Maternal smoking and DNA methylation abnormalities in children at early developmental stages

According to the World Health Organization, tobacco smoking is one of the leading risk factors for morbidity and premature mortality among middle-aged adults. Moreover, for women, it is the second most important risk factor after high blood pressure. While in adults smoking leads to the risk of developing various systemic diseases [1], the effects of smoking on ongoing pregnancy can lead to irreparable fetal malformations and affect the health and development of the newborn, as well as the quality of its life in general. Thus, smoking during pregnancy is known to be associated with respiratory failure and more pronounced asthma symptoms in childhood, low birth weight, orofacial fissures, sudden sensorineural hearing loss (SSHL), or sudden deafness, in newborns, otitis media, neurobehavioral disorders, etc. [2–7]. It has been established that smoking during pregnancy acts as an exogenous factor which interferes with the fetus development in a dose-dependent manner [8].

Traditional research studying the relative influence of genetic and environmental factors on individual variability in health and morbidity indicators tends to focus on the correlation between a disease, external factors and genotype. Studies conducted over recent decades in the field of human genomic have shown that epigenetic regulation of gene expression, particularly in the early stages of development, plays a significant role in the development of diseases and systemic disorders, along with the genome structural variation [9]. Epigenetic regulation is carried out through special mechanisms which control the functional activity of gene promoters (nucleic acid sequences, recognizable by RNA polymerase and transcription factors as a starting point for transcription) and other regulatory regions of the genome. These mechanisms include DNA methylation, histone modifications, chromatin remodelling and RNA interference [10]. Epigenetics (from the Greek epi - upon, above) focuses on these mechanisms while studying the patterns of heritable changes in gene expression that do not involve changes in the underlying DNA sequence.

This review focuses on one of the main epigenetic mechanisms, DNA methylation, and its role in human embryonic development, with a special emphasis on the impact of tobacco smoking during pregnancy on the child’s epigenetic status.

DNA methylation as an epigenetic mechanism

DNA methylation is one of the main and most studied epigenetic mechanisms [10]. This mechanism of gene expression regulation is involved in a number of key processes, including genomic imprinting, X-chromosome inactivation [11], repression of transposable elements [12] and tissue-specific gene expression [13] that underlies the structural and functional diversity of cells and tissues in the body. DNA methylation is primarily a process by which a methyl group (CH3) is added to one of the DNA’s nitrogenous bases, cytosine. This modification predominantly occurs in CpG dinucleotide sites (C means cytosine which is followed by G that stands for guanine and linked by a phosphate group, p). This site has a pair in the complementary strand of the DNA double helix. It makes it possible to restore methylation on the newly synthesized daughter strand after DNA replication (Fig. 1) that, in turn, allows maintaining the DNA methylation pattern over many cell generations. CpG-rich genomic regions are called CG islands (CGI), or methylation island. About 70% of human genes have a high CpG content in their promoter regions.

DNA methylation is catalysed by a group of enzymes, where DNA methyltransferases 1 and 3 (DNMT1 and DNMT3) play the major roles. DNMT1 has an intrinsic affinity for hemimethylated DNA. DNMT1 is responsible for the maintenance of methylation during DNA replication. It is particularly important for those tissues where active cell division occurs throughout the whole life. The DNMT3 family members interact with unmethylated DNA; they play an important role in establishing a tissue-specific de novo methylation pattern at the earliest stages of development. During embryogenesis DNMT3a and DNMT3b are the first enzymes, which establish a specific methylation pattern required for a particular structural and functional state of chromatin during ontogenesis [14]. The ratio of methylated and unmethylated cytosines required for normal development is maintained through both a balanced methyltranspherase activity and a process of removal of redundant methyl groups from DNA molecules referred to as demethylation (Fig. 2 ).

DNA methylation is a reversible process, which makes epigenetic inheritance flexible [15]. Unlike errors in the genome structure, methylation errors can be corrected or reduced, but that is what makes the system vulnerable. In turn, the nature made DNA methylation as a multiple stage process controlled by several monitoring systems. DNA demethylation can occur through a number of processes including a passive loss of methyl groups and an active removal of methyl groups, or DNA demethylation mediated by enzymes. In the developing human embryos, the demethylation with an intermediate oxidation state of methylcytosine to hydroxymethylcytosine plays an important role [16]. The proteins of TET family (Ten to Eleven Translocation enzymes, whose hyperactivity was first found in the cells with chromosomal translocation t(10;11)(q22;q23) in patients with acute myeloid leukemia) are essential for this process. They contribute to the rapid removal of methyl groups in CpG islands facilitating access to specific gene promoters for transcription factors.

It is known that 10-15% of CpG sites are methylated in a cell-specific manner [13]. This methylation pattern is inherited from one cell generation to another, but cells can adjust it and activate certain genes in response to external factors. Abnormalities and failures of this well-functioning and controlled system may lead to the development of diseases. Thus, an increase in the activity of demethylases has been identified as a contributing factor to the development of cancer. On the contrary, the suppression of their activity results in an increase in methylation level of CpG islands in gene promoters leading to the disruption of their work [17].

Due to their plasticity, DNA methylation patterns change with age in response to the influence of acquired experience and environment, lifestyle and socio-emotional factors [18, 19]. In a healthy human body methylation and demethylation are under strict control and help promptly react to the current needs of the body in changing environments. Due to its dynamic nature, DNA methylation is a widely used biomarker in the research aimed to investigate the role of the environment in the development of cancer and other diseases in the context of their molecular etiology [20–24].

DNA methylation dynamics during embryogenesis

While DNA methylation changes, in adults it can lead to the disruption of the functioning of individual cells and tissues or to their malignant transformation, under adverse conditions. In a developing organism, even a short-term disruption in coordinated epigenetic regulation may have a crucial impact on the development and may cause irreparable damages. The severity of such damages depends on the magnitude of epigenetic disruptions and the stage of development during which they occurred. Namely, during gonadogenesis and embryonic development the most rapid and dramatic dynamic changes occur when cell-specific methylation patterns are establishing via several waves of the global demethylation and de novo methylation [25, 26].

A new human birth is preceded by the formation of parental germ cells or gametes – an egg and a sperm. The DNA of primordial germ cells is considerably methylated; when cells migrate to undifferentiated gonads, the methylation level drops. During germ cell maturation, epigenetic information is largely erased as the DNA goes through active demethylation. These processes differ in oogenesis and spermatogenesis, and as a result the genetic information from parents is marked differently. These differences will persist in all offspring’s cells and will only be erased in the reproductive tract cells during gametogenesis [27, 28].

Immediately after fertilization, the second round of epigenetic reprogramming begins, which is accompanied by global changes in the DNA methylation and histone modifications. In a zygote, the level of DNA methylation decreases first in the paternal pronucleus and then in the maternal pronucleus [16, 27, 28]. Methylation levels continue to drop until a morula formation. This process involves both active and passive demethylation mechanisms. At the blastocyst stage, the process of remethylation starts, which provides a blastomere differentiation into the inner cell mass (ECM) and the trophectoderm (TE). After implantation, as a result of de novo methylation, epigenetic differences between ECM cells and TE are established. These newly established DNA methylation patterns are preserved during subsequent cell divisions; further differentiation of ECM cells into three germ layers is accompanied by the establishment of the layer-specific DNA methylation patterns through local methylation changes. Both maternal and paternal alleles have equal opportunities to be expressed in any cell of the offspring; for most genes, maternal and paternal alleles are expressed equally. However, several hundreds out of approximately twenty-five thousand human genes are subject to genomic imprinting, a process that causes genes to be expressed in a parent-of-origin-specific manner, and which involves differential methylation of maternal and paternal alleles.

The data on DNA methylation dynamics have been predominantly provided by the research using the model organisms. Altogether they indicate that there is an early development stage when epigenetic marks are removed, followed by the establishment of de novo methylation profiles and chromatin remodelling. In humans, reprogramming of the parental genomes occurs during the first cells divisions, and at the morula stage the establishment of new epigenetic signatures, which are specific for differentiating embryo tissues, begins. Any disruptions of epigenetic reprogramming, including those resulting from the negative impact of external factors, may disturb the ontogenesis program causing aberrant gene expression that, in turn, may lead to severe pathologies or atrophies, embryonic death or fetal malformations [16, 26].

Maternal smoking during pregnancy and offspring DNA methylation

There are numerous empirical publications and literature reviews focused on the effects of maternal smoking during pregnancy. In addition to indirect impact through the deterioration of the mother’s health, direct consequences of maternal smoking on fetal development are widely known in pediatrics [2]. Among those are the effects of nicotine accumulation in the fetal blood, amniotic fluid and breast milk (it should be noted that the concentration in the fetus is usually 15% higher than maternal levels). Nicotine effects are seen in every trimester of pregnancy: from spontaneous abortions during the first trimester to increased risk of premature birth and low birth weight during the last trimester. It has been shown that nicotine affects both factors determining the birth weight—gestational age and fetal growth rate. Both animal models and human research have revealed that nicotine increases the mother’s blood pressure and heart rate, while reducing uteroplacental blood flow. In addition, carbon monoxide in tobacco smoke forms carboxyhaemoglobin, which inhibits the release of oxygen into fetal tissues. Together with the effect of cancerogenic xenobiotics contained in cigarette smoke, this leads to toxication and hypoxia, affecting many systems of the developing fetus, especially the respiratory and nervous systems [2–4]. Consequently, this prevents normal development of the child, leading to congenital brain defects and malformations in other organs, and can lead to behavioural impairments manifested in crying for no reason, sleep disturbances, and later, in uncontrollable aggression, among others.

As we have pointed out, active smoking may disrupt DNA methylation, and the prenatal period is the most critical developmental stage for establishing an individual’s epigenetic status. The first few comparative studies showed that the genomes of blood cells in the preschool and primary school children with a history of maternal smoking during pregnancy were characterized by changes in the global and gene-specific DNA methylation [29]. Moreover, a study of the association between these changes in DNA methylation and the duration of exposure to tobacco smoke has found that the methylation signatures point to sustainable, rather than short-term effect of maternal smoking during pregnancy [30]. It has been shown that many epigenetic changes associated with maternal smoking have a long-term effect; they are persistent throughout adult life and are detectable around age 40 regardless of active smoking [31].

The studies of the effects of maternal smoking on offspring development utilize two different approaches. The first of them is a ‘retrospective’ approach, which examines the long-term epigenetic consequences comparing the cohorts of adults [31] and children (usually young ones) [29, 32] with a history of in utero exposure to maternal smoking with an age-matched comparison cohort of individuals without a history of maternal smoking during pregnancy. The second approach is based on studying the characteristics of DNA methylation in placental and/or cord blood cells [5, 7, 30, 33-39], and neonatal peripheral blood [40], or fetal tissues. The latter is commonly used in the research based on animal models [41]; when applied to humans, such studies are based on the tissues of aborted embryos and they are few and far between [34, 42-44].

It should be noted that similar to the research focused on the epigenetic effects of active smoking in adults, the tissue-specific epigenetic response of the fetus to maternal smoking has been barely studied. At the same time, available evidence indicates that such a differentiated response is in place. Thus, it has been found in methylation differences for certain genes (AHRR and CYP1A1) between placental and fetal tissues [37], and also noted in a slight overlapping between the sets of genes associated with maternal smoking in blood cells and brain tissues [42]. Another aspect of the problem is that research implementing fetal tissues is conducted on the epigenetic system which is extremely dynamic during the prenatal period, as it was outlined in the previous chapter. The timing of fetal development is extremely important because it is not known when exactly during development the specific epigenetic alterations caused by maternal smoking occur and stabilize, becoming detectable. Thus, a study on brain tissues from embryos in the second trimester showed that the stage (early or late) of the trimester characterized by rapid brain development rather than maternal smoking was the main factor differentiating individual epigenomes in the studied cohort [42]. In addition, the authors [42] registered the effects of maternal smoking on offspring DNA methylation, and the greatest effect of prenatal exposure to nicotine was assigned to the epigenetic changes associated with slower neuronal maturation and/or a decreasing number of mature neurons.

One of the first large-scale whole-epigenome studies published by B.R. Joubert with co-authors [30, 35] was conducted on a cohort of over a thousand newborns (Norwegian Mother and Child Cohort Study). Having analyzed DNA methylation in cord blood cells and assessed the impact of prenatal tobacco exposure based on circulating maternal cotinine (a nicotine metabolite and a stable biomarker of cigarette-smoking), the authors identified epigenetic changes in 10 genes significantly associated with maternal cotinine levels. The most important among them were the genes involved in the xenobiotic-detoxication, namely AHRR and CYP1A1. It is noteworthy that these two genes have been consistently detected in the association with active smoking in adults. A few years later these results were confirmed and were supported by additional findings in a number of studies [5, 32, 35, 36, 39, 40].

The Pregnancy and Childhood Epigenetics consortium (PACE) combined data across studies utilizing the same platform for whole-genome DNA methylation profiling—the Infinium HumanMethylation 450 array. They conducted a large-scale epigenome-wide meta-analysis using the data on whole-genome DNA methylation in cord blood from 13 birth cohort studies from the USA and Europe with a detailed history of maternal smoking during pregnancy [38]. To date, this is arguably the most extensive and comprehensive study of the effect of maternal smoking during pregnancy on DNA methylation in offspring. Using the results from older children cohorts, the researchers investigated the long-term epigenetic alterations triggered by prenatal exposure to nicotine with a correction for the effects of postnatal second-hand tobacco smoke exposure. They also took into account the pre-existing smoking history and smoking frequency during pregnancy, whether it was sustained or occasional smoking. The cohort-specific associations between maternal smoking during pregnancy and DNA methylation in offspring were a subject of the meta-analysis. For the functional significance of newly identified epigenetic alterations, the authors evaluated the associations between the methylation statuses and expression levels of a number of genes and performed the functional analysis of gene-networks involved in the response to prenatal exposure to nicotine.

The following observations and conclusions were made based on the results of this large-scale study [38].

1. Maternal smoking during pregnancy causes significant dysregulation in the fetal genome. This study revealed over six thousand CpG sites across the genome, which showed significant changes in methylation status in the children whose mothers smoked during pregnancy, with the AHRR gene taking a leading position with respect to the association significance (1.64x10-193).
2. The epigenetic alterations registered in newborns are the long-lasting epigenetic changes. Thus, although an attenuation of the effects was observed in older children, there was, nevertheless, a high concordance in the methylation statuses between the newborns and older children with a history of maternal smoking, statistically confirmed for 73% (four out of six thousand) epigenetic markers.
3. As expected, any maternal smoking pattern affects the fetus. Sustained maternal smoking, however, is characterized by a greater effect on epigenetic alterations that, in turn, may be associated with a higher impact of maternal smoking on the child’s health and development.
4. The fact that such negative health and developmental outcomes exist is well-known. However, the results of the functional analysis of the genes whose methylation changes as a result of prenatal exposure to tobacco smoke suggest that the negative outcomes may be partly due to epigenetic alterations driven by maternal smoking.. Thus, these genes are predominantly involved in the control of key developmental processes, such as growth and anatomical development, including those specifically related to embryonic morphogenesis (or genes, which are active during embryogenesis), of the nervous system development, cell growth and proliferation [38]. Epigenetic changes can undoubtedly be one of the molecular mechanisms linking the effects of prenatal exposure to nicotine with various negative consequences for the child’s health and development. These findings have been confirmed by further studies, albeit not numerous, on identifying the associations between maternal smoking – epigenome – phenotypes, which have included obesity [5], inadequate immune response [7] and the central nervous system developmental disorders [6].

Epigenetics of smoking. Open questions and future research

Smoking-related diseases continue to be a major public health concern, and understanding the mechanisms of the health effects of smoking is an important part of research. Genomic research of the past 10 years has shown that chronic exposure to tobacco smoke is an adverse environmental stimulus that is capable of modifying DNA methylation patterns that, in turn, may extend the effects of smoking onto gene expression and, eventually, lead to impairments and diseases associated with smoking. It has been found that the alterations in DNA-methylation driven by tobacco smoke can occur at all stages of development: in adult life, as a result of active smoking, and during prenatal development, as a result of maternal smoking. In adulthood, smoking affects methylation patterns, which are already established and are maintained throughout life and during cell division. While, prenatal exposure to tobacco smoke happens at a time when DNA methylation is highly dynamic, methylation patterns evolve, and the epigenetic status of the organism is setting up. As a result, prenatal effects of tobacco smoke are multiplied, having a significant impact on the program of embryonic development.

Tobacco smoke primarily affects the specific genomic regions and relevant genes; thus, identical alterations in methylation patterns of genes involved in chemical detoxication of tobacco smoke components (for example, AHRR and CYP1A1) are found in smokers’ somatic tissues, smoking mothers’ placenta and fetal tissues. In addition, smoking destabilizes methylation patterns at the global genomic level. It would also seem that the pattern of these alterations in different cells and tissues may vary in the effect on different gene cascades. This was observed both in somatic tissues of adult smokers and in differentiating embryonic tissues.

Despite significant progress in research of the epigenetics of smoking, many issues require further research; of those, the following two should be highlighted. Firstly, while an association between smoking and methylation disturbances in certain genes has been established, the epigenome-wide response to smoking, in terms of the genome dysregulation at the level of gene systems and gene cascades, remains understudied. Secondly, it has been shown that smoking can change DNA methylation in various tissues, and some of these changes may vary between the tissues [42, 45, 46]. However, the majority of studies were carried out on peripheral blood cells, which are the most accessible biological material. Thus, the issue of differential epigenomic response to smoking in various tissues and organs remains understudied.

Further research of the effect of smoking on the epigenome, including its impact on DNA methylation, has a high potential. Firstly, new findings in this field may contribute to understanding the molecular mechanisms of the impact of tobacco smoke on the body and the development of impairments and health issues associated with smoking. Secondly, the epigenetic alterations resulting in radical changes in the phenotype are of a particular interest because by nature they are either fully or partially reversible. This reversibility potentially allows running a scenario where the ‘epigenetic code’ can dictate the expression of a particular set of genes, essentially acting as an on/off switch. Such ‘epigenetic drugs’ have already been developed and used for certain neurological diseases and cancers [47, 48]. Regarding the smoking, any epigenetic drug, which could counteract its harmful effects, appears to be a creation of the distant future. Nevertheless, even today’s studies of the influence of certain substances on the epigenetic outcomes of maternal smoking in offspring have borne fruitful results. Thus, a randomized clinical study showed that adding vitamin C to the diet of a smoking pregnant woman helps normalize DNA methylation in the baby and reduce such a known effect of maternal smoking as respiratory problems in babies [34].