Intellectual disability (ID), is defined as IQ < 70 and is associated with functional deficits in adaptive behavior, such as daily-living skills, social skills and communication. It affects 1-3% of the general population and results from heterogeneous environmental, chromosomal and monogenic causes 1 . Besides the cognitive deficit, patients with ID often present with sleep disorders, which are persistent and a burden for the patients and their families. The most commonly reported disorders are delayed settling to sleep and frequent waking at night, with frequencies ranging from 58 up to 86% 2 3 4 . Several therapeutic strategies have been considered for treating sleep disturbances in ID. Among them, pharmacological use of melatonin was demonstrated to be efficient in several studies, recently reviewed in a meta-analysis 5 : exogenous melatonin appeared to decrease sleep latency and number of wakes per night, and increase total sleep time in individuals with ID. This positive effect of melatonin treatment could suggest that, in some patients with ID, endogenous melatonin level may not be sufficient to adequately set the sleep-wake cycles.

Melatonin is considered as a major biological signal of day-night rhythms, and thus a major endogenous "Zeitgeber" (time-giver). It is synthesized in the dark in the pineal gland from serotonin, first acetylated by aryl alkylamine N-acetyltransferase (AA-NAT) and then converted into melatonin by acetyl serotonin methyl transferase (ASMT also known as hydroxyindole O-methyltransferase or HIOMT). Besides sleep induction and circadian rhythms regulation, melatonin is also involved in various other physiologic functions, including immune response, antioxydative defense, metabolic regulations and memory 6 7 8 . Abnormal melatonin synthesis or signaling was reported as a risk factor for diverse medical conditions such as diabetes, circadian and psychiatric disorders 9 10 11 12 13 14 . Among these, autism spectrum disorders (ASD) - which are also often associated with ID and with sleep disorders - have been associated with low melatonin levels in at least four independent studies 15 16 17 18 . Melke et al 17 showed that melatonin deficit in patients with autism is correlated with low activity of the ASMT enzyme, and, in some patients, associated with mutations in the ASMT gene. This study provided the first insight into a molecular mechanism for melatonin deficit associated with neurodevelopmental disorders.

We hypothesized that patients with ID could carry deleterious ASMT mutations. If this was the case, these mutations might act as risk factors for sleep/circadian disorders and subsequently exacerbate the effect of independent genetic/environmental causes of ID. To address this question, we first screened the ASMT gene for rare variants in 361 patients with ID and 440 controls. For patients carrying ASMT mutations, we then measured the ASMT activity in B lymphoblastoid cell lines (BLCL) and if available provided information on sleep.

Alterations of the melatonin pathway have been suggested as susceptibility factors to developmental disorders, and especially to ASD 15 16 17 19 20 21 22 23 . The mechanisms leading to ASMT/melatonin deficit in humans are most likely diverse, including genetic/epigenetic alterations. The impact of melatonin deficit on sleep and on the susceptibility to developmental disorders (such as ASD or ID) also remains unclear. It may involve its role as a circadian synchronizer and sleep inducer, its effects on synaptic plasticity, and/or its antioxidant properties 8 14 19 24 25 . Melatonin deficit may also alter and/or desynchronize many physiological processes, and indirectly exacerbate other pathological processes.

We could identify predicted deleterious variants in a subgroup of patients with ID, including two deleterious splice site variants of ASMT found only in patients with ID. The splice site mutation in intron 7 (IVS7+1G>T) was never observed before. The splice site mutation in intron 5 (IVS5+2T>C) was previously identified in patients with ASD and reported to be more frequent in patients compared to controls (6/749 vs 1/861; p = 0.04) 17 20 22 . In addition, biochemical studies indicated that several of the variants, although present at the heterozygous state, were associated with low ASMT activity and might thus impair melatonin synthesis in vivo. These results were consistent with the biochemical studies performed by Melke et al. on families carrier of the L298F and IVS5+2T>C mutations and presenting with a dramatic decrease in ASMT activity and blood melatonin concentration 17 . Nevertheless, despite these interesting findings, we could not detect ASMT mutation enrichment in patients with ID compared to the controls. Our results were similar to those previously reported in patients with ASD, for whom no significant enrichment in ASMT rare variants was found 17 20 22 , although melatonin deficit is very frequently associated with this condition 15 16 , and is correlated with low ASMT activity in vivo 17 . Low ASMT activities were also observed in BLCL of some controls subjects who did not carry a coding mutation of ASMT. Low ASMT activity can thus be observed even in the absence of coding mutations. For example, SNPs within the promoter were associated with low ASMT mRNA levels 17 21 .

Several limitations exist in this study. First, the sample of patients with ID was initially collected for the identification of X-linked genes (e.g. families with multiple affected males). Therefore, this population is not representative of the broad diversity of patients with ID and is negatively biased for the identification of mutations in the ASMT gene, located on the pseudo-autosomal region 1 (PAR1) shared by the X and Y chromosomes. Another limitation is the sparse information that we could collect about sleep disorders. Further studies will be required to establish the precise link between ASMT variants, melatonin levels and sleep disorders.

This study does not support ASMT as a causative gene for ID since we observed no significant enrichment in the frequency of ASMT variants in ID compared to controls. Nevertheless, we could identify patients with deleterious ASMT mutations as well as decreased ASMT activity. Given the importance of sleep difficulties in patients with ID, for this subgroup of patients, melatonin supplementation might be beneficial.

The authors declare that they have no competing interests.

ASD: Autism Spectrum Disorders; ASMT: Acetyl-serotonin Methyl Transferase; BLCL: B Lymbhoblastoid cell lines; ID: Intellectual Disability; SNP: Single Nucleotide Polymorphism

SM, CP, SJ participated in the sequencing the ASMT gene in patients and controls; CP, HGB, JML measured the ASMT activity; HVE, KP, SB, FL, MR, DL, AT, VK, AdB and JC were involved in the patient's recruitments and database managements. AD, SJ and ML were involved in the control's recruitments and database managements. TB, RD and JC conceived of the study, and participated in its design and coordination. CP, RD and TB drafted the manuscript. All authors contributed to and have approved the final manuscript.