Background:
DNA methylation is an epigenetic mark which regulates gene expression and can be modified by environmental factors, possibly including toxic metals. Although their function is unclear, DNA contains large numbers of interdispersed repetitive elements (ALU and LINE-1) that are normally heavily methylated. Global methylation can be measured within these elements and changes in global methylation have been associated with diseases such as cancer and cardiovascular disease. Changes in global methylation may be representative of the effects of environmental factors on epigenetic marks and may explain observed phenomena such as programming and the latencies between exposure and disease onset.

Methods:
We measured global methylation on a subset of 783 Normative Aging Study subjects (all male) who had archived blood DNA samples. We measured patella and tibia lead levels by X-Ray fluorescence. Blood lead levels were measured by atomic absorption spectrometry. We determined global DNA methylation averages within CpG repeats of LINE-1 and ALU elements using PCR-Pyrosequencing on bisulfite-treated DNA. We constructed a series of multiple linear regression models using methylation with ALU or LINE-1 as the dependent variable and bone lead (tibia or patella in separate models) as the primary exposure markers. All models were adjusted for age at DNA collection, pack-years of smoking, education and blood Pb at time of DNA collection as covariates.

Results:
Overall mean (SD) values for global methylation as measured by ALU and LINE-1 were as follows: ALU-26.3% STD(1.1%); LINE-1- 76.8% STD(1.8%), respectively. Mean (SD) age was 72.9 (7.6) years. 32% of subjects had more than a high school education. Mean (SD) pack-years of smoking was 23.5 STD(26). In both the bivariate and full model patella lead levels were inversely associated with ALU (standardized Beta = -0.1; P = 0.03) but not LINE-1; standardized Beta = -0.04; P = 0.4). Results were unchanged if blood Pb was dropped from the model. Among the covariates only age was associated with global methylation (standardized Beta = -0.1; P = 0.01) Tibia lead and blood lead did not predict global methylation for either ALU or LINE-1.

Conclusion:
Patella lead levels inversely predict reduced global DNA methylation within ALU elements. We believe this is the first report in humans of an association between chronic lead exposure and DNA methylation. This finding may have implications for lead's mechanism of action in producing adverse health effects. Changes in DNA methylation could be a mechanism for long term programming effects as well. Further study is needed to confirm these findings and to determine whether DNA methylation marks (global and gene specific) are useful biomarkers of lead toxicity in epidemiologic studies.