|Home | About | Journals | Submit | Contact Us | Français|
Estrogens are essential for normal brain functions. The effects of estrogens on seizures are contradictory. More studies are necessary to determine under which conditions the estrogens have proconvulsant effects and when the estrogens may have beneficial action in patients with epilepsy.
Catamenial epilepsy is a condition during which seizures tend to cluster in relevance to menstrual cycle in some women with epilepsy (for detailed recent review see (Reddy 2009)). Herzog and colleagues recognized three patterns of catamenial epilepsy based on the higher seizure occurrence during the individual phases of the menstrual cycle: perimenstrual, periovulatory, and inadequate luteal phase patterns (Herzog, et al. 1997). The changes in seizure occurrence in distinct phases of the catamenial epilepsy have been notoriously simplified to the changes in the female sex hormone levels with estrogens increasing and progesterone suppressing neuronal excitability. Interestingly, most women with catamenial epilepsy report worsening of seizures during the perimenstrual phase. This phase is characterized by rapid decline in the levels of both hormones (estrogen and progesterone), as this is the signal for the initiation of menstruation. Low levels of estrogen and progesterone are the hallmark of this phase. Moreover, linking the catamenial epilepsy solely to estrogens and progesterone is an oversimplification since the underlying mechanisms for the increased neuronal excitability during the distinct phases of the menstrual cycle are clearly multifactorial (Reddy 2009). On the other hand, despite the argument whether the sex hormone contribution is a direct effect of sex hormones on neuronal receptor systems or an indirect effect as a result of hormone-induced changes in gene expression, the involvement of estrogens and progesterone in regulation of neuronal excitability is undisputable. We have been interested in determining the effects of estrogens on seizures and neuronal excitability, which represent the main focus of this manuscript.
Although it has been believed for many decades that estrogens induce excitatory actions in the CNS, recent studies show that the effects of estrogens can be both excitatory as well as inhibitory [for review see (Velíšková 2006, Velíšková 2007)]. This conclusion is supported by clinical and experimental data. Logothetis injected high doses of estrogen to women and such treatment induced rapid interictal epileptiform activity and exacerbated seizures during the premenstrual phase (Logothetis, et al. 1959). Later, Backström described a positive relationship between estrogen levels and seizure exacerbation (Backström 1976). However, in women with primary generalized epilepsy, at the peak of estrogen levels during ovarian cycle, the seizure rate was decreased (Jacono and Robertson 1987). Similarly, estrogen replacement therapy improved seizure control during menopause in some women (Harden, et al. 1999, Peebles, et al. 2000). In a pilot study, administration of estrogen did not exacerbate EEG ictal or interictal activity in any of the six tested women with epilepsy (Janoušek, et al. 2006). Most importantly, there is no clear evidence that oral contraceptives increase the risk of seizures in women with epilepsy.
In the animal models, topical administration of conjugated estrogens on the cortex produces epileptiform discharges (Marcus, et al. 1966) but topical application of β-estradiol on the cortex does not evoke any epileptiform discharges (Marcus, et al. 1966). Similarly, acute administration of estrogen has proconvulsant effects in different seizure models (Edwards, et al. 1999, Nicoletti, et al. 1985, Woolley and Timiras 1962) but chronic administration has no effects or may be even anticonvulsant (Kalkbrenner and Standley 2003, Reibel, et al. 2000, Tominaga, et al. 2001, Velíšková, et al. 2000). The differences in the effects of estrogens on seizures seem to be related, but not restricted, to the dose, treatment duration, estrogen species, seizure type/model, neurotransmitter system involved, sex, and interaction with progesterone.
To determine the effect of estrogens on seizures, we used β-estradiol (EB) administration in ovariectomized adult female rats. Following one week after ovariectomy, the animals were injected with oil (sterile peanut = controls) or EB. We compared different seizure models, different doses of EB, and different duration of EB administration (a single vs. multiple injections) using in vivo as well as in vitro approaches.
In vivo experiments: When using a low dose of EB (2 μg per rat per day), chronic EB administration delayed onset of clonic seizures induced by kainic acid or picrotoxin compared to oil controls (Table). We have also found that the dose of 2 μg of EB prevented mortality following status epilepticus (Velíšková 2007). This protective effect of EB was not related to its anticonvulsant effect since the treatment with EB did not prevent the progression of individual seizures to status epilepticus, its onset, severity, or duration. On the other hand, chronic EB administration did not affect the threshold to flurothyl- or lithium/pilocarpine-induced clonic seizures (Table 1). Finally, EB administration to gonadally intact females accelerated the onset of kainic acid-induced clonic seizures (Velíšková, 2007). Whether the effect was because of EB interaction with progesterone or simply because of supraphysiological levels of circulating estrogens remains to be determined.
Next we tested the effect of treatment duration. We compared a single EB injection with chronic EB administration (4 days). We used the picrotoxin-induced seizures since EB had an anticonvulsant effect in this seizure model. We found that a single EB dose (24 hours prior to seizure testing) had no effect on the onset of picrotoxin-induced seizures, which was in the contrast to the anticonvulsant effect of the paradigm of chronic EB administration (Figure 1).
In vitro experiments: We examined the effect of chronic EB administration (2 μg per rat per day) on the onset of epileptiform-like discharges in the CA1 and entorhinal cortex induced by bath application with [Mg2+]o=0 ACSF. There were no differences in the onset of epileptiform-like activity in CA1 or entorhinal cortex in slices from EB treated rats compared to oil controls (Velíšek, et al. 1999). Interestingly, however, we found that progesterone administration significantly delayed the onset of the discharges in the CA1 region (i.e., had an anticonvulsant effect) but profound proconvulsant effect in the entorhinal cortex (Velíšek, et al. 1999), illustrating that the effects of sex hormones on neuronal excitability depend on the structure of action.
Next we examined the effects of EB on paired pulse inhibition and facilitation in the dentate gyrus granule cells using stimulation of mixed perforant path. A single dose of EB 24 hours prior to the slice preparation did not change the paired pulse curve compared to oil-injected controls (Velíšková and Veliíšek 2007). Following the chronic EB administration, the early inhibition or intermediate facilitation of the paired pulse paradigm also were not affected. However, EB significantly enhanced the late inhibition of the paired pulse paradigm occurring between the interstimulus intervals 150–1000 ms (Velíšková and Veliíšek 2007). The EB-induced enhancement of the late inhibition could have been blocked by a neuropeptide Y-1 (NPY-Y1) receptor antagonist at the interstimulus intervals 150–300 ms and by a general metabotropic glutamate receptor antagonist (LY 341495 in μM concentrations) between the intervals 300–1000 ms (Velíšková and Veliíšek 2007). We have also demonstrated that a burst activity produced by electrical stimulation of mixed perforant path was inhibited in slices from EB-treated rats only when the frequency of the stimuli inducing this burst activity was 5 Hz but not 100 Hz. These findings further confirm that following EB administration there is an enhanced dentate gyrus gating for incoming activity within theta frequencies. Enhancement of synaptic inhibition at theta frequencies seems to play an important role in propagation of kainic acid-induced epileptiform activity from the entorhinal cortex into the hippocampus as the epileptiform activity involves increases in cortical neuronal activity at 2–5 Hz frequencies (Medvedev et al., 2000). Furthermore, in patients with temporal lobe epilepsy with the hypersynchronous seizure onset, deep electrode recordings from the hippocampus show increases in frequency of interictal spikes with spike-and-wave shape (corresponding to frequency of 3 Hz) during the very initial stage of a seizure (Engel, 1989). Thus, it is very likely that the enhanced dentate filtering of such frequencies leads to disruption of seizure propagation.
It is noteworthy to mention here that we also demonstrated that EB administration prior to status epilepticus significantly reduces seizure-induced damage of the vulnerable NPY neuronal population in the hilar region compared to oil controls experiencing similar seizure severity and duration (Velíšková 2007). The neuroprotective effect of EB can be blocked by intracerebroventricular administration of anti-NPY antibody suggesting a critical role of NPY in the seizure-induced damage and interactions between the EB and NPY system in the dentate gyrus (Velíšková 2007).
To summarize our data, here we attempt to explain the discrepancies in the reported data on the effects of estrogens on neuronal excitability. We show that the EB effects depend on the seizure model. While in some seizure models, chronic EB administration inducing EB levels within the physiological range leads to delayed seizure onset, in other seizure models EB does not affect or may even have a proconvulsant effect. We also demonstrate here that the EB-induced anticonvulsant effects seem to involve regulation of gene expression since the beneficial effects depend on treatment duration. In vitro studies show that the increased inhibition of granule cells following the chronic EB administration allows for better filtering of incoming seizure activity (especially between 3-5Hz) from the entorhinal cortex into the hippocampus and this feature may assist in the inhibition of seizure propagation. The EB-induced enhancement of the dentate gyrus filtering involves activation of NPY-Y1 and metabotropic glutamate receptors. Finally, estradiol within the physiological concentration range can protect hippocampal neurons against seizure-induced damage. These neuroprotective effects seem to involve transcriptional regulation of NPY.
In conclusion, we show that many effects of estrogens on seizures are secondary to their transcriptional action. Estrogens closely interact with molecules, which belong to the endogenous seizure-protecting mechanisms such as the NPY or metabotropic glutamate receptor systems and thus, their effects are essential for normal neuronal function.
The work has been supported in part by grants NS 056093, NS059504 and NS-20253 from NINDS and by the Harlem Children Society. The authors would like to thank Ms. Zunju Hu for excellent technical assistance.
Conflict of interest disclosure: None of the authors have any conflict of interest.