• Divison of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, USA. clara.bodelon@nih.gov.
• International Agency for Research on Cancer (IARC), Lyon, France.
• Cancer Epidemiology and Intelligence Division, Cancer Council Victoria, Melbourne, Victoria, Australia.
• Division of Cancer, Imperial College London, London, UK.
• Divison of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, USA.
• Genetic Epidemiology Laboratory, Department of Pathology, The University of Melbourne, Parkville, Australia.
• IIGM (Italian Institute for Genomic Medicine), Turin, Italy.
• Department of Clinical and Experimental Medicine, University of Pisa, Pisa, Italy.
• MRC-PHE Center for Environment and Health, School of Public Health, Imperial College, London, UK.
• CESP (U1018 INSERM, Équipe Générations et Santé), Facultés de médecine Université Paris-Sud, UVSQ, Université Paris-Saclay, Villejuif, France.

BACKGROUND: Environmental and genetic factors play an important role in the etiology of breast cancer. Several small blood-based DNA methylation studies have reported risk associations with methylation at individual CpGs and average methylation levels; however, these findings require validation in larger prospective cohort studies. To investigate the role of blood DNA methylation on breast cancer risk, we conducted a meta-analysis of four prospective cohort studies, including a total of 1663 incident cases and 1885 controls, the largest study of blood DNA methylation and breast cancer risk to date.

METHODS: We assessed associations with methylation at 365,145 CpGs present in the HumanMethylation450 (HM450K) Beadchip, after excluding CpGs that did not pass quality controls in all studies. Each of the four cohorts estimated odds ratios (ORs) and 95% confidence intervals (CI) for the association between each individual CpG and breast cancer risk. In addition, each study assessed the association between average methylation measures and breast cancer risk, adjusted and unadjusted for cell-type composition. Study-specific ORs were combined using fixed-effect meta-analysis with inverse variance weights. Stratified analyses were conducted by age at diagnosis (< 50, ≥ 50), estrogen receptor (ER) status (+/-), and time since blood collection (< 5, 5-10, > 10 years). The false discovery rate (q value) was used to account for multiple testing.

RESULTS: The average age at blood draw ranged from 52.2 to 62.2 years across the four cohorts. Median follow-up time ranged from 6.6 to 8.4 years. The methylation measured at individual CpGs was not associated with breast cancer risk (q value > 0.59). In addition, higher average methylation level was not associated with risk of breast cancer (OR = 0.94, 95% CI = 0.85, 1.05; P = 0.26; P for study heterogeneity = 0.86). We found no evidence of modification of this association by age at diagnosis (P = 0.17), ER status (P = 0.88), time since blood collection (P = 0.98), or CpG location (P = 0.98).

CONCLUSIONS: Our data indicate that DNA methylation measured in the blood prior to breast cancer diagnosis in predominantly postmenopausal women is unlikely to be associated with substantial breast cancer risk on the HM450K array. Larger studies or with greater methylation coverage are needed to determine if associations exist between blood DNA methylation and breast cancer risk.