Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

I’ve always envied people who sleep easily. Their brains must be cleaner, the floorboards of the skull well swept, all the little monsters closed up in a steamer trunk at the foot of the bed.

— David Benioff, City of Thieves

With insomnia, you’re never really awake; but you’re never really asleep.

— Chuck Palahniuk, Fight Club

It is difficult to overstate the importance of a good night’s sleep. Epidemiologic and experimental studies show that sleep restriction has deleterious consequences, including but not limited to impairments of mood, memory consolidation, attention/learning, creativity, athletic performance, inflammatory processes, clearance of neurotoxins, body weight/metabolism, and responsivity to stress, as well as hypertension, cancer, potential for accidents, and shortened life span. Despite the vast amount of research supporting the importance of a good night’s sleep, getting too little sleep, in the United States, for example, is quickly becoming the norm rather than the exception (Foundation 2006). The National Institutes of Health recommend that children and teens receive 9–10 h and adults 7–8 h of sleep each night. However, approximately one third of adults report getting less than 6 h of sleep nightly on average while approximately two thirds of high-school students sleep less than 8 h on a typical school night (Schoenborn et al. 2004).

Lifestyle changes can often effectively improve the quantity and quality of sleep (Montgomery and Dennis 2004). For some individuals, however, healthy sleep remains elusive despite their best efforts to implement these changes. Characterized by difficulty falling asleep, staying asleep, and/or poor sleep quality, insomnia is the most common sleep disorder, associated with incidence rates of 4–24 %, depending on the population sampled and diagnostic criteria used (Ohayon 2002; Mai and Buysse 2008; Morin et al. 2009; Ohayon and Guilleminault et al. 2010; Roth et al. 2011). The individual burden of insomnia is significant, as are its societal and economic costs, the latter estimated at more than $60 billion annually in the United States alone (Balter and Uhlenhuth 1991; Kuppermann et al. 1995; Roth and Ancoli-Israel 1999; Hajak et al. 2011; Kessler et al. 2011). Individuals suffer from impaired daytime functioning, deficits in work and cognitive performance, and mood disturbances. The societal consequences include increased healthcare costs and resources, reduced work productivity, worker absenteeism, and increased risk of accidents (Kessler et al. 2011). Effective treatments for sleep disorders are of extreme importance.

2 Current Standard of Care for Treating Insomnia

As a result of insomnia’s high prevalence and deleterious effects on functioning, well-tolerated treatments that are effective at promoting and maintaining sleep are a significant medical need. Approximately 4 % of the U.S. adult population over 20 years of age report taking a prescription sleep aid over the past 30 days (Chong et al. 2013). At the time of this writing, by far the most commonly prescribed pharmacologic treatments for insomnia are compounds that act as activators of the gamma-aminobutyric acid type A receptor (GABAA), including the non-benzodiazepines eszopiclone (Lunesta®) and zolpidem (Ambien®) (Roth et al. 2007).

Discovered well before GABAA had been identified and its mechanism of action understood, benzodiazepines and non-benzodiazepines act as allosteric modulators by binding to a site distinct from the endogenous ligand, GABA. Upon binding to GABAA, these compounds behave as positive allosteric modulators, increasing the ability of GABA to effectively open the chloride ion channel within the receptor, resulting in an influx of chloride ions and neuronal hyperpolarization (Mohler et al. 2002). As a result of decreasing the membrane potential of the neuron, the net effect of these compounds is to inhibit the propensity of neurons containing GABAA receptors to propagate action potentials.

Composed of five subunits, the GABAA receptor family exhibits exceptional diversity with heterologous combinations giving rise to a multitude of distinct subtypes. Specifically, seven subunit families (α1−6, β1−3, γ1−3, δ, ε, θ, ρ1−3) comprised of 18 subunits are found in the central nervous system (CNS). The different combinations of subunits produce heterogeneity in the channel kinetics, rate of desensitization, localization at and outside the synapse, and affinity for GABA (as well as exogenous ligands, including the hypnotics). In addition to the high level of structural diversity among the GABAA receptors, there is also a high level of diversity in CNS expression of the receptors and, therefore, the physiologic and behavioral consequences of positively modulating the receptor within each brain region. For example, the most common GABAA receptor, which contains the α1 subunit, is found at very high levels in the cerebral cortex, throughout the hippocampus, amygdala, globus pallidus, ventral pallidum, caudate nucleus and putamen, nucleus accumbens, most subnuclei of the thalamus, and olfactory bulb, and throughout the cerebellum (Mohler et al. 2002). Importantly, it appears that receptors containing the α1 subunit are responsible for the sedative hypnotic effects of the GABAA-positive allosteric modulators, as all these standard-of-care therapeutics bind to GABAA α1 subunit-containing receptors, and the sedative effects of these compounds are lost in mice with the α1 subunit genetically knocked out (Rudolph et al. 1999).

Besides the allosteric modulators of GABAA, a number of additional drugs with distinct mechanisms of action are used by patients with insomnia. These include compounds targeting serotonin signaling as well as drugs used for other indications (e.g., antidepressants and antipsychotics), over-the-counter treatments (e.g., diphenhydramine), and the recently approved melatonin receptor agonists. Because less is known about the mechanisms of action of these agents relative to the positive allosteric modulators for the GABAA receptor, this chapter will focus on comparing and contrasting the better understood allosteric modulators of GABAA with the dual orexin receptor antagonists (DORAs).

3 Orexin Receptor Antagonism—a Novel Mechanism for Treating Insomnia

In contrast to the discovery of GABAA positive allosteric modulators, which came prior to the identification of the GABAA receptor or the characterization of positive allosterism, the identification of drugs that block the orexin receptor represented a concerted effort based on an understanding of orexin genetics, and associated biology and function. The orexin peptides A and B (also referred to as hypocretin peptides) are generated from the same prepropeptide and were discovered simultaneously in 1998 by two different teams (de Lecea et al. 1998; Sakurai et al. 1998). The orexins and their receptors were subsequently characterized for their involvement in rodent and dog models of narcolepsy, as well as in human narcolepsy (Chemelli et al. 1999; Lin et al. 1999; Nishino et al. 2000). It was only a couple of years following these discoveries that the first selective orexin receptor antagonist was developed. The ensuing 15 years have seen a number of pharmaceutical companies dedicating resources to the development of orexin receptor antagonists (see accompanying chapters).

The distribution of orexinergic neurons and the orexin receptors 1 (OX1R) and 2 (OX2R) is much more spatially discrete relative to that of the GABAergic cells and GABAA receptors. For example, the human brain has approximately 70,000 orexin-synthesizing neurons (Thannickal et al. 2000), almost all of which can be found in two subnuclei of the hypothalamus, the lateral hypothalamic area and the posterior hypothalamus (Sakurai 2007). Orexin receptors (OX1R and OX2R) have overlapping and distinct localization expressed in key brain regions involved in arousal and vigilance (Marcus et al. 2001; Sakurai 2007). Importantly, orexin signaling rises during the normal active period sustaining wakefulness and falls silent during the normal sleep period (Gotter et al. 2013). In contrast, GABA is the major inhibitory neurotransmitter, synthesized ubiquitously throughout the brain. It has been estimated that GABAA is expressed in 20–50 % of the approximately 86 billion neurons in the human brain (Herculano-Houzel and Lent 2005; Herculano-Houzel 2012; Nutt and Malizia 2001). As a result of the much more restricted expression of orexin-producing neurons and orexin receptors, it was hypothesized that orexin receptor antagonists might promote sleep with far fewer side effects relative to globally acting GABAA-positive allosteric modulators.

Brisbare-Roche et al. (2007) were the first to demonstrate the sleep-promoting effects of an orexin receptor antagonist in humans. In their study, almorexant, a dual orexin receptor (OX1R and OX2R) antagonist (DORA), promoted sleep in rodents, dogs, and healthy humans (Brisbare-Roch et al. 2007). Although later shown to be active in patients with insomnia, development of almorexant was discontinued in 2011. A more recent DORA, suvorexant (Merck), has exhibited sleep-promoting effects in rodents, dogs, and humans, and is currently under evaluation by the U.S. Food and Drug Administration and other regulatory agencies after demonstrating efficacy in multiple phase 2 and 3 studies (Winrow et al. 2011; Herring et al. 2012; Michelson et al. 2013; Sun et al. 2013). Several other single or dual orexin receptor antagonists with sleep-promoting properties have also been identified.

It is now clear that DORAs are effective in promoting sleep. But how might DORAs differ from GABAA-positive allosteric modulators in a meaningful way for patients? The remainder of this chapter focuses on characterizing the differences between orexin receptor antagonism and allosteric modulation of GABAA with respect to (a) the nature of sleep produced by the two mechanisms of action, (b) motor and cognitive side effects, (c) arousability, and (d) interaction with alcohol. It is important to note that most of the work directly comparing these two mechanisms of action has been performed in preclinical species; how these findings will translate to humans is still largely unknown. As more clinical experience accumulates with orexin receptor antagonists, it will be crucial to further compare and contrast these two drug classes so that their potential benefits and liabilities can be better understood.

4 GABAA-Positive Allosteric Modulators and DORAs: Electroencephalographic Findings

Self-reported time to sleep onset, total sleep time, and wake time after sleep onset are three primary measures of sleep efficacy used in clinical studies. Both GABAA-positive allosteric modulators and DORAs impact these measures, helping individuals to fall asleep more rapidly and stay asleep. Despite these similarities, however, there are important distinctions in how each mechanism of action affects the state of the brain during sleep postdose.

Sleep can be characterized by changes in electrical activity in the brain, as measured by electrodes placed on or just below the scalp using electroencephalography (EEG). Additional measurements include eye movements using electrooculography and muscle activity using electromyography. Researchers employing these tools descriptively identified changes accompanying sleep as early as the 1930s (Loomis et al. 1936); more qualitative findings were published in the 1960s (Rechtschaffen and Kales 1968). At its simplest, these tools can be used to differentiate wakefulness from two different forms of sleep, non-rapid eye movement (NREM) (sometimes subdivided into non‒slow-wave and slow-wave sleep, or more subdivisions) and rapid eye movement (REM) sleep. More sophisticated analysis can further characterize distinct EEG signatures accompanying different phases of sleep.

Using EEG rather than the subjective measures of sleep described above, it has become evident that the electrophysiologic correlates of sleep induced by GABAA-positive allosteric modulators and DORAs differ significantly. Most clinical studies indicate that the former increase NREM (particularly light and slow-wave) sleep but decrease REM sleep (Brunner et al. 1991; Lancel 1999; Kanno et al. 2000; Dijk et al. 2010; Bettica et al. 2012), whereas the latter reportedly produce no effect on REM or NREM measures (Sun et al. 2013), or modest increases in REM with no change in NREM (slow-wave) sleep (Bettica et al. 2012; Herring et al. 2012). Although the function of REM and NREM sleep is controversial, less controversial is what subjects report when being woken from REM and NREM sleep. It is well documented that individuals report the most difficulty waking from NREM slow-wave sleep; when woken, they report feeling groggy and perform poorly on cognitive tests, a phenomenon referred to as sleep inertia (Lubin et al. 1976; Bonnet 1983). In contrast, when woken from REM sleep, individuals are more likely to report having been dreaming, and relative to slow-wave sleep, perform better on cognitive tasks and are more responsive in general (Broughton 1968; Scott 1968; Stones 1977; Dinges 1984). These findings might imply that there would be differences in capabilities upon awakening from DORA or GABAA-modulator treatment, a topic that will be addressed under the discussion below on arousability.

The effects of DORAs versus GABAA-positive allosteric modulators on spectral frequency during each of the sleep stages have also been characterized. In a human study, it was shown that the DORA SB-649868 had negligible effect on spectral frequency during either REM or NREM sleep compared to placebo. In contrast, zolpidem (10 mg) significantly impacted spectral frequency during NREM sleep compared to placebo, increasing power density in very low frequencies (0.25–1 Hz) and decreasing it in frequencies between 2.25 and 11 Hz (Bettica et al. 2012). We extended these findings in a preclinical study in which the effects of several doses of eszopiclone, zolpidem, and the dual orexin receptor antagonist DORA-22 were compared in rats following administration during the active or inactive phase (Fox et al. 2013). Doses of each compound were chosen to produce equivalent effects on total sleep. Both eszopiclone and zolpidem dose-dependently disrupted the spectral frequencies of active wake, NREM sleep, and REM sleep relative to vehicle, independent of when either compound was dosed. In contrast, only the highest dose of DORA-22 produced modest changes in the spectral profile of REM when given in the active, but not inactive phase (Fig. 1). These findings in both humans and preclinical species suggest that the spectral frequency changes produced by GABAA-positive allosteric modulators are quite different from those observed during nonpharmacologically induced sleep, whereas the effects of DORAs are quite similar.

Fig. 1
figure 1

Gamma aminobutyric acid (GABA A ) modulators (eszopiclone and zolpidem) dose-responsively alter electroencephalogram (EEG) spectral frequency of sleep/wake states in inactive-phase dosing, whereas DORA-22 produces sleep more akin to vehicle-treated animals. Shown are EEG spectral changes (as a ratio of treatment over vehicle) from 1 to 100 Hz within each sleep/wake state for each treatment at respective doses (low, mid, and high). Horizontal shaded area represents 1 (no change from vehicle) ±5 % as observed baseline dosing effect. Ninety-five percent confidence bounds were calculated, and circles indicate nonoverlapping areas above/below baseline dosing effect and 95 % confidence intervals for each sleep/wake state. All treatments and doses were collected as independent 3-day crossover study designs in male Sprague-Dawley rats (n = 16) during the inactive phase (ZT23). [Figure reproduced with permission from Fox et al. (2013)]

Full size image

5 GABAA-Positive Allosteric Modulators and DORAs: Potential Differences in Cognitive and Behavioral Effects

As alluded to above, GABA is the major inhibitory neurotransmitter in the brain and the GABAA α1 subunit–containing receptor, upon which the standard-of-care therapeutics interact, is found throughout the brain. Therefore, it is not surprising that these compounds produce a variety of effects in addition to impacting sleep. Most notably, compounds such as zolpidem and eszopiclone have been shown to produce cognitive disruption in human subjects, including deficits in attention and memory (Warot et al. 1987; Kuitunen et al. 1990; Balkin et al. 1992; Berlin et al. 1993; Roehrs et al. 1994; Allain et al. 1995; Wesensten et al. 1995; Rush and Griffiths 1996; Verster et al. 2002; Leufkens et al. 2009), and motor disturbances, such as ataxia and loss of balance (Drover 2004; Allain et al. 2005; Zammit et al. 2008; de Haas et al. 2010). These findings are aligned with the high level of GABAA α1 subunit-containing receptor found in the hippocampus and cortex, key regions involved in cognition, and the cerebellum and striatum, important for motor function. Many of these effects are also quite similar to those of ethanol, which also nonselectively impacts GABA receptor function.

Very few studies have directly compared the effects of GABAA-positive allosteric modulators versus DORAs on cognitive performance, but those that have suggest DORAs might be better tolerated. In a series of studies in rat and nonhuman primates, we evaluated the effects of zolpidem, eszopiclone, and DORA-22 on recognition memory in rats and working memory and attention in rhesus macaques (Uslaner et al. 2013). Compound testing was conducted shortly after dosing, while drug levels were at or near their maximal levels. The GABAA-positive allosteric modulators produced cognitive disruption in all three measures in both species at doses that were below or equal to those that promoted sleep. In contrast, DORA-22 increased sleep at doses 30-fold lower than the dose that impacted recognition memory, and no dose of DORA-22 tested impacted working memory or attention in rhesus macaques (Fig. 2). These findings are very similar to those of a more recent publication by Morairty et al. (2014), who demonstrated that, at doses producing equivalent levels of sleep, another DORA, almorexant, had no effect on spatial reference or spatial working memory in rats, whereas the effect of zolpidem was impairing. Finally, in rhesus macaques we have demonstrated that DORA-22 produced no next-day effects on working memory or attention, whereas diazepam produced next-day effects on both measures and eszopiclone produced effects on attention (Gotter et al. 2013). Importantly, in humans a dose of suvorexant five- or ten-fold greater than that necessary to promote sleep had effects on next-day subjective alertness (Sun et al. 2013). Future studies aimed at better characterizing the cognitive effects of these different mechanisms of action are warranted to determine the therapeutic window of each approach.

Fig. 2
figure 2

Percentage correct on a working memory task, delayed match to sample, in rhesus monkeys. Animals were given stimuli and needed to remember them over a short, medium, or long duration. Eszopiclone (3 and 10 mg/kg), diazepam (1, 5, and 10 mg/kg), and zolpidem (0.3 mg/kg) impaired performance, whereas no dose of DORA-22 impacted performance (asterisk indicates significantly different from vehicle). The minimum effective doses to produce sleep as measured by electroencephalogram were as follows: DORA-22 1 mg/kg, eszopiclone 3 mg/kg, diazepam 5 mg/kg, and zolpidem >3 mg/kg (as indicted by arrows). [Figure reproduced with permission from Uslaner et al. (2013)]

Full size image

In addition to potential differences in the cognitive impact of these drugs in subjects who are awake, it appears that DORAs and the GABAA-positive allosteric modulators may also differentially impact the ability of salient stimuli to arouse the individual when asleep. Specifically, it is important for survival and safety to be able to wake to a meaningful stimulus, such as the calling of one’s name, a fire alarm, or the sound of a predator, but to sleep through irrelevant stimuli, such as a bird chirping or background noise. We have demonstrated that dogs given vehicle or DORA-22 are much more likely to be awakened from sleep when exposed to a relevant, salient stimulus that has been previously paired with reward versus a nonrelevant, neutral stimulus (Tannenbaum et al. 2014). In stark contrast, rhesus macaques given eszopiclone or diazepam displayed no difference in their ability to awaken to a salient versus neutral stimulus, rarely waking to either (Tannenbaum et al. 2013) (Fig. 3). These results suggest that, at least in nonhuman primates, DORAs protect the ability to respond to salient stimuli when awake, whereas GABAA-positive allosteric modulators do not. Furthermore, these findings indicate that alternate arousal pathways responding to strong alerting signals remain intact while orexin signaling is silenced, as is observed when unmedicated subjects are awakened during normal sleep (and when endogenous orexin signaling is silent). This could have important consequences, as experimental data in human subjects show that GABAA-positive allosteric modulators impair the ability to arouse to salient stimuli, such as a fire alarm (Johnson et al. 1987; Mendelson et al. 1988).

Fig. 3
figure 3

Arousal from sleep to emotionally salient-conditioned or neutral acoustical stimuli at maximal nighttime drug exposure of a DORA-22, b eszopiclone, or c diazepam. a During DORA-22 sleep, rhesus monkeys (n = 12) woke to salient-conditioned acoustical stimuli significantly more than they woke to neutral stimuli. b During eszopiclone treatment, monkeys did not discriminate between salient and neutral stimuli; monkeys tended to sleep through both stimuli. c During diazepam treatment, monkeys did not discriminate between salient and neutral stimuli; at doses that induce sleep (5 and 10 mg/kg), monkeys slept through both stimuli. Data shown are mean ± standard error of the mean. ***P < 0.001. aNonsedating daytime dose as measured by electroencephalogram. [Figure based on data from a study presented by Tannenbaum et al. at the 27th Annual Meeting of the Associated Professional Sleep Societies (APSS), June 2013, Baltimore, MD (Tannenbaum et al. 2013]

Full size image

In addition to potentially differentiating with respect to cognitive performance, there is some evidence that the effects of these two mechanisms of action differ with respect to motor impairment. GABAA-positive allosteric modulators have been shown to produce motor disturbances, including loss of balance and an increased likelihood of falls, particularly in the elderly (Vermeeren 2004; Drover 2004; Allain et al. 2005; Zammit et al. 2008; de Haas et al. 2010). Far fewer clinical studies have characterized the motor effects of DORAs in humans. A very high dose of almorexant (1000 mg) was found to impact body sway; the effect was less pronounced than for zolpidem (10 mg), and it appeared to tolerate with repeated dosing (Hoever et al. 2010, 2012). In addition, whereas zolpidem produced reports of abnormal coordination and “feeling drunk,” no dose of almorexant produced such an effect (Hoever et al. 2010). However, other measures of sensorimotor processing, such as the ability of the eyes to smoothly follow an object or saccade, appear to be more sensitive to almorexant. Finally, interactions between alcohol and almorexant or suvorexant appear to be additive, but not synergistic, to any of the clinical effects produced by almorexant or suvorexant alone (Hoch et al. 2013).

Preclinical studies also suggest that DORAs might have less impact on motor coordination than GABAA-positive allosteric modulators. Using the rotarod performance test, in which the ability of an animal to maintain its balance on a rotating accelerating rod is assessed, two studies (Steiner et al. 2011; Ramirez et al. 2013) reported that two different DORAs, DORA-12 and almorexant, did not disrupt rotarod performance in rats at doses well above those necessary to produce sleep. In contrast, both studies showed that zolpidem and eszopiclone produced significant impairment in rotarod performance (Fig. 4). Finally, both studies showed that DORAs administered in the presence of ethanol did not impair rotarod performance, whereas Ramirez et al. (2013) showed that ethanol produced synergistic impairment when combined with zolpidem or eszopiclone. The interaction between ethanol and GABAA-positive allosteric modulators is likely due to the fact that these compounds act on the same receptor, whereas DORAs act on a distinct set of neuronal circuits.

Fig. 4
figure 4

Dose—response curve for latency to fall from rotarod for rats administered a GABAA receptor modulators, b ethanol, and c orexin receptor antagonists. Acute administration of zolpidem, eszopiclone, diazepam, and ethanol dose-dependently impaired rotarod performance. In contrast, the orexin receptor antagonists almorexant and DORA-12 did not impair rotarod performance after acute administration. Data shown are mean ± standard error of the mean. *P < 0.05, ***P < 0.001 versus vehicle-treated group. [Figure reproduced with permission from Ramirez et al. (2013)]

Full size image

6 Summary and Future Directions

Insomnia is a debilitating disorder that affects a significant proportion of the population and has substantial health and economic consequences. As reviewed in this chapter, the current standard of care, the GABAa positive allosteric modulators, were discovered before much was understood regarding the neural mechanisms underlying sleep. These compounds produce a range of effects in addition to increasing sleep, including reducing REM and disrupting the spectral frequency observed during normal sleep, impairing cognition and disrupting motor coordination. In comparison, the DORAs have only recently been developed and as a result have less clinical characterization, but may offer some potential advantages. The discovery and development of orexin receptor antagonists was a concerted effort of rational drug design stemming from a core understanding of orexin biology, genetics, sleep physiology, and neuroanatomy. It was hypothesized that DORAs would promote sleep without disrupting spectral frequency and impairing cognition and motor coordination, and preclinical and clinical data appear consistent with this hypothesis. More clinical experience is necessary to better understand the effect of DORAs in humans and how they compare with the standard of care. Suvorexant is currently under evaluation by regulatory agencies for the treatment of insomnia, and if approved, would represent the first DORA available for patients suffering from this sleep disorder.