Sleep disturbance is prevalent in aging populations, yet its association with ambient PM2.5 remains insufficiently characterized due to the lack of reliable molecular biomarkers. We leveraged 2350 middle-aged and older adults from the Guangxi Eco-Environmental Health and Aging Study to examine the association between PM2.5 exposure and sleep quality measured by the Pittsburgh Sleep Quality Index (PSQI) and to identify relevant DNA methylation signatures of circadian rhythm genes. Higher PM2.5 was significantly associated with worse sleep outcomes, with the strongest associations for the 2-month exposure window. Particularly, per 10 μg/m3 increase in the 2-month average PM2.5 was associated with a 1.33-point increase in PSQI score (95% CI: 0.95, 1.71), a 0.79-hour reduction in sleep duration (95% CI: -0.99, -0.59), and higher odds of poor sleep quality (odds ratio [OR] =2.17, 95% CI: 1.69, 2.80) and abnormal sleep duration (OR =1.59, 95% CI: 1.24, 2.03). A two-stage analysis of DNA methylation signatures identified 47 sleep-related CpG sites, of which 11 were selected using LASSO penalization to construct the sleep-quality-related methylation risk score (MRS). Mediation analyses further identified five CpGs annotated to NPAS2, PRKAG2, RORA, and CSNK2A2 that partially mediated the association between PM2.5 exposure and sleep quality (mediation proportions: 9.08%-15.98%), whereas the sleep MRS accounted for a higher mediation proportion than individual CpGs (25.03%). Our findings underscore ambient PM2.5 exposure as a relevant environmental factor linked to impaired sleep quality and highlight that circadian epigenetic signatures could serve as molecular markers of vulnerability to pollution-related sleep disturbances.
