Early life, the epigenome and human health

The paper by Schlinzig et al. (1) shows that global DNA methylation levels in cord white blood cells are higher in newborns delivered by Caesarean section (CS) than in those delivered by normal vaginal delivery. Interestingly, the levels of methylation in the CS-delivered babies decrease 3– 5 days after birth and become similar to that of the control group. This is the first set of data that demonstrates in humans the dynamic nature of the DNA methylation pattern in fully differentiated cells after birth. These data have important implications on our understanding of the impact of early life exposures on the DNA methylation patterns, gene expression programming and on how early life events could shape lifelong health trajectories in humans. DNA methylation in humans and other vertebrates occurs mostly in the dinucleotide sequence CG. A fraction of CG in the genome is methylated. The distribution of methylated and unmethylated CGs creates a pattern that is tissue specific (2). The classic accepted model has been that DNA methylation patterns change in a highly organized way during early development and differentiation, but that the pattern of methylation remains fixed thereafter. This model was founded on the belief that DNA methylation patterns are irreversible. It was believed that the only DNA methylation reaction that exists in somatic cells is maintenance DNA methylation, a semiconservative replication of DNA methylation by a DNMT1 a DNA methyltransferase that shows high preference to hemimethylated DNA and preserves the DNA methylation pattern during cell division (3). If indeed this model was true, then DNA methylation could not serve as a biological signal in a physiological time scale. This basic dogma has shaped the field of DNA methylation and as a consequence the biological questions that were asked for several decades. Almost a decade ago, Ramchandani et al. proposed that DNA methylation was a reversible biological signal (4). This raised the possibility that DNA methylation pattern plasticity might play a role in programming the genome in adaptive responses to changing environments early in life and perhaps throughout life. Several recent studies including the study by Schlinzig et al., in this issue, point to this possibility. DNA methylation patterns once formed in response to a significant exposure are retained as a stable mark in the genome. If indeed these alterations in DNA methylation happen to be in regulatory regions of genes, they would provide long-term programming of gene expression that could affect the phenotype later in life. DNA methylation changes would thus serve as a mark of early life exposure long after the original triggers are gone (5). Although different types of adult onset disease occur in humans, it is plausible that they are triggered by common basic mechanisms. This ‘early life environment’ hypothesis might provide a plausible explanation for the rapid emergence of certain late onset diseases such as obesity, type II diabetes and asthma in the past century (6). It is hard to believe that such an increase in the rate of disease was caused by a genetic drift. It is possible that changes in medical procedures, medications, nutrition and the social environment in early life contributed to the emergence of these diseases. The critical question is, what is a plausible mechanism that can mediate between ‘early life’ exposures and the phenotypes unravelled later in life? If genetic mechanisms are excluded, what about epigenetic mechanisms? Several studies pointed to these possibilities. A landmark study by the Jirtle group has shown how maternal diet could affect coat colour and other health indicators in the offspring and these were molecularly linked to DNA methylation changes (5). Other studies have shown that not only can maternal chemical-dietary environment have a long-term impact on the epigenome of the offspring, but the maternal behaviour can Acta Pædiatrica ISSN 0803–5253