Elsevier

Epilepsy Research

Volume 92, Issues 2–3, December 2010, Pages 163-169
Epilepsy Research

Comparative anticonvulsant efficacy in the corneal kindled mouse model of partial epilepsy: Correlation with other seizure and epilepsy models

https://doi.org/10.1016/j.eplepsyres.2010.09.002Get rights and content

Summary

Chronic electrical stimulation via corneal electrodes can rapidly yield large numbers of kindled mice with a seizure phenotype reflective of secondarily generalized partial seizures. The corneal kindled mouse model has been found to be a highly sensitive and efficient screening model for antiepileptic drug (AED) discovery. The present study further evaluates the utility of the corneal kindled mouse model as a tool for rapid screening of investigational AEDs. Results obtained with nine AEDs (valproic acid, lamotrigine, phenytoin, carbamazepine, levetiracetam, vigabatrin, topiramate, tiagabine, and ezogabine) with varying mechanisms of action and clinical spectrums, as well as six investigational compounds were evaluated in the corneal kindled mouse. ED50 values are compared to those obtained in the hippocampal kindled rat, the mouse maximal electroshock (MES) model, the 6 Hz partial psychomotor seizure model, and the subcutaneous pentylenetetrazol (scPTZ) test. The results obtained in the corneal kindled mouse demonstrate a positive correlation with those attained employing established preclinical models: MES (r2 = 0.9511), scPTZ (r2 = 0.9697), 6 Hz (r2 = 0.9519), and hippocampal kindling (r2 = 0.9037). The demonstrated predictive ability of the corneal kindled mouse model supports its use in the early evaluation of investigational AEDs.

Introduction

There are a number of animal models of evoked seizures and epilepsy currently employed in the search for new antiseizure drugs; however, no one model can successfully predict efficacy for every new AED. For this reason, a diverse set of preclinical models is incorporated in the initial screening of candidate compounds (White et al., 2002). The maximal electroshock (MES) seizure model, first utilized in 1937 (Putnam and Merritt, 1937), led to the discovery of the anticonvulsant properties of phenytoin (PHT). Following this discovery, the MES test became established as one of the primary preclinical screening tools utilized in the search for new AEDs. Other acute screening tools commonly used include the subcutaneous pentylenetetrazol, also known as metrazol, test (scPTZ or scMet) and the 6 Hz partial ‘psychomotor’ seizure test (White et al., 2002). Unfortunately, clinical results have challenged the predictive nature of both the MES and the scPTZ acute seizure models. For example, levetiracetam (LEV) was found to be inactive in the MES and scPTZ screens. In contrast, LEV was active in the rat amygdala kindling seizure model (Löscher et al., 1998), a chronic model of epilepsy. LEV is currently a clinically utilized AED with a novel mechanism of action; i.e., it binds with high affinity to a synaptic vesicle protein (SV2A) (Lynch et al., 2004). The inability of the primary screens to identify the anticonvulsant efficacy of LEV raised significant concerns regarding their capacity to identify drugs with novel mechanisms of action. The MES screen also failed to detect the efficacy of tiagabine (TGB) and vigabatrin (VGB), two clinically efficacious AEDs that work to increase synaptic GABA levels, albeit via different mechanisms. Furthermore, the MES model may determine efficacy that is not reproducible in humans. For example, NMDA antagonists are very effective in the MES screen, but their efficacy in human patients with epilepsy could not be demonstrated at doses that were devoid of psychiatric complications (Löscher and Hönack, 1991). The 6 Hz seizure test is non-discriminating at low stimulus currents and can be used as a screen to identify potential AEDs. As the stimulus intensity is increased, it becomes highly discriminating and can be used to differentiate an investigational drug from those currently available for the treatment of partial epilepsy. For example, at high stimulus intensities, PHT, carbamazepine (CBZ), lamotrigine (LTG), and topiramate (TPM) are all inactive at non-toxic doses. In contrast, with the exception of TPM, all of these compounds are active at lower stimulus intensities (Barton et al., 2001). Also, LEV is active in the 6 Hz test, while it is inactive in the other primary screens; e.g., the MES and scPTZ tests. Thus, the 6 Hz model is useful for differentiating the pharmacological profile of investigational compounds. The ideal preclinical model, as described by the National Institutes of Health (NIH)/NINDS/American Epilepsy Society (AES) Models II Workshop (2002), (1) would display spontaneous seizures following an appropriate post-insult latent period, (2) would be resistant to at least two existing AEDs, (3) would be amenable to high throughput screening, and (4) would reflect a pathophysiology and phenomology similar to human epilepsy (Stables et al., 2003). While the MES and scPTZ screens are amenable to high throughput screening, they do not meet any of the remaining criteria. The 6 Hz model possesses a differential pharmacological profile compared to MES and scPTZ tests. In this respect it can serve as a potential screen for pharmacoresistant epilepsy. However, the seizures are still acutely evoked and involve a normal animal, not an epileptic animal. The kindled rat model of partial epilepsy is a model of chronic hyperexcitability; however, it is not particularly amenable to high throughput screening and animals do not routinely display spontaneous seizures. Nonetheless, electrical kindling (focal seizures) in rats is a commonly used chronic model for temporal lobe epilepsy and is highly predictive of AEDs with clinical efficacy for partial seizures (Smith et al., 2007). In addition, the kindled rat offers a more informative look into the epileptic condition than do the naïve, pathologically normal animals utilized in acute primary screens. The advantage of using “seizure experienced” animals is that the pharmacology of AEDs can be altered by the epileptic pathology (Honack and Loscher, 1995). A good example of this is the lack of observed efficacy of LEV when evaluated in models of acute evoked seizures, such as the scPTZ and MES models, vs. the marked efficacy observed in the kindled rat model.

The corneal kindled mouse model possesses certain advantages as a primary screening model compared to the hippocampal kindled rat. While possessing increased clinical predictability, most electrical kindling models are limited by many factors, which render them impractical as primary screens. First, a significant amount of time is required to prepare and care for implanted rats. Second, a great deal of skilled labor is invested in stereotaxic surgeries, post-operative care, and kindling. Third, the means necessary to financially support the required resources and staff are limiting. Finally, because of the age and body mass of the rats utilized, a large amount of compound is required for dosing. The model therefore possesses both time- and labor-intensive preparation requirements (electrode implantation surgery and a recovery period in addition to the time required to conduct a study in a chronic model), and the cost of the model is high not only due to time and labor constraints but also to the chronic housing and feeding of multiple adult rats. Corneal kindling in mice has been proposed as a cost-effective alternative model for screening compounds for the treatment of partial epilepsy (Sangdee et al., 1982, Matagne and Klitgaard, 1998). The optic nerve is positioned such that the brain may be transcorneally stimulated, creating a non-invasive route for rapid kindling. In addition, the smaller animal size minimizes the feeding, housing, and dosage requirements that limit the usefulness of the electrically kindled rat model.

Section snippets

Animals

Adult male CF1 mice weighing a minimum of 18 g (Charles River, Kingston, WA) or male Sprague–Dawley rats weighing approximately 250 g at the onset of kindling (Charles River, Raleigh, NC) were used to determine ED50 values in each of the paradigms described. Animals were housed in a temperature-, humidity-, and light-controlled (12 h light:dark cycle) facility. Mice were group housed and rats housed individually and permitted free access to food and water. All experimental procedures were

Results

In the present study, 89.5% (34/38) of mice acquired a fully kindled state; i.e., 5 consecutive stage 5 seizures. As shown in Fig. 1, the kindling curve plateaus at an average seizure score of 4 after 26 stimulations due to the fact that the latency for some mice to reach a fully kindled state was prolonged an additional 12 days. Furthermore, at this point many animals were fully kindled and hence not receiving stimulations, thereby lowering the average seizure score. Of the 38 mice that

Discussion

The corneal kindled mouse model quickly and easily yields a population of chronically kindled animals available for serial use in the initial screening of AEDs. The use of this model is preferred as an initial screen to the electrical kindled rat model in that it conserves compound, time, and labor, thereby reducing costs. It also requires less technical skill since no surgical expertise is required. The corneal kindled mouse model is advantageous over the other primary screens, including the 6 

Acknowledgements

We wish to acknowledge the editorial assistance of Dr. Misty Smith and Dr. Erika Adkins and the excellent technical assistance of Kristi Johnson, Jennifer Huff, Rizvanna Khaleel, Tim Pruess, and Leslie Hart. This work was supported by NINDS Contract NO1-NS42359.

Cited by (96)

  • Loss of presenilin 2 age-dependently alters susceptibility to acute seizures and kindling acquisition

    2020, Neurobiology of Disease
    Citation Excerpt :

    Thus, the acute efficacy of various ASDs against the secondarily generalized kindled seizure was highly divergent within age, sex, and genotype. It is important to note that doses of the ASDs administered were selected to be below the known motor impairing doses in WT, young-adult CF-1 mice so as to minimize the potential for adverse effects (Barker-Haliski et al., 2017b; Rowley and White, 2010). Nonetheless, we observed significant and uncontrollable tremors following administration of ≥17 mg/kg LTG to aged male and female mice, regardless of genotype; an effect that was not apparent in any young adult mice treated with ≥17 mg/kg LTG.

View all citing articles on Scopus
View full text