Hypothalamo-Pituitary-Somatotrophic (HPS) System
GH stimulates protein anabolism and tissue growth. Its synthesis and secretion are stimulated by GHRH and ghrelin and are inhibited by somatostatin. All components of the HPS system participate in sleep regulation.
GHRH is an important endogenous sleep-promoting factor. In mice the GH receptor gene is found in the region of chromosome 13 which is linked to SWA (Franken et al., 2001). The expression of hypothalamic GHRH mRNA depends on a circadian rhythm. Its peak is found in rats at the onset of the light period. At this time sleep propensity reaches its maximum in these night active animals (Bredow et al., 1996). Hypothalamic GHRH levels are low in the morning, increase in the afternoon and decrease at night (Gardi et al., 1999). Calcium levels in GABAergic neurons cultured from rat fetal hypothalamus increase when they are perfused with GHRH (De et al., 2002). It appears that many hypothalamic GHRH responsive neurons are GABAergic.
After icv administration of GHRH, SWS increases in rats and rabbits (Ehlers et al., 1986, Obal et al., 1988). This effect is also found after GHRH injection into the medial preoptic area of rats (Zhang et al., 1999) and after its iv administration to rats (Obal et al., 1996). Also in young normal male subjects repetitive hourly iv injections of GHRH between 2200 and 0100 h prompt increases of SWS and GH and blunting of cortisol levels (Steiger et al., 1992). The sleep-promoting effect of GHRH in male subjects was confirmed after iv (Kerkhofs et al., 1993, Marshall et al., 1999) and intranasal (Perras et al., 1999a) administration of GHRH. The effects of GHRH on human sleep were investigated in three conditions with change of the GHRH/CRH ratio in favour of CRH in
(i) the early morning hours in young normal male subjects,
(ii) in healthy elderly men and women and
(iii) in patients with depression.
These studies showed the following results:
(i) After repetitive iv GHRH from 0400 to 0700 h no major changes of sleep EEG were found (Schier et al., 1997).
(ii) The sleep-promoting effect of iv GHRH was much weaker in elderly women and men (Guldner et al., 1997) than in young male subjects (Steiger et al., 1992).
(iii) A sexual dimorphism in the effects of iv GHRH on sleep EEG and HPA hormones was found in drugfree patients with depression of both sexes of a wide age range and in matched normal controls. Cortisol and ACTH concentrations were blunted in male patients and controls whereas in female patients and controls were elevated. Similarly wakefulness decreases and nonREM sleep increases in male patients and controls, whereas opposite sleep-impairing effects occur in women. These data point to a reciprocal antagonism of GHRH and CRH in male subjects, whereas the synergism is suggested in female subjects (Antonijevic et al., 2000b, 2000c).
In the rat after GHRH receptor antagonists (Obal et al., 1991) and antibodies to GHRH (Obal et al., 1992), nonREM sleep decreases. In the so- called supermice, the giant transgenic mice, GH is enhanced. In these mice nonREM sleep and REM sleep are elevated compared to normal mice (Hajdu et al., 2002). Dwarf rats showed deficits in the central GHRHergic transmission and reduced hypothalamic GHRH. In these animals the amount of nonREM sleep is reduced compared to control animals (Obal et al., 2001). Also in dwarf homozygous (lit/lit) mice possessing non-functional GHRH receptors, nonREM sleep and REM sleep are reduced (Obal & Krueger, 2004). These data show that the amount of nonREM sleep is high, when the amount of GHRH is high and vice versa GH deficiency and decreases in nonREM sleep are associated.
In spontaneous dwarf rats, GH deficiency is found due to mutation in the GH gene. In these animals plasma GH levels were almost undetectable. The expression of hypothalamic GHRH mRNA was increased, whereas GHRH receptor and somatostatin mRNAs were decreased. During the light period REM sleep was reduced, and nonREM sleep was enhanced compared to control rats. EEG delta and theta power decreased during nonREM sleep. After icv administration of GHRH, nonREM sleep increased in spontaneous dwarf rats and controls as well (Peterfi et al., 2006).
Unilateral administration of low doses of GHRH to the surface of the rat somatosensory cortex ipsilaterally decreased SWA, whereas higher dosages enhanced SWA. This effects of GHRH on EEG power occurred during nonREM sleep but not during REM sleep. Further the cortical forms of GHRH and its receptor were identical to those found in the hypothalamus and the pituitary, respectively. Cortical GHRH receptor mRNA protein levels did not vary across the light/dark cycle, whereas cortical GHRH mRNA increased with sleep deprivation. These data suggest that cortical GHRH and its receptor play a role in the regulation of localized SWA which is state dependent, as well as their hypothalamic role in nonREM-sleep regulation (Szentirmai et al., 2007c).
Sleep deprivation is the most potent stimulus for sleep homeostasis (Borbely et al., 1981). It is thought that GHRH mediates this stimulatory effect. Sleep promotion after sleep deprivation is inhibited by GHRH antibodies (Obal et al., 1992) and microinjections of a GHRH antagonist into the preoptic area of the rat (Zhang et al., 1999). Sleep deprivation causes a depletion of hypothalamic GHRH and decreased hypothalamic GHRH contents (Gardi et al., 1999). Furthermore hypothalamic GHRH mRNA increases and hypothalamic somatostatin decreases after sleep deprivation in rats (Toppila et al., 1997, Zhang et al., 1998). In humans during the recovery night following sleep deprivation, the nonREM sleep promoting effect of sleep deprivation was augmented by repetitive iv GHRH injections. Interestingly this effect was shared by CRH injections (Schussler et al., 2006b).
NonREM sleep is decreased by negative feedback inhibition of GHRH after administration of GH in humans (Mendelson et al., 1980) and animals (Obal & Krueger, 2004) or higher icv dosages of insulin-like growth factor-1 (IGS-1) (Obal et al., 1999). Since GH antagonism impairs sleep (Obal et al., 1997) also, GH appears to promote sleep. However chronic GH substitution in patients with acquired GH deficiency did not result in sleep-EEG changes (Schneider et al., 2005).
Systemic and icv administration of the somatostatin analogue octreotide reduced nonREM sleep and GH in rats (Beranek et al., 1999). Similarly in young normal male control subjects, intermittent wakefulness increases and SWS decreases after subcutaneous octreotide administration (Ziegenbein et al., 2004). Octreotide is long acting, and its potency is superior to exogenous somatostatin. This explains why sleep EEG remained unchanged after somatostatin administration in young subjects (Kupfer et al., 1992, Steiger et al., 1992). However the same dose of somatostatin which is ineffective in young men impairs sleep in normal elderly women and men (Frieboes et al., 1997). This difference is probably due to a decline of endogenous GHRH during ageing. In rats and cats, somatostatin inhibits GABAergic transmission in the sensory thalamus via presynaptic receptors (Leresche et al., 2000). This mechanism may contribute to the decrease of nonREM sleep after somatostatin. Taken together these data suggest a reciprocal interaction of GHRH on somatostatin in sleep regulation similarly to their role in the regulation of GH secretion.
Similar to GHRH repetitive iv, ghrelin administration increases SWS, SWA and GH in young normal male controls (Weikel et al., 2003). In contrast to the decrease of cortisol after GHRH in young men (Steiger et al., 1992), ACTH and cortisol levels are elevated after ghrelin (Weikel et al., 2003). During the recovery night after sleep deprivation, ghrelin secretion in normal subjects increased earlier than during the baseline night (Schussler et al., 2006a). This observation supports the hypothesis that ghrelin is a sleep-promoting substance. Similar to findings after GHRH, a sexual dimorphism was found as sleep remained unchanged after ghrelin in young normal women (Kluge et al., 2007a). Another similarity to GHRH is the lack of sleep-EEG changes after ghrelin administration during the early morning hours (Kluge et al., 2007b). In contrast to GHRH, ghrelin binds to the growth hormone secretagogue GHS receptor. Also synthetic GHSs modulate sleep. After iv administration of GHRP-6 sleep stage 2 increases (Frieboes et al., 1995), and after oral administration of MK-677 for one week a distinct sleep-promoting effect was found in young men whereas there are only weak effects in elderly controls (Copinschi et al., 1997). After iv hexarelin SWS and SWA decreased probably due to a change of the GHRH/CRH ratio in favour of CRH (Frieboes et al., 2004). Similar to the sleep-promoting effect of ghrelin in young normal men, nonREM sleep increases after systemic ghrelin in mice (Obal et al., 2003). An intact GHRH receptor is a prerequisite for this effect. In animals with nonfunctional GHRH receptors, sleep remains unchanged. A sleep-promoting effect of GHRH is furthermore supported by the observation that wakefulness increases and nonREM sleep decreases in ghrelin knockout mice in comparison to the wildtype (Szentirmai et al., 2007b). However after icv (Szentirmai et al., 2006) and intrahypothalamic (Szentirmai et al., 2007a) administration of ghrelin in rats wakefulness increased during the first two hours after injection. The increase of wakefulness during this interval may be related to increased feeding.
Gonadal Hormones
The menstrual cycle, pregnancy and the menopause reflect distinct changes in endocrine activity in women having some impact on sleep regulation. Only a few studies address this issue so far. Most previous studies of sleep regulation were performed in men or in male animals. One of the reasons why females are not included in these studies is the variability of sleep patterns influenced by hormonal changes due to the ovulational cycle (Kimura, 2005).
One consistent finding is a higher rate of insomnia found in women. According to a metaanalysis including more than a million subjects, women are at 41 % greater risk for developing insomnia than are men. Zhang and Wing (2006) reported that a greater risk of insomnia in women increases with age. In the elderly the risk almost doubles to a 73 % greater risk for insomnia in older women than in older men. It was discussed that this higher prevalence of insomnia in women probably does not emerge until puberty, suggesting a possible contribution of endocrine changes (Johnson et al., 2006).
After administration of gonadal hormones to adult animals only weak sleep-EEG changes were found (review: Manber & Armitage, 1999). After administration of chronic dosages of estradiol in transsexual men undergoing crossgender therapy, sleep stage 1 increased (Kunzel et al., 2000).
After the menopause sleep-endocrine changes associated with depression are accentuated. This is demonstrated by a comparison of sleep-endocrine data in pre- and postmenopausal female patients with depression and in matched normal controls. Cortisol levels are enhanced in the postmenopausal patients, whereas they are reduced in the postmenopausal controls. SWS decreases and REM density increases in post-, but not in premenopausal female patients. An inverse correlation exists between the decline in SWS and in sleep continuity and follicle-stimulating hormone (FSH) secretion in the patients. It appears likely that the menopause contributes to these sleep-EEG changes (Antonijevic et al., 2003). Estrogen replacement therapy via skin patch improves sleep as wakefulness decreases and REM sleep increases during the first two sleep cycles. The normal decrease of SWS and SWA from the first to the second nonREM period is restored (Antonijevic et al., 2000d). Similarly, estrogen supplement in old female rats is also capable of restoring imbalanced sleep patterns (Kimura & Inoue, 2003). Oral progesterone replacement for two weeks increases REM sleep and decreases intermittent waketime in postmenopausal women (Schussler et al., 2008b).
Other Peptides
After pulsatile iv thyrotropin-releasing hormone (TRH) sleep efficiency decreases (Hemmeter et al., 1998).
The peptide galanin is widely located in the mammalian brain. A cluster of GABAergic and galaninergic neurons was found in the ventrolateral preoptic area, which participates in nonREM sleep promotion (Saper et al., 2001). This fits with the observation that after repetitive iv galanin SWS increases in normal control subjects. Furthermore the duration of REM sleep periods is prolonged (Murck et al., 1997).
In animal models of anxiety, opposite effects of CRH and neuropeptide Y (NPY) were found (review: Steiger & Holsboer, 1997). After icv administration of NPY to rats, EEG spectral activity changes similarly to the effect of benzodiazepines (Ehlers et al., 1997a). The increase of REM latency after CRH is antagonized independently by NPY in rats (Ehlers et al., 1997b). In young normal men after repetitive iv NPY, sleep latency and the duration of the first REM period decreases, whereas stage 2 sleep and sleep period time increases. Furthermore cortisol and ACTH is blunted (Antonijevic et al., 2000a). In patients with depression of both gender with a wide age range and in matched controls, sleep latency is reduced and prolactin levels increases after NPY, whereas HPA hormones and other sleep-EEG variables remain unchanged (Held et al., 2006). It is thought that NPY participates in sleep regulation, particularly as a signal of sleep onset acting as an antagonist of CRH via the GABAA receptor.
Conclusions
Between sleep EEG and hormone secretion exists a bidirectional interaction. Some peptides and steroids participate in sleep regulation. Figure 1 depicts a model of peptidergic sleep regulation. It is thought that at least in male subjects a reciprocal interaction of GHRH and CRH plays a keyrole in sleep regulation. In male subjects GHRH promotes nonREM sleep, in younger subjects GHRH enhances particularly SWS and reduces cortisol secretion. In both gender GH was stimulated by GHRH. In contrast, CRH maintains wakefulness and enhances HPA hormones. In addition, CRH promotes REM sleep. Changes in the CRH:GHRH ratio in favour of CRH contribute to the similar sleep-endocrine changes during an acute episode of depression and during normal ageing. On the other hand, GHRH participates in the promotion of sleep after sleep deprivation. In women however GHRH exerts CRH-like sleep-impairing effects. Similar to the reciprocal role in GH secretion, GHRH and somatostatin influence sleep EEG in an opposite fashion, at least in male subjects. Somatostatin is beside of CRH another sleep-impairing peptide. In addition to GHRH, in males galanin and ghrelin promote nonREM sleep. In women however sleep remains unchanged after ghrelin. Ghrelin may act as an interface between the HPA and HPS systems. Galanin was found in clusters colocalized with GABA in the ventrolateral preoptic nucleus. Many GHRH-responsive neurons in the hypothalamus are GABAergic. Galanin, ghrelin and GHRH may either act synergistically, or all these peptides may promote nonREM sleep as part of a cascade. GABAergic neurons appear to mediate the effects of these peptides. The GABAA receptor appears to mediate also the effects of NPY, which acts as a major signal for sleep onset. Interestingly in most studies in patients with insomnia, cortisol levels are enhanced like in depressed patients. Since the risk to develop a depressive episode is highly elevated in patients with untreated insomnia, the question arises whether there are similarities in the pathophysiology and in the genetics, at least of certain, yet not determined subtypes of insomnia and depression. In both disorders elevated HPA hormones appear to be related to impaired sleep. The improvement of sleep after CRH-1 receptor antagonism in depressed patients leads to the question, whether also patients with primary insomnia may benefit from this therapy.
Figure 1. Model of peptidergic sleep-endocrine regulation. CRH, corticotropin-releasing hormone; GHRH, growth hormone-releasing hormone; NPY, neuropeptide Y; SRIF, somatostatin. Characteristic hypnograms and patterns of cortisol and GH secretion are shown in a young and in an elderly normal subject and in a depressed patient. It is thought that GHRH is released during the first half of the night, whereas CRH preponderates during the second half of the night. GHRH stimulates GH and SWS around sleep onset, whereas CRH is related to cortisol release and REMS in the morning hours. NPY is a signal for sleep onset. Besides of GHRH, galanin and ghrelin promote sleep, whereas somatostatin impairs sleep. During depression (CRH overactivity) and during normal ageing, similar changes of sleep-endocrine activity are found. Changes in the GHRH/CRH balance in favour of CRH appear to play a key role in these aberrances.
Beside of peptides steroids participate in sleep regulation. There appears to be a synergism of CRH and cortisol in the pathophysiology of sleep-EEG changes in patients with depression. The elevated risk for insomnia in menopausal women and the beneficial effects of estrogen and progesterone replacement therapy suggests a role of these hormones in impaired sleep in menopausal women. The therapeutic outcome after CRH-receptor antagonism in depressed patients and after estrogen and progesterone replacement in menopausal women and the observation that intranasal vasopressin impairs sleep in elderly subjects are promising hints for new therapies of insomnia in the future related to endocrinology. GHRH or GH secretagogues may help to counteract age-related changes of sleep. However it may be too late to start such treatment too late during the lifespan. A threshold age should be found when such an intervention should start.
