Proteomic Signatures Underlying Sleep, Circadian Activity Patterns, and Major Chronic Diseases

Abstract

Rest-activity rhythm and sleep may serve as potential intervention targets for a variety of diseases. However, the underlying biological mechanisms of rest-activity rhythm, sleep, and their proteomic associations with multiple diseases remain largely unexplored. Here, using data from approximately 10 000 participants in the UK Biobank with accelerometer-derived measures and proteomics profiles, we characterized the proteomic signatures of rest-activity and sleep and explored their associations with health outcomes. We found that the proteins associated with rest-activity and sleep were mainly enriched in inflammation, immune response and complement system. Most rest-activity and sleep measures, along with their associated proteomic signatures, were significantly associated with incident diabetes, cardiovascular disease, chronic kidney disease, respiratory diseases, and extended life expectancy. Several proteins, such as ADM and CA14, were observed to mediate multiple associations across distinct rest-activity and sleep measures. The impact of rest-activity and sleep measures on chronic diseases and mortality may be mediated through diverse biological pathways involving multiple proteins. These findings reveal potential mechanisms underlying these complex relationships and provide novel insights for the development of targeted intervention strategies.

Keywords: all‐cause mortality; chronic diseases; proteomics; rest‐activity rhythm; sleep.

Figure 1

The results of KEGG for proteins associated with rest‐activity and sleep measures. KEGG enrichment analyses were conducted for proteins that significantly associated with RA (1134 proteins), M10 (1301 proteins), L5 (92 proteins), IS (94 proteins), IV (126 proteins), Sleep efficiency (502 proteins), and Irregular sleep duration (259 proteins) in multivariable regression models. The bar plot shows the names of significant KEGG pathways with FDR correction less than 0.05. If more than 15 pathways meet this criterion, only the top 15 are displayed.

Figure 2

The results of GO for proteins associated with rest‐activity and sleep measures. GO enrichment analyses were conducted for proteins that significantly associated with RA (1134 proteins), M10 (1301 proteins), L5 (92 proteins), IS (94 proteins), IV (126 proteins), Sleep efficiency (502 proteins), and Irregular sleep duration (259 proteins) in multivariable regression models. The dot plot illustrates the top 10 processes for each GO category: Biological Process (BP), Cellular Component (CC), and Molecular Function (MF). The size of the dots represents the count of genes involved in each process, while the color indicates the FDR‐adjusted p‐value, with redder colors signifying higher significance.

Figure 3

Associations of rest‐activity and sleep measures and their corresponding proteomic signatures with health outcomes. The HRs comparing the 90th with the 10th percentile were represented by the squares and the 95% confidence intervals (95% CI) were represented by the error bars. The dashed lines represented HR = 1.0. The multivariable (MV) model was adjusted for age at start of accelerometry measurement (continuous, years), season at the time that accelerometry measurements started (Spring, Summer, Autumn, Winter), sex (male, female), self‐reported ethnicity (White, nonwhite), assessment center (England, Scotland, Wales), Townsend deprivation index (continuous), multivitamin use (yes, no), fasting time (continuous, hours), BMI (< 25, 25–30, 30 kg/m²), education attainment (College or University degree, Other, None of the above), smoking status (never, previous, current), drinking (daily or almost daily, three or four times a week, once or twice a week, one to three times a month, special occasions only, never), diet (continuous, 0–5 points), and history of shift work (yes, no). For the mutual adjustment, each rest‐activity and sleep measures and its corresponding proteomic signature were included simultaneously in the MV model to assess association independence. *The two‐sided P values corrected using the FDR method were less than 0.05.

Figure 4

Estimated life expectancy increase for the lowest and highest 10th percentiles of rest‐activity and sleep measures. For rest‐activity rhythm measures such as RA, M10, and IS, participants in the lowest 10% were used as the reference group, while for other metrics, the highest 10th percentile served as the reference. Analyses were adjusted for age at start of accelerometry measurement (continuous, years), season at the time that accelerometry measurements started (Spring, Summer, Autumn, Winter), sex (male, female), self‐reported ethnicity (White, nonwhite), assessment center (England, Scotland, Wales), Townsend deprivation index (continuous), multivitamin use (yes, no), fasting time (continuous, hours), BMI (< 25, 25–30, 30 kg/m²), education attainment (College or University degree, Other, None of the above), smoking status (never, previous, current), drinking (daily or almost daily, three or four times a week, once or twice a week, one to three times a month, special occasions only, never), diet (continuous, 0–5 points), and history of shift work (yes, no).

Figure 5

The estimated proportion of mediation for the proteins on the associations of rest‐activity and sleep measures with health outcomes. Only the top 10 ranked proteins in different mediations are shown. Analyses were adjusted for age at start of accelerometry measurement (continuous, years), season at the time that accelerometry measurements started (Spring, Summer, Autumn, Winter), sex (male, female), self‐reported ethnicity (White, nonwhite), assessment center (England, Scotland, Wales), Townsend deprivation index (continuous), multivitamin use (yes, no), fasting time (continuous, hours), BMI (< 25, 25–30, 30 kg/m²), education attainment (College or University degree, Other, None of the above), smoking status (never, previous, current), drinking (daily or almost daily, three or four times a week, once or twice a week, one to three times a month, special occasions only, never), diet (continuous, 0–5 points), and history of shift work (yes, no). Proteins included in the analysis were adjusted using FDR.