SEARCH

SEARCH BY CITATION

Keywords:

  • Ketongenic diet;
  • Epilepsy;
  • Phenytoin;
  • Valproic acid;
  • Neurotoxicity;
  • Rat

Abstract

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

Summary:  Purpose: The purpose of this study was to measure quantitatively the effectiveness of the ketogenic diet (KD) in comparison to two clinically important anticonvulsant drugs (AEDs), valproic acid (VPA) and phenytoin (PHT), and to evaluate possible associated neurotoxicity.

Methods: Rats were maintained on either a calorie-restricted, KD or calorie-restricted, rodent-chow diet for 3–5 weeks, after which neurobehavioral and seizure testing was completed. AEDs (either VPA or PHT) were injected acutely at the time to peak effect before neurotoxic and seizure assessment. Seizures were induced by timed infusion of pentylenetetrazole (PTZ) and maximal electroshock (MES).

Results: VPA protected from both MES- and PTZ-induced seizures, whereas the KD only elevated PTZ seizure threshold; PHT only attenuated MES-induced seizures. The KD was as effective as a high dose of VPA (i.e., 300 mg/kg) and combined treatment (i.e., KD + VPA) showed an additive increase in PTZ seizure threshold. No observed neurobehavioral deficits were associated with either diet treatment; however, drug-related side effects were noted with high doses of either VPA or PHT.

Conclusions: These data suggest that the KD ranks among VPA and PHT as an effective treatment for seizures, without observed drug-associated neurobehavioral contraindications. In combination with AEDs, our results indicate that the KD plus VPA work synergistically to increase seizure threshold, whereas the KD plus PHT may be complementary, elevating seizure threshold (KD) and reducing seizure severity (PHT). These findings may provide insights into future directions for rational polytherapy; however, it is important to be aware that the KD has been shown to elevate VPA-induced hepatotoxicity.

The ketogenic diet (KD) is a dietary treatment for human epilepsy that has been shown to be an effective means of managing difficult-to-control seizures for nearly 70 years (1–8). As Geyelin (9) reported that abstinence from food and drink could attenuate seizure activity, Wilder (1) subsequently developed a high-fat diet in an effort to mimic the physiology of fasting, circumvent the self-limiting nature of starvation, and attain the benefits of improved seizure control. The KD is predominantly fat (by weight and caloric content), with minimal proteins and carbohydrates. Historical reports of clinical efficacy have described a >50% reduction in seizure activity for 44–95% of the patients to whom the diet was administered (10). More recent, prospective KD studies show similar findings and indicate that the KD provides protection from various types of seizure activity (8,11).

Initial experimental studies focused on characterizing the putative anticonvulsant action of KD (e.g., 12). Early studies in mice (4,12) and rats (13) provided evidence that the KD attenuated seizure activity. There were no systematic studies of the KD in a single model, and generalizations have proven difficult. Since 1994, however, there has been a resurgence of interest in the diet and with that an increased interest in understanding the mechanisms by which the KD protects against seizures. In accordance with clinical reports (8,11,14), experimental evidence suggests that the KD acts predominantly to increase seizure resistance to a variety of acute seizure models. Specifically, the KD has been shown to increase the resistance to seizures induced by minimal electroconvulsive shock (13), by i.v. infusion of pentylenetetrazol (PTZ) (15,16), and by timed inhalation of flurothyl (17). KD-fed animals also have been shown to exhibit an increased afterdischarge threshold for electrically kindled rats (18).

Despite the long-standing clinical efficacy of the KD, the diet has not yet been evaluated quantitatively either in comparison with or in combination with other common pharmacologic treatments. This seems important for two reasons. First, a quantitative evaluation of the KD efficacy/neurotoxic effects in comparison with other antiepileptic drugs (AEDs) permits a direct valuation of KD effects. Second, a qualitative evaluation of KD efficacy/neurotoxic effects in comparison with other AEDs may provide insights into improved polytherapy, as children that begin KDs are typically taking multiple AEDs (mean, 6–7) at KD onset (8,11). Toward these aims, the present study was designed to analyze the efficacy of the KD in comparison to two commonly prescribed AEDs, valproic acid (VPA) and phenytoin (PHT), using standard AED-screening tests and parameters previously employed in this laboratory.

MATERIALS AND METHODS

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

Animals

Two hundred forty-three male Sprague–Dawley rats (Harlan Sprague–Dawley, Indianapolis, IN, U.S.A.) were housed three to five to a polycarbonate colony cage at a temperature of 21 ± 1°C, on an alternating 12:12 light/dark cycle with lights on at 07:00. Animals began diet treatment at either postnatal day 28 (P28; Experiment 1) or P37 (Experiment 2).

Diet treatment

All animals were maintained on rodent chow (Purina 5001) and water ad libitum before diet onset. Rats were age- and weight-matched and divided into two diet-treatment groups: a calorie-restricted ketogenic diet (KCR), and a calorie-restricted normal (rodent chow) diet (NCR). Both diets were calorie restricted [<90% of the daily requirement (13,19)] to minimize weight gain, as is described clinically (6). All animals were fed in calorically equivalent quantities to eliminate body weight and caloric intake as variables across groups. The KD implemented was one described previously (15), and was formulated to yield ≥78% of its calories from fat (Bio-Serv F3666, Frenchtown, NJ, U.S.A.). All rats were fasted 6–8 h before diet implementation and fed individually once each day, beginning between 15:00 and 17:00. Animals were allowed to feed for 2–3 h while water was provided ad libitum for all animals throughout the experiment. On days when neurotoxicity was tested and seizures were induced, animals were not fed until after seizure testing was completed.

Ketonemia

Ketonemia was assayed by measuring the levels of β-hydroxybutyric acid (BHB) present in blood. Blood was collected from the tail vein of the rat between 11:00 and 16:00 for all assays. BHB levels were measured immediately after specimen collection and determined spectrophotometrically using a StatSite meter (GDS Technologies, Elkhart, IN, U.S.A.).

Experiment 1

Drug Treatment

Young animals (P28 at onset) were used initially, as immature animals had been shown to respond most favorably to treatment with KDs (12,16,17) and would be most likely to show diet-related effects of drug treatment. Rats from each diet group (i.e., KCR or NCR) were subdivided into four drug-treatment groups. VPA was dissolved in 0.9% NaCl, sterile filtered (0.2 μm, Costar), and injected intraperitoneally (i.p.) in doses of 75 (n = 20), 150 (n = 40), and 300 (n = 20) mg/kg. Controls (n = 40) were given an equivalent volume of saline (i.p.). Seizure testing was completed after 15 min, the time to peak anticonvulsant effect for VPA (20).

Seizure testing

To compare the efficacy of the KD with VPA, we assessed seizure resistance by timed infusion of PTZ, a technique that affords a sensitive assay of seizure threshold using a minimal number of animals (20). After 35–39 days of diet treatment, PTZ (10 mg/ml) was infused into the tail vein via a 27-gauge infusion set at a constant rate (1.0 ml/min) using a Stoelting (Wood Dale, IL, U.S.A.) model 100 infusion pump. The infusion continued until the rat exhibited its first overt bilateral forelimb clonic seizure behavior. The total amount of drug infused at this time was calculated from the duration of the infusion and expressed as a dose (mg/kg) representing seizure threshold (20,21).

Experiment 2

Drug treatment

Animals from each of the two diet groups were subdivided into two drug-treatment groups, a VPA and a PHT group. For experiments with VPA, the following doses of VPA were injected: 0 (i.e., saline; n = 8), 150 (n = 32), or 300 mg/kg (n = 31). Seizure testing was completed as described, 15 min after injection of VPA (22). For experiments with PHT, animals were given one of three doses of PHT (i.p.): 0 (i.e., saline; n = 17), 50 (n = 17), or 75 mg/kg (n = 18). Controls were given an equivalent volume of saline (i.p.). Seizure testing was completed after 2 h, the time to peak anticonvulsant effect for PHT (23).

Doses of AEDs were chosen to be <TD50 values for rats [VPA, 365 mg/kg, (24); PHT >3,000 mg/kg (oral, based on ataxia), (25), 65.5 mg/kg (based on roto-rod) (23)] and to span much of the range of ED50 values for maximal electroconvulsive shock (MES)- and PTZ-induced seizures [VPA, 169 mg/kg for MES; 74 mg/kg for scPTZ (24)], [PHT, 30 mg/kg for MES; no protection ≤300 mg/kg for scPTZ (25)]. In the VPA and PHT experiments, the same animals were tested for responses to both MES and PTZ. Individual animals received the same doses of respective AEDs in both tests.

Seizure testing

For this experiment, seizures were induced in two ways, by timed infusion of PTZ (as in Experiment 1 above) and by MES; we did this to follow more closely the NIH-standard AED Drug Screening Protocol previously described (23). MESs were induced after 20 days of diet treatment and were immediately preceded by neurotoxicity testing (see later). PTZ seizures were induced 2 days after MES testing to allow sufficient time for recovery from seizures (26).

MESs were induced by injecting a current (150 mA for 0.2 s) via intraorbital electrodes using a constant-current device (Wahlquist, Salt Lake City, UT, U.S.A.). Before stimulation a drop of 0.9% saline was placed in each eye to ensure electrical contact and to minimize the chance of optical injury (26). The durations of hindlimb flexion and extension were recorded, and the extension-to-flexion (E/F) ratio was taken as a measure of seizure severity (27–29). Animals with an increased E/F ratio were considered to exhibit more severe seizures, as generalization was thought to occur more rapidly (26,28,29). All seizures [both chemical (PTZ) and electrical (MES)] were induced between 13:00 and 17:00 to minimize possible inconsistencies arising from circadian rhythms (30,31).

Neurotoxicity

In an effort to assess possible neurotoxic effects of diet alone, drug alone, or diet + drug treatment, neurobehavioral tests were performed at the appropriate time to peak neurotoxic effect. For some drugs, the reported time to peak neurotoxic effect is different from the time to peak anticonvulsant effect. PHT, for example, has a time to peak neurotoxic effect of 30 min compared with a time to peak anticonvulsant effect of 2 h (23). For VPA, however, the neurotoxic and anticonvulsant times to peak effects are both 15 min (22). Therefore, for VPA, we completed behavioral testing immediately before seizure testing, whereas seizure testing for PHT was done 90 min after neurotoxicity evaluation.

Six behavioral tests were performed after 20 days of diet treatment (i.e., only before the MES seizure test): (a) positional sense (b) righting reflex, (c) gait and stance, (d) muscle tone, (e) equilibrium, and (f) rotorod (29). For the positional test, the hind leg was pulled over the edge of a table to observe if the rat could quickly pull it back into a normal position. For the righting-reflex test, the rat was placed in a supine position to determine its ability to return to an upright stance. The gait-and-stance test noted any alteration in gait, such as a zig-zag or circular motion. Palpation of the upper part of the hind leg was used to assess muscle tone. For a test of equilibrium the rat was placed on the edge of a polycarbonate colony cage (∼1 cm in width), and the ability of the rat to walk along this edge was noted. The rotorod test was completed to supplement the aforementioned tests to determine any loss of motor-reflex coordination. The rotorod was 2.5 cm in diameter and revolved at 5.5 rpm. The inability to maintain equilibrium on the rod for ≥1 min (in three attempts) was taken as an indication of failure (29,32).

Neurotoxicity was scored in two different ways. First, tests were ranked either as pass or fail (30). Second, behavioral tests were also based on the following 3-point scale: 2, strong/quick response, 1, positive response, 0, failure (Tables 1 and 2). Failure in any one of these tests was considered an indication of a minimally neurotoxic dose for that animal and scored a 0.

Table 1.  Neurotoxicity for valproic acid–treated rats fed either calorie-restricted ketogenic or calorie-restricted normal diets
Behavioral testN + saline (n = 4)N + 150 (n = 16)N + 300 (n = 16)K + saline (n = 4)K + 150 (n = 16)K + 300 (n = 11)
  1. Animals were preinjected (i.p.) with saline, 150, or 300 mg/kg VPA 15 min before behavioral testing (20,23). Behaviors were quantitatively scored on a 3-point scale (see Methods). Testing was completed after 20 days of diet maintenance (at age P57), and mean scores are reported for each test.

  2. VPA, valproic acid.

  3. a,b Different letters are means that were determined to be significantly different from one another (only across rows; e.g., means or fractions labeled “a” were not different from others labeled “a” but were different from those labeled “b”). (For Mean combined score, p < 0.05, Tukey) (For No. Failures, p < 0.05, χ2).

Positional2.01.60.92.01.81.3
Righting2.02.01.92.02.02.0
Gait and  stance2.02.02.02.02.02.0
Muscle  tone2.01.60.82.01.20.5
Equil.2.01.60.91.81.70.9
Rotorod1.50.90.41.81.30.4
Mean  combined  score1.9b1.6b1.1a1.9b1.6b1.0a
 (±SD)(±0.08)(±0.24)(+0.36)(±0.16)(±0.20)(±0.39)
No. of  failures0.4b7/16a12/16a0/4b5/16a13/16a
Table 2.  Neurotoxicity for phenytoin-treated rats fed either a calorie-restricted ketogenic or a calorie-restricted normal diet
Behavioral testN + saline (n = 7)N + 50 (n = 7)N + 75 (n = 7)K + saline (n = 10)K + 50 (n = 10)K + 75 (n = 11)
  1. Animals were preinjected (i.p.) with saline, 50, or 75 mg/kg PHT 2 h before behavioral testing (23,24,26). Behaviors were quantitatively scored on a 3-point scale (see Methods). Testing was completed after 20 days of diet maintenance (at age P57).

  2. a,b Different letters are means that were determined to be significantly different from one another (compairsons are only within rows, as for Table 1). (For Mean combined score, p < 0.05, Tukey) (For No. Failures, p < 0.05, χ2).

Positional1.71.71.41.71.91.0
Righting2.01.71.71.91.71.3
Gait and  stance2.01.41.31.91.11.0
Muscle  tone2.01.71.41.31.40.6
Equil.1.91.30.91.80.80.7
Rotorod1.61.71.42.01.71.27
Mean  combined  score1.9a1.6a1.4b1.8a1.4a1.0b
 (±SD)(±0.06)(±0.27)(±0.38)(+0.12)(±0.20)(±0.39)
No. of  failures0/7a0/7a3/7a0/10a2/10a7/10b

Statistics

Statistical analyses were performed using SigmaStat and SAS. Unpaired t tests were used to compare mean BHB levels for each diet group. Analysis of variance (ANOVA) on ranks (Dunn's method) was used to compare means across different treatment groups (e.g., MES data, Figs. 2 and 4). ANOVA on ranks (Tukey method) was used to compare means with normal variance (e.g., PTZ data, Figs. 1, 3, and 5; mean combined neurotoxicity score, Tables 1 and 2). χ2 analyses were used to compare neurotoxicity differences (number of failures per test, Tables 1 and 2). Values were considered statistically different at p < 0.05. Drug and seizure groups were compared both within and across diets. Bar graphs were labeled with letters to distinguish significance; bars identified with different letters are signficantly different from each other. Those identified by the same letter are statistically indistinguishable.

image

Figure 2. Effects of valproate (VPA) on maximal electroconvulsive shock (MES) severity in rats fed either NCR or KCR diets. Animals were pretreated with 0 (i.e., saline), 150, or 300 mg/kg VPA, i.p., 15 min before seizure testing. Seizure severity was assessed as the extension-to-flexion (E/F) ratio after seizures were evoked by MES (see Methods). All animals were seizure naïve when tested. The number of animals successfully seizure-induced for each experimental group were as follows: NCR + saline, n = 4; NCR + 150 mg/kg VPA, n = 16; NCR + 300 mg/kg VPA, n = 13; KCR + saline, n = 3; KCR + 150 mg/kg VPA, n = 15; and, KCR + 300 mg/kg VPA, n = 15. Groups associated with different letters are indicative of means that were determined to be significantly different (p < 0.05, Dunn's), as for Fig. 1. The bar labeled “a,b” is not significantly different from any of those labeled either “a” or “b,” but those labeled “a” are different from those labeled “b.” Error bars are ±SEM.

Download figure to PowerPoint

image

Figure 4. Effects of phenytoin (PHT) on maximal electroconvuslive shock (MES) seizure severity in rats fed either NCR or KCR diets. All animals were seizure-naïve when tested. Animals were pretreated with 0 (i.e., saline), 50, or 75 mg/kg PHT (i.p.) 120 min before seizure testing. Seizure severity was assessed as the extension-to-flexion (E/F) ratio after seizures were evoked by MES (see Methods). The number of animals used successfully for each experimental group were as follows: NCR + saline, n = 7; NCR + 50 mg/kg PHT, n = 7; NCR + 75 mg/kg PHT, n = 7; KCR + saline, n = 10; KCR + 50 mg/kg PHT, n = 10; and, KCR + 75 mg/kg PHT, n = 11. Groups associated with different letters are indicative of means that were determined to be significantly different (p < 0.05, Dunn's), as for Fig. 1. Bar “a,b” is not different from any of the others, but all labeled “b” are different from bar “a.” Error bars are ±SEM.

Download figure to PowerPoint

image

Figure 1. Effects of valproate (VPA) on pentylenetetrazol (PTZ) seizure threshold in rats fed normal calorie restricted (NCR) or ketogenic calorie restricted (KCR) diets. Animals were pretreated with 0 (i.e., saline), 75, 150, or 300 mg/kg VPA; i.p., 15 min before seizure testing. Seizure threshold was assessed as the resistance to PTZ-induced seizures (see Methods). The number of successful infusions (i.e., animals with no interstitial edema) from each group were: NCR + 0, n = 15; NCR + 75 mg/kg VPA, n = 8; NCR + 150 mg/kg VPA, n = 13; NCR + 300 mg/kg VPA, n = 9; KCR + saline, n = 17; KCR + 75 mg/kg VPA, n = 5; and, KCR + 150 mg/kg VPA, n = 15; KCR + 300 mg/kg VPA, n = 6. Groups labeled with different letters represent means that were determined to be significantly different (p < 0.05, Tukey), that is, all bars labeled “a” are statistically equivalent but different from those labeled “b” or “c.” Error bars are ±SEM.

Download figure to PowerPoint

image

Figure 3. Effects of valproate (VPA) on pentylenetetrazol (PTZ) seizure threshold in rats fed either the NCR or KCR diet in rats tested in the maximal electroconvulsive shock (MES) protocol 2 days previously. Individual animals were pre-treated with same dose: 0 (i.e., saline), 150, or 300 mg/kg VPA, (i.p.) 15 min before seizure testing. Threshold was assessed as the resistance to PTZ-induced seizures. The number of successful infusions (i.e., animals with no interstitial edema) from each group were as follows: NCR + saline, n = 4; NCR + 150 mg/kg VPA, n = 10; NCR + 300 mg/kg VPA n = 6; KCR + saline, n = 3; KCR + 150 mg/kg VPA, n = 5; and, KCR + 300 mg/kg VPA, n = 6. Groups associated with different letters indicate means that were determined to be significantly different (p < 0.05, Tukey), as for Fig. 1. The bar labeled “b,c” is different from that labeled “a” but not different from those labeled “a,b,”“b” or “c.” Error bars are ±SEM.

Download figure to PowerPoint

image

Figure 5. Effects of phenytoin (PHT) on pentylenetetrazol (PTZ) seizure threshold in rats fed either NCR or KCR diets. Seizure testing was completed in the same group of animals (depicted in Fig. 4) previously exposed to maximal electroconvulsive shock (MES) testing 2 days earlier. Animals were pretreated with 0 (i.e., saline), 50, or 75 mg/kg PHT (i.p.) 120 min before seizure testing. Seizure threshold was assessed as the resistance to PTZ-induced seizures (see Methods). The number of successful infusions (i.e., animals with no interstitial edema) from each group were as follows: NCR + saline, n = 4; NCR+ 50 mg/kg PHT, n = 4; NCR + 75 mg/kg PHT, n = 5; KCR + saline, n = 7; KCR + 50 mg/kg PHT, n = 7; and, KCR + 75 mg/kg PHT, n = 6. Groups associated with different letters indicate means that were determined to be significantly different (p < 0.05, Tukey), as for previous figures. Error bars are ±SEM.

Download figure to PowerPoint

RESULTS

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

Experiment 1

Ketonemia

As the blood level of ketone bodies (specifically BHB) has typically been the best predictor of success for clinical KDs (e.g., 6,11), we determined the level of ketonemia for all animals (in both experiments). For animals that began diet treatment at age P28, blood BHB levels for KCR-fed rats (4.58 ± 1.36 mM, n = 17) were markedly greater than those of age-matched controls fed the NCR diet (0.52 ± 0.13 mM, n = 15) (p < 0.0001, unpaired t test).

In young animals, the KD is as efficacious as a maximal dose of VPA

Figure 1 shows three noteworthy results. First, VPA (300 mg/kg, i.p.) significantly increased PTZ seizure threshold in control animals fed an NCR diet (p < 0.05, Tukey). Second, administration of 300 mg/kg VPA to KCR-fed rats resulted in a significant increase in PTZ seizure threshold over KCR-fed animals treated with other doses of VPA (p < 0.05, Tukey). Third, ketogenic rats (without VPA) were as resistant to PTZ-evoked seizures as were control rats pretreated with the highest dose of VPA tested (p < 0.05, Tukey).

Experiment 2

Ketonemia

The levels of ketonemia for animals fed the KCR diet (P37 at onset) were significantly elevated above those of controls. Those rats maintained on the KCR diet exhibited blood BHB levels of 3.74 ± 1.91 mM (n = 12) compared with NCR-fed control values of 0.45 ± 0.13 mM (n = 10; p = 2 × 10−5, unpaired t test). Although statistically insignificant, the mean BHB level for P28 KCR-fed animals was slightly greater than the mean for the KCR-fed P37 animals (4.58 ± 0.33 mM vs. 3.67 ± 0.55 mM, respectively).

Valproic acid and KD
Valproic acid, but not KD, protects from MES-induced seizures:

VPA injection reduced MES-induced seizure severity compared with saline-injected controls (as measured by the E/F ratio; p < 0.05, Dunn's). Similar results were observed for animals fed the KCR (Fig. 2). Ketogenic rats injected with saline exhibited significantly more severe seizures (i.e., higher mean E/F ratio) than did animals fed the same diet and given 300 mg/kg VPA (p < 0.05 Dunn's). Compared with the control group (NCR), the KCR diet group did not show attenuated MES seizure severity; only pretreatment with VPA markedly reduced MES-induced hindlimb extension, and did so for both diet groups equally.

KD and VPA exhibit synergistic PTZ threshold effects:

We subsequently asked how KD-induced changes in PTZ seizure threshold compared with those of VPA in the same animals. In contrast to the failure of the KD to attenuate MES seizures, both KD and VPA treatment elevated PTZ seizure threshold (Fig. 3). Control animals injected with 300 mg/kg VPA were more seizure resistant than rats fed the same NCR diet and injected with saline (p < 0.05, Tukey). As was shown in Fig. 1, animals maintained on the KCR diet and given 300 mg/kg VPA had a markedly elevated PTZ seizure threshold compared with all other treatment groups except the NCR group given a maximal dose of VPA (p < 0.05, Tukey).

VPA, but not KD, induced neurobehavioral deficits:

Neurobehavioral performance was analyzed quantitatively by using a 3-point scale (see Methods) and by determination of the number of failures to compare potential contraindications for KD and/or VPA treatments (Table 1). For both diets, both doses of VPA resulted in an increased number of neurotoxicity-test failures. There was also an overall lower mean combined score, but only for the higher dose of VPA. It is important to note that, if the number of failures per treatment group were used as an end point [as indicated previously (29)], the neurotoxic deficits observed for both doses of VPA were more severe than those seen in animals fed either diet alone (i.e., a greater number of failures). Treatment with a KCR diet did not result in a greater number of neurobehavioral deficits than was exhibited by the NCR diet groups.

Phenytoin and KD
PHT, but not KD, protects from MES:

We also evaluated the anticonvulsant and neurotoxic effects of PHT in comparison to the KD. As shown in Fig. 4, KCR animals treated with either 50 or 75 mg/kg PHT exhibited markedly reduced E/F ratios compared with the KCR diet treatment alone (p < 0.05, Dunn's). As noted in the experiment testing VPA, saline-treated KCR animals tended to have more severe seizures than did saline-treated NCR animals (p > 0.05, Dunn's), an observation we reported previously (33) and for which we have no explanation. There also was a tendency for NCR-fed animals to show a reduced seizure severity when treated with either dose of PHT. There were no significant differences in MES-induced seizure severity for animals maintained on different diets and given the same dose of PHT. Although statistically insignificant, rats maintained on the NCR diet and treated with either 50 or 75 mg/kg of PHT showed less severe seizures than did animals fed the same diet without drug treatment (p > 0.05, Dunn's).

KD protects from PTZ:

PHT, unlike KD treatment, did not affect PTZ seizure threshold (Fig. 5). Regardless of dosage, pretreatment with PHT did not markedly affect PTZ seizure resistance for rats fed the NCR diet. Although the KCR diet group given 50 mg/kg PHT had a lower PTZ seizure threshold than did animals injected with 75 mg/kg PHT (p < 0.05, Tukey). KCR animals treated with saline exhibited PTZ seizure thresholds statistically similar to those of animals given 75 mg/kg PHT. This showed that larger doses of PHT do not markedly increase PTZ seizure threshold in KCR-fed animals. In contrast, treatment with the KCR diet did increase PTZ seizure resistance compared with respective NCR-fed groups (p < 0.05, Tukey).

PHT, but not KD, induced neurobehavioral deficits:

Drug treatment, but not diet treatment, generally resulted in neurobehavioral deficits (Table 2). NCR-fed rats treated with a 75 mg/kg dose of PHT had a lower mean combined score than did saline-injected animals fed the same NCR diet (p < 0.05, Tukey). Similar results were noted for the KCR-fed group; ketogenic animals treated with the higher dose of PHT had a significantly lower mean combined score than did saline-injected KCR-fed controls (p < 0.05, Tukey). There was a tendency for KCR-fed animals to show greater neurobehavioral side effects than NCR-fed rats given the same high dose of PHT (i.e., 75 mg/kg). This combination of treatments also resulted in a greater number of failures exhibited by the KCR + 75 mg/kg PHT group than for the NCR + 75 mg/kg PHT group (p < 0.05, χ2).

DISCUSSION

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

Valproic acid and KD

In comparing the KD with VPA and PHT, it was of interest to examine potential synergistic effects. In particular, given the earlier finding that the KD worsened MES-induced seizures (33), we sought to determine the role of each drug in KD animals exposed to MES. Our rationale, therefore, was both to compare the KD with these standard AEDs and to examine KD–drug combination, both in a test in which the KD was protective (PTZ threshold) and in one in which it was detrimental (MES).

The elevation of PTZ seizure threshold in KD-fed animals is age dependent (16). In these experiments, the increase in seizure threshold afforded by the KD was greater in young animals (P28 at diet onset) than it was in older animals (P37) fed the same diet. Although the P28 and P37 rats were fed the KD for different durations (35–39 and 20 days, respectively), we do not feel that differences in the duration of feeding at 20 days and beyond compromise these conclusions. Appleton and DeVivo (13) described a plateau in electroconvulsive threshold after 20 days, and PTZ seizure threshold has been shown to be dependent on the age at which the diet was initiated (16) but not on the duration of feeding (15). Interestingly, the KD was as effective as a 300-mg/kg dose of VPA in younger animals (P28), but was only as effective as a 150-mg/kg VPA dose in older animals (P37). These results show that the KD is not as effective in animals that start the diet at an older age, support earlier findings that the efficacy of the KD is age dependent (12,16,17,34), and suggest that there is a “window of opportunity” during which KD treatment is maximally efficacious.

The effects of VPA, unlike treatment with the KD, were not dependent on age. If one measures the incremental PTZ-threshold increase induced by VPA (i.e., for each 150 mg/kg given), there is an increase in PTZ threshold of approximately +20mg PTZ/kg for all ages tested. If one considers the VPA-induced percentage change, the effects are not much different. These observations indicate that the threshold elevation by VPA is not age dependent.

Interpretation of the effects of diet and drugs on MES is more difficult. As noted previously (33), MES-induced seizures were more severe in KD-fed animals compared with those fed normal rodent chow, an unanticipated finding we cannot explain. Whereas seizure severity for saline-injected animals was not significantly greater than that for NCR animals in this study (likely due to the small numbers of control animals used), VPA and PHT each reduced seizure severity similarly in both diet groups. For the VPA experiments, a possible complication is that the same animals were used for both MES- and PTZ-induced seizures. It is possible that AED levels had not declined to zero in the 2-day interval between these two tests, but two considerations argue against a major role for residual effects of either VPA or PHT. First, the plasma half-life of VPA has been reported as from 2.57 to 3.05 h in rats aged 18–901 days for a dose of 400 mg/kg i.p. (35) and 2.3 ± 0.7 h for an oral dose of 600 mg/kg (36). Although it remains possible that a more lasting physiologic effect remained, the 2-day interval represents 16 times the half-life, suggesting a negligible residual level of VPA. For PHT the half-life is dose dependent, just >1 h for an i.v. dose of 10 mg/kg but nearly 4.5 h at 50 mg/kg (37). Extrapolation to a dose of 75 mg/kg (even though the relationship is nonlinear) suggests a half-life of 6 h, representing 7 times the half-life for the higher dose over the 2-day interval. Second, the marginally increased seizure threshold observed with PTZ infusion in KD-fed rats, relative to rodent chow–fed rats, was nearly the same in control (saline) injected animals as it was in VPA-injected animals. A similar relationship was observed for PHT. These results are consistent with minimal residual drug or physiologic effects and further suggest a minimal role for either residual VPA or PHT.

A major finding of the present study is that VPA and the KD exhibit synergistic threshold effects. These additive effects were seen in both age groups tested. As we did not measure plasma or brain levels of either VPA or PHT, we suggest that at least four possibilities might account for this finding. First, plasma levels of free fatty acids (FFAs) and free VPA have been shown to be positively correlated (38). As elevated plasma concentrations of fatty acids are found in both KD-fed patients (39) and rats (13), higher FFA levels in the blood likely compete for VPA-binding sites in plasma. Displacement of VPA associated with a subsequent increase in free VPA by FFAs may, at least in part, account for our observations (40). Second, high concentrations of medium-chain fatty acids are known to disrupt cell membranes and open tight junctions in the intestine, as shown by increased flux of cefoxitin (41). If high concentrations of FFA associated with the KD also disrupt tight junctions in the BBB, VPA entry into the brain would increase. Third, conditions that elevate BHB have been shown to induce an up-regulation of MCT1, a transporter for BHB across the blood–brain barrier (42). It has also been shown that VPA uptake is mediated by a medium-chain fatty acid transporter (40). Consumption of a ketogenic diet might lead to upregulation of the medium-chain transporter and increased VPA transport or, alternatively, competition from elevated levels of fatty acids could lead to decreased VPA transport. Which effect might dominate is unclear. A fourth possibility is that the KD and VPA share a common mechanism(s) of action.

It is important to note, however, that the KD has been shown to enhance hepatotoxicity when combined with VPA (43). These authors reported that the KD, perhaps via carnitine deficiency or inhibition of oxidative phosphorylation, enhanced VPA toxicity in human patients (43). Although our data show that treatment with both KD and VPA may improve seizure control in animals acutely, extrapolation of these data to humans is perilous, and the hepatoxicity studies cited suggest that this therapeutic strategy, especially if instituted over the long term, may not be warranted in human patients.

Phenytoin and the KD

The KD and PHT are effectively different. PHT treatment has been shown to attenuate MES seizure activity, but to be ineffective in elevating PTZ seizure threshold (23). Conversely, the KD only elevates PTZ seizure threshold and is ineffective against MES-induced seizure activity (15,33,44,45). Our results confirm these findings (Figs. 4 and 5). Indeed, KD animals tend to exhibit more severe MES-induced seizure activity (33). Despite an incomplete understanding of mechanisms of seizures and AEDs, this work provides evidence favoring a rational polytherapy between the “complementary”actions of KD (raising threshold) and PHT (reducing severity) if the elevated neurotoxicity noted later can be tolerated.

Diet- versus drug-induced neurotoxicity

There were neurobehavioral deficits for high doses of both VPA and PHT, but not for KD treatment alone. In accordance with work showing that ketogenic animals did not perform differently from controls on spatial learning or exploratory behavior tasks (18), these data support the findings that there are few KD-associated cognitive side effects (14). The rotorod is, perhaps, the most challenging of the six behavioral tests used, and the combination of KD and PHT resulted in significantly more frequent failures (animals falling off the rod in each of three 1-min trials). Although the highest dose of PHT exceeds at least one published TD50(23), rats fed rodent chow did not experience an increase in failures when injected with this dose of PHT. The basis of this apparent neurotoxic synergy between the KD and high-dose PHT is unknown. Whereas the seizure-protective effects of the two treatments are complementary, the neurotoxicity finding suggests caution in combining the KD with higher doses of PHT.

Conclusions

In summary, there are three major findings in this study. First, the KD is as effective as a high dose of VPA (i.e., 300 mg/kg) in young animals (P28 at onset). Second, the KD acts synergistically with VPA to markedly elevate PTZ seizure threshold. Third, the KD improves seizure control without the associated neurobehavioral contraindications that accompany administration of high doses of either VPA or PHT. Collectively, these results suggest that the KD ranks among VPA and PHT as an effective treatment for seizures, with no obvious drug-associated neurobehavioral effects. These findings may provide insights into future directions of rational use of the KD and its potential role—and limitations—in polytherapy. Importantly, however, expectations based on recognition that the KD acts synergistically with VPA and complementary to PHT should be tempered by findings of metabolic disturbance associated with VPA monotherapy (46,47), possible hepatotoxicity (43), and some evidence (Table 2) for enhanced neurotoxicity between the KD and high doses of PHT.

Acknowledgments: We gratefully acknowledge the Georgetown University Department of Biology and the ARCS Foundation (K.J.B.) for their generous support of this work. We also thank Dr. Kenneth Kellar, Georgetown University Department of Pharmacology, for loan of the Wahlquist Stimulator. In addition, we thank Drs. Gail D. Anderson, Jong M. Rho, and Philip A. Schwartzkroin for helpful comments on earlier drafts of this manuscript.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES
  • 1
    Wilder RM. The effects of ketonemia on the course of epilepsy. Mayo Clin Bull 1921;2:3078.
  • 2
    Peterman MG. The ketogenic diet in epilepsy. JAMA 1925;84:197983.
  • 3
    Helmholtz HF & Keith HM. Eight years' experience with the ketogenic diet in the treatment of epilepsy. JAMA 1930;95:7079.
  • 4
    Millichap JG, Jones JD, Rudis BP. Mechanism of anticonvulsant action of ketogenic diet. Am J Dis Child 1964;107:593604.
  • 5
    Huttenlocher PR, Wilbourn AJ, Signore JM. Medium-chain triglycerides as a therapy for intractable childhood epilepsy. Neurology 1971;21:1097103.
  • 6
    Freeman JM, Kelly MT, Freeman JB. The epilepsy diet treatment: an introduction to the ketogenic diet. New York: Demos, 1994.
  • 7
    Kinsman SL, Vining EPG, Quaskey SA, et al. Efficacy of the ketogenic diet for intractable seizure disorders: review of 58 cases. Epilepsia 1992;33:11326.
  • 8
    Freeman JM, Vining EPG, Pillas DJ, et al. The efficacy of the ketogenic diet—1998: a prospective evaluation of intervention in 150 children. Pediatrics 1998;102:135863.
  • 9
    Geyelin HR. Fasting as a method for treating epilepsy. Med Rec 1921;99:10379.
  • 10
    Vining EPG. Clinical efficacy of the ketogenic diet. Epilepsy Res 1999;37:18190.DOI: 10.1016/S0920-1211(99)00070-4
  • 11
    Vining EPG, Freeman JM, Ballaban-Gil K, et al. A multicenter study of the ketogenic diet. Arch Neurol 1998;55:14337.
  • 12
    Uhlemann ER & Neims AH. Anticonvulsant properties of the ketogenic diet in mice. J Pharmacol Exp Ther 1972;180:2318.
  • 13
    Appleton DB & DeVivo DC. An animal model for the ketogenic diet. Epilepsia 1974;15:21127.
  • 14
    Swink TD, Vining EPG, Freeman JM. The ketogenic diet: 1997. Adv Pediatr 1997;44:297329.
  • 15
    Bough KJ & Eagles DA. A ketogenic diet increases the resistance to pentylenetetrazole-induced seizures in the rat. Epilepsia 1999;40:13843.
  • 16
    Bough KJ, Valiyil R, Han FT, et al. Seizure resistance is dependent upon age and caloric restriction in rats fed a ketogenic diet. Epilepsia 1999;35:218.
  • 17
    Rho JM, Kim DW, Robbins CA, et al. Age-dependent differences in flurothyl seizure sensitivity in mice treated with a ketogenic diet. Epilepsy Res 1999;37:23340.
  • 18
    Hori A, Tandon P, Holmes GL, et al. Ketogenic diet: effects on expression of kindled seizures and behavior in adult rats. Epilepsia 1997;38:7508.
  • 19
    Rogers AE. Nutrition. In: BakerHJ, LindseyJR, WeisbrothSH, eds. The Laboratory Rat. Vol I. Orlando: Academic Press, 1979:12352.
  • 20
    Pollack GM & Shen DD. A timed intravenous pentylenetetrazole infusion seizure model for quantitating the anticonvulsant effect of valproic acid in the rat. J Pharmacol Methods 1985;13:13546.
  • 21
    Foote F & Gale K. Proconvulsant effect of morphine on seizures induced by pentylenetetrazole in the rat. Eur J Pharm 1984;105:17984.
  • 22
    Bailleux V, Vallée L, Nuyts J-P, et al. Comparative anticonvulsant activity and neurotoxicity of 4-amino-N-(2,6-dimethylphenyl)phthalimide and prototype antiepileptic drugs in mice and rats. Epilepsia 1995;36:55965.
  • 23
    Krall RL, Penry JK, White BG, et al. Antiepileptic drug development: II. Anticonvulsant drug screening. Epilepsia 1978;19:40928.
  • 24
    Shenoy AK, Miyahara JT, Swinyard EA, et al. Comparative anticonvulsant activity and neurotoxicity of clobazam, diazepam, phenobarbital, and valproate in mice and rats. Epilepsia 1982;23:399408.
  • 25
    Swinyard EA, Sofia RD, Kupferberg HJ. Comparative anticonvulsant activity and neurotoxicity of felbamate and four prototype antiepileptic drugs in mice and rats. Epilepsia 1986;27:2734.
  • 26
    Swinyard EA. Assay of antiepileptic drug activity in experimental animals: standard tests. In: Mercier J, (ed. Anticonvulsant drugs: international encyclopedia of pharmacology and therapeutics. Section 19, Vol 1. New York: Pergamon Press, 1972:47–65.
  • 27
    Swinyard EA. Electrically induced convulsions. In: Purpura DP, Penry JK, Tower DB, et al. eds. Experimental models of epilepsy: a manual for the laboratory worker. New York: Raven Press, 1972:434–58.
  • 28
    Toman JEP, Swinyard EA, Goodman LS. Properties of maximal seizures and their alteration by anticonvulsant drugs and other agents. J Neurophysiol 1946;9:2319.
  • 29
    Swinyard EA, Brown WC, Goodman LS. Comparative assays of antiepileptic drugs in mice and rats. J Pharmacol Exp Ther 1952;106:31930.
  • 30
    Woolley DE & Timiras PS. Estrous and circadian periodicity and electroshock convulsions in rats. Am J Physiol 1962;202:37982.
  • 31
    Loscher W & Fiedler M. The role of technical, biological and pharmacological factors in the laboratory evaluation of anticonvulsant drugs. VI. Seasonal influences on maximal electroshock and pentylenetetrazole seizure thresholds. Epilepsy Res 1996;25:310.DOI: 10.1016/0920-1211(96)00022-8
  • 32
    Dunham MS & Miya TA. A note on a simple apparatus for detecting neurological deficit in rats and mice. J Am Pharm Assoc Sci Ed 1957;46:2089.
  • 33
    Bough KJ, Matthews PJ, Eagles DA. A ketogenic diet has different effects upon seizures induced by maximal electroshock and by pentylenetetrazole infusion. Epilepsy Res 2000;38:10514.
  • 34
    Otani K, Yamatodani A, Wada H, et al. Effect of ketogenic diet on convulsive threshold and brain monoamine levels in young mice. No Tu Hattatsu 1984;16:196204.
  • 35
    Mares P, Maresova S, Maresova D. Antimetrazol action and plasma levels of valproate in developing rats. Physiol Bohemoslovaca 1989;38:97107.
  • 36
    Binkerd PE, Rowland JM, Nau H, et al. Evaluation of valproic acid (VPA) developmental toxicity and pharmacokinetics in Sprague-Dawley rats. Fund Appl Toxicol 1988;11:48593.
  • 37
    Vicuna A, Lalka D, DuSouich P, et al. Dose dependence of the apparent half-life of phenytoin in the rat. Res Comm Chem Pathol Pharm 1980;28:311.
  • 38
    Albani F, Riva R, Procaccianti G, et al. Free fraction of valproic acid: in vitro time-dependent increase and correlation with free fatty acid concentration in human plasma and serum. Epilepsia 1983;24:6573.
  • 39
    Theda C, Woody RC, Naidu S, et al. Increased very long chain fatty acids in patients on a ketogenic diet: a cause of diagnostic confusion. J Pediatr 1993;122:7246.
  • 40
    Adkison KDK & Shen DD. Uptake of valproic acid into rat brain is mediated by a medium-chain fatty acid transporter. J Pharmacol Exp Ther 1996;276:1189200.
  • 41
    Van Hoogdalem EJ, Wackwitz AT, De Boer AG, et al. 3-Amino-1-hydroxypropylidene-1, 1-diphosphonate (APD): a novel enhancer of rectal cefoxitin absorption in rats. J Pharm Pharmacol 1989;5:33941.
  • 42
    Leino RL, Gerhart DZ, Drewes LR. Monocarboxylate transporter (MCT1) abundance in brains of suckling and adult rats: a quantitative electron microscopic immunogold study. Dev Brain Res 1999;113:4754.
  • 43
    Ballaban-Gil K, Callahan C, O'Dell C, et al. Complications of the ketogenic diet. Epilepsia 1998;39:7448.
  • 44
    Likhodii SS, Musa K, Mendonca A, et al. Dietary fat, ketosis, and seizure resistance in rats on the ketogenic diet. Epilepsia 2000;41:140019.
  • 45
    Thavendiranathan P, Mendonca A, Dell C, et al. The MCT ketogenic diet: effects on animal seizure models. Exp Neurol 2000;161:696703.DOI: 10.1006/exnr.1999.7298
  • 46
    Mortensen PR. Inhibition of fatty acid oxidation by valproate. Lancet 1980;2:8567.
  • 47
    Haas R, Chir B, Stumpf DA, et al. Inhibitory effects of sodium valproate on oxidative phosphorylation. Neurology 1981;31:14736.