Contrasting Effects of Zonisamide and Acetazolamide on Amygdaloid Kindling in Rats


Address correspondence and reprint requests to Dr. K. Hamada at Department of Neuropsychiatry, Faculty of Medicine, Kagoshima University, 8-35-1 Sakuragaoka, Kagoshima, 890-8520 Japan. E-mail:


Summary:  Purpose: Zonisamide (ZNS) and acetazolamide (AZM) are two antiepileptic drugs (AEDs) that differ in clinical efficacy. To elucidate the mechanisms of action of these compounds, we investigated their therapeutic and prophylactic effects in rats by using a kindling model of partial epilepsy.

Methods: Electrodes were implanted into the left amygdala of adult male Wistar rats. The animals were stimulated at the afterdischarge threshold until five stage 5 seizures were induced. The generalized seizure threshold was then determined. Therapeutic effects were examined in rats manifesting successive convulsions with near-threshold stimulation. To test prophylactic effects, drugs were administered intraperitoneally before daily kindling stimulation until the animal had a stage 5 seizure or reached day 18.

Results: ZNS (10–40 mg/kg; n = 6) suppressed kindled seizures in a dose-dependent manner. Repeated administration for 7 days produced tolerance to anticonvulsive effects. AZM (25–200 mg/kg; n = 7) showed limited therapeutic effect, alleviating only the clonic convulsion in stage 5 seizures and reducing afterdischarge duration. Secondary generalization was not significantly suppressed during repeated treatment (50–200 mg/kg; n = 6). ZNS, 25 or 40 mg/kg (n = 8), significantly retarded seizure development; 15.0 or 17.0 daily stimulations were required to produce a stage 5 seizure. AZM, 50–200 mg/kg (n = 6), also retarded seizure development, with 14.0–14.8 stimulations required.

Conclusions: ZNS exhibited modest therapeutic and prophylactic effects, whereas AZM showed mainly prophylactic effects. Hypotheses are presented that may explain the mechanisms of action of these drugs.

Zonisamide (ZNS) is a novel antiepileptic drug (AED) developed in Japan (1–3). In a large number of clinical trials, including controlled comparative studies with carbamazepine (CBZ) and valproate (VPA), this compound was shown to be effective in patients with various types of intractable epilepsies. ZNS has been widely used in Japan since 1989 and in South Korea since 1992 (4,5); in the United States, some trials were conducted in the mid-1980s, until occurrences of renal calculi led to suspension of these investigations. However, after completion of new studies and approval by the Food and Drug Administration, the drug is now available in the United States (6,7).

Because ZNS has a sulfonamide group on its side chain, an initial hypothesis held that its anticonvulsive effect would have a basis similar to that of acetazolamide (AZM): inhibition of carbonic anhydrase (CA) activity. However, ZNS was found to be 100–1,000 times less potent as a CA inhibitor than was AZM (8), suggesting that its anticonvulsant action would have a different mechanism. Furthermore, clinical reports have described these drugs as having differing antiepileptic effects: ZNS is a major AED appropriate for symptomatic, localization-related, and generalized epilepsies (4–7), whereas AZM is used as adjunctive therapy in partial epilepsies and in idiopathic generalized epilepsies, especially those with absence seizures (9–11). Therefore, judging from clinical experience, ZNS and AZM may have different anticonvulsive mechanisms of action, as shown in experimental animals, even though these drugs have been reported to exhibit similar anticonvulsive properties in conventional acute seizure models (1,8–11).

In our previous study of amygdala (AM) kindling in rats, ZNS showed a potent anticonvulsive effect comparable to that of phenytoin (PHT), CBZ, and phenobarbital (PB) (12). Conversely, some reports comparing the therapeutic effects of conventional AEDs in AM-kindled rats failed to show a potent anticonvulsive effect with AZM (13,14). To help elucidate the antiepileptic features of ZNS and AZM, we used a standardized protocol comparing the acute and chronic therapeutic effects of the two drugs in AM-kindled rats. We also examined the prophylactic effects of ZNS and AZM.


Animals and surgery

Male adult Wistar rats, aged 9–10 weeks at the time of surgery, were used. All experiments were conducted in conformity with the guidelines of our institution involving experimental animals. Rats were maintained under standard laboratory conditions with a 12-h light/dark cycle and free access to food and water. The experimental groups consisted of a minimum of six animals.

Details of the surgical technique have been described previously (12,15,16). In brief, under pentobarbital anesthesia (Nembutal, 45–50 mg/kg, i.p.), a tripolar electrode was stereotactically implanted into the left basolateral AM according to the atlas of Albe-Fessard et al. (17). Stainless steel screw scalp electrodes were implanted in the left frontal region for cortical electroencephalogram (EEG) recording, the right upper orbital region as a ground electrode, and 3.0 mm posterior to the lamboid suture as a reference electrode.

Kindling procedure

Two weeks after surgery, the rats were subjected to kindling. Bipolar stimulation consisting of 60-Hz sine waves over a period of 1 s was delivered through a constant-current unit. Stimulation of the AM began at 10 μA. The intensity was increased by 4-μA increments, with a 10-min interstimulus interval, until an afterdischarge (AD) was induced (maximum of four stimulations per day). Once the AD threshold was determined, it was kept constant for once-daily kindling.

Kindled seizures were classified according to Racine's five stages (18): stage 1, facial twitching, head version, or eye closure ipsilateral to stimulation; stage 2, head nodding accompanying mastication; stage 3, clonic forelimb convulsion; stage 4, rearing to a kangaroo-like posture; and stage 5, generalized convulsion, including falling down.

After five seizures of stage 5 intensity had been induced, the generalized seizure threshold (GST) was determined. The daily stimulus intensity was reduced by 4-μA decrements until the animal failed to respond with a stage 5 seizure. Then the stimulus intensity was increased by 2-μA steps until the animal again responded with a stage 5 seizure; this stimulus intensity was designated the GST.

Drug administration and concentration monitoring

For investigation of therapeutic effect, drugs were administered to fully kindled animals that showed three consecutive generalized convulsions at near-threshold stimulation (GST, +4 μA). Drug effects were analyzed in comparison with pretreatment baseline conditions, according to the following variables: seizure severity, AD duration, motor seizure duration, and latency to appearance of forelimb clonus. In six animals in each dose group, drug administration was repeated for 7 days, and the therapeutic effects were examined.

For investigation of prophylactic effect, animals were treated daily with one of the drugs or the corresponding control vehicle after initial ADT determination. When a stage 5 seizure failed to develop after 17 drug administrations (18 kindling stimulations), subsequent stimulations were delivered in the absence of drug until completion of kindling (rekindling session).

The sodium salt of ZNS (Dainippon Pharmaceutical Co, Ltd., Osaka, Japan) was diluted with physiologic saline, and the sodium salt of AZM (Diamox Parenteral; Lederle-Takeda Pharmaceutical Co., Ltd., Osaka, Japan) with distilled water. For each drug, at least three dose levels were tested in addition to a vehicle control. Each substance was administered i.p. either 30 min (AZM) or 45 min (ZNS) before stimulation, in a volume of 1.0 mg/kg. The same procedure was repeated for different doses in the same animal with at least a 2-week interval between doses. The order of testing for each dose was systematically counterbalanced among animals. The GST was reevaluated before each treatment session.

Venous blood was sampled immediately after each drug trial for analysis of drug concentration by high-performance liquid chromatography (HPLC) (15). After each daily drug treatment, behavioral changes in the animals were observed. In addition, responses to handling were described, with special attention to motor ataxia, muscle relaxation, and sedation.

Histology and data analysis

At the end of the experiment, rats were given an overdose of sodium pentobarbital and perfused with 10% buffered formalin. Brains were removed and immersed in formalin. Electrode locations were verified histologically. Statistical comparisons were made by using the Kruskal-Wallis test, with the Scheffé post hoc test for alteration of seizure stage. Analysis of variance (ANOVA) was determined by using the Bonferroni post hoc test for AD, motor seizure duration, and number of stimulations in the prophylactic experiment.


Short-term therapeutic effects

ZSM was administered at three dose levels. There were six rats in each dosage group and six in the control group. The results are summarized in Table 1. Generalized convulsions were suppressed in four rats in the 10-mg/kg group and in six rats in the 25-mg/kg and 40-mg/kg groups. When ZNS failed to suppress stage 5 seizures, seizure characteristics were not significantly affected. Seizure regression to stage 0 (no remarkable motor manifestations) was observed in two rats in the ZNS 10-mg/kg group, in four in the 25-mg/kg group, and in all six in the 40-mg/kg group. The AD duration also was reduced in a dose-dependent manner by increasing doses of ZNS and was completely suppressed in five rats in the 40-mg/kg group, without overt motor disturbance.

Table 1.  Acute effects of zonisamide (ZNS) on kindled seizures
AdministrationStagesAD duration
(% of pre)
01, 25
  • ZNS, zonisamide; AD, after discharge.

  • Each figure represents the number of rats that showed respective seizure stages (0, 1, 2 or 5). Brackets [stage 0–2] indicate suppression of secondary generalization.

  • a

     p < 0.01 versus control by Kruskall-Wallis test for seizure suppression, and by ANOVA with Bonferroni post-hoc test for AD duration.

Control (n = 6)[00]697.8 ± 9.0 
 10 mg/kg (n = 6)[22]243.9 ± 38.7a7.2 ± 1.1
 25 mg/kg (n = 6)[51]a04.7 ± 8.1a20.4 ± 1.7
 40 mg/kg (n = 6)[60]a01.7 ± 4.1a34.7 ± 1.1

AZM was administered at five dose levels, with seven rats in each dosage group and seven control animals. The results are summarized in Table 2. Significant suppression of generalized convulsions was not achieved at any dose. AZM failed to suppress stage 5 seizures; however, seizures were significantly altered, with reduced duration of AD and of forelimb clonus. Latency to the appearance of forelimb clonus was not significantly affected. Because the anticonvulsive effect of AZM was characterized by reductions in AD and motor seizure duration, the alterations in these parameters were calculated separately for a few rats showing nonsignificant suppression of secondary generalization. Overt behavioral change was not observed at any dose. Unlike the procedure in the ZNS group, the serum concentration was measured in only two animals in each AZM group, as indicated in the notes to Table 2.

Table 2.  Short-term effects of acetazolamide (AZM) on kindled seizures
AdministrationDuration of AD (% of pre)Forelimb clonus (%)
stages 4, 5
Stage 01, 24, 5
  • AZM serum level (μg/ml; n = 2): 25 mg/kg (22.8), 50 mg/kg (41.3), 100 mg/kg (139.1), 150 mg/kg (180.2), 200 mg/kg (249.9)

  • a

     p < 0.01 vs. control.

Control (n = 7)97.0 ± 9.691.5 ± 16.3(n = 7)
 25 mg/kg (n = 7)61.6 ± 17.0a76.1 ± 14.1(n = 7)
 50 mg/kg (n = 7)065.5 ± 15.7a60.2 ± 13.8a(n = 6)
 100 mg/kg (n = 7)6.265.9 ± 12.4a51.3 ± 19.6a(n = 6)
 150 mg/kg (n = 7)69.1 ± 12.1a49.2 ± 14.2a(n = 7)
 200 mg/kg (n = 7)056.4 ± 17.1a44.7 ± 12.5a(n = 6)

Long-term therapeutic effects

Repeated administration of ZNS over a period of 7 days (n = 6) gradually reduced its anticonvulsive potency at each dosage level (Fig. 1). Significant suppression of kindled seizures was observed only for the initial 1 or 2 days in the 25-mg/kg ZNS group, but suppression was preserved in the 40-mg/kg group (Fig. 1, Table 3). When ZNS failed to suppress generalized seizures, seizure parameters were not significantly affected at any time in the course of treatment.

Figure 1.

Therapeutic effects of zonisamide (ZNS) on kindled amygdaloid seizures. Bars show the average seizure stage. Lines indicate changes in afterdischarge (AD) duration as a percentage of baseline. Statistical comparisons were made for suppression of seizures using the Kruskall-Wallis test. *p < 0.01, **p < 0.05 vs. control.

Table 3.  Long-term effects of zonisamide (ZNS) on kindled seizures
AdministrationStagesAD duration
(% of pre)
01, 25
  • Therapeutic effects of ZNS after repeated administration of seven days.

  • a

     p < 0.05 vs. control.

Control (n = 6)[00]694.3 ± 18.2 
 10 mg/kg (n = 6)[01]590.2 ± 31.99.7 ± 1.1
 25 mg/kg (n = 6)[02]474.5 ± 40.625.3 ± 3.8
 40 mg/kg (n = 6)[31]a233.6 ± 43.0a35.9 ± 5.3

Conversely, significant suppression of secondarily generalized seizures never occurred during repeated treatment with AZM at 50, 100, or 200 mg/kg (n = 6), whereas one or two trials showed nonsignificant suppression in each group (Fig. 2). The characteristic anticonvulsive features of the drug [i.e., reductions in AD and motor seizure (forelimb clonus) duration] were not steady and consistent during the course of treatment (Fig. 2, Table 4).

Figure 2.

Alterations in therapeutic effects of acetazolamide (AZM) in kindled amygdaloid seizures. Bars show the average seizure stage. Lines indicate changes in afterdischarge (AD) duration as a percentage of baseline.

Table 4.  Long-term effects of acetazolamide (AZM) on kindled seizures
AdministrationDuration of AD (% of pre)Forelimb clonus (%)
Stage 01, 24, 5Stages 4, 5 
  1. Therapeutic effects of AZM after repeated administration for 7 days.

  2. a  p < 0.05; bp < 0.01 vs. control.

Control (n = 6)//103.0 ± 10.1103.1 ± 18.3(n = 6)
 50 mg/kg (n = 6)//77.3 ± 11.9a80.8 ± 18.8(n = 6)
 100 mg/kg (n = 6)//69.6 ± 21.2b55.5 ± 9.2b(n = 6)
 200 mg/kg (n = 6)//73.6 ± 15.4b59.4 ± 6.9b(n = 6)

Prophylactic effects

The effects of ZNS on the number of kindled AM seizures and on AD duration are compared with observations in the control group (Fig. 3). For the 25- and 40-mg/kg groups (n = 8), both types of kindled response were significantly reduced in the later 3trials. These two dosage groups required significantly more stimulations to develop an initial stage 5 seizure (Table 5). Three animals in each group remained at stage 2 after 18 stimulations and received subsequent stimulations in the absence of ZNS. In these animals, the first stage 5 seizure was attained with an average of 2.7 and 4.3 additional stimulations, respectively.

Figure 3.

Effects of zonisamide (ZNS) on development of kindled seizures. Pretreatment with ZNS at 25 or 40 mg/kg but no 10 mg/kg showed modest prophylactic effect. Each point represents the mean + SD. *p < 0.05, **p < 0.01 by the Kruskall-Wallis test with Scheffe post-hoc test for kindling stage and by ANOVA with Bonferroni post-hoc test for afterdischarge (AD) duration versus control.

Table 5.  Effects of zonisamide (ZNS) on kindling development
 Latency toReKdLevel
Stage 2Stage 5
  1. a  p < 0.05; bp < 0.01 vs. control by analysis of variance with Bonferroni's post-hoc test.

Control (n = 8)6.1 ± 2.49.6 ± 2.2   
 10 mg/kg
  (n = 6)
7.0 ± 1.412.0 ± 3.7  11.9 ± 3.3
 25 mg/kg
  (n = 8)
7.8 ± 4.615.0 ± 5.5a2.7(n = 3)25.1 ± 1.0
 40 mg/kg
  (n = 8)
9.3 ± 4.517.0 ± 5.6b4.3(n = 3)32.6 ± 4.0

AZM, 50–200 mg/kg (n = 6), significantly retarded the number and AD duration of kindled seizures in the middle and later trials (Fig. 4). Compared with controls, all AZM groups required significantly more stimulations to develop an initial stage 5 seizure (Table 6). Two animals in each of the 50- and 100-mg/kg groups remained at stage 2 after 18 stimulations and required an average of 3.5 and 2.0 additional stimulations, respectively, to develop a stage 5 seizure. One rat in the 200-mg/kg group remained at stage 4 on the last treatment day and developed a stage 5 seizure on the next day in the absence of AZM.

Figure 4.

Effects of acetazolamide (AZM) on development of kindled seizures. Pretreatment with AZM at 50–200 mg/kg showed significant prophylactic effects. *p < 0.05, **p < 0.01, by the Kruskall-Wallis test with Scheffe post-hoc test for kindling stage and by ANOVA with Bonferroni post-hoc test for afterdischarge (AD) duration.

Table 6.  Effects of acetazolamide (AMZ) on kindling development
 Latency toReKdLevel
Stage 2Stage 5
  • a

     p < 0.05 vs. control.

 (n = 8)
4.3 ± 1.59.0 ± 1.7   
 50 mg/kg
  (n = 6)
8.5 ± 4.714.7 ± 5.2a3.5(n = 2)53.1 ± 12.3
 100 mg/kg
  (n = 6)
7.5 ± 4.114.0 ± 4.8a2.0(n = 2)116.1 ± 12.1
 200 mg/kg
  (n = 6)
7.8 ± 3.114.8 ± 3.4b1.0(n = 1)221.1 ± 31.3


Short-term therapeutic effects

In the present study, short-term administration of ZNS suppressed kindled seizures in a dose-dependent manner: lower doses inhibited seizure generalization, and higher doses inhibited focal seizure activity. This observation is in agreement with our previous report comparing the therapeutic effects of ZNS with those of CBZ, PHT, and PB in AM-kindled rats by using an identical procedure (12). The effective serum concentration of ZNS obtained in this study was almost identical to that in our previous experiment (12) and also to findings in an earlier study assessing the therapeutic range of the drug by using conventional seizure/toxicity models (1). These concentrations also resemble therapeutic ZNS serum concentrations recommended in human clinical practice (4–7). Our previous study also indicated that near-threshold stimulation is more sensitive in detecting the anticonvulsive efficacy of ZNS than is suprathreshold stimulation, as was reported for CBZ (14,19) and more recently for PHT (20).

Previous reports comparing the therapeutic effects of conventional AEDs, including AZM, in AM-kindled rats with only suprathreshold stimulation failed to demonstrate a potent anticonvulsive effect of AZM, but AD or motor seizure duration nonetheless was reduced (13,14). In the present study, the therapeutic effect of AZM was reevaluated with near-threshold stimulation, and similar results were obtained. Thus the anticonvulsive effect of AZM on kindled seizures appears to be characterized by reduction of AD and motor seizure duration, whereas secondary generalization has not been found to be significantly suppressed, irrespective of the stimulus intensity applied. AZM serum concentrations in this study overlapped or exceeded the clinical therapeutic range for the drug (10,11).

Long-term therapeutic effects

Therapeutic effects were examined over time with repeated administration of the two drugs. Nonsignificant failure to progress to secondary generalization was occasionally observed even in control groups during the treatment period of 7 days. Similar variability in reproducing generalized seizures was noted in our previous series of kindling experiments using an identical procedure (16,21). With ZNS, development of tolerance in therapeutic effect was obvious, whereas significant anticonvulsive efficacy was still preserved in the 40-mg/kg group. Serum concentrations of ZNS on the final treatment day were almost the same as those seen with short-term administration.

In contrast, significant suppression of secondary generalization never occurred with AZM throughout the treatment period of 7 days. Significant reduction of AD and motor seizure duration was still observed on the final treatment day, and seizure latency was not significantly affected throughout the course. Blood was not sampled on the final day of treatment with AZM, but it is expected that the serum concentration remained fairly constant during repeated administration. This expectation is based on data obtained in the experiment evaluating prophylactic effect. Judging from the results of experiments with short- and long-term administration, we conclude that the anticonvulsive action of AZM is characterized by alleviation of final-stage convulsions, but not suppression of established secondary generalization.

Thus the therapeutic effects of ZNS and AZM, with the latter representing a typical CA inhibitor, are quite different in a kindling model of epilepsy. These contrasts indicate differing mechanisms of action for the two drugs. The present study clearly differentiated the anticonvulsive actions of ZNS and AZM, even though the drugs have been reported to exhibit similar anticonvulsive properties in conventional acute seizure models (1,8–10).

Prophylactic effects and mechanisms of action

The present experiments demonstrated moderate prophylactic effects of both ZNS and AZM, in agreement with previous preliminary reports (22,23). In the present study, moreover, alterations in anticonvulsive effects during repeated administration were taken into account. ZSM, at both 25 and 40 mg/kg, suppressed the development of kindled seizures in the later trials, although significant therapeutic effects were lost at the 10- and 25-mg/kg doses after 2 to 3 days of repeated administration. AZM also showed a modest prophylactic effect, compared with other AEDs, although the drug did not exhibit significant anticonvulsive effects with either one or several days of administration in kindled rats.

The therapeutic effect of ZNS was shown to be as potent as that of PHT, CBZ, and PB (12). The prophylactic efficacy of ZNS in this study is comparable to that of CBZ among the major AEDs examined so far; a moderate prophylactic effect has been reported with CBZ in cats (in rats the effect was not conclusive), more substantial prophylaxis has been found with PB, and no prophylaxis was observed with PHT in either animal species (22,24,25). Accordingly, the antiepileptic features of ZNS more closely resemble those of CBZ than of PHT or PB in the kindling model of partial epilepsy.

Conversely, AZM showed an interesting discrepancy between therapeutic and prophylactic effects. Among other AEDs, VPA and the benzodiazepines appear to have modest therapeutic and definite prophylactic effects (22,24,25). More specific γ-aminobutyric acid (GABA)ergic drugs (e.g., vigabatrin and tiagabine) have displayed moderate effects in both therapeutic and prophylactic experimental models (26,27). In relation to glutamatergic neurotransmission, both N-methyl-D-aspartate (NMDA)- and AMPA-receptor antagonists significantly retard the development of kindled seizures, and drugs that block AMPA receptors also show potent anticonvulsive effect in kindled animals. However, NMDA-receptor antagonists lack substantial therapeutic effect (28,29). The antiepileptic spectrum of AZM in the present study resembles that of NMDA-receptor antagonists, although with less prophylactic efficacy.

Inhibition of CA is believed to be a basic antiepileptic mechanism of AZM. Systemic acidosis is not the explanation for the anticonvulsant properties of AZM, but accumulation of CO2 in the brain is essential (9–11). Recently several neurophysiologic studies indicated a significant role for pH or hydrogen ion changes in modulation of neurotransmission in the central nervous system. Both GABAergic and glutamatergic transmission can be affected (30–32). These findings are intriguing in association with the antiepileptic spectrum of the AEDs mentioned earlier. More recent studies indicate that glutamatergic neurotransmission of the NMDA type is more important in this regard, suggesting a neuroprotective role for mildly acidic pH shifts (33–36). These studies support our notion that AZM shares some antiepileptic properties with NMDA antagonists in the kindling model. Possible interaction of AZM with this type of neurotransmission, and the clinical application of AZM as a neuroprotective agent, appear to warrant further investigation.

Acknowledgment: This research was supported by grants from the Ministry of Health and Welfare of Japan. Preliminary results were presented at the 8th International Bethel-Cleveland Clinic Epilepsy Symposium in April 1997. We thank Mrs. Kinoe Mieno and Mr. Yukio Abe for their technical assistance. Zonisamide was supplied by Dainippon Pharmaceutical Co., Ltd., and acetazolamide, by Lederle-Takeda Pharmaceutical Co., Ltd.