Brain washing and neural health: role of age, sleep, and the cerebrospinal fluid melatonin rhythm

Abstract

The brain lacks a classic lymphatic drainage system. How it is cleansed of damaged proteins, cellular debris, and molecular by-products has remained a mystery for decades. Recent discoveries have identified a hybrid system that includes cerebrospinal fluid (CSF)-filled perivascular spaces and classic lymph vessels in the dural covering of the brain and spinal cord that functionally cooperate to remove toxic and non-functional trash from the brain. These two components functioning together are referred to as the glymphatic system. We propose that the high levels of melatonin secreted by the pineal gland directly into the CSF play a role in flushing pathological molecules such as amyloid-β peptide (Aβ) from the brain via this network. Melatonin is a sleep-promoting agent, with waste clearance from the CNS being highest especially during slow wave sleep. Melatonin is also a potent and versatile antioxidant that prevents neural accumulation of oxidatively-damaged molecules which contribute to neurological decline. Due to its feedback actions on the suprachiasmatic nucleus, CSF melatonin rhythm functions to maintain optimal circadian rhythmicity, which is also critical for preserving neurocognitive health. Melatonin levels drop dramatically in the frail aged, potentially contributing to neurological failure and dementia. Melatonin supplementation in animal models of Alzheimer’s disease (AD) defers Aβ accumulation, enhances its clearance from the CNS, and prolongs animal survival. In AD patients, preliminary data show that melatonin use reduces neurobehavioral signs such as sundowning. Finally, melatonin controls the mitotic activity of neural stem cells in the subventricular zone, suggesting its involvement in neuronal renewal.

Introduction

The term “brainwashing”, also referred to as coercive persuasion, is conventionally defined as “a systematic effort to persuade a non-believer to accept a certain allegiance, command or doctrine”. In the context of the current review, however, “brain washing” is a phrase which has a different connotation, namely, referring to the flushing of toxic agents and destructive molecules out of the neural parenchyma. This requires a unique physiological association that exists between cerebrospinal fluid (CSF) circulating in the subarachnoid and periarterial spaces and its movement through the neuropil from where it removes toxic waste products. The contaminated CSF eventually enters lymph vessels situated in the dural covering of the brain from where it drains out of the cranial vault to the deep cervical lymph nodes.

The description of the neural-lymphatic network, now known as the glymphatic system [1–3], has partially demystified the association of the central nervous system (CNS) with the classic lymphatic drainage. For decades before this revelation, ambiguity existed as to how the CNS rids itself of unusable solutes and damaged molecules which, in other tissues, is accomplished by a network of lymphatic vessels that permeate these organs [4]. The glymphatic system involves the movement of CSF from the ventricles of the brain throughout the subarachnoid space and into periarterial spaces (historically known as Virchow-Robin spaces) that surround penetrating blood vessels which are unique to the brain [5] (Fig. 1, top.). CSF [referred to as interstitial fluid (ISF) after entering the substance of the brain] then presumably convectively percolates through the neural tissue where it leaches out discarded cellular solutes and extracellular debris. Thereafter, the contaminated fluid enters perivenous spaces which transport it to the subarachnoid space from where it is absorbed into conventional lymphatic vessels that are located in the overlying dura mater after which it is delivered to the deep cervical lymphatic network [6] (Fig. 1). While this system has been described and functionally documented [7, 8], there is still considerable controversy surrounding the details of this unusual interface between the CNS and its lymphatic drainage [9].

Fig. 1.

Proposed framework of the glymphatic system, the function of which is to remove solutes and waste products from the neural parenchyma. Top. Cerebrospinal fluid (CSF) and its constituents in the subarachnoid space move over the surface of the brain and spinal cord. Arteries that enter the neural tissue, known as penetrating arteries, are surrounded by a perivascular space (Virchow-Robin space) which also contains CSF allowing this fluid to enter the deepest regions of the neural parenchyma. From the periarterial space, CSF and its constituents enter the parenchyma to become interstitial fluid (ISF). The bulk flow of ISF moves through the extracellular space and flushes out solutes and molecular debris and eventually enters the space around penetrating veins (perivenous space). The toxin-laden CSF in the perivenous channels is carried to subarachnoid space on the surface of the brain from where the contaminants are absorbed into the lymph vessels in the overlying dura mater. The lymph is then drained from the cranial vault to the deep cervical lymph nodes. Thus, the glymphatic system serves the same purpose for the brain and spinal cord as the regular lymphatic vessels do for all other tissues. An example of a molecule that is removed from the brain as a result of the CSF/ISF/CSF movement is amyloid-β peptide (Aβ). Aβ is both flushed into the perivenous space (2) or it can also be metabolized by microglia (1). Bottom. On the left is illustrated the movement of periarterial CSF which is taken into the neural tissue, a transfer aided by the presence of aquaporin 4 (AQP4) water channels (green dots) which are abundant in the astrocytic feet which line the perivascular spaces. The convective flow of ISF through the neuropil leaches out metabolic debris and interstitial solutes and eventually collects in the perivenous space for eventual drainage from the brain. The arrows indicate the direction of flow of both the CSF in the perivascular spaces and blood in the penetrating artery and vein. The glymphatic system deteriorates in advanced age reducing its efficiency in eliminating damaged molecules from the brain. Also, it may function differently during the night than during the day (see text for details). Adapted from [9] and [212]

Herein, we describe the potential role of circadian pineal-derived melatonin, which is in high concentrations in the CSF at night [10–12], relative to brain protection. It reduces oxidative damage to neurons and glia and clears the brain of cellular by-products and toxic molecules to aid in the protection against neurodegeneration and cognitive decline [13, 14]. There is evidence from several sources that endogenously produced melatonin delays the onset of diminished neurological capacity and that its exogenous administration reduces neuronal loss and neurobehavioral changes associated with age-related CNS diseases [15, 16].

Cerebrospinal fluid production, movement, and absorption

The CSF bathes both the inside and outside of brain and spinal cord and was initially assumed to only buffer the CNS from possible damage resulting from trauma. Research during the last four decades has clearly shown, however, that this dynamic fluid has many additional critical functions in maintaining brain/spinal cord homeostasis [17]; nowhere is this more apparent than relative to the glymphatic system as will be elaborated in this report.

CSF is primarily a result of secretion from the four choroid plexuses that are present in the brain ventricles (Fig. 2), although other sources of CSF production are not completely precluded under some conditions. In addition to CSF production, there is a single report claiming that explants of porcine choroid plexus also synthesize melatonin which could potentially be released into the ventricular system [18]. Likewise, the pineal gland may also be the source of an estimated 10% of the total CSF in young individuals; this gland likely releases CSF into the posterodorsal third ventricle in tandem with the nocturnal discharge of melatonin into this fluid [19]. In the elderly, the quantity of CSF from all sources wanes as the contributing cells become dysfunctional. Under normal physiological conditions, the rate of CSF generation is balanced by the rate of absorption, so that movement of this fluid is constant. In the adult human, it is estimated that the rate of CSF production is on the order of 500 mL per 24-h period [20] and is renewed several times each day.

Fig. 2.

Top. Overview of the ventricular system of the brain. Cerebrospinal fluid (CSF) is continually secreted by the choroid plexuses in the lateral, third and fourth ventricles. CSF, which contains pineal-derived melatonin at night, moves steadily through the ventricles and enters the subarachnoid space through the bilateral foramina of Luschka and the medial foramen of Majendie. From the subarachnoid space melatonin enters the periarterial spaces (Virchow-Robin spaces) of the brain and spinal cord from where it is taken into the neural parenchyma. The boxes (A, B and C) are enlarged in the lower images. A An illustration of the association of the pineal gland with the posterodorsal aspect of the third ventricle (the pineal recess). The high nocturnal levels of melatonin in the CSF are derived directly from the pineal gland; there are several morphophysiological modifications of the gland and the recess which possibly support the release of melatonin directly into the CSF. The ependymal cells, which line the entire ventricular system, are absent in the pineal recess allowing the pinealocytes, the melatonin-synthesizing cells, to directly contact the CSF [38]. Also, the pineal gland contains non-endothelial lined web-like intercellular canaliculi into which melatonin is potentially released for transport directly to the CSF [213–215]. B The proximity of melatonin in the CSF of the third ventricle (3v), allows it access to the biological clock (suprachiasmatic nucleus) either by simple diffusion or via transfer by tanycytes (modified ependymal cells) which are especially numerous in the infundibular recess (represented in this image). C A representative depiction of the subventricular zone (SVZ) of the lateral ventricle. The SVZ is one of two areas of the brain that contain neural stem cells (NSC) which retain the ability to proliferate. The rhythmic proliferation of NSC is regulated by melatonin. The projection of some cells in the SVZ have processes that pass between the overlying ciliated ependyma to monitor the concentration of growth factors, cytokines, etc., in the CSF. There are significant differences in the cytoarchitecture of the SVZ among species. Blue cells = protoplasmic astrocytes; red cells = migratory neuroblasts; orange cells = mature neurons. Image C is an adaptation of a figure published by [216]

The movement of CSF through both the ventricles and subarachnoid space is reportedly aided by hydrostatic pressure, cardiac and pulmonary pulsations, and motile cilia on many ependymal cells throughout the ventricular system [17, 21]. This creates ebbs and flows of CSF such that constituents in this fluid can be shuttled between brain regions. From the lateral and third ventricles, CSF passes through the cerebral aqueduct of the mesencephalon and pons to enter the fourth ventricle from where it has access to the subarachnoid space via the paired foramina of Luschka and the single foramen of Majendie in the posterior cranial fossa (Fig. 2).

CSF absorption from the subarachnoid space involves a number of routes, i.e., into the superior sagittal sinus via the arachnoid granulations, into lymphatic vessels in the cerebral dura, and into the cervical lymphatics in the case of the spinal cord. Also, CSF which flows into the Virchow-Robin spaces that surround penetrating arteries enters the brain parenchyma to become ISF; this uptake process is aided by numerous aquaporin-4 (AQP4) channels on astrocytic end feet which line the periarterial spaces [22] (Fig. 1, bottom). This uptake does not deplete the total CSF volume since the ISF is subsequently released into the perivenous spaces from where it is transported back to the subarachnoid space on the surface of the brain (Fig. 1, top). Recent data indicate that the volume of CSF that enters the brain parenchyma from the Virchow-Robin spaces may be greater when neurons are exhibiting slow wave electrical activity, as occurs during sleep. At night, neuronal activity wanes, reducing the need for oxygen; this causes a diminished blood flow to select areas of the cerebrum leading to a transient pressure drop [23] Additional CSF then flows into these areas so a safe pressure is quickly re-established. The additional CSF circulating through the glymphatic network potentially allows for more melatonin-rich CSF to pass through the neuropil at night thereby improving waste clearance (Fig. 3).

Fig. 3.

A graphic representation of the presumed changes in the human glymphatic network and associated processes throughout the day and the night. The heavy red line at the top represents the melatonin concentration in the cerebrospinal fluid (CSF). The circadian variation in CSF melatonin varies from that in the blood in terms of the much greater nighttime amplitude and the sharper “on” and “off” at the light-to-dark and dark-to-light transitions. Each of the parameters listed exhibits a day:night difference. Greater nighttime CSF movement through the glymphatic system with higher waste clearance accompanies elevated melatonin, sleep and slow waves. In contrast, Aβ accumulation and neuronal and glial oxidative damage (ROS/RNS) are highest during the day and lower at night since melatonin functions as an antioxidant as shown by the nighttime rise in the total antioxidant potential (see Fig. 6). Regular circadian rhythms are also important in warding off neurodegeneration and melatonin stimulates neural stem cell (NSC) proliferation, which may assist in replacing lost neurons

Melatonin in the cerebrospinal fluid

Pineal melatonin synthesis is typically confined to the daily period of darkness in all vertebrates including the human [24]. Unlike in most other endocrine glands, this hormonal product is not stored in the gland but rather is immediately released when synthesized. As a result, the measurement of circulating melatonin levels is a dependable index on the ongoing synthetic activity of the pineal gland. The day/night rhythm in circulating blood melatonin levels has been identified in many species and is thought to exist in essentially all vertebrates with the exception perhaps of eyeless cave-dwelling fish, subterranean mammals, etc.

The circadian pineal and blood melatonin cycles are impacted by the seasonal duration in night length in animals kept under natural photoperiods with long daily durations of elevated blood melatonin during the shorter days of the winter [25] and in humans maintained under long night conditions [26]. Legally, blind humans who can still perceive light have a normal circadian melatonin rhythm synchronized by the light: dark changes [27]. By comparison, the melatonin cycle also persists in blind humans incapable of light perception but it free runs with a period length of about 25 h, rather than with a 24-h cycle common to sighted humans [28]. This inherent cycle of 25 h persists in profoundly blind individuals the lighting environment does not synchronize it. Normally sighted humans born with Smith–Magenis syndrome have an inverted rhythm which is not free-running such that high circulating melatonin levels are persistently associated with the daily light period with low blood levels at night [29, 30]; as a result of this perturbation, these individuals are chronically desynchronized [31].

While the primary route of pineal melatonin secretion has long been considered to be into the perfuse capillary bed that exists in this gland, nevertheless the nocturnal blood melatonin levels are uncommonly low (typically in the low pg/mL range in this fluid). By comparison, melatonin also has routinely been detected in the CSF with the nighttime melatonin concentrations in third ventricular CSF usually being much higher than concurrent blood levels, e.g., in the bovine calf and goat, suggesting the ventricular system is a major or primary route of melatonin release [10, 32] (Fig. 4); this equally applies to the human [33]. This route of pineal melatonin release was predicted almost five decades ago, but the strongest evidence for this has accumulated only recently [12, 34, 35]. The blood and CSF nocturnal melatonin rises are temporally correlated [36].

Fig. 4.

Concurrent melatonin levels, measured by radioimmunoassay, in the blood and the CSF collected from the third ventricle of bovine calves maintained in an inside enclosure under a strictly regulated light:dark cycle. The rhythms exhibit several distinct differences. As seen in other species, the amplitude of the nighttime rise in the CSF melatonin is much greater (average of 17-fold increase) than that in the plasma (average of sixfold increase), the rise occurs more rapidly in the CSF and is sustained at a higher level throughout the dark period. The mean values are the average of measurements made on six calves. The alternating white/black/white bar at the bottom identifies the light:dark cycle. Error bars were not included for figure simplification. Adapted from [10]

Tricoire and colleagues [21] provided the most complete study documenting the release of pineal melatonin directly into the third ventricle. In this elegant study, the authors simultaneously collected CSF using permanently implanted cannulas in the third ventricle from the area of the pineal recess and from the base of the ventricle near the infundibular recess of sheep. The measurement of melatonin concentrations in CSF samples collected simultaneously from these sites varied markedly, i.e., with very much higher levels in the CSF derived from the pineal recess compared to values in the lower third ventricle. These results strongly argue for the direct release, via the pineal recess, of melatonin into the third ventricle. To further document this, the authors sealed off the pineal recess with surgical glue; this led to a dramatic drop in CSF melatonin levels but did not impact concurrent circulating blood melatonin concentrations. There are also several morphological features of the pineal gland and modifications of the ependymal lining of the pineal recess that are consistent with the direct discharge of melatonin into the ventricular system [37, 38] (Fig. 2A). For ethical reasons, complete studies of this type that involve the third ventricular CSF have not been done in the human although a melatonin rhythm in human lumbar CSF that correlates closely with the changing blood melatonin cycle has been described [36]. Perhaps, as in animals, melatonin is also released directly into the CSF in humans; although less likely, pineal melatonin could initially be discharged into the blood with its eventual leakage into the CSF via the choroid plexus.

The study published by Bruce and colleagues [36] in which repeated lumbar CSF and simultaneous blood samples were collected from human volunteers over a 30-h period for the purpose of comparing their melatonin rhythms provides important information about the speed of movement of melatonin from the third ventricle to the lumbar cistern. Based on the curves, it is estimated that the rise in nocturnal lumbar CSF melatonin levels followed the increase in blood melatonin concentrations by roughly one hour. Considering the distance between the melatonin release site, i.e., the third ventricle, and the lumbar cistern, melatonin moves quickly throughout the ventricular and subarachnoid spaces. Thus, within a rather short interval after its discharge into the pineal recess, melatonin may perfuse the entire brain and spinal cord due to the movement of the CSF. The quick transfer of CSF-laden melatonin from the third ventricle to the lumbar cistern is consistent with studies that investigated the movement of fluorescent dextrans through the subarachnoid space and into the periarterial spaces of the mouse brain [37].

The results of Bruce et al. [36] also provide critical information about the regulation of the human pineal metabolic activity by the sympathetic nervous system. In a patient who was subjected to bilateral cervical peripheral sympathectomy to relieve hyperhidrosis, melatonin levels were either markedly reduced (blood) or undetectable (CSF). These data prove that the sympathetic neural input to the human pineal gland, as in other species, is driven by CNS signals acting through the peripheral autonomic nervous system [39, 40]. The residual quantities of melatonin in the blood of the ganglionectomized patient is not unexpected since many tissues, in addition to the pineal gland, produce melatonin which may leak (is not actively released) into the general circulation [41].

Cerebrospinal fluid flow aids in waste management

Sleep, by enhancing the movement of the CSF at night, carries out essential housekeeping functions in the brain [13, 35], namely, maintaining the intercellular and extracellular milieu of the brain. As noted, nighttime CSF also contains elevated levels of melatonin which may be involved with neural and glial protection and waste management [13, 35]. By clearing metabolic waste and damaged/misfolded proteins from the brain, the authors feel CSF flow may help to preserve neurocognitive physiology and reduce age-associated neurological decline. Melatonin may have several decisive roles in preserving brain cleanliness. These actions include scavenging of reactive oxygen and reactive nitrogen species to prevent the buildup of oxidatively-damaged molecules [42], stimulating the activity of neural antioxidant enzymes [43, 44] protecting the brain from bacteria/viruses [45], and curbing intracellular [46, 47] and extracellular debris and fluid/edema buildup [48–50].

Extracellular neurotoxic molecules such as amyloid-β (Aβ), a cardinal hallmark of Alzheimer’s disease (AD), are flushed from the brain during sleep [51] (Fig. 3). This washing of the brain to remove neural detritus is theoretically associated with increased CSF flow through specialized Virchow-Robin spaces surrounding penetrating arteries, its entrance into the brain parenchyma as ISF and its convective flow from the arterial pole to the venous pole [52] (Fig. 1). During the day when CSF flow is slow and melatonin levels are at their nadir, Aβ and intracellular hyperphosphorylated tau proteins accumulate [53] with their potential removal the following night [23]. A major explanation of some of the neuropathology associated with AD is the “Aβ cascade hypothesis”. This theory infers that when Aβ generation exceeds its clearance, neural deposits of this toxic protein increase which leads to extensive oxidative damage, neuroinflammation and cellular apoptosis. Alternating periods of damaged protein accumulation followed quickly by brain washing may be a key process in maintaining healthy neural tissue. Sleep deprivation exaggerates Aβ accumulation in the brain [54]. The reduced waste removal can be caused by a single night of sleeplessness or light exposure, e.g., in night shift workers. Such sleep disruptions are common for humans.

There are several theories to explain why solute flushing increases during sleep. One factor related to this is the increased extracellular space that accompanies sleep, which would enhance convective flow of ISF through the tissue. In addition, in rodents higher clearance rates of solutes from the brain occur when the animals are treated with anesthetics, which induce high EEG delta power, thus, implicating sleep dynamics as relevant to neural debris removal [55]. More recently, a human study revealed that large waves of CSF movement occur during NREM (non-rapid eye movement) sleep [23]. Importantly, the CSF waves are always preceded by neural slow wave activity. This coupling of CSF flow and neural activity implies these electrical changes may drive CSF movement via an action on blood flow. One feature not mentioned in these studies is the concurrence of restful sleep with elevated nocturnal melatonin concentrations in the CSF [36]. The importance of this to brain washing is discussed in detail below. CSF flow through the ventricles of the human brain, as mentioned above, is a result of the constant pulsating respiratory and cardiac cycles and perhaps other factors.

Many aspects of brain function undergo changes that are correlated with other fluctuations. Thus, neural slow waves are related to vascular dynamics, autonomic nervous system status, glial function, etc. As a basis for fMRIs, the depression in blood volume to localized areas occurs through neurovascular coupling [56, 57]. The localized reduced blood volume is correlated with enhanced CSF flow to these areas (Fig. 3). Lower blood volume due to vasoconstriction presumably relates to the degree of neural activity. As a result of these coupled alterations, the temporal changes in the vasculature are likely a major component of CSF dynamics [58].

There is evidence that advanced age alters the function of both the perivascular transport system as well as the dural lymphatics such that there is an estimated 40% reduction in human glymphatic clearance and solute disposal [59]. Likewise, sleep efficiency [60] and pineal melatonin [61, 62] production and circulating levels drop markedly in old age potentially contributing to an imbalance in the waste accumulation:waste clearance ratio in favor of accumulation. Importantly, sleep disorders are generally believed to contribute to impaired Alzheimer’s dementia and cognitive decline in the aged [63]. Also, melatonin is a commonly used agent for sleep promotion and in the elderly it often improves sleep quantity or quality; however, is efficacy in these individuals seems to depend on the physiological basis of the sleep disruption. Thus, improved sleep and the availability of youthful levels of melatonin may be beneficial in slowing age-related neurological decline due to its increased ability to remove toxic molecules from the brain each night. In the absence of a nightly melatonin rise resulting from nocturnal light exposure or advanced age, brain waste disposal may be compromised contributing to neural pathologies. Related to this is that individuals with dementias, e.g., AD, often have more significantly disturbed circadian rhythms and greater melatonin suppression than healthier elderly individuals [64, 65]. These perturbations may be further aggravated by their nocturnal locomotor activity which may expose them to light at night which would contribute to their circadian disruption and melatonin suppression [23, 66].

Melatonin: evidence suggesting it reduces neurocognitive decline

Melatonin has often been used as a supplement to alter the course of Aβ accumulation in experimental animals [67] and the associated behavioral signs of AD in humans [68]. The obvious sleep disturbances that accompany AD onset and progression could be a major factor in contributing to the piling up of Aβ in the brain of these individuals given that optimal sleep seems to be critical for flushing these extracellular misfolded proteins from the neural parenchyma via the glymphatic system [23, 69]. Furthermore, patients suffering with senile dementia seem to have severely disordered pineal melatonin synthesis and release [64, 70]. A common characteristic in the brain of patients suffering from neurodegenerative diseases is the massive cellular death due to apoptosis, which occurs in both neurons and glial cells in AD. In vitro experiments showed that melatonin reduces neuronal apoptosis including hippocampal neurons [71], the loss of which correlated with the early signs of AD. The reduction of neuronal apoptosis combined with improved sleep both likely contribute to melatonin’s memory-enhancing mechanisms [71].

Postmortem CSF melatonin levels in clinically diagnosed AD patients were significantly lower than those is age-matched non-demented elderly subjects [72]; moreover, AD patients identified as having two APOε4 alleles had lower melatonin levels than those of the genotype, APOε3/4. These results are hampered since they are single point measurements of patients who succumbed at different times of the day. Nevertheless, the overall collective evidence indicates AD patients experience a greater and possibly earlier lifetime drop in CSF melatonin concentrations. These changes may be a consequence of reduced neural signal the gland receives from the master circadian oscillator, the suprachiasmatic nucleus (SCN), which normally drives pineal melatonin synthesis and secretion, or due to the associated degenerative changes in the pineal gland itself [73]. The drop in CSF melatonin availability as a cleansing agent for the CNS via the glymphatic system in Alzheimer’s patients could potentially be a factor leading to an elevated accretion of Aβ [13, 14].

The marked reduction of pineal melatonin production in the aged has negative effects on neurophysiology that are independent of its actions related to removing Aβ via the glymphatic drainage (Fig. 3). Melatonin, as well as its metabolites [74], are relentless antioxidants that both directly extinguish ROS/RNS [75–77] as well as enzymatically remove them from the intracellular environment [78]. Harnessing these destructive reactants preserves neural and glial physiology as exemplified by the ability of the indoleamine to reduce oxidative stress in CNS tissues resulting from ischemia/reperfusion injury [79], neuroinflammation [80], heavy metal toxicity [81], trauma [82], and in many other situations. This action of melatonin relates most often to its preservation of endoplasmic reticulum health and mitochondrial integrity [83]. Mitochondria, in particular, are believed to be major sites at which melatonin impacts cellular physiology [84]. The faltering function of these organelles plays a critical role in neurodegenerative diseases [85].

In addition to serving as a component of the CSF for toxin removal, such as Aβ from the AD brain, melatonin has a variety of other actions by which it may delay progression of this debilitating condition (Fig. 5). The neural deposition of the 39–43 amino acid Aβ fragments are cleaved from the amyloid precursor protein (APP) by the action of secretases, with the accumulation of these fragments causing synaptic loss, molecular trafficking problems, inflammation and eventually neuron death. More than two decades ago it was shown that melatonin alters the cleavage of the APP [86]. Follow-up mechanistic studies have shown that melatonin enhances non-amyloidogenic processing and counteracts amyloidogenic modification of APP in cultured neurons by enhancing α-secretases causing the downregulation of the translation of β- and γ-secretases thereby reducing the formation of a major pathological feature of AD, namely toxic senile plaques [67].

Fig. 5.

A summary of the multiple processes modulated by melatonin that impact the development and progression of Alzheimer’s disease. Many of these actions are known also to apply to other age-related neurodegenerative conditions, e.g., Parkinson’s disease, the progress of which is slowed by melatonin treatment in experimental animal models of this condition. Reduction of amyloidogenesis. APP amyloid precursor protein, GSK-β glycogen synthase kinase-β, PS1 phosphatidylserine 1, P13K/Akt phosphoinositide-3-kinase/protein kinase B, PP2A protein phosphatase 2, PLC/DAG phospholipase C/diacylglycerol. Reduction in tauopathy. pTAU phosphorylated tau, NFT neurofibrillary tangles. Inhibition of apoptosis. Bax BCL2-associated X protein, MPTP mitochondrial permeability transition pore, Cyt c cytochrome c, AIF apoptosis-inducing factor, Bcl-2 apoptosis regulatory protein family. Scavenging of ROS/RNS. O₂⁻ superoxide anion radical, H₂O₂ hydrogen peroxide,^.OH hydroxyl radical, NO^. nitric oxide, ONOO⁻ peroxynitrite anion, ¹O₂ singlet oxygen, LOO^. lipid peroxyl radical. Modulation of antioxidants. MnSOD mitochondrial superoxide dismutase, CuSOD cytosolic superoxide dismutase, GPx glutathione peroxidase, CAT catalase, γGC gamma glutamyl cysteine synthetase. Inhibition of inflammation. IL-1β interleukin-1β, IL-6 interleukin-6, TNF-α tumor necrosis factor-α, GFAP glial fibrillary acidic protein, TRL-4 toll-like receptor-4, COX-2 cyclooxygenase-2, iNOS inducible nitric oxide synthase, MPO myeloperoxidase, LPO lipoxygenase. Red down arrows indicate downregulation, Green up arrows indicate upregulation. Because melatonin influences these processes as indicated, it protects against neuronal loss and improves neurobehavior and cognition. The loss of melatonin in the aged is believed to contribute to these processes thereby exacerbating neurodegenerative decline

Cell and transgenic rodent models of AD have provided clues of the possible mechanisms by which melatonin limits AD pathophysiology [87, 88]. In a transgenic mouse (Tg2576) model, melatonin inhibited the generation of β-sheets and reduced profibrillogenic activity due to apolipoprotein E4 [89, 90]. Moreover, melatonin reduced the accumulation of β-amyloid and/or increased its clearance as mice aged and also prolonged their survival [49]. In the context of the current review, the results reported by Pappolla et al. [91] are the exclusive data from animal studies related to the role of melatonin in clearing Aβ from the brain. Using the transgenic mouse model, which over-expresses APP, melatonin treatment led to an increased accumulation of Aβ in the cervical lymph nodes; specifically, soluble monomeric Aβ40 and, to a lesser degree, Aβ42. These lymph organs are the first to receive the toxin-laden lymph from the dural lymphatic vessels that drain the cerebral/cerebellar cortices via the glymphatic system [1, 3, 13, 14, 92]. In the non-transgenic, senescence-accelerated rat which develops sporadic signs of AD [93], similar to the majority of the AD cases that occur in humans, orally-administered melatonin prevented the neural pathologies including the aggregation of Aβ, synaptic degeneration, neural apoptosis and tau protein hyperphosphorylation [94].

Another feature potentially linking melatonin to Aβ removal from the neuropil is age. It has already been mentioned that blood and CSF melatonin levels become markedly diminished in many individuals as they age [13, 14, 95]; the reduction in endogenous melatonin production is often surmised to exaggerate the signs of aging and especially the progression of age-related diseases [96, 97], including AD [98, 99]. Based on available experimental evidence, the diminished melatonin availability would reduce Aβ clearance leading to a more rapid accumulation of the neural toxin which would contribute to the advancement of AD. This, in combination with the degenerative changes that the glymphatic system undergoes, as described in the previous section, could also be instrumental in the rate of AD development. In elderly mice, the number of dural lymph vessels as well as their diameter shrink by 20% and 35%, respectively, which contributes to a more than 50% reduction in the accumulation of toxic macromolecules in the cervical lymph nodes [100]. Also, the clearance of injected Aβ into brain of old mice is 40% lower than the removal of this toxin when injected into the CNS of young animals, perhaps in part related to the reduced amount of melatonin-rich CSF in the periarterial spaces that entered the brain parenchyma as ISF [59]. Clearly, the brain drainage network is severely compromised with advanced age, which also involves reduced melatonin levels. Whether either exogenous melatonin treatment late in life would preserve the number or the patency of the lymph vessels in the meninges has not been examined.

That supplemental melatonin may have value in altering the progression of AD in the human has often been suggested [68]. Especially in the aged, sleep disruption and weakening of the circadian system are common [101], perhaps in part related to the depressed CSF melatonin levels which precede the manifestation of cognitive impairments [72]. Sleep inefficiency is a reflection of brain aging and also contributes to the deterioration of overall health [102]. Given melatonin’s substantial free radical scavenging activities, its depression in the elderly leads to elevated oxidative and nitrosative damage in the CNS [103] with the drainage of the functionally impaired molecules via the failing glymphatic system also being compromised [104]. Moreover, immunosenescence, which negatively impacts neural physiology [105], routinely occurs as animals and humans advance in age. Thus, a number of age-associated physiological perturbations conspire to diminish neural function in the elderly, all of which also accompany the drop in melatonin and are improved by its exogenous supplementation (Fig. 5). For example, supplementing with melatonin improves nocturnal sleep in AD patients [106], strengthens their circadian system via its chronobiotic actions [15], reduces oxidative damage in the CNS [107], and supports the immune system [108]. These changes would be expected to defer the cognitive decline in AD patients and in animal models of this disease as has been reported [109–112]. In reference to the use of melatonin to potentially defer the signs of neurological decline in AD subjects, in the cited reports the daily dose of melatonin was less than 10 mg and the maximal duration of treatment was 3 years. At this point, there have been no randomized control trials of sufficient duration in which melatonin was exclusively used as a treatment to protect against any human neurodegenerative disease, including AD. Such long-term studies are badly needed to make a judgement concerning the utility of melatonin as a protector against dementia and neurocognitive decline.

Emerging evidence also shows that melatonin influences Aβ clearance in the humans who suffer with either familial or sporadic AD. While intracellular Aβ may be removed due to its ubiquitination and proteasomal degradation, broken down by the autophagosomal system or by enzymatic degradation [47, 113], in both familial and sporadic AD patients retarded flushing of extracellular Aβ from the brain is a prominent feature. Demented patients who possess the presenilin 1 (PS1) mutation experience both accelerated Aβ generation and also depressed Aβ removal via the glymphatic system [8]; conversely, sporadic AD patients most notably exhibit lowered Aβ clearance [114]. The importance of this lies in the fact that by far the largest number of AD patients suffer with the sporadic form of the disease [115]. Clearly, efficient removal of existing Aβ from the neural interstitium is a critical factor in the delayed progression of this debilitating neurodegenerative condition. Extracellular Aβ is also endocytosed and destroyed by microglia [116] (Fig. 1, top).

Another characteristic hallmark of the AD brain that interferes with neurogenesis and function is the hyperphosphorylation of tau; this protein polymerizes with the formation of intracellular neurofibrillary tangles [117]. The development of these tangles contributes to disordering of the cytoskeleton, a feature common in neurons of AD patients [118]. Melatonin treatment interferes with tau fibril formation and if fibrils are already formed, melatonin functions in their disassembly [119]. Mechanistically, melatonin’s efficacy in targeting tau protein is in part a result of its ability to alter the physical structure of the tau protofilament and filament. Since stimulation of melatonin receptors on cultured PC12 cells prevents Aβ-mediated tau phosphorylation, it is possible that melatonin also interferes with tau hyperphosphorylation in the AD brain via receptor-mediated mechanisms; neurons containing melatonin receptors are widespread in the CNS [120, 121]. A comprehensive evaluation of the role of melatonin in deferring AD pathology in both preclinical and clinical studies has been recently published [122]

Melatonin protects the brain from oxidative stress

Free radical damage is common in the CNS since the brain uses large amounts of oxygen (an estimated 20% of the inhaled oxygen although the brain is only 2% of the body weight). The associated oxidative stress is a major aspect of the pathogenesis in many neurodegenerative conditions. Examples include AD [110], Parkinson disease [123], amyotrophic lateral sclerosis [124], Huntington disease [125] and others, all of which exhibit extensive molecular damage inflicted by partially reduced highly reactive oxygen and nitrogen-based reactants. Melatonin, as well as the metabolites that are formed when it neutralizes toxic radicals, are powerful inhibitors of oxidative stress due to their ability to directly scavenge ROS/RNS as well as stimulate antioxidative enzymes [74, 126] (Fig. 5). Melatonin easily enters the neural parenchyma [127] and invariably reduces oxidative damage more effectively than classic antioxidants [128, 129] which are often excluded from the brain due to their inability to readily traverse the blood–brain barrier. This and other actions by which melatonin functions in protecting the brain from molecular destruction which result in cognitive and neurobehavioral deficits are summarized in Fig. 5. The maintenance of high levels of melatonin in the CSF, which enters the brain from the periarterial spaces surrounding penetrating arteries (Fig. 1, bottom), could play a major role in forestalling the loss of neurons and glia that normally accompanies advanced age.

The persistent decline of blood and CSF melatonin in aging humans [62, 72, 130] is associated with a concurrent drop in the total antioxidative capacity of the blood (Fig. 6). Hence, the ability of elderly humans to combat free radical damage may be diminished, thus, increasing the likelihood of neurological oxidative stress and dementia development [131]. At this point, melatonin is the only physiological molecule known to both potentially prevent Aβ accumulation and to drain it from the CNS as well as actively improve the reductive environment of neurons and glia. Given the low cost and the essential absence of toxicity of melatonin over a very large dose range, it would seem prudent to test whether melatonin supplementation would influence the progression of the disease in individuals diagnosed with AD or individuals in a high-risk population as well as against other neurodegenerative conditions.

Fig. 6.

Changes in blood melatonin levels and the total antioxidant status (TAS) of the blood throughout the life of humans. 120 volunteers participated in this study; they ranged in age from 2 to 80 years and were categorized by age into 10-year blocks. Daytime blood samples were collected at 1300 h and nighttime samples at 0100 h. High nighttime levels of melatonin correlate with a greater TSA of the blood and both parameters diminish dramatically in the aged. Error bars were not included for figure simplification. Adapted from [131]

The data published to date regarding the ability of supplemental melatonin to influence the development and progression of AD in humans, while suggestive [68–71, 99, 109] is not uniform in supporting this conclusion. Several narrative reviews and meta-analyses have come to somewhat different conclusions regarding the ability of melatonin to impede dementia development. Thus, after a comprehensive review of the literature, McCleery and Sharpley [132] concluded that melatonin improved sleep quality in Alzheimer’s patients but made no claims regarding its efficacy in altering the course of dementia. Similar evaluations of trial outcomes by Jansen et al. [133] and Majidazar and colleagues [134], however, report that melatonin has a demonstrated ability to defer AD progression. Seemingly major shortcomings of the performed trials are the short duration of treatment (several weeks to a couple of years) and the small doses of melatonin (2.5–10 mg daily) used. Considering the slow progression of AD, especially in its early and/or mild stages, longer term treatments would likely be necessary to either prove or disprove a relationship. Also, the dose of melatonin may be an issue. In other clinical trials where melatonin has been successfully employed to slow chronic neurological disease development, the doses of melatonin have been significantly higher and sometimes given for longer durations. Moreover, the interpretation of the results from these studies is complicated by the fact that, in addition to directly enhancing the reductive and inflammatory status of the CNS directly, melatonin also promotes sleep [135], regulates circadian rhythms [136] and influences gut microbiota [137], all of which may relate to its beneficial actions in retarding neurodegenerative pathologies.

Beyond the shortcomings of the data already mentioned which indicates a role for melatonin in brain washing, especially in relationship to Aβ disposal, critical information is lacking which would significantly strengthen what at this point is only a defensible hypothesis. Most of the data that confirm the significantly reduced levels of CSF melatonin in individuals suffering from AD come from single or several measurements obtained from patients without specific regard for time of sample collection, of the specific ventricle from which the sample was collected, or especially the lighting environment of the patient for an hour in advance of the CSF sample being obtained. Additionally, some samples were collected postmortem up to several hours after death of the patient. AD is a high complex disease which exhibits only poorly defined phases. Other than in very general terms, the Alzheimer’s stage has not been a consideration when CSF melatonin measurements were made. Finally, melatonin measurements of CSF from age-matched, normal, non-diseased human subjects which could be used as controls for samples collected from AD patients are even fewer. Obviously, studies in humans are impacted by circumstances that are not always in line with the goals of the information being sought.

Females have a significantly great chance of developing AD in comparison to elderly males. To date, however, no reproducible differences have been found relative to the amount of melatonin produced (as estimated by the amplitude of the nighttime peak), in the timing of its circadian rhythm or at the age where the rhythm begins to wane between males and females. Again, these studies are hampered by the difficulty in getting approval to collected repeated CSF samples for melatonin measurements over a carefully regulated light/dark cycle.

Other important investigations that are needed to strengthen the hypothesis related to melatonin and brain washing include examination of the correlations between aquaporin density in the astrocytic end feet that line the Virchow-Robin spaces and the transfer of melatonin into the neuropil. Moreover, while melatonin supplementation has been shown to defer Aβ accumulation in the brain of genetically sensitive rats, the specific interactions of melatonin with Aβ are unknown. The results of the animal studies regarding melatonin’s ability to reduce Aβ deposition in the brain are consistent and since melatonin diminishes concurrent with the accumulation of this protein along with preliminary data from human studies has prompted a number of investigators to conclude there is a causative relationship.

Melatonin influences the suprachiasmatic nucleus

Organism-wide behavioral and molecular rhythms are unequivocally essential for optimal health and for environmental adaptations [138–140]. This evolutionarily conserved, hierarchically organized clock network regulates the periodicity of systemic and cell biological functions and is under control of the master circadian oscillator, the suprachiasmatic nucleus (SCN) which is bilaterally located in the anterobasal hypothalamus just above the optic chiasm, as the name implies (Fig. 2B). Studies using the mouse and primates (including the human) have documented that, although the nucleus exhibits an inherent circadian rhythm, the intrinsic period of that cycle is close to 25 h and it is synchronized to 24 h due to a neural signal generated in the retinas and transferred to the SCN via the retinohypothalamic tract (RHT) [141, 142]. The retinal signal derives from a small subset of ganglion cells (the intrinsically-photosensitive retinal ganglion cells, ipRGC) which are equipped with a unique photopigment, melanopsin, that responds differentially to various wavelengths of light; this system functions independently of the visual system and is concerned with non-image forming functions [143, 144]. At the level of the SCN, the RHT signals via glutaminergic mechanisms and other processes, with wide variations in the response of neurons in different portions of the SCN.

The prevailing light:dark environment influences the circadian clock network in the SCN which in turn, via a multisynaptic central and peripheral sympathetic pathway, photoentrains melatonin synthesis and secretion by the mammalian pineal gland [145, 146]. Inhibition of the SCN by the light-mediated neural message received from the ipRGC results in downregulation of pineal melatonin production during the day. With the onset of darkness, the pinealocytes are disinhibited and melatonin production proceeds unabated with the release of melatonin into the blood and CSF accounting for the circadian rhythm in these fluids [147, 148].

Elevated nighttime melatonin aids in synchronizing clock networks in the SCN [149, 150]. After responding to a variety of inputs, the SCN sends neural and endocrine signals to peripheral slave oscillators in all cells for the regulation of physiological and behavioral responses, making them temporally appropriate [151, 152]. While melatonin receptor distribution in the CNS varies markedly among brain regions and species [153] in most mammals the SCN is rich in high-affinity, G-protein coupled MT1 melatonin receptors [154, 155] which mediate many of melatonin’s actions on intrinsic pacemaker mechanisms of the master circadian oscillator. There is an additional suggestion, however, that a direct receptor-independent interaction of melatonin with the proteasome of SCN cells could explain how melatonin influences the temporal variations in molecular timing genes [156]. The prevailing light:dark environment is a major driver of the electrical and metabolic activities of the SCN, which result in the ability of especially short wavelength light to impact phase shifts, etc.

Melatonin released from the pineal gland has feedback effects at the level of the master circadian clock [157, 158]. Since the blood melatonin rhythm has been well known for almost five decades and it has been described in detail in many mammalian species, it also was a foregone assumption that circulating blood melatonin levels were responsible for its influence on the clock network in the SCN and elsewhere [159, 160]. The rhythm of melatonin in the blood, however, generally increases rather slowly after darkness onset, peaks near mid-dark period and wanes in the latter half of the night. By comparison, the circadian CSF melatonin rhythm, in addition to often being of greater amplitude than in the blood, increases rapidly after darkness onset, remains at a relatively stable high level during the night, and falls precipitously at light onset [10] (Fig. 4). These nuances of the CSF melatonin rhythm make it a better candidate for serving as a synchronizer of the circadian clock, the SCN; since it more precisely identifies darkness onset and offset, we proposed that rather than blood melatonin levels it is the circadian CSF melatonin rhythm that aids in synchronizing the master circadian clock. Melatonin could easily gain access to sensitive SCN neurons, which possess the MT1 receptors, either by merely diffusing from the CSF in the supraoptic recess to the nearby responsive cells or by being transferred from the CSF to the SCN via tanycytes [161, 162] (Fig. 2, B). When exogenous administration of melatonin causes blood levels to rise significantly above those in the CSF, they presumably override the message provided by the simultaneously lower CSF melatonin values and then serve as a chronobiotic at the level of the SCN [38, 163]. This explains how peripherally injected melatonin can function as a pharmacological agent to influence circadian activities of the SCN.

There is a plethora of studies documenting the relationship between SCN and neurodegenerative diseases [122, 164, 165]. For instance, functional deterioration of SCN has a critical role in the AD pathophysiology and promotion of sleep disorders [166]. By analyzing immunohistochemically the SCN of aged and young subjects, Wu and colleagues [84] reported a significant drop in the number and density of MT1-expressing neurons in aged control; also, they identified a very low MT1 expression in the last neuropathological stages of AD (Braak stages V–VI). Maintenance of a stable circadian melatonin rhythm into advanced age would act on SCN neurons thereby reducing neuronal loss while preserving circadian rhythms throughout the body; preserved circadian rhythmicity is beneficial feature for delaying neurocognitive decline [186–188].

Melatonin regulates neural stem cell proliferation

In a mouse model that exhibits a tauopathy, mitotic activity is depressed in the two areas of the brain where active cell division normally takes place, that is, the subgranular zone (SGZ) of the hippocampus and the subventricular zone (SVZ) of the lateral ventricle [167, 168] (Fig. 2, C). While not specifically documented in this animal model, under non-pathological circumstances melatonin preserves neuroprogenitor cell proliferation in these zones [169, 170]. This is important since the cells renewed in these areas migrate to distant sites and replaced lost neurons due to some neuropathologies although the efficacy of replacing degraded neurons may vary with disease type [171–173]. In 2017, Mendivil-Perez and colleagues [174] documented melatonin’s ability to differentiate neural stem cells (NSC) into neurons and oligodendrocytes mainly through an increase in mitochondrial functions such as mitochondrial mass/DNA complexes, membrane potential, mitochondrial respiration, and ATP synthesis.

Melatonin in the CSF is directly involved in maintaining the proliferation of the NSC in these areas, especially in the SVZ niche. The cytoarchitecture of the neurogenic SVZ is such that it monitors the concentration of chemical messengers in the CSF (Fig. 2C). Processes of some SVZ astrocytes pass between the overlying ependymal cells and are in contact with the CSF [175]. Also, pharmacological evidence suggests the presence of melatonin receptors on the cells in the SGZ [170]. Finally, recent studies strongly support the involvement of pineal derived-melatonin in regulating cellular activity of the SVZ [176]. Using a minimally invasive in vivo technology in freely moving mice, this group followed the activation kinetics of the cells of the SVZ over a 2-month period. The mitotic activity of the cells was high during the day but the cells became relatively quiescent at night, with the darkness-induced inhibition being related to melatonin signaling. In mice maintained under continual darkness, cell division in the SVZ remained low over the 24-h period; conversely, under constant light conditions, the authors reported a marked rise in mitotic activity. The ability of melatonin to affect cell division in the SVZ relies on functional membrane melatonin receptors. We propose that the circadian melatonin rhythm in the lateral ventricles drives the rhythm in SVZ cell division. Experimental evidence had already revealed that the role of melatonin in regulating proliferation, survival, and differentiation of NSCs is dependent on histone acetylation, neurotrophic and transcription factors, MAPK/ERK signaling pathway, and apoptotic genes [177].

The ability of melatonin to reduce cellular proliferation in the SVZ [198] may also have implications for the initiation of the highly malignant and aggressive tumor, glioblastoma (grade 4 astrocytoma). This tumor type most often occurs in the elderly, subjects in which the nocturnal rise in melatonin is commonly severely dampened [62, 95, 130]. Several authors have either proposed or shown that melatonin dose-dependently inhibits the viability and migration of in vitro and in vivo glioblastoma cells and causes them to undergo apoptosis [178–181]. A mechanism proposed for this action involves melatonin’s upregulation of miR-16-5p which leads to a suppression of the proto-oncogene PIM1 [182]. Although the cell of origin of a glioblastoma remains a matter of debate, the stem cells of the SVZ have attracted the greatest attention [183]. Since the proliferative activity of these cells is controlled by melatonin [176], the CSF concentration of this indoleamine should be a focus of future research related to glioblastoma [184]. This cancer type is most common in the elderly, where endogenous melatonin levels are permanently depressed presumably allowing for increased mitotic activity of the glioma precursor cells in the SVZ.

Limitations and key experiments to be performed

While the data accumulated to date relative to the functions of the CSF circadian melatonin rhythm are certainly suggestive of its likely important role in influencing systemic circadian biology, including the sleep/wake cycle, via is actions on the SCN, there are gaps in the findings. What severely complicates the accumulation of definitive data are the difficulties with performing the necessary experiments. For example, monitoring the 24-h rhythm in CSF melatonin can only be performed on large mammals, e.g., sheep, goats, etc., since the accurate placement of cannulas directly into the very narrow third ventricle of rodents is difficult and, moreover, the amount of CSF that could be obtain from this ventricle is extremely small. Also, the performance of 24-h studies in humans is also essentially precluded by the guidelines regarding the use of human subjects; moreover, if the study could be accomplished, it would probably be a single individual and there would be confounding physiological conditions.

For CSF melatonin to influence the clockwork of the SCN, it would likely have to be transferred via the tanycytes into the nucleus. Whereas the tanycytes have been shown to transport many other molecules to and from the third ventricle, this has not been specifically shown for melatonin. The same shortcoming applies regarding the transfer of melatonin from the Virchow-Robin periarterial spaces into the neuropil. That this occurs is supported by the observations that melatonin treatment increased and Aβ content of the deep cervical lymph nodes which drain CSF from the glymphatic network. Confirmation of these findings using different approaches would reinforce the theory that CSF melatonin, via the glymphatic system, serves the function of “brain washing”.

It has long been felt that it is the blood melatonin cycle that has a feedback effect on the SCN to influence circadian biology. To prove that CSF melatonin is capable of also doing so. the peripheral and CSF melatonin rhythms could be extinguished by either surgically removing the pineal gland or exposing experimental animals to constant light. This could be followed by the controlled perfusion of melatonin into the third ventricle accompanied by monitoring circadian biology would aid in answering this question.

Safety of melatonin

Given that all living organisms, animal and plant, synthesized melatonin throughout most of their life, at physiological concentrations this important constituent is clearly not harmful. Indeed, it seems more likely that the reduction of melatonin in the aged is a contributor to senescence and functional decline since in both kingdoms lowest levels of melatonin correlate with changes that are generally associated with aging while supplementing with melatonin delays the associated processes [185, 186]. Even the issue of what constitutes a physiological level of melatonin is debated; normal levels are usually proposed to be those present in blood, which are in the low pg/mL range. However, melatonin is not in equilibrium in the organism with other bodily fluids typically having concentrations much higher than in simultaneously measured blood levels. Moreover, intracellular concentrations of melatonin exceed extracellular levels and may vary among organelles [187, 188]. It is assumed that the administration of exogenous melatonin to animals including humans invariably induces pharmacological concentrations in the blood at least in the short term. Melatonin has been extensively used by humans in pharmacological doses for long durations (years), e.g., for sleep. In the clinical trials where melatonin has been taken daily for one to three years have not uncovered any toxicities beyond some minor subjective parameters of sleepiness, headache, etc. [189–191]. Moreover, extremely large doses (300–1,000 mg daily) taken for shorter durations up to one month have revealed no substantial toxicities [192, 193]. Finally, drug interactions are generally not a serious consideration since every species used to examine such associations already contain melatonin when they receive the test drug. Occasional idiosyncratic reactions to melaotnin supplementation have been reported, i.e., diarrhea.

Concluding remarks and perspective

The multiple actions of melatonin in potentially preserving motor and cognitive skills have received extensive interest and investigation in the last two decades [67, 194]. Only recently, however, has melatonin circulating in the CSF gained significant investigative traction in reference to neural health [13, 14, 35, 195]. The latter subject became of interest when it was discovered that melatonin is released directly into the third ventricular CSF where is in high concentrations [10, 12, 21]. This finding coupled with the recent discovery of the movement of melatonin-rich nighttime CSF through the subarachnoid space and the glymphatic network from where melatonin likely has access to the neural parenchyma provided an additional stimulus for this route being a means by which it protects the brain from the conventional aged-related dementias. From the periarterial space melatonin moves through the blood–brain barrier (BBB), which is formed by endothelial cells, pericytes and astrocytic endfeet. The BBB excludes many molecules from entering the neural parenchyma but it is no impediment to melatonin, which readily crosses this morpho-physiologial barrier.

The CSF melatonin rhythm also impacts the circadian molecular network of the suprachiasmatic nucleus to conserve optimal circadian rhythmicity. Well-regulated circadian rhythms are essential in deferring mental deterioration. When used experimentally, melatonin proved to be a powerful reducing agent that protects the brain from oxidative damage and cellular loss in many neurodegenerative models including AD, PD and others [196, 197]. Because of its high requirement for oxygen, the CNS is subjected to an unrelenting and self-propagating bombardment by toxic partially-reduced oxygen and nitrogen species which eventually leads to the senescence and/or death of neurons which are post-mitotic [42, 44, 80, 84]; this illustrates the potential negative effect of the drop of endogenous CSF melatonin in the aged and underscores the need to maintain high levels of this neuroprotective agent for the purpose of delaying neurobehavioral deterioration.

Healthy mitochondria are a requirement for optimal bioenergetics of neurons and glia as well as for maintaining redox homeostasis and proper cellular signaling functions [198]. The neural accumulation of amyloid-β and p-tau compromises mitochondrial physiology, which contributes to the pathophysiological characteristics of AD. Dysfunctional mitochondria also typically generate increases amounts of ROS/RNS that create further damage which leads to loss of neurons common in the aged and diseased CNS [199, 200]. Melatonin is a well-known protector of mitochondrial physiology, metabolism and dynamics [201–203] including those in neurons and glia. Melatonin filtering through the neuropil enters brain cells and their mitochondria where it supports the Krebs cycle, oxidative phosphorylation and ATP synthesis [84, 201, 203]. In addition, since mitochondria, especially in the aging brain, produce an over-abundance of damaging partially-reduced oxygen species, melatonin protects this tissue from molecular destruction due to its direct free radical scavenging actions and well as its ability to stimulate antioxidative enzymes [44, 74, 76].

Intrathecal drug delivery is a common intervention for conditions such as chronic pain [204, 205]. These injections have not yet included melatonin although it is a well-studied analgesic [206, 207]. Moreover, in light of the remarkable recent observations that CSF from young animals when injected into their old counterparts improves their memory [208], the possibility that the addition of melatonin to the intrathecally injected CSF would also rejuvenate some aspects of brain function is worth considering. The restoration of memory involved an enhanced oligodendrocyte activity and myelination [208], actions well documented for melatonin [209]. Finally, whereas the beneficial neural functions of melatonin may be primarily relegated to the period of darkness since only then is CSF melatonin elevated, it is possible that vitamin D, which is synthesized during the day, may serve this function during the daily light period [210, 211].

Funding

This research did not receive any specific grant funding from agencies in the public, commercial or not-for-profit sectors.

Data availability

Enquiries about data availability should be directed to the authors.

Declarations

Conflict of interest

None of the authors declare a conflict of interest.