Placental mitochondrial DNA and CYP1A1 gene methylation as molecular signatures for tobacco smoke exposure in pregnant women and the relevance for birth weight

Despite a limited number of (epi)genomic studies in placental tissue and cord blood, we are improving our understanding of the molecular pathways underlying the association between gestational tobacco smoke exposure and low birth weight. Combining gene expression and epigenome-wide methylation arrays Suter et al. [26] showed that the expression of 623 genes and the methylation of 1024 CpG dinucleotides were significantly altered in placentas of smokers. For 438 genes significant correlations were revealed between methylation and gene expression, and their potential functions or mechanisms were explored using an Ingenuity Pathway Analysis. The authors found that the gene list was enriched for genes involved in functional pathways such as mitochondrial dysfunction, oxidative phosphorylation and hypoxia. Indeed, mitochondria, the “powerhouses” of cells, provide cellular energy via oxidative phosphorylation and are very sensitive to exposures that induce oxidative stress. The double stranded circular mtDNA, of which multiple copies are present in mitochondria, is vulnerable to reactive oxygen species (ROS) because of an inefficient DNA repair capacity and close proximity to the electron transport chain [33]. The estimated mutation rate of mtDNA is 5-10 times higher compared to nuclear DNA [34]. We showed that placental mtDNA content and methylation levels were responsive to tobacco smoke exposure during pregnancy indicating that mtDNA is a sensitive marker of mitochondrial damage and dysfunction as proposed by Sahin et al. [10]. In addition to other studies reporting changes in placental mtDNA content in smoking mothers [7, 8] or mothers exposed to air pollution [29], we provide here the first epidemiological evidence of altered methylation levels at specific loci of the mitochondrial genome of placental tissue in response to tobacco smoke exposure during pregnancy. We suggest that pollution-induced epigenetic modifications of the mitochondrial genome may prime alterations in mtDNA content by regulating mitochondrial function and biogenesis [35]. Damaged or non-functioning mitochondria are specifically degraded through mitophagy and could result in a depletion of mtDNA [36], which moreover may lead to changes in methylation patterns of a number of nuclear genes [12]. The sensitivity analysis showed that mtDNA content and mtDNA methylation correlated with methylation of CYP1A1 in placental tissue, which could be indicative of a relationship between mitochondrial dysfunction and the epigenetic landscape of the nuclear genome [11]. Whether mitochondrial dysfunction affects gene expression and methylation patterns of other genes needs to be elucidated.

An expanding body of evidence suggests that the epigenome of placental tissue and cord blood is sensitive to environmental exposures [13]. Epigenome-wide methylation studies are used to examine the epigenetic status of the human genome at many different loci in a number of individuals and also to assess whether any of these CpG loci are associated with a trait or an environmental pollutant [37]. A 450 K epigenome-wide methylation study by Joubert et al. [23] demonstrated differentially methylated detoxifying genes (AHRR and CYP1A1) in cord blood of newborns exposed to tobacco smoke during pregnancy. This finding was confirmed in another population of infants by analyzing whole blood obtained by a heel prick [25]. Maternal smoking as assessed by both self-report and cotinine levels in plasma showed higher methylation levels at different CpGs of CYP1A1 in cord blood [23]. Conversely, in placental tissue of smoker mothers, Suter et al. [19] observed hypomethylated CpG dinucleotides proximal to a xenobiotic response element (XRE); whereas, those distal from such elements did not demonstrate differential methylation. The authors calculated the total percentage of methylation for a distinct region of the promoter (−1411 to −1295 bp from the transcription start site) without taking into account the separate CpGs, unlike we did in our study. We observed lower methylation levels at a specific CpG site that lies adjacent to a XRE site in placental tissue of mothers who smoked during pregnancy. It is important to note that this specific CpG site harbors a C/G single nucleotide polymorphism (SNP: rs3809585 with allele frequencies C: 1.717% and G: 98.283%). We are confident that this SNP did not affect DNA methylation since all pyrograms confirmed a G nucleotide in the analyzed sequence. Interestingly, the study of Joubert et al. [23] in cord blood, the study of Suter et al. [19] in placental tissue, and our study in placental tissue, examined approximately the same region of interest and CpGs, however with different detection methods (Fig. 4). With the bisulfite pyrosequencing approach, we confirmed hypomethylation at a specific CpG of the CYP1A1 gene in placental tissue which is in contrast with the findings in cord blood [23]. Although we lacked meaningful gene expression data of CYP1A1 in our study, Suter et al. [19] previously showed that lower methylation levels in a region covering the XRE site were correlated with increased expression of CYP1A1 in placental tissue. Moreover, other studies demonstrated increased CYP1A1 mRNA [38] and protein [39] expression in human placentas in response to tobacco smoke exposure. Constituents of tobacco smoke such as PAHs enter cells and are recognized by the aryl hydrocarbon receptor (Ahr) causing its translocation to the nucleus and the formation of a heterodimer with the Ahr nuclear translocator protein (ARNT). This complex binds to genes with a XRE within the promoter and initiates expression of detoxifying enzymes involved in phase I and II xenobiotic metabolism [40].

A limitation of our study is the chance of exposure misclassification. Information about maternal smoking during pregnancy was based on self-report and is not verifiable. A possibility to overcome this limitation is the determination of the cotinine concentration in plasma or urine of the mother. Nevertheless, previous studies demonstrated that this would not be superior to self-reported smoking habit in pregnant women [2]. We acknowledge the fact that we cannot fully exclude residual or unmeasured confounding by other factors that could be associated with both tobacco smoke exposure and placental molecular signatures. Although a causal relationship exists between prenatal tobacco smoke exposure and low birth weight or preterm birth, not all infants exposed to tobacco smoke develop these adverse perinatal outcomes. It is therefore reasonable to assume that several interactions exists between tobacco smoke exposure and biochemical, genetic, and epigenetic factors which make the fetus more susceptible to changes in fetal programming.

Our findings are of clinical relevance because responses of mitochondrial DNA and changes in the fetal methylome are plausible alterations that may underlie the adverse effect of tobacco smoke exposure on birth weight. They increase our knowledge on the mechanisms of perturbations in the fetoplacental exchange that might lie at basis of low birth weight and, hence, may be used in the broader sense of clinical context.