Spontaneous sleep and homeostatic sleep regulation in ghrelin knockout mice

Éva Szentirmai, Levente Kapás, Yuxiang Sun, Roy G. Smith, James M. Krueger

Abstract

Ghrelin is well known for its feeding and growth hormone-releasing actions. It may also be involved in sleep regulation; intracerebroventricular administration and hypothalamic microinjections of ghrelin stimulate wakefulness in rats. Hypothalamic ghrelin, together with neuropeptide Y and orexin form a food intake-regulatory circuit. We hypothesized that this circuit also promotes arousal. To further investigate the role of ghrelin in the regulation of sleep-wakefulness, we characterized spontaneous and homeostatic sleep regulation in ghrelin knockout (KO) and wild-type (WT) mice. Both groups of mice exhibited similar diurnal rhythms with more sleep and less wakefulness during the light period. In ghrelin KO mice, spontaneous wakefulness and rapid-eye-movement sleep (REMS) were slightly elevated, and non-rapid-eye-movement sleep (NREMS) was reduced. KO mice had more fragmented NREMS than WT mice, as indicated by the shorter and greater number of NREMS episodes. Six hours of sleep deprivation induced rebound increases in NREMS and REMS and biphasic changes in electroencephalographic slow-wave activity (EEG SWA) in both genotypes. Ghrelin KO mice recovered from NREMS and REMS loss faster, and the delayed reduction in EEG SWA, occurring after sleep loss-enhanced increases in EEG SWA, was shorter-lasting compared with WT mice. These findings suggest that the basic sleep-wake regulatory mechanisms in ghrelin KO mice are not impaired and they are able to mount adequate rebound sleep in response to a homeostatic challenge. It is possible that redundancy in the arousal systems of the brain or activation of compensatory mechanisms during development allow for normal sleep-wake regulation in ghrelin KO mice.

  • electroencephalogram
  • non-rapid-eye-movement sleep
  • rapid-eye-movement sleep
  • sleep deprivation
  • slow-wave activity

ghrelin is a gut-brain peptide with potent stimulatory effects on food intake and growth hormone release (16). Outside the brain, the main source of ghrelin is the stomach (16). Gastric ghrelin release is inhibited by feeding; in fact, plasma ghrelin levels inversely correlate with food intake (5, 35). In the brain, ghrelin is mainly produced by neurons in the arcuate nucleus (ARC) and by a distinct hypothalamic neuronal population in the internuclear region (4). In the hypothalamus, ghrelin is involved in a functional circuit that also contains orexin and neuropeptide Y (NPY) and is known to play a role in the regulation of food intake. Ghrelin's actions are mediated through the growth hormone secretagogue receptor (GHS-R) (16, 26). In brain regions, such as the ARC, lateral hypothalamus (LH), dorsomedial hypothalamic nucleus, and paraventricular nucleus (PVN), which are implicated in the regulation of feeding and/or sleep-wake activity, the GHS-R mRNA is abundantly expressed (9, 19). Ghrelin may also serve a role in sleep-wake regulation. Hypothalamic ghrelin levels have clear diurnal rhythms that are affected by sleep deprivation (1). Intracerebroventricular administration (27) or local microinjections of ghrelin into the LH, medial preoptic area (MPA), and PVN induce wakefulness and stimulate food intake in rats (28). Similar to ghrelin, central administration of orexin (10) and NPY (29) stimulate not only food intake but also wakefulness in rats. We proposed that activation of the ghrelin-orexin-NPY circuit triggers behavioral manifestations, such as increased feeding and increased time awake, characteristic of the beginning of the activity period during the dark phase of normal rats.

The aim of the present experiments was to study how the lack of ghrelin affects spontaneous and homeostatic sleep regulation in mice. We investigated the spontaneous sleep-wake activity and sleep deprivation-induced sleep responses in ghrelin knockout (KO) and wild-type (WT) mice. We hypothesized that the congenital lack of a neuropeptide, potentially involved in the regulation of wakefulness, could lead to altered sleep patterns. We report that ghrelin KO mice have reduced duration of non-rapid-eye movement sleep (NREMS) and higher amounts of wakefulness and rapid-eye-movement sleep (REMS) accompanied by more fragmented sleep-wake architecture compared with WT mice.

METHODS

Animals and surgery.

Breeding pairs of ghrelin KO and WT mice with a C57BL6J/129SvEv genetic background, backcrossed to C57BL6J for 10 generations, were generated at Baylor College of Medicine (Houston, TX), and further bred at Washington State University. The procedures used to generate the mutant mice were described previously (25). Each mouse used was genotyped by Transnetyx (Cordova, TN), as described previously (25) at 3–4 wk of age from tail-tip samples, and they were 4- or 5-mo-old at the time of the experiments. The body weight of the WT and KO mice was 28.3 ± 0.4 g and 28.5 ± 0.5 g, respectively, with no significant differences between the two groups. Male mice were anesthetized with intraperitoneal injection of ketamine-xylazine mixture (87 and 13 mg/kg, respectively). The animals were implanted with cortical electroencephalographic (EEG) electrodes, placed over the frontal and parietal cortices, and electromyographic (EMG) electrodes in the dorsal neck muscles. The EEG and EMG electrodes were connected to a pedestal, which was fixed to the skull with dental cement. After surgery, mice were individually housed in a temperature-controlled (30°C ± 1) sound-attenuated environmental chamber under 12:12-h dark-light cycle (lights on at 0500) for habituation to the experimental conditions. During this 7- to 10-day recovery period, the animals were handled daily and connected to the recording cables. Water and food were available ad libitum throughout the experiment. Institutional guidelines for the care and use of research animals were followed and approved by the Institutional Animal Care and Use Committees of Washington State University.

Experimental schedule.

On the baseline day, EEG and EMG recordings of ghrelin KO (n = 9) and ghrelin WT (n = 9) mice were collected over a 24-h period beginning from dark onset to obtain baseline values of spontaneous sleep-wake activity. On the next day, mice were sleep-deprived by gentle handling in the last 6 hours of the light period. During the sleep deprivation, mice were aroused by knocking on the cage or by touching them with a brush when behavioral signs of sleep were observed. Sleep recordings were started immediately after the end of the sleep deprivation at the beginning of the dark period and lasted 23 h (recovery day 1). A second 23-h recovery day was recorded starting at the onset of the subsequent dark period (recovery day 2).

Sleep-wake recording and analyses.

Recording cables were attached to commutators, which were connected to amplifiers. The amplified EEG and EMG signals were digitized and recorded by computer. The EEG was filtered below 0.1 Hz and above 40 Hz. EMG activity was used as a measure of the vigilance state determination and was not further analyzed. The vigilance states (wakefulness, NREMS, and REMS) were visually determined off-line in 10-s epochs by using the conventional criteria, as described previously (1). The amounts of time spent in vigilance states were expressed in 12-h time blocks for the dark and 11-h time blocks for the light period. EEG power values during NREMS and REMS were integrated from artifact-free epochs by fast-Fourier transformation for consecutive 10-s epochs in the delta (0.5–4 Hz), theta (4.5–8 Hz), alpha (8.5–12 Hz), and beta (12.5–16 Hz) frequency bands. EEG power values in each frequency band were summed across the 23-h recording period and expressed as a percentage of total power for each vigilance state and for each mouse. In addition, power density values in the delta range during NREMS [also known as electroencephalographic slow-wave activity (EEG SWA) and used to characterize NREMS intensity] on the baseline day, were averaged across the entire 24-h recording period for each mouse to obtain a reference value. EEG SWA values for each experimental day are expressed as a percentage of this reference value in 2-h time blocks. The total number of wakefulness, NREMS, and REMS episodes and the average length of vigilance state episodes were determined for each 12-h dark and 11-h light period using the criterion that an episode should last at least 30 s. To analyze the kinetics of the rebound sleep on recovery day 1, we calculated the cumulative NREMS and REMS changes by subtracting the hourly NREMS and REMS amounts of the baseline day from the corresponding values of the recovery day 1. The NREMS and REMS cumulative changes were calculated for each mouse for each hour of the recovery day 1, and then the values were averaged within the groups.

Statistics.

The times spent in vigilance states and also the average duration of those vigilance state episodes of each individual mice were averaged within groups across the 12-h dark and, separately, across the 11-h light period. The numbers of vigilance state episodes were summed in the same time blocks. These calculations were performed separately for all three experimental days. The dark- and light-phase values were then analyzed across the three experimental days for both KO and WT mice by using three-way ANOVA. The factors for ANOVA were the genotype, the experimental day (i.e., baseline day, recovery day 1 and recovery day 2), and the phase of the day (i.e., light, dark) effects. The hourly values of the cumulative changes in NREMS and REMS on recovery day 1 were compared between genotypes by using two-way ANOVA (factors: genotype and time effects). The frequency distribution of EEG power density values on the baseline was analyzed separately for NREMS and REMS and compared between genotypes by two-way ANOVA (factors: genotype and frequency band effects). The effect of sleep deprivation on EEG SWA within genotypes was analyzed by ANOVA. Those hours during which a mouse did not have at least 5 min NREMS were not included in the SWA analysis, resulting in missing data points. Therefore, instead of using ANOVA for repeated measures, we used nonrepeated two-way ANOVA (factors: experimental day and the time effects). When ANOVAs indicated significant effects, post hoc comparisons were performed using the Student-Newman-Keuls (SNK) test or Student's t-test. A significance level of P < 0.05 was accepted.

RESULTS

NREMS.

Duration of NREMS showed a diurnal rhythm in both KO and WT mice [ANOVA, phase effect, F(1,96) = 88.0, P < 0.001; phase × genotype interaction F(1,96) = 0.12, not significant (ns)], with more time spent in NREMS during the light phase on the baseline day and recovery day 2 (Fig. 1). Overall, ghrelin KO mice had less NREMS than WT mice, as indicated by the significant genotype effect of ANOVA [F(1,96) = 11.6, P < 0.001]. The lack of significant genotype × experimental day interaction [F(2,96) = 0.1, ns] and genotype × phase interaction indicate that KO mice had less NREMS throughout the entire three experimental days, both during the dark and light period. Sleep deprivation induced a significant change in the amount of NREMS [ANOVA, experimental day effect, F(2,96) = 16.1, P < 0.001]. The effect of sleep deprivation did not differ between genotypes [ANOVA, genotype × experimental day interaction, F(2,96) = 0.1, ns]. The effects of sleep deprivation were confined to the dark phase of recovery day 1; thereafter, NREMS returned to normal [ANOVA, experimental day × phase interaction, F(2,96) = 25.6, P < 0.001; post hoc SNK test: baseline day dark vs. recovery day 1 dark, P < 0.05, recovery day 1 dark vs. recovery day 2 dark, P < 0.05, baseline day dark vs. recovery day 2 dark, ns]. During sleep deprivation, both KO and WT mice lost about the same amount of NREMS (Fig. 4). There was a significant difference in the dynamics of the NREMS rebound during the recovery day 1 between the two genotypes [ANOVA, genotype effect F(1,413) = 62.4, P < 0.001]. Post hoc analyses revealed that ghrelin KO mice recovered from the NREMS deficit significantly faster during the dark period on recovery day 1 compared with ghrelin WT mice. During the light periods, NREMS was not significantly different among the three experimental days (SNK test not significant in any comparison).

Fig. 1.

The time spent in non-rapid-eye movement sleep (NREMS) and the average duration of NREMS episodes values were averaged; the number of episodes was summed across the 12-h dark and 11-h light period, separately for all three experimental days in ghrelin knockout (KO) (solid bars) and wild-type (WT) (open bars) mice. Top: amount of NREMS. Middle: average episode duration. Bottom: number of NREMS episodes. D, dark period; L, light period. Values are presented as means ± SE. Asterisks denote significant differences between ghrelin KO and WT mice (P < 0.05, Student's t-test).

Fig. 2.

The amounts of rapid-eye movement sleep (REMS), the average duration and the total number of REMS episodes in ghrelin KO (solid bars) and WT (open bars) mice. See Fig. 1 for details.

Fig. 3.

The amounts of wakefulness, the average duration, and the total number of wakefulness episodes in ghrelin KO (solid bars) and WT (open bars) mice. See Fig. 1 for details.

Fig. 4.

Cumulative changes in the amounts of NREMS (top) and REMS (bottom) during the 6-h sleep deprivation and the subsequent 23-h recovery period in ghrelin KO (•) and WT (○) mice. Sleep changes are expressed as differences from baseline values on the baseline day. The first six time points represent the gradually accumulated sleep deficits during sleep deprivation. The next 23 time points show the cumulative changes in the amounts of NREMS and REMS during recovery day 1. Gray background represents the dark period. Values are expressed as means ± SE. Asterisks denote significant differences between ghrelin KO and WT mice (P < 0.05, Student-Neumann-Keuls test).

Reduced NREMS in ghrelin KO mice was due to shorter NREMS episode length [ANOVA, genotype effect, F(1,96) = 14.6, P < 0.001] (Fig. 1). NREMS episodes in KO mice were shorter throughout both phases of the three experimental days as indicated by the nonsignificant interactions of ANOVA [genotype × experimental day interaction, F(2,96) = 0.9, ns; experimental day × phase interaction, F(2,96) = 1.0, ns]. Because neither the experimental day effect [F(2,96) = 1.7, ns] nor the experimental day × genotype interaction [F(2,96) = 0.9, ns] showed significant difference, sleep deprivation-induced increase in NREMS time during the dark phase of the recovery day 1 cannot be due to an increase in NREMS episode duration.

There was an uneven diurnal distribution of the number of NREMS episodes with more NREMS episodes during the light period than in the dark [ANOVA, phase effect, F(1,96) = 46.5, P < 0.001] (Fig. 1). Overall, ghrelin KO mice had significantly more NREMS episodes than WT mice [ANOVA, genotype effect, F(1,96) = 10.2, P < 0.001]. This increased NREMS episode number, however, could not compensate for the shorter NREMS bout length; therefore, the NREMS time was significantly shorter in KO (as noted above) than in WT mice. The difference in the total number of NREMS episodes between the two groups of mice was independent of the experimental days [ANOVA, genotype × experimental day interaction, F(2, 96) = 1.0, ns] and of the phases of the experimental days [ANOVA, genotype × phase interaction, F(1,96) = 0.6, ns], indicating that ghrelin KO mice had more NREMS episodes during each day of recording during both the dark and light periods. After sleep deprivation, the number of NREMS episodes changed significantly [ANOVA, experimental day effect, F(2,96) = 5.3, P < 0.01]. NREMS episode number increased in both KO and WT mice [ANOVA, experimental day × genotype interaction, F(2,96) = 1.0, ns] but changed differently in the dark and light phases [ANOVA, experimental day × phase interaction, F(2,96) = 4.7, P < 0.01]. Post hoc analyses revealed that the increase in NREMS episode number was significantly higher during the dark phase of the recovery day 1 compared with the dark period of the baseline (SNK test, P < 0.05) and recovery day 2 (SNK test, P < 0.05). The number of NREMS episodes during the light periods of each experimental day was not significantly different from each other (SNK test, not significant in any comparison).

REMS.

The amount of REMS during the baseline day showed a diurnal distribution similar to NREMS in both KO and WT mice [ANOVA, phase effect, F(1,96) = 67.1, P < 0.001; phase × genotype interaction F(1,96) = 0.4, ns], with more time spent in REMS during the light than in the dark phase (Fig. 2). Overall, ghrelin KO mice had slightly but significantly more REMS [ANOVA, genotype effect, F(1,96) = 4.2, P < 0.05] across the three experimental days [ANOVA, genotype × experimental day interaction, F(2,96) = 0.1, ns] during both the dark and light phase, as indicated by the nonsignificant genotype × phase interaction of ANOVA. Sleep deprivation induced significant changes in the amount of REMS [ANOVA, experimental day effect, F(2,96) = 5.1, P < 0.01] in both groups of mice [ANOVA, experimental day × genotype effect, F(2,96) = 0.1, ns]. This effect was dependent on the phase of the experimental day [ANOVA, experimental day × phase interaction, F(2,96) = 12.0, P < 0.001]. Post hoc analyses revealed that the amount of REMS during the dark period of recovery day 1 was higher compared with the dark periods of the previous (SNK test, P < 0.05) and following day (SNK test, P < 0.05). During 6 h of sleep deprivation, both KO and WT mice accumulated about the same amount of REMS loss (Fig. 4). Cumulative changes in the amount of REMS during the recovery day 1 were significantly different between the two genotypes [ANOVA, genotype effect F(1,413) = 15.6, P < 0.001]. Post hoc analyses, however, failed to identify significant difference in any particular time point between KO and WT mice. There was no statistically significant difference in the amount of REMS during the light periods between any experimental days (SNK test, not significant in any comparison).

The higher amount of REMS in ghrelin KO mice was due to a higher number of REMS episodes [ANOVA, genotype effect, F(1,96) = 4.6, P < 0.05] (Fig. 2). The lack of significant genotype × experimental day [F(2,96) = 0.1, ns] and genotype × phase interaction [F(2,96) = 0.2, ns] indicates that ghrelin KO mice had significantly more REMS episodes during the dark and the light phases on all three experimental days. In response to sleep deprivation, the number of REMS episodes increased [ANOVA, experimental day effect, F(2,96) = 8.7, P < 0.001] in both KO and WT mice [experimental day × genotype interaction, F(2,96) = 0.1, ns]. The significant experimental day × phase interaction indicates that the sleep deprivation-induced changes in REMS episodes were phase dependent [ANOVA, F(2,96) = 12.6, P < 0.001]. Post hoc analyses revealed that the number of REMS episodes during the dark period following sleep deprivation was higher than during the dark period of the baseline (SNK test, P < 0.05) and recovery day 2 (SNK test, P < 0.05). The REMS episode numbers during the light periods of the three experimental days did not significantly differ from each other (SNK test, not significant in any comparison).

The average duration of REMS episodes did not show a diurnal variance [ANOVA, phase effect, F(1,96) = 2.2, ns] (Fig. 2). The difference in the amount of REMS between KO and WT mice cannot be attributed to the average duration of REMS because there was no statistically significant difference in the duration of REMS episodes between the two groups of mice [ANOVA, genotype effect, F(1,96) = 0.01, ns]. Sleep deprivation failed to induce any changes in the duration of REMS episodes [ANOVA, experimental day effect, F(2,96) = 3.1, ns].

Wakefulness.

There was a diurnal variation in wakefulness in ghrelin KO and WT mice, with more time in wakefulness during the dark phase [ANOVA, phase effect, F(1,96) = 277.5, P < 0.001; phase × genotype interaction, F(1,96) = 0.1, ns] (Fig. 3). Ghrelin KO mice overall had more wakefulness compared with ghrelin WT mice [ANOVA, genotype effect, F(1,96) = 7.5, P < 0.01]. The lack of significant genotype × experimental day interaction indicates [ANOVA, F(2,96) = 0.1, ns] that KO mice had more wakefulness throughout the three experimental days. The increased time spent awake in KO animals was present during both the dark and light periods [genotype × phase interaction, F(1,96) = 0.1, ns]. Sleep deprivation induced a decrease in the amount of wakefulness in both KO and WT mice during the subsequent day [ANOVA, experimental day effect, F(2,96) = 17.2, P < 0.001; experimental day × genotype interaction, F(2,96) = 0.1, ns]. The significant experimental day × phase interaction [F(2,96) = 28.6, P < 0.001] shows that the sleep deprivation-induced changes in the amount of wakefulness were different between the dark and light phases. Post hoc analyses revealed that the time spent awake during the dark period after sleep deprivation was less than during the dark period on the baseline (SNK test, P < 0.05) and recovery day 2 (SNK test, P < 0.05). The time in wakefulness during the light periods was significantly different among the three experimental days of recording (SNK test, not significant in any comparison).

The increased amount of wakefulness in ghrelin KO mice was due to the overall elevated number of wakefulness episodes in these mice [ANOVA, genotype effect, F(1,96) = 10.6, P < 0.001] (Fig. 3). KO mice had more wakefulness episodes compared with WT mice during both phases of each experimental day [ANOVA, genotype × experimental day interaction, F(2,96) = 1.2, ns; phase × genotype interaction, F(1,96) = 1.1, ns]. The number of wakefulness episodes was different between the dark and light phases [ANOVA, phase effect, F(1,96) = 42.6, P < 0.001] with more episodes during the light phase in both ghrelin KO and WT mice, as indicated by the nonsignificant phase × genotype interaction. In response to sleep deprivation, there was an increase in the number of wakefulness episodes [ANOVA, experimental day effect, F(2,96) = 5.1, P < 0.001]. These increases occurred in both groups of mice [ANOVA, experimental day × genotype interaction, F(2,96) = 1.2, ns]. The significant experimental day × phase interaction [ANOVA, F(2,96) = 4.6, P < 0.01] indicates that sleep deprivation affected the number of wakefulness episodes differentially during the dark and light period. The increase in the number of wakefulness episodes was confined to the dark period of the recovery day 1 (SNK test: baseline day dark vs. recovery day 1 dark, P < 0.05, recovery day 1 dark vs. recovery day 2 dark, P < 0.05, baseline day dark vs. recovery day 2 dark, not significant) by post hoc analyses. The wakefulness episode number during the light periods did not differ among the three experimental days (SNK test, not significant in any comparison).

The average duration of wakefulness episodes was not different between the two genotypes [ANOVA, genotype effect, F(1,96) = 0.6, ns]; therefore they do not contribute to the differences in wakefulness times seen between KO and WT mice (Fig. 3). Wakefulness episode duration did, however, vary across the experimental days in both groups of mice [ANOVA, phase effect, F(1,96) = 155.3, P < 0.001; phase × genotype effect, F(1,96) = 0.6, ns]. In response to sleep deprivation, the average duration of wakefulness episodes dropped significantly in both ghrelin KO and WT mice [ANOVA, experimental day effect, F(2,96) = 14.4, P < 0.001; experimental day × genotype interaction, F(2,96) = 1.1, ns]. The effect of sleep deprivation on the duration of wakefulness episodes was different between the dark and light phase [ANOVA, experimental day × phase interaction, F(2,96) = 19.4, P < 0.001]. Wakefulness episodes were shorter during the dark phase after sleep deprivation, compared with the dark phase of the baseline (SNK test, P < 0.05) and recovery day 2 (SNK test, P < 0.05). There was no significant difference in the average duration of wakefulness episodes during the light periods among the three experimental days (SNK test, not significant in any comparison).

EEG power density values and slow-wave activity.

The frequency distribution of EEG power values during NREMS and REMS did not differ between KO and WT mice on the baseline day, and there was no difference between the two genotypes on recovery day 1 [ANOVA, genotype effect for both NREMS and REMS and for both the baseline day and recovery day 1 F(1,64) = 0.0, ns; Table 1]. EEG-SWA on the baseline day had a diurnal rhythm with higher values during the dark phase and at the beginning of the light phase compared with the remaining part of the light phase in both groups of mice (Fig. 5). This diurnal distribution of SWA is consistent with those shown in other mouse strains (6). Six hours of sleep deprivation induced biphasic changes in EEG-SWA during the subsequent dark period in both ghrelin KO [ANOVA, experimental day effect, F(1,94) = 6.6, P < 0.05] and ghrelin WT [ANOVA, experimental day effect, F(1,84) = 5.2, P < 0.05] mice. Post hoc analyses identified a significant increase in EEG-SWA during the first 3-h time block after sleep deprivation in both sets of mice (SNK test, P < 0.05). Following the initial increase, there was a decrease in EEG-SWA values during NREMS (SNK test, P < 0.05). In ghrelin WT mice, EEG SWA during the light phase of the recovery day 1 [ANOVA, experimental day effect, F(1,84) = 5.3, P < 0.05] and during the dark phase of the recovery day 2 [ANOVA, experimental day effect, F(1,80) = 5.8, P < 0.05] was significantly different from the values of the baseline day. Sleep deprivation induced similar EEG-SWA changes in KO mice, but the effect was significantly shorter-lasting than in WT mice. The decrease in EEG-SWA was confined only to the first dark period after sleep deprivation, as indicated by the lack of significant experimental day effects of ANOVA for the rest of the recording period [for the light phase of recovery day 1, F(1,96) = 2.2, ns; for the dark phase of recovery day 2, F(1,93) = 2.5, ns; for the light phase of recovery day 2 F(1,96) = 0.3, ns].

Fig. 5.

EEG slow-wave activity (SWA) values on the baseline (○) and recovery days (•) in ghrelin WT (top) and KO (bottom) mice. SWA on the baseline day is double-plotted on each graph. Horizontal lines indicate significant treatment effect by ANOVA. Gray background represents the dark period. Values are presented as means ± SE. *Significant differences between baseline day and recovery days (P < 0.05, Student-Neumann-Keuls test).

View this table:
Table 1.

Electroencephalographic power distribution during non-rapid-eye movement sleep and rapid-eye movement sleep on the baseline day

DISCUSSION

The main finding of the present study is that ghrelin KO mice have a similar diurnal sleep-wake pattern as WT animals, but they have slightly more wakefulness and REMS and less NREMS. NREMS appears to be more fragmented in KO mice because they have a higher number of NREMS episodes, and the individual episodes are shorter than those of WT mice. These findings are consistent with the hypothesis that the lack of a peptide involved in wakefulness regulation affects sleep architecture. The direction of sleep changes, however, was different from the expected (discussed below). Sleep deprivation induces similar changes in the amount of rebound sleep and in NREMS intensity in both genotypes, suggesting that KO mice are also capable of mounting adequate homeostatic sleep responses.

Ghrelin has long been implicated in the regulation of feeding. Both peripheral and central administration of ghrelin strongly stimulate food intake in rodents (34, 38). Previous experiments in our laboratory revealed that ghrelin may also play a role in the regulation of sleep-wake activity. For example, intracerebroventricular administration of 1–5 μg ghrelin at the beginning of light onset induces a robust increase in wakefulness, accompanied by a decrease in NREMS and REMS in rats (27). Ghrelin also promotes wakefulness in rats that are not allowed to eat, suggesting that the increased wakefulness seen in these rats is not due to their increased eating activity. Local microinjections of 0.2–1 μg ghrelin into various hypothalamic nuclei also promote wakefulness in rats (28). Time awake increases and sleep decreases for 2 h after LH microinjection of ghrelin at light onset. Similar, but shorter-lasting effects are observed after ghrelin microinjection into the MPA and PVN in rats.

There is little information available concerning ghrelin's role in the regulation of sleep-wake activity in mice. Systemic injection of 400 μg/kg ghrelin at dark onset induces NREMS in control mice but has no effect in the mutant lit/lit mice (22). Our recent results indicate that intracerebroventricular administration of 0.2–1 μg ghrelin at light onset promotes wakefulness in mice (Szentirmai E, unpublished observation). Our findings that central injection of ghrelin induces wakefulness in two species and that hypothalamic ghrelin levels are elevated at dark onset (1) when rats are the most active suggest that hypothalamic ghrelin may have a role in promoting arousal. One would thus expect that the deletion of a peptide involved in promoting arousal would increase sleep. In the present study, however, the lack of ghrelin led to opposite changes; the amount of wakefulness increased and NREMS slightly but significantly decreased. There are several possible explanations for this inconsistency.

First, there are at least two distinct ghrelin pools in the body; ghrelin is secreted by the stomach (16) and also produced by neurons in the hypothalamus (4). Gastric ghrelin enters the circulation and acts as a gastrointestinal hormone, while hypothalamic ghrelin functions as a neurotransmitter/neuromodulator. Peripheral and central ghrelin may have different effects on sleep-wake activity. It is possible, that ghrelin, produced outside the brain, may have a sleep-promoting effect, as indicated by the increase in NREMS after intraperitoneal injection of ghrelin in mice (22) and also by some human studies (see below). Hypothalamic ghrelin, on the other hand, may be wakefulness-promoting as suggested by our previous findings (see previous paragraph).

Second, in addition to ghrelin, obestatin is also derived from the ghrelin gene. The effect of obestatin on food intake and sleep are the opposite of those of ghrelin; central administration of obestatin inhibits feeding (39) and induces sleep in rats (30). It is possible, that the slightly increased wakefulness and decreased NREMS in ghrelin KO mice reflects higher obestatin tone in WT mice.

Third, the sleep-modulating effect of ghrelin or ghrelin agonists may vary among species. While ghrelin induces wakefulness in rats consistently in several studies (27, 28, 33), in humans, both sleep-promoting and sleep-suppressing effects were reported. For example, repeated intravenous bolus injections of ghrelin increased time spent in slow-wave sleep (SWS) and decreased REMS in young healthy male subjects (36). Repeated intravenous injections of growth hormone-releasing peptide (GHRP)-6 induced a modest increase in stage 2 NREMS (8). Intravenous bolus injection of GHRP-2, failed to induce any change in sleep parameters, except a strong tendency toward increased amount of wakefulness in the first hour after injection (21). Hexarelin, a potent GHS-R agonist in terms of GH-secretion, has SWS- and EEG-SWA-suppressing activity in humans (7). In mice, systemic injection of ghrelin at dark onset was somnogenic (22). The interpretation of the possible species differences in the effect of ghrelin is complicated by the fact that the published effects of ghrelin in species other than rat, were after systemic administration of the peptide. Therefore, these apparent differences among species do not necessarily reflect real species differences; rather, they may be explained by the different administration protocols. In line with this notion, we found increased wakefulness after intracerebroventricular injection of ghrelin in mice.

The fourth possible explanation may lie in the redundancy of the arousal-promoting systems in the brain. Several mechanisms, e.g., the serotonergic, noradrenergic, histaminergic, orexinergic and basal forebrain cholinergic projections, are implicated in arousal. The acute stimulation of these wakefulness-promoting systems usually results in robust changes in the amounts of time spent awake. Chronic blockade or genetic deletion of an individual arousal system, however, does not cause major changes in the circadian rhythm or in the daily amount of spontaneous wakefulness or sleep. For example, chronic immunolesion of basal forebrain cholinergic neurons by 192 IgG-saporin does not affect the daily amount of wakefulness or the homeostatic response to sleep deprivation in rats (15). Several mice strains with the genetic deletion of an arousal-promoting mechanism, e.g., those lacking orexin (3), 5HT1B receptors (2), histamine H1-receptors (12, 14), norepinephrine (13), or histamine (23), exhibit near-normal duration of sleep. Interestingly, the normal sleep time is often accompanied by fragmented sleep architecture in some of these and other KO mice. For example, orexin (20), prion (32), leptin (17), and histidine decarboxylase enzyme (23) KO mice all exhibit fragmented sleep. These findings suggest that the overall, integrated function of arousal mechanisms to maintain normal amounts of wakefulness/sleep is not necessarily impaired by the chronic, especially genetic, deletion of a single component. This suggests a significant redundancy within the arousal systems and the possibility of developmental compensation.

Similar redundancy was observed in other functions also related to ghrelin. For example, food intake is stimulated by several peptides in the hypothalamus, such as NPY/agouti-related protein (AgRP), orexin and ghrelin. Ghrelin KO mice have normal size, growth rate, food intake, and body composition (25). Similarly to ghrelin, the genetic deletion of NPY (31), AgRP (24), or galanin (11), does not impair feeding and/or body weight in mice. Orexin KO mice tend to be hypophagic but maintain normal body weight (37).

Sleep deprivation induced the same amount of rebound increase in NREMS and REMS in both KO and WT mice. The increase in the amount of NREMS was accompanied by an initially elevated EEG-SWA, which was then followed by reduced EEG-SWA values in both genotypes. The time courses of the recovery NREMS and EEG-SWA changes appeared to be slightly different between the two groups. KO mice recovered from the NREMS and REMS loss more quickly and the delayed reduction in EEG-SWA was significantly shorter-lasting. Altogether, these findings suggest that the basic mechanisms required for mounting homeostatic sleep responses are intact, but the fine-tuning of these mechanisms may be affected in KO mice.

In summary, we have shown that the deletion of the ghrelin gene in mice does not lead to major changes in sleep. Although, they exhibit more fragmented NREMS and respond slightly differently to homeostatic challenge, their basic sleep-wake regulatory mechanisms do not appear to be impaired. Hypothalamic ghrelin, NPY, and orexin form a well-characterized circuit, which is implicated in the regulation of feeding. Previous findings suggest that this hypothalamic circuit may also be involved in the regulation of arousal. For example, light onset administration of NPY into the cerebral ventricle or into the LH induces a robust increase in wakefulness and a decrease in NREMS and REMS in rats (29). Central injection of orexin also promotes wakefulness in rodents (10), and orexin's critical role in maintaining arousal is well known (3, 18). We hypothesized that the activation of the hypothalamic ghrelin-NPY-orexin circuit has two main parallel outputs: one is the stimulation of eating and the other is increased arousal.

GRANTS

This work was supported by a grant from the National Institutes of Health (NS27250) to Dr. James M. Krueger and by RO1AG18895 and RO1AG1923 to Dr. Roy G. Smith.

Acknowledgments

We thank Richard A. Brown for breeding the mice.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract