Unraveling the Mechanisms Underlying Disordered Sleep in Alzheimer's Disease

Alzheimer's disease (AD) is a neurodegenerative disorder characterized by the aggregation of β-amyloid and tau. While cognitive impairment is the primary clinical manifestation of AD, patients also experience noncognitive symptoms, such as disordered sleep. These include sleep loss and fragmentation, which are associated with enhanced cognitive decline among patients (Wang and Holtzman, 2020). Sleep microstructure is also disrupted in AD, manifest as deficits in spindles and their coupling with slow oscillations (Wang and Holtzman, 2020). Importantly, converging evidence supports a bidirectional relationship between sleep disturbances and AD pathology, where poor sleep can promote aggregation of amyloid and tau; and in turn, this enhanced pathology worsens sleep. Given the role of sleep in cognition, studying the impact of AD pathology on sleep is essential for understanding how cognitive processes are impaired in this disease.

Although the mechanisms underlying disordered sleep in AD remain unclear, degeneration of neural circuits involved in sleep-wake control could give rise to these symptoms. Consistent with this, lesions of sleep-promoting nuclei are associated with sleep loss in rodents (Lu et al., 2000) and AD patients (Wang and Holtzman, 2020). Furthermore, sleep-wake regulating structures, such as the locus coeruleus (LC) and hypothalamus, are the first brain regions to harbor tau pathology in healthy aging and AD (Kam et al., 2023). Temporally aligned with this early pathology, disordered sleep can emerge before cognitive symptoms of AD (Wang and Holtzman, 2020). Congruent findings from animal studies demonstrate that sleep-wake circuits are vulnerable to pathology based on their connectivity and proximity with the LC (Kam et al., 2023), and progression of tauopathy to sleep centers coincides with the onset of sleep symptoms in mouse models of AD (Wang and Holtzman, 2020). Conversely, sleep can also aggravate AD pathology: amyloid and tau levels fluctuate across the sleep-wake cycle (Wang and Holtzman, 2020), while restricting sleep aggravates neuropathology and accelerates disease progression in animal models (Zhu et al., 2018).

Given these findings suggesting a modulatory role for sleep in AD, sleep is a promising target for treating those with AD or protecting individuals at risk. One way to increase sleep is by antagonizing the wake-promoting effects of orexin using dual orexin receptor antagonists (DORAs). DORAs increase sleep and reduce amyloid accumulation in mouse models (Wang and Holtzman, 2020). These findings were recently corroborated in clinical trials evaluating the effects of DORAs in AD patients (Lucey et al., 2023). Despite our growing understanding of sleep in AD, the interplay between degeneration and sleep remains elusive. Furthermore, although DORAs have been evaluated in rodent AD models of amyloid accumulation, they have not been tested in tau-based models of AD. To address these knowledge gaps, Kam et al. (2023) recently used behavioral, electrophysiological, and pharmacological approaches to understand the link between tau pathology and sleep.

To assess the role of AD pathology in age-related sleep disturbances, the authors recorded sleep-wake behaviors in young and old mice overexpressing mutant tau (PS19 mice). Their experiments revealed that, with aging, PS19 mice had decreased amounts of rapid eye movement (REM) and non-REM (NREM) sleep, whereas these measures increased or remained unchanged among wild-type mice. This age-dependent sleep loss stemmed from shortened sleep bouts and a corresponding increase in arousal frequency. Together, these results suggest that progressive tau pathology in PS19 mice results in decreased maintenance of NREM and REM sleep, which promotes frequent arousal from sleep and fragments the sleep-wake cycle.

Examining characteristics of NREM sleep in greater detail, the authors found that PS19 mice had defective sleep microstructure, characteristic of AD. Specifically, both the age of mice and their genotype had significant effects on the density of spindles and their coupling with slow oscillations during NREM sleep, such that old PS19 mice had the lowest spindle densities and most deranged spindle-slow wave coupling. Moreover, reduced spindle duration and power were evident in young PS19 mice relative to age-matched controls, at a time when these mice do not display hippocampal degeneration or cognitive impairment. These findings collectively suggest that abnormal spindle dynamics can emerge early on in disease and persist until old age, mirroring the symptomatic progression in AD (Wang and Holtzman, 2020).

Having uncovered disturbances to NREM sleep, the authors next asked whether REM sleep was similarly disrupted. Indeed, not only did old PS19 mice exhibit REM loss and fragmentation, but one-third of these animals lost muscle atonia, the motor quiescence that defines healthy REM sleep. That is, REM episodes in PS19 mice had excessive and exaggerated movements, consisting of behaviors, such as grasping and chewing, which were correlated with elevations in muscle activity. Intriguingly, this phenotype is characteristic of REM sleep behavior disorder (RBD), a parasomnia linked with neurodegenerative diseases, similarly characterized by REM without atonia (Iranzo et al., 2013). In summary, PS19 mice show dysregulated sleep-wake physiology, with impaired quantities and qualities of sleep.

Last, the authors explored whether DORA-12 treatment could alleviate sleep disturbances in old PS19 mice. Consistent with previous literature, the authors found that DORA-12 administration increased NREM and REM amounts, thus ameliorating sleep loss. However, DORA-induced increases in sleep were at the expense of its continuity, with the drug shortening NREM and REM bouts relative to vehicle-treated mice. Nonetheless, DORA-12 increased the density and duration of spindles during NREM sleep, and blunted the heightened EMG activity during REM sleep. Together, these results demonstrate that DORA-12 improved sleep disturbances in PS19 mice by promoting sleep, restoring spindle dynamics during NREM, and reinstating muscle suppression during REM sleep, despite enhanced sleep fragmentation.

Previous work showed that PS19 mice exhibit sleep loss and fragmentation and that these deficits are associated with degeneration of NREM- and REM-promoting structures, including the parafacial zone and sublaterodorsal nucleus (SLD). Kam et al. (2023) reported additional changes in NREM microstructure and REM motor activity in PS19 mice, likewise attributable to age-dependent degeneration of sleep-wake circuits (Fig. 1A). We speculate that, at 2-3 months, most sleep-wake regulating neurons remain intact, whereas few vulnerable regions, such as the LC and hypothalamus, succumb to tau pathology. During healthy NREM sleep, LC inactivity is required for spindle maintenance (Antila et al., 2022) and pathogenic tau causes neuronal hyperexcitability before cell death (Crimins et al., 2012). The fact that spindles were the main feature of sleep affected in young mice suggests early involvement of the LC. Specifically, in young PS19 mice, tau pathology may lead to aberrant LC recruitment during NREM sleep, disrupting thalamocortical circuits where spindles are generated and sustained. In contrast, other regions involved in sleep-wake control remain relatively spared from pathology; this intact circuitry is what allows for preserved sleep architecture and REM atonia in young PS19 mice.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Proposed brain circuits involved in sleep disturbances of PS19 mice. A, In young PS19 mice, tau pathology predominantly affects wake-promoting structures including, but not limited to, the LC, lateral hypothalamus (LH), and tuberomammillary nucleus (TMN). This early tau pathology may lead to abnormal firing of wake-promoting cells during sleep, consequently inhibiting the thalamic reticular nucleus (TRN), thalamic relay cells, and cortical regions involved in spindle control. With age, tau pathology expands to more regions, including to thalamocortical circuits, NREM-promoting neurons of the ventrolateral preoptic area (VLPO) and parafacial zone (PZ), as well as REM-promoting SLD neurons. Correspondingly, pathology in these regions would result in exacerbation of spindle deficits, decreased and fragmented sleep, and RBD symptoms as seen in old PS19 mice. B, We propose that DORA-12 treatment reduces sleep disturbances in old PS19 mice by antagonizing orexin signaling onto wake-promoting circuits. This would facilitate the invigoration of surviving spindle- and sleep-generating neurons, thus increasing sleep amount, as well as, respectively, restoring spindles and muscle suppression during NREM and REM.

With age, tau pathology becomes widespread, reaching more sleep-wake circuits in PS19 mice (Fig. 1A). By 9 months, tau pathology in aforementioned REM- and NREM-regulating structures underlies their sleep loss phenotype. Namely, the SLD controls the timing of individual REM episodes and generates the defining electrophysiological features of REM sleep itself, such as muscle atonia (Saper et al., 2010). Correspondingly, loss of SLD function is associated with RBD in rodents (Shen et al., 2020) and humans (Iranzo et al., 2013). Therefore, a likely explanation for the RBD phenotype in old PS19 mice is tau-mediated degeneration of SLD cells, resulting in REM without atonia. Thalamocortical circuits also develop pathology in old mice, which could exacerbate spindle deficits. Since disruptions to spindles can heighten sensitivity to arousal during sleep (Antila et al., 2022), spindle dysregulation could further deplete and fragment sleep by promoting arousal.

Kam et al. (2023) demonstrated that DORA-12 improves sleep disturbances linked with tau pathology. We propose that DORAs primarily act by reducing excitatory orexinergic tone onto wake-promoting structures (Fig. 1B). Because wake- and sleep-promoting circuits reciprocally inhibit each other across the sleep-wake cycle (Saper et al., 2010), DORA-induced suppression of wake-promoting areas would disinhibit NREM and REM centers and reinforce sleep. Another effect of increased SLD activity after DORA treatment would be enhanced muscle suppression during REM sleep, mitigating RBD symptoms. In this vein, orexin blockade in arousal centers, which suppress thalamocortical spindle circuitry, would disinhibit the latter and improve spindle coordination during NREM. On the other hand, exacerbated sleep fragmentation by DORA-12 can be explained by orexin blockade on surviving wake-maintaining neurons, leading to increased transitioning between states. Indeed, DORAs increase, but fragment, sleep in rodents (Kam et al., 2023).

In conclusion, these insights by Kam et al. (2023) extend our understanding of the relationship between degeneration and sleep in AD. They demonstrate that tau pathology underlies breakdown of sleep-wake homeostasis in PS19 mice. We posit that tau pathology within corresponding structures for sleep-wake control drives such sleep disturbances; further histologic analyses could help stage pathology with symptoms across the lifespan of PS19 mice, and clarify why RBD symptoms are more severe in some mice than others. The therapeutic effects of DORA-12 suggest several additional directions for future study. Having shown that tau pathology disrupts sleep, and given the bidirectionality between sleep and pathology, the question follows whether DORAs modify progression of AD. Examining pathology following DORA treatment and testing DORAs in young mice for potential neuroprotective outcomes would therefore be of interest in subsequent investigations.