Clues to Maternal Smoking Risks Uncovered

Blog Post created by Thomas3.20.2010 on Aug 14, 2012

Umbilical-cord blood from women who smoked during pregnancy had DNA modification in genes associated with detoxification of tobacco smoke, investigators reported.

An investigation of epigenome-wide DNA methylation in cord blood yielded statistically significant associations (P<0.05) with maternal cotinine levels (a biomarker of smoking) for 26 sites mapped to 10 genes. Analysis of a small replication cohort confirmed the findings for three genes, two of which have been implicated in removal of toxins from tobacco smoke.

"These results suggest that epigenetic mechanisms reflected by DNA methylation may underlie some of the well-documented impacts of maternal smoking on offspring," Stephanie J. London, MD, DrPH, of the National Institute of Environmental Health Sciences in Research Triangle Park, N.C., and co-authors reported online in Environmental Health Perspectives.


"Our identification of differential methylation in genes known to be involved in the response to tobacco-related compounds, in addition to a novel gene, demonstrates the value of using this approach to elucidate the epigenetic effects of in utero exposures."

Maternal smoking during pregnancy can cause multiple adverse effects on children. Understanding of the underlying mechanisms has remained incomplete.

Some evidence has suggested that epigenetic mechanisms, such as DNA methylation, might play a role in the adverse effects of maternal smoking. Supporting evidence has come primarily from preclinical models. London and colleagues sought to expand on the modest accumulation of data in humans.

The study included 1,062 participants in a Norwegian study of maternal folate status and asthma at age 3. The participants were selected on the basis of available cord blood for evaluation and complete data for maternal cotinine status.

For replication analysis of cord blood, investigators recruited 36 participants from an epigenetics study conducted at Duke University. Half of the mothers reported smoking during pregnancy.

Using commercial technology, investigators examined maternal plasma cotinine in relation to DNA methylation at 473,844 cytosine-guanine dinucleotide sites (CpGs) in the 1,062 cord-blood specimens from the Norwegian study the 36 specimens from the epigenetics study.


Analysis of the Norwegian cord blood samples revealed maternal plasma cotinine levels consistent with active smoking in 12.8% of the participants; 11.7% of the mothers reported smoking during pregnancy.

DNA methylation analysis produced statistically significant (P<1.067 x 10-7) associations with maternal plasma cotinine for 26 CpGs mapped to 10 genes. The 26 CpGs included:

  • 8 in the coding region of growth factor independent 1 transcription repressor (GFI1) on chromosome 1
  • 4 in the coding region of aryl-hydrocarbon receptor repressor (AHRR) on chromosome 5
  • 4 in a region upstream of cytochrome P450 isoform CYP1A1 on chromosome 15
  • 4 in the coding region of myosin 1G (MYO1G)


Relationships between DNA methylation and maternal plasma cotinine levels were inverse in some cases and direct in others.

CYP1A1 and AHRR previously have been implicated in tobacco smoking detoxification, but GFI1had not.


Comparison of self-reported smoking and DNA methylation showed that smoking was associated with reduced methylation in four CpGs in AHRR and increased methylation in CYP1A1 and GFI1CpGs.

The replication analysis revealed associations between maternal cotinine and DNA methylation consistent with those observed in the Norwegian discovery set, and that achieved statistical significance for AHRRCYP1A1, and GFI1.

The association with the greatest statistical significance in both the discovery and replication sets was AHRR cg05575921, wherein smokers had decreased DNA methylation compared with nonsmokers.

Despite the much smaller size of the replication set, "the replication P-values were systematically smaller than would be expected by chance (P<0.00011)," the authors wrote. "This suggests that it is exceedingly unlikely that the replication findings are false-positives and confirms the high degree of replication that we observed."