Abstract
Although the GABAergic benzodiazepines (BZDs) and Z-drugs (zolpidem, zopiclone, and zaleplon) are FDA-approved for insomnia disorders with a strong evidence base, they have many side effects, including cognitive impairment, tolerance, rebound insomnia upon discontinuation, car accidents/falls, abuse, and dependence liability. Consequently, the clinical use of off-label drugs and novel drugs that do not target the GABAergic system is increasing. The purpose of this review is to analyze the neurobiological and clinical evidence of pharmacological treatments of insomnia, excluding the BZDs and Z-drugs. We analyzed the melatonergic agonist drugs, agomelatine, prolonged-release melatonin, ramelteon, and tasimelteon; the dual orexin receptor antagonist suvorexant; the modulators of the α2δ subunit of voltage-sensitive calcium channels, gabapentin and pregabalin; the H1 antagonist, low-dose doxepin; and the histamine and serotonin receptor antagonists, amitriptyline, mirtazapine, trazodone, olanzapine, and quetiapine. The pharmacology and mechanism of action of these treatments and the evidence-base for the use of these drugs in clinical practice is outlined along with novel pipelines. There is evidence to recommend suvorexant and low-dose doxepin for sleep maintenance insomnia; there is also sufficient evidence to recommend ramelteon for sleep onset insomnia. Although there is limited evidence for the use of the quetiapine, trazodone, mirtazapine, amitriptyline, pregabalin, gabapentin, agomelatine, and olanzapine as treatments for insomnia disorder, these drugs may improve sleep while successfully treating comorbid disorders, with a different side effect profile than the BZDs and Z-drugs. The unique mechanism of action of each drug allows for a more personalized and targeted medical management of insomnia.
I. Introduction
A. Insomnia as a Public Health Burden
Insomnia is a significant public health burden, increasing work absenteeism and health care costs in a large proportion of the population. It causes altered cognition, emotional disturbances, and reduced quality of life (Zammit et al., 1999; Wickwire et al., 2016). Insomniacs commonly complain of irritability, daytime sleepiness, low energy and motivation, physical discomfort, and impaired cognitive functioning (Buysse et al., 2007; Fortier-Brochu et al., 2012; Morin and Jarrin, 2013), not to mention deficits in working memory, episodic memory, and some aspects of executive functioning (Fortier-Brochu et al., 2012).
The prevalence rate of insomnia in the general population has been estimated as low as 5% to as high as 50% (Ohayon, 2002; Morin and Jarrin, 2013). Most epidemiologic studies have found that about one-third of adults (30%–36%) report at least one symptom of insomnia, like difficulty initiating sleep or maintaining sleep; this rate drops to 10%–15% when daytime consequences, like excessive daytime sleepiness, are added to the definition (Ohayon, 2002; Morin and Jarrin, 2013). From 1999 to 2010, the number of prescriptions for any sleep medication increased by 293% (Ford et al., 2014). Strong increases in the percentage of office visits resulting in a prescription for second generation benzodiazepines or Z-drugs (zopiclone, zolpidem, or zaleplon) sleep medications (∼350%), benzodiazepine receptor agonists (∼430%), and any sleep medication (∼200%) were noted (Ford et al., 2014).
B. Changes in the Nosology of Insomnia in Diagnostic and Statistical Manual of Mental Disorders-V
The publication of the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) (American Psychiatric Association, 2013) fundamentally changed the landscape of sleep medicine and the diagnosis of insomnia. In DSM-IV, primary insomnia was distinguished from insomnia that is secondary to another diagnosis, including major depressive disorder and generalized anxiety disorder. DSM-IV understood secondary insomnia as a symptom of a primary psychiatric disease: the secondary insomnia was expected to normalize with treatment of the primary disorder (American Psychiatric Association, 2013). However, clinical research has established that this “secondary” insomnia is often resistant to treatment of the primary disorder: in the STAR*D trial, after remission with a course of citalopram therapy, 54.9% of remitters continued to experience midnocturnal insomnia and 71.7% continued to experience sleep disturbance in some form (Nierenberg et al., 2010). DSM-5 has eliminated primary insomnia as a diagnosis in favor of insomnia disorder, which may occur alongside other diagnoses like major depressive disorder. This revised definition obliges the clinician to treat insomnia as a comorbidity, rather than a symptom of a primary illness. In this review, we use the term “insomnia disorder,” except when a published study explicitly states that it analyzes patients with “primary insomnia.”
C. Clinical Guidelines for Insomnia Treatment and the Necessity of a Translational Approach
New evidence-based clinical practice guidelines for the treatment of insomnia disorder were recently published in The Journal of Clinical Sleep Medicine (Sateia et al., 2017a), representing the first comprehensive, systematic analysis of single agents for the treatment of insomnia disorder, developed using the GRADE methodology (Grading of Recommendations, Assessment, Development, and Evaluation) (Sateia et al., 2017b). Unfortunately, the level of evidence for all of the authors’ recommendations was “weak.” This evaluation means that the strength of the evidence in the published data were low. Notably, all of the recommended treatments for sleep onset insomnia besides ramelteon are Z-drugs or BZD hypnotics. For sleep maintenance insomnia, three of five of the treatment options are Z-drugs or BZDs.
The dearth of strong published evidence led us to adopt a translational approach in exploring treatments for insomnia disorder, integrating the neurobiological mechanism of action of each drug gleaned from basic science and integrating it with reported clinical data and current medical practice. This approach is currently the standard in pharmacological research and is a priority for federal and foundational grants. Although clinical research is critical for establishing evidence-based guidelines for treatment, knowledge gleaned from basic research can be helpful for the clinical judgment of the therapeutic efficacy of hypnotics and the treatment of psychiatric comorbidities (Comai et al., 2012a,b).
In this review, we will focus on drugs that are not BZDs or Z-drug, because an extensive literature already exists. Our approach will be translation, offering alternatives to BZDs and Z-drugs. We will also describe novel hypnotic compounds and pharmaceutical pipelines.
D. The Dark Side of Benzodiazepines and Z-Drugs and the Off-Label Use of Other Drugs
Although the market for insomnia medications continues to be dominated by BZDs and Z-drugs, both categories of drug have numerous problematic effects as short-term treatment and, in particular, as long-term therapy. BZDs are associated with hangover effects the next day, cognitive or memory impairment, the rapid development of tolerance, rebound insomnia upon discontinuation, car accidents or falls, and a substantial risk of abuse and dependence (Foy et al., 1995; Hemmelgarn et al., 1997; Soldatos et al., 1999; Ashton, 2005). A large proportion of people prescribed BZD drugs become chronic users. Furthermore, BZDs are a factor in approximately 5%–10% of car accidents, although the rate in individual studies varies from 1% to 65% (Thomas, 1998).
Z-drugs also cause cognitive impairments: case control studies find that BZD or Z-drug use approximately doubles the risk of being involved in a motor vehicle accident (Thomas, 1998; Gunja, 2013b). They can produce dependence (Lugoboni et al., 2014) as well as next-day cognitive, memory, psychomotor and balance impairments (Mets et al., 2011).
The problems with BZDs have led clinicians to prescribe other medications that are perceived to be less harmful or to be less liable to addiction. As an example, in the United States in 2002, the antidepressant medication trazodone was the most commonly prescribed medication for insomnia, with 34% more prescriptions than the most commonly prescribed FDA-approved treatment (Walsh, 2004). The Prescriber National Summary Report, Calendar Year 2014 pools data from all the Medicare recipients in the United States. In one cross-sectional study of American adults, 3% of 32,328 people used a “prescription medication commonly used for insomnia” in the previous month: 38% of those who received a hypnotic medication received Z-drugs, 31% trazodone, 17% BZDs, 11% quetiapine, and only 5% received doxepin (Bertisch et al., 2014). This study confirms that drugs prescribed off-label are very common in the treatment of insomnia, despite the low number of randomized, controlled trials (RCTs).
II. Sleep Architecture
In mammals, physiological sleep is divided into two strikingly distinct states known as non-rapid eye movement (NREM) and rapid eye movement (REM) sleep. Historically, NREM sleep was subdivided into four stages (stages 1, 2, 3, 4) defined according to different electroencephalogram (EEG) patterns (Rechtschaffen and Kales, 1968). According to the manual of the American Academy of Sleep Medicine published in 2007 (Iber et al., 2007), NREM sleep is now divided into three progressively deeper stages of sleep named stage N1, stage N2, and stage N3 (formerly stages 3 and 4). REM sleep is now officially referred to as stage R. The EEG pattern in NREM sleep is synchronous and presents characteristic waveforms: sleep spindles, K-complexes, and high-voltage slow waves. Stage N1 accounts for 2%–5% of total sleep time and is the phase of transition between the awake state and sleep. Stage N2 accounts for 45%–55% of total sleep time and occurs throughout the entire sleep period. The descent from stage N1 to stage N2 is characterized by a decrease in the frequency of the EEG trace paralleled by an increase in its amplitude. The EEG hallmark of N2 are theta waves. N2 is also characterized by the occasional occurrence of a series of high-frequency waves (8–14 Hz) known as sleep spindles, generated by interactions between thalamic and cortical neurons (De Gennaro and Ferrara, 2003), and fast and high-amplitude wave forms known as K-complexes also occur (Amzica and Steriade, 2002). According to Rechtschaffen and Kales’ (1968) criteria, a K-complex is defined as a negative slow wave immediately followed by a positive wave exceeding 0.5 seconds in duration.
As stage N2 sleep progresses, high-voltage, slow-wave activity appears as the subject enters stage N3. Stage N3, which corresponds to deep or delta-wave sleep and reflects slow-wave sleep (SWS), occurs mostly in the first third of the night and accounts for 5%–15% of total sleep time. During this sleep stage, there is a further fall in blood pressure, a slowing of breathing, and a reduction in body temperature, with reduced muscle activity, although muscles maintain their tonus and thus some movements can be observed.
Stage R or REM sleep is defined by low-amplitude desynchronized theta EEG activity and represents 20%–25% of total sleep time. It occurs in four to five episodes throughout the night and is characterized by complete disappearance of muscle tone paradoxically associated with a cortical but also hippocampal activation and rapid eye movements. Since REM sleep EEG activity closely or paradoxically resembled the EEG of alert-waking subjects, this sleep stage has been also referred as paradoxical sleep, particularly in studies conducted in animals.
In physiological conditions, the activity of the brain over the course of the night proceeds from waking through the three stages of NREM sleep and then REM sleep. NREM sleep and REM sleep continue to alternate through the night in a cyclical fashion, usually with a total of four to five sleep cycles throughout the night. Importantly, as sleep progresses, the time spent in stage N3 becomes shorter, whereas the time spent in REM gets longer. The average length of the first NREM-REM sleep cycle is between 70 and 100 minutes and that of the second and later cycles is about 90–120 minutes. Sleep architecture, including the duration of the different stages as well as the duration of a NREM-REM cycle, is strongly dependent on the subject’s age. With aging, humans tend to experience an increase in the latency to fall asleep, more fragmented sleep, and less time spent in SWS, particularly in the early cycles of sleep. One of the parameters providing information concerning sleep fragmentation is the “sleep efficiency index” that is a measure of the percentage of total time in bed actually spent sleeping and is calculated by the sum of the time spent in sleep stage N1, N2, N3, and REM, divided by the total time spent in bed. Specific details on the changes in the sleep structure occurring during aging are outside the scope of this paper but are analyzed in a comprehensive recent review written by Mander et al. (2017). Restorative sleep is not only dependent on an adequate duration of sleep; the physiological architecture of sleep must be conserved. Under certain conditions and with certain pharmacological treatments, the total duration of sleep may remain unchanged or increased, but deviations from normal sleep architecture generate increased sleep fragmentation. Disturbances in subjects’ sleep architecture results in a sense of having had nonrestorative sleep and is associated with next-day impairments in conducting daily activities. Unfortunately, most of the drugs currently used as hypnotics—in particular benzodiazepines, but also Z-compounds to a lesser extent—disturb sleep architecture (Bastien et al., 2003; Gunja, 2013a).
III. The Receptor-Mediated Mechanism of Action of Hypnotics
Drugs currently used to treat insomnia mainly act on specific ionotropic or G-protein-coupled receptors located in specific brain areas. Each receptor modulates different characteristics of sleep. It is thus important for the clinician to understand the mechanism of action of each hypnotic to target better its effect in individual patients. This translational approach helps to build a more personalized medicine, targeted to the patient, that overcomes the limitations of overarching clinical guidelines. While guidelines tend to homogenize the patient population, a translational approach based on the mechanism of action of each drug may help to target the individual patient and his or her particular comorbidities.
A. GABA Receptor
The most studied receptors in the treatment of insomnia are the GABAA receptors, GABA being the chief inhibitory neurotransmitter in the mammalian nervous system, where BZDs and Z-drugs act. BZDs and Z-drugs act as positive allosteric modulators at the GABAA binding site, potentiating GABA’s inhibitory effect (Stahl, 2008). The combination of GABA at the receptor’s agonist site and benzodiazepine-receptor agonists at the allosteric site increases the frequency of the chloride channel opening to an extent that does not occur with GABA alone (Stahl, 2008). The result is neuronal inhibition. Similar to other ligand-gated ion channels, the GABAA receptor is composed of five subunits belonging to different subunit classes (α1–6, β1–3, γ1–3, δ, ε, θ, π) that are distributed throughout the brain differentially; there is also some interindividual variability in their localization. For a detailed review on this topic, please see Olsen and Sieghart (2009). In this way, BZDs and Z-drugs exert their effects as sedatives, anxiolytics, anticonvulsants, muscle relaxants, and hypnotics. The main difference between BZDs and Z-drugs is in their receptorial affinities toward the different GABAA subunits. BDZs show similar affinity to the α1, α2, α3, and α5 receptor subtypes. In contrast, most of the Z-drugs show higher affinity for a subset of the alpha subunits, mainly the α1 receptor subtype, that seems to be specifically implicated in sleep but not in anxiety. Zaleplon, zopiclone, and zolpidem have high affinity and potency for the α1 subunit and low affinity and potency at α2 and α3 subunits; eszopiclone—the (S)-enantiomer of zopiclone—has high affinity and potency for the α2 and α3 subunits (Nutt and Stahl, 2010). Due to their selective agonism, Z-drugs mainly produce sedative and hypnotic properties and likely display improved tolerability over the BZDs (Wilson and Nutt, 2007).
This review will focus on alternate mechanisms of action that are not directly mediated through GABA receptors, with the exception of the gabapentinoids (pregabalin and gabapentin). Although pregabalin and gabapentin are analogs of GABA, they do not bind directly to GABAA or benzodiazepine receptors. Instead, they inhibit the α2δ-1 subunit voltage-dependent calcium channels. They have been found to increase SWS sleep in patients diagnosed with epilepsy and insomnia (Bazil et al., 2012) and healthy adults (Foldvary-Schaefer et al., 2002; Hindmarch et al., 2005). Further development of novel α2δ calcium channels like atagabalin (PD 200390) was pursued for the treatment of insomnia but then discontinued following unsatisfactory trial results (Springer Adis Insight, 2017).
B. Serotonin Receptors
The 5-HT-containing neurons of the dorsal raphe nuclei discharge maximally during waking and decrease their firing during SWS; they cease firing during REM sleep, similar to the norepinephrine-containing neurons of the locus coeruleus (Jones, 2005). The 5-HT1A agonist OH-DPAT, which decreases 5-HT firing activity by activating its autoreceptors (Gobbi et al., 2001) increases REM sleep (Portas et al., 1996). Although 5-HT1A receptors are autoreceptors located at the somatodendritic level, 5-HT1B receptors are autoreceptors localized at postsynaptic sites. 5-HT1B receptors are also used as heteroceptors in many cells throughout the brain to inhibit the release of neurotransmitters other than serotonin (Marek, 2010). Furthermore, serotonergic neurons attenuate cortical activation through inhibitory influences on other neurons of the activating systems, including acetylcholine-containing neurons (Jones, 2005).
C. Serotonin 2 Receptors
The 5-HT2 receptor is located in the prefrontal and orbitofrontal cortex, the (subgenual) anterior cingulate cortex, the occipital, and parietal cortex (van Dyck et al., 2000; Adams et al., 2004; Hinz et al., 2007); in the nucleus accumbens, olfactory tubercule, and the hippocampus (Pompeiano et al., 1994; Barnes and Sharp, 1999; López-Giménez et al., 2001); and in the locus coeruleus, areas that are important for both sleep modulation and mood regulation (Szabo and Blier, 2001).
At the cellular level, 5-HT2A receptors are located on apical dendrites on pyramidal cells and, particularly in subcortical regions, on local (GABAergic) interneurons (Jakab and Goldman-Rakic, 1998; Barnes and Sharp, 1999; Aghajanian and Sanders-Bush, 2002). Although the mechanism of action of 5-HT2A/2C receptor antagonists has yet to be fully elucidated, it is likely that they promote SWS via a reduction of inhibitory input to the cells of the ventrolateral preoptic nucleus that fire during sleep. Through postsynaptic 5-HT2A receptors on GABAergic cells of the reticular thalamus, serotonergic fibers from the dorsal raphe and supralemniscal nucleus (B9) modulate the reticular thalamus that in turn regulates sleep and wakefulness (Rodrìguez et al., 2011).
Recent studies with more subtype-selective 5-HT2A and 5-HT2C receptor ligands (antagonists and inverse agonists), as well as experiments in knockout (KO) mice, support a role for 5-HT2A receptor subtypes in promoting SWS and the 5-HT2C receptor in promoting REM (Popa et al., 2005).
Ritanserin, a potent antagonist of the serotonin receptors 5-HT2A and 5-HT2C, has been found to increase SWS in healthy volunteers to a greater degree than ketanserin (a drug with less potent effects as an antagonist of the 5-HT2C receptor, used clinically as an antihypertensive) (Sharpley et al., 1994). Ritanserin was never commercialized for safety problems, although it did show efficacy at increasing SWS and delta activity in young male poor sleepers (Viola et al., 2002), underlining the importance of these receptors in the regulation of SWS.
Interestingly, many hypnotic drugs prescribed off-label (trazodone, mirtazapine, olanzapine, quetiapine) act through 5-HT2A and 5-HT2C receptors to enhance sleep (Landolt and Wehrle, 2009). Trazodone, the first 5-HT2A antagonist, was initially developed as an antidepressant, but is currently one of the most common hypnotics prescribed in the clinic (Bertisch et al., 2014). Similarly, the 5- and 10-mg doses of olanzapine compared with placebo, significantly increased SWS, sleep continuity measures, and subjective sleep quality (Sharpley et al., 2000). See Tables 1 and 2 for the receptorial affinities of each drug and for their effects on sleep architecture, respectively.
Summary of main molecular targets of hypnotic drugs
Mechanism-of-action details can be found in the text of this review, in review articles (Sanger, 2004; Wilson and Argyropoulos, 2005; Sills, 2006; Stahl, 2009), and/or in the summary of product characteristics for each individual drug.
Summary of the effects of the different hypnotic drugs on the sleep architecture and their potential to induce dependence liability
D. Serotonin 1A Receptors
The 5-HT1A receptor is the main 5-HT autoreceptor, located at the somatodendritic level of the 5-HT neurons of the dorsal raphe as well as at synaptic terminals in the hippocampus and prefrontal cortex. 5-HT1 knockout (KO) mice have increased 5-HT firing activity (Richer et al., 2002) and decreased REM sleep, but not SWS (Boutrel et al., 2002); the 5-HT1A agonist OH-DPAT induces a decrease in REM sleep in the first 2 hours after injection followed by an increase in REM after 6–8 hours (Boutrel et al., 2002). 8-OH-DPAT (1.0–4.0 μg) injected into the dorsal raphe nucleus increased slow-wave sleep and decreased wakefulness, although its administration of subcutaneously induced biphasic effects such that low doses decreased wakefulness and increased slow-wave sleep while higher doses induced opposite effects, perhaps due to the opposing effects of the 5-HT1A autoreceptors and heteroreceptors (Monti and Jantos, 1992). Given these opposing and complex effects, employing 5-HT1A partial agonists is a rational approach to insomnia, given their capacity to act as agonists when the levels of endogenous agonist are low and as antagonists when the levels of endogenous agonist are high.
5-HT1B knockout mice have increased REM sleep and lower SWS during the light phase and lack REM sleep rebound after deprivation, suggesting that the blockage of 5-HT1B receptors increases REM and decreases NREM sleep (Boutrel et al., 1999). In agreement with this finding, in wild-type (WT) mice, the 5-HT1B agonists CP 94253 and RU 24969 induced a dose-dependent reduction of paradoxical sleep during the 2–6 hours after injection, whereas the 5-HT1B/1D antagonist GR 127935 enhanced paradoxical sleep (Boutrel et al., 1999).
Quetiapine and trazodone both act on 5-HT1A receptors as partial agonists (Richelson and Souder, 2000; Odagaki et al., 2005).
E. Noradrenaline Receptors
Similarly to 5-HT neurons, the norepinephrine-containing neurons of the locus coeruleus (LC) nuclei discharge maximally during waking and decrease their firing during SWS; they are nearly silent during paradoxical or REM sleep (Jones, 2005).
The most important adrenergic receptors implicated in sleep are α1 and α2 receptors. α2 Receptors are located at presynaptic terminals, acting as the main norepinephrine neuron autoreceptors, although they are also present at the postsynaptic terminal. The activation of the α2 autoreceptor decreases LC activity, while, in contrast, α2 receptor blockade using α2 antagonists increases the firing activity of LC neurons (Gobbi and Blier, 2005). In fact, although the α2 agonist clonidine decreases LC activity and thus promotes sleep (particularly SWS) while inhibiting REM (De Sarro et al., 1987; Berridge et al., 2012), there is evidence that the selective α2 antagonist yohimbine increases wakefulness, at least in rats (Mäkelä and Hilakivi, 1986).
However, selective knockdown of α2A-adrenergic receptors in the LC abolished α2 agonist dexmedetomidine-induced loss-of-righting-reflex, but not sedation (Zhang et al., 2015). These findings implicate other structures besides the LC in α2 receptorial regulation of arousal: these structures include the preoptic area, in which α2 receptors are situated on GABAergic interneurons, which influence sleep through reciprocal inhibitory projections to the W systems; or the prefrontal cortex, where α2 receptors are likewise situated on GABAergic neurons (Manns et al., 2003; Luppi et al., 2017).
Indeed, when GABAergic neurons (“OFF neurons”) containing α2 receptors located in the prefrontal cortex are stimulated during SWS, their firing activity ceases, which likely produces paradoxical or REM sleep (Manns et al., 2003). This complex α2 receptor-mediated mechanism may account for the manner in which the α2 antagonist mirtazapine promotes sleep, although it increases LC firing (Gobbi and Blier, 2005).
The α1 and beta adrenergic receptors are also involved in the regulation of arousal. While individual administration of the α1 blocker prazosin produces decreased behavioral arousal and individual administration of the beta noradrenergic antagonist timolol has no effect, coadministration of prazosin and timolol produces a substantial, synergistic increase in slow-wave firing activity with a corresponding strong sedative effect behaviorally (Berridge and Espana, 2000).
F. Dopamine Receptors
Dopamine neurons are also active during wakefulness and decrease during REM sleep and SWS (Monti and Monti, 2007). In particular, the dopamine D2 receptor is one of the main receptors involved in sleep regulation. D2KO mice exhibit a significant decrease in wakefulness, with a concomitant increase in NREM and REM sleep and a drastic decrease in low-frequency electroencephalogram delta power (0.75–2 Hz) of NREM sleep, especially during the first 4 hours following lights off. In agreement with these findings, the D2 antagonist raclopride mimicked these effects in WT mice (Qu et al., 2010). Similarly, the dopamine D2 receptor antagonists haloperidol and chlorpromazine have the tendency to induce sleepiness in human subjects, while blockers of the dopamine transporter like amphetamine and modafinil increase wakefulness by increasing extracellular levels of dopamine in the synapse (Schmitt and Reith, 2010). In contrast, paradoxically, the antiparkinsonian D2 agonist ropinirole also induces sleepiness, but mostly daytime sleepiness and sleep attacks during the day (Paus et al., 2003), likely through activation of D2 autoreceptors that decrease DA firing activity during the daytime.
G. Orexin Receptors
The orexins (OXs), also known as the hypocretins, are a pair of excitatory neuropeptide hormones with approximately 50% sequence homology: orexin-A and orexin-B (hypocretin-1 and -2). They are produced exclusively by a population of neurons in the lateral hypothalamic area (de Lecea et al., 1998; Sakurai et al., 1998). Physiological effects of the OXs in the brain result from the activation of two G-protein-coupled receptors, named orexin 1 (OX1) and orexin 2 (OX2) receptors (Sakurai et al., 1998). OX1 has one-order higher affinity for OX-A than for OX-B, whereas OX2 binds OX-A and OX-B with similar affinities.
The orexinergic system is known to promote behavioral arousal, increase food intake and locomotor activity (Sakurai et al., 1998; Nakamura et al., 2000), and induce wakefulness (Hagan et al., 1999). In addition, the OXs appear to regulate the stress response by increasing the activity of the hypothalamic-pituitary axis (Al-Barazanji et al., 2001). Furthermore, they are implicated in the pathophysiology and treatment of depressive-like behavior (Nollet et al., 2011).
KO animals that lack the prepro-orexin gene, OX2 gene, or orexin neurons have narcolepsy-like behavior, including fragmentation of sleep/wakefulness, direct transitions from wake to REM sleep, and sudden loss of muscle tone while still awake (cataplexy) (Chemelli et al., 1999; Lin et al., 1999; Hara et al., 2001; Willie et al., 2003). However, OX1 KO mice have mild or almost no abnormality in the regulation of sleep and wakefulness, suggesting that the orexin signal through OX2 has a more critical role in the regulation of sleep and wakefulness, especially in the maintenance of arousal. The expression pattern of orexin receptors matches the afferent projections of orexin neurons throughout the brain. The few studies using selective antagonists of the OX1 and OX2 receptors have demonstrated that selective blockade of OX2, but not OX1, increases REM and NREM. However, coadministration of the selective OX1 antagonist and the selective OX2 antagonist intensified the effect of OX2 blockade on REM and NREM (Dugovic et al., 2009). Other studies have agreed that OX1 may play a role in the regulation of sleep and arousal (Mieda et al., 2011).
A novel orexin antagonist (suvorexant) was recently put on the market and other selective orexin antagonists are under development (Winrow and Renger, 2014).
H. Melatonin Receptors
The neurohormone melatonin activates two G-protein-coupled receptors, MT1 and MT2. Melatonin is implicated in circadian rhythms and sleep regulation, but the differential role of its individual receptors remains undefined.
Melatonin receptors have a specific localization that implicates them in physiological functions related to sleep (Lacoste et al., 2015). MT2 receptors are located in the reticular thalamus, an area involved in modulating SWS (Steriade et al., 1993), as well as the substantia nigra (pars reticulata), supraoptic nucleus, red nucleus, and the CA2, CA3, CA4 areas of the hippocampus and SCN (Ekmekcioglu, 2006; Ochoa-Sanchez et al., 2011), while MT1 is located in the locus coeruleus, the dorsal raphe, and areas CA2 and CA3 of the hippocampus and SCN (Lacoste et al., 2015).
Recently our group has better characterized the differential role of each receptor in sleep function using MT1KO, MT2KO, and double MT1-MT2KO as well as selective MT2 ligands (Comai et al., 2013). MT2KO mice have a selective disruption of SWS during the inactive phase and increased wakefulness, whereas MT1KO mice have a selective disruption of REM during the inactive phase and an increase of NREM. These results elucidate the opposing and differential effects of the two receptors in the neurobiology of sleep.
Dual MT1-MT2 KO mice only show a small increase in wakefulness, without a difference in total sleep compared with their WT counterparts (Comai et al., 2013), establishing that the MT1 and MT2 receptors may have opposing roles. This interpretation fits with the fact that melatonin and drugs that bind both MT1 and MT2 only modify the time to induction of sleep, without a global effect on total sleep time or sleep architecture (Comai et al., 2015). Thus, the absolute benefit of melatonin compared with placebo is smaller than other pharmacological treatments for insomnia (Ferracioli-Oda et al., 2013). Similarly, ramelteon, a dual agonist of both MT1 and MT2, has only been approved by the FDA for the treatment of insomnia characterized by difficulty with sleep onset; it does not affect total sleep duration or increase SWS.
I. Histamine Receptors
The sole source of brain histamine is neurons localized in the hypothalamic tuberomammillary nuclei (Haas and Panula, 2003). These neurons project axons to the whole brain, although functionally distinct histaminergic neural circuits differentially influence individual brain regions. Four histamine receptors have been identified: H1, H2, H3, and H4 (Takahashi et al., 2002). In general, histamine modulates inflammatory responses through peripheral H1 receptors and modulates gastric acid secretion through peripheral H2 receptors. This led to the discovery and therapeutic use of potent selective H1 and H2 receptor antagonists. In contrast to the H1, H2, and H4 receptors, the H3 receptor is predominantly expressed in the CNS (Lovenberg et al., 1999; Oda et al., 2000), acting as an autoreceptor on presynaptic neurons and controlling histamine turnover. H3 receptors have also been shown to act as heteroreceptors in dopamine-, serotonin-, noradrenaline-, GABA-, and acetylcholine-containing neurons (Schlicker et al., 1994).
The H1 receptor is probably the most important physiological histamine target in the maintenance of waking. In animal studies, H1 receptor agonists increase wake duration (Passani and Blandina, 2011). H1 receptor KO mice show fewer incidents of brief awakening (<16-second periods), prolonged durations of NREM sleep episodes, a decreased number of state transitions between NREM sleep and wakefulness, and a shorter latency for initiating NREM sleep. When the H3 receptor antagonist ciproxifan was administered intraperitoneally to WT mice, wakefulness increased in the mice in a dose-dependent manner but did not increase at all in H1KO mice, highlighting the interdependent functional relationship between H1 and H3 receptors (Huang et al., 2006).
H3KO mice show clear signs of enhanced histaminergic neurotransmission and vigilance, with higher EEG θ power during spontaneous wakefulness and during behavioral tasks. During the dark period, they display deficient wakefulness and signs of sleep deterioration, such as pronounced sleep fragmentation and reduced cortical slow activity during SWS, which occurs due to a desensitization of postsynaptic histaminergic receptors as a result of constant histamine release (Gondard et al., 2013).
Prescription drugs like mirtazapine, quetiapine, and hydroxyzine—not to mention nonprescription sleep aids like diphenhydramine—act on histaminergic neurons.
J. Other Receptors
Other receptors involved in sleep, but which remain to be pharmacologically exploited, include the adenosine A1 and A2 receptors (Jacobson and Gao, 2006). Interestingly, the arousal effect of the adenosine antagonist caffeine is mediated through the adenosine A2A receptor, but not the A1 receptor (Huang et al., 2005).
Adenosine mediates the somnogenic effects of prior wakefulness and likewise plays an important role in the regulation of the duration and depth of sleep after wakefulness (reviewed by Greene et al. (2017). Pharmacological data suggest that A1A receptors are involved in the regulation of sleep, although a lack of A1A receptors is not sufficient to prevent homeostatic regulation of sleep (Stenberg et al., 2003). It is conceivable that although the A1A receptor is an important factor for sleep regulation in normal animals, other factors, such as the A2A receptor, may compensate for the absence of the A1A receptor when it is deleted in knockout models. Indeed, it has been shown that the A2A receptor has a key role in adenosine-mediated sleep-promoting effects (Urade et al., 2003).
Melanin-concentrating hormone (MCH) neurons are known to be active during REM sleep and the stimulation of these neurons promotes REM sleep; indeed, electrophysiological recordings of MCH neurons across the natural sleep-wake demonstrates that they do not fire during waking, fire occasionally during NREM sleep, and fire maximally during REM sleep (Hassani et al., 2009). Importantly, they are colocalized with orexin neurons in the lateral hypothalamic area and adjacent zona incerta but as unique cell populations spatially intermingled with each other (reviewed by Yamashita and Yamanaka (2017).
Importantly, when MCH neurons are active, they inhibit orexin neurons, and knockout of MCH peptide and the MCHR1 receptor in mice produces less REM and NREM sleep. Optogenetic studies have confirmed the role of MCH neurons in inducing REM sleep: optogenetic activation of these cells during NREM sleep produces REM, but activation during wakefulness produces no effect. MCH neurons also play a role in NREM sleep, because temporally controlled ablation of these cells increases wakefulness and decreases NREM sleep duration without affecting REM sleep (Tsunematsu et al., 2014).
K. From Receptors to Sleep Circuits.
The manner in which unique neurotransmitters and individual brain areas reciprocally interact is still not understood in its entirety. The neural circuits that generate arousal and sleep (both NREM and REM) remain to be completely elucidated.
Humans are diurnal mammals, with a circadian clock that promotes wakefulness during the day, even as homeostatic sleep drive builds up. Importantly, sleep timing is phase-linked to intrinsic circadian rhythm-controlled temperature rhythms as well as extrinsic light and dark signaling (Scammell et al., 2017).
In mammals, the circadian rhythm is organized by the suprachiasmatic nuclei (SCN). The retinohypothalamic tract, which contains the intrinsically photosensitive retinal ganglion cells and the photopigment melanopsin, projects directly and monosynaptically to the SCN via the optic nerve and the optic chiasm. The SCN, which is rich in MT1 and MT2 receptors (Lacoste et al., 2015), projects to the paraventricular nucleus, and the “darkness” signal is eventually relayed to sympathetic fibers that innervate the pineal gland, which produces melatonin in response to darkness. Melatonin then stimulates the brain’s MT2 receptors in the NREM sleep activating regions of the brain: the reticular thalamus and the preoptic areas, including both the ventrolateral preoptic area (vlPO) and the median preoptic nucleus (MNPO) (Ochoa-Sanchez et al., 2011; Lacoste et al., 2015). Specifically, the MNPO appears to regulate the firing activity of the vlPO (Chou et al., 2002). It has been shown that during the transition from wakefulness to sleep, the MNPO—which specifically contains neurons that fire during slow-wave and paradoxical sleep, with slow discharging activity <5 Hz—begin to fire not before, but after, sleep onset, with a gradual increase in discharge rate (Sakai, 2011). During NREM sleep, the vlPO sends inputs that act to reduce the activity of the orexinergic arousal system and the monoamine nuclei (including the ventral tegmental area containing DA neurons, the dorsal raphe containing 5-HT neurons, and the LC containing NE neurons) by releasing the inhibitory neurotransmitters GABA and galanin. As a feedback mechanism, vlPO neurons receive reciprocal inputs from the arousal nuclei, including the ventral tegmental area, dorsal raphe, and LC; the vlPO also receives input from the histaminergic tuberomammillary nucleus (Adamantidis et al., 2010).
The reticular thalamus (RT) is another area essential for NREM sleep: people suffering from fatal familial insomnia show thalamic disruption that inactivates their ability to sleep, which is paralleled by a dysfunction in melatonin production (Portaluppi et al., 1994). RT neurons discharge in burst activity exclusively during NREM, and thalamocortical pathways project this synchronous burst activity, intermingled with periods of silence, onto the cortex. This rhythmic firing activity generates the synchronized EEG pattern typical of SWS, which produces disconnection between the cortex and the outside world (Steriade and Timofeev, 2003). The RT is also rich in melatonin MT2 receptors, which are likely activated at the beginning of sleep (Ochoa-Sanchez et al., 2011). Disconnection between the prefrontal cortex and sensory input is greatest during Stage 4 of NREM sleep, when the frequency of the EEG trace is the lowest and its amplitude is the highest. Conversely, during wakefulness, the RT and thalamocortical neurons are depolarized by inputs from the reticular activating system of the brain stem and discharge instead with a tonic activity (adapted from Steriade et al., 1993; Purves et al., 2004).
REM sleep, in contrast, is regulated by other brain areas. Many researchers have hypothesized that REM sleep is mediated mostly through cholinergic neurons located in the lateralpontine tegmentum/peduncolopontine tegmental nuclei (LDT/PPT). These neurons are active during REM sleep and generate the cortical activation and atonia typical of this sleep stage and are inactive during NREM sleep. Indeed, LDT/PPT neurons send inputs to the ventromedial medulla, which inhibits motor neurons by releasing GABA and glycine into the spinal and brain stem motor neurons, producing atonia. LDT/PPT neurons are also the main source of acetylcholine to the thalamus: activation of this acetylcholine pathway depolarizes thalamic neurons, generating the cortical activation associated with REM sleep and dreaming. Other nuclei important for REM sleep regulation are 1) the sublaterodorsal nucleus (SDL), which produces GABA and glutamate and projects to the glycinergic/GABAergic premotor neurons in the ventromedial medulla and ventral horn of the spinal cord, and through these circuits likely inhibits motor neurons during REM sleep, and 2) the MCH-containing neurons that fire during REM sleep and decrease their activity during NREM sleep and wakefulness. The “REM-off versus REM-on” theory of REM sleep hypothesizes that during the REM-on period, LDT/PPT, SDL, and MCH neurons are active and inhibit monoamine neurons as well as motor neurons, while during the REM-off period, the vlPAG/LPT is inhibited by MCH neurons and other neurotransmitters (Saper et al., 2001; reviewed in España and Scammell, 2011).
Other cholinergic nuclei that are active during REM sleep and wakefulness include the basal forebrain and the lateral hypothalamus; these same nuclei are inhibited during NREM sleep.
The manner in which the brain alternates cycles of NREM and REM remains unknown, although some researchers have proposed the existence of a mutually inhibitory circuit between vlPAG/LPT and the SDL. Figure 1 is a schematic representation of nuclei important during sleep, illustrating the circuits that modulate the sleep/wake cycle and their respective receptors.
Brain areas involved in the regulation of sleep and wakefulness with their respective receptors. Top left, green: When the brain enters NREM, neurons of the arousal system decrease their firing activity. This includes the serotonin neurons of the dorsal raphe (DR), the dopaminergic neurons of the ventral tegmental area (VTA), and the noradrenergic neurons of the locus coeruleus (LC). These neurons are silent during REM. OX1- and OX2-containing orexinergic neurons of the lateral hypothalamus decrease their firing activity during NREM and REM. The histaminergic H1-containing neurons of the tuberomammillary nucleus (TMN) decrease their firing activity during sleep. During wakefulness, these arousal centers each send widespread ascending projections to the cerebral cortex, stimulating cortical desynchronization with high-frequency gamma and low-frequency theta rhythmic activity. Bottom left, black: The receptors likely responsible for the switch from wakefulness to NREM sleep are MT1 and MT2 receptors expressed in suprachiasmatic neurons, which receive inputs directly from the retinohypothalamic tract (RHT), influenced by light and external stimuli. The transition from NREM and REM is controlled by the ventrolateral periaqueductal gray area (vlPAG), containing melatonin MT2 receptors, GABA, and glutamate receptors. Top right, red: During NREM sleep, two nuclei are particularly active: the reticular thalamus (RT), containing melatonin MT2 and GABA receptors, which is responsible for thalamocortical input to the prefrontal cortex (showing synchronized activity during NREM), and the ventrolateral preoptic area (vlPAG), containing GABA and galanin receptors, which inhibits noradrenergic, serotonergic, cholinergic, histaminergic, and hypocretinergic neurons. These nuclei play a role in the “reciprocal inhibitory” model of the sleep-wake switch. Bottom right, blue: The vlPAG is a putative “REM ON” nucleus, switching the brain to the REM sleep mode. During REM, the sublateral nucleus (SLD), the basal forebrain (BF), and the lateral tegmentum/ pedunculopontine tegmentum (LDT/PPT, rich in acetylcholine receptors) and the ventromedial medulla (VM) neurons become particularly active. Note that the basal forebrain is active in REM and wakefulness and inhibited during NREM.
IV. Melatonergic Drugs
A. Agomelatine
1. Mechanism of Action
Similar to melatonin, agomelatine inhibits firing activity in the SCN, likely through its full agonist activity at MT1; it is also an agonist at the MT2 receptor (McAllister-Williams et al., 2010). Agomelatine has low affinity for the 5-HT1A and 5-HT2B receptors, and its effects are thought to be mediated by its antagonism of the 5-HT2C receptor, with a pKi of 6.2 at human receptors (Millan et al., 2003). In animals, chronic administration of agomelatine produces a dose-dependent increase in dopamine and norepinephrine levels in the frontal cortex, without an effect on serotonin (European Medicines Agency, 2008a); like many other antidepressants, agomelatine administration is associated with increased expression of brain-derived neurotrophic factor mRNA and enhanced neurogenesis in the hippocampus (Banasr et al., 2006). In one study, agomelatine given at the onset of the late phase induced no changes in rat polygraphic recordings. However, when it was administered shortly before dark phase, agomelatine (10 and 40 mg/kg) enhanced the duration of REM and SWS sleep and decreased the duration of the wake state for 3 hours (Descamps et al., 2009). A summary of the pharmacological targets of agomelatine is reported in Table 1.
2. Pharmacokinetics
Agomelatine is rapidly and well-absorbed following oral administration (>80%) (European Medicines Agency, 2016). However, absolute bioavailability is low and variation between individuals is substantial, with increased bioavailability in women compared with men. Elderly people likewise experience greater exposure to the drug, with AUC and Cmax 4- and 13-fold higher for patients ≥75 years old compared with patients <75 years old (European Medicines Agency, 2016). Smoking, oral contraceptive pills, and the presence of hepatic impairment likewise significantly affect agomelatine’s pharmacokinetics. The metabolism of agomelatine occurs mainly via hepatic CYP1A2, and CYP1A2 polymorphisms have been shown to significantly affect the pharmacokinetics the drug (Song et al., 2014).
3. Indications
Agomelatine is approved for use in the European Union and Canada for the treatment of major depressive disorder (MDD) but not approved in the United States (Sansone and Sansone, 2011). Clinical studies examining the hypnotic effects of agomelatine are detailed in Table 3.
Summary of studies assessing the effects of agomelatine (AGM) on sleep
4. Results in Insomnia Disorder
No studies of agomelatine as a treatment of insomnia disorder were found.
5. Other Results
The studies of agomelatine were conducted in people diagnosed with major depressive disorder. In a 6-week randomized, double-blinded comparison study (N = 332) in people diagnosed with major depressive disorder (Kasper et al., 2010), agomelatine was significantly superior to sertraline at improving sleep latency (P < 0.001) and sleep efficiency (P < 0.001); furthermore, symptoms of depression (P < 0.05) and anxiety (P < 0.05) improved significantly more with agomelatine than with sertraline. Another randomized, double-blinded study comparing agomelatine and escitalopram (N = 138) found that agomelatine resulted in a greater reduction in sleep latency than escitalopram from week 2 onward (Quera-Salva et al., 2011). Moreover, although escitalopram reduced the number of sleep cycles relative to baseline, agomelatine preserved the number of sleep cycles. Finally, a review that pooled the results from three randomized studies (N = 721) comparing agomelatine to SSRIs or venlafaxine (Quera-Salva et al., 2010) established that agomelatine increases SWS, improves sleep efficiency, and resynchronizes SWS to the first sleep cycle of the night in patients with major depressive disorder while not changing the amount or latency of REM sleep.
6. Conclusion
A summary of the effects of agomelatine on sleep architecture is presented in Table 2. There is good evidence that agomelatine is superior to other antidepressants at reducing sleep latency in patients with major depressive disorder based on one review and two randomized, double-blind comparison studies (Kasper et al., 2010; Quera-Salva et al., 2010, 2011). Based on these same studies, the evidence is weak for the use of agomelatine in insomnia.
B. Prolonged-Release Melatonin
1. Indications
Melatonin is FDA-approved as a dietary supplement with no dosage restriction. In Europe, prolonged-release melatonin 2 mg/day is approved for the treatment of insomnia in elderly patients.
PRM is a new option in the treatment arsenal for insomnia. It is targeted specifically toward older adults, potentially because endogenous melatonin production declines with age and PRM mimics the pharmacokinetics of endogenous melatonin (Lemoine and Zisapel, 2012). However, a 3-week RCT found that the effects of PRM in patients with low endogenous melatonin among all ages did not differ from placebo (Wade et al., 2010); in contrast, PRM significantly reduced sleep latency compared with the placebo in elderly patients irrespective of melatonin levels (−19.1 vs. −1.7 minutes). This finding supports the idea that PRM has targeted efficacy specifically among the elderly, the same group of patients for whom benzodiazepine treatment is discouraged due to the increased risk of falls, accidents, and cognitive impairment (“What’s Wrong,” 2004). Clinical studies examining the hypnotic effects of PRM are detailed in Table 4.
Summary of studies assessing the effects of the prolonged-release melatonin formulation (PRM) on sleep
2. Pharmacokinetics
Absorption of orally ingested melatonin is complete in healthy adults, although it may be decreased by up to 50% in the elderly (European Medicines Agency, 2017). Melatonin has linear pharmacokinetics over the dosage range of 2–8 mg, although bioavailability is only about 15%, and the rate of prolonged-release melatonin absorption is affected by food: the presence of food delayed the absorption of prolonged-release melatonin, resulting in a later and lower peak plasma concentration in the fed state. The metabolism of melatonin is mainly mediated by CYP1A enzymes, although exogenous administration of melatonin does not induce these enzymes, even at supratherapeutic dosages (European Medicines Agency, 2017).
3. Results in Insomnia Disorder
Four RCTs and one open-label trial found PRM effective in the treatment of primary insomnia in the elderly. Only one RCT, the largest one (N = 791), included patients below 55 years of age; this study demonstrated that PRM was only effective in the subgroup of patients over 55 years old, validating its specific efficacy among the elderly but not other groups (Wade et al., 2010). Among the elderly in this study, PRM reduced subjective sleep latency compared with baseline by −19.1 versus −1.7 minutes for placebo.
4. Other Results
One RCT (N = 80) in patients with mild to moderate Alzheimer’s disease with and without insomnia comorbidity found that patients treated with PRM had significantly superior cognitive performance during the trial; in contrast, PSQI scores did not significantly change in the study, although sleep efficiency was found significantly to improve in patients with and without comorbid insomnia (Wade et al., 2014).
5. Conclusion
A summary of the effects of PRM on sleep architecture is presented in Table 2. There is good evidence that PRM is effective in the treatment of insomnia disorder in adults over 55 years of age, based on four RCTs. There is also evidence that PRM is not effective in the treatment of primary insomnia in younger adults, based on one RCT (Wade et al., 2010).
C. Ramelteon
1. Mechanism of Action
Ramelteon is a potent and highly selective agonist at the MT1 and MT2 receptors with an affinity 3–16 times higher than that of melatonin (Kato et al., 2005). Its affinity for MT2 is eight times lower than its affinity for MT1. This binding profile distinguishes ramelteon from melatonin and tasimelteon, which both display more affinity for the MT1 receptor than the MT2 receptor (Lavedan et al., 2015). The hypnotic effect of ramelteon is mediated by its potent, long-lasting agonism of the melatonin receptors, because it does not exhibit affinity for benzodiazepine receptors, dopamine receptors, opiate receptors, ion channels, and does not affect the activity of various enzymes (Kato et al., 2005) (Table 1). While studies in rats and monkeys confirm that ramelteon reduces time to sleep onset without affecting total sleep time (Yukuhiro et al., 2004; Fisher et al., 2008), ramelteon increases total sleep time in cats (Miyamoto et al., 2004).
2. Indications
Ramelteon is FDA-approved for the treatment of insomnia characterized by difficulty with sleep onset (US FDA, 2010a). Notably, the European Medicines Agency initially rejected Takeda Pharmaceutical Company’s application (filed in March 2007) for lack of efficacy. Later, in September 2008, the company withdrew their Marketing Authorization Application to the Committee for Medicinal Products for Human Use (CHMP). The CHMP was concerned that the company had not demonstrated the effectiveness of ramelteon, because only one aspect of insomnia, the time to fall asleep, had been assessed in the trials (European Medicines Agency, 2008b). Furthermore, only one of the three studies that had been carried out in the natural setting found a significant difference in the time taken to fall asleep between patients taking ramelteon and those taking placebo, and this difference was considered too small to be clinically relevant. When other aspects of sleep were considered, ramelteon did not have any effect (Kuriyama et al., 2014). The CHMP was also concerned that Takeda had not demonstrated the long-term effectiveness of ramelteon (European Medicines Agency, 2008b). Clinical studies examining the hypnotic effects of ramelteon are detailed in Table 5.
Summary of studies assessing the effects of ramelteon (RMT) on sleep
3. Pharmacokinetics
At a dose range of 4–64 mg, ramelteon undergoes rapid, high first-pass metabolism and exhibits linear pharmacokinetics (US FDA, 2010a). However, the drug shows substantial intersubject variability in maximal serum concentration and area under the concentration curve. Median peak concentration occurs at about 0.75 hours after fasted oral administration. Although total absorption is at least 84%, absolute bioavailability is only 1.8% because of extensive first-pass metabolism (US FDA, 2010a). It has a half-life of about 1–2.6 hours. CYP1A2 is the major liver enzyme involved in the hepatic metabolism of ramelteon, although CYP2C and CYP3A4 are also involved to a lesser degree: the drug is extensively transformed to its hydroxylated M-II metabolite, with serum AUC values that average approximately 30 times those of the parent drug (Greenblatt et al., 2007). It has been argued that M-II, with its longer half-life and greater systemic exposure, may contribute significantly to the hypnotic effect of ramelteon: M-II has been shown to bind to human MT1 and MT2 receptors, although with lower affinity (Ki: 114 and 566 pmol/l for MT1 and MT2, respectively) (Nishiyama et al., 2014). Taking ramelteon with a high-fat meal changes its pharmacokinetics; the area under the concentration curve for a single 16 mg dose is 31% higher, whereas maximal concentration is 22% lower than when administered in a fasted state (US FDA, 2010b). For this reason, the US FDA does not recommend taking ramelteon after a high-fat meal. Moreover, clearance is significantly reduced in the elderly,
4. Results in Insomnia Disorder
Two meta-analyses found ramelteon effective at reducing subjective sleep latency time in primary insomnia (Liu and Wang, 2012; Kuriyama et al., 2014). The first study analyzed 4055 patients (Liu and Wang, 2012) and the second analyzed 5812 patients (Kuriyama et al., 2014). One, a pooled analysis of four trials comparing ramelteon to placebo, found that active treatment reduced subjective sleep latency by −4.22 minutes, 95% confidence interval −5.66 to −2.77 minutes (P < 0.00001) (Liu and Wang, 2012). The other had similar results for subjective sleep latency reduction, although it pooled results from 12 studies: −4.30 minutes, 95% confidence interval −7.01 to −1.58 minutes (Q = 23.64; df = 11) (Kuriyama et al., 2014). However, it did not find that ramelteon increased total sleep time significantly more than placebo (Kuriyama et al., 2014).
5. Conclusion
A summary of the effects of ramelteon on sleep architecture is presented in Table 2. There is strong evidence that ramelteon is effective in the treatment of insomnia disorder characterized by difficulty with sleep onset, based on two meta-analyses (Liu and Wang, 2012; Kuriyama et al., 2014).
D. Tasimelteon
1. Mechanism of Action
Tasimelteon displays comparable potency to melatonin at the MT1 receptor, whereas its affinity for MT2 is 2.1–4.4 times greater than its affinity for MT1 (Lavedan et al., 2015). Its agonism at these receptors is selective, as it lacks any other significant interactions with receptors or enzymes (Table 1).
2. Indications
Tasimelteon is the first FDA-approved treatment of non-24-hour sleep-wake disorder (non-24), for which it was granted orphan drug status. Initially, Vanda Pharmaceuticals evaluated the efficacy of tasimelteon in the treatment of insomnia in phase II and phase III studies (Vanda Pharmaceuticals Inc., 2008; Feeney et al., 2009), but the compound has only received regulatory approval for the treatment of non-24. Clinical studies examining the hypnotic effects of ramelteon are detailed in Table 6.
Summary of studies assessing the effects of tasimelteon (TMT) on sleep
3. Pharmacokinetics
The pharmacokinetics of tasimelteon is linear over dose ranges from 3 to 300 mg, with an absolute oral bioavailability of 38.3% and a mean half-life of 1.3 ± 0.4 hours (US FDA, 2014b). Like for ramelteon, taking tasimelteon with a high-fat meal lowers Cmax (by 44%) and delays Tmax (by 1.75 hours) compared with the fasted state. For this reason, the FDA recommends that the drug not be taken with food. CYP1A2 and CYP3A4 are the major isoenzymes associated with its hepatic metabolism, and major metabolites have 13-fold or less activity at melatonin receptors (US FDA, 2014b). Exposure to tasimelteon increases approximately twofold in the elderly compared with nonelderly adults and in women approximately 20%–30% compared with men. Smokers had approximately 40% lower exposure to tasimelteon than nonsmokers.
4. Results in Healthy Volunteers
Two short RCTs of tasimelteon in healthy volunteers have been published: a phase II RCT and a phase III RCT (again in healthy volunteers) of the drug as a treatment of transient insomnia induced by shifted sleep (Rajaratnam et al., 2009). In the phase II trial, the individuals were monitored for seven nights: three at baseline, three after a 5-hour advance of the sleep-wake schedule, and one night after treatment. In the phase II RCT, tasimelteon reduced sleep latency and increased sleep efficiency relative to placebo and shifted the plasma melatonin rhythm to an earlier hour. The phase III study had similar positive results.
5. Results in Insomnia Disorder
There were no published studies of tasimelteon in patients diagnosed with insomnia disorder. However, one clinical trial in patients with primary insomnia (NCT#00548340), which was published only as an abstract, found a significant mean change in latency to persistent sleep [standard error]: 45.0 [2.965] for tasimelteon 20 mg/day versus 46.4 [2.954] for tasimelteon 50 mg/day versus 28.3 [3.020] for placebo (Vanda Pharmaceuticals Inc., 2014).
6. Results in Other Conditions
Other trials were conducted in patients diagnosed with non-24 sleep-wake disorder (Lockley et al., 2015), for which tasimelteon has gained FDA approval as an orphan drug (Dhillon and Clarke, 2014). The investigators found that tasimelteon significantly improved entrainment.
7. Conclusion
A summary of the effects of tasimelteon on sleep architecture is presented in Table 2. There is poor-quality evidence for the use of tasimelteon in the treatment of insomnia disorder, based on two studies in healthy volunteers (Rajaratnam et al., 2009).
V. Orexin Receptor Antagonist Drugs
A. Suvorexant
1. Mechanism of Action
Suvorexant’s effect as a hypnotic is attributable to its selective antagonism of the orexin receptors OX1R and OX2R (Winrow et al., 2011; Yin et al., 2016) (Table 1). In vitro assay panels demonstrate suvorexant’s selectivity for the orexin receptors over 170 known receptors and enzymes (Cox et al., 2010). In mice, suvorexant has been shown to selectively increase REM sleep in the first 4 hours after dosing (Hoyer et al., 2013).
2. Indications
Suvorexant is FDA approved for the treatment of insomnia characterized by difficulty with sleep onset and/or sleep maintenance at the doses of 5, 10, 15, and 20 mg/day but not the higher doses studied in the clinical trials (US FDA, 2014a). The FDA chose not to approve suvorexant at the 30 or 40 mg/day doses studied in the phase III trials because of safety concerns, particularly next-day driving impairment at doses of 20 mg/d and higher (US FDA, 2014a). There were also a few reports of sleep paralysis and hallucinations, unconscious nighttime behaviors, and narcolepsy-like events among drug-treated subjects. Clinical studies examining the hypnotic effects of suvorexant are detailed in Table 7.
Summary of studies assessing the effects of suvorexant (SVX) on sleep
3. Pharmacokinetics
Exposure to suvorexant does not increase linearly over the dosage range 10–80 mg because the drug is absorbed less at higher doses (US FDA, 2014a). The mean bioavailability of suvorexant 10 mg is 82% and ingestion of the drug with food does not meaningfully affect AUC or Cmax but does delay Tmax by approximately 1.5 hours. Steady-state pharmacokinetics are achieved in 3 days, and the mean half-life of the drug is 12 hours (95% confidence interval: 12–13). Exposure to suvorexant is higher in women than in men, with AUC increased 17% and Cmax increased 9%, for suvorexant 40 mg.
4. Results in Insomnia Disorder
There are two systematic reviews of suvorexant as a treatment of primary insomnia (Citrome, 2014; Kishi et al., 2015). Although both reviews analyze the same four phase II and phase III trials in insomnia patients, one of them differs from the other by following PRISMA reporting guidelines (Kishi et al., 2015). Both systematic reviews analyzed the same four phase II and phase III RCTs and came to similar conclusions: that suvorexant was safe and effective for the treatment of insomnia. Suvorexant improved subjective total sleep time (weighted mean difference = −20.16, 95% confidence interval = −25.01 to −15.30, 1889 patients, three trials) and subjective time to sleep onset (weighted mean difference = −7.62, 95% confidence interval = −11.03 to −4.21, 1889 patients, three trials) (Kishi et al., 2015).
Subgroup analysis of approved (15 and 20 mg/day) versus unapproved (30 or 40 mg/day) found that for efficacy, the number needed to treat values versus placebo of suvorexant 40 and 30 mg/day and that of the 20 and 15 mg/day doses were the same: 8 (Citrome, 2014). However, for adverse effects, there were number needed to harm values versus placebo of 13 for the higher doses and 28 for the lower doses, indicating that the lower doses are better tolerated. Although the other systematic review did not specifically analyze approved versus unapproved doses, the authors performed an analysis in which they excluded the higher doses from the primary outcomes and found that suvorexant remained superior to placebo in subjective total sleep time and subjective time to sleep onset at 1 month (Kishi et al., 2015). Suvorexant was found to have an efficacy similar to the benzodiazepines, ramelteon, and sedating antidepressants at reducing symptoms of insomnia (Kishi et al., 2015).
5. Conclusion
A summary of the effects of suvorexant on sleep architecture is presented in Table 2. There is strong evidence that suvorexant is effective at reducing symptoms of insomnia disorder at doses 15–40 mg/day, based on two systematic reviews (Citrome, 2014; Kishi et al., 2015). Suvorexant exerts strong effects on increasing total sleep time. Lower doses may be preferred, per FDA guidelines, to minimize the risk of adverse effects.
VI. Antidepressant Drugs
Sedating antidepressants are commonly prescribed for insomnia: one analysis found that they were prescribed more often than the FDA-approved treatments for insomnia in 2002 (Walsh, 2004; McCall, 2016). Trazodone was the most commonly prescribed medication for insomnia in 2002, with 34% more prescriptions than the most commonly prescribed FDA-approved treatment (Walsh, 2004). In fact, there were 5.28 million prescriptions for antidepressants for insomnia and only 3.4 million prescriptions for FDA-approved hypnotics (Walsh, 2004).
A. Amitriptyline
1. Mechanism of Action
Amitriptyline is a tricyclic antidepressant with strong effects as a serotonergic reuptake inhibitor [SERT Ki (nM) = 3.13] (Vaishnavi et al., 2004) and moderate effects as a norepinephrine reuptake inhibitor [NET Ki (nM) = 22.4] (Tatsumi et al., 1997). Indeed, serotonergic-norepinephrine reuptake inhibitors lack hypnotic properties, and as thus, amitriptyline’s hypnotic effects are attributable to its profile as an H1 and H2 receptor antagonist as well as a 5-HT2A and 5-HT2C antagonist. Main molecular targets of amitriptyline are summarized in Table 1.
2. Indications
Amitriptyline is FDA approved for the treatment of major depressive disorder (Alphapharm, 2012). Clinical studies examining the hypnotic effects of amitriptyline are detailed in Table 8.
Summary of studies assessing the effects of amitryptiline (AMT) on sleep
3. Pharmacokinetics
Amitriptyline is well-absorbed with peak plasma concentrations occurring within 6 hours of oral administration (Alphapharm, 2012). The mean half-life of amitriptyline is 22.4 hours, whereas the mean half-life of its active metabolite nortriptyline is 26 hours. Amitriptyline is 96% bound to plasma proteins, and undergoes extensive first-pass metabolism in the liver to nortriptyline via N-demethylation mediated by CYP2C19 (Rudorfer and Potter, 1999). Other liver enzymes involved in its metabolism are CYP2D6 and CYP3A4 (Rudorfer and Potter, 1999). Genetic heterogeneity between patients affects the concentration of the drug in the body, particularly the ratio between amitriptyline and nortriptyline (Rudorfer and Potter, 1999).
4. Results in Insomnia Disorder
No studies of amitriptyline as a treatment of insomnia disorder were identified.
5. Other Results
Two studies of amitriptyline as a treatment of secondary insomnia were identified. One study examined amitriptyline’s effect on sleep in healthy volunteers when administered chronically (Hartmann and Cravens, 1973) and one study analyzed patients with opiate withdrawal insomnia (Srisurapanont and Jarusuraisin, 1998). The study in healthy volunteers found altered sleep patterns in the group given amitriptyline compared with placebo, with those given the drug displaying increased total sleep time, increased SWS, and less REM (“desynchronized” sleep) (Hartmann and Cravens, 1973). The other study compared amitriptyline 50 mg/day to lorazepam 1–4 mg/day in patients with opiate withdrawal insomnia, finding that although amitriptyline was likely as effective as lorazepam at relieving insomnia symptoms, it may have been associated with a hangover effect the next day (Srisurapanont and Jarusuraisin, 1998).
6. Conclusion
A summary of the effects of amitriptyline on sleep architecture is presented in Table 2. Based on amitriptyline’s effect in a study of healthy volunteers (Hartmann and Cravens, 1973) of increasing total sleep time and its efficacy in opiate withdrawal insomnia, there is weak evidence of its efficacy in the treatment of insomnia disorder.
B. Mirtazapine
1. Mechanism of Action
Mirtazapine is classified as a noradrenergic and specific serotonergic antidepressant, because it enhances adrenergic and serotonergic neurotransmission in a manner distinct from other classes of drugs. Its effects as a sedative and as a hypnotic are attributable to its blockade of the histamine H1 receptor. By antagonizing α2-autoreceptors it increases norepinephrine release; by antagonizing α2-heteroreceptors it increases serotonin release, although its effect on serotonergic systems is specific to 5-HT1A-mediated neurotransmission, because it also blocks the 5-HT2 and 5-HT3 receptors (Anttila and Leinonen, 2001) (See Table 1 for the list of mirtazapine’s main molecular targets). In mice, thermal hyperalgesia and sleep disturbance in a model of neuropathic pain were nearly completely normalized by mirtazapine administration (Enomoto et al., 2012).
2. Indications
Mirtazapine is FDA approved for the treatment of major depressive disorder in adults (US FDA, 2007b). Clinical studies examining the hypnotic effects of mirtazapine are detailed in Table 9.
Summary of studies assessing the effects of mirtazapine (MRT) on sleep
3. Pharmacokinetics
Mirtazapine is rapidly and completely absorbed and has a half-life of between 20 and 40 hours, with women exhibiting significantly longer elimination half-lives than men (mean half of life 37 hours for women vs. 26 hours for men) (US FDA, 2007b). Peak plasma concentration is reached within 2 hours of administration, and the presence or absence of food does not significantly affect its pharmacokinetics. Plasma levels are linear to dose over a dose range of 15–80 mg. The drug is 85% bound to plasma proteins. The drug’s absolute bioavailability is about 50%, and in vitro data indicate that cytochromes 2D6, 1A2, and 3A are responsible for the formation of its metabolites (US FDA, 2007b).
4. Results in Insomnia Disorder
No studies of mirtazapine as a treatment of primary insomnia were identified, except for a case series (Dolev, 2011) conducted in perimenopausal women who suffered from insomnia (N = 11) and who were not depressed by the HAM-D scale (Hamilton, 1960). In this study, the subjects were given mirtazapine 15 mg/day for 2–4 weeks followed by treatment with prolonged-release melatonin 2 mg/day concurrent with the tapering of mirtazapine over 1–3 months. Combination treatment with mirtazapine and melatonin during the tapering period reduced PSQI global scores from 14.45 ± 1 at baseline to 6.00 ± 0.7 at endpoint. Sleep latency as measured by PSQI question 2 decreased from 52.73 ± 14.04 minutes at baseline to 18.64 ± 2.87 minutes at endpoint.
5. Other Results
Five studies of mirtazapine as a treatment of secondary insomnia were identified. There was one open-label study (N = 36) in patients with advanced cancer and pain or other distressing symptoms, including insomnia (Theobald et al., 2002); one open-label trial (N = 6) in patients diagnosed with major depressive disorder and poor sleep quality (Winokur et al., 2000); one randomized trial (N = 19) comparing mirtazapine to fluoxetine in patients diagnosed with major depressive disorder and insomnia (Winokur et al., 2003); one open-label study (N = 53) in patients with cancer and comorbid MDD, anxiety disorders, or adjustment disorder (Cankurtaran et al., 2008); and one open-label study (N = 42) in patients with cancer and MDD (Kim et al., 2008). In these studies, mirtazapine was generally effective at reducing symptoms of insomnia; in the randomized trial, it was more effective than fluoxetine (Winokur et al., 2003).
6. Conclusion
A summary of the effects of mirtazapine on sleep architecture is presented in Table 2. There is weak evidence that mirtazapine is effective at reducing symptoms of insomnia disorder, based on one case series (Dolev, 2011) and the available open-label evidence of mirtazapine’s effectiveness in secondary insomnia.
C. Trazodone
1. Mechanism of Action
Trazodone’s effect as a hypnotic is attributable to its moderate antihistaminergic activity at the H1 receptor, its partial agonism at the 5HT1A receptor (Odagaki et al., 2005), its antagonism of the 5HT1C and 5HT2 receptors, and its antagonism of the postsynaptic α1-adrenergic receptor (Schatzberg and Nemeroff, 2009; McCall, 2016). It also exerts relatively weak, although specific, reuptake inhibition effects at the 5-HT transporter (see Table 1 for a summary of trazodone’s main targets). Thus, trazodone has a mixed profile as both an agonist and an antagonist of serotonin receptors. In rats, trazodone has been shown to increase NREMS without affecting REMS (Lelkes et al., 1994).
2. Indications
Trazodone is FDA approved for the treatment of depression. A survey revealed that trazodone was the first-line choice of 78% or psychiatrists when prescribing medications to treat SSRI-induced insomnia (Dording et al., 2002). As mentioned previously, in 2002, trazodone was the most commonly prescribed medication for insomnia, with 34% more prescriptions than the most commonly prescribed FDA-approved treatment (Walsh, 2004). Clinical studies examining the hypnotic effects of trazodone are detailed in Table 10.
Summary of studies assessing the effects of trazodone (TRZ) on sleep
3. Pharmacokinetics
Trazodone’s half-life is 7.0 ± 1.2 after multiple oral administration and shows linear pharmacokinetics within the dosage range of 50–150 mg/day (Nilsen et al., 1993). Its absorption is irregular in fasting subjects, but it is improved if the drug is taken after food. However, no differences are observed in the total amount of trazodone absorbed with and without food: its bioavailability values are 65 ± 6% and 63.4%, respectively (Nilsen and Dale, 1992). The drug is primarily metabolized by the liver enzyme CYP3A4, and inhibition of this enzyme by other drugs leads to high blood levels of trazodone (Rotzinger et al., 1998). CYP3A4 mediates the metabolism of trazodone to its main active metabolite, m-chlorophenylpiperazine, which has 5-HT2C agonist and 5-HT2A antagonistic properties (Rotzinger et al., 1998).
4. Results in Insomnia Disorder
One 2-week parallel group RCT (N = 306) in patients with primary insomnia was identified (Walsh et al., 1998). The study compared treatment with trazodone 50 mg/day to treatment with zolpidem 10 mg/day and placebo after a 1-week placebo lead-in period. During the first week, both drugs produced significantly shorter self-reported sleep latency and longer self-reported sleep duration than placebo. Sleep latency was significantly shorter with zolpidem than with trazodone. During week 2, only the zolpidem group maintained a significantly shorter sleep latency than the placebo group, and sleep duration did not vary significantly among groups.
5. Other Results
Two reviews on the use of trazodone as a treatment of insomnia were identified (James and Mendelson, 2004; Mendelson, 2005). Both reviews predominantly analyzed studies of trazodone in patients diagnosed with major depressive disorder. Although trazodone was shown to increase total sleep time in patients with MDD (James and Mendelson, 2004), there was limited evidence of its efficacy (Mendelson, 2005). The high rate of discontinuation due to adverse events, which included sedation, dizziness, and psychomotor impairment, make the risk-benefit ratio of trazodone therapy for insomnia uncertain (Mendelson, 2005). Furthermore, there is a risk of priapism in 1 out of 6000 patients treated with trazodone (James and Mendelson, 2004).
Nine RCTs on the use of trazodone in the treatment of insomnia secondary to other conditions were identified, of which three were conducted in patients with insomnia secondary to treatment with antidepressants (Nierenberg et al., 1994; Haffmans and Vos, 1999; Kaynak et al., 2004). One crossover RCT (N = 17) of trazodone 50 mg/day versus placebo in antidepressant-associated insomnia secondary to treatment with fluoxetine or bupropion (Nierenberg et al., 1994) and one 2-week crossover RCT (N = 12) of trazodone 100 mg/day versus placebo in patients with insomnia secondary to treatment with SSRIs (Kaynak et al., 2004) found trazodone effective, with significantly increased total sleep time, sleep efficiency, sleep continuity, and increased stage 3 and stage 4 sleep in the second trial. However, one smaller RCT (N = 7) of trazodone 50 mg/day in antidepressant-associated insomnia secondary to treatment with brofaromine (Haffmans and Vos, 1999) found that trazodone did not improve sleep latency, total sleep time, or time awake versus placebo, although it increased SWS. These results suggest that trazodone may be effective in cases of insomnia induced by SSRIs or bupropion, but not brofaromine, an antidepressant that was never brought to market.
A 6-week study of trazodone versus venlafaxine versus placebo in patients diagnosed with major depression was also identified (Cunningham et al., 1994): in this study, trazodone was more effective for improving sleep disturbance on the HAM-D, but venlafaxine was better at relieving cognitive disturbance and retardation. Trazodone caused more dizziness while venlafaxine caused more nausea.
Three studies of trazodone in the context of addiction were identified. One 4-week RCT (N = 16) of trazodone 50 mg/day versus placebo in patients with alcohol-induced insomnia and alcohol dependence (Le Bon et al., 2003) found that trazodone increased sleep efficiency. However, caution is warranted because a large 12-week RCT (N = 173) of trazodone 50–150 mg/day versus placebo in patients with alcohol dependence and sleep disturbances (Friedmann et al., 2008) found though trazodone reduced symptoms of insomnia, and the trazodone group experienced less improvement in the proportion of days abstinent during detoxification when receiving medication; furthermore, the trazodone group had an increase in the number of drinks per drinking day upon cessation of the study. A 6-month RCT (N = 137) of trazodone 50–150 mg/day during methadone maintenance treatment of opioid dependence (Stein et al., 2012) was negative, with placebo-treated subjects reporting significantly higher sleep quality ratings than trazodone-treated subjects.
Finally, a 2-week RCT (N = 30) of trazodone 50 mg/day in patients with Alzheimer’s disease found that trazodone-treated subjects slept significantly longer than placebo-treated subjects, although the drug did not have a detectable effect on cognition (Camargos et al., 2014).
6. Conclusion
A summary of the effects of trazodone on sleep architecture is presented in Table 2. There is good evidence that trazodone is effective at reducing symptoms of insomnia in patients with SSRI-induced insomnia based on two RCTs (Nierenberg et al., 1994; Kaynak et al., 2004). Trazodone use is discouraged in insomnia associated with opioid dependence or alcoholism based on one negative RCT in patients on methadone maintenance treatment (Stein et al., 2012) and the safety concerns in one RCT conducted in patients with alcoholism (Friedmann et al., 2008).
D. Low-Dose Doxepin (<6 mg/day)
1. Mechanism of Action
Doxepin is a tricyclic antidepressant with significant antihistaminic effects. Doxepin is the most potent antihistamine of the tricyclic antidepressants, with four times the potency of amitriptyline and 800 times the potency of diphenhydramine at the H1 receptor (Richelson, 1979; Gillman, 2007). At standard antidepressant doses, >75 mg/day, doxepin inhibits the reuptake of serotonin and norepinephrine and antagonizes cholinergic, histaminergic, and α-adrenergic activity. As a hypnotic, doxepin is used at low doses; at doses <10 mg/day, it theoretically affects only the histamine receptor, with no meaningful effects on the noradrenergic and serotonergic systems (McCall, 2016) (see Table 1).
2. Indications
Low-dose doxepin , under the brand name Silenor (Pernix Therapeutics, Morristown, NJ), is FDA approved for the treatment of insomnia characterized by difficulties with sleep maintenance to a maximum dose of 6 mg/day (US FDA, 2010b). Doxepin is also approved as an antidepressant in the treatment of “psychoneurotic patients with depression and/or anxiety.” Clinical studies examining the hypnotic effects of low-dose doxepin are detailed in Table 11.
Summary of studies assessing the effects of low-dose doxepine (DXP) on sleep
3. Pharmacokinetics
Low-dose doxepin has a terminal half-life of 15.3 hours, whereas the half-life of nordoxepin, Median to peak concentration (Tmax) of doxepin 6 mg occurs 3.5 hours after oral administration to healthy fasted subjects; Tmax is delayed by approximately 3 hours if the drug is taken with a high-fat meal, whereas AUC is increased by 41% and Cmax by 15% (US FDA, 2010b). For this reason, it is recommended that doxepin be taken without food, to minimize the risk of next day effects. The major liver enzymes responsible for the metabolism of doxepin are CYP2C19 and CYP2D6, whereas CYP1A2 and CYP2C9 are involved to a lesser extent. The drug is 80% bound to plasma proteins. Notably, doxepin interacts with cimetidine (causes a twofold increase in doxepin Cmax and AUC) and sertraline (causes AUC to be increased 21% and Cmax to be increased 32%) (US FDA, 2010b).
4. Results in Insomnia Disorder
One systematic review (Yeung et al., 2015), five published RCTs (Roth et al., 2007; Scharf et al., 2008a; Krystal et al., 2011, 2010; Lankford et al., 2012), and one unpublished RCT (Takeda Global Research & Development Center Inc., 2008) of low-dose doxepin as a treatment of primary insomnia were identified. The systematic review included two studies of doxepin at antidepressant doses (25–300 mg/day) that are excluded in this review. The authors did not perform meta-analysis of pooled results due to heterogeneity but confirmed that low-dose doxepin had a small to medium effect size versus placebo for sleep maintenance and sleep duration, but was ineffective at improving the time to sleep onset. The findings in the individual RCTs cited above generally came to similar conclusions as the recent systematic review, although some RCTs used subjective measurements and others used polysomnographic measurements.
The unpublished clinical trial is a double-dummy study of ramelteon + low-dose doxepin versus each drug as monotherapy versus placebo (Takeda Global Research & Development Center Inc., 2008). It found that ramelteon + low-dose doxepin was significantly more effective than ramelteon + placebo by polysomnography-measured wake time after sleep onset and total sleep time, as well as subjective wake time after sleep onset.
5. Other Results
A single night RCT of low-dose doxepin in healthy volunteers was also identified (Roth et al., 2010). In this trial, the investigators attempted to induce transient insomnia using the first-night effect as well as a 3-hour phase advance. Low-dose doxepin was effective at reducing latency to sleep and increasing total sleep time.
6. Conclusion
A summary of the effects of low-dose doxepin on sleep architecture is presented in Table 2. There is strong evidence that low-dose doxepin is effective at reducing symptoms of primary insomnia based on one systematic review (Yeung et al., 2015), five published RCTs (Roth et al., 2007; Scharf et al., 2008a; Krystal et al., 2011, 2010; Lankford et al., 2012), and one unpublished RCT (Takeda Global Research & Development Center Inc., 2008). Low-dose doxepin exerts a strong on improving sleep maintenance.
VII. Anticonvulsant Drugs
A. Gabapentin
1. Mechanism of Action
Gabapentin is an anticonvulsant drug that binds to the α2δ subunit of voltage-sensitive calcium channels (Gee et al., 1996) (Table 1). It crosses several lipid membrane barriers; in vitro, it has been shown to modulate the activity of the GABA synthesizing enzyme, glutamic acid decarboxylase, and the glutamate synthesizing enzyme, branched-chain amino acid transaminase (Taylor, 1997). Its modulation of the GABAergic and glutamatergic systems probably underlies its effect as a hypnotic and anxiolytic. Gabapentin is known to increase SWS without affecting other polygraphic variables and without causing increased drowsiness during the day (Foldvary-Schaefer et al., 2002). In mice, gabapentin alleviates sleep disturbances induced by a neuropathic pain-like condition (Takemura et al., 2011).
2. Indications
Gabapentin is FDA-approved for the management of postherpetic neuralgia in adults (US FDA, 2014c). It is FDA approved as adjunctive treatment of partial seizures with and without secondary generalization in patients over 12 years old with epilepsy and as adjunctive treatment of partial seizures in patients aged 3–12. Clinical studies examining the hypnotic effects of gabapentin are detailed in Table 12.
Summary of studies assessing the effects of gabapentin (GBP) on sleep
3. Pharmacokinetics
The bioavailability of gabapentin is inversely proportional to its daily dose: as dosage increases, bioavailability decreases. The bioavailability of gabapentin is 60%, 47%, 34%, 33%, and 27% following oral administration of 900, 1200, 2400, 3600, and 4800 mg/day of the drug given in three divided doses (US FDA, 2014c). Less than 3% of gabapentin is bound to plasma protein, and the drug is not appreciably metabolized in humans. Its elimination half-life is 5–7 hours and is unaltered by dose or following multiple dosing. Taking the drug with food has a small effect on its pharmacokinetics, increasing AUC and Cmax by 14% each.
4. Results in Insomnia Disorder
One open-label trial conducted in patients with primary insomnia was identified (Lo et al., 2010). This study (N = 18; mean dose of gabapentin 540 mg/day, range 200–900 mg/day) found polysomnographic evidence of increased sleep efficiency (80.00%–87.17%, P < 0.05) and SWS (10.47%–17.68%, P < 0.005) and decreased wake time after sleep onset (16.45%–7.84%, P < 0.05) (Lo et al., 2010). However, gabapentin did not significantly improve sleep onset latency (17.58–14.58 minutes, not significant).
5. Other Results
Five other studies of gabapentin were identified. Two of the four studies were RCTs: one was conducted in patients diagnosed with alcohol dependence and comorbid insomnia (Brower et al., 2008) and the other was a single-dose study conducted in patients with “occasional disturbed sleep” (Rosenberg et al., 2014). One open-label comparison study of gabapentin versus trazodone conducted in patients with alcohol dependence and “persistent insomnia” (Karam-Hage and Brower, 2003) and an open-label study conducted in children suffering from refractory insomnia comorbid to neurodevelopmental or neuropsychiatric disorders (Robinson and Malow, 2013) were also identified. Finally, one study in healthy subjects was identified (Foldvary-Schaefer et al., 2002).
In the RCT conducted in patients with alcohol dependence and comorbid insomnia (N = 21; gabapentin 1500 mg/day), treatment group did not predict changes in the Sleep Problems Questionnaire score (Brower et al., 2008). However, gabapentin treatment significantly reduced the risk of relapse to heavy drinking: at 12 weeks, 60% of the gabapentin group had relapsed compared with 100% of the placebo group. Similarly, in the single-dose study using a 5-hour phase advance model in patients with “occasional disturbed sleep,” gabapentin 250 mg/day and gabapentin 500 mg/day were not significantly superior to placebo at reducing latency to persistent sleep. However, the mean total sleep time was significantly greater for the gabapentin groups: 311.4 [8.4] min in the placebo group versus 356.5 [7.5] min in the gabapentin 250 mg/day group (P ≤ 0.001 compared with placebo) and 378.7 [7.3] min in the gabapentin 500 mg/day group (P ≤ 0.001 compared with placebo, P ≤ 0.01 compared with gabapentin 250 mg/day). Wake after sleep onset was significantly improved in the gabapentin groups, as was the proportion of time spent in SWS (stages 3 and 4) (Rosenberg et al., 2014). Finally, in the open-label comparison study of gabapentin and trazodone in alcoholism, gabapentin was significantly more effective.
6. Conclusion
A summary of the effects of gabapentin on sleep architecture is presented in Table 2. There is some evidence that gabapentin is effective in the treatment of insomnia disorder according to one open-label trial, although it did not significantly improve sleep onset latency (Lo et al., 2010).
B. Pregabalin
1. Mechanism of Action
Similar to gabapentin, pregabalin binds to α2δ subunit-containing voltage-gated calcium channels (Taylor et al., 2007). Pregabalin also modulates the influx of calcium at nerve terminals, which may accounts for its therapeutic benefit in neuropathic pain, seizures, and anxiety. The mechanism of action of pregabalin (Table 1) in improving sleep has not been completely elucidated, but it is known to be different from benzodiazepines as pregabalin is not active at GABA-A or benzodiazepine receptors (Taylor et al., 2007). Furthermore, although benzodiazepines typically reduce SWS, pregabalin has been found to increase it (Hindmarch et al., 2005). In rats, pregabalin has been found to increase NREMS and REMS, while markedly increasing the duration of NREMS episodes and reducing their number (Kubota et al., 2001). In one study, pregabalin increased NREMS in mice with a neuropathic pain-like condition, but not normal mice (Wang et al., 2015).
2. Indications
Pregabalin is FDA approved for the management of neuropathic pain associated with diabetic peripheral neuropathy, the management of postherpetic neuralgia, adjunctive therapy for adult patients with partial onset seizures, the management of fibromyalgia, and the management of neuropathic pain associated with spinal cord injury (US FDA, 2016). Clinical studies examining the hypnotic effects of gabapentin are detailed in Table 13.
Summary of studies assessing the effects of pregabalin (PGB) on sleep
3. Pharmacokinetics
Following oral administration, peak plasma concentrations of pregabalin occur within 1.5 hours; its bioavailability is greater than or equal to 90% and is independent of dose (US FDA, 2016). The half-life of pregabalin is about 6 hours. Taking pregabalin with food increases Tmax to approximately 3 hours and reduces Cmax by 25%–30%, although pregabalin can be taken with or without food. It does not bind to plasma proteins and undergoes negligible metabolism in humans (US FDA, 2016). Oral clearance tends to decrease with increasing age.
4. Results in Insomnia Disorder
No studies of pregabalin as a treatment of insomnia disorder were identified.
5. Other Results
Three reviews were found analyzing the use of pregabalin in patients with insomnia: in two, patients were primarily diagnosed with generalized anxiety disorder (Montgomery et al., 2009; Holsboer-Trachsler and Prieto, 2013), and in one, patients were primarily diagnosed with fibromyalgia (Russell et al., 2009). In the most recent review, pooled data from four RCTs (N = 1354) established that among patients with severe difficulty in falling asleep, remission was observed in 54.0% of the pregabalin group versus 29.8% of placebo group (Holsboer-Trachsler and Prieto, 2013). In those with severe difficulty in staying asleep, remission was observed in 54.2% of pregabalin group versus 26.7% of placebo group. Finally, in those with severe difficulty associated with waking up too early, remission was observed in 59.4% of pregabalin group versus 34.6% of placebo group. The fibromyalgia study (N = 1493, two RCTs) likewise found that pregabalin was significantly superior to placebo in reducing the burden of insomnia symptoms as measured using the Sleep Quality Diary and Medical Outcomes Study Subscales of Sleep Disturbance, Quantity of Sleep, and Sleep Problems Index (Russell et al., 2009). For more information about these psychometric scales, see Smith and Wegener (2003).
6. Conclusion
A summary of the effects of pregabalin on sleep architecture is presented in Table 2. There is good evidence based on two reviews (Montgomery et al., 2009; Holsboer-Trachsler and Prieto, 2013) that pregabalin is effective at reducing symptoms of insomnia in generalized anxiety disorder. There is also good evidence based on one review (Russell et al., 2009) that pregabalin is effective at reducing symptoms of insomnia in fibromyalgia. Based on these same reviews, there is weak evidence that pregabalin is effective in the treatment of insomnia disorder.
VIII. Atypical Antipsychotic Drugs
Sedating atypical antipsychotics, particularly quetiapine, are often used in the clinic for the management of insomnia disorder and insomnia symptoms that occur comorbid to psychiatric illness (Pringsheim and Gardner, 2014). Although they are effective in the management of bipolar and psychotic disorders, systematic reviews have found a lack of evidence for the use of sedating atypical antipsychotics and explicitly recommend against prescribing them for insomnia disorder (Thompson et al., 2016). The 2016 meta-analysis acknowledges that the use of atypical antipsychotics may be appropriate in patients who have failed other treatment modalities and who have a comorbid condition that could benefit from the primary action of the drug (based on consensus of experts in sleep medicine). Similarly, guidelines for the treatment of chronic insomnia report insufficient evidence for atypical antipsychotics as first-line therapy (Schutte-Rodin et al., 2008), but state that the medications may be suitable for patients with comorbid insomnia who may benefit from the primary action of these drugs as well as from the sedating effect.
A. Olanzapine
1. Mechanism of Action
Olanzapine is an atypical antipsychotic with affinity for the dopamine D1, D2, and D4 receptors; the serotonin 5-HT2A, 5-HT2C, and 5-HT3 receptors; the α1-adrenergic receptor; the histamine H1 receptor; and five muscarinic receptor subtypes (Bymaster et al., 1996). Its hypnotic effects are probably attributable to its strong antagonism of the H1 antagonism as well as its antagonism of serotonin receptors. In an assay of compounds tested at the histamine H1 receptor, olanzapine was the most potent compound Richelsen and Souder (2000) had tested of any class of compounds. Olanzapine seems to specifically increase SWS (Salin-Pascual et al., 1999; Giménez et al., 2007; Kluge et al., 2014). Main molecular targets of olanzapine are summarized in Table 1.
2. Indications
Olanzapine is FDA approved for the treatment of schizophrenia in adults and adolescents; for the acute treatment of manic or mixed episodes associated with bipolar I disorder and the maintenance treatment of bipolar I disorder in adults and adolescents; and, as an intramuscular injection, for the treatment of acute agitation associated with schizophrenia and bipolar I mania (US FDA, 2009). Clinical studies examining the hypnotic effects of olanzapine are detailed in Table 14.
Summary of studies assessing the effects of olanzapine (OLA) on sleep
3. Pharmacokinetics
Olanzapine reaches peak concentrations about 6 hours following oral administration, and its elimination half-life ranges from 21 to 54 hours, with a mean of 30 hours (US FDA, 2009). It is eliminated extensively by first-pass metabolism: about 40% of an oral dose is metabolized before it reaches the systemic circulation. Olanzapine is extensively metabolized by CYP1A2, CYP2D6, and the flavin monooxygenase system, although its metabolites do not display pharmacological activity at normal concentrations (US FDA, 2009). Its pharmacokinetics are not affected by food. The drug is 93% bound to plasma protein. Clearance of olanzapine is approximately 30% lower in women than in men, and the elimination half-life of olanzapine is about 1.5 times higher in subjects greater than 65 years old.
4. Results in Insomnia Disorder
No studies of olanzapine as a treatment of insomnia disorder were identified.
5. Other Results
Two studies that examined olanzapine’s effect on sleep (Salin-Pascual et al., 1999; Sharpley et al., 2000) and three studies on olanzapine as a treatment of secondary insomnia were identified (Jakovljević et al., 2003; Khazaie et al., 2010, 2013). Of the studies analyzing olanzapine’s effect on sleep, both were small: one was an open-label study conducted in 20 patients with schizophrenia, during which polysomnographic recordings were taken for 5 days, with patients only receiving olanzapine on two nights (Salin-Pascual et al., 1999) and one was a one-dose crossover RCT conducted in nine healthy male participants, with 7–14 days washout between doses (Sharpley et al., 2000). The crossover RCT examined olanzapine 5, 10 mg/day or placebo, whereas the study in patients with schizophrenia examined olanzapine 10 mg/day. Both studies found that olanzapine profoundly increased slow-wave sleep and increased total sleep time as acute treatment.
Of the studies in secondary insomnia, one was conducted in patients suffering from treatment-resistant posttraumatic stress disorder (PTSD) (Jakovljević et al., 2003), whereas the other two were in patients with paradoxical insomnia, also known as sleep state misperception (Khazaie et al., 2010, 2013). The first study was in patients suffering from intractable PTSD in which open-label olanzapine was added to their current medications (Jakovljević et al., 2003). The patients had recurrent nightmares and insomnia resistant to numerous medications. In all five cases, olanzapine resulted in a significant improvement in their symptoms. One randomized, open-label study was conducted in patients diagnosed with paradoxical insomnia, or sleep-state misperception (Khazaie et al., 2013). Patients suffering from this disorder complain of difficulties with initiating and maintaining sleep; however, the hallmark of paradoxical insomnia is that objective polysomnographic measures find that the patients are getting sufficient sleep. In this study, the investigators followed up on a case report study they had published in 2010 (Khazaie et al., 2010), in which they reported successful treatment of recalcitrant, paradoxical insomnia with olanzapine in a single patient. The group’s larger study compared treatment with two different atypical antipsychotics: olanzapine and risperidone. They found that, although both treatments were associated with significant improvements in subjective sleep quality, olanzapine was significantly superior to risperidone, as measured using the Pittsburgh Sleep Quality Index (PSQI) (Buysse et al., 1989).
6. Conclusion
A summary of the effects of olanzapine on sleep architecture is presented in Table 2. There is weak evidence that olanzapine acutely increases slow-wave sleep and total sleep time (Salin-Pascual et al., 1999; Sharpley et al., 2000), although this effect has not been confirmed in patients with insomnia. There is also evidence that olanzapine may be useful in 1) the treatment of insomnia associated with PTSD (Jakovljević et al., 2003) based on a case series and 2) paradoxical insomnia, based on a case report (Khazaie et al., 2010) and a randomized, open-label trial (Khazaie et al., 2013).
B. Quetiapine
1. Mechanism of Action
Quetiapine, a dibenzothiazepine derivative, is the atypical antipsychotic that displays the lowest D2 affinities (Richelson and Souder, 2000; Comai et al., 2012b). It shows antagonism at multiple neurotransmitter receptors, mainly 5-HT2A, 5-HT2c, H1, and D2 (Table 1). Its sedative and hypnotic properties are attributable to its antagonism of the histamine H1 receptor and various serotonin receptors. In monkeys, neither acute nor chronic administration of quetiapine improved sleep efficiency, whereas the first night after discontinuation, subjects had significantly decreased sleep efficiency and increases in nighttime activity (Brutcher and Nader, 2015).
2. Indications
Quetiapine is FDA approved for the treatment of schizophrenia in adults and adolescents, the treatment of bipolar mania in children and adolescents, and the treatment of bipolar depression in adults. Clinical studies examining the hypnotic effects of quetiapine are detailed in Table 15.
Summary of studies assessing the effects of quetiapine (QTP) on sleep
3. Pharmacokinetics
Quetiapine fumarate is rapidly absorbed after oral administration, reaching peak plasma concentrations within 1.5 hours (US FDA, 2017). The drug is 83% bound to serum proteins (DeVane and Nemeroff, 2001). Administration with food increases Cmax and AUC by 25% and 15%, respectively. The drug is mainly eliminated through hepatic metabolism, specifically CYP3A4, and its mean terminal elimination half-life is 6 hours. Oral clearance is reduced by 40% in subjects greater than 65 years of age, although sex does not affect its pharmacokinetics.
4. Results in Insomnia Disorder
One RCT was identified in patients diagnosed with primary insomnia (Tassniyom et al., 2010). Surprisingly, although observational evidence suggests that atypical antipsychotics like quetiapine are increasingly prescribed for insomnia, the 2010 study was the only RCT that was found in the literature of an atypical antipsychotic in primary insomnia, and it had a small sample size of only 13 patients who completed the study (Tassniyom et al., 2010). In the RCT, quetiapine 25 mg/day treatment was not significantly superior to placebo at increasing sleep time and reducing latency to sleep, although there was a trend toward the superiority of quetiapine (Tassniyom et al., 2010).
In contrast, an open-label trial of quetiapine 25–75 mg/day found that the drug was effective at reducing symptoms of insomnia, increasing total sleep time and reducing PSQI (Wiegand et al., 2008).
5. Other Results
Eleven other studies of quetiapine that included sleep parameters were identified: five were open label (Juri et al., 2005; Todder et al., 2006; Baune et al., 2007; Pasquini et al., 2009), three were randomized, placebo-controlled trials (Cohrs et al., 2004; Garakani et al., 2008; McElroy et al., 2010), one was a review (Anderson and Vande Griend, 2014) pooling many of the studies cited here, one was a naturalistic study (Šagud et al., 2006), one was a post hoc analysis of two RCTs (Endicott et al., 2008), and one was a retrospective study (Terán et al., 2008).
One RCT in 14 healthy male subjects found that quetiapine 25 and 100 mg/day significantly improved sleep induction and sleep continuity under standard and acoustic stress conditions (Cohrs et al., 2004). Active treatment with quetiapine also increased total sleep time, sleep efficiency, and subjective sleep quality.
In contrast to the results of the RCT in primary insomnia, the open-label studies in other conditions were generally positive, although they were conducted in a wide range of patient populations. The included studies analyzed patients diagnosed with Parkinson’s disease, treatment-resistant depression, bipolar disorder, breast cancer with tamoxifen-induced insomnia, and insomnia induced by detoxification from substance abuse. Unfortunately, a retrospective chart review (N = 43) found that quetiapine prescribed for insomnia at a mean dose of 120.3 ± 58.6 mg/day had adverse metabolic side effects (Cates et al., 2009). However, neither the RCT of quetiapine 25 mg/day study (Tassniyom et al., 2010) nor the open-label trial of quetiapine 25–75 mg/day (Wiegand et al., 2008) reported the rate of metabolic side effects, and the retrospective review did not perform subgroup analysis of the patients taking 25 mg/day (N = 4 at baseline).
6. Conclusion
A summary of the effects of quetiapine on sleep architecture is presented in Table 2. There is moderate evidence that quetiapine is not significantly effective for the treatment of primary insomnia, based on one small RCT (Tassniyom et al., 2010). Randomized studies support the usefulness of quetiapine in insomnia that is secondary to conditions for which quetiapine has an FDA-approved indication, like bipolar depression or unipolar depression (as augmentation). There is Level 1b evidence based on two RCTs (Calabrese et al., 2005; McElroy et al., 2010) that quetiapine is effective in the treatment of insomnia secondary to bipolar depression. There is Level 1b evidence based on one RCT (Garakani et al., 2008) and three open-label trials (Šagud et al., 2006; Todder et al., 2006; Baune et al., 2007) that quetiapine as augmentation of antidepressants is effective in reducing symptoms of insomnia in treatment-resistant depression.
IX. Discoveries, Novel Pathways, and Pipelines
The discoveries and pipelines in this section, constructed using data from a custom search of the Cortellis database, are an up-to-date (as of February 2017) snapshot of the current state of the research and development of insomnia medications.
A. Discoveries
1. Adenosine Receptor Agonist
YZG-331 is a promising sedative hypnotic and adenosine analog that exerts its effects by binding to the adenosine receptor. (See the Other Receptors section for a review of the pharmacology of A1A and A2A.)
2. Casein Kinase-1δ/ε
The casein kinase-1δ and casein kinase-1ε proteins are essential elements of the molecular oscillators known as circadian clocks (Lee et al., 2009). Their importance to the functioning of the mammalian circadian rhythm has spurred interest in casein kinase-1δ/ε inhibitors as potential clinical treatments of sleep disorders and other central nervous system disorders including neurodegenerative conditions (Perez et al., 2011).
3. Selective Melatonin MT2 Receptors
Studies on melatonin MT1 knockout, MT2 knockout, and double MT1-MT2 knockout mice have demonstrated that these two receptors have opposing or complementary functions. Whereas MT2 receptor activation promotes SWS, MT1 decreases SWS and increases REMS (Ochoa-Sanchez et al., 2011; Ochoa-Sanchez et al., 2014). This evidence prompted the development of novel selective MT2 agonists as hypnotics. The compound UCM765 has greater MT2 receptor affinity (pKi = 10.18) than melatonin (pKi = 9.59) and has about 100-fold higher affinity for the MT2 receptor than for the MT1 receptor (pKi = 8.28). UCM924 also displays MT2 affinity (pKi = 10.2) that is 300-fold higher than for MT1 (pKi = 6.75), with an intrinsic activity for MT1: IAr-hMT1 = 0.1; and for MT2: IAr-hMT2 = 0.4 (Rivara et al., 2009).
Both UCM765 (Ochoa-Sanchez et al., 2011) and UCM924 (Ochoa-Sanchez et al., 2014) increase SWS during the inactive phase of the day, without significant change in REMS or sleep architecture. The congener nonselective MT1-MT2 receptor UCM971 did not alter the 24-hour duration of wakefulness, NREMS, or REMS, but modified the number of episodes. MLT decreased (−37%) the latency to the first episode of NREMS and enhanced the power of NREMS delta band (+33%), but did not alter the duration of any of the three vigilance states or modify the duration of SWS (Ochoa-Sanchez et al., 2014). These data confirm the importance of targeting the MT2 receptor for hypnotic effects. UCM765 and UCM924 show a good safety profile and are currently under development for clinical studies.
4. Selective Orexin-2 Antagonist
Although dual orexin receptor antagonists like suvorexant are effective at promoting sleep, selective orexin-2 receptor blockers may preserve sleep architecture to a greater extent than dual antagonists (Bonaventure et al., 2015). Indeed, only orexin-2 but not orexin-1 is involved in the regulation of sleep (Dugovic et al., 2009). JNJ-42847922 is a novel orexin-2 antagonist shown to reduce the latency to NREM sleep and increase NREM sleep in the first 2 hours after administration, without affecting REM sleep in rats (Bonaventure et al., 2015). Importantly, the compound has been shown to reduce time to sleep onset and increase total sleep time after 7 days of chronic dosing (30 mg/kg). The compound did not produce conditioned-place preference or increase dopamine release in the nucleus accumbens, indicating that it lacks intrinsic motivational properties, in contrast to zolpidem. In a Phase I study in healthy human subjects, JNJ-42847922 (10–80 mg/day) significantly increased somnolence: 22 of 26 subjects (85%) receiving JNJ-42847922 reported somnolence as an adverse event, whereas only 3 of 13 subjects (23%) receiving placebo did (Bonaventure et al., 2015). The compound’s pharmacokinetic profile in humans was favorable, with a half-life of 2 hours. One subject reported experiencing sleep paralysis after receiving the 80 mg/day dose.
B. Pipelines
1. Lumateperone
Lumateperone is a mechanistically novel investigational antipsychotic drug with a unique pharmacological profile, showing very high 5-HT2A blocking activity (Ki = 0.54 nM) relative to its D2 modulating activity. The drug has a 60-fold difference between its affinity for 5HT2A and D2 receptors compared with a 12-fold difference for risperidone, a 12.4-fold difference for olanzapine, and a 0.18-fold difference for aripiprazole (Davis et al., 2015; Snyder et al., 2015). Lumateroperone functions as a modulator of the D2 receptor by partially agonizing presynaptic D2 receptors and antagonizing postsynaptic D2 receptors (Ki = 32 nM) (Snyder et al., 2015). Furthermore, the drug blocks the serotonin transporter with strong affinity (Ki = 62 nM) while having no affinity for the H1 histaminergic or muscarinic receptors (Snyder et al., 2015). Importantly, the drug’s D2 and SERT occupancy increase with dose (Davis et al., 2015); at low doses, the drug theoretically functions as a selective 5-HT2A blocker.
The company, Intra-Cellular Therapies (New York, NY), suggests on their website that lower doses of lumateperone could be useful in the treatment of sleep disorders, whereas higher doses are targeted to neuropsychiatric disease. A Phase 2 study (N = 18) of lumateperone as a treatment of insomnia characterized by sleep maintenance difficulties was discontinued early when the investigators found robust evidence of efficacy, with increased SWS and decreased wake after sleep time by polysomnography (Intra-Cellular Therapies, 2009). Furthermore, lumateperone did not impair next-day cognition as measured by Leeds Psychomotor Battery, Digit Symbol Substitution Test, or Word Pair Associates Test.
2. Piromelatine
Piromelatine is a unique drug that combines agonist activity at MT1 and MT2 with agonism at 5-HT1A/1D receptors (Laudon et al., 2012). Piromelatine was shown to have both hypnotic and antinociceptive effects by electroencephalogram (EEG) recordings and an animal model of neuropathic pain, partial sciatic nerve ligation (Liu et al., 2014). The drug was found to increase NREM sleep and decrease wakefulness in partial sciatic nerve ligation mice. Finally, the investigators demonstrated that the effect could be blocked by preadministration of a melatonin receptor antagonist, a 5-HT1A receptor antagonist, or an opiate receptor antagonist (Liu et al., 2014), implicating these receptors in the mechanism of action of the drug.
In 2013, Neurim Pharmaceuticals (Tel-Aviv, Israel) announced positive results from a phase II randomized clinical trial (N = 120) of piromelatine in primary insomnia (Neurim Pharmaceuticals, 2013). Active treatment with piromelatine 20 or 50 mg/day over 4 weeks resulted in significantly improved wake after sleep onset, sleep efficiency, and total sleep time. The Clinicaltrials.gov database lists a study currently recruiting patients entitled Safety and Efficacy of Piromelatine in Mild Alzheimer’s Disease Patients (ReCOGNITION); it also lists a completed study entitled The Effect of Neu-P11 on Symptoms in Patients with D-IBS. These studies indicate that Neurim Pharmaceuticals is exploring piromelatine’s potential efficacy in myriad conditions, including irritable bowel syndrome and Alzheimer’s disease.
X. Conclusions
In the last 20 years, preclinical and clinical research on sleep has expanded tremendously. The study of knockout mice for specific receptors has generated novel scientific knowledge of the unique role of each receptor in the regulation of sleep, the application of optogenetics to the study of sleep has elucidated new circuits, and the discovery of clock genes has generated insight into the cellular and molecular mechanisms that regulate sleep homeostasis.
In parallel, clinical studies have investigated how sleep architecture is differentially impaired in various neuropsychiatric diseases (including major depressive disorder, posttraumatic stress disorder, and Alzheimer disease) and the manner in which selective receptors' ligands can improve sleep quality and quantity.
Despite these advancements, BZDs continue to be widely prescribed, although their use, particularly in the elderly, is associated with an increased risk of falls, fractures, and emergency hospitalizations. Most or all BZDs and Z-drugs are available as generic drugs; as such, they are available to providers (private insurance companies, universal governmental health care systems) for a very low cost compared with innovative hypnotics (Tannenbaum et al., 2015; Goreveski et al., 2012). Innovative drugs cannot compete with the price of BZDs, discouraging academia and pharmaceutical research companies from investing in sleep medicine. Clinicians resort to prescribing drugs off-label, although these compounds often lack a strong evidence base for their use.
Substantial opportunity remains for pipelines that target the unmet needs in the insomnia market. Medicines with comparable efficacy and improved long-term safety would hold a competitive advantage over current first-line therapies, especially hypnotic without cognitive side-effects or not causing motor impairments the next day. Similarly, drugs that improve sleep quality by augmenting SWS would be viewed favorably by physicians. Moreover, drugs that selectively improve sleep in specific diseases would be also a new avenue for a personalized medicine.
Advancing drug discovery for insomnia and sleep disorders requires that the industry, in collaboration with regulatory authorities, clinical experts, and patient communities, engage in promoting and requiring better treatment of this condition. Moreover, there is a need to better define and agree on the nosology of sleep-related illnesses: the classifications of patient populations and the types of outcomes, other than sleep parameters, to be monitored among clients. Research into novel hypnotics may bolster the growing conception that sleep disorders are an integral part of, rather than secondary to, diseases such as depression and Alzheimer’s (Wafford and Ebert, 2008). Since the burden of insomnia and sleep disorders will likely increase in coming decades due to the aging of the population and the growing use of computer technologies (Fossum et al., 2014), research on novel hypnotics should be considered a priority.
Acknowledgments
We acknowledge Martina Dick for technical help in Fig. 1.
Authorship Contributions
Wrote or contributed to the writing of the manuscript: Atkin, Comai, Gobbi.
Footnotes
This work was supported by the Quebec Network on Suicide, Mood Disorders, and Related Disorders and the Canadian Depression Research and Intervention Network (CDRIN).
G.G. has received grants from the Canadian Institute of Health Research (CIHR), the Canada Foundation for Innovation (CFI), the Fonds de recherche du Québec – Santé (FRQS), and the Quebec Network on Suicide, Mood Disorders and Related Disorders. G.G. is an inventor and assignee in patents regarding the sleep properties of selective melatonin MT2 agonists.
Abbreviations
- AUC
- area under the curve
- BZDs
- benzodiazepines
- CHMP
- Committee for Medicinal Products for Human Use
- DSM
- Diagnostic and Statistical Manual of Mental Disorders
- EEG
- electroencephalogram
- 5-HT
- serotonin
- KO
- knockout
- LC
- locus coeruleus
- LDT/PPT
- lateralpontine tegmentum/pedunculopontine tegmental nuclei
- MDD
- major depressive disorder
- MNPO
- median preoptic nucleus
- NREM
- non-rapid eye movement sleep
- OX
- orexin
- PRM
- prolonged-release melatonin
- PSQI
- Pittsburgh Sleep Quality Index
- PTSD
- posttraumatic stress disorder
- RCT
- randomized-controlled trial
- REM
- rapid eye movement sleep
- RT
- reticular thalamus
- SCN
- suprachiasmatic nuclei
- SDL
- sublaterodorsal nuclei
- SSRI
- selective serotonin reuptake inhibitor
- SWS
- slow-wave sleep
- vlPO
- ventrolateral preoptic area
- WASO
- wake after sleep onset
- WT
- wild type.
- Copyright © 2018 by The American Society for Pharmacology and Experimental Therapeutics