SEARCH

SEARCH BY CITATION

Keywords:

  • Antiepileptic drug trials;
  • Topiramate;
  • Kainate

Abstract

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

Summary: Purpose: Potential antiepileptic drugs (AEDs) are typically screened on acute seizures in normal animals, such as those induced in the maximal electroshock and pentylenetet-razole models. As a proof-of-principle test, the present experiments used spontaneous epileptic seizures in kainate-treated rats to examine the efficacy of topiramate (TPM) with a repeated-measures, crossover protocol.

Methods: Kainic acid was administered in repeated low doses (5 mg/kg) every hour until each Sprague–Dawley rat experienced convulsive status epilepticus for >3 h. Six 1-month trials (n = 6–10 rats) assessed the effects of 0.3–100 mg/kg TPM on spontaneous seizures. Each trial involved six pairs of TPM and saline-control treatments administered as intraperitoneal injections on alternate days with a recovery day between each treatment day. Data analysis included a log transformation to compensate for the asymmetric distribution of values and the heterogeneous variances, which appeared to arise from clustering of seizures.

Results: A significant effect of TPM was observed for 12 h (i.e., two 6-h periods) after a 30-mg/kg injection, and full recovery from the drug effect was complete within 43 h. TPM exerted a significant effect at doses of 10, 30, and 100 mg/kg, and the effects of TPM (0.3–100 mg/kg) were dose dependent.

Conclusions: These data suggest that animal models with spontaneous seizures, such as kainate- and pilocarpine-treated rats, can be used efficiently for rapid testing of AEDs with a repeated-measures, crossover protocol. Furthermore, the results indicate that this design allows both dose–effect and time-course-of-recovery studies.

Traditional antiepileptic drug (AED) testing has used acute-seizure models based on chemical and electrical induction of seizures in otherwise normal animals. Efficacious AEDs may be ineffective in these models (1). Hypothetically, new AEDs that would be effective in pharmacoresistant epilepsy may be discovered by testing them in animal models with epileptic seizures, and these new AEDs may be ineffective in the acute-seizure models (1–3). Relatively little research has been conducted studying the effects of AEDs on spontaneous seizures in animals with injury-induced epilepsy. If epileptogenesis involves new mechanisms not present in the normal brain (e.g., altered receptor subunits or new circuits), then traditional AED testing in acute-seizure models may not identify effective versus ineffective drugs, because they are being tested on animals whose brains have not undergone the epileptogenic changes.

The National Institutes of Health (NIH)-sponsored “Models II Workshop” recommended that potential AEDs be tested on animals with chronic epilepsy (4). Although imperfect, these animals aim to “model” the condition of temporal lobe epilepsy. What is epileptogenic in these animal models and how these alterations may apply to human temporal lobe epilepsy is unknown. It has been hypothesized that the use of animals that have experienced epileptogenesis (and any changes that occur during this process) will more effectively detect new drugs (3,5–8). Chronic epilepsy models may be better able to predict the clinical success of experimental drugs because they produce spontaneous seizures, a chronic epileptic state, and histopathologic alterations qualitatively similar to the mesial temporal sclerosis observed in human temporal lobe epilepsy (9). Our laboratory recently generated evidence that chronic epilepsy models with spontaneous seizures, such as the kainate- and pilocarpine-induced epilepsy models, can be used to test the efficacy of AEDs (10). The present study in chronically epileptic rats attempted to develop an improved paradigm for testing AEDs that not only would provide dose–effect data, but also would allow time-course-of-recovery analyses.

Topiramate (TPM) is a broad-spectrum AED with multiple proposed uses and mechanisms of drug action (11–16). Possible mechanisms include antagonism of α-amino-3-hydroxy-5-methyl-isoxazole-4-propionate (AMPA)/kainate-type glutamate receptor–mediated inward currents (8,17), attenuation of voltage-dependent Na+ channels (18), negative modulation of L-type Ca2+ channels (19), augmentation of γ-aminobutyric acid (GABA)A receptor-mediated Cl currents (20), activation of K+ conductance (21), and inhibition of carbonic anhydrase (22). The diverse mechanisms of drug action exhibited by TPM allow a variety of clinical uses of the drug. In these studies, we conducted a proof-of-principle experiment to determine if a repeated-measures, crossover protocol could be used to perform both dose–effect and time-course-of-recovery analyses for intraperitoneal injections of TPM.

The experimental design in our previous study in rats with pilocarpine-induced epilepsy proved useful for comparing the effects of different AEDs (10). The important conceptual and practical problem in this previous study, however, was that the three drugs—1-(3-trifluoromethylphenyl) piperazine (TFMPP), phenobarbital (PB), and fluoxetine—each affected the frequency of spontaneous seizures for substantially different periods at the doses tested (i.e., TFMPP for ∼6 h, PB for slightly less than a day, and fluoxetine for substantially more than a day). In the present study, we have designed a new protocol that would also enable us to evaluate the duration of the anticonvulsant effect. We used a different but similar chemoconvulsant model of chronic epilepsy (i.e., the kainate-treated rat). A repeated-measures, crossover protocol allows each animal to be used as its own control, in spite of differences in baseline seizure frequency. In the present study, single AED injections were alternated with single vehicle-control injections, compared with the previous study in which the animals received each treatment for 5 consecutive days (10). In addition to being able to determine the dose–effect relations, we also were able to estimate recovery from the drug effect.

MATERIALS AND METHODS

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

Kainate treatment and status epilepticus

Adult Sprague–Dawley rats (150–200 g) were given intraperitoneal injections of kainate to induce status epilepticus, a model of injury-induced epilepsy. Kainate was administered in repeated, low doses (5 mg/kg) every hour until each rat (n = 8 rats per treatment series, eight separate kainate treatments) experienced convulsive status epilepticus for >3 h. Motor seizures were scored from III to V by using the Racine scale (23). A class III seizure was defined to include an erect tail, lordotic posturing, and forelimb clonus; a class IV seizure was defined by these mentioned behaviors, rearing on hind limbs, and continued forelimb clonus; and a class V seizure was characterized by the behaviors ascribed to a class IV seizure followed by a fall.

Each rat was injected subcutaneously with ∼2.5 ml lactated Ringers and received fresh fruit after the kainate-induced status epilepticus. The rats were housed in the vivarium with a standard 12-h light/dark cycle and provided food and water ad libitum for 5–6 months during direct monitoring of spontaneous seizures. The temperature of the vivarium was kept between 17° and 20°C, with a humidity of 44–78%. A trained technician monitored epileptic animals for intervals of 1–2 h (during a period when lights were on) for a total of 6 h/week before the rats were selected for the actual study on the effects of TPM. The Racine scale was used to score spontaneous seizures during direct observation in the same manner as it was previously during the kainate treatment. Rats with spontaneous, recurrent seizures were selected for drug trials. Once selected, animals to be included in the trials were video-monitored for 24 h/day before initiation of the testing protocol to eliminate rats that had infrequent spontaneous motor seizures.

Topiramate treatment and crossover protocol

Six 1-month trials were conducted by using a crossover protocol (Fig. 1) to assess the effect of 0.3, 1, 3, 10, 30, and 100 mg/kg TPM (n = 6–10 rats per trial). In total, 22 animals were used in all experiments, and some rats were tested in more than one trial. The 3-, 10-, and 100-mg/kg trials included eight rats, six of which were ovariectomized females, and two additional males. All other trials comprised solely male Sprague–Dawley rats. The mean duration of epileptogenesis (i.e., time from kainate treatment to initiation of protocol) for rats involved in each trial was 199 ± 19, 198 ± 15, 359 ± 11, 305 ± 11, 199 ± 4, and 331 ± 11 days for the 0.3-, 1-, 3-, 10-, 30-, and 100-mg/kg trials, respectively.

image

Figure 1. Experimental protocol. Topiramate (TPM) or saline was administered every other day for 25 days (e.g., Tuesday, Thursday, Saturday, Monday, Wednesday, and Friday); each kainate-treated rat served as its own control. The protocol shown was doubled in length to include six TPM vs. vehicle-control tests. Video recordings were made for 7 days/week, 24 h/day for the duration of the protocol (25 days) for each dose. The tapes that contained the actual injection began 1 h before and ended 7 h after the injection (i.e., tape was started at 8 a.m.; injection was at 9 a.m.).

Download figure to PowerPoint

Each trial involved six treatments of TPM and six saline treatments administered on alternate days (e.g., Sun., Tues., Thurs., Sat.) with a recovery day between the treatment days to allow possible persistent effects of high doses. TPM or saline was administered via intraperitoneal injection at 9 a.m. on treatment days. Analysis of seizure frequencies for the first 6-h epoch (10 a.m. to 4 p.m.) began 1 h after the injections were conducted to allow for possible effects that animal handling and injection may have on seizure frequency. To prepare TPM for injections, appropriate quantities (depending on dose) were mixed with 0.9% saline (pH 8) and stirred vigorously in a warm-water bath (40–60°C) until dissolved. Both drug treatments and saline-control (0.9% saline, pH 8) treatments were administered to rats based on the weight of the animal.

Continuous video-monitoring of spontaneous seizures

Each rat was video-monitored continuously for the entire protocol on 8-h videotapes by a Panasonic WV-BP334 black-and-white camera (G&G Technologies, Secaucus, NJ, U.S.A.). The Racine scale (23) was implemented, as before, to score seizure severity. A trained technician blinded to the treatments and dates viewed the animals' behavior. The videotapes were observed in the fast-forward mode for any activity suggestive of a seizure (running, jumping, rearing, lordosis, erect tail, etc.). If any seizure-like activity was seen, the tape was stopped, rewound, and watched in real time to evaluate any possible seizures (defined as earlier). All seizures meeting the criteria of a class III seizure (23) or greater were counted. The continuous video-monitoring did not detect subconvulsive seizures.

Seizure clusters

The occurrence of clusters of seizures during either saline or AED treatments would be expected to result in heterogeneous variance and an asymmetric distribution of the data. Others have defined normal seizure distributions by using the Poisson model (24,25), which describes the distribution of “random” events. By this definition, deviations from the Poisson model are representative of nonrandom seizure aggregations or seizure clusters. Seizures occasionally appeared to occur in clusters (Fig. 2), where an increased seizure frequency was observed within a specific time period (i.e., 6–18 seizures per 6-h epoch). Average interictal intervals within apparent clusters of seizures, regardless of dose or time, ranged between 9 min and 44 min. Seizure clusters were seen to varying degrees at several doses of TPM and during saline treatments.

image

Figure 2. Seizure clusters. A–F: Raster plots depicting the time of occurrence of seizures. A–C: Some seizures did not occur in clusters, even during high seizure frequency. D, E: High seizure frequency with short interictal intervals suggesting nonrandom seizure clusters, which supported the hypothesis that these seizures occurred in clusters. F: 16-h extension of the 6-h figure shown in E.

Download figure to PowerPoint

The log transformation and justifications for its use

Log transformations are commonly used in the analysis of seizure frequency to compensate for the asymmetric distribution of data and heterogeneous variance between groups, which presumably occurs because of seizure clustering (26). The data to be analyzed should meet a set of criteria for the use of a log transformation (27). These criteria include the following: the largest value of a set must be more than 3 times larger than the smallest value; and data are bound below by zero. A log transformation leads to a more symmetric distribution of the data where the variances are more homogeneous between groups.

The observed seizure clusters within the data and the criteria described earlier suggested that the data did not have a normal distribution. A plot of residuals versus predicted values from the original scale indicated that the variance increased with the mean; therefore, the data were analyzed in the log10(y + 0.1) scale. The relation between mean and variance (i.e., larger mean implies larger variance) identified in the original data permits the log transformation, and the transformed data more closely satisfy the assumption of homogeneity of variances. After log transformation, seizure frequencies for the TPM and saline treatments of individual AED tests were compared by using a repeated-measures analysis of variance (ANOVA) that included fixed effects for treatment, time, and the treatment-by-time interaction, as well as random effects for rats, rat-by-treatment, rat-by-time, and rat-by-treatment-by-time interactions. The analysis in SAS Proc Mixed (SAS Institute, 1999) used the restricted maximum likelihood estimation method, which pools random effects having negative variance estimates with the higher-order random interactions. Denominator degrees of freedom for F tests of effects and individual comparisons were estimated by using the Satterthwaite method. A repeated-measures analysis controlled for the effects of time within the protocol, because the kainate-induced epilepsy tends to become more severe over time (28). This approach essentially compared each seizure frequency for a particular AED treatment with the seizure frequency of a saline treatment directly before or after the AED treatment in question. Estimated relative seizure frequencies (i.e., a ratio of seizure frequency after AED injection per seizure frequency after vehicle injection) in the log scale were back-transformed after analysis to the original linear scale for presentation of results.

RESULTS

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

The time course of recovery from 30-mg/kg injections of topiramate

The first issue addressed in these experiments was to determine the duration of the effect of a 30-mg/kg injection of TPM on relative seizure frequency (Fig. 3). When each 6-h epoch was analyzed directly, a significant effect of TPM on seizure occurrence (p < 0.0001 and p = 0.03) was observed for the first two 6-h epochs (i.e., 12 h after the 30-mg/kg injection). The single injections of 30 mg/kg during the first 6-h epoch (Fig. 3A) reduced spontaneous motor seizures by approximately one half, producing a relative seizure frequency of 0.51 (±0.20). In the second 6-h epoch (Fig. 3A), seizure frequency was still reduced, with a relative seizure frequency of 0.68 (±0.24). No significant differences between TPM and saline treatments were found at any of the later time intervals. A linear regression analysis was performed to generate a “best-fit” line through the means of the relative seizure frequency of all epochs (Fig. 3B). With these data, the time point (i.e., x-values) at which the effect of TPM was completely recovered could be determined based on where the “best-fit” line intersected with a y-value equivalent to 1.0. These data suggest that the TPM injections did not have a persistent effect that interfered with the next sequential saline test.

image

Figure 3. A: Time course of recovery from 30-mg/kg injections of topiramate (TPM). TPM was significantly different (p < 0.05) from saline for the first and second 6-h epochs, 10 a.m. to 4 p.m. (p < 0.0001) and 4 p.m. to 10 p.m. (p = 0.0322). In this and the subsequent figure, the dashed line shows the baseline (i.e., no effect). The bracketed bars represent the periods of darkness (6 p.m. to 6 a.m.). Arrow, Time of drug/saline administration (9 a.m., every other day). Vertical bars, ±SEM. As shown with diagonal lines, the first epoch (2 a.m. to 8 a.m.) shares data with the last epoch (10 p.m. to 4 a.m.). B: A semi-log plot of the best-fit line through the mean values of the relative seizure frequency for the 6-h epochs (described earlier). The equation of the line is y = 0.01224x + 0.4673. Thus extrapolation of the line suggests that complete recovery from the effects of TPM occurred after ∼43 h. Ticks, Intervals of 0.1 on the y-axis. In this figure and in Fig. 4, an asterisk (*) indicates significant effects.

Download figure to PowerPoint

image

Figure 4. Dose-dependent effect of topiramate (TPM). A repeated-measures analysis of variance (ANOVA) was conducted to calculate significant differences (p < 0.05) between TPM and saline treatments. To compensate for clustering of seizures, a log transformation was performed. TPM had significant effects relative to saline at the concentrations of 10 (p = 0.02), 30 (p < 0.0001), and 100 (p < 0.0001) mg/kg.

Download figure to PowerPoint

TPM reduced seizure activity in a dose-dependent manner. The effect of different doses of TPM relative to saline treatment was compared by using a log transformation of the seizure frequency and a repeated-measures ANOVA. Six doses were tested, and they ranged from 0.3 mg/kg to 100 mg/kg. TPM exerted a significant effect relative to saline at the doses of 10 (p = 0.02), 30 (p < 0.0001), and 100 mg/kg (p < 0.0001). A dose of 30 mg/kg reduced spontaneous seizure frequency by approximately one half, whereas a dose of 100 mg/kg caused a slightly greater reduction in seizure frequency of 0.44 (±0.15). Therefore the effects of TPM (0.3–100 mg/kg) were found to be dose dependent.

An important end point of AED analysis, in addition to seizure frequency, is seizure severity. We addressed the question of whether the different doses of TPM reduced seizure severity. An analysis of the sum of seizure-severity scores (i.e., intensity-weighted seizure number) was conducted by using the same method as described earlier, but the data were analyzed in the log10 (y+1) scale. This analysis was performed on the seizure-severity score (23) for every seizure occurring in the first 6-h epoch (0.3–100 mg/kg). The doses of 10 mg/kg (p = 0.0024), 30 mg/kg (p = 0.0022), and 100 mg/kg (p = 0.0053) significantly reduced relative intensity–weighted seizure number (i.e., the ratio of the sum of seizure-severity scores after AED injection relative to the sum of seizure-severity scores after vehicle injection) during the first 6-h epoch. In addition, the single injections of 30 mg/kg TPM significantly (p = 0.0005) suppressed relative intensity–weighted seizure number by 0.53 (±0.18) over a 24-h period. Therefore TPM at 30 mg/kg caused a long-lasting reduction in seizure severity, and seizure intensity was reduced in a dose-dependent manner (0.3–100 mg/kg). Another analysis involved the ranking of intensity-weighted seizure numbers and a repeated-measures ANOVA. This type of analysis is normally performed on median values, but a repeated-measures, crossover paradigm does not provide median values. However, the analysis yielded similar p values, and significance was not altered with this strategy used to account for heterogeneous variance and asymmetric distribution.

DISCUSSION

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

Main findings of combined analyses

In the recovery-from-treatment analysis, the effect of 30-mg/kg TPM lasted 12–43 h (see later), and therefore had recovered before the next saline injection. When evaluated during the first 6-h epoch after intraperitoneal injection for the concentrations of 0.3, 1, 3, 10, and 100 kg, TPM showed a dose-dependent effect. TPM was found to cause a significant reduction in seizure frequency at 10 (p = 0.02), 30 (p < 0.0001), and 100 mg/kg (p < 0.0001). Seizure severity was also significantly reduced at doses of 10 mg/kg (p = 0.0024), 30 mg/kg (p = 0.0022), and 100 mg/kg (p = 0.0053).

The time course of recovery

Considerable translational research activity is being focused on pharmacoresistance and epileptogenesis (1,4). If one intends to use animal models with spontaneous recurrent seizures (e.g., kainate- and pilocarpine-treated rodents) to study pharmacoresistance and to test drugs for efficacy against epileptogenesis, one must first know both the dose–effect relations and the time course of drug action for seizure suppression. TPM was expected to have a significant effect during the first 6-h epoch, with a gradual decline thereafter. In rats and mice, the anticonvulsant activity peaks within 1 to 6 h after oral administration (13). The 30-mg/kg dose had a significant effect during the first 6-h epoch and significantly reduced seizure frequency for a second 6-h epoch (Fig. 3A; i.e., the drug effect was significant for ≥12 h). One concern was a persistent drug effect after the TPM injection that would last beyond the time of the subsequent control injection, which would be a potential confound in the protocol (i.e., the effect of TPM would still be present at the time of the saline injection). Analysis of the individual 6-h epochs during the subsequent recovery day revealed no significant difference between TPM and saline treatments. An analysis of the relation between time after injection of TPM and the relative seizure frequency (log10 axis of Fig. 3B) suggested that full recovery occurred ∼43 h after the TPM treatment. Therefore the 30-mg/kg injection of TPM reduced the frequency of spontaneous seizures in rats with kainate-induced epilepsy for ∼12 h, and the effects could have persisted for ≤43 h.

The repeated-measures, crossover protocol

A cross-over protocol (in which each animal serves as its own control) was used here because it accounts for differences between animals (i.e., variability in baseline seizure frequency) and the likelihood that seizures become more frequent over time (28). We typically used six crossover tests, but the protocol can be extended by using additional drug versus control tests to increase statistical power, while allowing a relatively small number of animals to be tested (as few as six animals were used for some doses). The experimental design allowed us to evaluate both the dose dependence of the drug effect and the time course of recovery from a single injection. The current protocol incorporated a recovery day to eliminate any possible additive effects of TPM, and it provided pharmacokinetic information in vivo that could not be addressed in the previous protocol (10). However, the potential for progressive drug accumulation is an important therapeutic consideration. One potential problem with the present protocol is that it may incorrectly show a drug to be ineffective (i.e., false negative) if the drug requires long-term accumulation. The protocol used here would need to be modified to study drugs that require accumulation over multiple days. Therefore the protocol used in the present experiments was more useful than the one used previously (10) for studies of the time course of recovery from AED treatment, but not necessarily for rapid dose–effect analyses.

Pharmacoresistance

Further experiments are needed to evaluate more rigorously whether this type of protocol will be useful for identifying drugs that are more efficacious for patients with pharmacoresistant epilepsy. The doses of 30 mg/kg and 100 mg/kg, which would be considered high doses from previous studies (29,30), only decreased the relative seizure frequency to 0.44 (±0.15) and did not block all seizures. Thus TPM appeared to have effects in these experiments similar to those with human intractable temporal lobe epilepsy. This suggests that rats with kainate-induced epilepsy are pharmacoresistant to TPM, but other studies with prolonged treatments are needed to test this hypothesis further. In the present study, each drug-versus-control test essentially evaluated the effect of a single injection of TPM relative to a single injection of vehicle. To test the hypothesis of pharmacoresistance, a similar repeated-measures design in which the animal receives the drug for several days in each repeat (e.g., 3 days of twice-daily injections followed by 3 days of recovery as a single AED vs. control test) would be more appropriate. Even if motor seizures were completely blocked by higher doses of TPM with this protocol, another important issue would be whether electrographic nonconvulsive seizures (i.e., equivalent to complex partial seizures) still persisted. Future experiments addressing pharmacoresistance also should include blood-concentration levels throughout the duration of an AED testing protocol. Differences in metabolism of AEDs between animals with kainate-induced epilepsy and age-matched control animals should be explored. Another factor possibly contributing to a reduced effect of AEDs in controlling the spontaneous seizures of rats with kainate-induced epilepsy is age-related metabolic changes, including renal insufficiency in aged rats (i.e., chronic progressive glomerulonephropathy) (31). Thus further studies are necessary to address the question of pharmacoresistance, but this approach should prove useful in further tests of potential AEDs. A new AED that is more effective in this protocol than TPM (and other presently available AEDs) may be more therapeutic in humans with intractable epilepsy.

Acknowledgments

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

Acknowledgment:  This study was supported by Johnson and Johnson Pharmaceutical Research and Development, LLC, and the National Institutes of Health (NS16683 and NS045144). We thank Dr. H. S. White for suggestions and comments.

REFERENCES

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