Metabotropic glutamate receptor function and regulation of sleep-wake cycles

Abstract

Metabotropic glutamate (mGlu) receptors are the most abundant family of G-protein coupled receptors and are widely expressed throughout the central nervous system (CNS). Alterations in glutamate homeostasis, including dysregulations in mGlu receptor function, have been indicated as key contributors to multiple CNS disorders. Fluctuations in mGlu receptor expression and function also occur across diurnal sleep-wake cycles. Sleep disturbances including insomnia are frequently comorbid with neuropsychiatric, neurodevelopmental, and neurodegenerative conditions. These often precede behavioral symptoms and/or correlate with symptom severity and relapse. Chronic sleep disturbances may also be a consequence of primary symptom progression and can exacerbate neurodegeneration in disorders including Alzheimer’s disease (AD). Thus, there is a bidirectional relationship between sleep disturbances and CNS disorders; disrupted sleep may serve as both a cause and a consequence of the disorder. Importantly, comorbid sleep disturbances are rarely a direct target of primary pharmacological treatments for neuropsychiatric disorders even though improving sleep can positively impact other symptom clusters. This chapter details known roles of mGlu receptor subtypes in both sleep-wake regulation and CNS disorders focusing on schizophrenia, major depressive disorder, post-traumatic stress disorder, AD, and substance use disorder (cocaine and opioid). In this chapter, preclinical electrophysiological, genetic, and pharmacological studies are described, and, when possible, human genetic, imaging, and post-mortem studies are also discussed. In addition to reviewing the important relationships between sleep, mGlu receptors, and CNS disorders, this chapter highlights the development of selective mGlu receptor ligands that hold promise for improving both primary symptoms and sleep disturbances.

1. Introduction

The word “sleep” holds different meaning and emotional weight for different individuals. While many people undergo an easy transition from waking to a peaceful state of altered consciousness, initiating and maintaining sleep is a fitful chore for others. Though there have been substantial advancements in understanding the underlying functions of homeostatic sleep, including its role in synaptic remodeling critical for memory consolidation (Manoach & Stickgold, 2019; Martin, Monroe, & Diering, 2019), many purposes remain unresolved. Although sleeping and waking states may be perceived as dichotomous processes, they are intimately intertwined. Acute sleep disruptions lead to decreased daytime arousal and cognitive impairments as well as increased fatigue or daytime somnolence. Further, chronic sleep disruptions including insomnia are often comorbid with neuropsychiatric, neurodevelopmental, and neurodegenerative disorders (Benca & Buysse, 2018; Benca, William, Thisted, & Gillin, 1992; Riemann, Krone, Wulff, & Nissen, 2020; Sprecher, Ferrarelli, & Benca, 2015; Wulff, Gatti, Wettstein, & Foster, 2010). Increasing evidence suggests a direct relationship between sleep disruptions and symptom onset and/or severity in central nervous system (CNS) disorders (Reeve, Sheaves, & Freeman, 2015; Riemann et al., 2020; Veatch et al., 2017; Wulff et al., 2010). Importantly, the Diagnostics and Statistic Manual, 5th Edition (DSM-5) delineated insomnia, defined by difficulty falling asleep and/or staying asleep, as a separate diagnosis from other mental disorders as opposed to a primary or secondary symptom, and, thus, sleep disturbances should be a primary target of pharmacotherapies (APA, 2013; Benca & Buysse, 2018).

Sleep is an intricately regulated process involving multiple circuits and neurotransmitter systems. These systems work in concert to actively inhibit arousal-related processes and shift from asynchronous electrical activity during wake to synchronous, slow oscillatory activity during sleep (Scammell, Arrigoni, & Lipton, 2017). Seminal studies employing region-specific brain lesions and genetic manipulations in murine models as well as pharmacological studies have identified roles of noradrenergic, dopaminergic, cholinergic, histaminergic, serotonergic and orexinergic systems in sleep and arousal (for review, see España & Scammell, 2011; Weber & Dan, 2016). More recent developments suggest that the glutamatergic system is also extensively involved in sleep-wake processes ( Jones, 2020; Martin et al., 2019; Saper & Fuller, 2017; Tononi & Cirelli, 2014). Although a critical regulator of several neurobiological processes including memory consolidation, the role of glutamate in sleep-specific processes is not as rigorously understood. With lines of evidence supporting a role of glutamatergic dysregulation in both CNS disorders and sleep impairments, examining the relationship between glutamate, sleep and CNS disorders is timely.

This chapter highlights research examining metabotropic glutamate (mGlu) receptor involvement in the regulation of sleep alone and in the context of CNS disorders. We focus on schizophrenia, major depressive disorder (MDD), post-traumatic stress disorder (PTSD), Alzheimer’s disease (AD), and substance use disorder (SUD). It is important to note that these disorders are selected examples and that sleep disturbances and disrupted mGlu function are associated with many other CNS disorders. In this chapter, we review preclinical electrophysiological, genetic, and pharmacological studies contributing to our understanding of how various mGlu receptors impact sleep architecture and duration. When possible, we also discuss related human genetic, imaging, and post-mortem studies. Ultimately, we work to shed light on the underappreciated contribution of the glutamate system to sleep and, thereby, comorbid sleep disruptions in CNS disorders. As remediation of sleep disruptions should be a complementary focus of drug development efforts, this chapter highlights novel pharmacotherapeutic approaches targeting mGlu receptors that may effectively address both sleep disturbances and primary symptoms of the neuropsychiatric disorder.

2. Overview of metabotropic glutamate receptors (mGlu receptors)

Glutamate, the major excitatory neurotransmitter in the CNS, regulates rapid excitatory neurotransmission via ionotropic glutamate (iGlu) receptors and slower neuromodulatory effects via mGlu receptors. iGlu receptors are ligand-gated ion channels subdivided into four functional classes: α-amino-3-hydroxy-5-methyl-iso-xazolepropionic acid (AMPA) receptors, N-methyl-d-aspartate (NMDA) receptors, kainate receptors, and GluD receptors (Hansen et al., 2021). mGlu receptors are G-protein coupled receptors (GPCRs) and are the most abundant receptor family in the CNS with 8 distinct subtypes (mGlu₁-mGlu₈). These class C GPCRs are widely distributed, located on both neurons and glial cells influencing glutamatergic, GABAergic, and other neuromodulatory synaptic transmission (Luessen & Conn, 2022; Maksymetz, Moran, & Conn, 2017; Niswender & Conn, 2010). mGlu receptors are subdivided into three groups (Group I: mGlu₁ and mGlu₅; Group II: mGlu₂ and mGlu₃; Group III: mGlu₄, mGlu₆, mGlu₇, and mGlu₈) based on sequence homology, G-protein coupling, and agonist selectivity (Conn & Pin, 1997; Niswender & Conn, 2010; Wang & Zhuo, 2012). Acknowledging the extensive distribution and neuromodulatory role of mGlu receptors throughout the CNS, it is not surprising that preclinical and clinical studies suggest altered mGlu receptor expression and function contributes to many neuropsychiatric, neurodevelopmental, and neurodegenerative disorders. In the following section, we provide a brief overview of the general functional and behavioral roles of the different mGlu receptor subtypes.

2.1. Group I mGlu receptor distribution and function

Group I mGlu receptors (mGlu_1/5) are widely distributed throughout the CNS in brain regions particularly relevant for learning, memory and motivation. mGlu₅ receptors are densely present in the cortex, striatum, and hippocampus whereas mGlu₁ receptors are found primarily in the cerebellum, thalamus, and hypothalamus (Niswender & Conn, 2010; Olive, 2010; Romano et al., 1995; Shigemoto et al., 1993; Shigemoto, Nakanishi, & Mizuno, 1992). These predominantly postsynaptic receptors are G_q/11-coupled and localized on glutamatergic and GABAergic terminals (Luessen & Conn, 2022; Niswender & Conn, 2010). Activation of G_q/11-coupled receptors canonically stimulates phospholipase C (PLC) and phosphoinositide hydrolysis leading to intracellular calcium mobilization. However, mGlu_1/5 receptors also activate several additional pathways including downstream effectors in the mitogen-activate protein kinase/extracellular protein kinase (MAPK/ERK) pathway (Luessen & Conn, 2022; Niswender & Conn, 2010; Page et al., 2006). Furthermore, mGlu_1/5 receptors are critical regulators of activity-dependent synaptic plasticity, influencing both long-term potentiation (LTP) and long-term depression (LTD), and this is in part attributed to their tight coupling with NMDARs (Anwyl, 1999; Joffe, Centanni, Jaramillo, Winder, & Conn, 2018; Lüscher & Huber, 2010; Lutzu & Castillo, 2021; Lu et al., 1997). As these functions are outside the scope of this chapter, readers are referred to several excellent reviews (Luessen & Conn, 2022; Niswender & Conn, 2010; Olive, 2010; Shigemoto et al., 1993). Given the role of group I mGlu receptors in regulating synaptic plasticity and, thereby, learning and memory (for review, see Olive, 2010), it is not surprising that they have been implicated in neurodegenerative and neuropsychiatric disorders associated with cognitive disruptions including AD (Bruno et al., 2000; Kumar, Dhull, & Mishra, 2015), schizophrenia (Kinney et al., 2003; Maksymetz et al., 2017; Nicoletti et al., 2019), Parkinson’s disease (Morin, Grégoire, Gomez-Mancilla, Gasparini, & di Paolo, 2010), MDD (Lindemann et al., 2015), and SUD (Gould et al., 2016; McGeehan & Olive, 2003; Veeneman et al., 2011). As will be described below, dysregulation of either mGlu₁ or mGlu₅ receptors also affects sleep which may further contribute to primary symptoms associated with these disorders (Aguilar, Strecker, Basheer, & Mcnally, 2020; Cavas, Scesa, & Navarro, 2013b; Holter et al., 2021).

2.2. Group II mGlu receptor distribution and function

Group II mGlu receptors (mGlu₂ and mGlu₃) are localized on both presynaptic and postsynaptic membranes and, similar to group I, are abundantly present in regions important for learning, memory, and motivation including the prefrontal cortex (PFC), hippocampus, striatum, and amygdala (Crupi, Impellizzeri, & Cuzzocrea, 2019; Luessen & Conn, 2022; Maksymetz et al., 2017). mGlu₃ receptors are also expressed on astrocytes, and this contributes to anti-inflammatory properties that may be neuroprotective (Bruno et al., 1998; Nicoletti et al., 2011). Group II mGlu receptors are G_i/o coupled and inhibit adenylyl cyclase and phosphatidyl inositol 3-kinase which reduces intracellular calcium mobilization. These receptors also inhibit MAPK/ERK pathways (Luessen & Conn, 2022; Nicoletti et al., 2011; Niswender & Conn, 2010). Importantly, the primary function of presynaptic mGlu_2/3 receptors is to work as autoreceptors and maintain glutamate homeostasis via inhibition of release at glutamatergic and GABAergic synapses following periods of glutamate efflux (Conn & Jones, 2009; Luessen & Conn, 2022; Maksymetz et al., 2017; Mazzitelli, Palazzo, Maione, & Neugebauer, 2018; Moussawi & Kalivas, 2010; Wright, Arnold, Wheeler, Ornstein, & Schoepp, 2001). In general, group II mGlu receptors contribute to synaptic plasticity by reducing synaptic excitability; receptor activation leads to decreased EPSC amplitude and induction of postsynaptic hippocampal LTD (Altinbilek & Manahan-Vaughan, 2009; Grueter & Winder, 2005; Joffe et al., 2020; Walker et al., 2015; Yokoi et al., 1996). Behaviorally, mGlu_2/3 receptors have been shown to influence spatial learning and working memory (de Filippis et al., 2015; Lyon et al., 2011). Findings indicate pharmacological activation of mGlu_2/3 receptors can improve attention and working memory in acute pharmacological models of cognitive disruptions (e.g. NMDAR antagonists) in both humans and animals (Greco, Invernizzi, & Carli, 2005; Griebel et al., 2016; Krystal et al., 2005; Moghaddam & Adams, 1998). Activation may also produce anxiolytic- and antidepressant-like effects in rodent models (Dogra & Conn, 2021). As such, group II mGlu receptors may be promising pharmacotherapeutic targets for multiple disorders including schizophrenia (Hackler et al., 2010; Hiyoshi, Hikichi, Karasawa, & Chaki, 2014; Maksymetz et al., 2017; Moghaddam & Adams, 1998; Sokolenko, Hudson, Nithianantharajah, & Jones, 2019), AD (Lee et al., 2004; Richards et al., 2010), MDD (Chaki, 2017), pain (Davidson et al., 2017; Jones, Eberle, Peters, Monn, & Shannon, 2005; Mazzitelli et al., 2018), and SUD (Cleva & Olive, 2012; Hao, Martin-Fardon, & Weiss, 2010; Rodd et al., 2006).

2.3. Group III distribution and function

Group III mGlu receptors (mGlu₄, mGlu₆, mGlu₇, mGlu₈) are localized pre- and post-synaptically on glutamatergic and GABAergic terminals with differential expression throughout the CNS (Luessen & Conn, 2022; Niswender & Conn, 2010). mGlu₄ and mGlu₈ receptors are located presynaptically with regionally-restricted expression in the cerebellum and hippocampus (Luessen & Conn, 2022; Niswender & Conn, 2010a; Zhai et al., 2002). mGlu₆ receptors, unlike others, have strict localization to the retina, particularly on ON bipolar cells, and, thus, are not relevant for this chapter (Crupi et al., 2019; Luessen & Conn, 2022; Niswender & Conn, 2010). Lastly, mGlu₇ receptors are presynaptic and widely distributed throughout the CNS with expression in regions including the hippocampus, thalamus, hypothalamus and amygdala. mGlu₇ is unique in that it has a very low affinity for glutamate and, thus, it serves to only function in times of high synaptic activity to fine-tune and prevent excessive glutamate levels (Niswender & Conn, 2010). In general, group III mGlu receptors contribute to spontaneous glutamate transmission and hippocampal-mediated long-term synaptic plasticity (Altinbilek & Manahan-Vaughan, 2007). For example, mGlu₄ knockout mice had enhanced hippocampal LTP, though the PFC was unaffected (Iscru et al., 2013). Furthermore, bath application of the mGlu₄ PAM foliglurax to corticostriatal slices reduced glutamatergic transmission, as shown through a reduction in spontaneous EPSC frequency (Calabrese et al., 2022). Pharmacological activation of mGlu₇ induced LTD in mossy fiber inputs to stratum lucidum interneurons (SLINs) (Pelkey, Lavezzari, Racca, Roche, & McBain, 2005). In Schaffer collateral-CA1 synapses, activation of mGlu₇ potentiated a submaximal level of LTP and inhibition of mGlu₇ prevented the induction of LTP in these synapses (Kalinichev et al., 2013; Klar et al., 2015). Lastly, activation of mGlu₈ reduced field EPSPs induced by lateral perforant path (LPP) afferents in hippocampal slices (Zhai et al., 2002). In line with their regions of expression and preclinical findings, mGlu₄ receptors are being explored for their pharmacotherapeutic potential in Parkinson’s disease and schizophrenia (Battaglia et al., 2006; Calabrese et al., 2022; Charvin, 2018; Luessen & Conn, 2022; Niswender & Conn, 2010). The mGlu₇ receptor has been a suggested target for disorders including AD (Crupi et al., 2019), neuropathic and inflammatory pain (Marabese et al., 2007; Palazzo, Fu, Ji, Maione, & Neugebauer, 2008), SUD (Li, Xi, & Markou, 2013), MDD and anxiety disorders (Bradley et al., 2012; O’Connor et al., 2013; Palucha & Pilc, 2007; Pałucha-Poniewiera & Pilc, 2013), and epilepsy (Girard et al., 2019; Sansig et al., 2001). Lastly, activation of mGlu₈ may be a promising therapeutic approach for anxiety disorders (Duvoisin et al., 2010; Linden et al., 2002; Niswender & Conn, 2010) and additional research supports mGlu₈ as a target for pain (Crupi et al., 2019; Marabese et al., 2007; Palazzo et al., 2008).

3. Sleep

Sleep is defined as an easily reversible state of reduced consciousness. However, from a neurobiological perspective, sleep is not passive. Rather, sleep is actively regulated by inhibition of arousal circuits which contrast the cortical activation and arousal that occurs during wake. The two-process model of homeostatic sleep regulation posits that sleep is driven by an interaction of sleep debt(Process S) with the circadian pacemaker (Process C). Process S increases during extended waking periods and decreases following periods of sleep, and the circadian pacemaker follows a similar oscillatory pattern, regulating sleep to specific times of day (Borbély, Daan, Wirz-Justice, & Deboer, 2016; Borbély & Wirz-Justice, 1982; Daan, Beersma, & Borbély, 1984). Circadian factors influencing sleep regulation are outside the scope of this chapter but are thoroughly detailed in both Logan and McClung (2019) and Rosenwasser and Turek (2015). There are many neuroanatomical and neurochemical systems that contribute to sleep-wake regulation, and we refer the reader to many excellent reviews (Brown, Basheer, McKenna, Strecker, & McCarley, 2012; España & Scammell, 2011; Saper, Fuller, Pedersen, Lu, & Scammell, 2010; Saper, Scammell, & Lu, 2005). The following sections provide a brief overview of human and nonhuman sleep and the neurochemical systems regulating sleep-wake cycles with a specific focus on glutamatergic involvement.

3.1. Sleep stages and characteristics

On average, healthy adults sleep 6–8h and undergo 4–5 sleep cycles per night (Hor & Tafti, 2009). Each cycle is comprised of non-rapid eye movement (NREM) and rapid eye movement (REM) sleep and lasts approximately 90min in duration (le Bon, Lanquart, Hein, & Loas, 2019). In humans, NREM sleep is divided into three stages, termed N1, N2, and N3. The American Academy of Sleep Medicine (AASM) defines N1 as the lightest stage of sleep, representing a transition between wake and sleep when external stimuli can still be processed (Iber, Ancoli-Israel, & Chesson, 2007). N2 sleep is classified as light sleep and N3 is considered deep sleep, or slow wave sleep (SWS). Overall, in one night of sleep, the percentage of time in N1, N2, and N3 is ~5%, 50% and 20% with the remaining 25% being REM sleep (Iber et al., 2007). As a result of increased sleep debt throughout the day, N3 stage sleep (and duration of bouts) is generally more prevalent in the first half of the night whereas duration of REM bouts are generally longer in the second half of the night. However, the actual duration and percentage of time in each stage changes across development and normal aging in healthy populations (Ohayon, Carskadon, Guilleminault, & Vitiello, 2004).

Sleep stages and transitions are identified using electroencephalography (EEG) to measure and determine changes in brain oscillatory activity. The transition from wake to N1 sleep is associated with a shift from high/mixed frequency, lower amplitude EEG patterns to predominately lower frequency, higher amplitude oscillations. This includes a decrease in alpha band [~8–13Hz] activity compared to waking states. N2 sleep is associated with sleep spindles (periodic synchronous burst firing in the sigma [~12–15Hz] range) and K-complexes (brief waveforms characterized by high-amplitude spike and rebound patterns). N3 sleep, or SWS, is associated with high amplitude slow wave activity (SWA, delta band, [~0–4Hz]). Lastly, REM sleep is associated with an increase in theta [~4–8Hz] activity relative to delta activity as well as an increase in beta [18–30Hz] and gamma [>30Hz] frequencies. Historically, REM sleep was termed paradoxical sleep as the EEG waveforms paradoxically resemble EEG activity during waking states (Boissard et al., 2002; España & Scammell, 2011).

Humans and nonhuman primates are monophasic sleepers whereas rodents are polyphasic sleepers, rapidly cycling through sleep-wake states over a 24-h period. Despite these differences, from a neurophysiological perspective, mechanisms regulating sleep-wake cycles are evolutionarily well-conserved. Rodents are nocturnal and spend roughly 80% of their time sleeping during the light (inactive) phase and 20% of their time sleeping during the dark (active) phase when under a typical 12 h:12h light:dark cycle. Due to the more rapid sleep cycle (~8–15min/cycle), sleep is typically classified into REM and NREM sleep. Although EEG activity in rodent sleep stages are comparable to humans, electromyography (EMG; muscle activity) is often used in rodents to accurately distinguish REM sleep, which demonstrates a distinct muscle atonia, from waking periods. Research in animals using lesions and genetic and pharmacological manipulations has provided extensive insight into the mechanisms regulating sleep-wake transitions. These studies have also elucidated roles of different sleep variables in maintaining healthy daily function and how irregularities may contribute to various CNS disorders.

3.2. Glutamate and the sleep-wake cycle

Historically, the ascending reticular activation system (ARAS) was solely implicated in regulating waking and arousal. Brainstem nuclei, that were primarily thought to be monoaminergic (dopamine, norepinephrine, serotonin, histamine) and cholinergic, innervate thalamic projections to the cortex. Additional projections to the hypothalamus and basal forebrain innervate orexin/hypocretin neurons and forebrain cholinergic neurons, respectively, which then further project to the cortex (Saper & Fuller, 2017). Early seminal studies by Moruzzi and Magoun (1949) demonstrated that inactivation of the reticular formation and ARAS produced a passive state resembling sleep. However, more recent research demonstrated that lesions to different components of the ARAS had a minimal, if any, effect on sleep-wake durations in rats, suggesting there must be additional contributors (for detailed review see Saper & Fuller, 2017). Glutamatergic projections from the brainstem parabrachial and pedunculopontine nuclei to the basal forebrain as well as projections from the hypothalamic sup-ramammillary area to the cortex have since been shown to contribute substantially to arousal and wake-promoting effects (see Saper & Fuller, 2017). Thus, multiple neurons, including glutamatergic neurons, in the brainstem, hypothalamus and basal forebrain regulate cortical arousal and waking.

Sleep promotion occurs via GABAergic inhibition of each of these wake-promoting nuclei. GABA neurons within the ventrolateral (VLPO) and median preoptic (MNPO) nuclei of the hypothalamus innervate all wake-regulating monoaminergic and cholinergic brainstem nuclei (Sherin, Elmquist, Torrealba, & Saper, 1998; Suntsova, Szymusiak, Alam, Guzman-Marin, & McGinty, 2002). Additional GABA neurons in the brainstem reticular formation inhibit glutamate neurons in the parabrachial nucleus (Anaclet et al., 2014). These GABA neurons are almost entirely inactive during waking periods, and NREM sleep is initiated when these neurons are activated to inhibit cortical activity. This shift in inhibitory balance from wake to sleep is in part driven by several metabolic and circadian influences including adenosine, prostaglandins, cytokines, and growth hormone-releasing hormone (Obal & Krueger, 2003). For example, extracellular concentrations of adenosine increase following prolonged periods of waking as active neurons hydrolyze ATP, and concentrations dissipate following sleep (Porkka-Heiskanen et al., 1997). Importantly, adenosine acts as a neuromodulator to inhibit arousal circuits and disinhibit VLPO GABA neurons thereby exerting sleep-promoting effects (Obal & Krueger, 2003).

There are multiple neuronal populations that contribute specifically to REM sleep. These include subsets of cholinergic neurons in the laterodorsal and pedunculopontine tegmentum (LTD/PPT) as well as the basal forebrain which promote cortical activation through acetylcholine release (Marrosu et al., 1995; Williams, Comisarow, Day, Fibiger, & Reiner, 1994). Additionally, neurons containing melanin-concentrating hormone and GABA neurons in the hypothalamus are active during REM sleep demonstrating inhibitory effects on wake-promoting nuclei (Verret et al., 2003). Lastly, a subset of glutamate neurons active in the sublaterodorsal (SLD) nucleus that project to the inhibitory neurons in the medulla and spinal cord are thought to contribute to muscle atonia during REM sleep (Boissard et al., 2002).

The distinct patterns of EEG activity that occur throughout sleep-wake cycles are derived from interactions between the subcortical systems described above (brainstem, basal forebrain, and hypothalamus) and the thalamus and cortex. Activity in the thalamocortical (TC) circuit also undergoes dynamic changes across sleep-wake cycles. The TC circuit is comprised of two groups of neurons. TC projection neurons are predominantly glutamatergic and relay sensory and motor information to the cortex. Thalamic relay neurons (TRNs) are predominantly GABAergic and are innervated by and can inhibit TC neurons (España & Scammell, 2011; Huguenard & McCormick, 2007). During wake and REM sleep, monoaminergic and cholinergic neurotransmitter release depolarizes thalamic neurons resulting in desynchronized fast, low-amplitude cortical oscillations and increasedsensitivity to incoming stimuli (Aston-Jones, Smith, Moorman, & Richardson, 2009; Hu, Steriade, & Deschênes, 1989). In contrast, the firing patterns shift during NREM sleep as these neurons become hyperpolarized and undergo synchronized burst firing, reducing responsivity to external stimuli (Brown et al., 2012; España & Scammell, 2011; Livingstone & Hubel, 1981; Llinás & Steriade, 2006; Mccormick & Bal, 1997; Steriade, Iosif, & Apostol, 1968). These reciprocally acting projections in part drive cortical delta waves and sleep spindles during NREM sleep (España & Scammell, 2011; Steriade, Domich, Oakson, & Deschfines, 1987). Additional inhibitory feedback loops between cortical glutamate neurons and thalamic GABA neurons further contribute to cortical oscillatory rhythms and sleep spindle generation (Steriade et al., 1987). Lastly, Yu et al. (2019) recently found that activation of glutamate neurons in the VTA projecting to both the nucleus accumbens (NAc) and the lateral hypothalamus increased time awake, whereas inhibition decreased waking bout duration and increased frequency of wake: NREM sleep transitions. In short, there are many overlapping circuits and neurotransmitter systems involved in the transition and maintenance of sleep-wake cycles and the associated oscillatory functions, and the glutamate system is a strong contributor.

Preclinical and clinical studies support a strong relationship between extracellular glutamate and the different sleep stages. EEG studies paired with simultaneous in vivo microdialysis recordings in the orbitofrontal cortex found extracellular glutamate concentrations to be highest during REM sleep, modest during waking periods and lowest during NREM sleep in rats (Lopez-Rodriguez, Medina-Ceja, Wilson, Jhung, & Morales-Villagran, 2007). Amperometric detection of glutamate in the cortex showed rapid increases in glutamate at the beginning of wake and REM sleep and rapid decreases during NREM sleep (Dash, Douglas, Vyazovskiy, Cirelli, & Tononi, 2009). Additional microdialysis studies revealed similar elevations in glutamate levels during wake in the pontine reticular formation (Watson, Lydic, & Baghdoyan, 2011), though glutamate concentrations detected in the thalamus were contrastingly highest during NREM sleep (Kékesi, Dobolyi, Salfay, Nyitrai, & Juhász, 1997). These changes in glutamate concentrations likely align with the shift from asynchronous cortical excitation during wake to synchronous TC regulation during NREM (España & Scammell, 2011). Additionally, evidence from rodent amperometry studies and human studies using proton magnetic resonance spectroscopy (¹H-MRS) indicates that glutamate levels and receptor availability are not only state-dependent but also sensitive to additional factors including general sleep history and disruptions in sleep homeostasis that occur following periods of extended waking (Dash et al., 2009; Weigend et al., 2019).

Lastly, the glutamate system has a major role in regulating state-dependent neurobiological changes in synaptic plasticity. The sleep homeostasis hypothesis suggests that a key role of sleep is to maintain a balance in synaptic strength. Notably, the number of synapses undergoing LTP and LTD shift between sleep and wake (Tononi & Cirelli, 2014). During time awake, as daily learning occurs, synapses progressively strengthen and undergo LTP-like potentiation. This is followed by synaptic weakening and LTD during sleep (Tononi & Cirelli, 2014; Vyazovskiy, Cirelli, Pfister-Genskow, Faraguna, & Tononi, 2008). This synaptic weakening functions predominantly to prevent saturation of synapses, which would result in an inability to form new connections; this synaptic weakening is also an important process in memory consolidation (Diering et al., 2017; Martin et al., 2019; Tononi & Cirelli, 2014). Supporting evidence for the morphological changes that occur in sleep and wake has emerged from studies using Drosophila. During time awake, an increase in synapse number as well as in scaffold and post-synaptic proteins Bruchpilot (BRP) and Discs-large (DLG), both key regulators of glutamate release, were reported. These changes were exacerbated in periods of extended wake as a result of sleep deprivation and reduced during periods of sleep (Bushey, Tononi, & Cirelli, 2011; Gilestro, Tononi, & Cirelli, 2009). Importantly, in rodents, both iGlu (GluA1-containing AMPARs) and mGlu receptors, (primarily group I mGlu receptors, detailed below), have been found to orchestrate these synaptic changes (Diering et al., 2017; Vyazovskiy et al., 2008). Briefly, synaptic expression of cortical and hippocampal GluA1-containing AMPARs is highest during wake and lowest during NREM sleep, which corresponds with electrophysiological findings of synaptic potentiation (Vyazovskiy et al., 2008). Cognitive impairments, a primary symptom of many CNS disorders, may be a direct result of aberrant regulation of these synaptic plasticity changes.

3.3. mGlu receptors and sleep

Many studies spanning from Drosophila to humans highlight important contributions of the different mGlu receptor subtypes in sleep-wake regulation. On the most basic level, genetic knockdown of the only mGlu receptor in Drosophila (DmGluRA) produced drastic changes in sleep-wake behavior, notably increasing time spent asleep during the day and decreasing time spent asleep at night (Ly, Lee, Strus, Prober, & Naidoo, 2020). Additional support for mGlu involvement in sleep regulation is derived from EEG studies in mice with a constitutive knockout of a single mGlu receptor subtype or altered function of a downstream signaling protein (see Table 1). Recent development of pharmacological tool compounds with mGlu receptor subtype specificity has also added to this literature base (see Table 2). The following section describes literature examining genetic and pharmacological manipulations in rodents, a majority of which involves the group I and II mGlu receptor classes. Although less well characterized, some evidence supports the role of group III mGlu receptors in sleep-wake regulation. Lastly, when available, relevant human postmortem, GWAS, and neuroimaging studies are also included.

Table 1.

Effects of subtype-specific mGlu receptor genetic alterations on sleep in rodents.

		Awake	Sleep	NREM				REM
Condition	Phase	Duration	Total duration	Duration	Bouts	Bout duration	Delta	Duration	Bouts	Bout duration	Theta
*PLCβ4−/− 1,2	Light	↓	—	n.s., ↑	n.s.	↑	↑	n.s.	↓	↑	↓
	Dark	↓	—	n.s., ↑	↑	↓	↑	↑	n.s.	↑	↓
TC-restricted PLCβ4−/− 1	Light	n.s.	—	↑	—	—	↑	n.s.	—	—	↓
	Dark	n.s.	—	n.s.	—	—	n.s.	↓	—	—	n.s.
*mGlu1 Mutation3	24 hr	—	↓	↓	—	—	—	n.s.	—	—	—
mGlu2/34	Light	—	↑	—	—	—	—	—	—	—	—
	Dark	—	n.s.	—	—	—	—	—	—	—	—
mGlu55,6,7	Light	n.s., ↓	—	n.s.	n.s.,↑	n.s.	n.s.,↓	↓	n.s.	n.s., ↓	↓
	Dark	n.s., ↑	—	n.s.	n.s.,↑	n.s., ↓	↓	n.s.	n.s.	n.s., ↓	↓
	24 hr	↓	—	↑	—	—	—	n.s.	—	—	—
mGlu78	Light	n.s.	—	n.s.	—	—	n.s.	n.s.	n.s.	↓	n.s.
	Dark	↑	—	↓	—	—	n.s.	n.s.	—	—	n.s.

^*PLCβ4^−/− is a downstream effector in the mGlu₁ pathway; mGlu₁ mutation is loss of function.

n.s., not significant; —, not determined.

¹Hong et al. (2016), ²Ikeda et al. (2009), ³Shi et al. (2021), ⁴Pritchett et al. (2012), ⁵Aguilar et al. (2020), ⁶Ahnaou, Raeymaekers, Steckler and Drinkenbrug (2015), ⁷Holst et al. (2017), ⁸Fisher et al. (2020).

Table 2.

Pharmacological studies examining subtype-specific mGlu receptor compounds on sleep in rodents.

	Awake	NREM				REM
Compound	Duration	Duration	Latency	Bouts	Bout duration	Duration	Latency	Bouts	Bout duration
mGlu5 receptor NAMS (within 4–6h post-dosing)
MPEP1,2 (1–20mg/kg, IP)	↑	↑	n.s.	↓	↑	↓	↑	↓	↓
MTEP1 (1–10 mg/kg, IP)	n.s.	↑	n.s.	↓	↑	↓	↑	n.s.	↓
GRN-5293 (0.32–1mg/kg, SC)	↑	↓	—	—	—	↓	—	—	—
Mavoglurant3 (1–3.2mg/kg, SC)	↑	↓	—	—	—	↓	—	—	—
Basimglurant4 (0.03–0.3mg/kg, PO)	↑	↓	↑	—	—	↓	↑	—	—
VU04242385 (1–30mg/kg, IP)	↑	↓	n.s.	—	—	↓	↑	—	—
M-5MPEP (partial NAM)5 (18–56.6 mg/kg, IP)	n.s.	n.s.	n.s.	—	—	↓	↑	—	—
mGlu5 Receptor PAMS (within 4–6h post-dosing)
LSN28146171,6 (2.5–10mg/kg, PO)	↑	↓	↑	↓	↓	↓	↑	↓	↓
LSN24633596 (0.3–3mg/kg, PO)	↑	↓	↑	—	—	↓	—	—	—
ADX472731,6 (10–300 mg/kg, PO)	↑	↓	↑	n.s.	↓	↓	↑	↓	n.s.
CDPPB6,7 (30mg/kg, IP)	↑ ; n.s.	↓	n.s.	—	—	↓	—	—	—
mGlu2/3 Receptor Antagonists/NAMS (within 4–6h post-dosing)
LY341495 (antagonist)8,9 (1–10mg/kg, SC)	↑	↓	↑	—	—	↓	↑	—	—
LY3020371 (antagonist)10 (3–10mg/kg, IP)	↑	↓	—	—	—	↓	—	—	—
Ro4491533 (NAM)8 (2.5–40 mg/kg, PO)	↑	↓	↑	—	—	↓	↑	—	—
mGlu2/3 Receptor Agonists/PAMS (within 4–6h post-dosing)
LY379268 (agonist)10,11 (10 mg/kg, PO10; 0.25–1 mg/kg SC11)	n.s. (dark) ↑(light)	n.s. (dark) ↓(light)	—	—	—	↓	—	—	—
LY354740 (agonist)12 (1–10mg/kg, SC)	n.s.	n.s.	n.s.	n.s.	n.s.	↓	↑	↓	↓
BINA (PAM)12 (1–40mg/kg, SC)	n.s.	n.s.	—	↑	↓	↓	↑	↓	↓
JNJ-42153605 (PAM)13 (3 mg/kg, PO)	n.s.	n.s.	—	—	—	↓	n.s.	—	—
JNJ-40411813 (PAM)14 (3–30mg/kg, PO)	↓	↑	—	—	—	↓	↑	—	—
JNJ-40068782 (PAM)15 (3–30mg/kg, PO)	n.s.	n.s.	—	—	—	↓	↑	—	—
THIIC (PAM; 12h post-dosing)16 (10–30mg/kg, PO)	—	↑	—	↑	—	↓	—	—	—
mGlu2/3 receptor agonists/PAMS (within 4–6h post-dosing)
(S)-3,4-dicarboxyphenylglycine17 (5–20mg/kg, IP)	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.

n.s., not significant; —, not determined; IP, intraperitoneal; SC, subcutaneous; PO, per os (oral).

¹Ahnaou, Langlois, Steckler, Bartolome-Nebreda and Drinkenburg (2015), ²Cavas et al. (2013b), ³Harvey et al. (2013), ⁴Lindemann et al. (2015), ⁵Holter et al. (2021), ⁶Gilmour et al. (2013), ⁷Parmentier-Batteur et al. (2012), ⁸Ahnaou, ver Donck, and Drinkenburg (2014), ⁹Feinberg, Schoepp, Hsieh, Darchia, and Campbell (2005), ¹⁰Wood et al. (2018), ¹¹Feinberg, Campbell, Schoepp, and Anderson (2002), ¹²Ahnaou et al. (2009), ¹³Cid et al. (2012), ¹⁴Lavreysen et al. (2015), ¹⁵Lavreysen et al. (2013), ¹⁶Fell et al. (2011), ¹⁷Cavas, Scesa, Martín-López, and Navarro (2017).

3.3.1. Group I mGlu receptors

Perhaps one of the most relevant contributions of mGlu_1/5 receptors is to maintain homeostasis in synaptic size and strength throughout the sleep-wake cycle via dynamic changes in the mGlu_1/5/Homer protein complex. Homer family proteins (Homer1, Homer2, and Homer3) are adaptor proteins that each contain several isoforms as a product of alternative splicing (Shiraishi-Yamaguchi & Furuichi, 2007). Homer protein variants are differentially recruited to the synapse during sleep and wake and, thereby, influence downstream effector systems of mGlu_1/5. During wake, mGlu_1/5 receptors are anchored to the excitatory post synaptic density via long-form Homer proteins, which contain both an Ena/VASP (EVP) domain and a coiled-coil domain. These long-form Homer proteins link mGlu_1/5 receptors to their primary signaling effector, the inositol triphosphate (IP3) receptor, in the endoplasmic reticulum as well as Shank1-3 proteins, facilitating cell excitability and increased GluA1-containing AMPAR levels during arousal states (Diering et al., 2017; Martin et al., 2019). Following prolonged waking periods, the activity-dependent immediate early gene Homer1a, a short-form protein containing only an EVP domain, is targeted to the synapse (Martin et al., 2019; Shiraishi-Yamaguchi & Furuichi, 2007). Homer1a uncouples the mGlu_1/5 receptor from the IP3 receptor, resulting in signaling through alternate pathways and reduced synaptic GluA1 (Kammermeier & Worley, 2007; Ronesi & Huber, 2008). While the specific downstream pathways of the mGlu_1/5/Homer1a protein complex are not yet extensively studied, there is support for increased MAPK/ERK signaling (Diering et al., 2017; Martin et al., 2019). Importantly, regulation of synaptic levels of Homer1a has been found to be mediated in part by neuromodulators including noradrenaline and adenosine which suppress and promote Homer1a synapse targeting, respectively (Diering et al., 2017). Although Homer1a mRNA is highest during wake, cortical synaptic levels are highest during sleep and states of increased sleep drive suggesting Homer1a levels are activity-dependent and contribute to sleep-maintained synaptic homeostasis (Diering et al., 2017; Maret et al., 2007). Overall, this process, termed homeostatic scaling-down, yields restorative benefits, increasing memory consolidation and opportunities for future learning (de Vivo et al., 2017; Diering et al., 2017; Martin et al., 2019; Tononi & Cirelli, 2014).

3.3.1.1. mGlu₁ receptors and sleep

In general, mGlu₁ has been implicated in modulating sleep duration. Recently, two loss-of-function mutations of GRM1, the gene encoding mGlu₁, were identified in two families diagnosed with familial natural short sleep (FNSS) (Shi et al., 2021). Individuals with FNSS can sleep less than 6h a day (compared to the average 8h of sleep) with no known adverse consequences (Aeschbach et al., 2003). When these mutations were introduced into mice, significant decreases in total and NREM sleep time were found with no effects on REM sleep time (Shi et al., 2021). No other studies to date have directly examined region-or circuit specific GRM1 genetic knockouts to further elucidate their involvement in sleep regulation. However, immunohistochemistry studies have shown that mGlu₁ expression was higher in rodents following a 12-h sleep deprivation period (Tadavarty, Rajput, Wong, Kumar, & Sastry, 2011), suggesting that mGlu₁ receptors may also serve a role in homeostatic regulation of sleep.

Additional studies examining downstream signaling pathways of mGlu₁ have elucidated the important role of this receptor in sleep via expression on TC neurons (Shigemoto et al., 1992). Activation of mGlu₁ leads to downstream PLC signaling, specifically the PLCβ4 isoform in TC neurons (Watanabe et al., 1998). Thus, to investigate the function of mGlu₁ in sleep, studies have examined sleep in PLCβ4-deficient (PLCβ4^−/−) mice. Both global PLCβ4^−/− and TC-restricted PLCβ4^−/− mice displayed longer total NREM sleep time as well as longer NREM and REM episodes. Delta power was also increased during NREM sleep (Hong et al., 2016). Alternately, a separate study reported no differences in NREM sleep in PLCβ4^−/− mice compared to WT mice but did find increased total REM duration, abnormal wake-to-REM transitions, and altered ultradian body temperature rhythms (Ikeda et al., 2009). Altogether, though the GRM1 mutation studies and studies using PLCβ4^−/−-deficient mice suggest different roles of mGlu₁ in sleep, it is clear that expression of mGlu₁ and the associated signaling pathways in TC circuits is likely involved in the transitions and maintenance of wake and sleep. Pharmacological studies examining receptor subtype-selective mGlu₁ compounds on sleep-wake architecture are also lacking, in large part due to the unavailability of selective ligands until recently (Luessen & Conn, 2022).

3.3.1.2. mGlu₅ receptors and sleep

Similar to the mGlu₁ receptor, human and animal studies suggest the mGlu₅ receptor serves a role in homeostatic regulation of sleep, with several lines of evidence supporting a role in sleep need/drive. Positron Emission Tomography (PET) imaging studies in healthy humans reported global increases in mGlu₅ receptor availability following a sleep deprivation period (Hefti et al., 2013; Holst et al., 2017; Weigend et al., 2019). mGlu₅ receptor availability was positively correlated with both subjective and objective measures of sleepiness as well as delta power during NREM sleep (Holst et al., 2017). Multiple EEG studies comparing mGlu₅ knockout (KO) mice with wild type (WT) mice corroborate these findings. First, mGlu₅ receptor KO mice showed less time awake and greater duration of NREM sleep during the active phase, though sleep was reportedly more fragmented (Aguilar et al., 2020). Alterations in spectral power were also found in mGlu₅ KO mice including decreased delta power during NREM, sleep, theta power during REM sleep, and alpha power during waking periods. These spectral changes could have implications in sleep quality and sleep homeostasis, sleep-dependent memory consolidation, and TRN function, respectively (Aguilar et al., 2020). Interestingly, some (Ahnaou, Langlois, et al., 2015; Ahnaou, Raeymaekers, et al., 2015; Holst et al., 2017) but not all (Aguilar et al., 2020) studies reported a lack of build-up in delta power following increased time awake across the dark (active) phase and following sleep deprivation, suggesting altered homeostatic sleep drive in KO animals. Additionally, one study reported extended waking periods following sleep deprivation in KO animals in contrast to the expected rebound sleep found in WT animals (Holst et al., 2017). However, a separate study reported a more rapid sleep rebound in mGlu₅ KO mice (Aguilar et al., 2020). Lastly, mGlu₅ KO mice showed higher basal gamma power during wake in both the light and dark phase (Ahnaou, Langlois, et al., 2015; Ahnaou, Raeymaekers, et al., 2015) as well as during NREM and REM sleep (Aguilar et al., 2020) suggesting heightened arousal and inability to shift from asynchronous to synchronous oscillatory activity.

Pharmacological studies using subtype-selective compounds have further aided in our understanding of the contributions of the mGlu₅ receptor to sleep regulation and underlying brain oscillatory activity (see Table 2 summarizing effects of mGlu selective compounds on sleep). Most highly selective compounds targeting mGlu₅ receptors include negative and positive allosteric modulators (NAMs and PAMs, respectively). NAMs and PAMs bind to an allosteric site and either reduce or enhance receptor function, respectively when a ligand binds to the orthosteric binding site, (Niswender & Conn, 2010). In rodent EEG studies, traditional mGlu₅ NAMs administered in the light cycle including MPEP and MTEP dose-dependently increased NREM sleep time and bout duration as well as REM latency and decreased REM duration within the 4–6h post-administration (Ahnaou, Langlois, et al., 2015; Ahnaou, Raeymaekers, et al., 2015; Cavas et al., 2013b; Harvey et al., 2013). Increased NREM duration, but not decreased REM duration, is consistent with results from mGlu₅ KO studies. Novel, more selective mGlu₅ compounds have also been examined. In contrast to MTEP and MPEP, basimglurant (RO4917523, RG7090; administered 2h into the dark cycle) and VU0424238 (administered 2h into the light cycle), both demonstrated initial but transient wake-promoting effects (Holter et al., 2021; Lindemann et al., 2015). However, as the wake-promoting effects of VU0424238 dissipated, selective reductions in REM sleep persisted for 8–10h following administration (Holter et al., 2021). Furthermore, all mGlu₅ NAMs mentioned above modulated quantitative EEG by producing increases in NREM delta power, which is possibly reflective of sleep drive, though this finding is inconsistent with studies in mGlu₅ KO mice (Harvey et al., 2013; Holter et al., 2021; Lindemann et al., 2015). However, similar to KO mice, multiple mGlu₅ NAMs increased high frequency gamma power during waking periods, which may be reflective of increased arousal (Ahnaou, Langlois, et al., 2015; Ahnaou, Raeymaekers, et al., 2015; Holter et al., 2021). Interestingly, M-5MPEP, a partial mGlu₅ NAM (at full receptor occupancy, there is only 50% functional inhibition), selectively decreased REM sleep without affecting quantitative EEG (Holter et al., 2021; Rodriguez et al., 2005). Lastly, in contrast, to mGlu₅ NAMs, mGlu₅ PAMs administered in the light phase including LSN2814617, CDPPB and ADX47273 demonstrated wake-promoting effects for up to 7h following administration (Ahnaou, Langlois, et al., 2015; Ahnaou, Raeymaekers, et al., 2015; Gilmour et al., 2013; Parmentier-Batteur et al., 2012). In the case of LSN2814617 (3mg/kg), while wake and REM sleep effects normalized at around 6h post-administration, decreases in deep sleep duration did not normalize until the onset of the dark cycle at 10h post-administration (Ahnaou, Langlois, et al., 2015; Ahnaou, Raeymaekers, et al., 2015). It is important to note tolerance developed to the wake-promoting effects of CDPPB, such that sleep-wake durations and latencies were not different from baseline following 7 days of repeated administration (Parmentier-Batteur et al., 2012).

Overall, mGlu₅ receptor activity influences multiple stages of the sleep-wake cycle. Human studies support a relationship between mGlu₅ receptor availability and sleep homeostasis/sleep need. In rodent models, both mGlu₅ KO mice and pharmacological studies confirm altered sleep-wake patterns. Of note, differences in selectivity, potency, and pharmacokinetics of mGlu₅ compounds need to be considered as these may contribute to a state-specific effect (e.g. selective REM sleep reduction) or initial wake-promotion that may artificially reduce both NREM and REM sleep; off-target effects of some compounds, especially MTEP and MPEP, may also be of influence (Montana et al., 2009).

3.3.2. Group II mGlu receptors and sleep

Similar to group I mGlu receptors, substantial preclinical literature describes the modulatory role of group II mGlu receptors on sleep. Using video analysis to determine sleep vs waking periods, Grm2/3^−/− mice had significant reductions in sleep time during both the light and dark phase as well as fragmented sleep (increased number of sleep bouts and decreased bout duration) (Pritchett et al., 2015). However, two additional studies using EEG reported no marked differences between Grm2/3^−/− and Grm2^−/− mice in waking, NREM and REM durations compared to WT mice (Ahnaou et al., 2009; Wood et al., 2018).

Pharmacological modulation of mGlu_2/3 receptors, however, has produced robust changes in sleep/wake patterns. Activation of mGlu_2/3 during the light cycle via agonists including LY379268 and LY354740 selectively increased REM sleep latency and reduced REM sleep duration without impacting wake or NREM sleep for 7–12h following administration (Ahnaou et al., 2009; Feinberg et al., 2002). mGlu_2/3 agonists also reduced high frequency oscillatory activity in the 10–50Hz range indicating that, in addition to selectively reducing REM sleep, these compounds also reduce cortical arousal during wake (Ahnaou et al., 2009; Feinberg et al., 2002). Importantly, these effects were absent in Grm2^−/− mice suggesting that a reduction in REM sleep is specifically attributed to activation of mGlu₂ and not mGlu₃ receptors (Ahnaou et al., 2009; Wood et al., 2018). mGlu_2/3 agonists have similar effects on sleep when compared to mGlu_2/3 PAMs including BINA, JNJ-42153605, THIIC and JNJ-40411813, although the latter two compounds also increased sleep duration and bout length for 8–12h following administration (Ahnaou et al., 2009; Cid et al., 2012; Fell et al., 2011; Lavreysen etal., 2013). Effects of mGlu_2/3 PAMs were also absent in Grm2^−/− mice (Ahnaou et al., 2009; Wood et al., 2018). In contrast, mGlu_2/3 antagonists (LY341495 and LY3020371) administered in both the light and dark phase and NAM (Ro-4491533) increased wake duration (both active and passive) and non-selectively decreased total sleep time. These wake-promoting effects were long-lasting, such that administration in the light cycle resulted in sustained changes through the beginning of the dark cycle (Ahnaou et al., 2014). Likely tied to increased arousal, increased gamma power was reported following administration, and this opposes effects to receptor activation (Ahnaou et al., 2014; Feinberg et al., 2005; Wood et al., 2018). Interestingly, reduced, yet still measurable, effects of the mGlu_2/3 antagonist LY3020371 and NAM Ro-4491533 were present in mGlu2^−/− rats suggesting that both mGlu₂ and mGlu₃ contribute to wake promotion (Wood et al., 2018). As noted above, acute pharmacological studies recapitulate some, but not all, findings from knockout mice.

3.3.3. Group III mGlu receptors and sleep

Unlike group I and II mGlu receptors, the influence of group III mGlu receptors on sleep-wake regulation are understudied. To our knowledge, no study has examined the influence of mGlu₄ receptors on sleep-wake architecture in KO mice or with selective ligands despite an extensive catalogue of pharmacological compounds and widespread interest in mGlu₄ receptors with regard to Parkinson’s Disease (Hopkins, Lindsley, & Niswender, 2009). Furthermore, research examining the influence mGlu₇ and mGlu₈ receptors on sleep regulation is scarce. A single study reported increased time awake and decreased NREM sleep duration in mice with a GRM7 deletion, the gene that encodes mGlu₇, compared to WT mice. Additionally, though these Grm7^−/− mice displayed no change in overall REM sleep duration, they demonstrated decreased average REM bout duration and increased bout numbers, suggesting changes in the maintenance of REM sleep (Fisher et al., 2020). In part attributed to the scarcity of selective compounds, pharmacological studies are also limited. The mGlu₇ PAM AMN082 produced differential effects on sleep in healthy rats. Acute administration of lower doses (5 and 10mg/kg) during the light phase increased NREM sleep and decreased time awake whereas a high dose (20mg/kg) increased time awake and decreased both NREM and REM sleep (Cavas, Scesa, & Navarro, 2013a). However, similar effects were present when examined in Grm7^−/− mice (Ahnaou, Raeyemaekers, Huysmans, & Drinkenburg, 2016) suggesting off-target effects. Thus, there is still a need to examine more selective compounds to better characterize the role for mGlu₇ in sleep regulation. Lastly, a single study reports that the mGlu₈ agonist (S)-3,4-DCPG did not affect any sleep parameter in healthy rats (Cavas et al., 2017).

Overall, the individual mGlu receptors have differential roles in sleep/wake regulation. In general, both pharmacological and knockout models examining group I receptors identified contributions to both NREM and REM sleep time. Further, in conjunction with human findings, several rodent studies identified changes in spectral activity that reflected contributions to homeostatic sleep drive. In contrast, group II receptors appear to have less profound contributions in EEG knockout studies, and, in pharmacological studies, many compounds appear to produce REM-selective inhibition (agonists/PAMs) or robust wake-promotion (antagonists/NAMs). Lastly, while some roles of group III receptors have been reported, the literature is still in early stages with genetic and pharmacological advancements necessary to further our understanding of their role (or lack thereof ) in regulating sleep.

3.4. mGlu receptors, neuropsychiatric disorders and sleep

Comorbid sleep disturbances including insomnia have been found in the majority of CNS disorders (Benca & Buysse, 2018; Winkelman & de Lecea, 2020; Wulff et al., 2010). Although primarily diagnosed subjectively and defined as difficulty initiating or maintaining sleep, insomnia is also associated with aberrant quantitative EEG characteristics including increased high frequency (beta/gamma) activity at sleep onset and during NREM sleep (for review, see Zhao et al., 2021). Importantly, the DSM-5 delineated insomnia as a separate diagnosis from other mental disorders as opposed to a primary or secondary symptom (APA, 2013). Thus, as a separate diagnosis, insomnia should also be a primary target of pharmacotherapies (Benca & Buysse, 2018). Furthermore, the DSM-5 emphasizes the bidirectional relationship between insomnia and neuropsychiatric illness, where both disorders can serve as a risk factor for the development of one another (Benca & Buysse, 2018; Winkelman & de Lecea, 2020). Support for insomnia as a risk factor stemmed from data collected from the National Institute of Mental Health’s Epidemiologic Catchment Area Survey. This survey collected comprehensive information on mental disorders in the United States and included questions about sleep disturbances in addition to current psychiatric symptoms. Findings from this survey suggested that insomnia increased the risk of first episode diagnosis of MDD (Ford & Kamerow, 1989). As discussed below, preexisting insomnia has also been implicated in the onset of SUD, PTSD, schizophrenia, and AD (Irwin & Vitiello, 2019; Maher, Rego, & Asnis, 2006; Manoach & Stickgold, 2009; Neylan et al., 2021; Riemann, Berger, & Voderholzer, 2001; Roehrs, Sibai, & Roth, 2021; Zhang et al., 2022).

It is important to understand how the glutamate system may contribute to the comorbid and bidirectional relationship between sleep and CNS disorders. As there are ongoing efforts to develop mGlu receptor subtype-selective treatments for these conditions, current and future research should work to determine if mGlu receptor compounds can target both sleep disturbances and the comorbid neuropsychiatric disorders. In the following sections, we present preclinical and clinical literature detailing known alterations in glutamatergic function associated with schizophrenia, MDD, PTSD, AD, and SUD as well as the comorbid sleep disruptions frequently linked with these conditions. As the field is still expanding, we build on the predominately preclinical pharmacological data above and speculate on possible mGlu receptor manipulations/strategies to improve sleep disturbances in addition to other symptoms of the CNS disorder.

3.4.1. Schizophrenia

Schizophrenia is a neuropsychiatric disorder that affects around 1% of the human population and is characterized by three primary symptoms clusters: positive, negative, and cognitive ( Jones, Watson, & Fone, 2011; Manoach & Stickgold, 2019; Winship et al., 2019). Sleep disruptions represent an additional underappreciated symptom cluster (Manoach & Stickgold, 2019). Sleep disturbances are reported in both medicated and drug-naïve patients, affecting ~80% of all patients, and are linked with symptom prevalence and severity (Chen et al., 2020; Cohrs, 2008; Korenic et al., 2020; Manoach & Stickgold, 2019; Sprecher et al., 2015). Patients commonly experience insomnia and related increases in NREM sleep latency as well as reduced sleep quality (e.g. reduced delta power in N3 sleep) and sleep efficiency (defined as the ratio of time spent asleep when in bed) (Benca & Buysse, 2018; Manoach & Stickgold, 2019; Sprecher et al., 2015). Decreased REM sleep latency has also been found during acute psychosis, though reported changes in REM sleep duration in patients with schizophrenia are inconsistent throughout literature (Pritchett et al., 2012; Sprecher et al., 2015). Importantly, sleep disturbances can precede the onset of schizophrenia and associated symptoms. Insomnia has been found to precede the onset of positive symptoms including hallucinations and paranoia (for review, see Reeve et al., 2015), and, similarly, decreased NREM sleep duration and quality has been associated with the onset of negative and cognitive symptoms (Göder et al., 2004; Göder et al., 2006; Keshavan et al., 1995). Sleep spindles, which are important for memory consolidation and cognition, also occur less frequently in medicated patients with schizophrenia relative to healthy humans (Göder et al., 2006, 2015; Manoach & Stickgold, 2019; Wamsley et al., 2012). Together, altered sleep duration and architecture are directly associated with all primary symptom clusters of schizophrenia.

Current antipsychotic medications, which exert therapeutic effects via blockade of dopamine D2 and serotonin 5-HT2A receptors, affect sleep and can contribute to results of sleep studies conducted in medicated patients. Both first- and second-generation antipsychotics increase total sleep time in healthy humans and in patients with schizophrenia in part by decreasing monoaminergic mediated arousal circuits, and these sedative effects may persist during wake (Göder et al., 2004; Manoach et al., 2010; Manoach & Stickgold, 2009). Although literature is mixed on specific changes, second generation antipsychotics have commonly been found to increase both NREM and REM sleep time in patients with schizophrenia (for review and more detailed descriptions, see Cohrs, 2008). Importantly, antipsychotic medications are used as off-label agents to treat insomnia (Benca & Buysse, 2018). Thus, although many of the current FDA-approved medications can increase sleep in patients with schizophrenia, novel, more effective pharmacotherapies that target cognitive impairments and have a lower adverse effect risk than current antipsychotics are needed.

Although antipsychotics were initially developed on the premise that excessive dopamine in the striatum contributed to the positive symptoms, growing evidence supports glutamatergic dysregulation, specifically N-methyl-d-aspartate receptor (NMDAR) hypofunction, as a key contributor to multiple symptoms of schizophrenia (Coyle, 2006; Javitt, 2007; Jones et al., 2011; Kantrowitz & Javitt, 2012). This hypothesis stemmed from early studies that reported psychotic-like behavior in healthy humans following administration of NMDAR antagonists (Davies & Beech, 1960; Luby, Cohen, Rosenbaum, Gottlieb, & Kelley, 1959). Since then, multiple studies have reported similarities in brain function, psychotic-like behaviors, and cognitive impairments between healthy humans administered NMDAR-antagonists and patients with schizophrenia, and NMDAR antagonists have been integrated into many preclinical studies to model all symptom clusters (for review, see Javitt, 2007; Jones et al., 2011). Genetic, functional imaging, and post-mortem brain tissue studies have provided substantial evidence of neurobiological alterations in the glutamate system in patients with schizophrenia (Hu, MacDonald, Elswick, & Sweet, 2015; Korenic et al., 2020). For example, altered pyramidal cell morphology was found in the cortex of post-mortem brain tissue, including decreased dendritic spine density, dendritic length, and somal size (Black et al., 2004; Glantz & Lewis, 2000; Kalus, Müller, Zuschratter, & Senitz, 2000; Pierri, Volk, Auh, Sampson, & Lewis, 2001, 2003). Changes in enzymes that have a role in glutamate synthesis and metabolism have also been reported. Protein expression of glutamine synthetase was lower and activity of phosphate-activated glutaminase (PAG) was higher in the PFC. These enzymes are responsible for converting glutamate to glutamine and vice versa suggesting impairment in glutamate metabolism in patients with schizophrenia compared to healthy controls (Burbaeva et al., 2003; Gluck, Thomas, Davis, & Haroutunian, 2002). Importantly, studies have found that alterations in glutamate neurotransmission correspond with associated sleep impairments. Using [¹H]-MRS, Korenic et al., 2020 found poor sleep quality in patients with schizophrenia was associated with higher Glx (glutamate and glutamine) levels in the hippocampus and lower levels of glutamate in the anterior cingulate cortex. The glutamate levels in the anterior cingulate cortex correlated with lower total scores on the Brief Psychiatric Rating Scale (BPRS). Moreover, poor sleep quality corresponded with heightened severity of positive symptoms. Taken together, these clinical studies firmly support a relationship between glutamate dysregulation, schizophrenia, and sleep.

Alterations in mGlu receptors have also been identified. Post-mortem tissue from patients with schizophrenia revealed higher protein expression of mGlu₅ and several mGlu₅ trafficking proteins in the hippocampus (Matosin et al., 2015). Conversely, a PET imaging study reported lower cortical mGlu₅ binding potential in patients with schizophrenia compared to healthy controls and lower binding was associated with an increase in negative symptoms and worse cognitive performance (Régio Brambilla et al., 2020). Lastly, polymorphisms of GRM5, the gene encoding the mGlu₅ receptor, have also been reported in patients with schizophrenia (Matosin et al., 2018). This relationship between mGlu₅ and symptoms of schizophrenia is further supported by pharmacological studies. Some mGlu₅ NAMs, including fenobam, induced psychotomimetic-like effects in both humans and animals. These compounds were also associated with increased resting-state high frequency gamma power, which is commonly found in patients with schizophrenia (Ahnaou, Langlois, et al., 2015; Ahnaou, Raeymaekers, et al., 2015; Holter et al., 2021; Homayoun, Stefani, Adams, Tamagan, & Moghaddam, 2004; Jacob et al., 2009). The mGlu₅ receptor is coupled to the NMDAR (Niswender & Conn, 2010; Shigemoto et al., 1993), and, thus, one hypothesized mechanism through which mGlu₅ NAMs produce these psychotomimetic-like effects is via downstream NMDAR inhibition (Sengmany & Gregory, 2016). Furthermore, as described above, global knockout of mGlu₅ in mice recapitulated some sleep deficits similar to those in patients with schizophrenia (Aguilar et al., 2020). Taken together, these preclinical and clinical studies support an underlying role of mGlu₅ hypofunction in schizophrenia symptomology.

While overwhelming evidence suggests NMDAR dysregulation may underlie all symptom clusters of schizophrenia, directly targeting NMDARs as a therapeutic approach has been met with adverse effects including seizures. Thus, modulating mGlu receptors represents a more tolerable approach than direct NMDAR activation, and all groups of mGlu receptors have gained preclinical support as potential therapeutic targets (Dogra & Conn, 2022; Luessen & Conn, 2022). Activation/allosteric modulation of group I mGlu receptors has shown promise through increasing presynaptic glutamate release and postsynaptic NMDAR function. Support for the use of both mGlu₁ and mGlu₅ PAMs spans multiple preclinical models of all primary symptom clusters. For example, compounds targeting either receptor have reversed amphetamine and NMDAR antagonist-induced hyperlocomotion, deficits in paired pulse inhibition, and deficits in multiple cognitive assessments (for review, seeDogra & Conn, 2022; Maksymetz et al., 2017). Although mGlu₅ PAMs increase time awake and reduce sleep in healthy rodents, the potential for these compounds to modulate sleep spindles still warrants examination of the sleep-altering effects in animal models of schizophrenia. As noted above, patients with schizophrenia and mGlu₅ KO mice showed lower sleep spindle frequencies (Aguilar et al., 2020; Göder et al., 2006, 2015; Manoach & Stickgold, 2019; Wamsley et al., 2012). Sleep spindles are generated in the TRN (Bazhenov, Timofeev, Steriade, & Sejnowski, 2000; Ferrarelli & Tononi, 2017) and are in part mediated by group I mGlu receptor activation (Sun et al., 2016). Thus, both mGlu₁ and mGlu₅ PAMs may enhance sleep spindle activity and, thereby, memory consolidation. Patients with schizophrenia also report excessive daytime sleepiness which may be a product of insomnia, reduced sleep quality, and/or the sedative effects of current antipsychotics (Cohrs, 2008; Miller, 2004; Sharma, 2016). Increasing arousal/time awake during the day could aid with sleep consolidation at night, similar to the primary goal of the cognitive behavioral therapy approach for insomnia (Haynes, Talbert, Fox, & Close, 2018). These can also be employed as an adjunct treatment to counteract the sedative effects of many antipsychotics. mGlu₅ PAMs have not yet been examined on sleep-wake activity in preclinical models of schizophrenia, though these studies are critical to determine potential of these compounds to normalize sleep disturbances. It is also important to consider time of dosing in future studies to mitigate undesired wake promoting/sleep-disrupting effects.

Group II mGlu receptors are also promising pharmacotherapeutic targets for schizophrenia. NMDAR hypofunction induced by ketamine and PCP ultimately leads to increased levels of synaptic glutamate in regions including the PFC (Lorrain, Baccei, Bristow, Anderson, & Varney, 2003; Maksymetz et al., 2017; Moghaddam & Jackson, 2003). Activation of mGlu_2/3 presynaptic autoreceptors causes a reduction in synaptic glutamate concentrations (Dogra & Conn, 2022; Maksymetz et al., 2017) and, thus, may be a viable mechanism to attenuate this aberrant glutamate efflux. Similar to group I PAMs, mGlu_2/3 agonists and PAMs have shown efficacy in preclinical models of all primary symptom clusters of schizophrenia (Benneyworth et al., 2007; Galicietal., 2006; Greco etal., 2005; Griebel et al., 2016; Hackler et al., 2010; Kawaura, Karasawa, & Hikichi, 2016; Maksymetz et al., 2017; Moghaddam & Adams, 1998). Following extensive support from preclinical studies, the mGlu_2/3 agonist LY2140023 advanced into clinical trials for treatment of schizophrenia but had mixed results in phase II studies. However, significant improvements in positive and negative symptoms were reported in some trials supporting activation of mGlu_2/3 as a mechanism for further pursuit (for review, see Maksymetz et al., 2017). Unlike mGlu₅ PAMs, mGlu_2/3 PAMs may influence sleep directly by reducing extracellular glutamate levels. Newer, more selective compounds increased NREM sleep time and bout length (Fell et al., 2011; Lavreysen et al., 2013) and, thus, may be of benefit to patients experiencing insomnia. Furthermore, although not directly investigated, mGlu₂ receptors are also located in the TRN and regulate signaling (Crabtree, Lodge, Bashir, & Isaac, 2013) suggesting a possible, but not yet studied, influence on sleep spindle activity.

In summary, both mGlu₅ and mGlu_2/3 PAMs represent promising therapeutic approaches for multiple symptoms of schizophrenia. While speculative, these compounds may appropriately address associated sleep disturbances. It is important to note that preclinical studies also support targeting group III mGlu receptors, specifically with mGlu₄ PAMs and mGlu₇ NAMs, for all primary symptom clusters of schizophrenia (for review, see Dogra & Conn, 2022 and Luessen & Conn, 2022). However, as reiterated throughout this chapter, further research is needed to understand the potential of group III compounds to target schizophrenia-associated sleep disturbances.

3.4.2. Major depressive disorder

Major depressive disorder (MDD) is characterized by the DSM-5 as having at least five symptoms including depressed mood, loss of interest in activities, fatigue, and sleep disturbances for a 2-week period (APA, 2013). Sleep disturbances are reported in around 90% of patients with MDD, and, while the majority of these patients report insomnia, there is a smaller subset of patients (15–35%) who report hypersomnia (Armitage, 2007; Reynolds & Kupfer, 1987; Steiger & Pawlowski, 2019). Further, there appears to be a strong bidirectional relationship between depression and insomnia. First, in an epidemiological study examining an adult sample from the UK population, insomnia was linked to preexisting depression (Morphy, Dunn, Lewis, Boardman, & Croft, 2007). Multiple epidemiological studies have also found preexisting insomnia was associated with the development of depression (Fang, Tu, Sheng, & Shao, 2019; Ford & Kamerow, 1989; Morphy et al., 2007; Riemann et al., 2020). However, it is important to note that while the majority of patients with depression have comorbid insomnia, only around 20% of patients with insomnia are diagnosed with depression (Breslau, Roth, Rosenthal, & Andreski, 1996). Polysomnography studies reveal that, in general, patients with depression experience a reduction in overall sleep time and display increased NREM sleep latency alongside decreased NREM sleep duration and increased wake after sleep onset (WASO) (Benca et al., 1992; Steiger & Kimura, 2010; Wichniak, Wierzbicka, & Jernajczyk, 2013; Wichniak, Wierzbicka, Walęcka, & Jernajczyk, 2017). This is paired with reduced delta power during NREM sleep (Armitage, 1995; Kupfer, Reynolds, Ulrich, & Grochocinski, 1986; Steiger & Kimura, 2010). Interestingly, in contrast to NREM sleep, patients with MDD concurrently experience an increase in REM sleep duration and density and a reduction in REM sleep latency (Foster, Kupfer, Coble, & McPartland, 1976; Steiger & Pawlowski, 2019; Wichniak et al., 2013). Although variable, most antidepressant medications selectively decrease REM sleep and increase REM sleep latency without producing concomitant decreases in NREM sleep (Steiger & Pawlowski, 2019; Wichniak et al., 2017). Antidepressants have also been found to increase NREM sleep delta power (Kupfer, Ehlers, Pollock, Swami Nathan, & Perel, 1989; Steiger & Kimura, 2010).Importantly, the ability of medications to mitigate REM sleep abnormalities has been shown to correspond with treatment efficacy, reiterating the importance of improving sleep for overall treatment outcome (Riemann et al., 2001).

There is substantial evidence for glutamatergic contributions to the pathology of MDD and some even suggest there should be a glutamate hypothesis of depression (Sanacora, Treccani, & Popoli, 2012). Several studies in clinical populations have shown differences in glutamate levels in the CSF, plasma and platelets, and post-mortem brain tissue between individuals with MDD and healthy individuals. For example, increased glutamate levels in the plasma have been found in patients with MDD relative to a healthy control group (Altamura et al., 1993; Kim, Schmid-Burgk, Claus, & Kornhuber, 1982; Mauri et al., 1998). In contrast, decreased glutamate levels were reported in the CSF in patients with refractory affective disorder (includes both unipolar and bipolar patients) (Frye, Tsai, Huggins, Coyle, & Post, 2007), although earlier studies examining glutamine levels in the CSF found increases in patients with unipolar and bipolar depression (Levine et al., 2000). In studies examining post-mortem tissue, increased glutamate levels were found in the frontal cortex (Hashimoto, Sawa, & Iyo, 2007) and dorsolateral prefrontal cortex (Lan et al., 2009). Lastly, several studies have examined glutamate levels in patients with MDD using ¹H-MRS. Collectively, a meta-analysis reported consistent decreases in concentrations of glutamate both globally and specifically in the anterior cingulate cortex (Luykx et al., 2012). In short, these clinical findings have pointed to clear dysregulations in glutamate levels in patients with MDD, and the sampling methodology, brain-region, and/or heterogeneity associated with MDD may have influenced varying findings across studies.

Targeting the glutamate system through both iGlu and mGlu receptors has been explored for depression. Early findings suggest that ketamine, an NMDAR antagonist, produced rapid-acting antidepressant effects, and NMDAR antagonism remains a promising therapeutic approach (Berman et al., 2000; Raab-Graham, Workman, Namjoshi, & Niere, 2016; Trullas & Skolnick, 1990; Zarate et al., 2006). Of relevance, ketamine administration to individuals with MDD decreased mGlu₅ receptor availability in several brain regions and hippocampal mGlu₅ availability correlated with Montgomery-Å sberg Depression Rating Scale (MADRS) total score (Esterlis et al., 2018). Thus, long-standing research also supports targeting mGlu receptors as a treatment approach for MDD. Functional antagonism of mGlu₅ displayed therapeutic potential in preclinical models that have predictive validity for antidepressant medications including the forced swim test (FST) and tail suspension test (TST) (Felts et al., 2017; Gould et al., 2016; Hughes et al., 2013; Kato et al., 2015; Lindemann et al., 2015). Importantly, extensive preclinical work suggests mGlu₅ NAMs may be beneficial for treating sleep disturbances associated with MDD. For example, the mGlu₅ NAM basimglurant significantly increased REM latency and decreased REM sleep and the REM/NREM ratio in healthy rats when administered during the active phase. Furthermore, basimglurant increased delta power during NREM sleep. This profile is similar to other mGlu₅ NAMs (e.g. Holter et al., 2021) and consistent with several clinically prescribed antidepressants (Steiger & Pawlowski, 2019; Wichniak et al., 2013, 2017). Furthermore, some mGlu₅ NAMs have progressed into clinical trials for MDD, though historically with negative outcomes due to the adverse effects noted above (e.g. fenobam, Pecknold, McClure, Appeltauer, Wrzesinski, & Allan, 1982). However, basimglurant advanced to phase II clinical trials as an adjunctive therapy for depression. Although the primary endpoint criteria, change in MADRS total score, was not met with this compound, significant results were found in secondary endpoint criteria including other self-report measures of depressive symptoms (Quiroz et al., 2016). It is critical to note that sleep was not evaluated in this clinical trial.

Clinical evidence, although mixed, also supports a role of group II mGlu receptors in MDD. One study reported increased mGlu_2/3 protein levels in the PFC of postmortem tissue (Feyissa et al., 2010) while a PET imaging study reported significantly lower binding in the anterior cingulate cortex in patients with MDD suggesting the relationship between mGlu_2/3 and MDD is complex (Mcomish et al., 2016). Both agonists and antagonists of group II mGlu receptors have been pursued in preclinical and clinical studies for their antidepressant-like potential. Antagonism of the mGlu_2/3 receptor has been most heavily pursued because these compounds have similar mechanisms to ketamine. Ultimately, similar to subanesthetic doses of ketamine, administration of an mGlu_2/3 antagonist should lead to increased glutamate release because presynaptic mGlu_2/3 receptors function as autoreceptors (Chaki, Koike, & Fukumoto, 2019). Aside from sleep, preclinical studies showed mGlu_2/3 antagonists and NAMs display antidepressant-like effects in preclinical models including the FST and TST (Chaki et al., 2004; Fukumoto, Iijima, & Chaki, 2016; Witkin et al., 2016). Interestingly, both mGlu₂ and mGlu₃ contribute to this effect because selective compounds VU6001966 (mGlu₂ NAM) and VU650786 (mGlu₃ NAM) both decreased immobility time in the FST ( Joffe et al., 2020). Additionally, the mGlu_2/3 NAM decoglurant advanced to clinical trials as an adjunctive treatment to SSRIs or SNRIs but was unsuccessful in meeting primary or secondary endpoint criteria. However, complexities with the clinical trial including a high placebo response made it difficult to adequately evaluate the compound (Umbricht et al., 2020). Importantly, mGlu_2/3 antagonists may be relevant to target sleep disturbances in patients with depression because of their ability to reduce REM sleep. However, applicability of mGlu_2/3 antagonists may be population-dependent, perhaps better suited for the sub-population presenting with hypersomnia, as mGlu_2/3 antagonists also have wake-promoting effects.

mGlu_2/3 agonists and PAMs may have greater potential for targeting sleep disturbances in MDD. Several of these compounds have also displayed antidepressant-like effects in preclinical screens including the FST (Fell et al., 2011). Furthermore, the mGlu_2/3 agonist LY379268 demonstrated promising antidepressant-like effects when administered as an adjunctive treatment to fluoxetine and may help bridge the gap between treatment initiation and delayed onset of the therapeutic effects of classic antidepressants (Matrisciano et al., 2007). Importantly, similar to mGlu_2/3 antagonists, these compounds also significantly decrease REM sleep time and increase REM latency, but, unlike mGlu_2/3 antagonists, most do not promote wake (Ahnaou et al., 2009; Cid et al., 2012). Thus, this approach may be more beneficial to those who are concurrently experiencing insomnia. Because elevated extracellular glutamate concentrations were found in the orbitofrontal cortex during REM sleep in the rat (Lopez-Rodriguez et al., 2007), activation of mGlu_2/3 agonists may be a more suitable approach to correct REM sleep abnormalities in MDD by reducing overall synaptic glutamate levels and selectively suppressing REM sleep (Ahnaou et al., 2009).

In summary, reducing glutamate function via mGlu₅ NAMs or mGlu_2/3 PAMs may be a viable approach to target primary symptoms and sleep disruptions in MDD. As evaluated with basimglurant, the ability of new compounds to normalize sleep disruptions associated with MDD should be integrated into future clinical trial design.

3.4.3. Post-traumatic stress disorder

Post-traumatic stress disorder (PTSD) affects roughly 4% of the population and is associated with fear, anxiety, avoidance, negative mood, intrusive thoughts, hyperarousal, and sleep disturbances, including insomnia, that persist following a traumatic event (Kessler, Wai, Demler, & Walters, 2005; Kilpatrick et al., 2013). While objective measures of sleep within the PTSD population are inconclusive, perhaps due to the heterogeneity of individuals and their associated trauma, many studies report increased light sleep, decreased SWS, increased REM density, and decreased REM bout length (Kobayashi, Boarts, & Delahanty, 2007; Kobayashi, Huntley, Lavela, & Mellman, 2012; Lewis et al., 2020; Woodward, Leskin, & Sheikh, 2002). Subjectively, patients with PTSD report sleep disturbances including nightmares, sleep terrors, limb movements, and dream enactment (Baird et al., 2018). Additionally, the prevalence of sleep disordered breathing (e.g. sleep apnea) and sleep movement disorders, including periodic limb movement disorder and REM movement disorders, are higher among those with PTSD (Baird et al., 2018; Maher et al., 2006; Neylan et al., 2021). Sleep disturbances have also been found to correlate with prevalence and severity of other symptoms. For example, insomnia severity correlated with intrusion and hyperarousal (Lauterbach, Behnke, & McSweeney, 2011). The frequency and severity of nightmares was associated with a higher prevalence of hopelessness and suicidal behaviors (Littlewood, Gooding, Panagioti, & Kyle, 2016). Further, in support of a bidirectional relationship, it is well-established that preexisting sleep disturbances influence the development of PTSD. Insomnia and stress reactivity prior to a motor vehicle collision predicted subsequent PTSD diagnosis (Neylan et al., 2021). Additionally, some studies evaluating self-reported sleep measures including nightmares and insomnia prior to deployment corresponded to subsequent PTSD development in combat veterans (van Liempt, van Zuiden, Westenberg, Super, & Vermetten, 2013; Wang et al., 2019). Given the important role of sleep in regulating stress, emotion, and learning and memory, it is not surprising that dysregulated sleep may exacerbate the development and progression of PTSD.

Glutamatergic function has been shown to be altered in individuals with PTSD which may directly contribute to PTSD symptoms including altered fear reactivity, hyperarousal, and dysregulated sleep. Using ¹³C-acetate MRS, patients with PTSD showed reduced glutamatergic synaptic strength in the PFC compared to healthy controls (Averill et al., 2022). Similarly, in rodents, the single prolonged stress model reduced glutamatergic excitation of pyramidal neurons in the infralimbic medial PFC (Nawreen, Baccei, & Herman, 2021). Another study using ¹H-MRS imaging in humans reported lower GABA and increased glutamate levels in the cortex in patients with PTSD, and these findings correlated with higher subjective ratings of insomnia (Meyerhoff, Mon, Metzler, & Neylan, 2014). In further support of research that suggests sleep disturbances may exacerbate neurobiological factors in PTSD, studies conducted in healthy humans found a single night of sleep deprivation altered functional connectivity between the amygdala and PFC and was associated with aberrant amygdalar reactivity (Killgore, 2013; Yoo, Gujar, Hu, Jolesz, & Walker, 2007). These data support the prevailing theory that inadequate top-down PFC inhibitory control leads to hyperactivity in the amygdala and over-reactivity to stressful stimuli (Rauch, Shin, & Phelps, 2006). Furthermore, as many of the sleep disturbances and other symptoms derived from hyperarousal occur in part because of increased glutamate activation of the hippocampus and amygdala, putative pharmacotherapeutics should attenuate glutamate to restore excitatory:inhibitory balance.

Current treatments for PTSD include cognitive behavioral therapy and exposure therapy, which are aimed to help individuals re-process traumatic events and re-structure responses in an effort to attenuate hyperarousal and responsivity to environmental stimuli (American Psychological Association, 2017). Treatment of the emotion-related symptoms of PTSD have been found to positively impact sleep. For example, subjective and objective measures of sleep improved in patients with PTSD following successful outcomes of trauma-targeted psychotherapies (e.g. remission of emotional symptoms) (Rousseau et al., 2021). While severity of sleep disturbances may be reduced following behavioral treatment approaches, they are not fully resolved (Taylor et al., 2020). There are only two FDA-approved pharmacological treatments for PTSD, the antidepressant medications sertraline and paroxetine (Brady et al., 2000; Krystal et al., 2017; Marshall, Beebe, Oldham, & Zaninelli, 2001). However, both antidepressants, in addition to commonly used anxiolytics, target depressed mood and anxiety symptoms rather than directly targeting hyperarousal or sleep disturbances. Since NMDAR hyperactivity has been associated with this hyperarousal and PTSD formation after trauma, NMDAR antagonists including ketamine have been proposed to both target this hyperarousal and provide rapid antidepressant effects (Liriano, Hatten, & Schwartz, 2019). A single dose of ketamine decreased subjective measures of intrusion, avoidance, and hyperarousal in patients with PTSD (Feder et al., 2014). Additionally, ketamine reduced the likelihood of developing PTSD after burn injury (McGhee, Maani, Garza, Gaylord, & Black, 2008). The partial NMDAR agonist D-cycloserine has also been investigated and was administered alongside exposure therapy to enhance extinction learning (Baker, Cates, & Luthin, 2017; George et al., 2018). Lastly, while sleep aids such as benzodiazepines, zolpidem, buspirone, and gabapentin have been tested in addition to cognitive behavioral therapy for PTSD-induced sleep disturbances, results are variable (Maher et al., 2006). Thus, as reiterated throughout, novel, disorder-specific medications that target sleep are still needed.

Emerging evidence supports targeting both group I and group II mGlu receptors for symptoms of PTSD and associated sleep disturbances. Importantly, unlike clinical and preclinical studies for other CNS disorders, mGlu receptor compounds have been assessed for their ability to improve sleep disturbances in the context of animal models of traumatic stress. First, functional antagonism of the mGlu₅ receptor may be effective in treating both sleep disruptions and hyperarousal associated with PTSD. In vivo PET imaging showed that patients with PTSD have higher mGlu₅ receptor availability in the hippocampus relative to healthy controls, and mGlu₅ receptor availability correlated with heightened avoidance symptoms and suicidal ideations (Davis et al., 2019; Esterlis, Holmes, Sharma, Krystal, & Delorenzo, 2018; Holmes et al., 2017). In preclinical rodent models, treatments targeting mGlu₅ receptors also show promise in treating PTSD due to their role in fear, avoidance, and extinction learning (Riedel, Casabona, Platt, MacPhail, & Nicoletti, 2000). As previously discussed, mGlu₅ NAMs inhibit arousal and have been shown to selectively reduce REM sleep, and may modulate sleep disruptions induced by single prolonged stress in rodents (Nedelcovych et al., 2015). Given that the hippocampus influences memory consolidation and REM sleep and the amygdala regulates emotional processing (Rasch & Born, 2013; Wang, Liu, Li, Qu, & Huang, 2021), it is theorized that hippocampal and amygdalar hyperactivity during REM sleep decreases extinction learning, increases emotional response to trauma, and increases PTSD symptom manifestation (Bottary et al., 2020; Cominski, Jiao, Catuzzi, Stewart, & Pang, 2014; Pace-Schott, Germain, & Milad, 2015; Straus et al., 2018). Thus, mGlu₅ NAMs may be a two-fold treatment approach that facilitates extinction learning and targets the sleep disturbances associated with PTSD.

To date, most work has focused on mGlu₅ as a potential treatment for PTSD and associated sleep impairments. However, activation of group II receptors also produces selective decreases in REM sleep and preclinical studies support group II receptors as a therapeutic target for PTSD. Given their-expression in regions including the amygdala, both mGlu₂ and mGlu₃ receptors have been implicated in fear memory and extinction learning (Kim et al., 2015; Linden et al., 2005; Potter, Zanos, & Gould, 2020; Sweeten, Adkins, Wellman, & Sanford, 2021). Of relevance, Sweeten et al., 2021 explored the therapeutic potential of mGlu_2/3 agonist LY379268 for sleep disturbances associated with fear extinction and found that a microinjection of LY379268 to the basolateral amygdala normalized the increased REM response associated with context re-exposure. Thus, group II receptors remain an exciting area of future research and should be further pursued. Lastly, it is important to note group III receptors (notably mGlu₇ and mGlu₈) also modulate learning and memory, including aversive and extinction learning, though their effects on sleep disturbances have not yet been studied (Fendt et al., 2013; O’Connor et al., 2019).

3.4.4. Alzheimer’s disease

As the most common form of dementia, Alzheimer’s Disease (AD) is characterized by neurodegeneration, aggregation of extracellular amyloid β (Aβ) into amyloid plaques, and the aggregation of intracellular tau into neurofibrillary tangles (NFTs), which begin to accumulate approximately 15 years before cognitive decline ( Jack et al., 2010; Jack et al., 2016). Importantly, sleep disruptions are common in both clinical populations (even in prodromal stages) and preclinical models, and these have correlated with cognitive decline (Lim, Kowgier, Yu, Buchman, & Bennett, 2013; Potvin et al., 2012; Romanella et al., 2021). In clinical populations, total sleep time and SWS percentage were decreased and REM sleep latency and waking after sleep onset (WASO) were increased in patients with AD when compared to healthy controls (Bliwise et al., 1989; Zhang et al., 2022). Furthermore, a recent meta-analysis examining polysomnographic sleep changes in patients with AD found a significant positive relationship between SWS and REM sleep duration and Mini-Mental Sate Examination (MMSE) scores, such that reduced sleep duration correlated with lower MMSE scores (Zhang et al., 2022).

Electroencephalography (EEG) studies in patients with AD have reported several age-related differences, especially in spectral waveforms, that may directly contribute to cognitive dysfunction and AD pathology (Lyashenko, Poluektov, Levin, & Pchelina, 2016; Scullin & Bliwise, 2015a, 2015b; Song et al., 2015; Xie et al., 2013; Zhang et al., 2022). Patients with AD demonstrated increased power in low frequency (delta/theta) bands and decreased alpha power when awake compared to healthy aged or young cohorts (Czigler et al., 2008; Rondina et al., 2016). These waveforms each have distinct roles in cognitive processes and, thus, may correspond with AD-associated cognitive decline. Additionally, findings of reduced delta power (0.5–4Hz) during NREM sleep, as well as sleep spindle abnormalities including reduced density, amplitude, and duration, have been reported in patients with AD; these further correlated with impaired cognitive performance (Scullin & Bliwise, 2015b; Weng, Lei, & Yu, 2020).

Importantly, disrupted sleep has been implicated as both a cause and a consequence of AD pathology. Homeostatic sleep is important for Aβ clearance, and, in healthy adult participants, disrupted slow wave activity was found to increase AD pathology and Aβ levels (Ju et al., 2017; Lucey et al., 2019). Furthermore, healthy individuals with amyloid deposition demonstrated worsened sleep efficiency compared to individuals without amyloid deposition (Ju et al., 2013). Murine models of AD-like pathology also support a relationship between sleep and pathological development. For example, in transgenic mice expressing the human amyloid precursor protein (APP), Aβ accumulation was observed in brain interstitial fluid following sleep deprivation (Kang et al., 2009). Additionally, sleep disturbances associated with AD may also contribute to sundown syndrome, a phenomenon in which neuropsychiatric symptoms such as delusions, irritation, and depression are exacerbated during the late afternoon and night hours (Canevelli et al., 2016). Although the underlying causes of these are not well understood, possible contributing factors include abnormal sleep schedules (e.g. reduced nighttime sleep and increased daytime napping) as well as circadian disruptions (e.g. degeneration of the SCN and decreased melatonin release) (Canevelli et al., 2016; Srinivasan, Pandi-Perumal, Cardinali, Poeggeler, & Hardeland, 2006; Stopa et al., 1999). Taken together, identifying at risk individuals with sleep disturbances early in life may present a relevant treatment window to slow or prevent pathological progression.

Recent evidence points to a glutamate hypothesis of AD, which is supported by the increased release and decreased clearance of synaptic glutamate (Revett, Baker, Jhamandas, & Kar, 2013; Tok, Ahnaou, & Drinkenburg, 2021). These alterations in glutamate levels may contribute to network hyperexcitability and, thereby, pathological development (Tok et al., 2021). Glutamate receptors have also been targeted for the treatment of AD. Aβ-mediated neurotoxicity has been hypothesized to occur via an NMDAR-activity dependent mechanism, and, thus, memantine, a non-competitive NMDAR antagonist, has been FDA-approved as a treatment for moderate to severe AD (Kumar et al., 2015; Miguel-Hidalgo, Alvarez, Cacabelos, & Quack, 2002; Rogawski & Wenk, 2003; Shankar et al., 2007). However, memantine is limited in that it only relieves symptoms and does not improve existing pathology (Matsunaga et al., 2018).

Several studies support targeting both group I and group II mGlu receptors for AD. Because mGlu₅ is coupled to the NMDAR (Niswender & Conn, 2010; Shigemoto et al., 1993), reducing mGlu₅ function may exert neuroprotective effects by blocking Aβ-mediated neurotoxicity, similar to memantine, with lesser adverse effects (Kumar et al., 2015). Furthermore, mGlu₅ has been shown to directly engage with Aβ oligomers (Aβo). One possible contributor to AD pathology is Aβo binding to cellular prion protein (PrP^c) which was shown in neuronal cultures to ultimately induce synaptic dysfunction (Laurén, Gimbel, Nygaard, Gilbert, & Strittmatter, 2009). Interaction of the Aβo-PrP^c complex with mGlu₅ induces downstream calcium mobilization and increases Aβ formation, and these effects may be associated with deficits in learning and memory (Hamilton, Esseltine, Devries, Cregan, & Ferguson, 2014; Um et al., 2013). In support of this interaction, both deletion of mGlu₅ and blockade of mGlu₅ with NAMs including MTEP in a mouse model of AD (APPswe/PS1ΔE9 mice) reversed deficits in spatial learning and working memory and prevented Aβo-associated changes in synaptic density (Budgett, Bakker, Sergeev, Bennett, & Bradley, 2022; Hamilton et al., 2014, 2016; Um et al., 2013). Lastly, there is preclinical support for both inhibition and activation of group II mGlu receptors. Support for inhibition as a therapeutic approach stems from findings that mGlu₂ receptor activation increased accumulation of Aβ42 in cortical synaptoneurosomes (Kim et al., 2010). Pharmacological activation of mGlu₂ receptors also contributed to microglial neurotoxicity, whereas blockade of mGlu₂ receptors reduced microglia apoptosis (Taylor, Diemel, Cuzner, & Pocock, 2002). Importantly, increased mGlu₂ receptor expression was found in hippocampal neurons in post-mortem tissue from patients with AD when compared to age-matched controls suggesting a close relationship between mGlu₂ and AD pathogenesis (Lee et al., 2004). Conversely, there is some support for mGlu_2/3 receptor activation for the treatment of AD. Caraci et al., 2011 found activation of both mGlu₂ and mGlu₃ receptors with LY379268 exerted neuroprotective effects, reducing Aβ-mediated neuronal death. Further, neuroprotective effects of LY379268 did not occur in cells lacking mGlu₃ receptors and selective activation of mGlu₂ receptors increased Aβ-mediated neurotoxicity suggesting a role for mGlu₃ receptor activation in these neuroprotective effects (Caraci et al., 2011; Caraci, Nicoletti, & Copani, 2018).

mGlu₅ NAMs may also hold promise for targeting sleep disturbances associated with AD. Aforementioned effects of mGlu₅ NAMs on sleep, primarily increases in NREM sleep and NREM sleep quality, could mitigate the accelerated Aβ deposition caused by sleep disruptions and, thereby, slow cognitive impairment and behavioral effects (Ahnaou, Langlois, et al., 2015; Ahnaou, Raeymaekers, et al., 2015; Cavas et al., 2013b). Furthermore, while speculative, arousal enhancing effects of some mGlu₅ NAMs may be favorable to target sundowning effects if administered at the correct time of day. For example, VU042438 has initial wake-promoting effects followed by compensatory increases in NREM sleep and increased NREM sleep quality (Holter et al., 2021). This profile could normalize sleep patterns that are perceived contributors to sundowning pathophysiology, serving to prevent daytime napping and increase nighttime sleep. However, further research is needed to determine if effects of mGlu₅ NAMs on sleep in healthy rodents will translate similarly to rodents in models of AD-pathology and humans with and without AD. Lastly, while both inhibition and activation of mGlu_2/3 receptors has been explored for the possible benefits in targeting AD pathology, the effects of these compounds on sleep in healthy animals (both decreased REM sleep) appear to counteract the desired effects in patients with Alzheimer’s disease. Nonetheless, examining these compounds on sleep in the context of AD is still important to acknowledge and address in future studies.

3.4.5. Substance use disorders

Current research suggests a bidirectional relationship between sleep and substance use/misuse. On one hand, sleep disturbances increase the risk of licit and illicit drug use which contributes to the subsequent development of substance use disorders (SUDs). For example, reduced sleep duration or quality as well as social jetlag (reduced sleep time on weekends) during adolescence is associated with increased problematic drinking and continued illicit substance use in adulthood (Dolsen & Harvey, 2017; Hasler & Pedersen, 2020; Roehrs et al., 2021; Troxel et al., 2021). Conversely, substance use also induces both acute and long-lasting sleep disruptions (Koob & Colrain, 2020; Roehrs et al., 2021; Valentino & Volkow, 2020). Although all illicit substances, as well as alcohol, cannabis, and nicotine alter sleep, specific disturbances and severity vary depending on a number of factors including, but not limited to, substance, duration, and frequency/typography of use. Chronic drug exposure associated with SUD often leads to long-lasting sleep disruptions including insomnia that last for weeks to months into abstinence. Importantly, sleep disturbances during abstinence have been correlated with increased craving and risk of relapse (Brower, 2003; Dolsen & Harvey, 2017; Lydon-Staley et al., 2017). Thus, as with other CNS disorders, targeting the sleep disruptions associated with SUD represents a promising treatment approach to not only prevent progression from recreational use to SUD but also to reduce relapse following abstinence. While most substances affect sleep (Hser et al., 2017; Jaehne, Loessl, Bárkai, Riemann, & Hornyak, 2009; Mondino et al., 2021), the following sections focus on relevant clinical and preclinical literature describing effects of cocaine and opioids on sleep.

3.4.5.1. Cocaine use disorder (CUD)

There is growing preclinical and clinical evidence for the presence of sleep disturbances throughout various stages of cocaine use disorder (CUD). As a stimulant, acute use of cocaine is associated with increased arousal and time awake and decreased total sleep time across species (Bjorness & Greene, 2018, 2021; Post, Gillin, Wyatt, & Goodwin, 1974; Schierenbeck, Riemann, Berger, & Hornyak, 2008). However, abstinence following chronic use is associated with persistent alterations in sleep. In general, polysomnography studies show a decrease in REM sleep on the first day of abstinence (Schierenbeck et al., 2008). Sleep disturbances further into abstinence include decreased total and SWS time, decreased REM latency, and increased REM duration, which may be a REM rebound effect (Bjorness & Greene, 2022; Schierenbeck et al., 2008). During early abstinence (<1 week), objective polysomnography and subjective self-report measures of sleep concurrently indicate decreased sleep time, increased sleep latency, and decreased sleep quality. Interestingly, although most patients continue to display insomnia-like symptoms on objective measures during later abstinence periods (>2 weeks), subjective reports do not align, as patients report either unchanged or improved sleep (Morgan et al., 2006; Morgan, Perry, Cho, Krystal, & D’Souza, 2006; Morgan & Malison, 2007; Pace-Schott et al., 2005). Importantly, the DSM-5 characterizes these sleep disturbances as a subset of stimulant-induced disorders, which are symptoms that last months to years into abstinence (APA, 2013).

Surprisingly, few studies to date have examined cocaine-associated sleep alterations in animal models of CUD. One nonhuman primate study using actigraphy-based assessments reported sleep fragmentation and decreased NREM sleep during abstinence following cocaine self-administration (Cortes et al., 2016). Further, only two rodent studies have used polysomnography to examine effects of repeated cocaine exposure on sleep. Compared to a drug-naïve baseline, Chen and colleagues reported sleep fragmentation and, in contrast to human reports, both decreased NREM and REM sleep following one week of cocaine self-administration (Chen etal., 2015). These effects were only in part recapitulated in a study examining sleep following one week of repeated, non-contingent administration of cocaine in mice (Bjorness & Greene, 2022). Thus, this suggests frequency, dose, duration, and route of administration (contingent vs non-contingent) may be influential factors. Lastly, reduced REM sleep was correlated with increased operant responding to a cocaine cue (incubation paradigm; a preclinical model of drug seeking) and restoring REM sleep reduced cue-associated responding (Chen et al., 2015). Thus, this study established a direct relationship between cocaine induced sleep-disruptions and future drug-related behaviors.

Impaired glutamate homeostasis has been widely implicated in CUD. The primary reinforcing effects of cocaine are attributed to the blockade of the dopamine transporter in the mesocorticolimbic dopamine pathway (Di Chiara & Imperato, 1988), and chronic cocaine self-administration is associated with a hypodopaminergic state (Goldstein & Volkow, 2002; Gould, Porrino, & Nader, 2012; Mateo, Lack, Morgan, Roberts, & Jones, 2005; Siciliano, Ferris, & Jones, 2016). However, cocaine self-administration also induces long-lasting glutamate-mediated increases in synaptic plasticity in mesocorticolimbic regions including the VTA, NAc, and PFC (Niedzielska-Andres et al., 2021; van Huijstee & Mansvelder, 2014). During abstinence, exposure to cocaine-paired cues increases cortical glutamate concentrations in the NAc, and this excessive glutamate-mediated response drives increased DA release (Korpi et al., 2015; Shin, Gadzhanova, Roughead, Ward, & Pont, 2016), a purported mechanism contributing to craving. Excessive cue-reactivity following extended periods of abstinence is associated with an accumulation of calcium-permeable AMPA receptors (CP-AMPARs; for review, see Wolf, 2016), and this process has been shown to be regulated via mGlu₁ receptor activity (McCutcheon et al., 2011). Importantly, Chen and colleagues (2015) demonstrated that altering sleep impacted CP-AMPARs and subsequent cocaine-related behaviors in rats. Specifically, cocaine-induced REM sleep disruptions correlated with the accumulation of CP-AMPARs and restoring REM sleep during abstinence reduced CP-AMPARs and cocaine cue-reactivity (Chen et al., 2015). Alterations in mGlu receptors are also present during abstinence following cocaine self-administration. Rats showed lower mGlu_2/3 receptor expression (Logan et al., 2020), density (Pomierny-Chamiolo, Miszkiel, Frankowska, & Mizera, 2017), and function (Moussawi et al., 2011) in the NAc as well as lower mGlu₅ receptor expression in regions including the NAc and striatum (Hao et al., 2010; Knackstedt, Trantham-Davidson, & Schwendt, 2014; Swanson, Baker, Carson, Worley, & Kalivas, 2001).

3.4.5.2. Opioid use disorder (OUD)

Understanding the direct effects of opioids on human sleep is complicated by a number of factors including the type/frequency of use (prescription use, misuse, OUD), the presence or absence of pain, and whether studies use subjective or objective measures. Studies conducted in healthy, pain-free humans reported that opioid administration increased lighter NREM sleep stages (N1 and N2) and decreased N3 and REM sleep (Dimsdale, Norman, DeJardin, & Wallace, 2007; Shaw, Lavigne, & Mayer, 2006). Prescription opioid-dependent users with chronic pain experience subjective and objective (actigraphy-based) decreases in sleep duration and sleep quality (Hartwell, Pfeifer, McCauley, Moran-Santa Maria, & Back, 2014), and, in some instances, prescription opioids have been shown to further exacerbate sleep disturbances associated with chronic pain (Robertson et al., 2016). Moreover, illicit opioid use may affect sleep differently than FDA-approved opioids for the treatment of pain and OUD based on differences in potency/efficacy, route of administration, and related pharmacokinetic factors. In studies that used the Pittsburg Sleep Quality Index (PQSI) to assess sleep, 70–84% of sampled participants in methadone maintenance therapy were poor sleepers, and PQSI scores correlated with methadone dosage (Hsu et al., 2012; Stein et al., 2004). These studies are not exhaustive but highlight the complexity of interpreting opioid-related effects on sleep and reiterate that sleep is negatively impacted across conditions.

Specifically, within the context of OUD, sleep impairments including insomnia are some of the most common complaints during acute (days to weeks) and protracted (weeks to months) withdrawal (Dunn, Huhn, Bergeria, Gipson, & Weerts, 2019). Other common symptoms reported during abstinence periods in non-treatment seeking individuals with OUD include restless sleep, fatigue, and unusual dreaming (Hartwell et al., 2014; Tripathi, Rao, Dhawan, Jain, & Sinha, 2020). For example, increased N1 and reduced N2 sleep was associated with relapse in heroin detoxification patients at a 6-month follow-up assessment (Rady, Mekky, Moulokheya, & Elsheshai, 2020). Sleep disruption can also lead to decreased pain tolerance, which may influence subsequent opioid use (Eacret, Veasey, & Blendy, 2020). Together, similar to cocaine, chronic opioid use is associated with long-lasting sleep alterations and influences the likelihood of relapse.

Animal studies provide the opportunity to directly assess the effects of opioids on sleep, controlling for a number of the aforementioned confounding variables associated with human studies. In rats, acute morphine administration reduced NREM and REM sleep, yet NREM, but not REM, sleep duration returned to normal within three consecutive days of administration (Khazan, Weeks, & Schroeder, 1967). Acute effects of opioids are largely consistent between humans and rodents, except that opioids increased delta power during NREM sleep in rodents whereas opioids decreased delta power during N3 sleep in humans (Gauthier, Guzick, Brummett, Baghdoyan, & Lydic, 2011; Shaw et al., 2006). Additionally, actigraphy-based assessments in non-human primates were used to examine acute effects of the FDA-approved treatments for OUD on sleep-like parameters including sleep latency and efficiency (Berro, Zamarripa, Talley, Freeman, & Rowlett, 2022). Interestingly, while methadone (full mu opioid receptor agonist) and buprenorphine (partial mu opioid receptor agonist) both increased and decreased sleep latency and efficiency, respectively, disruptive effects of buprenorphine were longer lasting. In contrast, naltrexone (mu opioid receptor antagonist) reduced sleep latency and increased efficiency (Berro et al., 2022). While these studies were conducted in non-opioid dependent monkeys and did not employ polysomnography, they elegantly distinguished effects of activating or inhibiting opioid receptors on sleep-like parameters.

Similar to CUD, few preclinical studies have examined effects of chronic opioid self-administration and abstinence on sleep/wake cycles. In the only study to our knowledge, rats given 6-hr access to self-administer heroin during the light cycle for 2 weeks demonstrated a complete reversal in sleep pattern compared to a saline control group. REM and NREM sleep were increased during the dark period but decreased during the light period (Coffey, Guan, Grigson, & Fang, 2016). While NREM sleep duration returned to baseline levels in both the light and dark period within three days of abstinence, REM sleep impairments in the dark period persisted. These findings suggest altered recovery duration and possible underlying neurobiological alterations following heroin self-administration. Future preclinical research is needed to understand, in a well-controlled setting, how chronic opioid self-administration affects sleep and how FDA-approved opioid treatments affect sleep during abstinence.

There are many known and posited mechanisms through which opioids affect sleep-wake regulation. Reinforcing effects (e.g. abuse liability) of opioids are attributed to activation of mu-opioid receptors (MORs). MORs are broadly distributed throughout the CNS, influencing reward, pain, stress reactivity, mood, cognition and sleep (for review, see Welsch, Bailly, Darcq, & Kieffer, 2020). Due to the broad receptor distribution, opioids likely affect sleep-wake regulation via multiple circuits including excitation of the mesolimbic dopamine system and direct activation of MORs on hypocretin-containing hypothalamic neurons (Eacret et al., 2020). Additionally, opioids influence the suprachiasmatic nucleus affecting arousal and circadian rhythm and disrupt sleep by inhibiting respiratory activity and contributing to sleep apnea (Bergum, Berezin, & Vigh, 2022; Rosen et al., 2019; Wang & Teichtahl, 2007). While these are noted mechanisms through which acute exposure may alter sleep-wake regulation, similar to CUD, neurobiological alterations spanning many of these circuits may contribute to the long-lasting sleep disturbances associated with OUD.

Extensive evidence implicates excessive glutamate associated with OUD as a contributor to opioid-associated withdrawal symptoms and risk for relapse. In fact, hyperglutamatergic function may be a critical component of withdrawal. Glutamate concentrations within the locus coeruleus (LC), a noradrenergic nucleus in the brainstem, have been shown to increase during naloxone-precipitated withdrawal in opioid-dependent rats (Feng, Zhang, Rockhold, & Ho, 1995) and acute injections of glutamate into the LC induced behavioral signs of withdrawal similar to naloxone in morphine-dependent animals (Medrano, Mendiguren, & Pineda, 2015; Sekiya, Nakagawa, Ozawa, Minami, & Satoh, 2004; Tokuyama, Wakabayashi, & Ho, 1995; Zhu, Rockhold, & Ho, 1998). Given that the LC is largely implicated in OUD and associated withdrawal, as well as a key contributor for wake-promoting circuits, reducing glutamatergic function may be a viable therapeutic approach for multiple symptoms of OUD during abstinence including sleep disturbances. Additionally, like cocaine, long-lasting glutamatergic-mediated synaptic plasticity changes likely contribute to drug craving and risk of relapse (van Huijstee & Mansvelder, 2014). While less well characterized, recent evidence shows increased CP-AMPARs in the NAc that contribute to drug seeking during abstinence following oxycodone self-administration, which is a finding similar to cocaine studies (Wong, Zimbelman, Milovanovic, Wolf, & Stefanik, 2022).

3.4.5.3. Modulating mGlu receptors as treatments for SUD

There are currently 3 FDA-approved pharmacological treatments for OUD and no pharmacological treatments for CUD. While each is successful in some regard, all have adverse risk factors and none directly target sleep disturbances. Decades of research have examined compounds targeting various mGlu receptor subtypes as treatments for SUD. For example, mGlu₅ antagonists/NAMs as well as mGlu_2/3 PAMs have shown promise in reducing cocaine and opioid-related behaviors in rodent and nonhuman primate models of drug taking or seeking (Amato et al., 2013; Gould et al., 2016; Gould, Felts, & Jones, 2017; Johnson & Lovinger, 2020; Keck et al., 2013; Kumaresan et al., 2009; Lee, Platt, Rowlett, Adewale, & Spealman, 2005; Lou, Chen, Liu, Ruan, & Zhou, 2014; Salling, Grassetti, Ferrera, Martinez, & Foltin, 2021; van der Kam, de Vry, & Tzschentke, 2007). Moreover, many of these compounds have demonstrated anxiolytic-like and antidepressant-like activity suggesting potential to alleviate anxiety/depressive symptoms often present during withdrawal (Gould et al., 2016). Lastly, cocaine abstinence-induced accumulation of CP-AMPARs can be restored followed administration of an mGlu₁ receptor agonist (McCutcheon et al., 2011), although this has not been examined in the context of sleep, nor extended to studies involving opioids.

A causal relationship has not yet been established between sleep disruptions and altered mGlu receptor function during abstinence following cocaine or opioid self-administration. However, based on KO and pharmacological studies in drug-naïve rodents (Tables 1 and 2), we can speculate on possible treatment approaches. Results from several studies suggest targeting mGlu₁ and mGlu_2/3 receptors for CUD and OUD may benefit both abstinence-related behaviors and abstinence-induced sleep disturbances including insomnia and decreases in REM sleep. For example, mGlu₁ agonists may alter CP-AMPARs during abstinence in part by altering sleep. CP-AMPARs have a strong role in regulating sleep and sleep-maintained synaptic homeostasis and, thus, balancing these may normalize the restorative benefits of sleep including memory consolidation (for review, see Shepherd, 2012 and Martin et al., 2019). Additionally, some novel mGlu_2/3 PAMs may be able to target aberrant decreases in NREM sleep associated with abstinence. While mGlu₅ NAMs are promising for reducing drug taking/seeking, these compounds may not produce the desired effects on sleep as it relates to findings in some, but not all, rodent and human studies (i.e. mGlu₅ NAMs reduce NREM and REM sleep when increased NREM and REM sleep is needed). However, conflicting evidence between clinical and preclinical studies that show increased and decreased REM sleep during abstinence, respectively, indicates the appropriate way to target REM sleep still needs further investigation. Additionally, this does not discount the ability of these compounds to normalize possible sleep disturbances associated with stimulant-induced disorders including anxiety and depression. Lastly, though there is minimal literature surrounding group III mGlu receptors and sleep, it is important to note that activating mGlu₇ and mGlu₈ receptors was effective in attenuating self-administration, reinstatement, and withdrawal as it relates to cocaine and opioids (Hajasova, Canestrelli, Acher, Noble, & Marie, 2018; Kahvandi et al., 2021; Li et al., 2013; Luessen & Conn, 2022; Pałucha-Poniewiera & Pilc, 2013; Salling et al., 2021). Together, these studies provide compelling rationale to further examine group I-III mGlu receptor compounds in multiple preclinical models of SUD, with a particular need to examine associated impacts on sleep disturbances.

4. Discussion

As described in this chapter and summarized in Tables 3 and 4, sleep disturbances including insomnia are prevalent in many neuropsychiatric disorders, and there is a clear bidirectional relationship between the two. Sleep disturbances can precede traditional symptom onset and/or correlate with symptom severity or relapse (e.g. schizophrenia, MDD, SUD, AD). Sleep disturbances may also contribute to or exacerbate neuropathology associated with disorders including AD. Although disrupted sleep is common in both subjective and objective reports, primary pharmacological treatments for neuropsychiatric disorders have historically disregarded sleep as a modifiable symptom. Importantly, the DSM-5 distinguished CNS disorders and sleep disturbances as separate diagnoses. While common sleep medications including benzodiazepines and z-drugs are often prescribed for many neuropsychiatric disorders, they all decrease arousal by increasing inhibitory GABA activity (Benca & Buysse, 2018). Thus, depending on the disorder, these compounds may not directly affect the disrupted neurobiology/circuits implicated in disrupted sleep-wake activity in a disease-specific manner. Given that many CNS disorders have seen a dearth of novel treatments, future advancements should investigate sleep modifying pharmacotherapies in parallel with strategies to mitigate other symptoms.

Table 3.

Objective measures of sleep disruptions in neuropsychiatric disorders.

Neuropsychiatric condition		NREM		REM					Awakenings		qEEG	Sleep spindles		Comorbidity
		Duration	Latency	Duration	latency	Density	Bout Length	%	WASO	Sleep fragmentation	NREM Delta	Number	Density	Sleep disorders
Schizophrenia 1,4–13		↓	↑	n.s	↓	↑	—	n.s	↑	↑	↓	↓	↓	↑
Major Depressive Disorder 1,14–24		↓	↑	↑	↓	↑	↑	↑	↑	↑	↕	↓	↓	↑
Post-traumatic stress disorder 25–39		↓ ; n.s	↑	↕ ; n.s	n.s	↑	↓	n.s	↑ ; n.s	—	↓	—	↑	↑
Alzheimer’s Disease 1,40–42		↓	↑	↓ ; n.s	↑	↓	—	↓	↑	↑	↓	↓	—	↑
Substance Use Disorders: Withdrawal and Abstinence	Cocaine (Early Abstinence) 2,3,43–50	↑	↓	↑	↑	—	—	↑	—	—	—	—	—	‐
	Cocaine (Later Abstinence) 2,3,43–49	↓	↑	↓	↑ ; n.s	—	—	↕	↑	↑	↓	—	—	-
	Opioid (w/o medication assistance) 2,3,50–57	↓	↑	↓	↑	—	—	—	—	—	↓	—	—	↑

n.s., not significant; —, not determined; qEEG, quantitative EEG; REM Density, number of REMbouts per minute of sleep; % REM, duration of REM out of total sleep time; WASO, time awake after sleep onset until final awakening; Early abstinence, <1 week; Later abstinence, >2 weeks.

¹Benca et al. (1992), ²Angarita, Emadi, Hodges, and Morgan (2016), ³Dolsen and Harvey (2017), ⁴Cohrs (2008), ⁵Gerstenberg et al. (2020), ⁶Göder et al. (2006), ⁷Hodges, Pittman, and Morgan (2017), ⁸Kaskie, Graziano, and Ferrarelli (2017), ⁹Keshavan et al. (1998), ¹⁰Manoach et al. (2004), ¹¹Manoach and Stickgold (2009), ¹²Manoach and Stickgold (2019), ¹³Sprecher et al. (2015), ¹⁴Armitage, Hudson, Trivedi, and Rush (1995), ¹⁵Armitage, Hoffmann, Fitch, Morel, and Bonato (1995), ¹⁶Bovy et al. (2021), ¹⁷Coble, Foster, and Kupfer (1976), ¹⁸Steiger, Pawlowski, and Kimura (2015), ¹⁹Steiger and Kimura (2010), ²⁰Steiger and Pawlowski (2019), ²¹Steiger et al. (2015), ²²Tesler et al. (2016), ²³Wichniak, Antczak, Wierzbicka, and Jernajczyk (2002), ²⁴Wichniak et al. (2013), ²⁵Baird et al. (2018), ²⁶Khazaie and Masoudi (2016), ²⁷Kobayashi et al. (2007), Kobayashi et al. (2012), ²⁹Krakow, Ulibarri, Moore, and McIver (2015), ³⁰Laniepce et al. (2020), ³¹Le Bon et al. (1997), ³²Leeman-markowski et al. (2020), ³³Lewis et al. (2020), ³⁴Mellman, Bustamante, Fins, Pigeon, and Nolan (2002), ³⁵Mellman, Pigeon, Nowell, and Nolan (2007), ³⁶Mellman, Kobayashi, Lavela, Wilson, and Hall Brown (2014), ³⁷Neylan et al. (2021), ³⁸Swift (2020), 39Woodward et al. (2002), ⁴⁰Bliwise (1993), ⁴¹Scullin and Bliwise (2015a, 2015b), ⁴²Zhang et al. (2022), ⁴³Bjorness and Greene (2021), ⁴⁴Chen et al. (2015), ⁴⁵Johanson, Roehrs, Schuh, and Warbasse (1999), ⁴⁶Matuskey, Pittman, Forselius, Malison, and Morgan (2011), ⁴⁷Morgan, Pace-Schott, et al. (2006), Morgan, Perry, et al., 2006, ⁴⁸Schierenbeck et al. (2008), ⁴⁹Valladares and Irwin (2007), ⁵⁰Chen, Ting, Wu, Lin, and Gossop (2017), ⁵¹Eckert and Yaggi (2022), ⁵²Gauthier et al. (2011), ⁵³Kay (1975), ⁵⁴Khazan and Colasanti (1972), ⁵⁵Rosen et al. (2019), ⁵⁶Shaw et al. (2006), ⁵⁷Wang and Teichtahl (2007).

Table 4.

Subjective assessments of sleep disruptions in neuropsychiatric disorders.

Neuropsychiatric condition		Sleep				Nightmares	Limb movement	Daytime napping
		Duration	Latency Quality		WASO
Schizophrenia 2–7		↕	↑	↓	↑	↑	—	↑
Major depressive disorder 8–11		↕	↑	↓	↑	—	—	↑
Post-traumatic stress disorder 12–16		↓	↑	↓	↑	↑	↑	—
Alzheimer’s disease 17–22		↓	↑	↓	↑	—	—	↑
Substance use disorders: Withdrawal and abstinence	Cocaine (Early abstinence) 1,23–27	↓	↑	↓	—	—	—	—
	Cocaine (Later abstinence) 1,23–27	↑	↓	↑	—	—	↑	—
	Opioid (w/o medication assistance) 1,28–30	↓	↑	↓	↑	↑	↑	↑

Abbreviations: n.s., not significant; —, not determined. WASO, time awake after sleep onset until final awakening, Early abstinence, <1 week; Later abstinence, >2 weeks.

¹Angarita et al. (2016), ²Alam and Chengappa (2011), ³Chiu, Ree, Janca, and Waters (2016), ⁴Chouinard, Poulin, Stip, and Godbout (2004), ⁵Dule, Ahmed, Tessema, and Soboka (2020), ⁶Göder et al. (2006), ⁷Manoach and Stickgold (2019), ⁸Ağargün, Kara, and Solmaz (1997), ⁹Lopez, Hoffmann, Armitage, and Are (2010), ¹⁰Maglione et al. (2014), ¹¹Steiger and Pawlowski (2019), ¹²Denis et al. (2021), ¹³Kobayashi et al. (2012), ¹⁴Westermeyer et al. (2010), ¹⁵Woodward et al. (2002), ¹⁶Yeh et al. (2021), ¹⁷Borges et al. (2021), ¹⁸Kim, Ahn, Min, Yoo, and Choi (2021), ¹⁹Kuo, Hsiao, Lo, and Nikolai (2021), ²⁰Li et al. (2022), ²¹Scullin and Bliwise (2015a, 2015b), ²²Zhang et al. (2022), ²³Hodges et al. (2017), ²⁴Matuskey et al. (2011), ²⁵Morgan, Perry, et al. (2006), ²⁶Morgan and Malison (2007), ²⁷Schierenbeck et al. (2008), ²⁸Chen et al. (2017), ²⁹Gupta, Ali, and Ray (2018), ³⁰Sharkey et al. (2011).

As highlighted throughout, glutamate is a key regulator of homeostatic sleep, and glutamate dysfunction is commonly associated with many CNS disorders. Thus, it is important to investigate how glutamate dysfunction may directly contribute to sleep disturbances associated with specific CNS disorders. This chapter provided an overview of known roles of glutamatergic and, more specifically, mGlu receptor regulation of the sleep-wake cycle. Key contributions of multiple mGlu receptor subtypes in sleep regulation are discussed alone and in context of multiple CNS disorders. This chapter also highlights current knowledge gaps and speculates on plausible mGlu receptor subtypes to target that mitigate both the primary symptoms of a specific disorder and the associated sleep disturbances. We argue that, while possible to identify a singular treatment to augment sleep across multiple conditions, given the heterogeneity of CNS disorders, unique treatments based on neurobiological alterations may have greater therapeutic efficacy across multiple symptoms.

As summarized in Table 1 and 2, much of our understanding of mGlu receptor contributions to sleep regulation stem from constitutive knockout studies or studies using acute pharmacological challenges with subtype selective (or preferring) compounds. While these studies are vital to our basic understanding, methodological considerations should be discussed to assist with accurate interpretation of findings. For example, the absence of altered baseline sleep between wild type mice and mice with genetically altered receptor expression/function should not entirely rule out the possible involvement of that receptor influencing sleep regulation. Given the highly conserved circuitry regulating the sleep-wake cycle, constitutive knockout may result in compensatory functional and circuit-level changes that mask subtle differences. Moreover, baseline sleep comparisons reflect normal sleep regulation but not the mechanisms regulating sleep following a perturbation (sleep restriction/deprivation, stressor, etc.). For example, mGlu₅ studies found disruptions in homeostatic sleep drive in KO animals following a sleep deprivation period (Ahnaou, Raeymaekers, et al., 2015; Holst et al., 2017). Importantly, while many studies have examined direct effects of mGlu receptor knockouts (notably mGlu_2/3, mGlu₅, mGlu₇) or altered downstream signaling (e.g. PLCβ4, a downstream effector of mGlu₁) as shown in Table 1, direct effects of mGlu₁, mGlu₄, and mGlu₈ have not yet been examined. Furthermore, recent advances in neuroscience methods afford the exciting opportunity to examine sleep-wake activity changes at baseline and in response to perturbations. These include studies with conditional genetic alterations in mice as well as circuit/cell-specific modulation via opto/chemogenetic techniques.

Similar methodological considerations should be reviewed when designing/interpreting pharmacological studies including time of dosing and selectivity of pharmacological effects. In rodent studies, compounds administered during the dark (active) phase can provide information regarding general sleep-promoting effects (or sedative effects based on quantitative assessments of spectral power distribution) whereas compound administration during the light (quiescent) phase may better align with investigating wake-promoting/sleep-disrupting effects, NREM:REM ratios/latencies, and delta power during NREM sleep. Given the known fluctuations in both neurotransmitters and receptor cycling across diurnal periods, compounds may have divergent profiles on sleep-wake duration and architecture depending on time of dosing. Examining pharmacological effects on rodent sleep-wake behavior in both periods is even more critical when considering that recently developed compounds with increased mGlu receptor subtype selectivity predominately involve allosteric modulators (for review, see Luessen & Conn, 2022). Unlike compounds that bind to the orthosteric site, that activate/inhibit receptor function regardless of endogenous physiological conditions, allosteric modulators require the presence of an endogenous agonist (e.g. glutamate) to augment intracellular signaling. In this regard, allosteric modulation may provide unprecedented temporal selectivity, effectively modulating signaling when endogenous glutamate is present at higher concentrations (wake and REM) without altering sleep/wake state when glutamate is lower (NREM). Importantly, most pharmacological studies included in this chapter examined acute but not chronic pharmacological effects on sleep-wake cycles. Many studies have also not examined effects of these compounds on sleep in higher order, gyrencephalic species with more analogous sleep-wake patterns to humans. Lastly, it must be noted that, given the absence of literature examining mGlu receptor modulation on sleep in preclinical models of a CNS disorder, we have speculated on most potential pharmacotherapeutic applications. Future studies are thus needed to expand on findings from pharmacological assessments in otherwise healthy animals and incorporate these methodologies into animal models of disease states and/or in clinical populations to confirm/refute these hypotheses.

In summary, normalizing sleep disturbances associated with CNS disorders represents a novel treatment approach. Improving sleep may serve as a critical component in the therapeutic outcome, and restoration of homeostatic sleep may have several downstream benefits for other primary symptoms of a CNS disorder. Recent development of mGlu receptor-subtype selective ligands presents an opportunity to further interrogate the role of group I, II and III mGlu receptor regulation of the sleep-wake cycle with separate or synergistic effects on other symptoms. In general, group I mGlu receptors appear to be involved in homeostatic sleep pressure in addition to influences in sleep duration (Aguilar et al., 2020; Martin et al., 2019) whereas group II receptors appear to be more so involved in cortical arousal and selective effects on REM sleep following activation (Ahnaou et al., 2009; Pritchett et al., 2015). Fewer studies have examined group III, though their role (or lack thereof) will likely be determined in future studies with the recent development of more subtype-selective ligands (Lin et al., 2022; Luessen & Conn, 2022).

Importantly, the potential for specific mGlu receptors and the associated pharmacological effect (activation or inhibition) to ameliorate sleep disturbances and primary symptoms varies across disorders. For example, where mGlu₅ agonists/PAMs may have pharmacotherapeutic relevance for sleep disturbances and symptoms in schizophrenia (Maksymetz et al., 2017), mGlu₅ antagonists/NAMs may be more relevant for MDD, PTSD, AD, and SUD (Hamilton et al., 2014; Luessen & Conn, 2022). Likewise, both activation and inhibition of mGlu_2/3 may be therapeutically relevant for MDD (Dogra & Conn, 2021) whereas activation may be most relevant with regard to schizophrenia (Maksymetz et al., 2017). The current, most effective therapies for both schizophrenia and depression influence sleep, predominantly through increasing sleep duration (antipsychotics; Cohrs, 2008) or selectively decreasing REM sleep (antidepressants; Steiger & Pawlowski, 2019; Wichniak et al., 2017). Thus, systematically examining sleep as a primary or secondary endpoint should be considered in the development of novel therapies. Recent preclinical studies investigating mGlu₅ NAMs and mGlu_2/3 PAMs have shown antidepressant-like effects, both in classic models and on sleep (Ahnaou et al., 2009; Gould et al., 2016; Holter et al., 2021; Lindemann et al., 2015). Several compounds targeting mGlu receptors have advanced to clinical trials for disorders including schizophrenia and MDD. For example, both mGlu₅ NAMs (Quiroz et al., 2016) and mGlu_2/3 antagonists (see Maksymetz et al., 2017 for discussion) have advanced to phase II for the treatment of MDD but were stalled due to a lack of significant effects on primary endpoints despite promising results on secondary endpoints. As multiple development programs progress forward with novel mGlu receptor subtype selective compounds for various indications, incorporating sleep measures into preclinical and clinical studies (although costly) should be considered. Lastly, while this chapter highlights the importance of mGlu receptor function on sleep in the context of schizophrenia, MDD, PTSD, SUD, and AD, there is impetus for targeting sleep disturbances as a treatment approach for numerous other neurodegenerative and neurodevelopmental disorders.