Sleep/Wake Behavior and EEG Signatures of the TgF344-AD Rat Model at the Prodromal Stage

Abstract

Transgenic modification of the two most common genes (APPsw, PS1ΔE9) related to familial Alzheimer's disease (AD) in rats has produced a rodent model that develops pathognomonic signs of AD without genetic tau-protein modification. We used 17-month-old AD rats (n = 8) and age-matched controls (AC, n = 7) to evaluate differences in sleep behavior and EEG features during wakefulness (WAKE), non-rapid eye movement sleep (NREM), and rapid eye movement sleep (REM) over 24-h EEG recording (12:12h dark-light cycle). We discovered that AD rats had more sleep-wake transitions and an increased probability of shorter REM and NREM bouts. AD rats also expressed a more uniform distribution of the relative spectral power. Through analysis of information content in the EEG using entropy of difference, AD animals demonstrated less EEG information during WAKE, but more information during NREM. This seems to indicate a limited range of changes in EEG activity that could be caused by an AD-induced change in inhibitory network function as reflected by increased GABAAR-β2 expression but no increase in GAD-67 in AD animals. In conclusion, this transgenic rat model of Alzheimer's disease demonstrates less obvious EEG features of WAKE during wakefulness and less canonical features of sleep during sleep.

Keywords: Alzheimer’s; electroencephalography; sleep architecture.

Figure 1

Hypnogram of the AC (top) and AD (bottom) animals over the entire 24 h observation period derived from the 10 s scoring with lights turned off after the 12th hour. White indicates WAKE, light blue (AC) or light red (AD) indicates NREM, and blue (AC) or red (AB) indicates REM. The x-axis is Zeitgeber time (ZT) with ZT = 0 indicating the start of the inactive “lights-on” phase at 7 am. The grey rectangle indicates the active “lights-off” phase. #: number

Figure 2

There were no significant differences in (A) WAKE (F_WAKE = 0.03, p = 0.857), (B) NREM sleep (F_NREM = 0.41, p = 0.522), and (C) REM sleep (F_REM = 3.00, p = 0.085) between the control (AC, blue, n = 7) and Alzheimer’s (AD, red, n = 8) animals. The distribution plots were derived from the 10 s scoring. The x-axis is Zeitgeber time (ZT) with ZT = 0 indicating the start of the inactive “lights-on” phase at 7 am. NREMS: NREM sleep; REMS: REM Sleep.

Figure 3

Cumulative probability plots of the different vigilance states derived from the 10 s scoring in logarithmic presentation of the inactive (left) and active (right) periods for (A) WAKE, (B) NREM sleep (NREMS), and (C) REM sleep (REMS). Blue: AC group, red: AD group. The bout lengths of WAKE (inactive) and NREM sleep (active and inactive) are significantly different among the groups. Except for the WAKE bouts in the active phase, the bout length distributions are significantly different. In general, the AC graphs are approaching cumulative probability 1 more gradually indicating a more uniform distribution of bout lengths than in the AD rats. The red and blue areas indicate the bout duration corridor; where 80% of the bouts were less than that time epoch.

Figure 4

Dot plots inclusive mean (dashed line) and median (solid line) of transitions between WAKE and SLEEP of the AC group (n = 7, blue) and the AD (n = 8, red) group. (A) The plot represents the transition for the entire 12 h of the active or inactive period, scored in 10 s epochs. For the inactive period, we derived AUC values reflecting a fair effect. For the WAKE to SLEEP transition, AUC was 0.71 (95% CI: 0.41–0.96). (B) For the 2 h episodes (scored in 4 s epochs) used for quantitative EEG analysis, the AD animals showed a higher number of transitions as well.

Figure 5

Plots of the median of the relative spectral power for the control group (AC, n = 7, for REM: n = 5, blue) and the Alzheimer’s (AD, n = 8, for REM: n = 6, red) group as recorded from the rostral electrode. The light blue and red shaded areas indicate the median absolute deviation. The accompanying AUC plots including the 95% confidence intervals represent the results of the statistical analysis. Dots indicate the calculated AUC, x signs indicate the limits of the 95% confidence intervals. A filled dot indicates a significant difference as determined by the 95% confidence interval exclusive 0.5. The areas in gray indicate an AUC > 0.7, i.e., an AUC value corresponding to an at least acceptable effect. REMS: REM sleep; NREMS: NREM sleep

Figure 6

Comparison of entropy of difference (EoD) between WAKE and NREM sleep (NREMS). Combined box and dot plots of the EoD (τ = 1), the comparison within groups between WAKE and NREM sleep. Blue indicates the control (AC) and red—the Alzheimer’s (AD) group. We observed significant EoD differences (as indicated by AUC) in the AC group, but not in the AD group. The “+” indicate outlier for the box plot.

Figure 7

Dot plots of the relative change in protein expression normalized to beta-actin expression. (A) GAD-67 expression was not different between the groups indicating no significant change (AUC: 0.60, 95% CI: 0.2 to 1) in the amount of GABA produced; note that the highest value in the AC group and the lowest value in the AD group could be considered outliers as evaluated by the Grubbs’s test. (B) GABA_AR β2 expression was significantly (AUC = 1) higher in the AD group. This indicates an increased number of GABA_AR in the AD group. * indicates significance

Figure 8

Exemplary EEG raw traces for the animals from (A) the age-matched controls and (B) the AD rats at different vigilance states—WAKE, NREM sleep, and REM sleep—that were recorded from the caudal or rostral EEG lead. NREMS: NREM sleep; REMS: REM sleep.