Background/objectives: There is increasing evidence of a relationship between blood DNA methylation and body mass index (BMI). We aimed to assess associations of BMI with individual methylation measures (CpGs) through a cross-sectional genome-wide DNA methylation association study and a longitudinal analysis of repeated measurements over time.
Subjects/methods: Using the Illumina Infinium HumanMethylation450 BeadChip, DNA methylation measures were determined in baseline peripheral blood samples from 5361 adults recruited to the Melbourne Collaborative Cohort Study (MCCS) and selected for nested case-control studies, 2586 because they were subsequently diagnosed with cancer (cases) and 2775 as controls. For a subset of 1088 controls, these measures were repeated using blood samples collected at wave 2 follow-up, a median of 11 years later; weight was measured at both time points. Associations between BMI and blood DNA methylation were assessed using linear mixed-effects regression models adjusted for batch effects and potential confounders. These were applied to cases and controls separately, with results combined through fixed-effects meta-analysis.
Results: Cross-sectional analysis identified 310 CpGs associated with BMI with P<1.0 × 10-7, 225 of which had not been reported previously. Of these 225 novel associations, 172 were replicated (P<0.05) using the Atherosclerosis Risk in Communities (ARIC) study. We also replicated using MCCS data (P<0.05) 335 of 392 associations previously reported with P<1.0 × 10-7, including 60 that had not been replicated before. Associations between change in BMI and change in methylation were observed for 34 of the 310 strongest signals in our cross-sectional analysis, including 7 that had not been replicated using the ARIC study.
Conclusions: Together, these findings suggest that BMI is associated with blood DNA methylation at a large number of CpGs across the genome, several of which are located in or near genes involved in ATP-binding cassette transportation, tumour necrosis factor signalling, insulin resistance and lipid metabolism.