Early-life inorganic arsenic exposure influences not only child health and development but also health in later life. The adverse effects of arsenic may be mediated by epigenetic mechanisms, as there are indications that arsenic causes altered DNA methylation of cancer-related genes. The objective was to assess effects of arsenic on genome-wide DNA methylation in newborns. We studied 127 mothers and cord blood of their infants. Arsenic exposure in early and late pregnancy was assessed by concentrations of arsenic metabolites in maternal urine, measured by high performance liquid chromatography-inductively coupled plasma mass spectrometry. Genome-wide 5-methylcytosine methylation in mononuclear cells from cord blood was analyzed by Infinium HumanMethylation450K BeadChip. Urinary arsenic in early gestation was associated with cord blood DNA methylation (Kolmogorov–Smirnov test, P-value<10–15), with more pronounced effects in boys than in girls. In boys, 372 (74%) of the 500 top CpG sites showed lower methylation with increasing arsenic exposure (rS-values>−0.62), but in girls only 207 (41%) showed inverse correlation (rS-values>−0.54). Three CpG sites in boys (cg15255455, cg13659051 and cg17646418), but none in girls, were significantly correlated with arsenic after adjustment for multiple comparisons. The associations between arsenic and DNA methylation were robust in multivariable-adjusted linear regression models. Much weaker associations were observed with arsenic exposure in late compared with early gestation. Pathway analysis showed overrepresentation of affected cancer-related genes in boys, but not in girls. In conclusion, early prenatal arsenic exposure appears to decrease DNA methylation in boys. Associations between early exposure and DNA methylation might reflect interference with de novo DNA methylation.

Maternal high-fat diet consumption and obesity have been shown to program long-term obesity and lead to impaired glucose tolerance in offspring. Many rodent studies, however, use non-purified, cereal-based diets as the control for purified high-fat diets. In this study, primiparous ICR mice were fed purified control diet (10–11 kcal% from fat of lard or butter origin) and lard (45 or 60 kcal% fat) or butter (32 or 60 kcal% fat)-based high-fat diets for 4 weeks before mating, throughout pregnancy, and for 2 weeks of nursing. Before mating, female mice fed the 32 and 60% butter-based high-fat diets exhibited impaired glucose tolerance but those females fed the lard-based diets showed normal glucose disposal following a glucose challenge. High-fat diet consumption by female mice of all groups decreased lean to fat mass ratios during the 4th week of diet treatment compared with those mice consuming the 10–11% fat diets. All females were bred to male mice and pregnancy and offspring outcomes were monitored. The body weight of pups born to 45% lard-fed dams was significantly increased before weaning, but only female offspring born to 32% butter-fed dams exhibited long-term body weight increases. Offspring glucose tolerance and body composition were measured for at least 1 year. Minimal, if any, differences were observed in the offspring parameters. These results suggest that many variables should be considered when designing future high-fat diet feeding and maternal obesity studies in mice.

Lifecourse trajectories of clinical or anthropological attributes are useful for identifying how our early-life experiences influence later-life morbidity and mortality. Researchers often use growth mixture models (GMMs) to estimate such phenomena. It is common to place constrains on the random part of the GMM to improve parsimony or to aid convergence, but this can lead to an autoregressive structure that distorts the nature of the mixtures and subsequent model interpretation. This is especially true if changes in the outcome within individuals are gradual compared with the magnitude of differences between individuals. This is not widely appreciated, nor is its impact well understood. Using repeat measures of body mass index (BMI) for 1528 US adolescents, we estimated GMMs that required variance–covariance constraints to attain convergence. We contrasted constrained models with and without an autocorrelation structure to assess the impact this had on the ideal number of latent classes, their size and composition. We also contrasted model options using simulations. When the GMM variance–covariance structure was constrained, a within-class autocorrelation structure emerged. When not modelled explicitly, this led to poorer model fit and models that differed substantially in the ideal number of latent classes, as well as class size and composition. Failure to carefully consider the random structure of data within a GMM framework may lead to erroneous model inferences, especially for outcomes with greater within-person than between-person homogeneity, such as BMI. It is crucial to reflect on the underlying data generation processes when building such models.