Evaluation of a DNA methylation-based measure of chronic inflammation in two generations of adults in metropolitan Cebu, Philippines

Proxy measures of chronic inflammation derived from DNA methylation (DNAm) data have emerged as promising predictors of cardiometabolic disease in high income countries. The performance of these measures in lower and middle income settings experiencing higher levels of endemic infections as well as rising rates of chronic degenerative diseases is not known. A DNAm-based proxy for C-reactive protein (DNAm-CRP) was calculated from 1468 CpG sites on the Infinium MethylationEPIC v1.0 array applied to genomic DNA from leukocytes in young adults (N = 1665; 20-22 years) and older women (N = 1070; 35-68 years) in the Philippines. CRP was determined in plasma using a high sensitivity immunoturbidimetric assay. Pearson correlation and least squares regression were implemented to evaluate the strength of association between DNAm-CRP and plasma CRP, and to investigate patterns of association between DNAm-CRP and established predictors of chronic inflammation. For younger adults, the correlation between DNAm-CRP and log-transformed CRP was 0.41, and DNAm-CRP explained 17.2% of the variance in CRP. For older women, the correlation was 0.47, with 22.7% explained variance in CRP. For both cohorts larger waist circumference was associated with higher DNAm-CRP. Recent infectious symptoms and leukocyte composition were both significant predictors of DNAm-CRP. DNAm-CRP may be a useful tool for research on chronic inflammation across a range of epidemiological and ecological settings globally, but future applications should consider how recent infections and the distribution of leukocyte subsets may confound or mediate associations of interest.