Female reproductive hormones alter sleep architecture in ovariectomized rats

Abstract

Study objectives: Treating ovariectomized rats with physiological levels of estradiol and/or progesterone affects aspects of both baseline (24 h) sleep and recovery (18 h) sleep after 6 h of sleep deprivation. We have extended the analysis of these effects by examining several additional parameters of sleep architecture using the same data set as in our previous study (Deurveilher et al. SLEEP 2009;32(7):865-877).

Design: Sleep in ovariectomized rats implanted with oil, 17 β-estradiol and/or progesterone capsules was recorded using EEG and EMG before, during, and after 6 h of sleep deprivation during the light phase of a 12/12 h light/dark cycle.

Measurements and results: During the baseline dark, but not light, phase, treatments with estradiol alone or combined with progesterone decreased the mean duration of non-rapid eye movement sleep (NREMS) episodes and the number of REMS episodes, while also increasing brief awakenings, consistent with the previously reported lower baseline NREMS and REMS amounts. Following sleep deprivation, the hormonal treatments caused a larger percentage increase from baseline in the mean durations of NREMS and REMS episodes, and a larger percentage decrease in brief awakenings, consistent with the previously reported larger increase in recovery REMS amount. There were no hormonal effects on NREMS and REMS EEG power values, other than on recovery NREMS delta power, as previously reported.

Conclusions: Physiological levels of estradiol and/or progesterone in female rats modulate sleep architecture differently at baseline and after acute sleep loss, fragmenting baseline sleep while consolidating recovery sleep. These hormones also play a role in the diurnal pattern of NREMS maintenance.

Keywords: Ovariectomy; estradiol; progesterone; sleep architecture; sleep deprivation; sleep homeostasis.

Figure 1

The number (A) and mean duration (B) of non-rapid eye movement sleep (NREMS) episodes observed in ovariectomized (OVX) female and intact male (M) rats during the baseline 12 h light (white bars) and 12 h dark phases (black bars) and during 12 h of recovery sleep opportunity (dark phase; gray bars) following 6 h of sleep deprivation. OVX female rats had been treated for two weeks with oil, low estradiol (LE), high estradiol (HE), low progesterone (LP), or low estradiol plus low progesterone (LEP) doses. The number of NREMS episodes (A) differed significantly between the light and dark phases at baseline. During the recovery dark phase, the number of NREMS episodes (A) remained similar to that during the baseline dark phase. The mean duration of NREMS episodes (B) was shorter in the HE, LEP, and M groups than in the Oil and the LP groups during the dark phase, and it was shorter in the dark vs. the light phase in the LE, HE, LEP, and M groups only. During the recovery dark phase, the mean duration of NREMS episodes (B) was significantly increased compared to the corresponding baseline in the LE, HE, LEP, and M groups only. Data are shown as means + standard errors of the mean (SEM) for 8 animals per group. ^xdifferent from baseline light phase; #different from baseline dark phase; ^adifferent from Oil; ^bdifferent from LE; ^cdifferent from HE; ^ddifferent from LP; ^edifferent from LEP; all P < 0.05 (Tukey-Kramer post hoc comparisons).

Figure 2

The number (A) and mean duration (B) of rapid-eye movement sleep (REMS) episodes during baseline recordings for the 12 h light (white bars) and 12 h dark phases (black bars) and during 12 h of recovery sleep opportunity (dark phase; gray bars) following 6 h of sleep deprivation. See the legend of Figure 1 for further information regarding the hormonal treatment groups. The number of REMS episodes (A) in the dark phase was lower in the HE, LP, LEP, and M groups compared to the Oil group under both baseline and recovery conditions. The mean duration of REMS episodes (B) was significantly longer during the recovery dark phase relative to the baseline dark phase in the HE and LEP groups only. ^xdifferent from baseline light phase; #different from baseline dark phase; ^adifferent from Oil; all P < 0.05 (Tukey-Kramer post hoc comparisons).

Figure 3

The frequency distributions of NREMS (A) and REMS (B) episodes as a function of episode duration (s) during the baseline dark phase in the Oil, LE, HE, LP, LEP, and M groups. All episodes of NREMS and REMS were divided into 9 consecutive bins of increasing duration on a logarithmic scale. There were significant group differences in the number of NREMS episodes ≤ 70 s (F_5,42 = 3.81; P < 0.01) and the number of REMS episodes ≤ 150 s (F_5,42 = 4.71, P < 0.01). Specifically, the Oil and LP groups had fewer NREMS episodes ≤ 70 s than the M group (P < 0.05 vs. M), while the HE, LEP and M groups had fewer REMS episodes ≤ 150 s than the Oil group (P < 0.05 vs. Oil). Data are shown as means ± SEM of 8 animals per group. * indicates the combined bins for which the main factor of Group was significant (P < 0.05, ANOVA).

Figure 4

The number of brief awakenings from sleep during the baseline light (white bars) and dark (black bars) phases and during the recovery dark phase (gray bars) in the Oil, LE, HE, LP, LEP, and M group. A brief awakening was defined as a 10-s epoch of wakefulness bounded by NREMS (or rarely REMS) and expressed per hour of sleep time. At baseline, the number of brief awakenings was higher in the LEP group than in the Oil and LP groups during the dark phase, and it was higher during the dark than the light phase in the LE, HE, LEP, and M groups. During the recovery dark phase, the number of brief awakenings was significantly decreased compared to the corresponding baseline in the HE, LEP, and M groups. Data are shown as means + SEM of 8 animals per group. ^xdifferent from baseline light phase; #different from baseline dark phase; ^adifferent from Oil; ^ddifferent from LP; ^edifferent from LEP; all P < 0.05 (Tukey-Kramer post hoc comparisons).

Figure 5

Power spectra of the EEG during NREMS (A) and REMS (B) in the baseline light (left) and dark (right) phases in the Oil, LE, HE, LP, LEP, and M groups. EEG power was normalized to the total power (0.5-50 Hz) for each 0.5 Hz bin in each animal. The power spectra during NREMS and REMS in the light and dark phases were similar among groups, except for a trend toward lower REMS theta power in the HE group. Data are shown as means ± SEM of 8 animals per group.

Figure 6

Changes during the recovery dark period relative to time-matched baselines in the numbers of NREMS (A) and REMS (B) episodes, the mean duration of NREMS (C) and REMS (D) episodes, and the numbers of brief awakenings (E) in the Oil (black bars), LE, HE, LP, LEP, and M (white bars) groups. The percentage increases in mean duration of both NREMS and REMS episodes were greater in the HE and LEP groups than in the Oil group. The percentage decrease in the number of brief awakenings was larger in the LEP group than in the Oil group. Data (means + SEM) are expressed as percentages of the corresponding baseline period; n = 8 per group. ^adifferent from Oil; all P < 0.05 (Tukey-Kramer post hoc comparisons).