P35 shows no Timm stain in the inner molecular layer (IML) of the dentate. Only rare staining is visible in a rat that underwent 8 h of PPS at the age of P21 and was killed 3 months later (Fig. 2C). Clearly discernible IML staining is seen in sections representing mature animals that underwent LiPC SE at P21 (Fig. 2D), 8 h of PPS at P35, or 30 min of PPS as an adult. The group mean of the GVD scores for these rats (136, 111, and 129, respectively) is greater than that for those that underwent PPS at P21. The fraction of animals that developed chronic spontaneous seizures after each treatment is compared with the dentate inhibition scores and the GVD resulting from MFS in Fig. 2E.

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Our initial evaluation of SE-induced damage in developing rats involved determination of s-NSE. We were interested in this enzyme because of the published human data on adults after SE (35,36). Furthermore, recognizing the inherent difficultly in quantifying the total amount of neuronal injury in the brain, we used s-NSE as a biochemical correlate that takes into account differing regional vulnerabilities in the entire CNS, ignores the type of the injury itself (necrosis or apoptosis), and represents elevations corresponding to the appearance of neuronal injury.

The age-related damage seen in the rats with elevated s-NSE appeared to show particularly interesting patterns in the hippocampus. A detailed examination of the injury and also a study of substance P (SP) expression as a consequence of SE (37) revealed region-specific vulnerabilities that were age dependent and also paralleled changes in SP expression. By using multiple techniques sensitive to distinctive mechanisms of neuronal damage, we were able to demonstrate that some of these cells undergo apoptosis. With further maturation, apoptosis seemed quite rare, as indeed has been confirmed by Fujikawa et al. (38).

Some rats that had undergone LiPC SE at age 3 weeks began to display spontaneous seizures. These rats had demonstrated significant hippocampal hilar damage as well as damage to extrahippocampal structures such as the amygdala, thalamus, entorhinal cortex, and cortex in response to LiPC SE. In a separate set of studies using the PPS model of SE, we observed that similarly aged rats (P20) responded very differently, with even 8 h of stimulation rarely resulting in self-sustained SE (SSSE) (39). However, a significant number of P35 animals did so. The very fact that some of the P35 animals could tolerate 8 h of PPS without developing SSSE confirmed that even by this age (prepubescent to pubescent), synaptic maturation has not reached an adult level.

Some of the rats that underwent PPS at P35 did develop spontaneous seizures, and they had also shown more widespread damage than those that underwent PPS at P20. We hypothesized that the difference in the epileptogenic potential between LiPC SE and 8 h of PPS in the P21 rat may be attributable to differences in the recruitment of extrahippocampal circuits by these two methods. To demonstrate this difference, we undertook an immunohistochemical study for c-Jun, a transcription factor product of an immediate-early gene, c-jun, that may reflect early recruitment of epileptogenic areas preceding the onset of neuronal injury. We confirmed that PPS on a P21 pup resulted in activation restricted to the hippocampus, and extrahippocampal spread could be demonstrated well in this method only on the P35 or adult animals (40). In contrast, LiPC SE caused extensive hippocampal and extrahippocampal expression of c-Jun, consistent with widespread activation of circuits. Bertram (41,42) investigated the role of such extrahippocampal involvement and found the involvement of midline thalamic structures to be crucial in the development of limbic epilepsy.

We also wanted to examine for changes in plasticity in the epileptic state and chose both an anatomic marker, MFS, and a physiologic method, dentate inhibition as measured by population response to short interstimulus interval (40 ms) PPS of the perforant path. The most established anatomic method is one that involves the Timm stain for zinc, which is abundant in the mossy fiber terminals and synapses. Sprouting and neosynapse formation by mossy fibers is readily discerned by this technique. PPI, meanwhile, is an electrophysiologic correlate of the strength of recurrent inhibition in the dentate gyrus (43). Our studies demonstrated that the two phenomena, anatomic evidence of sprouting and enhanced inhibition to paired-pulse stimuli, correlated well with the evolution of an epileptic state.

In summary, our data suggest that several factors influence the development of an epileptic state. We propose a schematic with the central feature to be the condition of chronic epilepsy (CE) in Fig. 3. A key concept expressed by this diagram is the possibility that areas of overlap can vary with the relative contributions of each component, which in turn modify the area representing the epileptic state. No single component is likely responsible for CE by itself. For example, neurodegeneration alone cannot result in spontaneous seizures due to the presence of the silent period after limbic SE. Plastic changes (neurogenesis), meanwhile, have been seen in naïve mice placed in an enriched environment (44). Although numerous studies show an overlap of these two areas, a current study using older animals that undergo both these changes reported the incidence of CE to be almost certain, approximating 90% (see Mazarati et al. same issue). The stage of development of the brain is an important factor as well, but it appears that some provocations are more epileptogenic with advancing maturation (KA, LiPC, PPS), whereas others (hypoxia in P10–12 pups, CRH in P10–12 pups) may be more likely to cause injury and/or lead to plasticity, resulting in epileptogenicity in younger animals.

image

Figure 3. Venn diagram of relations between acute, chronic, and maturational issues resulting in epileptogenesis.

Download figure to PowerPoint

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES