Melatonin as a Novel Interventional Candidate for Fragile X Syndrome with Autism Spectrum Disorder in Humans

In terms of social interaction and communication, ASD is the one of the most frequently studied developmental disabilities. Children with autism commonly exhibit stereotyped behaviors within the framework of restricted and repetitive interests. Although a number of studies have attempted to elucidate the causes of autism, an exact etiology has yet to be clearly defined. Additionally, because autism is associated with several complex conditions that involve genetic predispositions and environmental triggers, a clear treatment strategy has not been suggested either. The ASD diagnosis was proposed in the fifth revision of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; American Psychiatric Association 2013 [51]) because children with autism do not show a uniform set of symptoms but, rather, a unique constellation of features particular to each individual. Therefore, the older term “autism”, which depicts a specific category of diagnoses, is being replaced by the newer term “ASD”, which better describes a postulated spectrum disorder that encompasses multiple conditions.

FXS is the most common genetic cause of autism [52] as it affects approximately 1 in 3600 males and 1 in 4000–6000 females [53]. FXS is caused by a mutation of the fmr1 gene at Xq27.3 on the X chromosome [2]. The mutation of the fmr1 gene is induced by methylation at the frm1 promoter region and is associated with the expansion of the CGG triplet sequence in the 5′-untranslated region (UTR). As a result, FMRP levels are lower or absent (Figure 1). Depending on the triplet repeat mutation of the fmr1 gene, the fmr1 alleles are classified as normal, pre-mutation, and full mutation. In typical alleles, the fmr1 gene contains 5–54 CGG repeats (most commonly 30 repeats) while the pre-mutation alleles range from 55 to 200 CGG repeats. Additionally, pre-mutation fmr1 alleles are unstable and can become fully mutated alleles via maternal transmission [54,55].

ASD refers to a constellation of neurodevelopmental disorders that manifest with particular behavioral characteristics. Some studies have reported that severe sleep problems are more frequent in children with autism than in typically developing children [60,61,62] and that ASD children with sleep problems tend to exhibit overactive and stereotypical behaviors. Other studies have suggested that the abnormal regulation of melatonin may be related to sleep disorders in ASD because the sleep-wake cycle is related to circadian rhythms, which are modulated by melatonin. Abnormalities in the production of melatonin might be the cause of sleep disturbances because consecutive sleep disorder has been attributed to the dysregulation of melatonin synthesis, sensitization to environmental stimuli, and behavioral insomnia syndromes. Several sleep studies have suggested that there is a correlation between sleep problems and the physiological roles of melatonin but the relationship between melatonin levels in the blood and melatonin synthesis in pinealocytes has yet to be clearly established.

The estimated prevalence of sleep problems in individuals with FXS with ASD is approximately 80% higher than that of the general population [63] and can lead to circadian variations and altered glucose homeostasis in this population [64]. In experimental studies, mice lacking fmr1 exhibit abnormal circadian behavioral rhythms such as a loss of rhythmic activity in a light:dark (L/D) cycle and a shorter free running period in constant darkness (DD) [9,10]. Additionally, the altered expression of the clock component has been observed in FXS animal models. The overexpression of FMRP via transfection assays increases PER1- and PER2-mediated BMAL1 (Brain muscle aryl hydrocarbon receptor nuclear translocator-like protein 1)–NPAS2 (Neuronal Per-Arnt-Sim domain protein 2) transcriptional activity [9], which suggests that FMRP is an essential component involved in the regulation of rhythmic circadian behaviors. Accordingly, Drosophila lacking the fmr1 gene exhibit altered circadian rhythms [10]. Taken together, these results indicate that fragile X-related proteins might be associated with the induction of abnormal sleep patterns in FXS due to alterations in circadian genes; they may also play a critical role in the regulation of circadian output pathways.

Morphological analyses have consistently identified neuronal abnormalities and immature dendritic spines in most fmr1 KO mice. Spines on neuronal dendrites with membranous protrusions have a bulbous head and a thin neck and most of these spines are associated with synaptic strength and/or the transfer of electrical signals to the axon terminal [65]. Dendritic spines release various receptor-related neurotransmitters and neurotrophic factors that enable synaptic transmission. Morphologically immature dendritic spines have been observed in fmr1 KO mice that may lack the expression of FMRP proteins [56]. FXS patients with abnormal dendritic spine structures tend to exhibit an intellectual disability [66] and these individuals also have morphologically distinct dendritic spines such that they are longer or shorter, thinner, and fewer than those in typically developing individuals. This type of dendritic spine dysmorphogenesis is related to intellectual disabilities [67,68], and the observation of abnormal spine morphologies in FXS mice implies that their characteristics during early development can be identified [69]. In FXS, structural and functional abnormalities in dendritic spines are induced by the silencing of the fmr1 gene, and the resultant absence of FMRP may alter the morphology and synaptic number of dendritic spines [58,70] (Figure 1). The specific role of FMRP synthesized near synapses is associated with the regulation of synaptic structure and function [59]. FMRP, which is an mRNA binding protein, stimulates the synthesis of synaptic proteins by influencing synaptic plasticity. Thus, the loss of FMRP in FXS due to fmr1 gene silencing may imply the presence of neurodevelopmental abnormalities.

Melatonin is a circadian synchronizer that is predominantly synthesized in the pineal gland at night. A major role of melatonin involves the regulation of biological signals associated with the L/D cycle. Many studies have demonstrated the beneficial effects of melatonin, its antioxidative and neuroprotective effects, and its involvement in neuronal plasticity and network remodeling. Melatonin synthesis begins during periods of darkness via the serotonin/N-acetyl serotonin (NAS)/melatonin pathway. First, the amino acid tryptophan is uptaken into the pineal gland, then tryptophan is converted into serotonin, which is converted into NAS by N-acetyltransferase (AANAT), and, finally, NAS is converted into melatonin by acetylserotonin N-methyltransferase (ASMT [71]).

Clinical studies have reported decreased levels of melatonin in the blood of individuals with FXS and ASD [19,20]. However, other studies have reported the overproduction of melatonin, which may occur to compensate pineal gland overstimulation following increased sympathetic nervous system activity [63], such as occurs with FXS [72]. Melatonin deficiencies are caused by dysfunction in its synthesis and are reflected in altered circadian rhythms (Figure 2). Sleep disturbances may be caused by significantly lower levels of melatonin as well as by significant decreases in AANAT, which is responsible for converting serotonin into NAS. Because melatonin has anxiolytic effects, ASD in conjunction with an impaired serotonin/NAS/melatonin pathway may result in circadian problems. Recent studies have reported that disruptions of the serotonin/NAS/melatonin pathway are highly sensitive and may be a useful biomarker for ASD [73,74].

In individuals with FXS, deficits in neuronal plasticity lead to problems in learning, memory, and cognition. Recent studies have reported that FMRP modulates the number, function, and maturation of synapses and is associated with protein synthesis-dependent synaptic plasticity [71]. Because FMRP is an influential regulator of protein synthesis in dendrites, the synaptic changes associated with synaptic plasticity can be observed in FXS [75] (Figure 2). In particular, FMRP is localized in neuronal dendrites and synapses where it is thought to play a role in the regulation of local protein synthesis, such as for metabotropic glutamate receptors (mGluRs) via mRNA trafficking; thus, FMRP regulates mGluR activity as a shuttle of its mRNA [74,75]. On the other hand, decreases in FMRP expression influence long-term synaptic enhancement and are highly correlated with developmental disorders.

The activation of mGluRs induces the postsynaptic internalization of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPA) receptors, which is modulated by the rapid translation of proteins involved in long-term depression (LTD) [76]. Lasting patterned stimuli affect neuronal synaptic plasticity achieved by both long-term potentiation (LTP) and LTD and these long-lasting enhancements tend to improve learning and memory. LTP alters synaptic connectivity by changing the morphology of postsynaptic neurons, while LTD lowers postsynaptic receptor density and leads to the elimination of old memories, which allows for the formation of new connections via the LTP process. Both of these dynamic processes are necessary to maintain the efficient development of synaptic networks and allow for the continuous receipt of new information by eliminating older less important memories. However, excess LTD proteins have been reported in mouse models of FXS, which is important because FMRP acts as an inhibitor of the translation of LTP proteins. Because the activation of mGluRs triggers the over-synthesis of LTD proteins due to a lack of FMRP, neuronal synapses with elongated and weak spine morphologies are expressed in the hippocampus and cerebellum [77]. Taken together, these findings indicate that abnormal neuronal synaptic structures negatively impact synaptic plasticity. Thus, it is possible that the inhibition of mGluRs contributes to the suppression of mGluR/LTD signaling rather than FMRP.

Experimental studies have observed excessive mGluR activation in fmr1 KO mice [78]. Under normal conditions, FMRP is thought to be a translational repressor and to negatively regulate mGluR [79] because it is an RNA-binding protein involved in the transcriptional regulation and transport of specific mRNAs [80]. Additionally, FMRP can be highly localized in the cytoplasm of neurons and dendritic spines and acts to regulate the translation of ribosomes [81]. However, under dysfunctional FMRP conditions, mGluR activity is altered by disruptions in intracellular signaling and, subsequently, the absence of FMRP accelerates excessive mGluR5 signaling [82]. In turn, excessive mGluR5 may inhibit melatonin synthesis, as reported previously [83]. Recent studies have proposed the existence of abnormalities in melatonin secretion and circadian patterns in individuals with FXS with ASD that are likely to be due to excessive signaling via mGluRs. These receptors are a type of G-protein coupled receptor (GPCR) that can be classified into three groups (Groups I, II, and III) based on receptor structure and physiological activity [83]. mGluR Group I, which includes mGluR1 and mGluR5, is coupled to the Gq-protein subtype that activates phospholipase C [84]. mGluR Group II includes mGluR2 and mGLuR3, while Group III includes mGluRs 4, 6, 7, and 8; these groups are negatively linked with the Gi- and Go-protein subtypes, which inhibit adenylyl cyclase and suppress the formation of cyclic adenosine monophosphate (cAMP) [85].

Norepinephrine (NE)-dependent melatonin synthesis, which plays a role in the regulation of circadian rhythms and alleviates epilepsy, is suppressed by the release of glutamate [86]. mGluR Group II, particularly mGluR3, negatively regulates melatonin synthesis in pinealocytes [87] and it is known that mGluR3 and mGluR5 are expressed in pinealocytes and are involved in the negative regulation of melatonin synthesis via the inhibition of cAMP cascade. Similarly, Group II agonists suppress melatonin synthesis and prevent AANAT activity in the rat pineal gland [88]. Thus, irregularities in the synthesis of melatonin in FXS may be linked to the absence of fmr1, which regulates the expression of FMRP and glutamate receptors.

Sleep disorders are common in patients with neurological diseases [89,90,91]. Additionally, the modulating properties of melatonin in terms of sleep patterns and circadian rhythms are associated with the development of ASD [92,93]. Individuals with ASD appear to be susceptible to sleep disorders and survey research has indicated that the prevalence rate of sleep issues can be as high as 89% in children with ASD and as high as 77% in children with FXS with ASD [63]. Melatonin supplements regulate the sleep-wake cycle and have been shown to alleviate sleep problems in clinical research studies. Recent studies have shown that levels of melatonin or melatonin metabolites are significantly lower in children with ASD than in typically developing children [17,19]. Melatonin is an endogenous neurohormone produced in pinealocytes, which are neuroendocrine cells, and is widely used in clinics to treat insomnia in children because it is inexpensive, efficient for treating sleep problems, and has no side effects [94]. For example, a study investigating the efficacy of melatonin in 107 children with autism (age range: 2–18 years old) found that 25% of the treated children no longer reported sleep problems and that 60% of the treated children reported improvements in sleep [95]. It has also been shown that autistic children have lower melatonin levels than typically developing children [96].

Chronic sleep disorders are typically associated with learning and behavior issues in ASD individuals, and the functional consequences of abnormal melatonin levels in individuals with FXS may also include learning and memory problems. In fmr1 KO mice, abnormality of dendritic spines can be seen [97], which is known to be harmful to memory function [98]. However, several studies have reported that melatonin facilitates synaptic plasticity and enhances the mechanisms underlying learning and memory [24,99,100]. Taken together, these findings suggest that there is a correlation between the loss of neuroplasticity and the malfunctioning of, or irregularities in, melatonin production in FXS.

Low melatonin levels are associated with altered activity in the GABAergic system [13]. γ-aminobutyric acid (GABA) is the main inhibitory neurotransmitter in the central nervous system (CNS) and, accordingly, induces the relaxation of the brain and sleep. Melatonin stimulates GABAergic activity in the brain and, thus, abnormal melatonin levels influence the onset and length of sleep [101]. Similarly, alterations in the circadian clock mechanism due to abnormal melatonin synthesis can affect the sleep-wake cycle. Recently, studies using animal models of autism have indicated that clock and clock-related genes may interact in the ASD phenotype and studies using fmr1 KO mice have implicated clock proteins in sleep alterations in FXS. Furthermore, melatonin is helpful for treating sleep problems in cases of autism with oxidative stress and the physical alterations of axons and dendritic spines [92,102].

Individuals with FXS have a higher risk of neurological diseases, such as seizures, which are an important characteristic of autism and significantly associated with FXS. Clinical survey data have revealed that epilepsy occurs in 10–20% of children with FXS [103]. However, a recent study found that melatonin, which is used to treat sleep disorders and does not cause side effects, can effectively regulate severe epilepsy [104] as well as suppress its incidence.

Epilepsy is a neurological disease that is accompanied by biochemical responses resulting from brain injuries and chemical imbalances. In particular, free radicals are linked to seizure initiation [105]. The excessive production of free radicals contributes to brain damage in patients with neuropathological conditions such as stroke, Alzheimer’s disease, and Parkinson’s disease [106]. Oxidative stress occurs in mitochondria following seizures, which constitute a primary cause of oxidative stress that is critical to neuronal cell fates. Recently, melatonin was discussed in relation to epileptic seizures [107,108,109] because this neurohormone is known to act as an antioxidant and free radical scavenger. A large study reported that melatonin may be a promising anticonvulsant candidate due to its capability to induce antiepileptic activity, and a clinical study showed that low baseline levels of melatonin are observed in patients who suffer from epilepsy but that these levels dramatically increase after a seizure. Although these findings indicate that melatonin may play a role in the regulation of seizures, other studies have reported that melatonin might actually increase the risk of seizure. The latter findings are supported by evidence showing that melatonin may affect hippocampal excitability and can increase one’s susceptibility for seizures by lowering the seizure threshold [110]. Thus, the role of melatonin in epilepsy remains under debate.

Recently, it was reported that melatonin acts as a neuroprotective agent against neurological injuries [111,112,113,114]. The neuroprotective effects of melatonin have been demonstrated by in vivo studies that co-administered injections of melatonin and dexamethasone, the latter of which is known to be neurotoxic to cells in the hippocampus. The animals that received these injections exhibited decreased numbers of abnormal hippocampal cells and different histological properties compared to the vehicle group. Melatonin is an endogenous neurohormone that regulates several biological functions but exogenous melatonin has also been shown to prevent neuronal cell death and improve cognitive dysfunction [115]. The neuroprotective effects of melatonin have also been demonstrated in cases of acute global cerebral ischemia and hypoxic ischemia [116]. Most cerebral ischemic models produce a significant loss of neurons in the hippocampal CA1 to CA4 regions. In the case of children who experienced perinatal hypoxic ischemia, the resulting brain injuries were directly caused by neuronal cell death and then indirectly by chronic neuropathic conditions such as cerebral palsy, intellectual disabilities, and epilepsy [117]. On the other hand, brain damage due to cerebral ischemia is primarily induced by reductions in the blood supply and treatment may include oxygen supplementation. When the blood supply becomes blocked, the cellular metabolic system converts to anaerobic metabolism and, consequently, depletions in adenosine triphosphate (ATP), the accumulation of lactic acid, and the cellular input of calcium may occur.

However, melatonin treatment can protect the brain from damage and delay neuronal cell death. During reperfusion after cerebral artery occlusion, the overproduction of free radicals triggers the activation of oxidative stress but melatonin may suppress this type of brain damage. Because melatonin contributes to various forms of neuroplasticity, including learning, memory, and recovery from brain damage, the manner in which this hormone acts as a neuroplastic agent after autism should be elucidated. Interestingly, the delayed stabilization and abnormal morphological features of dendritic spines are the main characteristics of FXS, and these are related to impaired synaptic signaling and connections. Disruptions in the pruning of excitatory synapses and hyperconnectivity have been observed in FXS patients and the fmr1 KO mouse model due to loss of postsynaptic FMRP [118]. Deficits in FMRP in conjunction with incomplete pruning induces cell-to-cell hyperconnections in synapses. Moreover, the connection pruning process that occurs during early development is essential for the formation of normal neuronal circuits. In contrast, in FXS, FMRP dysfunction results in hyperconnectivity and an excessive number of synapses that lead to autistic features. Melatonin has been associated with neurogenesis and microtubule polymerization in dendrites and, thus, melatonin may stimulate dendrite maturation and affect neuroregeneration.