Sleep phenotyping in a mouse model of extreme trait anxiety

Abstract

Background: There is accumulating evidence that anxiety impairs sleep. However, due to high sleep variability in anxiety disorders, it has been difficult to state particular changes in sleep parameters caused by anxiety. Sleep profiling in an animal model with extremely high vs. low levels of trait anxiety might serve to further define sleep patterns associated with this psychopathology.

Methodology/principal findings: Sleep-wake behavior in mouse lines with high (HAB), low (LAB) and normal (NAB) anxiety-related behaviors was monitored for 24 h during baseline and recovery after 6 h sleep deprivation (SD). The amounts of each vigilance state, sleep architecture, and EEG spectral variations were compared between the mouse lines. In comparison to NAB mice, HAB mice slept more and exhibited consistently increased delta power during non-rapid eye movement (NREM) sleep. Their sleep patterns were characterized by heavy fragmentation, reduced maintenance of wakefulness, and frequent intrusions of rapid eye movement (REM) sleep. In contrast, LAB mice showed a robust sleep-wake rhythm with remarkably prolonged sleep latency and a long, persistent period of wakefulness. In addition, the accumulation of delta power after SD was impaired in the LAB line, as compared to HAB mice.

Conclusions/significance: Sleep-wake patterns were significantly different between HAB and LAB mice, indicating that the genetic predisposition to extremes in trait anxiety leaves a biological scar on sleep quality. The enhanced sleep demand observed in HAB mice, with a strong drive toward REM sleep, may resemble a unique phenotype reflecting not only elevated anxiety but also a depression-like attribute.

Figure 1. Sleep-wake distribution in the mouse model of trait anxiety.

(A) Time-course changes in sleep-wake patterns and EEG delta power during NREM sleep of NAB (n = 7), LAB (n = 10) and HAB (n = 10) mice under baseline conditions. Data points represent 2 h means ± SEM of time spent in wake, NREM sleep (NREMS) and REM sleep (REMS). The delta power during NREM sleep is represented as the mean value ± SEM of normalized EEG power densities in the frequency range of 0.5–4.0 Hz restricted to NREM sleep. Two-way analysis of variance (ANOVA) revealed significant effects of ‘line’ and ‘time’ for all vigilance states and delta power (P<0.001) and their interaction for all vigilance states across 24 h (wake, P<0.001; NREMS, P<0.001; REMS, P<0.05). (B) Percentage of time spent in wake, NREM and REM sleep and the normalized EEG power of delta band during the 12 h light and dark period. Values are the 12 h means ± SEM. *P<0.05, **P<0.001, assessed by one-way ANOVA with the factor ‘line’ followed by Bonferroni's test. The light-dark (LD) differences were analyzed as well. Two-way ANOVA revealed significant effects of ‘line’ and ‘LD interval’ for all vigilance states (P<0.001) and their interaction for all vigilance states (P<0.05). The direct comparisons between lines of the 12 h LD dynamics in circadian amplitudes were performed by one-way ANOVA with the factor ‘line’ followed by Tukey's test (wake, NREMS, P<0.001; REMS, P<0.05).

Figure 2. Sleep architecture of the mouse model of trait anxiety.

(A) Representative 24 h hypnograms under baseline conditions of a NAB, LAB, or HAB mouse. The white-black bars below the hypnograms indicate the 12 h light-dark cycle. (W: wakefulness, N: NREM sleep, R: REM sleep) (B – D) Average numbers (B) and durations (C) of wakefulness, NREM sleep (NREMS) and REM sleep (REMS) bouts and state transitions (D) during the 12 h light and dark period. Values are the means ± SEM. NAB (n = 7), LAB (n = 10), and HAB (n = 10). Left and right panels indicate values for the light and dark periods, respectively. *P<0.05, **P<0.001, assessed by one-way ANOVA with the factor ‘line’ followed by Bonferroni's test.

Figure 3. EEG spectra during wakefulness, NREM and REM sleep in the mouse model of trait anxiety.

NAB (n = 7), LAB (n = 10), HAB (n = 10). (A) Spectral distribution of EEG power densities averaged from all 4-s epochs scored as wake, NREM sleep (NREMS) and REM sleep (REMS) over 24 h of recording under baseline conditions. The power of every 0.25 Hz bin was first averaged across the individual vigilance state and then normalized as a group by calculating the percentage of each bin from the total power across 24 h (0.25–32 Hz). All figures show spectral distributions of the EEG power density in the frequency range of 0.25 to 25 Hz. Two-way ANOVA with the factors ‘frequency’ and ‘line’ followed by Bonferroni's test revealed significant effects of ‘frequency’ and an interaction between ‘frequency’ and ‘line’ for each vigilance state (P<0.0001), whereas the ‘line’ effects were not significant. For REM sleep, the EEG theta peak frequency, between 6 and 9 Hz, is shown in the insertion. *P<0.05, assessed by one-way ANOVA with the factor ‘line’ followed by Bonferroni's test. (B) Comparisons of EEG power density within the delta (0.5–5 Hz), theta (6 – 9 Hz), sigma (10 – 15 Hz), and beta (16 – 23 Hz) bands in the 12 h light and dark period for each vigilance state. Data are presented as the 12 h means ± SEM. The EEG power of each frequency band was normalized by total EEG power during 24 h. *P<0.05, **P<0.001, assessed by one-way ANOVA with the factor ‘line’ followed by Bonferroni's test.

Figure 4. Effects of sleep deprivation on sleep recovery and EEG delta power changes during NREM sleep.

(A) Time-course changes in NREM sleep (NREMS) and REM sleep (REMS) and the delta power of slow-wave activity (0.5–5 Hz) during NREM sleep are expressed as the 2 h means ± SEM during baseline recordings (closed circles) and sleep deprivation (SD, open circles) in NAB (n = 7), LAB (n = 10), HAB (n = 10) mice. SD began at the onset of the light period and lasted for 6 h. Two-way ANOVA among the three mouse lines revealed a significant effect of ‘line’ on EEG delta power in NREM sleep during the post-SD 6–12 h during the light period (P<0.001) and on NREM and REM sleep and EEG delta power during the post-SD 13–18 h during the dark period (P<0.05, P<0.05, P<0.001). * P<0.05, ** P<0.001, assessed by two-way ANOVA with the factors ‘line’ and ‘interval’ followed by Bonferroni's test.