Functional Annotation and Cross-study Comparison of DNA Methylation Regions uniquely associated with 24-hour Blood Pressure Phenotypes in African Americans.

Epigenetics may reflect the interactions between DNA and environment and lifestyle and DNA methylation marks have been implicated in the development and progression of hypertension. In an African American (AA) cohort of hypertensives and normotensives (n = 281), we have recently discovered up to 72 methylation regions (MRs) in whole blood DNA that were significantly associated with 24-h BP phenotypes (False Discovery Rate (FDR) < 0.05) via reduced representation bisulfite sequencing (RRBS). One-time clinic BP measurements were not associated with MRs. Upon adjustments for covariates age, sex, and body mass index, 13 MRs remained associated with 24-h BP phenotypes (FDR < 0.05), each explaining 6.5% - 9.4% of BP variance. Analysis of the chr19 MR in an independent cohort (n=117) using Bisulfite ULtrapLEx Targeted Sequencing (BULLET-Seq), a new method that we developed, validated its association with 24-h avg BP phenotypes (FDR < 0.05). In the current study, we functionally annotated the MRs that we identified as uniquely associated with 24-h BP phenotypes and examined CpG sites previously reported to be associated with BP. Two MRs were near chr1 linkage disequilibrium (LD) regions within genes, PRDM16 and CASZ1, containing BP-associated sentinel single nucleotide polymorphisms (SNPs). Although no MRs were found to directly overlap with reported BP-relevant expression or methylation quantitative trait loci (eQTLs or meQTLs), many overlapped with enhancer regions, CTCF-binding sites, or genes related to hypertension and stroke. To demonstrate consistency with prior studies, we found associations between reported clinic BP-associated methylation array chip CpGs and the 24-h BP phenotypes in our study. Of the 16 BP-associated CpGs from two reports that were detected in both the methylation chip and our RRBS data, several CpGs were significantly associated with 24-h and clinic BP phenotypes in our cohort (p < 0.05) but not after adjustment for both covariates and FDR. A chr7 CpG (cg06340364) showed significant negative associations with all studied BP phenotypes except nighttime and clinic DBPs (p < 0.05, unadjusted). These investigations highlight the value of regional DNA methylation and the ability to capture the overall effect of environmental and genetic factors that influence BPs over a 24-h period which have functional epigenomic relevance.