Smith-Magenis syndrome (SMS; OMIM 182290) is a genomic disorder characterized by multiple congenital anomalies, developmental delay, intellectual disability, and a distinct behavioral phenotype including aggressiveness, self-injury, and sleep disturbances [Edelman et al., 2007; Greenberg et al., 1996; Smith et al., 1986]. Most patients with SMS (~75%) harbor a recurrent 3.7 megabase (Mb) microdeletion of 17p11.2, mediated by flanking low copy repeats (LCRs) [Bi et al., 2003; Chen et al., 1997; Potocki et al., 2003; Shaw et al., 2002]. . The retinoic acid-induced gene 1 (RAI1) is one of approximately 25 genes within the SMS critical region [Bi et al., 2002; Vlangos et al., 2003], and heterozygous RAI1 mutations have been reported in individuals who resemble SMS patients clinically, but who lack a 17p11.2 deletion; thus, haploinsufficiency of RAI1 is considered the cause of most of the manifestations of SMS [Bi et al., 2004, 2006; Girirajan et al., 2005, 2006; Slager et al., 2003]. Nonetheless, the prevalence of some phenotypes, for example cardiovascular malformations and hearing loss, differ between SMS deletion patients and those with RAI1 point mutations [Edelman et al., 2007; Girirajan et al., 2006], suggesting that other loci within 17p11.2 and/or genomic rearrangement itself [Ricard et al., 2010] may contribute to or modify the SMS phenotype.

Subjective sleep disturbances are well documented in SMS [Smith et al., 1998]. Additionally, objective sleep disturbances, including multiple awakenings, decreased or increased percentage of REM sleep, and decreased total sleep time have been detected by polysomnography [Greenberg et al., 1996; Potocki et al., 2000] and actigraphy [De Leersnyder et al., 2001a; Gropman et al., 2006]. Inverted circadian rhythmicity of melatonin in SMS deletion patients has also been reported [De Leersnyder et al., 2001a; Potocki et al., 2000].

Persons with RAI1 mutations also have a high rate of subjective sleep disturbance (14 of 14 subjects described in [Bi et al., 2004, 2006; Girirajan et al., 2005, 2006; Slager et al., 2003]), statistically equivalent to that of individuals with the common (3.7 Mb) SMS deletion (39/39 in [Potocki et al., 2003]). It is unknown if RAI1 mutations lead to an alteration of melatonin rhythmicity. Given melatonin’s role in sleep-wake patterning [Brzezinski 1997], we hypothesized that haploinsufficiency of RAI1 alone may lead to inversion of melatonin rhythmicity, potentially explaining disordered sleep in individuals with this genotype. In this report we document an inverted circadian rhythm of melatonin in a child and an adult who each harbor a RAI1 mutation.

Individuals with clinical suspicion for SMS were enrolled in a multidisciplinary clinical study, approved by the Institutional Review Board of Baylor College of Medicine, at Texas Children’s Hospital in Houston, USA; informed consent was provided in each case. This comprehensive clinical protocol included physical examination, polysomnography, and urine collection to measure 6-sulfatoxymelatonin (6-hydroxymelatonin sulfate, aMT6s), the major excreted metabolite of melatonin [Reiter, 1991; Potocki et al., 2000]. Molecular characterization and some phenotypic details of these subjects have been described previously [Bi et al., 2004, 2006; Chen et al., 1997; Liburd et al., 2001; Potocki et al., 2000, 2003; Slager et al., 2003]. Briefly, individuals with clinical suspicion for SMS were screened by FISH for deletions on chromosome 17p11.2. Subjects who did not have a del(17)(p11.2p11.2) by FISH underwent sequencing of the RAI1 gene, which identified a heterozygous nonsense mutation in subject 1106, an 11-year-old girl [Bi et al., 2004], and a heterozygous frameshift mutation in subject 526 (listed as SMS156 in Slager et al., [2003]), a 27-year-old woman. Both of these subjects experienced subjective sleep disturbances by parental report. Subject 1106 took melatonin (6 mg at bedtime), which was discontinued 3 weeks prior to the study. Subject 526 took no medications.

Subjects were evaluated in the sleep laboratory at Texas Children’s Hospital [Potocki et al., 2000]. Briefly, the subjects underwent continuous 21-channel polysomnographic monitoring as well as in-person and video behavioral monitoring for one night. Sleep staging was determined using standard criteria. A multiple sleep latency test (MSLT) was performed the following day. The subjects’ spontaneously voided urine was collected at several time points and analyzed for aMT6s concentration [Potocki et al., 2000]. Urine from a healthy girl and a healthy man, analyzed similarly, served as controls [Reiter et al., 1996].

Both subjects with RAI1 mutations demonstrated sleep disturbances by polysomnography similar to those of subjects with the SMS common deletion (Table I). MSLT findings were normal (data not shown). The rhythm of urinary aMT6s concentration, a surrogate for serum melatonin concentration [Reiter, 1991], was found to be altered in both subjects with RAI1 mutations (Fig. 1). A normal circadian rhythm of urinary aMT6s concentration peaks in the early morning (Fig. 1a), resulting from a nocturnal maximum of melatonin production. In contrast, the urinary aMT6s concentration in both child (1106) and adult (526) with RAI1 point mutations (Fig. 1b) displayed an inverted pattern with a daytime maximum, similar to age-matched subjects with the SMS common deletion (1206 and 1123, respectively [Potocki et al., 2000]) shown for comparison (Fig. 1c). As subjects 1106 and 1206 and the younger control individual are children, their overall levels of aMT6s are expectedly higher than those of the adult control individual or of subjects 526 and 1123 [Waldhauser et al., 1984].

Subjective sleep disturbance is common in SMS, corroborated by objective abnormalities on polysomnography [Greenberg et al., 1991, 1996]. SMS deletion patients have been reported to have an inverted circadian rhythm of melatonin [De Leersnyder et al., 2001a; Potocki et al., 2000]. Our results indicate that SMS patients with mutations in RAI1, a gene mapping within the SMS critical region, have similarly altered melatonin rhythmicity.

It is unknown whether RAI1 is the only gene within the SMS critical region for which reduced dosage leads to altered melatonin rhythmicity. For example, RASD1, also located within the SMS critical region, is a modulator of the responsiveness of the core (suprachiasmatic nucleus) circadian clock to photic and nonphotic inputs [Cheng et al., 2004]. It is also unknown 1) whether alteration of melatonin rhythmicity is solely responsible for the sleep disturbances in SMS, 2) whether this alteration accompanies or causes similar disturbance of the core circadian clock, peripheral clocks, or other clock outputs, and 3) most basically, whether it constitutes a true inversion of rhythmicity or simply a phase advance or delay of approximately half of a day.