SEARCH

SEARCH BY CITATION

Keywords:

  • Ketogenic diet;
  • Calorie restriction;
  • Lithium-pilocarpine;
  • Epileptogenesis;
  • Neuroprotection

Summary

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References

Purpose: Although the number of antiepileptic drugs (AEDs) is increasing, none displays neuroprotective or antiepileptogenic properties that could prevent status epilepticus (SE)–induced drug-resistant epilepsy. Ketogenic diet (KD) and calorie restriction (CR) are proposed as alternative treatments in epilepsy. Our goal was to assess the neuroprotective or antiepileptogenic effect of these diets in a well-characterized model of mesial temporal lobe epilepsy following initial SE induced by lithium-pilocarpine in adult rats.

Methods: Seventy-five P50 male Wistar rats were fed a specific diet: normocalorie carbohydrate (NC), hypocalorie carbohydrate (HC), normocalorie ketogenic (NK), or hypocalorie ketogenic (HK). Rats were subjected to lithium-pilocarpine SE, except six NC to constitute a control group for histology (C). Four rats per group were implanted with epidural electrodes to record electroencephalography (EEG) during SE and the next six following days. From the seventh day, the animals were video-recorded 10 h daily to determine latency to epilepsy onset. Neuronal loss in hippocampus and parahippocampal cortices was analyzed 1 month after the first spontaneous seizure.

Results: After lithium-pilocarpine injection, neither KD nor CR modified SE features or latency to epilepsy. In hippocampal layers, KD or CR exhibited a neuroprotective potential without cooperative effect. Parahippocampal cortices were not protected by the diets.

Conclusion: The antiepileptic effect of KD and/or CR is overwhelmed by lithium-pilocarpine injection. The isolated protection of hippocampal layers induced by KD or CR or their association failed to modify the course of epileptogenesis.

More than 10 new antiepileptic drugs (AEDs) have been developed during the last two decades. They especially improved patients’ quality of life but failed to demonstrate a neuroprotective or antiepileptogenic effect. In the peculiar case of chronic drug-resistant epilepsy, a neuroprotective therapy design has to be acceptable for a long period with few adverse effects. The ketogenic diet (KD) could match these features. For almost a century, the KD has been a commonly used alternative therapy to manage drug-resistant epilepsies, especially in childhood (Swink et al., 1997; Vining et al., 1998; Hartman & Vining, 2007; Cross, 2009). The KD is a high-fat, low-protein plus low-carbohydrate diet with a usual ratio by weight of 4:1 fats to carbohydrates plus proteins. The antiepileptic effect of the KD increases for the first 2 weeks after its onset (Freeman et al., 2000). A moderate (5–10%) calorie restriction can be a part of the KD in clinical practice (Bough et al., 2000a,b).

Animal studies support the antiepileptic efficacy of the KD. In rodents, the KD increases epileptic seizure threshold significantly about 1–2 weeks after initiation (Appleton & DeVivo, 1974; Bough & Eagles, 1999; Rho et al., 1999; Bough et al., 2006). In addition, a potential neuroprotective effect of KD is suggested by its ability to reduce cellular loss and mossy fiber sprouting after status epilepticus (SE) (Muller-Schwarze et al., 1999; Su et al., 2000; Noh et al., 2003; Sullivan et al., 2004).

The aim of our study was to assess the antiepileptogenic efficacy of a calorie-restricted KD in the model of limbic epilepsy subsequent to lithium-pilocarpine–induced SE in rats. Because the antiepileptic potential of a calorie-restricted carbohydrate diet is also reported in animal models (Bough et al., 2003; Eagles et al., 2003; Raffo et al., 2008), our additional goal was to determine the respective impact of the KD and calorie restriction (CR) alone or combined.

Methods

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References

Animals, breeding, and diets

Animals were kept in uncrowded breeding facilities at 22°C under a 12 h/12 h normal light/dark cycle (lights on at 07:00 h). All animal experimentation was performed in accordance with the rules of the European Community Council Directive of November 24, 1986 (86/609/EEC), and the French Department of Agriculture (License No. 67–97). The experimental protocol was approved by the ethical Animal Research Committee Board of University Louis Pasteur (CREMEAS, #AL/02/05/03/07), and all efforts were made to minimize animal suffering.

Seventy-five P50 male Wistar rats were randomly divided into four weight-matched groups. Each group was fed a specific diet according to calorie intake and composition: normocalorie carbohydrate (NC), hypocalorie carbohydrate (HC), normocalorie ketogenic (NK), or hypocalorie ketogenic (HK). Water was given ad libitum. Animals were kept in separate cages to control their individual calorie intake. Appropriate amounts of each diet were provided every morning on the basis of daily animal’s weight. Food was maintained constant for the first 3–4 days, even if the animals lost weight. This procedure was applied to allow specific adaptation to the KD. Normocalorie intake consisted of 100% of the daily calorie requirement, that is, 0.3 kcal/g (Rogers, 1979). Hypocalorie intake consisted of a 15% calorie-restricted diet. Two feeding products were used: the usual laboratory certified UAR A04C carbohydrate rodent diet (UAR, Villemoisson-sur-Orge, France) and a pharmaceutical ketogenic product: Ketocal (Nutricia, North America, Gaithersburg, MD, U.S.A.) characterized by a 4:1 ratio of lipids to carbohydrate plus protein, with 100% long chain triglycerides, and balanced in vitamins and oligoelements. The carbohydrate, protein, and lipid detailed composition of both diets is indicated in our previous report (Raffo et al., 2008).

Measurement of blood glucose and β-hydroxybutyrate

Before the injection of methyl scopolamine, two drops of blood were taken from the tip of the tail and used for the measurement of blood glucose and β-hydroxybutyrate levels, respectively. We used the Accu-Chek glucometer (Roche Diagnostics, Meylan, France) for glucose measurement and the β-ketone strip test on the Medisense Optium Xceed reader (Abbott France Division Medisense, Rungis, France) for β-hydroxybutyrate.

Status epilepticus

SE was induced at posnatal day 71 (71), after 3 weeks of the specific diets. From the NC group, six rats were not subjected to SE in order to constitute a control (C) group for histologic procedures. SE was induced by injection of lithium chloride (3 mEq/kg, i.p.; Sigma-Aldrich, St. Louis, MO, U.S.A.) 20 h before the subcutaneous injection of pilocarpine hydrochloride (25 mg/kg, s.c.; Sigma-Aldrich). Rats received methyl-scopolamine bromide (1 mg/kg, s.c.; Sigma-Aldrich) 30 min before pilocarpine to reduce the peripheral consequences of the cholinergic agonist. By 2 h after the onset of SE, all rats received an intramuscular injection of diazepam (2.5 mg/kg, i.m., Valium; Roche, Basel, Switzerland). The control group received lithium and saline instead of pilocarpine and diazepam. Controlled diets were maintained for 1 week after SE.

EEG recording

One week before SE, four rats per group were implanted with epidural electrodes. Rats were anesthetized by i.p. injection of 37 mg/kg ketamine (Ketalar; Pfizer, La Jolla, CA, U.S.A.) and 5.5 mg/kg xylazine (Rompun; Bayer, Leverkusen, Germany). Four screw electrodes were implanted bilaterally over the frontoparietal cortex, and one reference electrode was implanted over the cerebellum. All EEG acquisitions were computerized (Coherence, Deltamed, France) and performed in freely moving animals placed in acrylic glass boxes. Each animal was recorded 1 h before the lithium injection for baseline. SE EEG was recorded during 8 h, starting 30 min before pilocarpine injection. On the subsequent 6 days, EEG was recorded 2 h per day, starting at 10:00 h to avoid any influence of the circadian cycle.

Behavioral assessment of spontaneous recurrent seizure occurrence

From the seventh day after SE, all rats were video-recorded 10 times per 24 h 7 days a week during the lighted period of the animal housing. Fast playing analysis of the videos allowed identification of stage 3–5 spontaneous recurrent seizure (SRS). The duration of the latent phase of the lithium-pilocarpine model was counted from the day of SE to the day of the first spontaneous limbic seizure that marks the onset of the chronic phase. For the analysis of the latency to the first SRS, video-recording was used preferably to video-EEG recording, since a recent study performed on more than 40 rats showed that the detection of the first SRS by the two methods differs at most by 24–72 h (François J & Nehlig A, unpublished data).

Histology

All animals were sacrificed 30 days after the onset of the chronic phase, after deep pentobarbital anesthesia. Animals that did not exhibit seizures 10 weeks after SE were sacrificed at this end point. Brains were dissected out and frozen in methylbutane chilled to −30°C. Coronal brain sections of 30-μm thickness were obtained in a cryostat. Every fifth section was collected along the whole brain. After thionine staining, neuronal counting was assessed by a standardized two-dimensional (2D) technique. A 200× computerized picture was taken for each structure of interest: CA1 and CA3 layers of the hippocampus (CA1, CA3), hilus of the dentate gyrus (DG), ventral entorhinal cortex (VEntCx), dorsal entorhinal cortex (DEntCx), ventral piriform cortex (VPirCx), and dorsal piriform cortex (DPirCx). The average neuronal soma size as well as the nucleus density of the thionine staining was determined in each brain structure. The structure was manually outlined before computerized counting of the neurons to avoid subjectivity of visual counting (MCID analysis software, 2D grain count module, InterFocus Imaging, Cambridge, United Kingdom). Each structure was bilaterally read twice for each animal by an observer blinded to the animal’s treatment diet, and the average of these measurements was considered for each animal.

Statistical analysis

Acute mortality during SE and incidence of chronic epilepsy were compared between the SE groups using chi-square test. Blood glucose and β-hydroxybutyrate levels and neuronal densities of each structure of interest were compared by means of the multivariate analysis of variance (MANOVA) test for multiple comparisons followed by a post hoc Fisher’s protected least significant difference (PLSD) test between the NC, HC, NK, HK, and C groups. Within the SE groups, the interaction between the calorie restriction and KD on neuronal loss was assessed by a two-way ANOVA.

Results

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References

Rats were distributed in five groups: C (n = 6), NC (n = 15), HC (n = 14), NK (n = 16), and HK (n = 24). As already reported by our group, the design of the different animals feeding procedure permitted good diet acceptance and healthy breeding (Raffo et al., 2008).

Ketosis and SE features

Concentrations of β-hydroxybutyrate and gulcose are given in mM (mean ± SD). The concentration of β-hydroxybutyrate was significantly higher in the ketogenic (1.9 ± 0.8 and 2.6 ± 0.7 for NK and HK rats, respectively) compared to the carbohydrate groups (0.1 ± 0.1 and 0.2 ± 0.1 for NC and HC rats, respectively). Plasma glucose levels were in the same range as in our previous study (Raffo et al., 2008), with significantly higher levels in carbohydrate-fed animals (6.9 ± 1.2 and 6.4 ± 2.1 for NC and HC rats, respectively) compared to rats receiving the KD (6.1 ± 0.9 and 4.7 ± 0.7 for NK and HK rats, respectively).

The clinical signs of lithium-pilocarpine–induced SE in this study were identical to those of our previous reports (Fernandes et al., 1999; Dube et al., 2000a, 2001; Francois et al., 2006). Almost all the rats in each group experienced SE after pilocarpine injection, without any difference with the diet (14/15 NC, 12/14 HC, 16/16 NK, and 23/24 HK). Rats of all groups exhibited the same behavioral features after pilocarpine injection, without difference in the latency to each parameter. Within 5 min after pilocarpine injection, rats developed diarrhea, piloerection, and other signs of cholinergic stimulation. During the following 15–20 min, rats exhibited head bobbing, scratching, chewing, and exploratory behavior. Recurrent seizures started around 20–30 min after pilocarpine administration with associated episodes of head and bilateral forelimb myoclonus with rearing and falling. They progressed to SE at about 40–50 min after pilocarpine. SE was considered to start after three consecutive stage 5 seizures (clonic seizures, rearing, and falling) (Tursky et al., 1989). During SE, clonic seizures occurred continuously for 8–10 h after pilocarpine. After the end of SE, occasional recurrent seizures occurred over the following 24 h. In the first 24 h following SE onset, 5/15 NC, 6/14 HC, 4/16 NK, and 11/24 HK rats died. The difference in acute mortality between all ketogenic and carbohydrate-fed animals, calorie-restricted or not, was not statistically significantly. A higher mortality rate occurred in the HK compared to NK animals (p < 0.01) (Fig. 1). Mortality in this group was the consequence of sustained tonic seizures during the first 2 days after pilocarpine injection.

image

Figure 1.   Lithium-pilocarpine injection: global outcome regarding the diet. The bar graphs represent the number of animals that did not develop SE (green), that died from status epilepticus (SE, yellow), or that became epileptic [spontaneous recurrent seizures (SRS), orange] or not (blue) in the four diet groups (NC, HC, NK, and HK). No statistically significant difference in acute mortality was seen between ketogenic and carbohydratefed animals.

Download figure to PowerPoint

EEG recordings and clinical correlations

Whatever the animal’s diet, no significant difference could be seen in SE features (Fig. 2). After pilocarpine injection, the first spikes occurred within 17.0 ± 1.1 min (NC 17.2 ± 2.6; HC 16.4 ± 2.7; NK 16.9 ± 1.1; HK 17.5 ± 2.4) concurrent with arrests of any motor activity. High amplitude recruiting spike activity started between 31.5 ± 7.2 min after pilocarpine (NC 33.8 ± 6.7; HC 31.5 ± 8.7; NK 33.0 ± 5.2; HK 27.8 ± 9.2), while rearings occurred. This spike activity became continuous in <10 min in all animals, with an average spike frequency of 10.4 ± 0.5 Hz (NC 10.4 ± 0.6; HC 10.5 ± 0.5; NK 10.8 ± 0.6; HK 10.2 ± 0.5). Within the fourth hour after SE, spikes parceled and led to periodic epileptiform discharges. All along this period animals exhibited no adapted motor behavior but suffered from loss of postural tonus, body shivering, and mild jerks.

image

Figure 2.   Electroencephalographic patterns after lithium-pilocarpine injection. The recordings were performed using a right anteroposterior EEG derivation. Day 1: (A) Baseline; (B) First spikes; (C) High amplitude recruiting spike activity becomes continuous; (DF) Spikes parcel and lead to periodic epileptiform discharges. Day 2: (G) Frequent bursts of spikes and waves and sharp waves. Day 3: (H) EEG reorganizes. Interictal spikes decrease to disappear by the sixth day.

Download figure to PowerPoint

The day following SE, EEG showed frequent bursts of spike-and-waves and sharp waves, without motor manifestations. These EEG discharges occurred over an average duration of 9.8 ± 2.7 min/h (NC 10.8 ± 3.0; HC 9.3 ± 3.3; NK 9.8 ± 3.1; HK 9.5 ± 2.8). Rats recovered postural tonus but remained behaviorally impaired, mainly with respect to food intake. The second day after SE onset, EEG reorganized, whereas animals recovered feeding and drinking behaviors. The frequency of EEG interictal spikes decreased in the succeeding days to almost disappear by the sixth day.

Chronic epilepsy onset

After a clinically silent period, 9/9 NC, 6/6 HC, 10/12 NK, and 11/12 HK became epileptic. Only three rats receiving the KD (NK = 2; HK = 1) did not exhibit seizures after 70 days of monitoring. There was no statistically significant difference in the incidence of epilepsy between the groups (Fig. 1).

The clinical features of the recurrent seizures were similar in all groups. The first seizures were partial limbic seizures. With time, these seizure trends generalized. The average latency to the first limbic seizure was variable (7–50 days) and reached a mean value of 20.9 ± 10.7 and 26.3 ± 16.3 in NC and HC rats, respectively, and 14.5 ± 5.5 and 16.7 ± 9.5 in NK and HK rats, respectively. Because of the large variability, there was no significant difference in this latency according to the animals’ diet at the time of SE.

Histology

Neuronal density was measured in seven of the most damaged brain structures in the lithium-pilocarpine model. Neuronal loss constitutes the first event underlying epileptogenesis in this model of mesial temporal lobe epilepsy. It occurs mainly in CA1, CA3, the hilus of the dentate gyrus, and the dorsal and ventral entorhinal and piriform cortices.

In parahippocampal cortices the type of diet did not affect neuronal loss subsequent to lithium-pilocarpine SE. In the DPirCx and VPirCx, whatever the diet, neuronal cell layers II and III were damaged and did not permit cell counting (Fig. 3). Significant damage occurred in all SE groups compared to naive controls in DEntCx and VEntCx (p < 0.01). The extent of neuronal loss ranged from 29.9–61.9% in the DEntCx and from 43.5–61.7% in the VEntCx without any significant difference according to the diet (Fig. 4).

image

Figure 3.   Brain damage induced by lithium-pilocarpine status epilepticus (SE). Coronal sections stained with thionine staining. Composite pictures from Nikon optical-MCID analyzer acquisition. (A) Control rat. (B) Epileptic rat after SE. Whatever the diet, SE induced almost complete neuronal loss in the piriform cortex.

Download figure to PowerPoint

image

Figure 4.   Effect of calorie restriction and/or ketogenic diet on neuronal loss in entorhinal cortices after lithium-pilocarpine SE. Compared to naive controls, NC rats showed significant neuronal loss in ventral (yellow bars) and dorsal (orange bars) entorhinal cortices (p < 0.01). Whatever the diet, SE induced highly significant neuronal loss in these structures. *p < 0.05; **p < 0.01: statistically significant differences from controls (C). Abbreviations: C, controls; NC, normocalorie carbohydrate; HC, hypocalorie carbohydrate; NK, normocalorie ketogenic; HK, hypocalorie ketogenic; DEntCx, dorsal entorhinal cortex; VEntCx, ventral entorhinal cortex.

Download figure to PowerPoint

In the hippocampal layers, highly significant differences could be seen between neuronal densities during the chronic phase of epilepsy according to the diet (MANOVA, p < 0.01). NC animals exhibited significant neuronal loss in CA3 compared to naive controls (C) (−31.9%; p < 0.01). A trend toward neuronal loss in CA3 was seen in HC rats (−12.3%; p = 0.07). Interestingly, in the two KD groups no neuronal loss occurred in CA3. The average neuronal density after SE was not significantly different in both KD-fed groups compared to naive controls (NK + 15.6%; HK +7.2%). Consequently, a strong significant difference was observed in neuronal CA3 density between the ketogenic rats compared to carbohydrate-fed rats subjected to lithium-pilocarpine SE (NK, p < 0.01; HK, p < 0.01) (Fig. 5). Two-way ANOVA analysis showed the main effect of the KD in reducing neuronal loss in CA3 (p < 0.001) without effect of calorie restriction and without cooperative effect between the two strategies.

image

Figure 5.   Effect of calorie restriction and/or ketogenic diet on neuronal loss in hippocampal cell layers after lithium-pilocarpine SE. Significant neuronal loss occurred in the hippocampal layers CA3 and CA1 and the hilus of the dentate gyrus in normocalorie carbohydrate-fed animals compared to naive controls. In CA3 (dark green bars) as well as CA1 (light green bars) reduced neuronal loss after SE was shown with both normocalorie and calorie-restricted KDs. In the hilus (red bars), calorie restriction in the carbohydrate diet as well as the KD exhibited a neuroprotective potential that was no longer found with the hypocalorie ketogenic diet. *p < 0.05, **p < 0.01: statistically significant differences from C. #p < 0.05, ##p < 0.01: statistically significant differences from NC. Abbreviations: C, controls; NC, normocalorie carbohydrate; HC, hypocalorie carbohydrate; NK, normocalorie ketogenic; HK, hypocalorie ketogenic; Hilus, hilus of the dentate gyrus.

Download figure to PowerPoint

In CA1 similar effects of the two diet strategies were seen. Compared to C, NC exhibited significant neuronal loss in CA1 after SE (−52.3%; p < 0.01), whereas no significant neuronal loss was seen in the others groups (HC −12.3%; NK +12.1%; HK +7.02%). In CA1, hypocalorie carbohydrate as well as the two KD-fed groups demonstrated significantly reduced neuronal loss compared to NC animals (Fig. 5). A two-way ANOVA analysis confirmed the main neuroprotective effect of the KD (p < 0.001) in CA1 without interaction or effect of calorie restriction by itself.

A significant reduction in the average surface of the hilus of the dentate gyrus area was documented in all SE groups compared to naive controls (p < 0.01). This area was reduced by 33.9–40.8% without difference due to diet. Furthermore, neuronal density in the hilus was almost reduced by half in NC animals (−45.9%), whereas HC and NK animals showed a preservation of this density (Fig. 5). A two-way ANOVA showed that neither KD nor calorie restriction exhibited a main effect on neuronal density in the hilus, but an unexpected negative interaction between the two strategies was significant (p < 0.001). In other words, the association of calorie restriction and KD did not protect hilar neurons despite the neuroprotective effect induced by the two strategies used in isolation.

Discussion

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References

Breeding and diets

Good acceptance and healthy animal breeding despite lower weights in the NK and HK groups were already reported with the same experimental feeding designs. In this previous experiment, we also demonstrated significantly higher blood levels of both β-hydroxybutyrate and lower glycemia in the ketogenic groups compared to the carbohydrate ones. These differences were of higher amplitude in the HK group than in the NK one, as reported previously (Raffo et al., 2008).

Electroclinical features of status epilepticus

Neither calorie restriction nor the KD modified the acute phase of lithium-pilocarpine SE. In each group, features and latencies of the clinical events as well as EEG patterns were similar to those previously described in this model (Clifford et al., 1987; Cavalheiro, 1995; Dube et al., 2001).

The inability of the KD to prevent lithium-pilocarpine SE was expected. If the KD was shown to increase seizure thresholds in different animal models (Loscher & Fiedler, 1996; Bough & Eagles, 1999; Eagles et al., 2003; Raffo et al., 2008), no reduction in seizure severity was demonstrated either with maximal electroshock stimulation in rats (Hori et al., 1997), audiogenic seizures in magnesium-deficient rats (Mahoney et al., 1983), or pentylenetetrazole (PTZ) infusion in rats (Raffo et al., 2008). A proconvulsant effect of the KD diet was even suggested by several authors in maximal PTZ seizures in rats but remains controversial (Mahoney et al., 1983; Bough et al., 2000a; Thavendiranathan et al., 2000, 2003). The latter group raised the hypothesis that, in accordance with Davenport’s report (Davenport & Davenport, 1948), calorie restriction is the key factor for this proconvulsant effect, whatever the type of diet. This is consistent with our finding that calorie-restricted animals exhibited a trend to higher acute mortality than normocalorie fed ones, and this difference was a consequence of severe tonic seizures during SE.

We would like to emphasize seizure threshold and severity as two different criteria. Caloric-restricted diets seem to act by enhancing fast γ-aminobutyric acid (GABA)ergic inhibition (Erecinska et al., 1996; Yudkoff et al., 2001; Bough et al., 2003). Bough et al. already reported more severe seizures induced by maximal electroshocks in calorie-restricted animals (Bough et al., 2000b), although immature rats subjected to calorie restriction exhibited higher PTZ thresholds (Bough et al., 1999). The same authors used in vivo extracellular experimental field responses to angular stimulation to determine population spikes (PS) thresholds in the dentate gyrus. They were able to show a higher PS amplitude for a given stimulus intensity with a carbohydrate or ketogenic calorie-restricted diet compared to a carbohydrate diet provided ad libitum (Bough et al., 2003). Interestingly, they concluded that both calorie-restricted diets (ketogenic or carbohydrate) have anticonvulsant properties and pointed out that the probable enhanced inhibition in the dentate gyrus could be overcome by a high-level excitatory stimulus. We hypothesize that lithium-pilocarpine–induced synaptic excitability could be such an overwhelming stimulus. This phenomenon may explain the paradoxical finding of the ability of the KD in increasing seizure threshold while failing to prevent, or even worsening, severe seizures.

Neuroprotection

The aim of our experimental design was not to study different sensitivities to lithium-pilocarpine induced by the diets; such a study would necessitate progressively increasing injected doses of pilocarpine (Hirsch et al., 1992). Conversely the similar electroclinical features of the initial SE allowed us to compare histologic damage and the subsequent phase of epileptogenesis according to the diet.

Parahippocampal cortices were highly damaged. The piriform cortex is usually entirely destroyed after lithium-pilocarpine SE, and that was also the case in the present experiment, whatever the diet. The ventral entorhinal cortex exhibited significantly higher neuronal loss than the dorsal entorhinal cortex in each group, as already shown in animals fed with the carbohydrate diet ab libitum (Francois et al., 2006).

In the hilus of the dentate gyrus, not only was the neuronal density after lithium-pilocarpine SE at least twice as low as in controls, but also the average surface of this structure was reduced in all groups by two-fold or more. No diet strategy was able to reduce this neuronal loss, and we also found a negative effect of the combination of the KD and calorie restriction in this structure in terms of neuronal density. In neurodegenerative disorders, a neuroprotective effect of the KD is mediated by the ability of neurons to afford a very high metabolic level (Gasior et al., 2006; Maalouf et al., 2007). The lithium-pilocarpine SE induces neuronal loss in highly metabolic challenged structures like the hippocampal complex and the parahippocampal cortices (Turski et al., 1983, 1989; Cavalheiro, 1995; Dube et al., 2000b, 2001; Santiago et al., 2006). In these extreme conditions, glucose oxidative potential is unable to support the acute and sustained metabolic demand that could be partly fulfilled by the alternative energetic pathway provided by the ketone bodies.

Calorie restriction was not neuroprotective when the carbohydrate diet was used, and never demonstrated a positive interaction with the KD. The KD (NK and HK) displayed neuroprotection in CA3 and CA1. Similarly, the KD also displayed neuroprotective properties when introduced 2 days after SE induced by kainic acid, reducing the occurrence of spontaneous seizures and abnormal sprouting of dentate granule mossy fibers (Muller-Schwarze et al., 1999; Su et al., 2000). Spontaneous seizures and mossy fiber sprouting were also reduced by the KD in epileptic Kv1.1 potassium channel gene null mutant mice (Rho et al., 2000). Several putative mechanisms involved in neuronal loss after epileptic seizures have been considered to be influenced by the KD (Maalouf et al., 2009). Among those mitochondrial dysfunction may lead to the production of toxic reactive oxygen species (ROS). The KD has been shown to increase the expression and activity of mitochondrial uncoupling proteins and thereby reduce neuronal damage induced by ROS (Sullivan et al., 2004). Another neuroprotective effect of the KD in kainic acid seizures is mediated by the inhibition of caspase-3-mediated apoptosis (Noh et al., 2003).

Epileptogenesis

Neither calorie restriction nor KD, or the combination of the two strategies, modified the clinical course of temporal lobe epilepsy after lithium-pilocarpine SE in terms of incidence, latency, or clinical expression of recurrent seizures. The features we observed were similar to the ones already described in this model (Leite et al., 1990; Cavalheiro, 1995; Andre et al., 2000a,b). Neuroprotection afforded only by the KD to hippocampal structures was similar to the effect of AEDs like vigabatrin and topiramate or amygdala kindling in the same SE model. As here, these neuroprotection strategies with effects limited to Ammon’s horn failed to influence epileptogenesis (Andre et al., 2000b, 2001; Rigoulot et al., 2004). However, no causal link between neuronal loss and epileptogenesis has been established in this model. It is not yet clear whether the degree of neuroprotection in the hippocampus resulting from the strategies mentioned earlier, including the KD, is insufficient to modify the course of the disease or if the hippocampus plays only a secondary role in the genesis of spontaneous seizures. In fact, CA1 seems to act more as a relay in the entorhinal-hippocampal loop than as an area promoting epilepsy, despite the network reorganization occurring in this structure (Lehmann et al., 2001).

Other AEDs, like pregabalin, did protect only parahippocampal cortices and delayed epilepsy onset after SE (Andre et al., 2003). The piriform and entorhinal cortices are injured as early as 2–6 h after SE onset (Roch et al., 2002a) and in 21 day-old rats subjected to SE; no SRS are observed in the absence of an early signal in these cortices (Roch et al., 2002b). A recent finding of our group was the ability of carisbamate to protect hippocampal and parahippocampal structures after lithium-pilocarpine SE, which prevented the occurrence of spontaneous motor seizures (François et al., 2005). Based on these findings, the association of KD and a parahippocampal protective drug could constitute rationale polytherapy in some types of drug-resistant symptomatic epilepsies, for example temporal lobe epilepsy (TLE).

In amygdala kindling in rats, a model of almost nonlesional TLE, the KD protected against the focal generation of seizures during the first 2 weeks of stimulation but, as in the present study, did not prevent seizure spread (Hori et al., 1997). This finding is in line with the data of our previous study showing that the KD is able to prevent only seizures of moderate severity (Raffo et al., 2008). On the other hand, the KD displayed antiepileptogenic properties in PTZ kindling in mice, but again mainly on seizures of low severity (Hansen et al., 2009), and a 15% calorie restriction delayed epileptogenesis in EL mice, but quite severe 30% calorie restriction was necessary to prevent epileptogenesis (which is more severe than the paradigm used here) (Greene et al., 2001).

Conclusion

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References

Systemic lithium-pilocarpine injection can overcome the antiepileptic effect of normocalorie or hypocalorie KDs. After lithium-pilocarpine SE, calorie restriction as well as the KD exhibited a neuroprotective effect in CA1 and CA3, without cooperative effect. The association of CR and KD even showed a paradoxically worsening effect on hilar neuronal loss. In parahippocampal cortices each diet failed to prevent damage. As shown with pharmaceutical drugs the protection of hippocampal layers by KD or CR or their association failed to modify the clinical course of epileptogenesis after an initial lithium-pilocarpine SE.

Acknowledgments

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References

This work was supported by a grant from INSERM (U 666). The authors are grateful to SHS for providing freely Ketocal. This protocol has received written consent from ethical Animal Research Committee Board of University Louis Pasteur in Strasbourg (CREMEAS, #AL/02/05/03/07).

Disclosure

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References

None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

References

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References