A Rat Model of Epilepsy in Women: A Tool to Study Physiological Interactions between Endocrine Systems and Seizures

Abstract

Epilepsy in women is influenced by endocrine status and antiepileptic drugs, but without an animal model, the effects of endocrine variables and antiepileptic drugs cannot be easily dissociated from the influence of epilepsy itself. Animal models have had limited utility because experimentally induced seizures typically result in reproductive failure. This study was conducted to develop an improved animal model. The muscarinic convulsant pilocarpine was used to elicit status epilepticus (SE) in adult female Sprague Dawley rats. The selective estrogen receptor modulator raloxifene was administered 30 min before pilocarpine. An anticonvulsant barbiturate, pentobarbital, was injected 5–10 min after the onset of SE and at least once thereafter to minimize acute convulsions. Mortality, morbidity, estrous cyclicity, and the ultimate success of the procedure (i.e. induction of recurrent, spontaneous seizures) were monitored. The combination of raloxifene and pentobarbital led to significantly improved estrous cyclicity compared with previous methods. Animals treated with raloxifene and pentobarbital became epileptic, as defined by the recurrence of spontaneous convulsions in the weeks after SE. The results of this study provide an improved animal model to examine the interactions between seizures and ovarian hormone secretion. The results also suggest that treatment of SE with raloxifene may benefit women with SE.

In adult female rats, estrous cyclicity is preserved in the pilocarpine model of epilepsy if raloxifene is administered prior to pilocarpine, and pentobarbital is used rapidly after pilocarpine-induced status epilepticus.

Catamenial epilepsy, defined by seizures associated with specific stages of the menstrual cycle, has been extensively studied (1,2,3,4). The etiology of this disorder, however, remains poorly understood (5,6). A major problem is the lack of an animal model because current animal models of epilepsy usually disrupt estrous cycles. Thus, in kindled rats, ovarian cyclicity ceases once a stimulus elicits a consistent seizure (7). If status epilepticus (SE) is used to induce an epileptic state, many animals do not survive, and the survivors typically do not maintain normal estrous cycles (8,9,10). Therefore, the relationship between ovarian secretion of estradiol and progesterone and spontaneous seizures cannot be explored, which impedes research relevant to women with epilepsy.

Anticonvulsants such as benzodiazepines can be used 1–3 h after the onset of SE to reduce mortality and morbidity associated with SE in males (11,12,13,14). However, in females, this procedure does not preserve reproductive function reliably (15). In the present study, we asked whether a more aggressive, multifactorial approach might work better. Raloxifene, a selective estrogen receptor modulator, has been shown to protect the hippocampus from the excitotoxic effects of kainic acid (16). We reasoned that raloxifene might be protective against pilocarpine-induced damage to hypothalamic centers controlling reproductive function. We also attempted to decrease SE severity more than preexisting methods, by administering a strong anticonvulsant (pentobarbital), rapidly after the onset of SE. The results demonstrate that mortality and reproductive function after SE are improved, and epileptogenesis (development of recurrent spontaneous convulsive seizures) still occurs, providing an improved model to study interactions between epilepsy and ovarian function.

Materials and Methods

Subjects

Adult Sprague Dawley rats were obtained from Taconic Farms (Germantown, NY) or Charles River Laboratories (Wilmington, MA). Rats were housed two to three per cage at 64–75 F and 40–75% relative humidity, in standard opaque cages with corn cob bedding and a 12-h light, 12-h dark cycle (lights on, 0700 h). Food (Purina 5001 chow; WF Fisher, Somerville, NJ) and water were provided ad libitum. Cages with adult male rats (one to two per cage) were placed between cages of female rats to facilitate regular estrous cycles (17).

All procedures were approved by the Institutional Animal Care and Use Committee and met the guidelines of the National Institutes of Health. Chemicals were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise specified.

Vaginal cytology

Starting at 50 d of age, female rats were examined daily between approximately 0900 and 1200 h to assess the relative frequency of cell types in the vaginal epithelium in a 30-μl sample by vaginal lavage (18). Animals that failed to show three consecutive 4-d cycles, approximately 10% of rats, were excluded from the study. Animals with 5-d cycles were not included because our animals did not exhibit 5-d estrous cycles until they approached reproductive senescence.

The pattern of vaginal cytology used to define a 4-d estrous cycle was based on previous evaluations of serum estradiol and progesterone levels in normal female rats (18). The pattern of vaginal cytology that correlated with normal hormone values was a 2-d period when epithelial or cornified epithelial cells dominated the vaginal sample (the proestrus-estrus transition), followed by a 2-d period when leukocytes were dominant (diestrus).

An acyclic pattern was defined by at least two concurrent estrous cycles during which this 4-d pattern was absent. In the majority of cases, vaginal cytology was dominated by leukocytes for prolonged periods of time (persistent diestrus). More rarely, animals entered persistent estrus, characterized by cornified cells in the vaginal fluid (7,19). In a few cases, a third type of acyclicity was observed, characterized by random fluctuations between diestrus and estrus cell types. Because the basis of these different acyclic patterns in the epileptic female rat is not yet understood, animals were characterized only as cyclic or acyclic in Results.

Drug administration

Animals were housed two to three per cage and brought to a separate room, in which they were isolated into individual cages with fresh bedding. They were allowed to acclimate for 30–60 min, and then atropine methylbromide (1 mg/kg, sc) was injected. Thirty minutes later pilocarpine hydrochloride (350 mg/kg, sc) was injected. SE was defined as a stage 4–5 seizure that was followed by continuous tonic-clonic movements of the torso, head, and tail. The onset of SE was defined as the time of that stage 4–5 seizure.

After SE, diazepam (10 mg/kg, ip; Henry Schein, Melville, NY) was injected in some animals at 1 h after the stage 4–5 seizure that defined its onset. In others, pentobarbital sodium (30 mg/kg, ip, diluted in distilled water or Nembutal; Henry Schein) was injected 5–10 min after the onset of SE. Subsequent doses of pentobarbital (up to three doses total) were 5 or 10 mg/kg, ip. The additional doses were administered 30–60 min after the previous dose if animals continued to exhibit convulsions that involved the entire body. If convulsions were severe (animals rose and fell), 10 mg/kg was administered. If convulsions were not as severe (animals remained prone), 5 mg/kg were used. Once convulsions were reduced to the point that only the head was involved, no further doses of pentobarbital were administered. The effects of pentobarbital sodium could not be discriminated from Nembutal, and total dose was comparable per animal (range 30–50 mg/kg), but pentobarbital sodium was more difficult to maintain in solution, making it a less desirable choice.

Rats were treated with raloxifene hydrochloride (1 or 5 mg/kg, sc) or vehicle (dimethyl sulfoxide, sc) 30 min before pilocarpine administration. Doses were chosen based on the literature describing their effects in vivo (16,20). Higher doses than 5 mg/kg were not tested because of their poor solubility and indications that high doses did not increase protection, at least in vitro (21).

After SE, dextrose-lactate Ringer’s solution (Henry Schein) was provided by mouth (∼1 ml) and injected sc (2.5 ml) during the time when diazepam- or pentobarbital-induced anesthesia was maximal. For the subsequent days, animals were housed two to three per cage, with many portions of apple placed at the base of the cage. Food pellets, moistened with water, were placed at the base of the cage. Approximately 1 ml dextrose-lactate Ringer’s solution was provided by mouth twice per day. All animals were fed, handled, and housed in a similar manner.

Animals were observed in their home cage from 0800–1800 h each day for the duration of the study, and vaginal lavages was conducted each day between 0900 and 2400 h. Observations were made for 5–10 min throughout the day, and seizures were scored using the Racine scale (22). Because of the difficulty in discriminating normal movement from stages 1–2 on this scale, only severe seizures (stages 3–5) were scored. Stage 3 seizures were defined by unilateral forelimb clonus, with rearing; stage 4 was defined by bilateral forelimb clonus with rearing; and stage 5 was defined as a stage 4 seizure that was followed by loss of posture (22).

Data analysis

Statistical comparisons were made using Quantitative skills statistics software (Qsetup, version 4.0.0.3; Pantaray Software Systems, Hanegev, Israel). Means ± ses are reported. P criterion was 0.05 (two tailed).

Results

Effects of raloxifene pretreatment on SE and estrous cyclicity

Subjects

Adult female rats (aged 66–89 d, 255–280 g) were administered pilocarpine. Animals were chosen without regard for the stage of the estrous cycle because a previous study demonstrated no differences in mortality or morbidity among animals that were injected at different stages of the estrous cycle (15,23).

Mortality and morbidity

Mortality and morbidity after the onset of SE was evaluated by comparing three groups: animals that were pretreated with raloxifene (1 or 5 mg/kg) or vehicle. In both the vehicle-treated and raloxifene-treated animals, diazepam (10 mg/kg) was administered ip. Mortality was defined as animals that died during SE, typically because of respiratory failure during a severe tonic-clonic seizure or animals that were found dead in their cage within 48 h after the onset of SE. Morbidity was defined as animals that survived but did not move often, groom, eat, or drink on their own for 48 h after SE. A longer time was not used to judge these animals because preliminary studies showed that the females that exhibit this behavior do not recover except in rare instances, even if aggressive intervention is instigated (hydration by injection and manual oral delivery, handling, and additional nourishment).

Morbidity was reduced in raloxifene-pretreated rats compared with vehicle treatment (Table 1). There was no statistical difference between the effects of the two doses of raloxifene (Table 1).

Table 1.

Effects of raloxifene pretreatment on mortality, morbidity, and cyclicity of SE

	Control + Dzp		Raloxifene + Dzp
	Untreated	Vehicle	1 mg	5 mg	Total	P value
Mortality
n	6/34	2/15	0/12	0/7	0/19	0.49198
Percent	18	13	0	0	0
Morbidity
n	16/34	8/15	1/12	1/7	2/19	0.03004a
Percent	47	53	8	14	10
Total mortality and morbidity
n	22/34	10/15	1/12	1/7	2/19	0.00304a
Percent	65	67	8	14	10
Cyclicity
n	7/12	2/5	4/8	3/5	7/13	0.85603
Percent	58	20	50	60	54

Effects of raloxifene pretreatment on mortality, morbidity, and cyclicity following pilocarpine-induced SE in adult female rats. Total refers to the pooled data from both doses of raloxifene. For statistical comparisons, a two-tailed 2 × 3 exact test was used, and the groups were vehicle pretreatment, 1 mg/kg raloxifene, and 5 mg/kg raloxifene. Data from untreated rats that were evaluated previously (15) and illustrate similar findings are also shown (untreated) but were not included in statistical comparisons. Mortality after SE as well as total mortality was reduced by raloxifene pretreatment, but cyclicity was not affected significantly. Dzp, Diazepam. 

^aStatistical significance. Due to the low numbers of surviving animals, the statistical comparisons of cyclicity were also conducted by pooling raloxifene-pretreated animals (both doses) and comparing the pooled data to vehicle-pretreated animals, and this approach also did not demonstrate statistical significance (P = 0.6471). 

Although the incidence of SE was not significantly different between rats that were injected with raloxifene or vehicle and there was no difference in the latency to the first stage 4–5 seizure, the number of stage 4–5 seizures that occurred between pilocarpine injection and the onset of SE was reduced by raloxifene (Fig. 1). The latency to SE was delayed by raloxifene but only at the higher dose (5 mg/kg; Fig. 1).

Figure 1.

Effects of raloxifene on SE. A, There was no significant difference in the incidence of SE in rats pretreated with vehicle (Veh), raloxifene (Ral; 1 or 5 mg/kg) before pilocarpine (2 × 3 exact test, two-tailed P = 0.8295). Sample sizes are listed above each bar. B, There was a significant decrease in the number of stage 4–5 seizures that occurred before SE in the three groups (one way ANOVA followed by Tukey post hoc tests, P < 0.05 for vehicle vs. 1 mg/kg and P < 0.05 for vehicle vs. 5 mg/kg). Sample size is listed at the base of each bar. C, There was no significant difference in the latency to the first stage 4–5 seizure in the three treatment groups (one way ANOVA followed by Tukey post hoc tests; P > 0.05 for all post hoc comparisons). D, There was a significant increase in the latency to SE after raloxifene treatment but only at the highest dose (one way ANOVA followed by Tukey post hoc tests, P > 0.05 for vehicle vs. 1 mg/kg raloxifene; P < 0.05 for vehicle vs. 5 mg/kg raloxifene). Asterisk indicates statistical significance.

Estrous cyclicity

Animals that were pretreated with raloxifene and then experienced SE, followed by diazepam administration, were compared with another group treated identically, except vehicle was used instead of raloxifene. Acyclicity developed in raloxifene-treated rats at a frequency similar to vehicle-treated animals (Table 1).

Effects of raloxifene and pentobarbital on mortality and estrous cyclicity

Rats pretreated with 1 mg/kg raloxifene were administered pentobarbital after SE. Pentobarbital (30 mg/kg, ip) was administered 5–10 min after the onset of SE. Earlier times were not used because the definition of SE requires 5 min to ensure that seizures do not terminate. Subsequent doses (see Materials and Methods) were injected until tonic-clonic movements were limited to the head, i.e. the torso and tail no longer exhibited tonic-clonic movements.

Doses of pentobarbital were chosen based on pilot studies showing that higher doses produced respiratory depression. Other anesthetics (chloral hydrate, phenobarbital) were not used because studies in male rats showed they were not more effective (24). Similarly, multiple doses of diazepam were not used because preliminary studies in four animals showed that multiple doses did not reduce motor behaviors during SE as well as pentobarbital, and ovarian cyclicity was lost in subsequent weeks.

Morbidity and the ability to sustain regular estrous cycles were greatly improved after raloxifene and pentobarbital, compared with raloxifene and diazepam (Table 2 and Fig. 2). Additional experiments using 2 mg/kg (n = 5) or 5 mg/kg (n = 2) raloxifene and pentobarbital demonstrated similar effects: the animals survived SE (n = 7/7) and most animals demonstrated regular estrous cycles (n = 5/7) for up to 3 months.

Table 2.

Effects of raloxifene and pentobarbital on mortality, morbidity, and cyclicity after SE

	Veh+Dzp	Ral+Dzp	Ral+Pb	P value
Mortality
n	2/15	0/12	0/12	0.17814
Percent	13	0	0
Morbidity
n	8/15	1/12	2/12	0.03017a
Percent	53	8	16
Total, mortality and morbidity
n	10/15	1/12	2/12	0.00214a
Percent	67	8	16
Cyclicity
n	2/5	3/11	9/10	0.01006a
Percent	20	33	90

Comparison of treatments to modify effects of SE in female rats. Animals were either vehicle pretreated and injected with diazepam (Dzp) after SE began or were raloxifene (Ral) pretreated and injected with diazepam after SE began (Ral+Dzp) or were raloxifene pretreated and injected with pentobarbital after the onset of SE (Ral+Pb). Raloxifene dose was 1 mg/kg for all groups. Statistical comparisons demonstrated that there were differences in the three experimental groups for morbidity, total mortality and morbidity (pooled), and cyclicity (whether animals maintained estrous cycles for at least 3 wk after SE). Statistical comparisons used 2 × 3 exact tests and two-tailed P values are reported. 

^aStatistical significance. 

Figure 2.

Effects of raloxifene and pilocarpine on outcome after SE. A, Raloxifene (1 mg/kg) pretreatment before pilocarpine (Ral) or raloxifene pretreatment with pentobarbital posttreatment after SE (Ral+Pb) decreased mortality and morbidity relative to controls (Con; vehicle pretreatment and diazepam after SE). The total mortality and morbidity after SE (Table 2) is shown. Sample sizes are listed at the base of each bar. B, Raloxifene did not increase estrous cyclicity after SE compared with control, but raloxifene and pentobarbital significantly increased the percent of animals that were cyclic after SE. For statistical comparisons, see Table 2. Asterisk indicates statistical significance.

In all animals that had SE after treatment with raloxifene (1 or 5 mg/kg) plus pentobarbital, motor seizures were observed in the subsequent 2 months. Each animal had at least two spontaneous seizures that reached stage 3. Two animals that did not exhibit regular estrous cycles developed severe spontaneous seizures (stage 5), which are associated with acyclicity in the kindling model (7).

Discussion

These results demonstrate that simple modifications to a common method to induce epileptogenesis in rodents can allow the method to be used in females without loss of ovarian cyclicity. The modification leads to reduced morbidity, improving the approach. Furthermore, the modifications allow regular estrous cycles to continue as spontaneous seizures develop, although seizures are not usually severe. This model is advantageous because it potentially allows studies of the interactions between the female reproductive cycle and epileptic seizures to be pursued. Many authors have hypothesized that cyclical changes in estradiol or progesterone are related to epileptic seizures (3,4,25,26,27,28). Testing these hypotheses has, however, been stymied by the lack of an animal model of epilepsy in cycling females. Such an animal model will be useful not only to evaluate hypotheses about mechanisms of hormone-sensitive seizures but also for preclinical studies evaluating anticonvulsant efficacy and effects on reproductive function.

Although the combination of raloxifene and pentobarbital improved estrous cyclicity compared with raloxifene and diazepam, it is not clear whether pentobarbital, used by itself, would be just as effective as the combination. In pilot studies, females that were treated only with pentobarbital after SE, i.e. without raloxifene pretreatment, exhibited minimal grooming, movement, food intake, and drinking in the 48 h after SE. Therefore, the combination of both raloxifene and pentobarbital appears to be important to meet the objective, inducing epilepsy, but limiting distress, and preserving reproductive function.

Raloxifene pretreatment

The results show that raloxifene has a protective role in reducing morbidity after pilocarpine-induced SE. These results are consistent with other studies that have shown that raloxifene is neuroprotective (29,30). However, no other study, to our knowledge, has suggested that morbidity can be reduced by pretreatment with raloxifene. Morbidity appears to be associated with cessation of food and water intake that does not recovery spontaneously. The latter condition could be due to a dysregulation of the homeostatic mechanisms controlling food and water intake, suggesting a role for the hypothalamus in which these behaviors are controlled, and damage after seizures has been shown (31,32).

Raloxifene may also have exerted an anticonvulsant effect, reducing the severity of SE, because some indices of SE were influenced by raloxifene. Specifically, raloxifene decreased the numbers of seizures before the onset of SE and increased the latency to SE, although the latter occurred only after administration of the highest dose. One could argue that the reduction in the number of stage 4–5 seizures before SE is not necessarily a reflection of an anticonvulsant action, because SE ultimately developed, but there was an increased delay to SE, suggesting a weak anticonvulsant effect. One mechanism to explain the delay in SE is that raloxifene may have protected the blood brain barrier, which is more vulnerable in the female than the male (33), thereby slowing access of pilocarpine to the brain. Finally, raloxifene may have also modulated muscarinic receptor activation by pilocarpine. Estradiol induces muscarinic receptors in the hypothalamus nuclei and suppresses appetite (34). Therefore, blockade of estrogen receptors may help to reduce the dysregulation of hypothalamic function induced by pilocarpine. Further studies using electroencephalograms would help define the interaction between raloxifene and SE.

Whatever the mechanisms, these results suggest that raloxifene might be a useful alternative for hormone replacement in postmenopausal women with epilepsy, in which estrogens are often avoided because they are considered proconvulsant. Raloxifene is already used to prevent osteoporosis in normal postmenopausal women because it is estrogenic in bone but decreases the risk of cancer in the breast and reproductive tract (35,36). Our data suggest that raloxifene should be evaluated for women with epilepsy.