A Dopamine Receptor D2-Type Agonist Attenuates the Ability of Stress to Alter Sleep in Mice

Abstract

Although sleep disruptions that accompany stress reduce quality of life and deteriorate health, the mechanisms through which stress alters sleep remain obscure. Psychological stress can alter sleep in a variety of ways, but it has been shown to be particularly influential on rapid eye movement (REM) sleep. Prolactin (PRL), a sexually dimorphic, stress-sensitive hormone whose basal levels are higher in females, has somnogenic effects on REM sleep. In the current study, we examined the relationship between PRL secretion and REM sleep after restraint stress to determine whether: 1) the ability of stress to increase REM sleep is PRL-dependent, and 2) fluctuating PRL levels underlie sex differences in sleep responses to stress. Because dopamine D2 receptors in the pituitary gland are the primary regulator of PRL secretion, D2 receptor agonist, 1-[(6-allylergolin-8β-yl)-carbonyl]-1-[3-(dimethylamino) propyl]-3-ethylurea (cabergoline), was used to attenuate PRL levels in mice before 1 hour of restraint stress. Mice were implanted with electroencephalographic/electromyographic recording electrodes and received an ip injection of either 0.3-mg/kg cabergoline or vehicle before a control procedure of 1 hour of sleep deprivation by gentle handling during the light phase. Six days after the control procedure, mice received cabergoline or vehicle 15 minutes before 1 hour of restraint stress. Cabergoline blocked the ability of restraint stress to increase REM sleep amount in males but did not alter REM sleep amount after stress in females even though it reduced basal REM sleep amount in female controls. These data provide evidence that the ability for restraint stress to increase REM sleep is dependent on PRL and that sex differences in REM sleep amount may be driven by PRL.

The mechanisms that underlie the ability of stress to cause sleep disruptions are not well understood. In 1991, it was reported that restraint, a psychological stressor, causes an increase in rapid eye movement (REM) sleep in rats (1). In rats, 2 hours of restraint stress applied near the onset of activity increases REM sleep amount, and to a lesser degree non-REM (NREM) sleep amount, during the remainder of the active phase (1). A potential mechanism for this effect of stress on REM sleep was revealed by a study examining the ability of restraint stress to increase REM sleep in 2 mouse strains, C57BL/6J and BALB/cJ (2). In the study, the ability of restraint stress to increase REM sleep in C57BL/6J male mice was accompanied by a concomitant increase in serum prolactin (PRL) levels, but in the BALB/cJ male mice, restraint stress had no effects on REM sleep or PRL (2). These studies raise the possibility that PRL may mediate the effects of stress on sleep. Notably, PRL is one of a small number of “sleep” hormones or hormones whose daily secretion depends on sleep. Unlike most rhythmic hormones whose secretion is under the control of circadian timing, PRL exhibits peaks in secretion only during extended bouts of sleep (3). Furthermore, PRL is a somnogenic hormone that promotes REM sleep. In rats, systemic or intrahypothalamic injections of PRL acutely enhance REM sleep (4), and the ability of vasoactive intestinal polypeptide to increase REM sleep is inhibited by PRL immunoneutralization (5).

In adult rats and mice, basal PRL levels are higher in females than in males. These sex differences are driven primarily by the positive effect of gonadal estradiol, which acts at the pituitary gland to increase PRL release (6, 7). Restraint stress also increases PRL levels in female rats (5) and mice (2). However, in females, this response seems to be blunted. To date, only 1 published study has sought to directly examine this sex difference in the PRL response to restraint (8). This study found that after 60 minutes of restraint, PRL release was greater in male rats. Furthermore, this effect was to a degree that, after restraint, the sex difference in PRL levels was eliminated. Interestingly, mice also exhibit a sex difference in the REM sleep in response to restraint stress, where stress-induced REM sleep is greater in males than females (9, 10).

We hypothesized that the ability of stressful stimuli to increase REM sleep amount is PRL dependent. In this study, we first confirmed previously discussed findings demonstrating that both PRL and REM sleep exhibit sex differences after restraint stress. We then used the D2 receptor agonist, 1-[(6-allylergolin-8β-yl)-carbonyl]-1-[3-(dimethylamino) propyl]-3-ethylurea (cabergoline), which offers a direct means of PRL inhibition, to test whether PRL is required for the increase in REM sleep after restraint stress. Pituitary PRL is regulated by the inhibitory actions of dopamine released from efferents of the arcuate nucleus of the hypothalamus. This dopamine acts directly on D2 receptors in lactotrophic cells of the anterior pituitary. Cabergoline, a dopamine D2 receptor agonist, inhibits PRL release, reducing circulating PRL levels. The effect has been reported to last for up to 6 days (11).

Materials and Methods

Animals

Female and male C57BL/6J mice were purchased from The Jackson Laboratory and maintained by the Center for Laboratory Animal Resources at Morehouse School of Medicine. Mice were housed individually and maintained under a 12-hour light, 12-hour dark cycle at a temperature of 23.0°C with food and water available ad libitum. All protocols and procedures used in this project were in accordance with principles outlined in the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by Morehouse School of Medicine Institutional Animal Care and Use Committee.

Surgeries

Under isofluorane anesthesia, mice (10–14 wk of age) received electrodes for electroencephalographic (EEG) and electromyographic (EMG) recording by either a telemetric (4 females, 5 males) or tethered (5 females, 6 males) recording system. For telemetric recording, 2 stainless steel recording screws (SmallParts, Inc) were positioned contralaterally to each other on the skull surface. The first was located 1 mm anterior to bregma and 0.5 mm right of the central suture, whereas the second was located 0.5 mm posterior to λ and 1 mm left of the central suture. EMG electrodes were placed in the nuchal muscle. EEG and EMG electrodes leads were attached to a telemetric transmitter (Data Systems International) surgically implanted into a sc “pocket” made from the electrode incision over the skull and extending caudally along the dorsal surface of the thorax. For tethered recordings, animals received a prefabricated head mount (Pinnacle Technologies) as a part of an EEG/EMG tether electrode recording system. Four screw electrodes were positioned using this headmount: 2 contralateral screws (2.5 mm in length) over the frontal cortex and 2 screws (3.0 mm in length) contralateral over the parietal cortex (1.0 mm anterior to bregma/0.5 mm right of the central suture and 0.5 mm posterior to λ/1.0 mm left of the central suture). Two Teflon-coated stainless steel wires served as EMG electrodes and were placed bilaterally in the nuchal muscle, 1.5–2.5 mm from the midline. Animals were allowed a minimum of 14 days to recover from the surgical procedure before entering any experiment.

Sleep-wake (EEG/EMG) recordings

Female mice underwent sleep recording and treatment in an environmental chamber separate from males. Animals recorded using the tethered system were transferred to a recording chamber and attached to a low-resistance commutator mounted over the cage (Pinnacle Technologies). Five to seven days of habituation to the recording chamber and tether (if present) were allowed before the experiment. Tethered recordings were acquired and analyzed using Sirenia software (Pinnacle Systems). Animals implanted with telemeters were recorded via a platform receiver placed under the cage. Waveforms in these animals were recorded and analyzed using Neuroscore software (Data Sciences International) (Figure 1).

Figure 1.

Representative EEG and EMG recordings are shown for each vigilance state: 1) wake, 2) NREM sleep, and 3) REM sleep. Each tracing represents a 10-second time span. Wake consists of a high-amplitude EMG and mixed-frequency, low-amplitude EEG. NREM sleep consists of a low-amplitude EMG and low-frequency, high-amplitude EEG. REM sleep consists of a very low-amplitude EMG and high-frequency, low-amplitude EEG.

Experimental paradigms

Sleep deprivation

A sleep deprivation procedure was performed for 1 hour from zeitgeber time (ZT) 5 to ZT 6 (5 h after light onset) to control for sleep loss that occurs during restraint stress. Sleep deprivation via gentle handling was accomplished by tapping on the outside of, or placing novel objects in, the cage (12). This is the control condition for stress. Mice had free access to food and water during sleep deprivation.

Restraint stress

Restraint stress was performed for 1 hour from ZT 5 to ZT 6. Mice were confined to a 50-mL conical tube or a plastic restraint cone taped at the tail end. The mice remained in their home cage during restraint.

Drug

Cabergoline was generously supplied by Pfizer. Additional cabergoline was purchased from Sigma-Aldrich. Cabergoline was dissolved in 100% pharmasolve and then diluted with 20% β-cyclodextrin in water to yield a final concentration of 0.15–0.5-mg/mL cabergoline. Mice received a 0.3-mg/kg ip injection of cabergoline or vehicle. All drugs were prepared within 48 hours of experiment and stored at 4°C. Solutions were allowed to reach at room temperature before injection.

Enzyme-linked immunosorbent assays

Quantification of PRL levels was performed using species (mouse/rat) specific PRL ELISA kits (Calbiotech) in accordance with manufacturer's protocol. Briefly, trunk or submandibular blood was collected in dry ice-chilled centrifuge tubes containing EDTA. The blood was centrifuged at 4°C for 15 minutes at 2600g, and the supernatant was stored at −80°C for later analysis. Samples were run in triplicate and values averaged.

The experimental protocol was a 3-factor design, in which male and female mice were subjected to 1 of 2 conditions (restraint stress or control) and underwent potential pharmacologic intervention with either vehicle or the dopamine agonist, cabergoline.

Data analysis and power spectra

Four-second and 10-second epochs EEG and EMG recordings were hand scored by a trained analyst as wake (low-voltage, high-frequency EEG; high-amplitude EMG), NREM sleep (high-voltage, mixed-frequency EEG; low-amplitude EMG), or REM sleep (mixed-frequency EEG with a predominance of θ activity [6–10 Hz]; very low-amplitude EMG). Raw EEG epochs scored as NREM sleep and REM sleep were subjected to Fast Fourier Transform in order to determine spectral power in the δ (0.5–4 Hz) and the θ (5–8 Hz) band. Power values in each band are expressed relative to total EEG power.

Sleep fragmentation was also analyzed in this study. Sleep fragmentation is characterized by sleep interruptions and arousals during sleep opportunities, which can account for fluctuations in REM and NREM sleep amount. Because a single sleep cycle consists of a continuous bout of NREM sleep followed by a bout of REM sleep and then a brief arousal, the duration and frequency of sleep cycles can be effective measures of sleep fragmentation. In fact, the primary measures of sleep fragmentation over a given time period include: 1) the number of continuous bouts of NREM and REM sleep, 2) the mean duration of NREM and REM sleep bouts, 3) the number of brief arousals, and 4) the number of transitions (stage shifts) between sleep-wake states.

θ Power (5–7 Hz) is an important marker of REM sleep in humans and rodents. However, θ waves can also be seen in rodents during active movement (13, 14). θ Bands are generated in the hippocampus, medial thalamus, and other cortical regions.

Statistics

Statistical comparisons were made using ANOVA with full-factorial designs. Mouse sex, drug-vehicle, restraint-gentle handling, and phase were used as fixed factors in four-way designs for all response variables except PRL concentration; for PRL concentrations, phase was not included in the analysis, and three-way ANOVA were used to test for differences in means. Post hoc comparisons were made using the Tukey's honest significant difference procedure or Student's t tests for dependent or independent samples, when indicated. Student's t test for dependent or independent samples was used, when indicated. A significance level of P < .05 was set for all full model (ie, ANOVA) comparisons and Tukey's honest significant difference. Dunn-Sidak corrections were used when making pair-wise comparisons with Student's t tests. Data are presented as mean ± SEM.

Results

Reducing PRL secretion attenuates REM sleep amount in restrained males

Consistent with previous reports (9), restraint stress increased REM sleep amount in male mice during the subsequent 18 hours of recovery sleep from ZT 6 to ZT 24 (Figure 2A). Most this increase occurred during the dark phase. The primary variables tested in this investigation were sex (male/female), treatment (restraint/control), drug (cabergoline/vehicle), and phase (light/dark). Four-way ANOVA revealed a main effect of treatment (F_(1,83) = 6.20, P = .014). The effect of drug was not slight but not significant (F_(1,83) = 3.75, P = .056). However, there were significant interactions; sex × treatment: F_(1,83) = 6.96, P = .010; sex × treatment × drug: F_(1,83) = 8.52, P = .004; and sex × treatment × drug × phase: F_(1,83) = 4.21, P = .043. The main effect of stress revealed that restrained mice had more REM sleep than handled controls. Follow up post hoc tests of the interactions revealed that restrained males had 40% (24.55 min) more REM sleep during the dark phase than handled controls (t₁₁ = 2.47, P = .031). Restrained males injected with cabergoline (dopamine D2 agonist) had less REM sleep (Figure 2A) during the 18-hour recovery period than restrained males treated with vehicle.

Figure 2.

Sleep in mice. REM Sleep in males (A), REM sleep in females (B), NREM sleep in males (C), and NREM sleep in females (D). C57BL/6J mice received vehicle or cabergoline (ip) at ZT 4.75 followed by 1 hour of gentle handling or 1 hour of restraint stress from ZT 5 to ZT 6. Bars represent the 6-hour light, 12-hour dark phase and 18-hour total after treatment. M, male; F, female; VH, vehicle-treated and gentle handling; VR, vehicle-treated and restraint; CR, cabergoline-treated and gentle handling; CR, cabergoline-treated and restraint. *, P < .05 vehicle-treated restrained animals vs cabergoline-treated restrained animals; **, P < .005; ***, P < .001; ^, P < .05 vs vehicle-treated animals; +, P < .05 cabergoline-treated handled animals vs cabergoline-treated restrained animals.

This effect of cabergoline was apparent during dark but not during the light phase. During the light phase, REM sleep amount in cabergoline-injected restrained males was not different than in vehicle-injected males (t₁₁ = 0.24, P = .815). During the dark phase, however, cabergoline-injected restrained males had 46.1% (28.6 min) less REM sleep than vehicle-injected males (t₁₁ = 2.37, P = .037). Cabergoline had no effects on REM sleep in handled male mice (Figure 2A). REM sleep amount in cabergoline-injected handled males was similar to that of vehicle-injected handled males during the light (t₁₁ = 0.597, P = .562) and dark (t₁₁ = 0.265, P = .456) phases.

Cabergoline reduced baseline REM sleep in females

Cabergoline reduced REM sleep amount in female controls (Figure 2B). Cabergoline-injected females that were handled had 50.66% (23.6 min) less REM sleep during the dark phase (t₁₂ = 4.27, P = .001) than vehicle-injected handled females (Figure 2B). During the light phase, there were no differences in REM sleep amount (t₁₂ = 0.262, P = .953) between these groups. Cabergoline had no influences in restrained females. Female REM sleep amounts were not different (light phase: t₁₂ = 0.148, P = .885 and dark phase: t₁₂ = 0.509, P = .620 from vehicle-injected females after restraint) (Figure 2B). Restraint stress had no effects on REM sleep amount during the light (t₁₂ = 0.022, P = .983) or dark (t₁₂ = 0.024, P = .981) phases in vehicle-injected females (Figure 2B). During the light phase, restraint had no effect (t₁₂ = 0.021, P = .935) in cabergoline-injected females, but during the dark phase, restrained females had 57.72% (31.17 min) more REM sleep (t₁₀ = 3.50, P = .006) than female controls (Figure 2B). It is important to note that this effect was produced by the decrease of REM sleep by cabergoline in female controls.

Cabergoline increased NREM sleep amount in male controls

Restraint stress had no influences on NREM sleep amount in male mice. Four-way ANOVA revealed a main effect of sex (F_(1,84) = 4.48, P = .037) and no other effects or interactions. The effect of sex revealed that male mice had (51.1 min) more NREM sleep than females. This finding has been reported previously (19). NREM sleep amount in restrained males was not different from handled males during the light or dark phases (Figure 2C). Cabergoline had no effects on NREM sleep amount (Figure 2C), or total sleep amount in restrained or control males during 18 hours of recovery.

Cabergoline had no effect on NREM sleep in females

Cabergoline had no influences on NREM sleep amount in restrained females during the light or dark phases (Figure 2D). Restraint stress, when compared with handling, had no effects on NREM sleep amount in females during the light or dark phases (Figure 2D). Cabergoline-injected, handled females did not exhibit any changes in NREM sleep during the light or dark phases when compared with vehicle-injected handled females (Figure 2D), although they had lower levels of REM sleep (Figure 2B).

Cabergoline increased REM sleep bout number in restrained females and decreased NREM sleep bouts in control females

Cabergoline increased REM sleep bout number in restrained females (F_(1,11) = 5.913, P = .033). Specifically, cabergoline-injected female handled controls had 77.56% fewer REM sleep bouts than cabergoline-injected females that were restrained (Figure 3B). The most significant reduction in REM sleep bout number occurred during the light phase, in which cabergoline-injected female handled mice had 67.3% less REM sleep bouts (F_(1,11) = 12.892, P = .004) than cabergoline-injected females that were restrained, although the greatest number in reduction of REM sleep bouts occurred during the dark phase (82.3% fewer REM sleep bouts; F_(1,11) = 3.667, P = .082). Cabergoline-injected handled females had fewer REM sleep bouts than vehicle-injected handled females during the light (F_(1,11) = 5.746, P = .035) and dark (F_(1,11) = 7.929, P = .017) phases (Figure 3B).

Figure 3.

REM sleep bouts and REMS bout duration in mice. REMS bouts in male mice (A), REMS bouts in female mice (b), REMS bout duration in male mice (C), and REMS bout duration in female mice (D). *, P < .05 restrained animals; +, P < .05 between cabergoline-treated (cabergoline-treated handled and cabergoline-treated restrained) animals; #, P < .05 between handled (vehicle-treated handled and cabergoline-treated handled) animals; double indicators, P < .01; triple indicators, P < .001.

Cabergoline did not alter REM sleep bout duration in restrained females during the light (F_(1,13) = 0.060, P = .811) or dark (F_(1,13) = 1.018, P = .331) phases (Figure 3D). Cabergoline did not alter REM sleep bout duration in handled females during the light (F_(1,11) = 0.535, P = .480) or dark (F_(1,11) = 0.003, P = .960) phases (Figure 3D). Restraint did not alter REM bout duration in cabergoline-injected females during the light (F_(1,11) = 0.002, P = .970) or dark (F_(1,11) = 2.006, P = .184) phases (Figure 3D). Restraint did not alter REM bout duration in vehicle-injected females during the light (F_(1,13) = 2.124, P = .169) or dark (F_(1,13) = 0.076, P = .786) phases (Figure 3D). Consequently, there was no net change in REM sleep amount.

There were effects of cabergoline on the number of NREM sleep bouts in handled females during the light (F_(1,10) = 5.012, P = .049) but not dark (F_(1,10) = 4.706, P = .055) phases (Figure 4B). Overall, female handled mice that received cabergoline had 44.96% fewer NREM sleep bouts (F_(1,10) = 5.632, P = .039) than those that received vehicle (Figure 4B). However, this effect was offset by a reduction in mean NREM duration in cabergoline-injected handled females (Figure 4D), who had shorter (F_(1,11) = 18.251, P = .001) NREM sleep bouts during the dark phase than vehicle-injected controls. There were no effects of cabergoline on the number of NREM sleep bouts in restrained females during the light (F_(1,13) = 0.002, P = .968) or dark (F_(1,13) = 0.813, P = .384) phases (Figure 4B). There were no effects of cabergoline on NREM sleep bout duration in restrained females during the light (F_(1,13) = 0.003, P = .955) or dark (F_(1,13) = 0.227, P = .640) phases (Figure 4D).

Figure 4.

NREM sleep bouts and NREMS bout duration in mice. NREMS bouts in male mice (A), NREMS bouts in female mice (B), NREMS bout duration in male mice (C), and NREMS bout duration in female mice (D). +, P < .05 vs cabergoline-treated animals, two-way ANOVA; #, P < .05 vs handled animals, two-way ANOVA; +++, P < .001, two-way ANOVA.

Restraint stress increased NREM sleep bout number during the light (F_(1,9) = 5.044, P = .05) and dark (F_(1,9) = 6.439, P = .032) phases in cabergoline-injected females (Figure 4B). This effect was offset by a reduction in NREM sleep bout duration during the dark phase (F_(1,10) = 6.054, P = .034) in restrained females that received cabergoline. However, there was no effect (F_(1,10) = 3.058, P = .111) of restraint during the light phase (Figure 4D). There were no effects of restraint on NREM sleep bout number during the light (F_(1,13) = 0.250, P = .625) or dark (F_(1,13) = 0.317, P = .583) phases or mean NREM sleep bout duration during the light (F_(1,12) = 0.026, P = .874) or dark (F_(1,12) = 3.795, P = .075) phases in vehicle-injected females (see figure 6 below).

Cabergoline decreased bouts of REM sleep in restrained males and had no effect on NREM sleep bouts.

Cabergoline significantly reduced the number of REM sleep bouts only in restrained males (main effect of treatment; F_(1,10) = 5.91, P = .033). The restrained group treated with cabergoline was significantly reduced compared with both restrained animals treated with vehicle and control animals treated with cabergoline. There were no differences in NREM bout numbers between any of the male groups.

Restraint stress increased the number of brief arousals and stage shifts in cabergoline-injected females

Restraint stress increased brief arousals in cabergoline-injected females. Cabergoline-injected restrained females had 58.7% more brief arousals (F_(1,10) = 5.984, P = .034) than cabergoline-injected handled females (Figure 5B). This was mainly due to differences during the dark phase (F_(1,10) = 9.885, P = .010). The number of brief arousals in the light phase in the cabergoline-injected restrained females were slightly but not significantly higher (F_(1,10) = 2.630, P = .136) (Figure 5B) than in cabergoline-injected controls. Restraint stress had no effects on brief arousals in vehicle-injected females during the light (F_(1,13) = 1.872, P = .194) or dark (F_(1,13) = 1.341, P = .268) phases (Figure 5B). Cabergoline did not significantly alter the number of brief arousals between handled (light [F_(1,11) = 1.196, P = .297], dark [F_(1,11) = 1.148, P = .307]) or restrained (light [F_(1,11) = 0.555, P = .469], dark [F_(1,11) = 0.399, P = .539]) females (Figure 5B).

Figure 5.

Number of brief arousals in (A) male and (B) female mice. Number of stage shifts in (C) male and (D) female mice. +, P < .05 between cabergoline-treated (cabergoline-treated handled and cabergoline-treated restrained) animals; ^, P < .05 between vehicle-treated (vehicle-treated handled and vehicle-treated restrained) animals; double indicators, P < .01.

The effect was almost identical for transitions (stage shifts) between sleep-wake states (Figure 5D). Cabergoline-injected females that were restrained had 75.45% more stage shifts than cabergoline-injected and handled females (Figure 5D). This effect was due mainly to significant differences in the number of stage shifts during the dark phase (F_(1,9) = 6.186, P = .035) and in total (F_(1,9) = 6.641, P = .030). This effect is consistent with the increase in REM sleep amount and REM sleep bout number observed in this group. The effects of restraint stress on stage shifts in cabergoline-injected females during the light phase (F_(1,9) = 4.849, P = .055) were close to but not significant. There were no effects of restraint stress on stage shifts in vehicle-injected females during the light (F_(1,13) = 0.741, P = .405) or dark (F_(1,13) = 0.403, P = .537) phases (Figure 5D). Cabergoline did not alter stage shifts in restrained (light [F_(1,11) = 0.233, P = .639], dark [F_(1,11) = 0.958, P = .349]) female mice. Cabergoline did impact stage shifts in handled female mice overall (F_(1,9) = 6.056, P = .036), due mostly to changes that took place during the light phase (light [F_(1,9) = 5.784, P = .040], dark [F_(1,9) = 4.001, P = .077]) (Figure 5D).

Restraint stress increased the number of brief arousals in vehicle-injected males

Brief arousals were significantly increased in restrained males treated with vehicle when compared with handled animals vehicle-treated (main effect of treatment; F_(1,10) = 5.63, P = .023). This effect was only significant in the dark phase. This increase during the dark phase after restraint was not significant in cabergoline-treated animals. No significant differences in stage shifts were found between the male treatment groups.

Restraint stress increased REM θ power in cabergoline-injected females

Cabergoline had no influences on NREM sleep δ power in restrained (F_(1,5) = 0.042, P = .845) or handled (F_(1,6) = 0.739, P = .423) males (Figure 6A). Despite the negative effects of cabergoline on REM sleep in restrained males, spectral analysis of EEG during REM sleep did not reveal any effects of cabergoline on spectral power in the predominant bandwidth for the θ sleep state in restrained (F_(1,5) = 0.275, P = .623) or handled (F_(1,6) = 0.003, P = .960) males (Figure 6C). There were no effects of restraint stress on NREM sleep δ power in cabergoline-injected (F_(1,7) = 0.059, P = .819) or vehicle-injected (F_(1,7) = 2.258, P = .177) male mice (Figure 6A). There were no effects of restraint stress on REM sleep θ power in cabergoline-injected (F_(1,4) = 0.682, P = .455) or vehicle-injected (F_(1,7) = 0.681, P = .436) males (Figure 6C).

Figure 6.

Relative δ and θ power. δ Power in males (A), δ power in females (B), θ power in males (C), θ power in females (D). *, P < .05 between restrained (vehicle-treated restrained and cabergoline-treated restrained) animals; ^, P < .05 between vehicle-treated (vehicle-treated handled and vehicle-treated restrained) animals. Double indicators, P < .01.

Restraint stress increased REM θ power in cabergoline-injected females

Cabergoline had no influences on NREM sleep δ power in restrained (F_(1,11) = 0.812, P = .387) or handled (F_(1,11) = 0.033, P = .858) females (Figure 6B). However, cabergoline affected restrained females resulting in higher REM sleep θ power overall (F_(1,9) = 7.277, P = .024) compared with vehicle-injected restrained females. Cabergoline did not affect handled females (F_(1,10) = 4.826, P = .053) (Figure 6D). There were no effects of restraint stress on NREM sleep δ power in cabergoline-injected (F_(1,10) = 0.003, P = .961) or vehicle-injected (F_(1,12) = 0.473, P = .505) females (Figure 6B). Additionally, restraint stress had no effect on REM sleep θ power in cabergoline-injected (F_(1,10) = 2.757, P = .128) females, but restraint stress did decrease REM sleep θ power in vehicle-injected females (F_(1,10) = 4.668, P = .041) (Figure 6D).

Cabergoline reduced PRL levels in male and female mice

We examined the effect of cabergoline on PRL levels in naïve mice that were not subjected to any treatments. In female mice, cabergoline injection reduced baseline PRL levels 68.6% (F_(1,12) = 12.079, P = .005) from 16.2 ± 3.8 to 5.1 ng/mL within 2 hours of injection. After a 7-day recovery period, PRL levels returned to values that were not different from baseline (20.7 ± 4.7 ng/mL; F_(1,12) = 2.042, P = .179). In male mice, cabergoline reduced baseline PRL levels (98.5%; F_(1,6) = 13.192, P = .011) from 5.8 ± 1.3 to 0.08 ng/mL within 2 hours of injection. After a 7-day recovery period, PRL levels returned to values that were not different from baseline (5.0 ± 0.60 ng/mL; F_(1,6) = 0.715, P = .430).

Next, we measured the effects of cabergoline on PRL under our experimental conditions. Three-way ANOVA revealed a main effect of sex (F_(1,150) = 10.02, P = .002), a main effect of treatment (F_(1,150) = 7.10, P = .009), and a main effect of drug (F_(1,150) = 88.67, P < .001) (Figure 7). These effects reveal that in all mice from all groups, all females had 17.24% higher PRL concentrations than all males, all restrained mice had 12.03% higher PRL concentrations than all handled controls, and all cabergoline-injected mice had 42.87% lower PRL concentrations than vehicle-injected mice (Figure 7). Three-way ANOVA also revealed no interactions, although there was a slight trend when we tested sex × drug (F_(1,150) = 3.78, P = .054).

Figure 7.

PRL concentrations in (A) male mice or (B) female mice. C57BL/6J mice received a vehicle or cabergoline ip injection at ZT 4.75 followed by 1 hour of restraint stress or gentle handling from ZT 5 to ZT 6. ***, P < .001.

Discussion

Sex differences in the REM sleep response to restraint stress are due to differences in PRL levels between males and females

We confirmed that restraint stress increased REM sleep amount in male mice. In female, however, we observed an increase in REM sleep amount after restraint stress only in animals treated with cabergoline. Hence, after restraint stress, the ability of cabergoline to alter REM sleep exhibited a sex difference. In males, cabergoline reduced REM sleep amount after restraint stress. The opposite occurred in females where cabergoline reduced REM sleep amount in handled controls but not after restraint stress. Interestingly, PRL levels in restrained males were similar to those in female controls (Figure 7). Additionally, the total REM sleep amount in cabergoline-injected female controls was less than the total REM sleep amount in vehicle-injected male controls (Figure 7). These data indicate that when PRL levels are reduced in females, sex differences in REM sleep are revealed. That is, in females, cabergoline unmasks a sex difference in the REM-response to restraint stress.

In males, cabergoline did not influence REM sleep in controls, but it blocked the ability of restraint stress to increase REM sleep. These findings provide strong evidence that basal REM sleep amount in males is not dependent on PRL. However, the ability of stress to increase REM sleep amount is PRL dependent.

The effects of PRL on sleep-wake fragmentation were more apparent in female mice than in males. In mice that received cabergoline, females had less REM sleep fragmentation after gentle handling than males, more REM sleep fragmentation after restraint stress than males, and more fragmentation of NREM and total sleep after restraint stress than males. The increase in REM sleep bout number in restrained females after cabergoline injection was surprising, because the ability of cabergoline to reduce PRL was expected to mitigate REM sleep phenotypes, not enhance them. Cabergoline-injected female controls had fewer REM sleep bouts than vehicle-injected handled female controls. Therefore, the REM sleep reduction in cabergoline-injected controls was caused by fewer bouts of REM sleep.

Taken together, the results of this study suggest that higher levels of baseline PRL in female mice: 1) mask a baseline sex difference in REM sleep amount, and 2) inhibit the ability of restraint stress to increase REM sleep. Reduction of baseline PRL levels by cabergoline in controls reduced REM sleep amounts in females revealing a difference in the REM sleep response to restraint. Consistent with this hypothesis, there was no sex difference in PRL levels (P = .516) between restrained male mice and control females (Figure 4, A and B).

A number of reports have examined the somnogenic effects of PRL on REM sleep. Male PRL knockout mice have reduced REM sleep compared with wild-type or heterozygous littermates, and sc injection of PRL increases REM sleep amount in PRL knockout mice to levels similar to that of wild-type or heterozygous controls (15). Both peripheral and central injections of PRL result in increased REM sleep. Injection of PRL systemically in rats, cats, and rabbits, into the hypothalamus of rats and rabbits, and into the cerebroventricular area of the rat brain increases REM sleep amount (2). Furthermore, ether stress stimulates PRL secretion (16) and enhances REM sleep in wild-type mice but not in PRL knockout mice. Taken together, these studies suggest that PRL is an integral part of the REM sleep regulating mechanism.

PRL regulation of sleep provides clues for potential roles of PRL in females

Two unexpected influences of cabergoline in this study were the increase in NREM sleep amount observed in control males and the increase in REM θ power in restrained females. The effect in NREM males is paradoxical, because another group of researchers reported that dopamine decreases NREM sleep amount in rats (17). The increase was moderate and only significant when assessed over the duration of the dark phase, but still enigmatic. However, the concomitant and slight decrease of NREM δ power in control males suggests that sleep pressure may have been slightly reduced. Therefore, the increase of NREM sleep amount was likely not attributable to an increase in NREM sleep pressure. It is more likely that the reduced PRL levels in cabergoline-injected males were associated with a slight reduction of vigilance. Because PRL is a stress hormone, it is not surprising that hypoprolactinemia has been correlated with anxiolyitc properties and muscle fatigue (18). To date, few studies examining PRL and its relationship to vigilance or anxiety have been conducted in males.

Taken together, the results of this study predict a model of sex differences and sexual dimorphism in sleep in which PRL has a pivotal role. The premise of this model is grounded in the finding that although male mice have more total sleep (less wake) and more NREM sleep than female mice, there are no sex differences in REM sleep amount (9, 19).

This finding is enigmatic, because REM sleep fluctuations are often similar to NREM sleep fluctuations. Although there are several agents (such as PRL) that can increase REM sleep with minimal influences on NREM sleep, increases of NREM sleep are often associated with comparable increases in REM sleep. Therefore, it is notable that males have higher levels of NREM sleep than females but do not have concomitant higher levels of REM sleep. The explanation for this observation may lie in the following model that emerges from the results of the current study.

Baseline PRL levels are higher in females than males. Therefore, although females should have less REM sleep than males (because of the sex difference in NREM sleep amount), higher baseline levels of PRL confer additional amounts of REM sleep. This increases baseline REM sleep amount and masks the endogenous sex difference in REM sleep. This model is supported by the surprising observation in the current study that cabergoline reduced REM sleep amount in female controls but not in male controls. In fact, cabergoline-injected female controls had less REM sleep than male controls, recapitulating the sex difference in NREM sleep and total sleep reported by our group previously (19). In this model, once female PRL levels were reduced by cabergoline, and similar to baseline levels in males, the sex difference in REM sleep amount was revealed and consistent with sex differences in NREM and total sleep. This model is also supported by the observation that cabergoline-injected females exhibited an increase of REM sleep after restraint stress. This finding suggests that once baseline PRL levels in females are reduced to levels observed in males, they exhibit a male-like REM sleep response to restraint stress.

Few human studies have examined the involvement of PRL in the ability of stress to alter sleep, and of these studies, few have included women. This model has important implications in humans for the role of PRL in modulating sleep responses to stress, especially in women, for 2 reasons: 1) PRL may confer protective properties to REM sleep deprivation in women under reproductive stress, and 2) higher baseline REM sleep amounts are associated with more consolidated sleep, which would benefit women with high levels of endogenous PRL during pregnancy and postpartum recovery.