Prevention or modification of epileptogenesis after brain insults: experimental approaches and translational research

Abstract

Diverse brain insults, including traumatic brain injury, stroke, infections, tumors, neurodegenerative diseases, and prolonged acute symptomatic seizures, such as complex febrile seizures or status epilepticus (SE), can induce "epileptogenesis," a process by which normal brain tissue is transformed into tissue capable of generating spontaneous recurrent seizures. Furthermore, epileptogenesis operates in cryptogenic causes of epilepsy. In view of the accumulating information about cellular and molecular mechanisms of epileptogenesis, it should be possible to intervene in this process before the onset of seizures and thereby either prevent the development of epilepsy in patients at risk or increase the potential for better long-term outcome, which constitutes a major clinical need. For identifying pharmacological interventions that prevent, interrupt or reverse the epileptogenic process in people at risk, two groups of animal models, kindling and SE-induced recurrent seizures, have been recommended as potentially useful tools. Furthermore, genetic rodent models of epileptogenesis are increasingly used in assessing antiepileptogenic treatments. Two approaches have been used in these different model categories: screening of clinically established antiepileptic drugs (AEDs) for antiepileptogenic or disease-modifying potential, and targeting the key causal mechanisms that underlie epileptogenesis. The first approach indicated that among various AEDs, topiramate, levetiracetam, carisbamate, and valproate may be the most promising. On the basis of these experimental findings, two ongoing clinical trials will address the antiepileptogenic potential of topiramate and levetiracetam in patients with traumatic brain injury, hopefully translating laboratory discoveries into successful therapies. The second approach has highlighted neurodegeneration, inflammation and up-regulation of immune responses, and neuronal hyperexcitability as potential targets for antiepileptogenesis or disease modification. This article reviews these areas of progress and discusses the challenges associated with discovery of antiepileptogenic therapies.

Fig. 1.

Steps in the development and progression of temporal lobe epilepsy and possible therapeutic interventions. The term epileptogenesis includes processes that take place before the first spontaneous seizure occurs to render the epileptic brain susceptible to spontaneous recurrent seizures and processes that intensify seizures and make them more refractory to therapy (progression). It is important to note that the concept of a multistep process of epileptogenesis illustrated in this figure bears similarities to the multistep process of carcinogenesis with initiation (DNA damage), repair of damage or failure to repair, promotion to tumor, and progression to malignancy and metastasis (Löscher and Liburdy, 1998). See section II for further explanation and discussion. [Adapted from Löscher W, Gernert M, and Heinemann U (2008) Cell and gene therapies in epilepsy—promising avenues or blind alleys? Trends Neurosci 31:62–73. Copyright © 2008 Elsevier Science. Used with permission.]

Fig. 2.

Schematic illustration of an experimental protocol to evaluate drug effects on kindling acquisition. Note that three categories of drug effects are analyzed: 1) drug is administered before each stimulation and the effects on kindling acquisition are determined relative to vehicle controls; 2) kindling is continued after washout of drug; 3) anticonvulsant drug effects are studied in fully kindled rats.

Fig. 3.

Schematic illustration of an experimental protocol to evaluate antiepileptogenic (or disease-modifying) drug effects by prophylactic drug treatment after a status epilepticus.

Fig. 4.

Lack of direct relationship between loss of dentate hilus neurons and development of spontaneous recurrent seizures (SRS) in three rat models of temporal lobe epilepsy. In all models, the occurrence of SRS was monitored in the chronic epileptic phase, and rats with or without observed SRS were differentiated. Hilus neurons were quantified with counting frames in serial sections, using stereological methods (see Brandt et al., 2003b for details). Sham controls were used for comparison. Each symbol illustrates the neuronal density in the hilus of one rat. The median of the individual data are indicated by horizontal line. In A, SE was induced by systemic administration of kainate (10 mg/kg i.p.) and terminated after 90 min by diazepam. Six rats were treated with 0.1 mg/kg MK-801 immediately after diazepam. Data are from six sham controls, seven rats with SE plus vehicle, and six rats with SE plus MK-801. Analysis of data by nonparametric analysis of variance (Kruskal-Wallis test) indicated a significant difference between means (P = 0.0082). Post hoc analysis by Dunn's test indicated that only the SE-vehicle rats differed significantly from controls (P < 0.01), suggesting a neuroprotective effect of MK-801. However, all except one of the MK-801-treated rats developed SRS. Data were reanalyzed from the study of Brandt et al. (2003a). In B, SE was induced by sustained electrical stimulation of the BLA. Data are from 6 sham controls and 26 SE rats (18 with SRS and 8 without observed SRS). The asterisk indicates a significant difference between the two groups (P = 0.0003). When SE rats with SRS (median neuronal density 4223 neurons/mm³) and without SRS (8042 neurons/mm³) were compared with controls (10,598 neurons/mm³), only the group with SRS differed significantly from controls (P < 0.001). However, note that several rats with SRS had neuronal densities within control range, indicating no direct relationship between hilar cell loss and development of SRS. Data are from the study of Brandt et al. (2003b) and unpublished experiments. In C, rats were kindled via the BLA and then further stimulated twice daily for up to approximately 280 stimulations (“overkindling”) until SRS were observed in approximately 50% of rats. Data are from 10 sham controls and 21 overkindled rats (10 with SRS and 11 without observed SRS). The asterisk indicates a significant difference between the two groups (P = 0.0011). When overkindled rats with SRS (median neuronal density 6294 neurons/mm³) and without SRS (7693 neurons/mm³) were compared with controls (10,371 neurons/mm³), only the group with SRS differed significantly from controls (P < 0.01). Data were reanalyzed from the study of Brandt et al. (2004).

Fig. 5.

Illustration of a dose-finding experiment for an experimental trial with the COX-2 inhibitor parecoxib (Pcb) in the pilocarpine model of TLE in rats. A lithium-pilocarpine–induced SE was terminated after 60 min by diazepam plus phenobarbital. Twenty-four hours after SE induction, PGE₂ levels were significantly increased in hippocampus, piriform cortex, amygdala, and frontal cortex. Additional groups of rats were treated with Pcb at either 1 or 5 mg/kg. Pcb was administered 1, 7, and 22 h after onset of SE. Both dosing protocols completely prevented the increase in PGE₂ after SE, thus allowing rational selection of a dosing protocol for an epilepsy prevention trial. Data are means ± S.E.M. of four to six rats per group; statistical analysis was performed by analysis of variance with post hoc Bonferroni test (*, P < 0.05). Data are from unpublished experiments (N. Polascheck, M. Bankstahl, W. Löscher).