Chronic Sleep Disturbances Alters Sleep Structure and Tau Phosphorylation in AβPP/PS1 AD Mice and Their Wild-Type Littermates

Abstract

Background:

Emerging evidence indicates that sleep disorders are the common non-cognitive symptoms of Alzheimer’s disease (AD), and they may contribute to the pathogenesis of this disease.

Objective:

In this study, we aim to investigate the effect of chronic sleep deprivation (CSD) on AD-related pathologies with a focus on tau phosphorylation and the underlying DNA methylation regulation.

Methods:

AβPP^swe/PS1^ΔE9 AD mice and their wild-type (WT) littermates were subjected to a two-month CSD followed by electroencephalography and electromyography recording. The mice were examined for learning and memory evaluation, then pathological, biochemical, and epigenetic assessments including western blotting, immunofluorescence, dot blotting, and bisulfite sequencing.

Results:

The results show that CSD caused sleep disturbances shown as sleep pattern change, poor sleep maintenance, and increased sleep fragmentation. CSD increased tau phosphorylation at different sites and increased the level of tau kinases in AD and WT mice. The increased expression of cyclin-dependent kinase 5 (CDK5) may result from decreased DNA methylation of CpG sites in the promoter region of CDK5 gene, which might be associated with the downregulation of DNA methyltransferase 3A and 3B.

Conclusion:

CSD altered AD-related tau phosphorylation through epigenetic modification of tau kinase gene. The findings in this study may give insights into the mechanisms underlying the effects of sleep disturbances on AD pathology and provide new therapeutic targets for the treatment of this disease.

Keywords

Alzheimer’s disease DNA methylation epigenetics sleep disturbances tau pathology

Introduction

Sleep is a vital physiological process of the human body characterized by several cycles of different sleep stages, including rapid eye movement (REM) sleep and non-REM (NREM) sleep [1]. It is believed that good sleep may be beneficial to memory consolidation [2], while insufficient or irregular sleep is linked to neurodegenerative diseases such as Alzheimer’s disease (AD) [3, 4]. Epidemiological data show that up to 60% of AD patients have symptoms of sleep disturbances, which are associated with the severity of cognitive decline [5, 6]. AD pathologies can worsen sleep disorders by affecting sleep-related brain regions and nuclei. Studies have reported that the basal forebrain involved in sleep regulation can be affected by amyloid-β (Aβ) pathology in the early stage of Aβ deposition [7, 8], and the arousal-promoting system can be exposed to tau in Braak tau pathology 0 stage [9, 10]. The LC region involved in REM sleep regulation is also one of the first regions affected by tau pathology [8, 11]. A study using the AβPP/PS1 mouse model showed that diurnal fluctuation of brain interstitial fluid (ISF) Aβ was disrupted in the hippocampus when Aβ deposition developed, and Aβ₄₂ immunization normalized the sleep– wake cycle and the diurnal fluctuation of ISF Aβ [12]. Disruption of the sleep– wake cycle was also reported in 3xTg and 5xFAD AD animal models [13, 14]. In Tg2576 mice, REM sleep impairment may be associated with a decrease in pedunculopontine cholinergic neurons [15]. Autopsy studies have found that loss of optic nerve axons, especially loss of melanin ganglion cells which are photoreceptors that drive the circadian rhythm, is associated with Aβ deposition [16]. It may be one of the mechanisms of circadian rhythm disruption in AD patients. On the other hand, increasing evidence suggests that sleep disorders may increase the risk of AD and contribute to AD-related pathologies, such as Aβ deposition and tau phosphorylation [17 –20]. It has been reported that the sleep– wake cycle may regulate Aβ level and its clearance in the brain [21 –23], and sleep-dependent global brain activity is involved in the clearance of AD-related brain waste [24]. The sleep– wake cycle may also affect tau pathology [25, 26]. A study showed that brain ISF tau exhibited diurnal fluctuation, with higher levels when mice were awake, and the ISF tau level increased after manual sleep deprivation [27]. Moreover, increasing excitatory neuronal activity significantly increased ISF tau levels in mice, indicating that the sleep– wake cycle may regulate extracellular tau levels [28]. In human subjects, sleep is also linked to tau pathology. A study indicated that worse in-home sleep quality was significantly associated with higher cerebrospinal fluid tau levels [29], and spindle density during NREM stage 2 was negatively correlated with cerebrospinal fluid tau level in humans [30]. In animal models of AD, chronic sleep deprivation (CSD) may influence AD pathologies [25]. A study reported that CSD might aggravate memory impairment and increase Aβ level and tau phosphorylation in 3xTg AD mice [31]. Another study showed that CSD increased tau protein level and impaired synaptic function in 3xTg AD mice [32]. Our previous study and studies from others have also found that CSD might accelerate AD-related Aβ pathology [33 –35]. Moreover, CSD is also reported to aggregate pathological changes in other neurodegenerative disease, for example, in Parkinson’s disease CSD exacerbated PD-like phenotype such as motor deficits, striatal dopamine level reduction, and dopaminergic neuron degeneration [36].

The pathological hallmarks of AD are Aβ plaques and neurofibrillary tangles (NFTs). Aβ plaques are composed of Aβ deposition, degenerative neural structures, and activated microglia and astrocytes [37, 38]. The amyloid hypothesis proposes that Aβ derived from amyloid-β protein precursor (AβPP) after sequential cleavages by β- and γ-secretases plays a key role in neurodegeneration in AD [39]. NFTs are the intracellular aggregation of hyperphosphorylated tau caused by dysregulation of tau kinases, including glycogen synthase kinase-3β (GSK-3β) and cyclin-dependent-like kinase-5 (CDK5) [40, 41]. Our previous study reported that two-month CSD exacerbated learning and memory disability, altered AβPP processing, increased Aβ production and plaque deposition, and increased tau phosphorylation in AβPP^swe/PS1^ΔE9 AD mice [35]. In addition, CSD also caused neuronal mitochondrial damage and apoptosis in the hippocampus of experimental mice. Interestingly, all these behavioral, biochemical, and pathological changes were long-lasting and irreversible [35]. In our previous study mentioned above, the effect of CSD on Aβ pathology has been investigated. Studies and interventions involving tau pathology have nowadays attracted great attention [42]. Thus, in this study, we focus on the effect of CSD on tau phosphorylation in AD mice.

Non-genetic factors, such as hypoxia and sleep disorders, are thought to interact with AD pathological changes through epigenetic modifications. DNA methylation is a well-studied mechanism of epigenetic modification, which is mainly catalyzed by DNA methyltransferases (DNMTs) through transferring a methyl group to the DNA molecule on the C-5 position of the cytosine ring [43]. DNA methylation establishes at CpG islands in the gene promoter regions and may physically hinder the binding of transcriptional factors via methyl-CpG-binding proteins, which strongly represses the transcription [44]. Our group previously documented that chronic hypoxia may increase Aβ production through demethylation of γ-secretase by downregulating DNMT3B [45]. In this study, we investigated the impact of CSD on AD-related tau pathology and its underlying epigenetic mechanism. The findings in this study may give new insight into the epigenetic mechanisms of the effect of sleep disturbances on the pathogenesis of AD and provide novel targets for the future treatment of this disease.

Methods

Animals

AβPP^swe/PS1^ΔE9 AD mice were obtained from the Jackson Laboratory (No. 004462, Bar Harbor, MA, USA). This AD mouse model has been most widely used in the field of AD research. With the expression of a mouse/human AβPP (Mo/HuAβPP695swe) transgene, the mice are allowed to secrete human Aβ peptide and deposit it as plaques in the brain by six months of age. All mice were kept in a 12 h– 12 h light– dark cycle with lights on at 7:00 a.m., 22±1°C room temperature, and 50±10% relative humidity. Male AD and wild-type (WT) mice at 4–4.5 months were randomized into four groups with or without CSD: n = 13 in AD+CSD and WT+CSD, and n = 15 in AD+NSD and WT+NSD. The CSD mice were sleep-deprived by the modified multiple platform method for two month and then subjected to polygraphic recording of electroencephalography (EEG) and electromyography (EMG). The mice were also examined with behavioral tests for learning and memory evaluation (see Supplementary Material) and finally sacrificed for further experiments. The timeline of the whole experiment is summarized in Supplementary Figure 1. Animal care and experimental procedures were carried out in accordance with the Laboratory Animal Care Guidelines approved by the Dalian Medical University Institutional Animal Care Committee. All protocols were approved by the Dalian Medical University Institutional Animal Care Committee.

Fig. 1

Sleep– wake profile and power density changes in AD and WT mice after CSD. Total time spent in wakefulness, REM, and NREM sleep in 24 h and during 12 h dark and 12 h light sessions in AD and WT mice after CSD (A). Time course changes in wakefulness, REM, and NREM sleep (B). EEG power density of REM and NREM sleep during 12 h dark and 12 h light sessions (C). Data are shown as mean±standard error and analyzed by two-way ANOVA. n = 3 in AD+CSD and WT+CSD; n = 5 in AD+NSD and WT+NSD. (A) *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; and (B,C) & means AD+CSD versus AD+NSD, $ means AD+CSD versus WT+NSD, @ means AD+NSD versus WT+NSD, and # means WT+CSD versus WT+NSD. Exact p values are shown in Supplementary Tables 3 and 4.

Chronic sleep deprivation

AD and WT CSD mice were sleep-deprived by using the modified multiple platform method for two months, as described previously (from 12:00 p.m. to 8:00 a.m. the next day) [35, 46]. Briefly, CSD mice were kept on 24 small platforms (3 cm in diameter and 5 cm in height) in a tank with water 2 cm in depth and 1 cm beneath the platforms, while NSD mice were kept in a similar water tank with large platforms (11.5 cm in diameter). Food and water were available ad libitum throughout the study.

Polygraphic recording and sleep pattern analysis

Mice were arranged for polysomnographic recordings after simultaneous implantation of EEG and EMG electrodes. The detailed procedure of implant surgery and SleepSign software (Kissei Comtec, Nagano, Japan) usage has been described previously [47]. In brief, the mice were first anesthetized with an isoflurane/oxygen mixture (2–3%). After anesthesia, two EEG electrodes (stainless steel screws with about 1 mm in diameter) were implanted over the cortex through the skull at +1 mm from bregma, ±1.5 mm lateral from midline. Two insulated stainless steel, Teflon-coated wires were placed into bilateral trapezius muscles as EMG electrodes. After a recovery period for 10 days individually, the mice were habituated to the recording cable for four days. The polygraphic recordings were monitored in mice for 24 h by using SleepSign software (Kissei Comtec, Nagano, Japan). EEG data were analyzed for a full 24 h day– night cycle, and then separately investigated for dark and light phases.

Immunofluorescent staining

The procedures of immunofluorescent staining were carried out as described in our previous studies [48, 49]. Briefly, the mouse brains were fixed, dehydrated, and sliced. The brain slices were blocked and incubated at 4°C overnight with the primary antibodies as shown in Supplementary Table 1, and then incubated with the secondary antibody (Supplementary Table 1). Three randomly selected fields per slice were photographed. The images were then analyzed by ImageJ software.

Western blotting

The western blotting protocol was described in our previous studies [35, 45]. Briefly, hippocampal and cortex tissues were dissected and collected on ice, then sonicated and centrifuged. The supernatant was detected with protein concentration and used for further analysis. After SDS-PAGE gel running and transfer to the PVDF membrane, the membrane was blocked and incubated at 4°C overnight with the primary antibodies listed in Supplementary Table 1. Then the membrane was incubated with secondary antibodies (Supplementary Table 1). The protein bands on the membrane were then pictured and quantified with FluorChem Q system (ProteinSimple, California, USA).

Dot blotting

The procedures for dot blotting were carried out as our previous protocol [50]. Genome DNA was obtained from the hippocampal and cortex tissues using phenol-chloroform extraction and then serially diluted, denatured, and spotted on the nylon membrane. After ultraviolet cross-linked and blocking, the membrane was incubated at 4°C overnight with 5-methylcytidine antibody (Supplementary Table 1). Then the membrane was incubated with the secondary antibody (Supplementary Table 1). The dots were detected and quantified using FluorChem Q system (ProteinSimple, California, USA).

Bisulfite polymerase chain reaction and sequencing

Genome DNA extracted from the hippocampal and cortex tissues was also used for bisulfite polymerase chain reaction (PCR) and sequencing as described previously [50]. The genome DNA was bisulfited and amplified by PCR reaction with primers shown in Supplementary Table 2. After agarose gel electrophoresis and purification, the PCR products were inserted into a vector and amplified in E. coli for sequencing. The DNA methylation percentage of each CpG site within the amplified region was analyzed by an online software (https://services.ibc.uni-stuttgart.de/BDPC/BISMA/).

Statistical analyses

All statistical data in this study were shown as mean±standard error of the mean (SEM). Two-way analysis of variance (ANOVA) was used in the statistical analyses with Tukey’s multiple comparisons test for post hoc analyses. Statistical significance was determined when p values were less than 0.05. The n values in the text and the figure legends were the number of animals referred to the statistical analyses.

Results

Chronic sleep deprivation causes sleep EEG pattern and sleep structure changes in AD and WT mice

To investigate sleep EEG changes caused by CSD, we recorded 24 h ambulatory EEG in both AD and WT mice. As shown in Fig. 1A and 1B, we found a significant decrease in REM sleep in AD and WT mice after CSD. The AD and WT CSD mice exhibited 95.3% (p = 0.0043) and 93.7% (p = 0.0005) decreases in REM sleep compared to NSD mice, respectively, in 24 h. During the 12 h dark session, WT+CSD mice showed a 14.7% decrease in wakefulness (p = 0.002) and 2.36-fold increase in NREM sleep (p < 0.0001). REM sleep was decreased significantly in both AD and WT mice during the 12 h dark session (93.0% decrease in AD mice, p = 0.0095; 100% decrease in WT mice, p = 0.0041). During the 12 h light session, AD and WT mice showed 95.5% (p = 0.0373) and 92.4% (p = 0.0026) decreases in REM sleep, respectively. Moreover, REM and NREM power density showed that θ and δ power were decreased significantly in AD and WT mice after CSD (Fig. 1C). The exact p values for the results in Fig. 1B and 1C are presented in Supplementary Tables 3 and 4.

Furthermore, we analyzed the sleep– wake structure of AD and WT mice after CSD. We found that AD and WT CSD mice exhibited significant decreases in 128 and 256 s episodes of REM sleep (p = 0.0002 in AD mice, p < 0.0001 in WT mice in 128 s episode, p = 0.0002 in AD mice, p = 0.0258 in WT mice in 256 s episode in 12 h dark and p < 0.0001 in AD mice, p < 0.0001 in WT mice in 128 s episode, p < 0.0001 in AD mice, p < 0.0001 in WT mice in 256 s episode in 12 h light) and increases in 32 and 64 s episodes of NREM sleep during 12 h dark and 12 h light sessions (p = 0.0003 in AD mice, p < 0.0001 in WT mice in 32 s episode, p < 0.0001 in AD mice, p = 0.0006 in WT mice in 64 s episode in 12 h dark and p < 0.0001 in AD mice, p = 0.0036 in WT mice in 32 s episode, p < 0.0001 in AD mice, p = 0.0331 in WT mice in 64 s episode in 12 h light), implying the poor sleep maintenance in AD and WT mice after CSD (Fig. 2A). During 12 h dark and 12 h light sessions, the number of wake (p = 0.009 in 12 h dark and p = 0.0025 in 12 h light) and NREM (p = 0.0096 in 12 h dark and p = 0.0027 in 12 h light) sleep episodes increased, and the mean duration of wake (p < 0.0001 in 12 h dark and p = 0.0041 in 12 h light) was decreased in AD mice, implying sleep fragmentation after CSD. The number of wake (p = 0.015) and NREM sleep (p = 0.0164) episodes also increased during the 12 h dark session in WT mice after CSD (Fig. 2B). As shown in Fig. 2 C, AD+CSD mice showed increases in NREM– wake (p = 0.0005 in 12 h dark and p < 0.0001 12 h light) and wake– NREM (p = 0.0017 in 12 h dark and p < 0.0001 in 12 h light) transitions during 12 h dark and 12 h light sessions, indicating decreased sleep stability and increased arousal in sleep in AD mice after CSD. WT+CSD mice also showed increases in NREM– wake (p < 0.0018) and wake– NREM (p = 0.0075) transition during the 12 h dark session and NREM– wake transitions (p = 0.019) during the 12 h light session.

Fig. 2

Sleep– wake structure changes in AD and WT mice after CSD. Numbers of sleep bouts (A); episode number and mean duration (B); and transition number (C) in REM and NREM sleep during 12 h dark and 12 h light sessions in AD and WT mice after CSD. Data are shown as mean±standard error and analyzed by two-way ANOVA. n = 3 in AD+CSD and WT+CSD; n = 5 in AD+NSD and WT+NSD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Chronic sleep deprivation aggravates cognitive decline in AD mice and induces cognitive impairment in WT mice

To investigate whether CSD caused cognitive changes, we examined AD and WT mice with a battery of cognitive behavioral tests (Supplementary Figure 2). In the Morris water maze test, the AD and WT CSD mice showed increased escape latency during the training (shown in Supplementary Figure 2A and the exact p values are presented in Supplementary Table 5) and during the probe test, CSD mice spent less time staying in the target quadrant (Supplementary Figure 2B, p = 0.0106 in AD mice and p < 0.0001 in WT mice) with less count in the target position (Supplementary Figure 2C, p = 0.033 in AD mice) indicating impaired spatial learning and memory after CSD. The novel object recognition test (p = 0.0482 in AD mice and p = 0.0371 in WT mice) and Y-maze (p = 0.0308 in AD mice and p = 0.0001 in WT mice) further exhibited decreases in short-term recognition memory and working memory after CSD, respectively (Supplementary Figure 2D, E). The RotaRod test indicated that AD and WT mice had similar balance and motor abilities (Supplementary Figure 2F, G). The performance in cognitive tests was independent of the motor abilities of mice.

Chronic sleep deprivation results in an aggravation of AD-related Aβ and tau pathologies in AD and WT mice

To evaluate the changes in Aβ pathology after CSD, we further performed 6E10 immunostaining in the AD mouse hippocampus and cortex. As shown in Supplementary Figure 3, Aβ plaque deposition was significantly increased in the AD mouse hippocampus and cortex. Quantitation showed increased areas occupied by plaques (p = 0.0149 in hippocampus and p = 0.0437 in cortex). These results imply that CSD increased Aβ pathology in this AD mouse model.

As for tau pathology, we stained p-tau T181, T212, T231, and S396 in the hippocampus and cortex of both AD and WT mice. As shown in Fig. 3, tau phosphorylation at T181, T212, T231, and S396 was increased in the AD and WT mouse hippocampus and cortex after CSD (T181: p = 0.0058 in AD mice, p = 0.0086 in WT mice in hippocampus and p < 0.0001 in AD mice, p = 0.0162 in WT mice in cortex; T212: p = 0.01 in AD mice, p = 0.0096 in WT mice in hippocampus and p = 0.0029 in AD mice in cortex; T231: p = 0.0018 in AD mice, p = 0.0019 in WT mice in hippocampus and p = 0.005 in AD mice, p = 0.0015 in WT mice in cortex; S396: p = 0.0023 in AD mice, p = 0.0017 in WT mice in hippocampus and p = 0.0031 in AD mice in cortex). We further analyzed the protein levels of total tau (t-tau) (Fig. 4A,a) and phosphorylated tau (p-tau) at T181 (Fig. 4B,b), T212 (Fig. 4C,c), T231 (Fig. 4D,d), and S396 (Fig. 4E,e). The results show that t-tau level did not change, but p-tau was significantly increased at T181, T212, T231, and S396 sites in the cortex (T181: p = 0.0003 in AD mice, p = 0.004 in WT mice; T212: p = 0.0005 in AD mice, p < 0.0001 in WT mice; T231: p = 0.0006 in AD mice, p = 0.0366 in WT mice; S396: p < 0.0001 in AD mice, p = 0.0003 in WT mice) and T231 and S396 sites in the hippocampus (T231: p = 0.0367 in AD mice, p = 0.0358 in WT mice; S396: p = 0.0388 in AD mice) of AD and WT mice after CSD. Moreover, we tested protein levels of tau kinases CDK5 and GSK3β. We found that protein levels of CDK5 (p = 0.0238 in hippocampus, p < 0.0001 in cortex) and its activator p25 (p = 0.0123 in hippocampus, p < 0.0001 in cortex) were significantly induced in the AD mouse hippocampus and cortex and CDK5 (p = 0.0057) and p25 (p = 0.0003) were also increased in the WT mouse cortex after CSD (Fig. 5A,a1,a2). The level of GSK3β was also increased in the AD and WT mouse hippocampus and cortex (p = 0.0431 in AD mice, p = 0.0071 in WT mice in hippocampus and p < 0.0001 in AD mice, p = 0.0001 in WT mice in cortex), but the level of p-GSK3? at Y216 was not significantly changed after CSD (Fig. 5B,b1,b2). These results indicate that CSD might increase tau phosphorylation by upregulating the kinases CDK5 and GSK3β.

Fig. 3

Tau phosphorylation in hippocampus and cortex of mice after CSD. P-tau T181 staining and quantitation of positive staining in hippocampus and cortex of AD and WT mice (A,a). P-tau T212 staining and quantitation of positive staining in hippocampus and cortex of AD and WT mice (B,b). P-tau T231 staining and quantitation of positive staining in hippocampus and cortex of AD and WT mice (C,c). P-tau S396 staining and quantitation of positive staining in hippocampus and cortex of AD and WT mice (D,d). Data are shown as mean±standard error and analyzed by two-way ANOVA. n = 4. Scale bars: left, 50μm; right, 20μm. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Fig. 4

Protein levels of t-tau and p-tau in hippocampus and cortex of AD and WT mice after CSD. The protein levels were detected by western blotting: t-tau (A,a); p-tau at T181 (B,b); T212 (C,c); T231 (D,d); and S396 (E,e). Data are shown as mean±standard error and analyzed by two-way ANOVA. n = 3. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Chronic sleep deprivation reduces DNA methylation in the promoter of tau kinase gene through downregulating DNA methyltransferases in AD and WT mice

To investigate whether CSD increased tau phosphorylation and the levels of the kinases through epigenetic modifications, we assessed the whole genome DNA methylation and the methylation level of CpG sites in promoters of tau kinase genes. The results show that CSD decreased the whole genome DNA methylation (Fig. 5C) and also the methylation level of CpG sites in CDK5 promoter (Fig. 5D) in the AD and WT mouse hippocampus and cortex.

Fig. 5

Protein levels of tau kinases, whole genome DNA methylation, and methylation level of CpG sites in the promoters of tau kinase genes in hippocampus and cortex of AD and WT mice after CSD. The levels of CDK5 and its activator p25 (A,a1,a2) and the level of GSK3β and p-GSK3β (B,b1,b2) were detected by western blotting. The level of whole genome DNA methylation was detected by dot blotting of 5-methylcytosine antibody (C). The methylation level of CpG sites in the promoters of tau kinase genes CDK5 and GSK3B was detected by bisulfite DNA PCR and sequencing (D). Data are shown as mean±standard error and analyzed by two-way ANOVA. n = 3 in (A,B,C), n = 4 in (D). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

DNMTs are mainly responsible for DNA methylation. We further determined the protein levels of DNMT1 (Fig. 6A,a), DNMT3A (Fig. 6B,b), and DNMT3B (Fig. 6C,c). We found that DNMT3A and -3B levels were decreased significantly in the AD and WT mouse hippocampus and cortex after CSD (DNMT3A: p = 0.0069 in AD mice, p = 0.0113 in WT mice in hippocampus and p = 0.0125 in AD mice, p = 0.0108 in WT mice in cortex; DNMT3B: p = 0.0090 in AD mice, p = 0.0411 in WT mice in hippocampus and p = 0.0271 in AD mice, p = 0.0266 in WT mice in cortex). These results indicate that CSD might increase the tau kinase CDK5 protein level through reducing the methylation level of CpG sites in CDK5 promoter by downregulating DNMT3A and -3B.

Fig. 6

Protein levels of DNMTs in hippocampus and cortex of AD and WT mice after CSD. The protein levels were detected by western blotting: DNMT1 (A,a); DNMT2 (B,b); DNMT3A (C,c); and DNMT3B (D,d). Data are shown as mean±standard error and analyzed by two-way ANOVA. n = 3. *p < 0.05, **p < 0.01.

Discussion

In the present study, we applied two-month CSD in AβPP/PS1 AD mice and their WT littermates, followed by EEG and EMG recording, behavioral, pathological, biochemical, and epigenetic tests. The results showed that CSD changed sleep patterns and decreased sleep quality, sleep maintenance, sleep stability, and increased arousal in sleep and fragmentation of sleep in AD and WT mice. CSD also increased Aβ plaques and tau phosphorylation at T181, T212, T231, and S396 sites, but there were differences observed between hippocampus and cortex in tau phosphorylation at T181 and T212. This might be associated with different response to CSD since there are diversity of sleep patterns in different brain regions [51] or different brain regions have a dissimilar development of tau pathology. The protein levels of the tau kinases, CDK5 and GSK3β were induced after CSD, along with decreased DNA methylation in the promoter region of the CDK5 gene. This DNA methylation change was associated with decreased levels of DNMT3A and 3B, indicating that CSD may aggravate tau phosphorylation through the regulation of DNA methylation.

Acute sleep deprivation (ASD) and CSD could both contribute to the pathological changes of AD. Studies have indicated that three-day to one-week ASD could alter neuronal activities and oscillations, change synaptic plasticity and signaling, and cause memory impairment in different animal models [52 –54]. However, a study in C57BL6 mice found that tau was hyperphosphorylated during sleep but not during 6 h acute sleep deprivation due to inhibition of protein phosphatase 2A by hypothermia during sleep [55]. In human subjects, one-night or one-week ASD affected declarative and emotional memory consolidation, increased false memory, and caused slow cognitive performance [56 –59]. One-night ASD also influenced Aβ clearance and increased Aβ burden in hippocampus of human subjects [60].

Epigenetic modifications, such as DNA methylation and histone modification, altering DNA accessibility and chromatin structure, change heritable and non-heritable traits without altering the DNA sequence, thereby regulating patterns of gene expression [61]. It has been reported that epigenetic mechanisms are associated with cognitive impairment and AD-related pathologies [62, 63]. Alteration in CpG methylation and transcriptional changes in DNMTs were found as a result of Aβ₄₂ accumulation in 5xFAD mice, and there were also changes in histone methyltransferases and histone deacetylase, which were related to cognition and memory [62]. Moreover, DNA methylation is also reported to be associated with tau pathology and modulates tau-induced neurotoxicity [64, 65]. In the present study, we found that two-month CSD increased protein levels of tau kinases GSK3β and CDK5, with a reduction of whole genome DNA methylation and a decreased methylation level of CpG sites in the promoter of CDK5 gene. However, p-GSK3β level was not significantly changed after CSD. CSD might be associated with the expression but not the post-translational modification of this kinase. Increased protein level of GSK-3β without changes of methylation at CpG sites in the promoter might result from other mechanisms that needed further investigation. Interestingly, our data show similar results in both AD and WT mice to some extent. Changes in WT mice without Aβ-related gene mutation might demonstrate the effects of sleep disturbances in sporadic AD. At the same time, DNA methylation is also considered as a mechanism for age-related circadian changes. Alterations in CpG methylation profiles are now considered as a canonical epigenetic feature of aging [66, 67]. There might be a potential interaction between DNA methylation and circadian rhythm. A study found that CpG methylation affected the expression of clock genes in cancer cells, which indicates that CpG methylation might exert a function within the mammalian circadian oscillator [68]. Researchers also found that infusion of a methyltransferase inhibitor into the suprachiasmatic nucleus (SCN) suppressed circadian changes, and SCN utilized DNA methylation as a mechanism to drive circadian clock plasticity [69]. Our recent study showed that CSD caused abnormal expression of clock genes including Bmal1, Clock, and Cry1 in the circadian rhythm-related nuclei, which was associated with tau pathology [46]. Therefore, we hypothesize that changes in sleep patterns after CSD may also result from changes in DNA methylation levels. Epigenetic mechanisms of sleep pattern changes after CSD need further investigation. The findings of the current study provide new evidence for the potential link between sleep disturbances, DNA methylation, and tau pathology and provide new insights into the pathogenesis of sporadic AD. In sporadic AD, non-genetic factors such as sleep disorders might also play a crucial role in the pathogenesis of this disease.

During different sleep stages, studies have reported that NREM sleep was associated with tau pathology [30, 70], and a role of REM sleep was also recently identified [71]. In the present study, the multiple platform method was used, which was designed to mainly deprive animals of REM sleep but also affected NREM sleep [72]. We found that CSD caused sleep disturbances such as increased sleep fragmentation and poor sleep maintenance. CSD led to changes in sleep patterns and structure rather than simple deprivation of REM sleep. Moreover, the water environment when using the modified multiple platform method and sleep deprivation itself are stressful to mice. Although NSD mice were placed in a similar water environment, the effect of stress could not be excluded. Chronic stress is also a risk factor for aging and AD; it causes synaptic loss and neuroinflammation in brain regions affected by AD pathology [73]. The main limitation of the current study is the lack of rescue experiment, we plan to upregulate DNMT3A and -3B to increase DNA methylation and decrease CDK5 level in future investigation.

In summary, our data show that CSD caused sleep disruptions and aggravated cognitive impairments in AD mice. CSD increased tau kinase expression through epigenetic modifications by downregulating DNMT3A and -3B, resulting in an increase in tau phosphorylation. Our results reveal an epigenetic mechanism underlying the effect of non-genetic risk factors on AD pathogenesis and might provide new therapeutic targets for future AD treatment.

Footnotes

ACKNOWLEDGMENTS

The authors have no acknowledgments to report.

FUNDING

This work was supported in part by funding from Shanghai municipal central government funds for guiding local scientific and technological development (YDZX20213100001002).

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

DATA AVAILABILITY

All data generated from the experiments of this study are included in the manuscript and its supplementary materials.

The supplementary material is available in the electronic version of this article: .

References

Tamaki

, Wang

, Barnes-Diana

, Guo

, Berard

, Walsh

, Watanabe

, Sasaki

(2020) Complementary contributions of non-REM and REM sleep to visual learning. Nat Neurosci 23, 1150–1156.

Zeng

, Lin

, Wang

, Hu

(2021) Sleep’s short-term memory preservation and long-term affect depotentiation effect in emotional memory consolidation: Behavioral and EEG evidence. Sleep 44, zsab155.

Falgas

, Walsh

, Neylan

, Grinberg

(2021) Deepen into sleep and wake patterns across Alzheimer’s disease phenotypes. Alzheimers Dement 17, 1403–1406.

Liu

, Niu

, Liu

, Cheng

, Le

(2021) Recent progress in non-motor features of Parkinson’s disease with a focus on circadian rhythm dysregulation. Neurosci Bull 37, 1010–1024.

Zhang

, Zhong

, Li

, Chang

, Le

(2017) The missing link between sleep disorders and age-related dementia: Recent evidence and plausible mechanisms. J Neural Transm (Vienna) 124, 559–568.

Proserpio

, Arnaldi

, Nobili

(2018) Integrating sleep and Alzheimer’s disease pathophysiology: Hints for sleep disorders management. J Alzheimers Dis 63, 871–886.

, Chung

, Zhang

, Zhong

, Ma

, Chang

, Weissbourd

, Sakai

, Luo

, Nishino

, Dan

(2015) Basal forebrain circuit for sleep-wake control. Nat Neurosci 18, 1641–1647.

Goedert

(2015) Neurodegeneration. Alzheimer’s and Parkinson’s diseases: The prion concept in relation to assembled Abeta, tau, and alpha-synuclein. Science 349, 1255555.

Braak

, Thal

, Ghebremedhin

, Del Tredici

(2011) Stages of the pathologic process in Alzheimer disease: Age categories from 1 to 100 years. J Neuropathol Exp Neurol 70, 960–969.

10.

Stratmann

, Heinsen

, Korf

, Del Turco

, Ghebremedhin

, Seidel

, Bouzrou

, Grinberg

, Bohl

, Wharton

, den Dunnen

, Rub

(2016) Precortical phase of Alzheimer’s disease (AD)-related tau cytoskeletal pathology. Brain Pathol 26, 371–386.

11.

Fraigne

, Torontali

, Snow

, Peever

(2015) REM sleep at its core - circuits, neurotransmitters, and pathophysiology. Front Neurol 6, 123.

12.

Roh

, Huang

, Bero

, Kasten

, Stewart

, Bateman

, Holtzman

(2012) Disruption of the sleep-wake cycle and diurnal fluctuation of beta-amyloid in mice with Alzheimer’s disease pathology. Sci Transl Med 4, 150ra122.

13.

Sterniczuk

, Dyck

, Laferla

, Antle

(2010) Characterization of the 3xTg-AD mouse model of Alzheimer’s disease: Part 1. Circadian changes. Brain Res 1348, 139–148.

14.

Sethi

, Joshi

, Webb

, Beckett

, Donohue

, Murphy

, O’Hara

, Duncan

(2015) Increased fragmentation of sleep-wake cycles in the 5XFAD mouse model of Alzheimer’s disease. Neuroscience 290, 80–89.

15.

Zhang

, Veasey

, Wood

, Leng

, Kaminski

, Leight

, Abel

, Lee

, Trojanowski

(2005) Impaired rapid eye movement sleep in the Tg2576 APP murine model of Alzheimer’s disease with injury to pedunculopontine cholinergic neurons. Am J Pathol 167, 1361–1369.

16.

La Morgia

, Ross-Cisneros

, Koronyo

, Hannibal

, Gallassi

, Cantalupo

, Sambati

, Pan

, Tozer

, Barboni

, Provini

, Avanzini

, Carbonelli

, Pelosi

, Chui

, Liguori

, Baruzzi

, Koronyo-Hamaoui

, Sadun

, Carelli

(2016) Melanopsin retinal ganglion cell loss in Alzheimer disease. Ann Neurol 79, 90–109.

17.

Insel

, Mohlenhoff

, Neylan

, Krystal

, Mackin

(2021) Association of sleep and beta-amyloid pathology among older cognitively unimpaired adults. JAMA Netw Open 4, e2117573.

18.

Grydeland

, Sederevicius

, Wang

, Bartres-Faz

, Bertram

, Dobricic

, Duzel

, Ebmeier

, Lindenberger

, Nyberg

, Pudas

, Sexton

, Sole-Padulles

, Sorensen

, Walhovd

, Fjell

(2021) Self-reported sleep relates to microstructural hippocampal decline in beta-amyloid positive adults beyond genetic risk. Sleep 44, .zsab110.

19.

Musiek

, Xiong

, Holtzman

(2015) Sleep, circadian rhythms, and the pathogenesis of Alzheimer disease. Exp Mol Med 47, e148.

20.

Winer

, Morehouse

, Fenton

, Harrison

, Ayangma

, Reed

, Kumar

, Baker

, Jagust

, Walker

(2021) Tau and beta-amyloid burden predict actigraphy-measured and self-reported impairment and misperception of human sleep. J Neurosci 41, 7687–7696.

21.

Kang

, Lim

, Bateman

, Lee

, Smyth

, Cirrito

, Fujiki

, Nishino

, Holtzman

(2009) Amyloid-beta dynamics are regulated by orexin and the sleep-wake cycle. Science 326, 1005–1007.

22.

Huang

, Potter

, Sigurdson

, Santacruz

, Shih

, Ju

, Kasten

, Morris

, Mintun

, Duntley

, Bateman

(2012) Effects of age and amyloid deposition on Abeta dynamics in the human central nervous system. Arch Neurol 69, 51–58.

23.

Xie

, Kang

, Xu

, Chen

, Liao

, Thiyagarajan

, O’Donnell

, Christensen

, Nicholson

, Iliff

, Takano

, Deane

, Nedergaard

(2013) Sleep drives metabolite clearance from the adult brain. Science 342, 373–377.

24.

Han

, Chen

, Belkin-Rosen

, Gu

, Luo

, Buxton

, Liu

, Alzheimer’s Disease Neuroimaging Initiative (2021) Reduced coupling between cerebrospinal fluid flow and global brain activity is linked to Alzheimer disease-related pathology. PLoS Biol 19, e3001233.

25.

Wang

, Holtzman

(2020) Bidirectional relationship between sleep and Alzheimer’s disease: Role of amyloid, tau, and other factors. Neuropsychopharmacology 45, 104–120.

26.

Liguori

, Spanetta

, Izzi

, Franchini

, Nuccetelli

, Sancesario

, Di Santo

, Bernardini

, Mercuri

, Placidi

(2020) Sleep-wake cycle in Alzheimer’s disease is associated with tau pathology and orexin dysregulation. J Alzheimers Dis 74, 501–508.

27.

Holth

, Fritschi

, Wang

, Pedersen

, Cirrito

, Mahan

, Finn

, Manis

, Geerling

, Fuller

, Lucey

, Holtzman

(2019) The sleep-wake cycle regulates brain interstitial fluid tau in mice and CSF tau in humans. Science 363, 880–884.

28.

Yamada

, Holth

, Liao

, Stewart

, Mahan

, Jiang

, Cirrito

, Patel

, Hochgrafe

, Mandelkow

, Holtzman

(2014) Neuronal activity regulates extracellular tau}. J Exp Med 211, 387–393.

29.

, Ooms

, Sutphen

, Macauley

, Zangrilli

, Jerome

, Fagan

, Mignot

, Zempel

, Claassen

, Holtzman

(2017) Slow wave sleep disruption increases cerebrospinal fluid amyloid-beta levels. Brain 140, 2104–2111.

30.

Kam

, Parekh

, Sharma

, Andrade

, Lewin

, Castillo

, Bubu

, Chua

, Miller

, Mullins

, Glodzik

, Mosconi

, Gosselin

, Prathamesh

, Chen

, Blennow

, Zetterberg

, Bagchi

, Cavedoni

, Rapoport

, Ayappa

, de Leon

, Petkova

, Varga

, Osorio

(2019) Sleep oscillation-specific associations with Alzheimer’s disease CSF biomarkers: Novel roles for sleep spindles and tau. Mol Neurodegener 14, 10.

31.

Rothman

, Herdener

, Frankola

, Mughal

, Mattson

(2013) Chronic mild sleep restriction accentuates contextual memory impairments, and accumulations of cortical Abeta and pTau in a mouse model of Alzheimer’s disease. Brain Res 1529, 200–208.

32.

Di Meco

, Joshi

, Pratico

(2014) Sleep deprivation impairs memory, tau metabolism, and synaptic integrity of a mouse model of Alzheimer’s disease with plaques and tangles. Neurobiol Aging 35, 1813–1820.

33.

Chen

, Huang

, Yang

, Zeng

, Zhang

, Wang

, Chen

, Li

, Zhang

(2017) Sleep deprivation accelerates the progression of alzheimer’s disease by influencing Abeta-related metabolism. Neurosci Lett 650, 146–152.

34.

Minakawa

, Miyazaki

, Maruo

, Yagihara

, Fujita

, Wada

, Nagai

(2017) Chronic sleep fragmentation exacerbates amyloid beta deposition in Alzheimer’s disease model mice. Neurosci Lett 653, 362–369.

35.

Qiu

, Zhong

, Liu

, Zhang

, Li

, Le

(2016) Chronic sleep deprivation exacerbates learning-memory disability and Alzheimer’s disease-like pathologies in AbetaPP(swe)/PS1(DeltaE9) mice. J Alzheimers Dis 50, 669–685.

36.

Liu

, Yu

, Wang

, Li

, Cheng

, Al-Nusaif

, Le

(2022) Altered motor performance, sleep EEG, and Parkinson’s disease pathology induced by chronic sleep deprivation in Lrrk2(G2019S) mice. Neurosci Bull 38, 1170–1182.

37.

Fakhoury

(2018) Microglia and astrocytes in Alzheimer’s disease: Implications for therapy. Curr Neuropharmacol 16, 508–518.

38.

Murphy

, LeVine

3rd , (2010) Alzheimer’s disease and the amyloid-beta peptide. J Alzheimers Dis 19, 311–323.

39.

Karran

, De Strooper

(2022) The amyloid hypothesis in Alzheimer disease: New insights from new therapeutics. Nat Rev Drug Discov 21, 306–318.

40.

Iqbal

, Liu

, Gong

(2016) Tau and neurodegenerative disease: The story so far. Nat Rev Neurol 12, 15–27.

41.

Giovannini

, Smeralda

, Jouanne

, Sopkova-de Oliveira Santos

, Catto

, Voisin-Chiret

(2022) Tau protein aggregation: Key features to improve drug discovery screening. Drug Discov Today 27, 1284–1297.

42.

Guo

, Li

, Zeng

L-H

, Tan

(2022) Tau-targeting therapy in Alzheimer’s disease: Critical advances and future opportunities. Ageing Neurodegener Dis 2, 11.

43.

Jones

(2012) Functions of DNA methylation: Islands, start sites, gene bodies and beyond. Nat Rev Genet 13, 484–492.

44.

Ehrlich

, Lacey

(2013) DNA methylation and differentiation: Silencing, upregulation and modulation of gene expression. Epigenomics 5, 553–568.

45.

Liu

, Qiu

, Yang

, Ni

, Le

(2016) Chronic hypoxia facilitates Alzheimer’s disease through demethylation of gamma-secretase by downregulating DNA methyltransferase 3b. Alzheimers Dement 12, 130–143.

46.

Niu

, Zhang

, Xu

, Yang

, Li

, Liu

, Le

(2021) Chronic sleep deprivation altered the expression of circadian clock genes and aggravated Alzheimer’s disease neuropathology. Brain Pathol 32, e13028.

47.

, Li

, Luo

, Wang

, Chen

, Qu

, Huang

(2015) Gelsemine alleviates both neuropathic pain and sleep disturbance in partial sciatic nerve ligation mice. Acta Pharmacol Sin 36, 1308–1317.

48.

Zhang

, Zhong

, Qi

, Li

, Cheng

, Liu

, Le

(2018) Impacts of acute hypoxia on Alzheimer’s disease-like pathologies in APP(swe)/PS1(dE9) mice and their wild type littermates. Front Neurosci 12, 314.

49.

Zhang

, Zhong

, Li

, Fu

, Cheng

, Cai

, Le

(2017) Acute hypoxia induced an imbalanced M1/M2 activation of microglia through NF-kappaB signaling in Alzheimer’s disease mice and wild-type littermates. Front Aging Neurosci 9, 282.

50.

Zhong

, Zhang

, Yang

, Li

, Xu

, Lan

, Cyganek

, El-Battrawy

, Zhou

, Akin

, Borggrefe

(2022) Epigenetic mechanism of L-type calcium channel beta-subunit downregulation in short QT human induced pluripotent stem cell-derived cardiomyocytes with CACNB2 mutation. Europace 24, 2028–2036.

51.

Petrovic

, Radovanovic

, Saponjic

(2021) Diversity of simultaneous sleep in the motor cortex and hippocampus in rats. J Sleep Res 30, e13090.

52.

Duan

, Liu

, Wang

, Wu

, Gao

, Zhang

(2016) Histone acetylation regulation in sleep deprivation-induced spatial memory impairment. Neurochem Res 41, 2223–2232.

53.

Havekes

, Park

, Tudor

, Luczak

, Hansen

, Ferri

, Bruinenberg

, Poplawski

, Day

, Aton

, Radwanska

, Meerlo

, Houslay

, Baillie

, Abel

(2016) Sleep deprivation causes memory deficits by negatively impacting neuronal connectivity in hippocampal area CA1. Elife 5, e13424.

54.

Tabassum

, Misrani

, Tabassum

, Ahmed

, Yang

, Long

(2021) Disrupted prefrontal neuronal oscillations and morphology induced by sleep deprivation in young APP/PS1 transgenic AD mice. Brain Res Bull 166, 12–20.

55.

Guisle

, Gratuze

, Petry

, Morin

, Keraudren

, Whittington

, Hebert

, Mongrain

, Planel

(2020) Circadian and sleep/wake-dependent variations in tau phosphorylation are driven by temperature. Sleep 43, zsz266.

56.

Wiesner

, Pulst

, Krause

, Elsner

, Baving

, Pedersen

, Prehn-Kristensen

, Goder

(2015) The effect of selective REM-sleep deprivation on the consolidation and affective evaluation of emotional memories. Neurobiol Learn Mem 122, 131–141.

57.

Nir

, Andrillon

, Marmelshtein

, Suthana

, Cirelli

, Tononi

, Fried

(2017) Selective neuronal lapses precede human cognitive lapses following sleep deprivation. Nat Med 23, 1474–1480.

58.

, Chong

, Ganesan

, Leong

, Chee

(2016) Sleep deprivation increases formation of false memory. J Sleep Res 25, 673–682.

59.

, Bennion

, Chee

(2016) Sleep restriction can attenuate prioritization benefits on declarative memory consolidation. J Sleep Res 25, 664–672.

60.

Shokri-Kojori

, Wang

, Wiers

, Demiral

, Guo

, Kim

, Lindgren

, Ramirez

, Zehra

, Freeman

, Miller

, Manza

, Srivastava

, De Santi

, Tomasi

, Benveniste

, Volkow

(2018) beta-Amyloid accumulation in the human brain after one night of sleep deprivation. Proc Natl Acad Sci U S A 115, 4483–4488.

61.

Liu

, Le

(2014) Epigenetic modifications of chronic hypoxia-mediated neurodegeneration in Alzheimer’s disease. Transl Neurodegener 3, 7.

62.

Grinan-Ferre

, Sarroca

, Ivanova

, Puigoriol-Illamola

, Aguado

, Camins

, Sanfeliu

, Pallas

(2016) Epigenetic mechanisms underlying cognitive impairment and Alzheimer disease hallmarks in 5XFAD mice. Aging (Albany NY) 8, 664–684.

63.

Watson

, Roussos

, Garg

, Ho

, Azam

, Katsel

, Haroutunian

, Sharp

(2016) Genome-wide DNA methylation profiling in the superior temporal gyrus reveals epigenetic signatures associated with Alzheimer’s disease. Genome Med 8, 5.

64.

Bernstein

, Lin

, Street

, Lin

, Dai

, Yu

, Bao

, Gearing

, Lah

, Nelson

, He

, Levey

, Mulle

, Duan

, Jin

(2016) 5-Hydroxymethylation-associated epigenetic modifiers of Alzheimer’s disease modulate Tau-induced neurotoxicity. Hum Mol Genet 25, 2437–2450.

65.

Mano

, Nagata

, Nonaka

, Tarutani

, Imamura

, Hashimoto

, Bannai

, Koshi-Mano

, Tsuchida

, Ohtomo

, Takahashi-Fujigasaki

, Yamashita

, Ohyagi

, Yamasaki

, Tsuji

, Tamaoka

, Ikeuchi

, Saido

, Iwatsubo

, Ushijima

, Murayama

, Hasegawa

, Iwata

(2017) Neuron-specific methylome analysis reveals epigenetic regulation and tau-related dysfunction of BRCA1 in Alzheimer’s disease. Proc Natl Acad Sci U S A 114, E9645–E9654.

66.

Horvath

, Raj

(2018) DNA methylation-based biomarkers and the epigenetic clock theory of ageing. Nat Rev Genet 19, 371–384.

67.

Lopez-Otin

, Blasco

, Partridge

, Serrano

, Kroemer

(2013) The hallmarks of aging. Cell 153, 1194–1217.

68.

Ripperger

, Merrow

(2011) Perfect timing: Epigenetic regulation of the circadian clock. FEBS Lett 585, 1406–1411.

69.

Azzi

, Dallmann

, Casserly

, Rehrauer

, Patrignani

, Maier

, Kramer

, Brown

(2014) Circadian behavior is light-reprogrammed by plastic DNA methylation. Nat Neurosci 17, 377–382.

70.

Lucey

, McCullough

, Landsness

, Toedebusch

, McLeland

, Zaza

, Fagan

, McCue

, Xiong

, Morris

, Benzinger

TLS

, Holtzman

(2019) Reduced non-rapid eye movement sleep is associated with tau pathology in early Alzheimer’s disease. Sci Transl Med 11, eaau6550.

71.

de Oliveira

, Cella

, Locker

, Ravindran

KKG

, Mendis

, Wafford

, Gilmour

, Dijk

, Winsky-Sommerer

(2022) Improved sleep, memory, and cellular pathological features of tauopathy, including the NLRP3 inflammasome, after chronic administration of trazodone in rTg4510 mice. J Neurosci 42, 3494–3509.

72.

Colavito

, Fabene

, Grassi-Zucconi

, Pifferi

, Lamberty

, Bentivoglio

, Bertini

(2013) Experimental sleep deprivation as a tool to test memory deficits in rodents. Front Syst Neurosci 7, 106.

73.

Lyons

, Bartolomucci

(2020) Stress and Alzheimer’s disease: A senescence link? Neurosci Biobehav Rev 115, 285–298.