The brains of higher animals perform a staggering array of tasks at the same time, posing a fundamental control problem: how does the brain arrange different actions to generate meaningful behaviour? For example, how does it generate a heightened state of alertness when needed, such as during hunger, but reduce wakefulness when it may not be translated into efficient outcomes, for example during darkness? The hypothalamus, an almond-sized area of the brain in humans, has long been recognised as a fundamental orchestrator of diverse behaviours and adaptive responses (Refs Reference Saper, Scammell and Lu1, Reference Morton2). This neuron-dense structure integrates a wide range of peripheral and sensory signals, while interfacing with a multitude of circuits with more specialised functions. Recently, hypothalamic cells containing orexins (also known as hypocretins; HCRTs) have emerged as critical and multifaceted players in hypothalamic behavioural co-ordination. In this article, we review recent data suggesting that orexin neurons are important for a successful partnership between arousal states and changes in ventilation.

Orexins: arousal signals from the lateral hypothalamic area

Orexins are a pair of peptide neurotransmitters (orexin-A and -B) produced from the same precursor (Refs Reference de Lecea3, Reference Sakurai4). In the mammalian brain, orexin-containing neurons (orexin neurons) are thought to be located solely within the hypothalamus. While only a few thousand orexin neurons are found in this restricted location, they project to virtually the entire brain, except the cerebellum (Refs Reference de Lecea3, Reference Peyron5, Reference Sakurai6). The orexin peptides act on specific G-protein-coupled receptors (OX1R/HCRTR1 and OX2R/HCRTR2), resulting primarily in electrical excitation of the receptor-expressing cell (Refs Reference Sakurai6, Reference Burdakov7). The two receptors have different specificities for the orexins (OX2R is nonselective whereas OX1R prefers orexin-A) and different distributions in the central nervous system (CNS) (Ref. Reference Sakurai6), alluding to the so far unexplored possibility of distinct physiological roles for each orexin.

Orexins play crucial roles in the maintenance of wakefulness, energy homeostasis, and reward-seeking behaviour (Refs Reference Mignot, Taheri and Nishino8, Reference Willie9, Reference Harris and Aston-Jones10, Reference de Lecea11). Abnormalities in orexin signalling – produced by deletion of orexin peptides, orexin neurons or OX2Rs – produce narcolepsy (Refs Reference Lin12, Reference Chemelli13, Reference Thannickal14, Reference Peyron15, Reference Nishino16, Reference Hara17), a disease characterised by irresistible attacks of sleep, cataplexy, and hypnagogic hallucinations. Although orexins stimulate feeding, selective orexin cell ablation leads to obesity due to reduced energy expenditure (Ref. Reference Hara17). Furthermore, mice lacking orexins fail to exhibit normal foraging-like activity during food shortage (Ref. Reference Yamanaka18).

Neural pathways involved in the effects of orexins on wakefulness and feeding have been recently reviewed in detail (Ref. Reference Sakurai6). Briefly, orexin neurons send excitatory inputs to all key components of the classical reticular activating system, as well as directly to attention-promoting regions of the cortex. Orexin cell activity, when monitored using microdialysis, FOS expression, or electrophysiology, appears to be maximal during active wakefulness and minimal during slow-wave sleep (Refs Reference Estabrooke19, Reference Kiyashchenko20, Reference Lee, Hassani and Jones21, Reference Mileykovskiy, Kiyashchenko and Siegel22). Interestingly, stimuli potentially associated with danger, such as sudden loud noises, cause rapid firing in orexin cells (Ref. Reference Mileykovskiy, Kiyashchenko and Siegel22), suggesting that the orexin system may enhance arousal in fight-or-flight situations. Overall, evidence points towards orexin as a key stabiliser of the ‘wake’ position in the sleep–wake flip–flop switch (Ref. Reference Saper, Chou and Scammell23).

The regulation of orexin cell firing is a subject of ongoing research. Known stimulators of orexin cell activity include the hunger hormone ghrelin, while inhibitors include the satiety signals leptin and glucose (Ref. Reference Yamanaka18). Transmitters of other ‘arousal’ systems, such as noradrenaline and serotonin, also inhibit orexin cell firing (Refs Reference Li24, Reference Yamanaka25), presumably acting as a negative feedback system and possibly allowing orexin cells to take over the maintenance of wakefulness when the activity of aminergic systems wanes (Ref. Reference Zeitzer26).

The lateral hypothalamic area and breathing: historical links

The involvement of the lateral hypothalamus in both cognitive arousal and breathing was recognised many decades ago. In the 1920s, the Viennese neurologist Constantin von Economo performed post-mortem studies of brain lesions of patients with encephalitis lethargica, and predicted that the lateral hypothalamus is the source of a critical wake-promoting signal, now thought to come from orexin neurons (reviewed in detail in Ref. Reference Saper, Chou and Scammell23). In relation to breathing, a key prediction of a stimulatory role for the lateral hypothalamus was made by Redgate and Gellhorn in the 1950s. These authors used high-frequency currents to produce localised lesions in the lateral hypothalamus of lightly anaesthetised cats, while monitoring the rate and depth of respiration (Ref. Reference Redgate and Gellhorn27). The lesions resulted in an immediate decrease in the rate and/or depth of respiration, and these effects increased with the size and number of lateral hypothalamic lesions (Ref. Reference Redgate and Gellhorn27). Chemical inhibition of lateral hypothalamic activity, caused by injection of barbiturates, also reduced respiratory activity (Ref. Reference Redgate and Gellhorn27). On the basis of these data, Redgate and Gellhorn concluded that ‘impulses from the lateral hypothalamus exert a tonic facilitatory action on the respiratory centre’ (Ref. Reference Redgate and Gellhorn27). This is supported by recent data revealing an anatomical ‘hotspot’ for breathing stimulation very close to the hypothalamic area where orexin neurons are found (Ref. Reference Tanaka and McAllen28).

Anatomical links between orexin neurons and breathing centres

Apart from heightening cognitive arousal, orexins have also been found to stimulate the sympathetic system, increasing heart rate and blood pressure (Ref. Reference Zhang29), possibly generating a defence or stress-type behavioural response. Do orexins also stimulate breathing as part of their orchestration of the ‘fight-or-flight’ response? The cardiovascular effects of orexin were observed when it was applied to the rostral ventrolateral medulla (RVLM) (Refs Reference Samson30, Reference Shirasaka31, Reference Machado32, Reference Chen33), an area that is also critical for respiratory rhythmogenesis (Refs Reference Feldman34, Reference Hilaire, Bou and Monteau35). Immunohistochemical evidence documented orexin-A-immunoreactive nerve fibres in close proximity of RVLM sympathoexcitatory neurons projecting to the spinal cord (Ref. Reference Machado32), and within the pre-Bötzinger region (part of the respiratory rhythm generator), the nucleus tractus solitarius (area containing inspiratory cells responsive to sensory afferents), and the hypoglossal and phrenic nuclei (areas that control swallowing and diaphragm movements) (Refs Reference Volgin, Saghir and Kubin36, Reference Fung37, Reference Krout, Mettenleiter and Loewy38, Reference Young39). These studies were complemented by demonstration of OXR1 expression within the pre-Bötzinger region of the RVLM and on phrenic motor neurons (Ref. Reference Young39). Indirect pathways by which orexins may modulate respiration were also found: for example, there are orexinergic projections to the raphe nuclei in the brainstem that regulate respiratory long-term facilitation (Ref. Reference Terada40). Thus, a number of anatomical pathways through which orexin neurons may modulate respiration are now known. The relative importance of these pathways in the stimulation of breathing remains to be determined. The percentage of orexin neurons directly projecting to the pre-Bötzinger region and the phrenic nucleus appears to be rather small: 0.5% and 2.9%, respectively (Ref. Reference Young39). This may imply either that only a small subset of orexin cells regulates breathing or that the majority of orexin neurons control breathing through indirect projections. The inputs and outputs of orexin neurons that may be relevant to breathing are summarised in Figure 1.

Figure 1.

Inputs and outputs of orexin neurons that may be relevant to respiratory control. (a) Orexin neurons in the lateral hypothalamus send projections both ‘upwards’, to arousal-regulating regions such as the thalamus and the cortex (among other areas – see Ref. Reference Peyron5), and ‘downwards’ to brain stem nuclei directly or indirectly involved in respiratory control: the raphe nuclei, nucleus tractus solitarius (Refs Reference Smith76, Reference Smith, Connolly and Ferguson77), the rostral ventrolateral medulla (Refs Reference Machado32, Reference Young39) and the phrenic and hypoglossal nuclei (Refs Reference Volgin, Saghir and Kubin36, Reference Fung37, Reference Young39). In turn, orexin neurons receive anatomical inputs from the amygdala and bed nucleus of stria terminalis (BNST), the dorsomedial hypothalamic nucleus (DMH), GABAergic neurons in the ventrolateral preoptic area (VLPO), and many other hypothalamic and extrahypothalamic areas (see Refs Reference Sakurai52, Reference Yoshida53). Orexin cells are also able to act as sensors of body energy levels, and acid and CO₂ (see Fig. 2). (b) Saggital mouse brain section indicating the location of input and output nuclei mentioned in part a. Abbreviation: GABA, gamma-aminobutyric acid.

Effects of stimulation of orexin receptors in breathing centres

Central administration of orexin-A in mice increases the tidal volume (the amount of air inhaled or exhaled during normal ventilation) at concentrations previously shown to induce other orexin-stimulated physiological effects (Refs Reference Young39, Reference Zhang, Fukuda and Kuwaki41). This appears to be an orexin-specific effect, localised selectively to actions on the pre-Bötzinger complex and spinal cord (Ref. Reference Young39). Secondary increases in respiratory frequency and blood pressure do occur but these may be due to later indirect mechanisms (Ref. Reference Zhang, Fukuda and Kuwaki41). In addition to the expression of OX1R in breathing centres (Ref. Reference Young39), involvement of OX2R is also likely, since in some experiments orexin-B elicits greater stimulation of ventilation than orexin-A (Ref. Reference Deng42), and OX2R mRNA has been found on hypoglossal motoneurons (Ref. Reference Volgin, Saghir and Kubin36).

Effects of genetic knockout of orexins on breathing

Further evidence linking the orexin system and breathing comes from mice with genetic deletion of the gene encoding prepro-orexin, the peptide precursor for both orexin-A and orexin-B. These animals lack orexins and develop narcolepsy–cataplexy (Ref. Reference Chemelli13). In a series of recent papers, Kuwaki and colleagues identified several clear abnormalities in the respiration of orexin-knockout mice. Increases in breathing caused by increased atmospheric CO₂ levels (hypercapnia) were reduced by ∼50% in the orexin knockouts (Ref. Reference Nakamura43). Chemical stimulation of the lateral hypothalamus resulted in weaker respiratory increases in orexin-knockout mice than in wild-type mice (Ref. Reference Kayaba44). The orexin-knockout mice also displayed a markedly increased frequency of spontaneous sleep apneas (Ref. Reference Nakamura43). Interestingly, basal ventilation remained similar in orexin-knockout and wild-type mice, irrespective of whether they were asleep or awake (Ref. Reference Kuwaki45), suggesting that either the role of orexin in basal respiration is compensated for by other systems in the knockouts, or orexins do not set the basal breathing tone but selectively stimulate breathing during stressful situations. A full description of known respiratory abnormalities in orexin-knockout mice is given in Ref. 45.

Effects of breathing-related signals on orexin neurons

As mentioned above, orexins do not seem to contribute to basal respiration, but become increasingly important for appropriate respiratory stimulation during hypercapnia (Ref. Reference Nakamura43). This raises the question of how orexin neurons ‘know’ when to become involved in respiration. One theoretical possibility is that they can directly sense homeostatic parameters controlled by breathing. A key chemical variable controlled by breathing is the extracellular levels of H⁺ and CO₂, which in the body are inextricably linked by the reaction H⁺ + HCO₃⁻ ↔ CO₂ + H₂O (Ref. Reference Chesler46). Falling pH (or rising CO₂) is a powerful stimulus for both respiration and behavioural arousal (Refs Reference Nattie47, Reference Putnam, Filosa and Ritucci48, Reference Richerson49). Patch-clamp recordings of the membrane potential of orexin neurons in mouse brain slices showed that the firing rate of these cells is exquisitely sensitive to the ambient levels of H⁺ and CO₂ (Fig. 2). Physiological acidosis (e.g. pH = 7) depolarises orexin neurons and increases their electrical activity, while alkalosis (e.g. pH = 7.4) causes hyperpolarisation and an inhibition of orexin cell firing (Fig. 2). These effects appear to be mediated, at least in part, by acid-induced inhibition of background K⁺ currents in the membrane of orexin cells (Ref. Reference Williams50). Although these currents display some functional properties of TASK channels (Ref. Reference Williams50), their molecular identity remains to be determined.

Figure 2.

Activation of orexin neurons by acidification in brain slices. (a) Orexin neurons (green), identified in a mouse brain slice by transgenic green fluorescent protein expression, are shown with a cartoon of a patch-clamp pipette used to measure their membrane potential (full details in Ref. Reference Williams50). (b) CO₂ effects on firing, membrane potential, and membrane resistance of an orexin neuron. (c) pH effects on firing, membrane potential, and membrane resistance of an orexin neuron. (d) Effects of small extracellular pH changes on firing and membrane potential of an orexin neuron. Parts b–d are reprinted from Ref. Reference Williams50 [Williams, R.H. et al. (2007) Control of hypothalamic orexin neurons by acid and CO₂. Proc Natl Acad Sci U S A 104, 10685–10690 (© 2007 by The National Academy of Sciences of the USA)]. Since acid and CO₂ levels in the brain are controlled by breathing, modulation of orexin neurons by these signals may communicate breathing state to many brain areas through the widespread projections of orexin cells.

Because even small changes in orexin cell firing (e.g. 5 Hz) can stimulate wakefulness (Ref. Reference Adamantidis51), it is tempting to speculate that breathing can regulate states of consciousness by controlling the firing rate of orexin neurons via pH. This theory remains to be examined, for example by comparing breathing and pH-induced changes in cognitive arousal in control and orexin-knockout animals. It also remains to be determined whether different subsets of orexin neurons (Ref. Reference Harris and Aston-Jones10) are differentially sensitive to acid and CO₂; indeed, our preliminary experiments suggest that some orexin neurons are considerably more sensitive to changes in pH than others (D.Burdakov, unpublished). While many aspects require further investigation, we would like to propose that orexin neurons are part of the central pH-sensing network that also includes the classical pH-sensing neurons in the brainstem (Refs Reference Nattie47, Reference Putnam, Filosa and Ritucci48, Reference Richerson49). It seems likely that orexin neurons also integrate other signals relevant to breathing; for example, they receive afferents from the amygdala and the bed nucleus of stria terminalis (Refs Reference Sakurai52, Reference Yoshida53), which may be important for regulation of breathing during emotional stimuli.

Clinical implications

Blunted hypercapnic respiratory response in orexin-knockout mice can be partially restored by administration of orexin, and, conversely, the orexin antagonist SB-334867 (selective for OX1R) reduces the hypercapnic response in wild-type mice (Ref. Reference Deng42). These results indicate that respiration can be effectively altered by drugs targeting the orexin system, potentially suggesting new strategies for treating respiratory disorders.

Overview

Orexin neurons are emerging as a neurophysiological link between breathing and arousal states. They are connected to both the central respiratory nuclei and arousal regions, and may serve to match ventilation to changes in states of consciousness. Although demonstrated only in vitro so far, the chemosensing ability of these cells (Fig. 2) may play a role in the generation of appropriate responses to hypercapnic insults. This could potentially explain the weakened hypercapnic chemoreflex in orexin-knockout mice, which can be partly mimicked in wild-type mice by blocking orexin receptors. By sensing breathing-controlled variables such as pH, orexin neurons may communicate the respiratory status to higher brain areas, potentially providing a bidirectional control circuit linking breathing and overall brain state. It remains to be determined how the breathing-related functions of the orexin system coexist with its other attributes such as glucose sensing (Ref. Reference Yamanaka18) and stimulation of appetite (Ref. Reference Sakurai4). It would also be important to address whether imbalances in energy or sleep homeostasis could have a knock-on effect on the ability of orexin cells to regulate respiration.