Abstract Introduction The onset of puberty is associated with a shift in the circadian timing of sleep leading to delayed sleep initiation [i.e., later sleep onset time (SOT)] driven by later bedtimes and/or longer sleep onset latency (SOL). Subjective sleep initiation, as per self-reports, and objective sleep initiation, as per actigraphy (ACT) or polysomnography (PSG), assess equally relevant but different domains of the same physiological process. Several GWAS have identified genes associated with sleep and circadian phenotypes; however, little is known regarding the epigenetic basis and significance of delayed sleep initiation in adolescence, a critical developmental period. Methods We analyzed data from 263 adolescents from the Penn State Child Cohort follow-up study who had complete subjective (self-reported questionnaires), at-home 7-night ACT, and in-lab 9-hour PSG data for bedtime, SOL and SOT. Epigenome-wide single-nucleotide resolution of DNA methylation in cytosine-phosphate-guanine (CpG) sites and surrounding regions were obtained from peripheral leukocytes. Linear regression assessed the association between bedtime, SOL and SOT with site-specific methylation levels, adjusting for sex, age, race/ethnicity, body mass index, and psychotropic medication use. P-values were adjusted using the Benjamini & Hochberg method to correct for false discovery rate and, thus, q-values are reported. Results Exome-wide analysis showed differential methylation in 1450 unique genes across the 9 sleep measurements, while GWAS-informed analysis yielded 57 genes. Gene hits were identified for bedtime (PSG), SOL (subjective, ACT and PSG) and SOT (subjective and PSG): 14 genes were associated with both subjective and PSG-measured SOL, 34 with both ACT- and PSG-measured SOL, and one (TBC1D22A) with subjective, ACT- and PSG-measured SOL. One gene (ABTB2) was associated with subjective and PSG-measured SOT. Conclusion Objective and subjective sleep initiation in adolescents is associated with altered DNA methylation in genes previously identified in adult GWAS of sleep and circadian phenotypes. Our data provides evidence for a potential epigenetic link between habitual (subjective and ACT) SOL and in-lab SOT and DNAm in genes involved in circadian regulation (i.e., RASD1, RAI1), metabolism (i.e., FADS1, WNK1, SLC5A6), and neuropsychiatric disorders (i.e., PRR7, SDK1, FAM172A). Support (If Any) National Institutes of Health (R01HL136587, R01MH118308, UL1TR000127)
Read full abstract