The composition of each diet is shown in Table 1 and in the Methods. No adverse effects of the diets were observed in either R-fed mouse group. Despite the 20–23% body weight reduction, mice in both R-fed groups appeared healthy and were more active than the mice in the UR-fed groups as assessed by ambulatory and grooming behavior. With the exception of oily fur, the KD-fed mice appeared active and healthy throughout the study as previously found 20. No signs of vitamin or mineral deficiency were observed in the R-fed mice according to standard criteria for mice 56. These findings are consistent with the well-recognized health benefits of mild to moderate caloric restriction in rodents 57, and support our previous findings that both the KD and a moderate CR are well tolerated by EL mice 2021.

All mice were matched for age (approximately 210 days) and body weight (approximately 31.0 ± 1.5 g) before the start of the dietary treatment (Fig. 1). All mice lost approximately 7–9% of their body weight during the 14 hr fast. Body weight remained relatively stable over the nine-week treatment period in both UR-fed mouse groups (Fig. 1). The 20–23% body weight reduction was achieved in the R-fed groups after about two weeks of gradual food restriction. However, more difficulty was encountered initially in maintaining a stable body weight reduction for the KD-R group than for the SD-R group. This difficulty may result from the high caloric content of the KD that produces greater body weight changes per calorie adjustment than the SD. We also estimated that the degree of CR necessary to maintain the 20–23% body weight reduction was about 38–45% for the SD and about 45–52% for the KD.

All mice had at least 15 recurrent seizures before the start of dietary treatment (arrow, Fig. 1). The seizures occurred occasionally during routine cage changing prior to the pre-trial period and regularly from handling during the pre-trial test period. Seizure susceptibility was analyzed in all mouse groups after the R-fed mice achieved a stable body weight reduction, i.e., week five of treatment (Figs. 1 and 2). Seizure susceptibility was high for both UR-fed groups throughout the study. In both R-fed groups, seizure susceptibility decreased from 1.0 to about 0.3 after two weeks and remained significantly lower than that of the UR-fed control groups from treatment weeks 5–12 (Fig. 2). Only a single mouse in the KD-R group had a break-through seizure on week 8. Taken together, our findings show that seizure management in EL mice is more associated with the amount than with the origin of dietary calories.

Plasma glucose levels were analyzed in all mouse groups after the R-fed mice achieved a stable body weight reduction (Figs. 1 and 3). Glucose levels remained high for both UR-fed groups throughout the study and were stable over treatment weeks 5–12. However, plasma glucose levels were somewhat lower (about 8 mM) in both UR-fed groups between treatment weeks 3–5 compared to the pre-trial glucose levels (about 10 mM). This reduction might result from a combination of repetitive handling, seizures, blood collection, and the initial fast. In both R-fed mouse groups, the plasma glucose levels decreased from about 10 mM to about 5.0 mM after three weeks and remained significantly lower than those of their respective UR-fed control groups (Fig. 3).

Plasma β-hydroxybutyrate levels were also analyzed in all mouse groups after the R-fed mice achieved a stable body weight reduction (Figs. 1 and 4). These levels remained low in the SD-UR group throughout the study and were stable for treatment weeks 5–12. β-hydroxybutyrate levels were significantly higher in the R-fed groups than in their respective UR-fed control groups (Fig. 4). These levels were also significantly higher in the KD-UR group than in the SD-UR group. The levels increased from about 0.4 mM to about 1.7 mM in the SD-R group and to about 3.0 mM in the KD-R group. These findings demonstrate that circulating β-hydroxybutyrate levels were inversely related to circulating glucose levels and that elevated β-hydroxybutyrate levels alone are not associated with seizure susceptibility.

The relationship between body weight, food intake, plasma glucose levels, plasma β-hydroxybutyrate levels, and seizure susceptibility was determined using Pearson bivariate correlation analysis (Table 2). All variables were significantly (p < 0.01) correlated with each other. Positive correlations were found among body weight, food intake, glucose, and seizure susceptibility. On the other hand, β-hydroxybutyrate was negatively correlated with all variables. The correlations among glucose, β-hydroxybutyrate, and seizure susceptibility were also apparent from the data in Figures 2,3,4. Plasma glucose was significantly (p < 0.001) associated with seizure susceptibility in the EL mouse, as determined by chi-square analysis (Fig. 5). These results support our previous findings that glucose levels are predictive of seizure susceptibility in adult EL mice 2129.

Binary logistic regression was also used to determine the relationship between seizure susceptibility, plasma glucose, and plasma β-hydroxybutyrate levels when mice were fed either the SD and/or the KD. The data indicate that regardless of diet, glucose could predict seizure susceptibility with an approximate 75 to 78 % accuracy (Table 3). Although β-hydroxybutyrate could also predict seizure susceptibility, we previously showed that β-hydroxybutyrate levels were dependent on and were inversely related to plasma glucose levels 21.

The inbred EL/Suz (EL) mice were originally obtained from J. Suzuki (Tokyo Institute of Psychiatry). The mice were maintained in the Boston College Animal Care Facility as an inbred strain by brother × sister mating. The mice were group housed (prior to initiation of study) in plastic cages with Sani-chip bedding (P.J. Murphy Forest Products Corp., Montville, N.J.) and kept on a 12-hr light/dark cycle at approximately 22°C. Cotton nesting pads were provided for warmth when animals were individually housed. All cages and water bottles were changed once per week. Only females were used for these studies as adult males die sporadically with age from acute uremia poisoning due to urinary retention 70. The procedures for animal use were in strict accordance with the NIH Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care Committee.

Seizure onset in EL mice is generally between 60–70 days of age as previously described 6. These seizures occur occasionally during routine cage changing. Our recently developed seizure handling protocol was used to regularly induce seizure susceptibility in EL mice 621. Briefly, the testing procedure included repetitive handling and simulated the stress normally associated with weekly cage changing, i.e., picking the mouse up by the tail for short intervals and transferring it to a clean cage with fresh bedding. The test included two trials that were separated by 30 min. In each trial, a single mouse was held by the tail for 30 sec at approximately 10–15 cm above the bedding of its home cage. After 30 sec, the mouse was placed into a clean cage with fresh bedding for 2 min. The mouse was then held again for 15 sec before being returned to its home cage. Trial 2 was performed even if the mouse experienced a seizure in trial 1. The epileptic seizures commenced during holding or soon after the mice were placed on the clean bedding. Mice that developed an epileptic seizure while handling were placed immediately in either the clean cage or their home cage depending on the testing stage. Mice were tested each week for a total of 13 measurements over a 12-week period using this method. Mice were undisturbed between testing phases (no cage changing) and testing was performed between 12 to 3 pm.

Mice were designated seizure susceptible if they experienced a generalized seizure during seizure testing. Generalized seizures in EL mice involve loss of postural equilibrium and consciousness, together with excessive salivation, head, limb, and chewing/swallowing automatisms. An erect forward-arching Straub tail, indicative of spinal cord activation, was also seen in most mice having generalized seizures. Mice that displayed only vocalization and twitching without progression to generalized seizure were not considered seizure susceptible 621. Seizure susceptibility scores were generated for each mouse according to the seizure severity scores previously described 6. Mice having a score of 4 or 5 were assigned a susceptibility score of 1.0, whereas mice having a seizure severity score less than 4 were given a susceptibility score of 0. The seizure susceptibility for each mouse was then averaged over multiple tests and the mean seizure susceptibility for a mouse dietary group was determined.

All mice received PROLAB RMH3000 chow diet (LabDiet, Richmond, IN, USA) prior to the experiment. This is the standard food pellet diet (SD) and contained a balance of mouse nutritional ingredients. According to the manufacturer's specification, this diet delivers 4.4 Kcal/g gross energy, where fat, carbohydrate, protein, and fiber comprised 55 g, 520 g, 225 g, and 45 g/Kg of the diet, respectively. The ketogenic diet (KD) was obtained from the Zeigler Bros., Inc. (Gardners, PA, USA) in butter-like form and also contained a balance of mouse nutritional ingredients. According to the manufacturer's specification, the KD delivers 7.8 Kcal/g gross energy, where fat, carbohydrate, protein, and fiber comprised 700 g, 0 g, 128 g, and 109 g/Kg of the diet, respectively. The fat in this diet was derived from lard and the diet had a ketogenic ratio (fats: proteins + carbohydrates) of 5.48:1. The individual % composition of each dietary energy component for the SD and KD diet is shown on Table 1.

Seizure susceptibility, body weight, and food intake was measured four times over a three-week period in 24 singly caged female EL mice (approximately 210 days of age). All mice received the SD during the pre-trial period and food intake was determined by subtracting the weight of food pellets remaining in the food hopper after one week from the initial amount given (200 g). The difference was then divided by seven to estimate the average daily food intake. Thus, all mice were highly seizure susceptible at the initiation of the diet therapy.

After the three-week pre-trial period, the mice were placed into four groups (n = 6 mice/group) where the average body weight of each group was similar (about 31.0 ± 1.5 g) (Fig. 1). All mice were then fasted for 14 hr to establish a similar metabolic set point at the start of the experiment (arrow, Fig. 1). The mice in each group were then given one of four diets to include: 1) the standard diet fed ad libitum or unrestricted (SD-UR), 2) the KD fed ad libitum or unrestricted (KD-UR), 3) the SD restricted to achieve a 20–23% body weight reduction from the pre-trial weight (SD-R), and 4) the KD restricted to achieve a 20–23% body weight reduction from the pre-trial weight (KD-R). Each mouse in the two R groups served as its own control for body weight reduction. Based on food intake and body weight during the pre-trial period, food in the R-fed mouse groups was reduced until each mouse achieved the target weight reduction of a 20–23%. In other words, the daily amount of food given to each R mouse was reduced gradually until it reached 77–80% of its initial (pre-trial) body weight.

The mice in the SD-UR group received 200 g of food in the hopper/week as in the pre-trial period. For mice in the SD-R group, weighed food pellets were dropped directly into each cage for easy access. The KD was administered to the mice in a modified plastic Falcon tissue culture dish (60 mm × 15 mm). The dish edges were shaved to reduce the height from 15 mm to about 6 mm. After placing about 5 g of KD in the dish for the KD-UR mice, the dish with the weighed KD was inverted and placed on the top of the food hopper. An empty water bottle was placed on top of the dish to prevent dish movement during animal feeding. The butter-like consistency adhered the KD to the inverted dish. This feeding apparatus allowed the mice easy access to the KD and prevented KD contact with bedding material. After about 24 hr, the amount of KD consumed was determined and another 5 grams of fresh KD were added to the dish. The KD was therefore given fresh every day without moving or disturbing the mice. The total amount of KD consumed per day was summed each week and divided by 7 to obtain the average weekly food intake of each mouse. For the KD-R mice, a calculated restricted amount of KD was placed directly on top of the food hopper bars for easy access. The R-fed mice licked the bars clean of the KD.

Blood was collected approximately 1 h after seizure testing except for the pre-trial period where blood was not collected. Blood was first collected from all mice about 24 hr prior to the initiation of the 14 hr fast (Fig. 1). Mice were anesthetized with isoflurane, USP (Halocarbon, River Edge, NJ, USA) and blood was collected in heparinized tubes by puncture of the retro-orbital sinus using a borosilated capillary tube (FHC, Bowdoinham, ME, USA). The blood was centrifuged at 6,000 × g for 10 min, the plasma was collected, and aliquots were stored at -80°C until analysis. Plasma glucose concentration was measured spectrophotometrically using the Trinder Assay (Sigma-Aldrich, St. Louis, MO, USA). Plasma β-hydroxybutyrate concentration was measured using either the Stanbio β-Hydroxybutyrate LiquiColor^® procedure (Stanbio, Boerne, TX, USA), or a modification of the Williamson et al procedure 71.

Both ANOVA and a two-tailed t test were used to evaluate the significance of differences of body weight, seizure susceptibility, plasma glucose levels, and plasma β-hydroxybutyrate levels between unrestricted and restricted groups. Chi-square analysis was performed on the association between glucose and seizures. Pearson bivariate correlation analysis (SPSS software) was used to determine the relationship between body weight, food intake, plasma glucose levels, plasma β-hydroxybutyrate levels, and seizure susceptibility. Binary logistic regression (SPSS) was used to determine the relationship between seizure susceptibility, plasma glucose, and β-hydroxybutyrate levels on mice fed either the SD or the KD. Differences were considered significant at p < 0.01. All values are expressed as mean ± SEM. All statistical data were presented according to the recommendations of Lang et al., 72.