Orexins and the cardiovascular events of awakening

ABSTRACT

This brief review aims to provide an updated account of the cardiovascular events of awakening, proposing a testable conceptual framework that links these events with the neural control of sleep and the autonomic nervous system, with focus on the hypothalamic orexin (hypocretin) neurons. Awakening from non-rapid-eye-movement sleep entails coordinated changes in brain and cardiovascular activity: the neural “flip–flop” switch that governs state transitions becomes biased toward the ascending arousal systems, arterial blood pressure and heart rate rise toward waking values, and distal skin temperature falls. Arterial blood pressure and skin temperature are sensed by baroreceptors and thermoreceptors and may positively feedback on the brain wake–sleep switch, thus contributing to sharpen, coordinate, and stabilize awakening. These effects may be enhanced by the hypothalamic orexin neurons, which may modulate the changes in blood pressure, heart rate, and skin temperature upon awakening, while biasing the wake–sleep switch toward wakefulness through direct neural projections. A deeper understanding of the cardiovascular events of awakening and of their links with skin temperature and the wake–sleep neural switch may lead to better treatments options for patients with narcolepsy type 1, who lack the orexin neurons.

KEYWORDS:

• arousal
• arterial blood pressure
• heart rate
• hypocretin
• narcolepsy
• skin temperature
• sleep

Introduction

This brief review aims to provide an updated account of the cardiovascular events of awakening, and propose a testable conceptual framework that links these events with the neural control of sleep and the autonomic nervous system, with focus on the hypothalamic orexin neurons.

Sleep is a heterogeneous behavior, with two main states termed non-rapid-eye-movement (NREM) sleep and rapid-eye-movement sleep. These states can be objectively discriminated in human subjects¹ and experimental models such as mice and rats² based on the recordings of the electrical activity produced by brain neurons (electroencephalogram), skeletal muscles (electromyogram), and eye movements (electrooculogram). Rapid-eye-movement sleep occurs for a minor fraction of total sleep time and shows a prominent neurophysiological and autonomic variability, whose underlying factors remain poorly understood (cf. refs. 3 and 4). As a result, the cardiovascular events of awakening from NREM sleep are better understood than those of awakening from rapid-eye-movement sleep, and are the only addressed in this review.

The first section of this review thus summarizes the available evidence on tonic cardiovascular control during NREM sleep. The time-dependent changes in arterial blood pressure (ABP), heart rate (HR), and skin temperature upon awakening from NREM sleep are then examined on this basis. In particular, evidence is discussed linking these changes to the hypothalamic neurons that release the orexin peptides (also named hypocretins), involved both in sleep and in autonomic control.⁵ Last, this review raises the hypothesis that awakening may be reinforced and sharpened by its accompanying cardiovascular events through neural positive-feedback control circuits modulated by orexin neurons.

Cardiovascular control during NREM sleep

The values of ABP are approximately 10% lower during NREM sleep than during wakefulness,^3,4 which implies lower values of cardiac output and/or lower values of systemic vascular resistance (Fig. 1). Measurements have indeed shown that cardiac output may be lower during NREM sleep than during wakefulness because of lower values of HR^6,7 caused by a higher cardiac parasympathetic tone.⁸ On the other hand, the sympathetic nerve activity (SNA) to the skeletal muscles,^9-11 kidneys,¹² and skin,^10,11,13 which supposedly exerts vasoconstrictor effects, has been found lower during NREM sleep than during wakefulness. This may contribute to explain changes in skin temperature upon awakening, as discussed in the next section. However, direct evidence that vascular resistance is lower during NREM sleep than during wakefulness is scanty and limited to the skeletal muscles and kidneys.^14,15 Conversely, the resistance of the coronary,¹⁶ cerebral,¹⁷ and spinal cord¹⁸ vascular beds is reportedly higher during NREM sleep than during wakefulness, most likely because of flow–metabolism coupling.^16,19,20 These pieces of evidence are still insufficient to determine how much the lower values of ABP during NREM sleep compared with wakefulness result from lower values of systemic vascular resistance versus lower values of cardiac output.

Figure 1. Schematic diagram summarizing the autonomic and hemodynamic changes that may underlie the tonic decrease in arterial blood pressure during non-rapid-eye-movement sleep (NREM) with respect to wakefulness (WAKE). The question mark emphasizes variability in published experimental evidence. CNS, central nervous system.

The central neural pathways responsible for the specific features of autonomic cardiovascular control during NREM sleep are also largely uncharted. A set of testable hypotheses has been recently proposed.⁴ The lower values of SNA to the skeletal muscles and kidneys during NREM sleep compared with wakefulness may result from: (a) inhibition of a “stress pathway” that involves the hypothalamic paraventricular nucleus,²¹ a master autonomic and endocrine controller;²² (b) inhibition of a “command neuron” pathway that involves the pedunculopontine nucleus of the rostral pons,^23,24 part of the mesencephalic locomotor region;²⁵ and (c) potentiation of a “baroreflex pathway”²⁶ that involves the pontine parabrachial nucleus²⁷ and the medullary nucleus of the solitary tract.²⁸ The lower values of skin SNA during NREM sleep compared with wakefulness may result from inhibition of a “thermoregulatory pathway” for heat generation and retention, which includes the hypothalamic median preoptic nucleus, the hypothalamic medial preoptic area, and the rostral medullary raphe.²⁹ These four pathways may be controlled by a “sleep center” in the hypothalamic ventrolateral preoptic nucleus,³⁰ which is a key structure promoting NREM sleep.³¹ All of these structures receive synaptic projections from the group of hypothalamic neurons that release the orexin peptides.^32,33 The loss of orexin neurons underlies narcolepsy with cataplexy,³⁴ a severe neurological disorder presently classified as narcolepsy type 1 (NT1).³⁵

The cardiovascular events of awakening from NREM sleep

Changes in ABP and HR upon awakening from NREM sleep

The most direct approach to the assessment of the cardiovascular events of awakening is a coherent averaging of cardiovascular variables synchronized at the transition from NREM sleep to wakefulness. Unfortunately, data obtained with this direct approach are still quite limited. Fig. 2A shows changes in ABP and HR upon awakening from NREM sleep studied with coherent averaging in middle-aged wild-type and orexin-ataxin3 (ORX-ATX3) transgenic mice.³⁶ ORX-ATX3 mice express ataxin3, a neurotoxin, under control of the orexin gene promoter.³⁷ As a result, these mice undergo a progressive loss of orexin neurons, which becomes virtually complete in young adulthood.^37,38 These mice may therefore be considered a model of NT1 with good construct validity.

Figure 2. Panel A shows changes (Δ) in mean arterial blood pressure (ABP) and heart rate (HR) upon awakening, i.e., on passing from non-rapid-eye-movement sleep (NREM) to wakefulness (WAKE). ΔABP and ΔHR were computed as % changes from a baseline 20-s period of NREM before awakening (left horizontal bar). Statistical analysis was performed on ΔABP and ΔHR in a 20-s period after awakening (horizontal right bar). The gray shading at time 0 marks the time of awakening. Data are mean values ± SEM in `middle-aged (10–11 months) orexin-ataxin3 transgenic mice (ORX-ATX) with postnatal loss of hypocretin/orexin neurons (n = 9; black: ABP; red: HR), a mouse model of narcolepsy type 1, and age-matched wild-type control mice (WT, n = 8; gray: ABP; blue: HR). * and †, P < 0.05, TG versus WT for ΔABP and ΔHR, respectively (t-test). Reproduced from ref. 36, with permission. Panel B shows ΔHR upon awakening from NREM (mean values ± SEM) in NT1 patients and control subjects (CTRL), with n = 12 per group, excluding epochs with apneas, leg movements and arousals transitions. *, P < 0.05 for ΔHR upon awakening, NT1 patients versus controls. Redrawn from ref. 46, with permission.

In the experiment shown in Fig. 2A, regardless of the mouse strain, HR reached its waking values sooner upon awakening than ABP did. As discussed previously, there is evidence that HR is lower during NREM sleep than during wakefulness because of higher cardiac parasympathetic tone.⁸ The HR response to parasympathetic activity is much faster than responses to SNA.³⁹ The different time course of the changes in ABP and HR upon awakening from NREM sleep is thus consistent with cardiac parasympathetic withdrawal upon awakening.

As just discussed, ABP and HR increased upon awakening with different time courses. However, this mismatch was transient and limited approximatively to 30 s after awakening (Fig. 2A). Thereafter, both ABP and HR eventually increased approximatively 15% of their respective values during NREM sleep in wild-type mice.³⁶ This is in line with the observations that ABP is approximately 10% lower during NREM sleep than during wakefulness in different species,^3,4 and that changes in HR during NREM sleep may contribute to determine those of ABP by modifying cardiac output.^6,7 Inspection of Fig. 2A also indicates that ABP and HR increased upon awakening significantly less in ORX-ATX3 mice, which lack the orexin neurons, than in wild-type mice.³⁶ This observation fits with previous evidence³⁸ that ABP was higher during sleep, but not during wakefulness, in either young adult ORX-ATX3 mice lacking orexin neurons³⁷ or young adult orexin knockout (ORX-KO) mice lacking orexin peptides⁴⁰ than in wild-type mice. There is substantial evidence that orexin neurons co-release dynorphin and glutamate, and more contrasting evidence that at least some of these neurons co-express nociceptin/orphanin FQ, galanin, neurotensin, and GABA (cf. ref. 41 for a recent review). The role of these orexin co-transmitters in the cardiovascular control during sleep is unknown. However, the finding of similar ABP derangements in ORX-ATX3 mice and ORX-KO mice with respect to wild-type mice suggests that the lack of orexin release, as opposed to the lack of other transmitters co-released by orexin neurons, is sufficient to alter ABP during sleep.³⁸ These data,³⁸ as well as those reported in Fig. 2A,³⁶ were obtained on mice with a “pure” C57Bl6/J genetic background. Conversely, a later study on young adult ORX-ATX3 mice on a mixed (75% C57Bl/6J and 25% DBA/2J) genetic background indicated that, on passing from wakefulness to NREM sleep, these mice showed a significantly blunted decrease in ABP but not in heart period (the reciprocal of HR) compared with wild-type mice.⁴² This was because ABP was significantly lower in these hybrid ORX-ATX3 mice than in hybrid wild-type mice during wakefulness, but not during NREM sleep.⁴² Other experiments, performed without taking sleep into account on ORX-KO⁴³ and ORX-ATX3⁴⁴ mice with mixed genetic background, reported lower values of ABP during behaviorally-determined wakefulness in these mice compared with wild-type mice. Experiments on ORX-ATX3 rats found lower values of ABP compared with wild-type rats, with similar increases in ABP and HR upon awakening from NREM sleep.⁴⁵ Based on these pieces of evidence, the conclusion that the lack of orexin neurons blunts the increases in ABP and HR upon awakening from NREM sleep is still to be considered as controversial in mice and rats. Nonetheless, this conclusion was shared by a recent study that used coherent averaging of HR upon awakening from NREM sleep in NT1 patients (Fig. 2B), even after exclusion of the epochs of NREM sleep with apneas, leg movements and arousals transitions.⁴⁶ This is relevant because previous work on NT1 patients found that the increase in ABP upon morning awakening was blunted in these patients compared with normal controls.⁴⁷ This difference was attributed, at least in part, to the higher occurrence of periodic leg movements during sleep and the greater sleep fragmentation before awakening in NT1 patients.⁴⁷ Blunted differences in ABP and HR between NREM sleep and wakefulness¹³ and a more common non-dipping pattern of ABP^48,49 have also been reported in NT1 patients compared with control subjects. However, other studies found no significant differences in ABP during wakefulness and sleep⁵⁰ or a decrease in ABP during wakefulness⁵¹ in NT1 patients compared with controls. It should be remarked that NT1 is not caused by the isolated deficiency of orexin signaling, but, rather, by the loss of the orexin neurons.⁵² Thus, the phenotype of NT1 patients may result not only from the loss of orexin signaling, but also from the loss of the other transmitters co-released by the orexin neurons.⁴¹ Taken together, these pieces of evidence on rodent models of orexin deficiency and on NT1 patients highlight a substantial variability among the results of studies, which would appear carefully conducted. This variability may thus point to powerful biological modifiers of the cardiovascular effects of orexin neuron deficiency. Nonetheless, some regularities do emerge: Orexin neuron deficiency causes normal or reduced values of ABP during wakefulness, normal or increased values of ABP during sleep, normal or increased values of HR during wakefulness and sleep, and normal or reduced differences in ABP and HR between wakefulness and sleep. The available evidence is, thus, at least consistent with the view that orexin neuron deficiency may blunt the increases in ABP and HR upon awakening from NREM sleep. The magnitude of this effect likely depends on one or more biological modifiers, which still remain undefined, as do the underlying autonomic, hemodynamic, and central neural mechanisms.

Changes in skin temperature upon awakening from NREM sleep

Changes in skin temperature are a relatively overlooked result of the cardiovascular events of awakening. A careful study on rats (Fig. 3) revealed that at 21°C ambient temperature, the tail skin temperature sharply decreased upon awakening from NREM sleep in the face of increasing core body temperature.⁵³ These data may reflect the rapid return of SNA to waking values, in light of the previously discussed observations that skin SNA is lower during NREM sleep than during wakefulness.^10,11,13 The resultant distal skin vasoconstriction may contribute to increase core body temperature by promoting heat conservation. Fig. 3 also shows that in rats, differences in core body temperature and tail skin temperature upon awakening were much less marked at 29°C than at 21°C ambient temperature.⁵³ This may indicate that at 29°C ambient temperature, the tail skin of rats is already maximally vasodilated during NREM sleep.⁵³ Replication of these findings at different skin sites and in different species, including mice, would improve our understanding of the link between the cardiovascular events of awakening and thermoregulation. The thermoregulatory responses of the distal/hairless (tail) skin and the proximal/hairy (back) skin of rats are controlled by independent neural pathways that originate from the rostral medullary raphe.⁵⁴ These pathways may thus be engaged differently upon awakening. Accordingly, human subjects reportedly show a relative vasodilation of distal (wrist and ankles) skin compared with proximal (infraclavicular and sternal) skin during sleep, so that the temperature difference between distal and proximal skin sites, termed the distal-proximal skin temperature gradient (DPG), becomes less negative and tends to 0 during sleep.⁵⁵ Upon awakening from sleep, conversely, the DPG rapidly becomes negative again, mainly because of a decrease in distal skin temperature.⁵⁶ A study on human subjects claimed that the more negative are the values of DPG during daytime activities, the greater is the nocturnal ABP dipping.⁵⁵ This suggests that changes in skin SNA may play a key causal role in driving the changes in ABP during NREM sleep. However, evidence on experimental models does not appear consistent with this suggestion. The results of the study on rats shown in Fig. 3 indicate that the decrease in tail skin temperature upon awakening from NREM sleep was much smaller and shorter lasting at 29°C than at 21°C ambient temperature.⁵³ Conversely, work on mice indicated that the difference in ABP between NREM sleep and wakefulness was greater at 30°C than at 20°C ambient temperature.⁴² Notably, the cardiovascular effects of changes in ambient temperature are even greater in mice than in rats, likely because mice have a higher ratio between body surface area and volume.⁵⁷ Thus, it remains unclear whether increases in skin SNA may contribute to increase systemic vascular resistance and ABP upon awakening from NREM sleep.

Figure 3. Panel A shows changes (Δ) in peritoneal (core) and distal skin (tail) temperature (T) upon awakening, i.e., on passing from non-rapid-eye-movement sleep (NREM) to wakefulness (WAKE) in wild-type rats at ambient temperature (T_AMB) of 21 °C and 29 °C. The gray shading at time 0 marks the time of awakening. Data are means for seven rats. Error bars and statistics are omitted for clarity. Modified from ref. 53, with permission.

Alterations of skin temperature have been reported in NT1 patients during a multiple sleep latency test, which is designed to detect excessive daytime sleepiness by measuring sleep latency during repeated opportunities for daytime naps.⁵⁸ In particular, the DPG during wakefulness was reportedly less negative in NT1 patients than in control subjects because of higher distal and lower proximal skin temperatures. Once asleep, NT1 patients maintained their elevated distal skin temperature, while their proximal skin temperature increased toward the values in controls.⁵⁸ However, other studies based on 24-h temperature profiles reported that NT1 patients have lower proximal skin temperature but normal distal skin temperature during the daytime⁵⁹ or higher proximal and higher distal skin temperature in the morning⁶⁰ compared with control subjects. Recordings of skin SNA in NT1 patients indicated a normal decrease during NREM sleep compared with wakefulness,¹³ and data on ORX-ATX3 rats have not shown significant differences in the effect of sleep on tail skin temperature compared with wild-type rats.⁴⁵ Contrasting results were also obtained by studies of the 24-h profiles of core body temperature in NT1 patients^59-61 and by studies of sleep-related changes in core body temperature between NREM sleep and wakefulness in NT1 patients, ORX-ATX3 mice, and ORX-KO mice.^61-63 More recent experiments on ORX-ATX3 rats indicated that the integrity of orexin neurons is important for the full expression of cold defense responses, such as brown adipose tissue thermogenesis, behavioral activity, and tail skin vasoconstriction.⁶⁴ Studies of thermoregulation in ORX-ATX3 mice and ORX-KO mice led to the conclusion that orexin neurons regulate body temperature through an orexin co-transmitter.⁶⁵ Although the identity of this co-transmitter is still uncertain, this hypothesis is biologically plausible. Accordingly, as previously mentioned, orexin neurons co-release dynorphin and glutamate, and potentially also nociceptin/orphanin FQ, galanin, neurotensin, and GABA.⁴¹ However, a report on mice showed that the cardiovascular effects of mild cold stress did not differ in any wake-sleep state between ORX-ATX3 mice, who lack the whole orexin neurons, and wild-type mice.⁴² To summarize, orexin neuron deficiency in NT1 patients may alter the skin temperature upon awakening from NREM sleep in the sense of higher distal skin temperature during wakefulness, but this effect has not received direct support by experiments on animal models. Again, this effect seems to depend on one or more undefined biological modifiers, and its autonomic, hemodynamic, and central neural mechanisms remain unclear.

Integration of the neural and cardiovascular events of awakening

Orexin neurons may be part of a positive feedback neural circuit that prevents uncontrolled transitions between wake–sleep states

Transitions between wake–sleep states result from coordinated changes in the activity of multiple neuronal populations. A series of neuronal unit recordings in mice by Takahashi and coworkers^66-69 led to the conclusion that awakening from NREM sleep was heralded by increases in the activity of the wakefulness-specific noradrenergic (A6) neurons of the pontine locus coeruleus.⁶⁶ These increases preceded the activation of wakefulness-active neurons in the basal forebrain and the hypothalamic preoptic area.⁶⁹ Increases in the activity of the hypothalamic orexin neurons lagged behind, but occurred still before the onset of electroencephalographic desynchronization, which marks awakening.⁶⁸

The projections from the locus coeruleus to the sympathetic⁷⁰ and pre-sympathetic⁷¹ neurons are relatively limited. Chemical stimulation of the locus coeruleus reportedly decreases ABP mainly because of skeletal muscle and kidney vasodilation.⁷² The locus coeruleus is not necessary for thermogenesis in response to noradrenaline and cold exposure, but may play a role in the fever responses to prostaglandin E2.⁷³ On the other hand, the increased discharge of locus coeruleus neurons that occurs during wakefulness⁶⁶ appears sufficient to cause awakening, and may be necessary for the normal duration of the wakefulness episodes.⁷⁴ The neurons of the locus coeruleus project to³² and inhibit⁷⁵ neurons of the hypothalamic ventrolateral preoptic nucleus, which, as discussed above, is a key structure for NREM sleep.³¹ In turn, neurons of the ventrolateral preoptic nucleus send inhibitory projections to the locus coeruleus.³⁰ Reciprocal inhibitory projections may also exist between the locus coeruleus and the hypothalamic median preoptic nucleus,^76,77 which is involved in thermoregulatory control,²⁹ and includes sleep-active neurons⁷⁸ that decrease firing soon after awakening from NREM sleep.⁷⁹ Reciprocal inhibitory connections between the ventrolateral preoptic nucleus and ascending arousal systems, including the noradrenergic neurons of the locus coeruleus, the serotonergic neurons of the dorsal raphe, and the glutamatergic neurons of the parabrachial nucleus, have been proposed by Saper to constitute the basic wake–sleep switch^80,81 (Fig. 4A, red arrows). This neural switch is thought to operate as a flip–flop switch, an electronic device designed to cause sharp and complete state transitions.^80,81 This flip–flop switch may be stabilized by the orexin neurons, which send excitatory projections to the locus coeruleus, dorsal raphe, and parabrachial nucleus.³³ In particular, the locus coeruleus receives the most dense arborization of orexin axons in the brainstem.⁸² Orexin projections to the locus coeruleus are key for the effects of orexins on arousal⁸³ and the control of rapid-eye-movement sleep.⁸⁴ The locus coeruleus may be relevant for the pathophysiology of NT1 as a result of the lack of orexin input and of chronic, possibly compensatory, alterations in GABAergic input.⁸⁵ The orexin neurons also project to the ventrolateral preoptic nucleus.³² However, orexin microinjections in the ventrolateral preoptic nucleus increase wakefulness, suggesting that orexin neuron projections exert a net inhibitory effect on neurons of the ventrolateral preoptic nucleus through local circuits.⁸⁶ In turn, the orexin neurons receive inhibitory projections from the hypothalamic ventrolateral preoptic nucleus⁸⁷ and excitatory projections from the parabrachial nucleus,⁸⁸ but minimal or absent input from the dorsal raphe and locus coeruleus.⁸⁷ The orexin neurons may thus be part of a positive feedback neural circuit that stabilizes the wake–sleep flip–flop switch upon awakening from NREM sleep. In particular, the increased activity of orexin neurons upon awakening may further increase arousal by enhancing the activation of ascending arousal systems while inhibiting neurons of the ventrolateral preoptic nucleus. In addition, orexin directly modulates the thalamo-cortical system, which plays a key role in supporting waking and consciousness. In particular, the orexin input switches thalamic neuron firing from the burst mode characteristic of NREM sleep to the tonic mode characteristic of wakefulness and rapid-eye-movement sleep.⁸⁹

Figure 4. Panel A shows a simplified schematic model of the basic wake–sleep switch. DR, dorsal raphe; LC, locus coeruleus; ORX, hypocretin/orexin neurons of the hypothalamus; PB, parabrachial nucleus; VLPO, ventrolateral preoptic nucleus. Red and blue arrows indicate inhibitory and excitatory connections, respectively. Modified from ref. 81, with permission. Panel B shows a schematic diagram of the hypothetical integration between the central neural mechanisms and the cardiovascular events of awakening. EEG, electroencephalogram; NTS, nucleus of the solitary tract. Modified from ref. 4, with permission.

The increase in ABP upon awakening may be part of a positive feedback control that prevents uncontrolled transitions back to sleep

Phasic increases in ABP may promote awakening from NREM sleep.⁹⁰ In particular, 10–20 mmHg increases in ABP obtained with pharmacological^91,92 or mechanical^91,93 means increased the probability of arousal from NREM sleep in experiments on dogs,⁹¹ lambs,⁹³ and human subjects.⁹² At least in lambs, this phenomenon was neither an experimental artifact nor a byproduct of changes in brain circulation, because it was prevented by sectioning baroreceptor afferents.⁹³ The nucleus of the solitary tract, which is the first neural relay of the ascending baroreflex pathway,²⁶ is strongly and reciprocally connected to the parabrachial nucleus.^26,94 As previously mentioned, the parabrachial nucleus is, functionally, part of the ascending arousal system critical for the maintenance of wakefulness.^81,95 In particular, there is evidence that glutamatergic neurons in the medial and lateral parabrachial nucleus play important roles in promoting spontaneous wakefulness and in mediating hypercapnic arousal from NREM sleep, respectively.⁹⁶ The parabrachial nucleus projects to the basal forebrain⁹⁵ and the thalamus,⁹⁷ which critically underlie the waking state, as well as to the orexin neurons.⁸⁸ These considerations suggest that efferent connections from the nucleus of the solitary tract to the parabrachial nucleus may promote awakening in response to increases in ABP.⁹⁰ The activity of parabrachial nucleus neurons is generally lower during NREM sleep than during wakefulness,^27,98 potentially because of inhibition by the hypothalamic ventrolateral preoptic nucleus³⁰ and, possibly, also by local neurons.⁹⁹ Increases in parabrachial nucleus activity may inhibit the baroreflex.^100-102 Thus, the decrease in parabrachial nucleus activity during NREM sleep may contribute to decrease SNA through potentiation of the neural “baroreflex pathway”⁴ previously mentioned. It follows that the increases in ABP upon awakening may be part of a positive feedback control that contributes to stabilize the wake–sleep neural switch (Fig. 4B). As previously discussed, the increase in ABP upon awakening may be blunted in the absence of the orexin neurons. This raises the hypothesis that the orexin neurons may indirectly enhance the positive feedback control triggered by the increase in ABP upon awakening, which would be in synergy with their direct effect on the wake–sleep neural switch. Conversely, weak increases in ABP upon awakening may contribute to destabilize sleep–wake transitions in NT1 patients and in animal models who lack orexin signaling.

The decrease in distal skin temperature upon awakening may be part of another positive feedback control that prevents uncontrolled transitions back to sleep

Distal, non-hairy skin is thought to provide a limited contribution to the feedback for thermoregulation.¹⁰³ However, skin temperature can play a significant role in the regulation of sleep and vigilance.¹⁰⁴ Information on skin temperature is transmitted via a neural pathway that includes the dorsal root ganglia, spinal lamina I neurons, the parabrachial nucleus, and the hypothalamic median preoptic nucleus.²⁹ As previously discussed, the DPG rapidly becomes negative upon awakening because of the decrease in distal skin temperature.⁵⁶ It has been shown that the time course of the decrease in DPG after awakening significantly and positively correlates with that of the decrease in subjective sleepiness.⁵⁶ This suggests that subjective sleep inertia is dissipated as the distal skin progressively cools down relative to the proximal skin after awakening.⁵⁶ Conversely, the DPG at the time when lights are switched off was shown to correlate negatively and significantly with sleep onset latency. This indicates that the warmer the distal skin is, at least relative to the proximal skin, the shorter it takes to fall asleep.¹⁰⁵ Subsequent experiments have supported a causal link between skin temperature and wake–sleep control. In particular, distal skin warming by less than 0.5°C without increases in core body temperature strongly enhanced NREM sleep in elderly subjects with insomnia complaints,¹⁰⁶ while similar decreases in distal skin temperature increased the time that NT1 patients were able to stay awake.¹⁰⁷ These pieces of evidence raise the hypothesis that distal skin cooling caused by increased skin SNA upon awakening may promote wakefulness. This would constitute another positive feedback loop that indirectly stabilizes the sleep–wake neural switch upon awakening. The orexin neurons may indirectly strengthen this positive feedback control based on distal skin cooling upon awakening. This would be in synergy with the potential role of orexin neurons on the other positive feedback control based on the ABP increase upon awakening, as well as with the direct neural effects of orexin neurons on the wake–sleep switch. Conversely, the persistence of relatively high distal skin temperature during wakefulness in NT1 patients, who lack the orexin neurons, may weaken this positive feedback loop, thus contributing to destabilize sleep–wake transitions (Fig. 4B).

Conclusions

The experimental evidence reviewed in the previous sections indicates that awakening from NREM sleep entails coordinated changes in brain and cardiovascular activity. At the brain level, the activity of reciprocally inhibitory circuits, which constitute the basic wake–sleep switch, becomes biased toward the ascending arousal systems. At the cardiovascular level, variables such as ABP, HR, and skin temperature, which also depend on cardiovascular control, reach their waking values with characteristic time courses. Much critical evidence is still missing. However, the available data are at least consistent with the hypothesis that the changes in ABP and skin temperature upon awakening, sensed by baroreceptors and thermoreceptors, may positively feedback on the brain wake–sleep switch, and thereby contribute to sharpen, coordinate, and stabilize awakening. These effects may be enhanced by hypothalamic orexin neurons, which may modulate the changes in ABP and skin temperature upon awakening, and, at the same time, bias the wake–sleep switch through direct neural projections. These hypotheses are testable, albeit with the technical complexity required by the study of cardiovascular control without perturbation of the wake–sleep behavior. In particular, studies on NT1 patients and experimental models with impaired orexin signaling should involve simultaneous coherent averaging of ABP, HR, and skin temperature across transitions from NREM sleep to wakefulness. Independent experimental manipulation of ABP and skin temperature upon awakening with physical or pharmacological approaches is needed to provide critical insights on causal links. Further research is also needed to understand which biologic factors underlie the reported variability in the effects of orexin deficiency on cardiovascular control during wakefulness and sleep. A deeper understanding of the cardiovascular events of awakening and of their links with skin temperature and the wake–sleep neural switch may lead to better treatments options for NT1 patients, who lack the orexin neurons.

Abbreviations

ABP	=	arterial blood pressure
DPG	=	distal–proximal skin temperature gradient
HR	=	heart rate
NREM	=	non-rapid-eye-movement
NT1	=	narcolepsy type 1
ORX-ATX3	=	hypocretin/orexin-ataxin3
ORX-KO	=	hypocretin/orexin knockout
SNA	=	sympathetic nerve activity

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.