2.1 Orexin neuropeptides and their receptors

The orexin system consists of orexin-A and orexin-B peptides, along with orexin receptors 1 and 2 (OX1R, OX2R) (de Lecea et al., 1998; Sakurai et al., 1998). Orexins are produced exclusively in the lateral/perifornical and dorsomedial regions of the hypothalamus (Teske & Mavanji, 2012) and project ubiquitously throughout the brain and spinal cord. In addition, orexin is detected in several peripheral tissues (Teske & Mavanji, 2012). Like orexin A and B peptides, OX1R and OX2R are observed throughout the central neuraxis and in discrete peripheral regions (Cluderay et al., 2002; Hervieu et al., 2001; Johren et al., 2002; Trivedi et al., 1998). Interestingly, both OX1R and OX2R are observed in human adipose tissue, and treatment with orexin enhances peroxisome proliferator-activated receptor gamma (PPAR-γ) expression, along with enhancing browning of adipose tissue. In addition, orexin treatment of adipose tissue results in increased release of glycerol, indicating higher lipolysis (Digby et al., 2006). These studies indicate that targeting adipose tissue with orexin may be a potential therapy for metabolic dysfunctions such as obesity and hyperglycemia (Digby et al., 2006). Even though OX1R and OX2R are concurrently present, they are unevenly distributed within the brain and spinal cord, suggesting that they may play differential physiological roles (Brown et al., 2002; Cluderay et al., 2002; Digby et al., 2006; Hervieu et al., 2001; Mavanji et al., 2010; Ohno & Sakurai, 2008; Teske & Mavanji, 2012; Trivedi et al., 1998). Affinity of orexin A is equal for both OXR's, whereas orexin B's affinity is five times greater for OX2R compared with that for OX1R (Sakurai et al., 1998). Orexins have excitatory as well as inhibitory postsynaptic effects (de Lecea et al., 1998). The orexin field projects to several brain areas implicated in physical activity, including the locus coeruleus, dorsal raphe nucleus (DRN), and substantia nigra (C. Kotz et al., 2012). Earlier studies indicate that orexin receptors (OXRs) have varying affinities for OXA and OXB, and orexin receptor binding activates several G-proteins, such as Gq, Gs, and Gi/o, indicating complex intracellular orexin signaling cascades (Sakurai et al., 1998; L. Zhang et al., 2009).

The differential affinity of orexins for its receptors may influence the effects of orexins on physiological functions. For example, narcoleptic mice with underlying orexin deficiency are prone to weight gain, and rats with higher orexin sensitivity do not gain weight when fed a high-fat diet (Kakizaki et al., 2019; Teske et al., 2006). Yet, it is difficult to parse out the role of individual OXRs in susceptibility to weight gain. Numerous studies suggest that the stimulation of OX1 and OX2 receptors results in independent functional outcomes. For example, earlier studies showed that the sleep disorder canine narcolepsy is caused by a mutation in the OX2R gene (Sakurai, 2014). In addition, it has been demonstrated that OX2R KO mice and double orexin-receptor KO mice exhibit a narcoleptic phenotype, whereas OX1R KO mice showed only a mild fragmentation of sleep/wake states (Sakurai, 2014; D. Zhang et al., 2021). Moreover, while the effects of orexin-A on wake promotion and sleep suppression were attenuated in both OX2R and OX1R KO mice, substantially greater reductions were observed in OX2R KO mice, indicating that sleep appears to be primarily regulated by OX2R and to a lesser extent by OX1R (D. Zhang et al., 2021). In contrast, food intake, emotion, autonomic regulation, and reward-related behaviors are shown to be most closely tied to OX1R receptors (Sakurai, 2014).

Several studies have demonstrated the role of OX1R in eating behavior. (Cason & Aston-Jones, 2013a, 2013b; D. L. Choi et al., 2012). The OX1R antagonist SB334867 attenuates home-cage and OXA-induced feeding (Haynes et al., 2002; Nair et al., 2008; Sakurai, 2014); high-fat pellet self-administration in food restricted rats, and ad-libitum fed mice (Cason & Aston-Jones, 2013a, 2013b; Sharf et al., 2010); binge-like consumption of palatable food in mice (Alcaraz-Iborra et al., 2014) and rats (Freeman et al., 2021); and cue-driven food consumption (Cason & Aston-Jones, 2013a, 2013b; Cole et al., 2020; Kay et al., 2014). Similarly, knockdown of OX1R in paraventricular nucleus of the thalamus (PVT) using OX1R shRNA reduces palatable food intake in mice (D. L. Choi et al., 2012), and high-fat food-conditioned place preference was inhibited by fourth ventricular administration of SB334867 (Kay et al., 2014). A rat study demonstrated that OX1R signaling mediates the consolidation and recall of Pavlovian cue–food association, as well as its extinction (Keefer et al., 2016), and oral GSK1059865, a selective OX1R antagonist, and i.p. SB33487 reduced binge eating in mice and female rats, respectively (Alcaraz-Iborra et al., 2014; Piccoli et al., 2012).

A recent study showed that lack of OX1R results in obesity resistance in mice given a high-fat diet, whereas mice lacking OX2R showed lower thermogenesis when receiving a high-fat diet. Moreover, mice deficient in either OX1R or OX2R gained weight equivalent to that observed in narcoleptic mice, that lack orexin (Kakizaki et al., 2019). These results may appear to be contradictory, but demonstrate the concept that signaling of each orexin receptor may uniquely influence energy balance. Importantly, results from studies in gene knock-out animal models need to be interpreted with caution, as compensatory responses at the intact OX receptor may yield inconclusive outcomes. The study by Kakizaki et al., also showed suppression of diet-induced obesity (DIO) in wild type (WT) mice in the presence of a running wheel, an effect which was attenuated in orexin-deficient mice, indicating that orexin neuron signaling interacts with both diet and exercise in body weight regulation (Kakizaki et al., 2019). In sum, orexin signaling increases not only food intake but also energy expenditure, and an increase in the net orexin tone generally results in thermogenesis and decreased body weight gain (Sakurai, 2014).

2.2 Orexin, spontaneous physical activity, and energy expenditure

Inherent biological mechanisms and environment influence susceptibility to obesity in humans and animals (C. M. Kotz et al., 2017). Susceptibility to DIO and co-morbidity of obesity varies extensively between individuals (Jordan et al., 2012; C. M. Kotz et al., 2017). Variability in obesity susceptibility in response to diet is partially determined by thermogenesis resulting from SPA, referred to as non-exercise induced thermogenesis (NEAT) (C. M. Kotz et al., 2017). Spontaneous physical activity in humans presents as fidgeting, standing, and ambulating (Garland Jr. et al., 2011; Vanltallie, 2001). It is believed that SPA reflects non-goal-oriented activity (such as high intensity voluntary exercise), but emanates from a subconscious drive for movement (C. M. Kotz & Levine, 2005; Levine et al., 2006). Approximately 30% of daily energy expenditure (EE) is attributed to SPA and NEAT in humans (C. M. Kotz et al., 2017), and is a major determinant of individual susceptibility to DIO (Levine et al., 1999). Some individuals increase their SPA and NEAT to resist obesity when overfed, whereas others do not, suggesting that SPA and NEAT variability among individuals critically contribute to energy homeostasis (Levine et al., 1999). Similar to human studies, most animal studies support the idea that SPA and NEAT confer protection against obesity. Mechanisms controlling NEAT and exercise involve orexin neurons (Garland Jr. et al., 2011; Nixon et al., 2012; Teske et al., 2008), and our previous work shows that the orexin system is central for the regulation of vigilance states, SPA, NEAT, and energy balance (C. M. Kotz et al., 2017), and that LH orexin is protective against weight gain (Bunney et al., 2017; DePorter et al., 2017; Levin et al., 1997; Z. Liu et al., 2020; Teske et al., 2006). For instance, earlier studies using rats that were selectively bred as obesity-prone (OP) or obesity resistant (OR, based on weight gain profiles following exposure to a high fat diet) (Levin et al., 1997), indicated that OR rats show enhanced intrinsic SPA, and orexin-induced SPA as compared with that of control and OP rats (Teske et al., 2006). In another study, we used high activity (HA) and low activity (LA) rats, where rats were classified based on their intrinsic SPA, into HA or LA rats. This study showed that similar to OR rats, HA rats resist obesity following exposure to high-energy diet in comparison to that of LA animals, and exhibit higher behavioral sensitivity to orexin (Perez-Leighton et al., 2012). Thus, both OR and HA rats demonstrate that an individual's propensity for SPA significantly determines resistance to diet-induced obesity. In a recent study, OR mice (mouse strain phenotypically identical to OR rats) increased their SPA when exposed to a cafeteria diet, without exhibiting any difference in their SPA prior to cafeteria diet exposure (Gac et al., 2015). On the other hand, in obesity susceptible C57 mice, high fat diet (HFD) feeding decreased their SPA and NEAT (Moretto et al., 2017), indicating that increased SPA and NEAT promotes obesity resistance. Moreover, physical exercise increases plasma orexin A levels, which activates the sympathetic nervous system and energy expenditure (C. M. Kotz et al., 2017; Monda et al., 2020; Polito et al., 2020). Further, exercise increases the activation of orexin neurons (James et al., 2014), and dual orexin receptor antagonist (DORA) oral administration decreases SPA for 8 h and core body temperature (CBT) for 4 h, with the CBT response being independent of SPA. Similarly, exercise-induced enhancement of CBT via treadmill running was blunted after DORA administration, further supporting a role of orexin in thermoregulation during exercise (Martin et al., 2019). Thus, it is possible that a secondary increase in OXA during exercise promotes thermogenesis.

Activity of orexin neurons is entrained to waking, as well as to external and internal signals such as fasting and caloric restriction (Alam et al., 2005) indicating their prominent role in sleep/wake and energy homeostasis regulation. Orexin neuron activity is influenced by metabolic state indicators and intra-hypothalamic and extra-hypothalamic inputs (glucose, leptin, and amino acids), such that their firing is higher during the wake state and fasting (Gac et al., 2015; Moretto et al., 2017; Sadowska et al., 2017a, 2017b). In addition, orexin neurons are sensitive to ATP and lactate levels, and thus act as energy sensors (Z. W. Liu et al., 2011; Parsons & Hirasawa, 2010). Orexin neurons influence homeostatic and physiological behaviors such as food intake, attention, sleep/wake cycle, locomotion, addiction, learning, and memory through widespread neuronal projections to regions that mediate these phenomena (C. M. Kotz et al., 2017; Mavanji et al., 2015; Stanojlovic, Pallais, & Kotz, 2019, Stanojlovic et al., 2021; Stanojlovic, Pallais, Lee, & Kotz, 2019; Stanojlovic, Pallais Yllescas Jr., Mavanji, & Kotz, 2019). Importantly, orexin neurons increase SPA in both sexes (Bunney et al., 2017; Zink et al., 2018), which is mediated by several brain sites (Kiwaki et al., 2004; C. M. Kotz et al., 2006; Teske et al., 2013; Thorpe et al., 2006; Thorpe & Kotz, 2005) and GABAergic neurons within the LH (C. M. Kotz et al., 2006). Orexins have also been shown to enhance eating behavior, but this is a short-term effect that is quickly compensated by reduced intake, resulting in no overall long-term changes in food intake following orexin stimulation (Kiwaki et al., 2004; C. M. Kotz et al., 2002; Nishino et al., 2001). In addition, central injection of orexin or brain orexin overexpression decreases body weight gain in animals (Funato et al., 2009; Novak & Levine, 2009; Perez-Leighton et al., 2012). Obesity resistant rats have greater orexin receptor gene expression and orexin behavioral sensitivity (Perez-Leighton et al., 2012; Teske et al., 2006; Teske et al., 2008); are more physically active; and have consolidated sleep (Mavanji et al., 2010; Teske et al., 2006), indicating a protective effect of orexin against obesity. As detailed above, physical activity results in energy expenditure and orexin consistently promotes SPA (Teske & Mavanji, 2012). Physically active individuals exhibit higher plasma levels of orexin (Hao et al., 2017). Further supporting the role of orexin in obesity resistance, it has been shown that orexin A injected into the rostral LH and paraventricular nucleus (Kiwaki et al., 2004; Novak et al., 2006) increases NEAT (C. M. Kotz et al., 2017; C. M. Kotz et al., 2002; C. M. Kotz et al., 2006; Teske et al., 2010), and repeated LH orexin injections injection reduces adiposity (Perez-Leighton et al., 2013). In addition, orexin injected into tuberomammillary nucleus, medial preoptic area, DRN, nucleus accumbens, substantia nigra, ventrolateral preoptic area (VLPO), and locus coeruleus (España et al., 2001; Kiwaki et al., 2004; C. M. Kotz, 2006; C. M. Kotz et al., 2008; C. M. Kotz et al., 2002; Mavanji et al., 2015; Novak et al., 2006; Novak & Levine, 2009; Teske et al., 2010; Teske et al., 2006; Teske et al., 2013; Thorpe & Kotz, 2005) enhances SPA and EE. Moreover, micro-infusion of orexin into several brain regions increased EMG activity and muscle tone (Kiyashchenko et al., 2002; Mileykovskiy et al., 2002, 2005; Peever et al., 2003; Teske & Mavanji, 2012). In a recent study, we showed that optogenetic stimulation of orexin neurons increases SPA (C. M. Kotz et al., 2017). Using designer receptors exclusively activated by designer drugs (DREADD), we showed that chemogenetic orexin neuron activation increases SPA and NEAT, and prevents obesity in mice given a high-fat diet (C. M. Kotz et al., 2017; Zink et al., 2018). Moreover, DREADD-induced activation of the orexin neuronal field mitigated aging-related reductions in SPA and EE (Stanojlovic, Pallais Yllescas Jr., Mavanji, & Kotz, 2019). Conversely, orexin antagonists reduce SPA and NEAT (Martin et al., 2019; Mavanji et al., 2015). Similarly, humans with obesity and animal models of obesity exhibit lower physical activity, reduced sleep quality, and orexin levels (in hypothalamus and plasma), whereas enhanced sleep quality and plasma orexin levels are observed following weight loss (Teske & Mavanji, 2012; C. Kotz et al., 2012). A recent study showed that the orexin neurons are necessary for movement initiation, as optogenetic-silencing of these cells reduced movement initiation, without affecting ongoing movement (Karnani et al., 2020). In addition, orexin is largely sympathoexcitatory (Teske & Mavanji, 2012). Studies in rodents demonstrate that orexin enhanced sympathetic outflow as indicated by elevated blood pressure and heart rate (Teske & Mavanji, 2012), increased renal sympathetic nerve activity including to the brown adipose tissue (BAT) (Straat et al., 2020), and elevated plasma epinephrine, noradrenaline release, and firing rate of sympathetic nerves (Teske & Mavanji, 2012). This indicates that orexin-induced NEAT has “extra” caloric expenditure which may be via sympathetic output to BAT (Morrison et al., 2014). Thus, orexins are important for energy balance, via their afferent projections throughout the CNS, including to areas crucial for the regulation of physical activity, such as the DRN, locus coeruleus, and substantia nigra (C. Kotz et al., 2012).

Orexin-deficient humans with narcolepsy, and animal models of orexin loss show a propensity for weight gain, despite decreased caloric intake (Hara et al., 2001). Over-expression of orexin results in obesity resistance in mice (Funato et al., 2009), and existing studies suggest that obesity resistance in animal models is associated with; (a) higher behavioral sensitivity to orexin A (Teske et al., 2006), (b) greater LH prepro-orexin expression, and (c) higher rostral LH orexin A sensitivity to enhance SPA (Perez-Leighton et al., 2012). As described above, OXR subtypes in many brain regions are involved in orexin-dependent regulation of energy expenditure (C. M. Kotz et al., 2017), including the rostral LH (C. M. Kotz et al., 2017), ventral lateral preoptic area, DRN, substantia nigra, and locus coeruleus (Hao et al., 2017; C. Kotz et al., 2012; C. M. Kotz et al., 2017) (Figure 1).

DRN-injected orexin increases SPA. Plots A and B represent the first and second hour post-injection time intervals. Two-way repeated measures ANOVA (dose as repeated) indicates significant effects of different doses of orexin A at 1 and 2 h post-injection. Post-hoc testing by Holmes comparison indicates that all doses are significantly different from baseline at 1 h, and the 250 and 500 pmol dose are significantly greater at 1–2 h post-injection. Bars with differing superscripts indicate that the representative means are significantly different from each other, p < 0.05; N = 11