Circadian disruption and sleep disorders in neurodegeneration

Sleep-wake activity is regulated by internally driven rhythm of the circadian clock. Disruption of the sleep-wake cycle has been found in a variety of PD and AD models with multiple assessment methods, including running wheels, infrared beams, piezoelectric systems, and electrophysiological measures such as EEG and EMG. Sleep-wake disturbances in PD and AD seem to be associated with genetic mutations. For instance, Drosophila with pink1 (phosphatase and tensin homolog-induced putative kinase 1) and parkin mutants show fragmentation of sleep [12]. Heterozygous (D409V/WT) GBA1 (glucocerebrosidase 1) mutant mice show increased NREM sleep and reduced REM sleep durations [13]. Additionally, α-synuclein BAC transgenic mice exhibit REM sleep without atonia (RSWA), which is a key feature of REM sleep behavior disorder (RBD). Regarding the sleep alterations in animal models of PD and AD, Fifel, Medeiros and their colleagues have comprehensively discussed this topic, together with the strengths and limitations [14, 15].

The sleep-wake activity may be related to proteostasis in the brains of patients with neurodegenerative diseases. A few studies have shown that the sleep-wake activity regulates the clearance of misfolded proteins in the glymphatic system of the brain [7]. Interrupted sleep has also been shown to increase the level of Aβ protein in the cerebrospinal fluid [16, 17]. One possible explanation is that Aβ or α-synuclein induces sleep disorders, which in turn hinders their clearance by the lymphatic system, and ultimately accelerates the pathological progression of the disease.

The rest-activity rhythm is often characterized using nonparametric analysis of actigraphy data. The circadian rhythm disruption in PD is characterized by a reduction in the amplitude of the circadian rhythm, resulting in an overall flattening of the rhythm, but with no significant shift in circadian phases [4, 21]. One recent study quantified the actigraphy data, and found that the activity rhythms are associated with disease severity and fluctuations of symptom intensity [22].

Non-motor symptoms in PD are associated to a certain extent with an impaired circadian rhythm. Increasing lines of evidence have shown that circadian dysfunction plays a role in cognitive impairment associated with PD, either by directly affecting cognition or indirectly by exerting effects on sleep and alertness [23]. Several other functions (emotional disorders and gastrointestinal dysfunctions) that are relevant for optimal behavior are known to be altered in PD. However, their link with the circadian clock remains poorly understood [4].

Sleep disorders are the most common non-motor symptoms and comprise the entire spectrum of sleep disorders, mainly presenting with disorders of regulation of sleep and wakefulness (such as insomnia and daytime sleepiness), parasomnias (mainly REM sleep behavior disorders, but also, albeit more rarely, sleepwalking and overlap parasomnia), sleep-related movement disorders (restless leg syndrome, RLS) and sleep‑disordered breathing (SDB) [24]. Insomnia is a frequent symptom in PD and most PD patients often complain about sleep fragmentation and early awakening. The prevalence of insomnia in PD increases over time and requires periodic assessments. PD patients with depressive symptoms, motor fluctuations and the use of higher doses of dopamine agonists tend to suffer more severe insomnia [25]. Interestingly, insomnia is often one of the characteristics of PD where motor symptoms are improved upon awakening from sleep and prior to drug intake. However, the underlying mechanism of this phenomenon (“sleep benefit”) is unclear, and there is no direct evidence for its correlation with the circadian type predominance [26]. Excessive daytime sleepiness (EDS) is a major health hazard in PD, affecting 21%–76% of PD patients with an incidence of 6% per year [27]. In our previous study employing 586 PD patients, male sex, disease duration and depression were found to be main risk factors for EDS in PD patients, while depression was a predictive factor for poor night-time sleep quality in all PD patients, whether they were male or female, and had early- or late-onset PD [28]. Circadian dysfunction may underlie the excessive sleepiness in PD. Compared with PD patients without EDS, patients with EDS have significantly lower amplitude of the melatonin rhythm and 24-h area-under-the-curve for circulating melatonin level [29]. RBD is a parasomnia, characterized by dream-enacting behaviors (DEB) and nightmares linked to RSWA. Compared with PD without RBD, PD with RBD exhibits clinical heterogeneity of motor and non-motor symptoms. More specifically, both DEB and RSWA are associated with the severity of PD. DEB symptom might fluctuate or disappear over time whereas RSWA may continue to develop as PD progresses [30]. RLS has one essential diagnostic characteristic, the presence of circadian variation of symptoms. The circadian clock-controlled gene, Tef, is also associated with sleep disturbances in PD, including RLS symptoms [31]. SDB is not more frequent in PD than in the general population. We previously explored the clinical characteristics of PD with comorbid obstructive sleep apnea (OSA), and found that age and male gender are risk factors for OSA in PD [32]. OSA may exacerbate neurodegenerative processes in PD. It is hypothesized that the OSA-related intermittent hypoxemia leads to oxidative stress, neuroinflammation, cerebrovascular effects, disruption of the glymphatic function by sleep fragmentation, and changes in the integrity of the blood–brain barrier [33].

Nocturia is a condition that is caused by the dampening or reversal of the daily pattern of urine excretion, which is associated with poor quality of life, falls, and institutionalization in PD. The prevalence of nocturia ranges between 76% and 86% according to a previous questionnaire survey [34]. Recent advances in circadian biology and sleep science have raised the importance of considering nocturia as a form of circadian dysfunction, with a focus on the influence of circadian genes on the bladder as studies have demonstrated circadian gene cycling in all levels of bladder tissue, and abnormal nocturnal urine production [35].

The physiological facets of circadian dysfunctions in PD include dysfunction of the autonomic nervous system and dysrhythmias of neuroendocrine secretion.

Rhythmic abnormalities in the autonomic nerve function in PD are well recognized, including reversal or even a full arrhythmia of blood pressure and heart rate variability (HRV), and impairment of the core body temperature (CBT) rhythm [20]. Nocturnal hypertension is nearly ubiquitous in PD, and most PD patients exhibit either a blunted nocturnal fall of blood pressure or higher blood pressure during the night than during the day (known as reverse dipping), leading to the difference between daytime and night-time BP closed [36]. HRV, an index for the autonomic, especially parasympathetic functions, is lower all day but higher at night in individuals with PD than in healthy controls [37]. Research has shown that impaired HRV may be related with the disease severity, motor symptom duration and dopaminergic dose in PD [38]. PD patients also show disruptions of circadian thermoregulation, with significant reductions in the mesor (the mean value around which the core temperature rhythm oscillates) of the CBT rhythm and dampened CBT rhythms, both of which are strongly correlated with REM sleep [39].

Melatonin synthesis and corticosteroid secretion are directly or indirectly regulated by SCN, and can be used as markers of the central clockwork to reflect the endogenous rhythmicity [40]. Two well-known studies have examined rhythms of melatonin in PD. One study showed a significantly diminished amplitude and amount of melatonin secretion in PD patients (n = 20) receiving stable dopaminergic therapy compared with controls (n = 15). Among PD patients, those with EDS exhibit the most prominent impairment in circadian melatonin secretion [29]. The other study also suggested reduced circulating melatonin levels in patients newly diagnosed with PD (n = 30), compared with the matched controls (n = 15) [41]. Neither of these studies found any difference in the timing of melatonin onset or offset [29, 41]. Different from these observations, some other studies [42] found an earlier peak of nocturnal melatonin level in PD patients receiving levodopa than in the control group. However, secretion and diurnal rhythmicity of some other circadian-modulated hormones are unaffected in PD, including growth hormone, thyroid stimulating hormone, prolactin, as well as certain fat tissue-associated hormones [43, 44].

Physiological facets of circadian dysfunctions in AD mainly include dysfunctions of the autonomic nervous system and dysrhythmias of hormone secretion.

Dysregulations of the autonomic function in AD primarily manifest as orthostatic hypotension [83], non-dipping or reverse dipping in mainly systolic blood pressure [84], reduction in HRV index [85] and a phase delay in CBT [86]. Intriguingly, Kim et al. reported that HRV might be a potentially useful tool for early differentiation between AD and DLB [87].

Additionally, AD patients exhibit reductions of melatonin levels and a phase delay in melatonin secretion [88, 89]. Findings in HD are largely inconsistent with those in AD [90, 91]. As for cortisol, elevated cortisol levels and a phase-advanced cortisol-secretion pattern are observed in AD patients [92, 93], which bear some resemblance to those in late-stage HD [94]. In addition to higher cortisol levels, early-stage HD also displays an increased amplitude in the cortisol secretion rhythm [90].

The mechanisms linking circadian rhythms, sleep and AD have not come to a definitive conclusion. Mounting evidence has suggested a bidirectional relationship between them. Previous studies indicated that circadian dysfunctions worsen neurodegeneration in AD through cholinergic disturbances and melatonin loss. Autopsy studies have revealed loss of neurons in the SCN of AD patients, particularly neurons expressing vasopressin, melatonin receptor type 1 and vasoactive intestinal peptide (VIP) [95, 96]. In addition, the decreased melatonin levels [97] are associated with the rest-activity rhythm disorder [95, 98]. Moreover, AD patients with circadian dysfunctions show loss of ipRGCs [82, 99], which was reported to be associated with Aβ deposition in one study [82]. Additionally, AD patients or mouse models with cholinergic disturbances demonstrate impaired circadian pattern [100, 101]. Melatonin suppresses Aβ generation and amyloid fibril formation. There is evidence showing a reduction of melatonin level in the prodromal and progressive stages of AD [102]. On the contrary, circadian disruptions promote neurodegeneration in AD. Sleep deprivation or disruption significantly increases neuroinflammation and subsequent Aβ production in the cerebrospinal fluid (CSF) of AD patients [103], increases Aβ and phosphorylated tau (ptau) levels in transgenic AD mouse models [104, 105], and decreases glymphatic flow and Aβ clearance in human CSF and in animal models [104, 106], which may lead to further progression of neurodegeneration in AD. Notably, chronic sleep disruption may increase the dissemination of tau protein in neural networks [107]. Musiek et al. have reported severe astrogliosis, oxidative injury and synaptic degeneration in Bmal1-deleted mice, indicating that circadian dysregulation of the neuronal redox homeostasis may also contribute to neurodegeneration in AD [1].

Circadian disruption and sleep disorders are also observed in other neurodegenerative disorders, including MSA, DLB, frontotemporal dementia (FTD), and HD.

Although no α-synuclein deposition is found in either the SCN or the pineal gland in MSA cases, the circadian dysfunction may be secondary to degeneration of other systems, such as the autonomic networks [108]. The circadian rhythm is regulated by VIP-expressing neurons, which are more involved in autonomic control and depleted in the SCN of patients with MSA [95]. MSA has a wider impairment of circadian regulation of endocrine and autonomic functions, such as plasma cortisol concentration [109], blood pressure, gastric myoelectrical activity [110], and nocturnal polyuria.

Patients with DLB present with decreased amplitude of CBT during night and more severe daytime sleepiness than controls, as well as more frequent RBD than AD and healthy controls [111]. Notably, inclusion of RBD has been proven to improve the diagnostic accuracy of DLB [112].

Although neurodegenerative disorders exhibit fragmentation of sleep, patients with FTD [113] show phase-advanced activity rhythm while HD patients [114] are with phase delay.

Circadian activity disturbance, initially considered as symptoms of neurodegeneration, is believed to be involved in the occurrence or progression of neurodegenerative process. In other words, circadian disruption may precede clinical symptomology or add to the risk of neurodegenerative diseases.

Only few studies have examined the temporal association between circadian rhythm disruption and PD [159–161]. In a prospective study including 2930 men, Len et al. found that those who self-reported napping time of more than 1 h/day had a twofold risk of developing PD, compared to those with no EDS or having napping < 1 h/day [159]. From the same cohort, they also reported that weakened circadian rhythmicity was associated with increased risk of incident PD and such association remained significant even after excluding PD diagnosis in the first 2 years of follow-up [160]. These findings indicated that circadian rhythm disruption could be a prodromal marker for PD. To minimize the effect of reverse causality and potential confounding factors, recent mendelian randomization study using UK Biobank from European ancestry found that morning chronotype had an inverse causal effect while M10 (average activity during the most active 10 consecutive hours of the day) had positive causal effect on the later onset age of PD [161]. They also noticed that better sleep efficiency was causally associated with a decreased AD risk. However, temporal associations between circadian rhythm and PD or other neurodegenerative diseases such as HD and amyotrophic lateral sclerosis have not been thoroughly investigated as in AD and dementia. In addition, behavioral indicators like actigraphy do not necessarily parallel endogenous circadian biomarkers which maintain 24-h oscillations even in the case of sleep disturbance. Therefore, comprehensive evaluations of both behavioral and biological markers of circadian rhythm are needed to provide more convincing and detailed evidence on circadian rhythm disruption in PD-related neurodegeneration.

In a case–control study, more evident circadian misalignment, assessed by PSG and dim light melatonin onset, was noticed in MCI patients before AD diagnosis [162]. A recent study by Musiek et al. including 189 healthy elders showed that those with preclinical AD, as assessed by PET, demonstrated increased rest-activity rhythm fragmentation independent of aging and sex [163]. This finding suggested that circadian disruption is involved early in AD pathogenesis. In line with clinical findings, research on animal models revealed disruption of daily Aβ oscillations in the hippocampal interstitial fluid and acceleration of amyloid plaque accumulation in mice with disturbed circadian rhythm [164]. Inhibition of the circadian repressor REV-ERB is also associated with enhanced transcription of core clock gene BMAL1 and increased Aβ clearance [165]. These observations add to the evidence of the potential role of circadian regulation in AD.

The temporal association between circadian rhythm and AD or dementia has been explored in many epidemiological studies [166–172]. Habits of dysregulated circadian rhythm, such as long-time night-shift or delayed rising time, are associated with increased risk of dementia among healthy populations [166, 167]. Assessment of behavioral indicators of circadian rhythm disruption via actigraphy can more objectively reveal the association [173]. Based on the same cohort of Study of Osteoporotic Fractures including 1283 women, Tranah et al. [174] and Walsh et al. [168] both found an association of decreased circadian amplitude with higher risk of cognitive decline or dementia during the 5-year follow-up. Besides, they reported worse cognitive decline in association with phase delay, different from the report by Rogers-Soeder et al. [169], which showed phase-advanced acrophase in the men-only cohort. In a prospective cohort study including 1401 healthy older adults, lower amplitude of 24-h activity rhythm and higher intradaily variability for hourly fragmentation of activity rhythm are associated with higher risk of developing AD dementia, while lower interdaily stability of 24-h activity rhythm predicts higher risk of transition from MCI to AD dementia [170]. The observed parameters also worsen as dementia deteriorates, indicating that the relationship between circadian disruption and AD progression is bidirectional. To specify the affected aspects of circadian rhythm, novel or refined analysis has been applied in some studies [171, 172]. A 24-h time-limited difference of circadian rhythm was observed in individuals with preclinical PD in various periods in comparison to healthy controls [172]. By using parametric and nonparametric analysis, reduced overall rhythmicity, lower amplitude and activity level, and later activity timing were revealed to be associated with development of MCI and dementia [175]. However, there are also some studies showing no association of 24-h activity rhythm fragmentation with dementia risk or preclinical AD [172]. Discrepancies in these findings could be due to different sample sizes, study populations or statistical methods.

Although sleep-wake disturbance is commonly found in PD, the rhythmic change throughout disease process has been rarely noticed. Using continuous actigraphy, a study showed that PD patients with higher Hoehn and Yahr stage are significantly more active later in the day [22], indicating an alteration of circadian rhythm in relation to disease severity. Additionally, the coexistent sleep disorders in PD might increase the variability of circadian rhythm. In a study of 15 iRBD patients, 31 PD patients and 6 DLB patients [111], Raupach et al. observed an inverse correlation between the CBT amplitude and RBD severity. Interestingly, the alterations in CBT were absent in PD patients free of RBD. This suggests that certain circadian rhythms might be particularly linked to RBD pathology and a further exploration of prodromal sleep disorders in PD is necessary. When discussing the longitudinal contribution of PD progression in relation to sleep disorders or circadian changes, the use of anti-PD medications is also of great importance [4]. In other words, whether these alterations result from dopaminergic treatment or from PD deterioration itself should be clarified. For example, a previous study found that only PD patients with levodopa-induced motor complications showed decreased ratio of melatonin secretion at night while such change was not observed in untreated patients or treated PD patients without motor complication [42]. Two other sleep disorders in PD, EDS and RLS, may also be consequences of dopaminergic therapy when symptoms aggregate as disease progresses. Drug-naive PD patients showed increased EDS severity from baseline to year 3 while no change was observed in the healthy control group. Meanwhile, the influence of dopaminergic medications on EDS was dose-dependent at years 2 and 3 of study follow-up [176]. The prevalence of RLS in untreated PD patients does not differ significantly from that of the general population, but is significantly increased in those receiving dopaminergic medication [177].

Across the disease continuum, studies using functional and biological indicators of circadian disruption in AD showed severer or irregular changes [88, 162, 178]. The severity of increased fragmentation of rest-activity cycle and overnight activity correlates with AD severity, with disturbances being most prominent in institutionalized patients [179]. The phase delay identified in AD patients is more prominent in those with advanced AD pathology [180]. As for core body temperature, a significant phase advance in circadian thermoregulation was observed in MCI patients compared to the healthy group [178] while phase delay was noticed in AD patients [180]. Similarly, for biological indicators such as melatonin level, MCI patients have a phase advance in melatonin secretion whereas those with mild-to-moderate AD exhibit a phase delay [88, 162]. These observations indicate that the neurodegenerative process can alter circadian rhythmicity and the latter can become more irregular as pathology deteriorates. Another question is that whether it is AD itself, but not aging, that promotes circadian disruption and sleep disorders. Initial view contends that individuals with AD have circadian alterations similar to those seen in healthy older adults, but with higher severity [181]. However, a recent study by Musiek et al. showed that preclinical amyloid pathology is associated with worse circadian fragmentation regardless of age [163]. This suggests that despite the same fragmentated circadian pattern, AD and normal aging drive circadian dysfunction in separate ways. Moreover, ageing alone is associated with a dramatically increased prevalence of preclinical AD to 30%–40% [182]. That is to say, in several epidemiological studies examining circadian dysregulation/sleep disorders preceding AD symptomology [168, 171, 175], the enrolled participants of advanced age might be already in the early stage of AD, which provides an interpretation for the observation that an underlying preclinical AD pathology may in turn lead to circadian and sleep dysfunction. The estimated prevalence of overall sleep disorders in individuals of AD is about 39% [46]. With self-reported assessment and overnight PSG, Hita-Yañez et al. found patients with MCI already showing disturbed sleep at both objective and subjective levels [183]. This work was further supported by a subsequent study showing significant relationship between Aβ deposition and sleep quality in preclinical AD [184]. As these cross-sectional observations were from an early stage of AD, this could be interpreted, from one perspective, that sleep is particularly sensitive to AD pathology. As dementia worsens, concurrent sleep disorders in AD such as EDS can be aggregated through potential mediating factors [184]. Comorbid depression and decreased social engagement resulting from impaired cognition in AD may add extra burden to EDS severity [185]. For SDB, around half of individuals with AD experience OSA and show five times higher odds of having OSA compared to age-matched cognitively intact individuals [184]. However, the longitudinal association between OSA or other specific sleep disorders (e.g., RLS) and AD progression has not been well explored to date [186].

As studies focusing on alterations of circadian rhythm/sleep disorder after diagnosis of neurodegenerative diseases are generally lacking, more studies are needed to longitudinally examine alterations of circadian rhythm and sleep disorders with progression of neurodegenerative diseases.

Non-pharmacological approaches are considered as the first-line therapies for the management of circadian dysfunction and sleep disorders in neurodegenerative diseases. Treatment plans must be tailored at individual level. Prior to treatment, a comprehensive set of clinical, neuropsychological, neuroimaging, and electrophysiological assessments should be conducted, and the circadian disruption as well as the causes and subtypes of sleep disorders need to be carefully evaluated. Notably, sleep hygiene is recommended as an important behavioral and environmental practice in every treatment plan to promote better-quality sleep in PD [187] and AD [188] patients. As different medications for different types of sleep disorder may interact with each other, it is necessary to identify and treat coexisting or primary sleep disorders before making medication plans (Fig. 4).

Chronotherapy is a therapeutic approach that incorporates an individual's circadian rhythm into disease treatment. It is based on the principle of prescribing drugs according to the different characteristics of an individual's circadian rhythm or combining both pharmacological and non-pharmacological approaches. Characterized by maximizing drug effectiveness and minimizing its side effects, chronotherapy was initially used mainly in the treatment of hypertension and in oncology. Recent studies have found that chronotherapy has promising applications in neurodegenerative diseases [194, 224, 225].

A study suggests that combining BLT with melatonin in demented patients may increase sleep efficiency, attenuate agitated behavior and even improve nocturnal restlessness for 3.5 years [224]. Dowling et al. found that 1 h of morning light exposure (2500 lx) for 10 weeks together with 5-mg melatonin in the evening significantly increases the daytime awake time and activity levels and strengthens the rest-activity rhythm of AD patients, compared to the light therapy alone [225]. MuCurry et al. found that combination treatment (walking, light, and sleep education) and each treatment alone have similar effects in improving sleep outcomes in AD patients [194]. Preliminary studies of chronotherapy in neurodegenerative diseases have shown promising findings. Personalized treatment plans are essential for effective implementation of chronotherapy, as the melatonin secretion curve varies from individual to individual. Therefore, the timing of pharmacological agents such as melatonin should be personalized according to the individual’s circadian rhythm. The implementation time and dose are also to be studied in the future [226].

Animal studies have shown that rTMS has positive effects on neural regeneration and neuroprotection through inhibiting apoptotic cell death, as well as regulating neurotransmitters and neurotrophic factors [227]. Light therapy decreases oxidative stress markers [228] and removes Aβ via the lymphatic system of the brain [229] in AD mouse models, and reduces the loss of dopaminergic cells and increases tyrosine hydroxylase-positive cells in PD mouse models [230].

Orexin is a neuropeptide that contributes to the regulation of the sleep-wake cycle by increasing the arousal level and maintaining wakefulness. Suvorexant, one of the orexin receptor antagonists, has been approved to treat insomnia in elderly adults. Animal experiments showed that suvorexant can reduce amyloid-β plaques, improve synaptic plasticity, and restore the circadian phosphorylated CREB (cyclic AMP-response element binding protein) expression in the hippocampus of APP/PS1 mice [231].

Furthermore, some small-molecule modulators have been developed to restore the disrupted circadian system. For example, casein kinase 1 δ/ε inhibitor CKI-7 [232] significantly reduces endogenous Aβ peptide production [233], thus playing an important role in neuroprotection. Rev-erbα is a core negative component of the circadian clock and modulates the cellular clock and energy metabolism [234]. Rev-erbα knock-out mice show disrupted diurnal patterns [235]. The agonists (GSK4112) and antagonists (SR8278) of Rev-erbα could correct the abnormal circadian rhythms [232], providing biological evidence for future trials of these small-molecule modulators as a therapeutic for neurodegeneration.

In addition, traditional Chinese herbs or herbal extracts and non-pharmacological interventions are proven beneficial to patients with neurdegenerative disorders, by exerting antioxidant and anti-inflammatory effects [236, 237]. Clinically, several studies showed that acupuncture [238] and Yang-Xue-Qing-Nao granules [239] improve sleep quality in PD patients. Future research should focus on the quality control of traditional Chinese medicine studies and figuring out the pharmacological mechanisms of main active ingredients.

Not applicable.

Not applicable.

All authors claim that there are no conflicts of interest.

Not applicable.