Ketogenic diet treatment abolishes seizure periodicity and improves diurnal rhythmicity in epileptic Kcna1-null mice

Abstract

Introduction: Seizures are known to perturb circadian rhythms in humans as well as in animal models of epilepsy. However, it is unknown whether treatment of the underlying epilepsy restores normal biologic rhythms. We asked whether: (1) seizure activity is characterized by diurnal rhythmicity, (2) chronically epileptic mice exhibit impaired rest-activity rhythms, and (3) treatment with the anticonvulsant ketogenic diet (KD) improves such perturbations.

Methods: Chronically epileptic Kcna1-null mice were fed either a standard diet (SD) or KD for 4 weeks and subjected to continuous video-EEG (electroencephalography) and actigraphy monitoring for 3-5 days to assess seizure activity and rest-activity cycles.

Results: Seizure activity in Kcna1-null mice demonstrated diurnal rhythmicity, peaking at zeitgeber (ZT)2.30 +/- 1.52. Rest-activity rhythms of epileptic mice were significantly disrupted. Whereas locomotor activity for wild-type mice peaked at ZT15.45 +/- 0.28 (ZT14:26-ZT16:51), peak activity of epileptic mice was more unpredictable, occurring over a 12.4 h range (ZT06:33-ZT18:57). In six of nine epileptic mice, peak activity was delayed to ZT17.42 +/- 0.38, whereas peak activity was advanced to ZT10.00 +/- 1.26 in the remaining mice. Treatment with the KD abolished seizure periodicity and restored the rest-activity rhythm to values resembling those of wild-type mice (i.e., activity peaking at ZT16.73 +/- 0.67).

Conclusions: Kcna1-null mice experience seizures with 24-h periodicity and impaired circadian behavior. KD reduces the number and periodicity of seizures and restores normal behavioral rhythms, suggesting that this nonpharmacologic therapy may benefit biologic rhythm disturbances in epileptic patients.

Figure 1

Cosinar analyses indicate that seizures in chronically epileptic kcna1-null mice exhibit diurnal rhythmicity. Seizure incidence for each animal was tallied and binned into 4-hr periods (i.e. the ZT00:00 time-point includes seizures that occurred from ZT00:00-03:59). The data points in this figure reflect the number of seizures that occurred during each 4-hr bin relative to the total number of seizures for each animal. Seizure scores were calculated for individual mice and are expressed as the mean percent total ± SEM. The period (P) of the cosinor analysis was fixed to 24 hr. * p < 0.05 when compared to ZT00:00 using a One-factor ANOVA. ZT, zeitgeber time.

Figure 2

Kcna1-null mice have disrupted rest-activity rhythms. Representative double-plotted activity graphs reflecting activity of (A) one wild-type and (B) one kcna1-null mouse during five consecutive days. The open bars above the activity graphs indicate the light phases, during which the mice are more restful; the solid bars indicate dark phases, during which the animals are more active. The activity pattern of WT mice predictably has the highest density of activity during the dark phases (arrows), whereas the onset of rest-activity phases of Kcna1-null mice is more random. These data are presented in the line graphs below. Each peak-trough cycle reflects one 24-hr period (demarked by the vertical hashes). The horizontal blue dashed lines drawn from the peaks and troughs of the WT highlight the abnormal rest-activity cyclicity of the epileptic mice. ZT, zeitgeber time.

Figure 3

(A) Activity scores were collapsed across days into a 24-hr period and fit to a cosine function (n = 9 WT, n = 9 Kcna1-null mice). The peak activity for WT mice occurred at ZT15:28 ± 0:18, while activity of Kcna1-null mice was either phase delayed or phase advanced. (B) Kcna1-null mice had extended activity periods when compared to WT mice. Epileptic mice with advanced peak activity had a significantly longer activity period. ‘Delayed’ and ‘advanced’ labels below bar graph indicate the population of mice whose peak activity was either phase advanced or phase delayed. (C) Seizure occurrence and activity rhythms were inversely correlated. Activity scores were collapsed across days into a 24-hr period and averaged across animals in 2-hr bins (n = 9). Seizure scores are expressed as the mean percent total ± SEM (described in Methods and Figure 1 legend). WT, wild-type mice; ZT, zeitgeber time; p < 0.05.

Figure 4

Ketogenic diet (KD) treatment reduces seizure frequency. (A) The total number of seizures that occurred during light (ZT00:00-11:59) and dark phases (ZT12:00-23:59) were examined. KD-treated mice had fewer seizures than those fed a standard diet (SD). (B) Further analyses examining the average number of seizures per animal per day per 4hr time bin indicated that KD treatment specifically reduced seizures, specifically during ZT00:00-04:00. In addition, the number of seizures in KD-treated mice was similar at all times of day. * Differs significantly from KD; ** SD ZT16:00 and 20:00 differ significantly from SD ZT0:00 and 04:00; p < 0.05. ZT, zeitgeber time.

Figure 5

Rest-activity patterns are restored in epileptic mice fed the KD. (A) Representative activity and line graphs of one KD-treated epileptic mouse to show restored diurnal rest-activity patterns. (B) There were no differences in peak activity or activity period of WT and KD-treated Kcna1-null mice. ZT, zeitgeber time.