Analysis of C-reactive protein omics-measures associates methylation risk score with sleep health and related health outcomes

Abstract

Introduction: DNA methylation (DNAm) predictors of high sensitivity C-reactive protein (CRP) offer a stable and accurate means of assessing chronic inflammation, bypassing the CRP protein fluctuations secondary to acute illness. Poor sleep health is associated with elevated inflammation (including elevated blood CRP levels) which may explain associations of sleep insufficiency with metabolic, cardiovascular and neurological diseases. Our study aims to characterize the relationships among sleep health phenotypes and CRP markers -blood, genetic, and epigenetic indicators-within the Hispanic Community Health Study/Study of Latinos (HCHS/SOL).

Methods: In HCHS/SOL, methylation risk scores (MRS)-CRP and polygenetic risk score (PRS)-CRP were constructed separately as weighted sums of methylation beta values or allele counts, respectively, for each individual. Sleep health phenotypes were measured using self-reported questionnaires and objective measurements. Survey-weighted linear regression established the association between the multiple sleep phenotypes (obstructive sleep apnea (OSA), sleep duration, insomnia and excessive sleepiness symptom), cognitive assessments, diabetes and hypertension with CRP markers while adjusting for age, sex, BMI, study center, and the first five principal components of genetic ancestry in HCHS/SOL.

Results: We included 2221 HCHS/SOL participants (age range 37-76 yrs, 65.7% female) in the analysis. Both the MRS-CRP (95% confidence interval (CI): 0.32-0.42, p = 3.3 × 10^-38) and the PRS-CRP (95% CI: 0.15-0.25, p = 1 × 10^-14) were associated with blood CRP level. Moreover, MRS-CRP was associated with sleep health phenotypes (OSA, long sleep duration) and related conditions (diabetes and hypertension), while PRS-CRP markers were not associated with these traits. Circulating CRP level was associated with sleep duration and diabetes. Associations between OSA traits and metabolic comorbidities weakened after adjusting for MRS-CRP, most strongly for diabetes, and least for hypertension.

Conclusions: MRS-CRP is a promising estimate for systemic and chronic inflammation as reflected by circulating CRP levels, which either mediates or serves as a common cause of the association between sleep phenotypes and related comorbidities, especially in the presence of diabetes.

Figure 1.

Study design. Best-performing polygenic risk score (PRS) of C-reactive protein (CRP) was first selected in MESA as the one explaining the highest variance in circulating CRP. Then association of circulating CRP with Methylation risk score (MRS) and PRS constructed in SOL-INCA was confirmed before evaluating the association of these CRP measures with sleep and other health outcomes.

Figure 2.

Association of polygenic risk score (PRS) and methylation risk score (MRS) for C-reactive protein (CRP) with blood CRP level.

Figure 3.

Forest plot showing associations between C-reactive protein (CRP) measures and measured health outcomes. A) obstructive sleep apnea (OSA), sleep duration and cognitive traits; B) odds ratio for binary cardio-metabolic and other sleep outcomes. From left to right model coefficients or odds ratio in case of binary outcomes, 95% confidence interval (95% CI), p value (p.val) and FDR corrected q value (q.val). MRS: methylation risk score; PRS: polygenic risk score