Lead Exposure Biosensors from Epigenome-Wide Blood DNA-Methylation in Adults

摘要

Background: Lead is a ubiquitous toxicant with a numerous chronic health effects. Between 2002-2011 more than 11,000 U.S. adults suffered acute lead-poisoning (>5 μg/dL). Acute symptoms are often misinterpreted or unrecognized and for this reason blood lead levels are likely under tested. Patella bone lead-a reliable measure of chronic cumulative exposure- is a highly specialized measure, requires exposure to X-radiation, and is available only in a small number of centers. Objective: To develop novel and safe biosensors reflecting individual current (blood) and cumulative (patella) lead exposure using whole blood DNA methylation profiles. Methods: We developed and tested two biosensors using data from epigenome-wide whole blood DNA methylation levels in men of the Normative Aging Study (NAS) cohort via the Illumina 450K bead chip. We selected methylation sites most responsive to lead exposure via robust regressions-adjusted for socio-demographic, lifestyle information and white cell proportions-and constructed the final biosensors using elastic nets. We used 10-fold cross-validation to test the biosensors. Results: Most NAS participants were exposed to relatively low levels of lead (median blood lead (Interquartile-range IQR) = 3 (3) μg/dL, median patella bone lead (IQR) 25.2 (20.2) μg/g). The elastic nets selected 1177 and 599 sites within the epigenome with non-null coefficients to estimate the blood biosensor and bone biosensor, respectively. Among the selected sites of both biosensors, only eight sites overlapped. The biosensors estimated lead levels that were highly correlated with actual values of blood (R2=0.92) and bone lead (R2= 0.96). Discrepancies were found mostly at extreme lead levels (in blood>7.5μg/dL, in bone>50ug/g). Conclusions: Our epigenetic lead biosensors are novel tools of measuring individual lead levels and may help identify lead-poisoning and its health impacts when available.