The hypocretins/orexins: integrators of multiple physiological functions

Effectors of orexinergic neurons: the monoamines

Monoaminergic neurons constitute one side in the flip/flop model of the sleep-wake cycle (Saper et al., 2010); they excite the neocortex but inhibit sleep centres to promote wakefulness. Importantly, these monoaminergic neurons present in the tuberomammillary nucleus (TMN, histaminergic), locus coeruleus (LC, noradrenergic), dorsal raphe nuclei (DRN, 5-hydroxytryptaminergic) and ventral periaqueductal gray matter (vPAG, dopaminergic) receive dense projections of orexinergic neurons (Peyron et al., 1998; Saper et al., 2005) and, consistent with the various roles of orexin (Marcus et al., 2001), the LC mainly expresses OX₁ receptors, the TMN mainly OX₂ receptors and the DRN expresses both OX₁ and OX₂ receptors. Mieda et al. (2011) showed that the effects of i.c.v. orexin-A (Hcrt-1) on wakefulness and NREM sleep are significantly attenuated in both KO mice as compared with wild-type mice, with the effect being substantially larger in OX₂ receptor-KO mice than in OX₁ receptor-KO mice. These results suggest that although the OX₂ receptor-mediated pathway has a pivotal role in the promotion of wakefulness, the OX₁ receptor is also involved in promoting arousal. However, both receptors appear to be similarly involved in REM sleep suppression.

Orexinergic neurons exhibit parallel firing patterns with monoaminergic neurons that exhibit tonic firing during wakefulness especially during active wakefulness, mild firing during slow-wave sleep, and then are silent during REM sleep (Estabrooke et al., 2001; Lee et al., 2005; Mileykovskiy et al., 2005), except during the transition to wakefulness, when intensive firing is observed (Sakurai, 2007). These data are also consistent with the oscillation of extracellular orexin-A levels that peak during the waking state and decrease to about half their maximum levels during sleep (Yoshida et al., 2001; Zeitzer et al., 2003). These observations suggest that the orexin system stabilizes the state of wakefulness by driving the arousal system (Saper et al., 2010).

Indeed, in vitro electrophysiological studies showed that orexin activates the TMN histaminergic (Bayer et al., 2001; Eriksson et al., 2001; Huang et al., 2001), LC noradrenergic (Hagan et al., 1999) and DRN 5-hydroxytryptaminergic (Liu et al., 2002) neurons. Furthermore, in vivo experiments revealed the involvement of the LC and OX₁ receptors in the LC (Bourgin et al., 2000), as well as the histamine H₁ receptor (Huang et al., 2001) and OX₂ receptor signalling in the TMN (Mochizuki et al., 2011), in orexin- induced arousal. Although, in contrast, recent studies have shown that the orexin-mediated sleep-to-wake transition in mice is not dependent on the histaminergic system (Carter et al., 2009a) and mice deficient in both H₁ receptors and OX₁ receptors display normal sleep/wakefulness patterns (Hondo et al., 2010).

Moreover, Lu and Greco (2006) demonstrated that loss of dopaminergic neurons in vPAG results in a 20% reduction in wakefulness accompanied by an increase in NREM, REM sleep. This finding is supported by a recent report (Kaur et al., 2009) that demonstrated activation of the orexin-vPAG circuit suppresses REM sleep but not non-REM sleep. Orexinergic neurons also receive inhibitory innervation from noradrenergic (Li et al., 2002), 5-hydroxytrptaminergic (Yamanaka et al., 2003b; Kumar et al., 2007) and dopaminergic (Yamanaka et al., 2006) inputs, whereas histamine has little, if any, effect (Yamanaka et al., 2003b). However, the effect of noradrenergic innervation on the activity of orexinergic neurons remains controversial, as some reports show excitatory effects in rats and others demonstrate an inhibitory action (Grivel et al., 2005).

Orexin activates the monoaminergic neurons (Carter et al., 2012), which project to the LH, basal forebrain, as well as cerebral cortex to exert its arousal effect. This notion is supported by a recent study, which showed that mice deficient in the histamine synthesizing enzyme histidine decarboxylase (HDC -/-mice) but not OX receptor -/-mice exhibit abnormal cortical electroencephalographic activity (Anaclet et al., 2009). Interestingly, optogenetic activation (photostimulation) of orexin neurons is not affected in HDC KO mice (Carter et al., 2009a), suggesting that histamine is not essential for orexin-mediated awakenings.

Furthermore, the cholinergic neurons in the pedunculopontine tegmental nucleus/laterodorsal tegmental nucleus (PPT/LDT) fire rapidly during wakefulness and REM sleep but slowly during NREM sleep (Saper et al., 2005), suggesting that they help to maintain cortical activation in the states of wakefulness and REM sleep. Injecting orexin-A into the laterodorsal tegmental nucleus (LDT) results in a significant increase in wakefulness but a decrease of the amount rather than the duration of REM sleep (Xi et al., 2001). In addition, orexin-A has been shown to elicit long-lasting excitation in both cholinergic and non-cholinergic neurons of the LDT (Takahashi et al., 2002). Interestingly, the cholinergic synaptic transmission of LDT neurons is augmented in OX receptor-KO mice, which may induce cataplexy (Kalogiannis et al., 2010). In addition, the (pedunculopontine tegmental nucleus) PPT may also be involved in orexin-induced modulation of wakefulness, as orexin-A has been shown to inhibit the activation of PPT cholinergic neurons, which induces REM sleep onset and atonia (Takakusaki et al., 2004) via GABAergic neurons in substantia nigra pars reticulata and the PPT itself (Takakusaki et al., 2005).

In vitro studies have shown that carbachol, a cholinergic agonist, excites orexinergic neurons (Bayer et al., 2005). In addition, i.c.v. administration of orexin-A (Piper et al., 2000) or its local application into the LC (Bourgin et al., 2000), basal forebrain (Espana et al., 2001; Thakkar et al., 2001) or lateral preoptic area (Methippara et al., 2000) increases the waking time at the expense of sleep. In summary, orexin-induced arousal is modulated not only by monoaminergic neurons, but also by cholinergic neurons in the PPT/ LDT and basal forebrain.

Importantly, the orexin system may be modulated by the circadian clock and homeostatic states (Deboer et al., 2004; Carter et al., 2009a; Appelbaum et al., 2010). Even though there is no evidence of a direct synaptic connection between the suprachiasmatic nucleus (SCN) and orexin cells, the circadian clock drives the orexin system through the output circuits of the SCN (Deurveilher and Semba, 2005). Additionally, local modulation of orexinergic neurons induced by orexin release (Li et al., 2002; Yamanaka et al., 2010), melanin-concentrating hormone (MCH) (Rao et al., 2008; Hassani et al., 2009) or neurons expressing an active form of the leptin receptor (Leinninger et al., 2011) may also be important in the circadian stabilization of a proper sleep-wake cycle. Modulation of orexin activity by energy homeostasis will be expanded in the Energy homeostasis section. Intrinsic plasticity mechanisms may regulate the firing probability of orexin cells during the day and at night (Appelbaum et al., 2010). During the wakefulness period, tonic excitation of orexin neurons may be enhanced when certain stressors are present, like emotional stimulation, which involves the limbic input (Tsujino and Sakurai, 2009). Adamantidis et al. (Adamantidis and de Lecea, 2008a,b2008b) suggested a dual mode of action of orexin: phasic activity lasting 1–10 s that is mostly responsible for the state transitions, and a circadially-regulated oscillation that encodes superimposed information about metabolic and circadian state.

Pain

The analgesic properties of orexin peptides have been well-established with i.v., intrathecal or i.c.v. injection approaches in mouse and rat models of thermal (hot-plate, tail-flick, paw-withdrawal), mechanical (tail pressure, partial sciatic nerve ligation), chemical (formalin, carrageenan, capsaicin and abdominal stretch) nociception and (or) hyperalgesia (Bingham et al., 2001; Yamamoto et al., 2002; 2003a,b,). In all of those models, the orexin-induced analgesic effects were suppressed by the OX₁ receptor antagonist SB-334867 but not by naloxone, an opioid receptor inverse agonist, suggesting that regulation of nociception by orexin is independent of the opiate system (Figure 2). In addition to the direct projection of orexin neurons to the spinal cord (van den Pol, 1999), orexin signalling in the posterior hypothalamic area (Bartsch et al., 2004), the pontine reticular nucleus, oral part (Watson et al., 2010), and periaqueductal gray matter have been shown to be important for its antinociceptive effects. Moreover, orexin plays a significant role in the regulation of stress-induced analgesia (SIA), coordinating with the nociceptin/orphanin FQ systems (Xie et al., 2008; Gerashchenko et al., 2011). Consistent with these data, preproorexin KO mice exhibit hyperalgesia and less SIA (Watanabe et al., 2005). The involvement of orexin in pain is also supported by clinical observations, which have shown there is an association between changes in the Ox receptors and headaches (see below) (Rainero et al., 2004) and a recent multicentre case-control study revealed that chronic pain is more common in patients with narcolepsy with cataplexy than in the controls (Dauvilliers et al., 2011).

Olfaction

The participation of orexin in olfactory function is suggested by the presence of orexin neurons and receptors at all levels of the olfactory system, and their ability to modulate the excitability of olfactory sensory and relay neurons (Caillol et al., 2003; Gorojankina et al., 2007). Indeed, i.c.v. injection of orexin-A increases the olfactory sensitivity to isoamyl acetate (Julliard et al., 2007) and food odour (Prud'homme et al., 2009), although it is unclear whether these increases in sensory perception are related to modulation of brain reward function. Importantly, in studies in humans, it has been established that olfactory dysfunction is a feature of narcolepsy with or without cataplexy (Stiasny-Kolster et al., 2007; Buskova et al., 2010), and that this state could be reversed by intranasal orexin-A (Baier et al., 2008).

HPA axis

Administration of orexin-A, i.c.v., increases plasma adrenocorticotropic hormone (ACTH), corticosterone concentrations, as well as CRF mRNA levels and fos expression in the parvocellular cells of PVN (Al-Barazanji et al., 2001). This orexin-induced activation of the HPA axis involves both OX receptors (Samson and Taylor, 2001; Samson et al., 2002; Ferguson and Samson, 2003), and NPY-dependent (Jaszberenyi et al., 2001) and NPY-independent (Moreno et al., 2005) pathways. Interestingly, these responses are attenuated by pregnancy or CRF antagonists (Jaszberenyi et al., 2000). Moreover, ACTH production is reduced in narcoleptic patients (Kok et al., 2002).

However, orexin-A suppresses the CRF-stimulated ACTH secretion (Samson and Taylor, 2001). Moreover, orexin-A directly stimulates corticosterone secretion (Malendowicz et al., 1999; 2001,; Ziolkowska et al., 2005), and stimulates (Kawada et al., 2003), inhibits (Nanmoku et al., 2000), or has no effect (Mazzocchi et al., 2001) on catecholamine release from the adrenal medulla or phaeochromocytoma cells. In addition, orexin also modulates the proliferation of normal adrenocortical cells or adenomatous cells (Spinazzi et al., 2005a,b2005b).

Whether orexin peptides can be produced by extrahypothalamic sources is still not known. A significant body of literature indicates peptide activity and receptor expression in peripheral organs including the adrenal gland and the gonads. However, there is still no consensus about the origin and the quality of orexin immunoreactive signals in plasma (Arihara et al., 2001; Dalal et al., 2001; Adam et al., 2002; Higuchi et al., 2002; Igarashi et al., 2003; Busquets et al., 2004; Baranowska et al., 2005).

Taken together, the existing data suggest that the orexinergic system at least centrally, is an important modulator of CRF responses and, by extension, of the HPA axis, with potential implications in stress, anxiety and panic disorders.

HPG axis

Administration of orexins, i.c.v., increases LH secretion in ovariectomized (OVX) female rats in the presence of oestradiol and progesterone (Brunton and Russell, 2003), consistent with the observation from orexin-deficient narcolepsy patients exhibit diminished LH release (Kok et al., 2004). However, in the absence of oestradiol or progesterone, orexins suppressed LH release in OVX rats (Pu et al., 2000). Potential mechanisms for this effect may involve NPY (Kiyokawa et al., 2011), CRF (Iwasa et al., 2007), β-endorphin (Irahara et al., 2001) and oestrogen (Furuta et al., 2002). In addition to its ovarian steroid-dependent actions, orexin-A has region-dependent effects on LH release; it increases LH release when microinjected into the rostral preoptic area but reduces LH release when microinjected into the medial preoptic area or arcuate/median eminence, an effect which can be inhibited by the selective OX₁ receptor antagonist SB-334867 (Small et al., 2003).

Notably, orexins also exert their endocrine effects at the testicular level, as orexin-A directly stimulates testosterone secretion in rat testis. In contrast, expressions of key Sertoli cell genes and DNA synthesis in specific stages of the seminiferous epithelium are inhibited by orexin-A (Barreiro et al., 2004; 2005,). In addition, i.p. injection of an OX₁ receptor antagonist (SB-334867) and/or a selective OX₂ receptor antagonist (JNJ-10397049) decreased pro-oestrus gonadotropins and ova number, accompanied by haematological (hyperaemic and/or haemorrhagic) reaction of ovaries with more preovulatory follicles and less corpora lutea (Silveyra et al., 2007). Among the peripheral actions of the orexin peptides, those reported in the gonads seem to have the strongest and most reliable responses. No reports of decreased fertility in narcoleptics have been published, but whether orexin-deficiency in narcolepsy also applies to peripheral sources is unclear.

HPT axis

Orexin-A, i.c.v., inhibits the release of hypothalamic thyrotropin-releasing hormone and subsequently reduces the plasma levels of thyrotropin (TSH). However, orexin-A failed to affect TSH release at the adenohypophysis level, and the plasma thyroid hormone levels showed no changes after peripheral or intra-PVN administration of orexin-A (Mitsuma et al., 1999; Samson and Taylor, 2001; Russell et al., 2002). In addition, narcoleptic patients exhibit decreased plasma TSH concentrations (Kok et al., 2005) but normal plasma thyroid levels (Chabas et al., 2007). In contrast to orexin-A, i.c.v. administration of orexin-B increases the plasma TSH level (Jones et al., 2001). It is possible that imbalances in TSH levels are the cause of metabolic alterations and obesity in adolescent narcoleptics.

GH

In early studies, orexin-A was reported not to stimulate GH release directly (Xu et al., 2002) but to enhance GH-releasing hormone (GHRH)-stimulated GH secretion (Xu et al., 2002) in primary cultured ovine somatotrophs, whereas orexin-B directly increased GH secretion. Furthermore, both orexin-A (Seoane et al., 2004) and orexin-B (Barb and Matteri, 2005) failed to affect the GH secretion from pituitary cells. Administration of orexin-A, i.c.v., was found to reduce the plasma GH levels (Hagan et al., 1999) by inhibiting spontaneous GH secretion (Seoane et al., 2004). Moreover, orexin-A stimulates somatostatin release (Russell et al., 2000) and increases the somatostatin mRNA content in the hypothalamic periventricular nucleus in a GH-dependent manner (Lopez et al., 2004) but selectively decreases the GHRH mRNA in PVN. In addition, orexin-A markedly blunted the GH responses to ghrelin but not GHRH (Seoane et al., 2004). Finally, orexin-deficient narcoleptic patients exhibit normal rates of basal and pulsatile GH secretion (Overeem et al., 2003).

Prolactin

Orexins decrease basal plasma prolactin levels (Jones et al., 2001), and reduce the domperidone-induced plasma prolactin concentrations (Russell et al., 2000), probably through a mechanism involving NPY (Hsueh et al., 2002). In addition, orexin-A restores the fasting-induced abolition of prolactin and LH surges, and anti-orexin-A antisera also abolish these surges in normally fed rats (Kohsaka et al., 2001). Interestingly, the study on the pituitary explants showed that the regulation of prolactin secretion by orexin-A is day-length dependent, namely, orexin-A increases prolactin secretion during the long days (May) but decreases it during the short days (December) (Molik et al., 2008). However, a recent clinical study showed that prolactin secretion is no different in narcolepsy patients than in matched controls (Donjacour et al., 2011), consistent with a lack of effect of both orexin-A and orexin-B on prolactin secretion at the pituitary level (Russell et al., 2000; Russell et al., 2001; Samson and Taylor, 2001).

Circulatory system

Orexin-deficient mice exhibit lower arterial blood pressure, heart rate and sympathetic tone (Kayaba et al., 2003). Furthermore, i.c.v., intracisternally or intrathecally applied orexin increases the mean arterial pressure (MAP), heart rate (HR), renal sympathetic nerve activity and plasma catecholamine or vasopressin levels (Samson et al., 1999; 2005,; Shirasaka et al., 1999; 2002,; Matsumura et al., 2001; Hirota et al., 2003), effects that are blocked or attenuated by the OX₁ receptor antagonist SB-334867 (Hirota et al., 2003; Shahid et al., 2011). However, i.v. injections of orexin-A have no effect on sympathetic activity (Matsumura et al., 2001), suggesting that the cardiac effects of orexin are mediated centrally. Consistently, microinjections of orexin-A into the rostral ventrolateral medulla (Huang et al., 2010) or rostral ventromedial medulla (Ciriello and de Oliveira, 2003) elicit cardiovascular excitatory responses through the activation of both OX₁ and OX₂ receptors (Huang et al., 2010).

However, orexin-A signalling in the nucleus ambiguus (NA) (de Oliveira and Ciriello, 2003) and subfornical organ (Smith et al., 2007) has been shown to produce bradycardia responses, which are, respectively, mediated by an elevation of vagal excitation and a reduction of sympathetic tone. In contrast, it has also been found that orexin-A enhances the inhibitory input and attenuates excitatory synapses to vagal neurons in the NA (Dergacheva et al., 2005). Moreover, the cardiac effects of orexin in the NTS are both dose- and site-dependent, as orexin increases MAP and HR at higher doses (>20 pmol) but reduces these variables at a lower dose (5 pmol). Furthermore, microinjections of orexin into the caudal lateral and medial subnuclei of the NTS decrease both the MAP and HR (de Oliveira et al., b2003), whereas pressor and tachycardiac effects were obtained when orexin was injected into the commissural nucleus of NTS (Smith et al., 2002).

Respiratory effects

Several studies with preproorexin-KO and orexinergic neuron-ablated mice have demonstrated the essential role orexin plays in the hypercapnic chemoreflex response, as well as in phrenic and ventilatory long-term facilitation (Deng et al., 2007; Nakamura et al., 2007; Terada et al., 2008). Also, pharmacological experiments have shown that i.c.v., intracisternal or intrathecal administration of orexin has the ability to elevate respiratory frequency, tidal volume and minute ventilation (Shahid et al., 2011). Furthermore, orexin signalling on OX₁ receptors in the retrotrapezoid nucleus contributes to the control of the hypercapnic chemoreflex (Dias et al., 2009). In addition, when applied to the pre-Botzinger region and phrenic nuclei, orexin augments the phrenic nerve discharge and, subsequently, the electromyographic activity of the diaphragm (Liu et al., 2010). Moreover, injection of orexin-B into pontine Kölliker-Fuse nucleus results in an increase in respiratory frequency and facilitation of upper airway patency (Dutschmann et al., 2007). In parallel, it may be worth noting that orexinergic neurons are strongly inhibited by anaesthetics and orexin-KO mice show delayed emergence from isofluorane anaesthesia (Kelz et al., 2008). However, ambient levels of H⁺ and CO₂ can significantly enhance the activation of orexinergic neurons (Williams et al., 2007; Williams and Burdakov, 2008). Unlike in mice, chemoresponsiveness in humans is independent of OX₁ receptors, as the different ventilatory responses to hypoxia observed in humans with narcolepsy-cataplexy have been associated with the HLA-DQB1*0602 allele that segregates with narcolepsy, but not an orexin deficiency (Han, 2012). Therefore, in humans ventilatory responses to hypoxia may be mediated by other factors or immune components independent of orexin.

It has been suggested that orexinergic neurons are involved in sleep apnoea syndrome, as patients show increased plasma levels of orexin-A (Igarashi et al., 2003; Busquets et al., 2004). Hypercapnia and associated reflexes may increase the activity of orexinergic neurons during sleep, facilitating microarousals and a cascade of sympathetic activity that results in elevated blood pressure during the night. Thus, OX receptor antagonists could be used to prevent these peaks of blood pressure in mild sleep apnoea.

Regulation of digestive activity

Early work showed that orexin immunoreactivity is present in intestinal tissue (Kirchgessner and Liu, 1999). Intracisternal or intraventromedial hypothalamus, but not i.p., injections of orexin-A stimulate gastric acid secretion by activating the vagal system through OX₁ receptors (Takahashi et al., 1999; Yamada et al., 2005; Eliassi et al., 2009; Kermani and Eliassi, 2012). Moreover, activation of OX₁ receptors in the dorsal motor nucleus of the vagus results in facilitation of vagal pancreatic efferent nerve activities (Wu et al., 2004), stimulating pancreatic exocrine secretion (Miyasaka et al., 2002). Administration of orexin-A, i.a., increases duodenal secretion in normal fed but not in fasted animals, by an effect that is independent of cholinergic pathways (Flemstrom et al., 2003; Bengtsson et al., 2007). In addition, orexin-A can modify gastrointestinal motility, including gastric emptying, gastric interdigestive motility (Naslund et al., 2002; Ehrstrom et al., 2005a,b2005b; Bulbul et al., 2010), and enteric peristalsis (Satoh et al., 2006), as well as colonic motility (Kirchgessner and Liu, 1999; Nozu et al., 2011). It is noteworthy that orexin exerts region-specific effects on gastric contractility and relaxation both at the central and peripheral levels, by mechanisms involving ACh and NO respectively (Kobashi et al., 2002; Krowicki et al., 2002; Baccari et al., 2009; Baccari, 2010). Furthermore, orexin-A shows gastroprotective effects against stress-induced (Brzozowski et al., 2008), ischaemia-reperfusion-induced (Bulbul et al., 2008) or ethanol-induced (Yamada et al., 2007) gastric damage.

Urinary activity

The presence of orexin-A and its receptors has been shown in human kidneys and urine (Takahashi et al., 2006), as well as in the bovine urethroprostatic complex (Russo et al., 2008). These findings are supported by results from physiological studies, which demonstrated that orexin-A is involved in the pelvic-urethral reflex (Peng et al., 2008) and the micturition reflex (Kobayashi et al., 2009).