Identification of miRNAs Involved in Foetal Growth Restriction Due to Maternal Smoking during Pregnancy

Introduction: Smoking during pregnancy is associated with reduced foetal growth, amongst other effects. Epigenetic modification in the foetus and placenta during embryonic development as a result of changes in the function of miRNAs is one of the pathophysiological mechanisms responsible for this. This dysregulation may be due to environmental changes or toxins such as tobacco. Objective: To study the impact of smoking during pregnancy and its role in intrauterine growth restriction via hypermethylated miRNAs. Materials and methods: The differences in methylation patterns for miRNAs in umbilical cord blood from low-birth-weight newborns of smoking mothers were compared with those from normal-weight newborns using MedIP-seq (StarArray). Results: Seven hypermethylated miRNAs were identified in the epigenetic study of cord blood from low-birth-weight newborns of smoking mothers in our sample. The miRNAs found to be hypermethylated were: MIR7-1, MIR3918, MIR1244-1, MIR4721, MIR25, MIR93, MIR3656. Conclusion: Intrauterine exposure to tobacco induces hypermethylation-mediated miRNA silencing in low-birth-weight newborns by modifying the expression of factors involved in vascular development, growth, and adaptation to hypoxia.


Introduction
Epigenetics refers to the modification of gene expression without altering the DNA sequence [1][2][3][4]. DNA methylation is one of the main mechanisms involved in regulating gene expression. Under normal conditions, methylation patterns are specific for each tissue in the body and, therefore, for each species, and can be inherited. During embryonic development (from the eighth cell division onwards), cells undergo new methylation processes, which may also appear in somatic cells of adults. An inversely proportional relationship can be established between DNA methylation and gene expression [1] and may be subject to changes induced by the environment or external toxic factors, such as tobacco, malnutrition or contamination [2,4,5].
Intrauterine growth restriction (IUGR) is a heterogeneous entity that encompasses numerous aetiologies and requires a factor inhibiting intrauterine foetal growth. Its incidence varies between populations, and it is estimated that around 30 million children are born annually worldwide with a low weight for their gestational age. Maternal malnutrition is the leading cause of IUGR worldwide, followed by the intake of toxins, especially tobacco, both actively and passively [6]. A study conducted by our group revealed that the most significant idiopathic aetiological factors related to IUGR are tobacco, maternal stress, total number of months worked during pregnancy, hours worked per day and time standing [7].
Tobacco is a negative factor for foetal growth. Smoking affects length, weight and head circumference and its effects can be detected by week 30 of pregnancy, with a weight reduction of approximately 250 g having been reported. However, the literature on the link between maternal smoking and restricted foetal growth is not completely consistent, as a reduction in foetal growth has been reported in some, but not all, studies, and others have documented direct tissue damage and impaired placental circulation [8].
Epidemiological studies in humans and experimental animals have shown that a suboptimal intrauterine and breastfeeding environment affects growth and predisposes the newborn to other diseases in adult life. Maternal smoking during pregnancy may modify the methylation pattern of foetal DNA in genes involved in foetal growth and development [9] and may also induce epigenetic modifications and alter gene expression during pregnancy, thereby affecting prenatal growth, foetal and neonatal phenotype and susceptibility to developing diseases as adults such as diabetes mellitus, obesity, dyslipidaemia, asthma, behavioural disorders and cardiovascular disease, amongst others. A positive association between maternal smoking and obesity or overweight in offspring has been demonstrated, whereas the relationship with type 2 diabetes remains uncertain [10].
Epigenomic wide association studies (EWAS) have shown the association between tobacco and DNA methylation in cord blood and newborns, with some of these alterations being reversible postnatally. In addition, recent studies have shown that some of these modifications to the DNA methylation pattern may persist postnatally until adolescence [11]. The identification of epigenetic changes associated with tobacco exposure could be used in the future to identify the population susceptible to developing chronic diseases as adults. As such, study of the methylation in these regions may be an indicator of prenatal harm. The possible reversibility of these epigenetic modifications opens up a window of opportunity to modulate and change the evolution of prenatally "programmed" chronic diseases.
Exposure to tobacco smoke during pregnancy may be associated with an increase in miRNA genes methylation and therefore altered miRNA expression in newborns [12]. miRNAs regulate gene expression post-transcriptionally by inducing gene silencing. They bind to messenger RNAs due to their complementarity, thereby preventing their translation [4]. They also act as internal regulators for a large number of genes, including the retrotransposons required for the stability of DNA or regulation of processes such as cell growth and differentiation [13].
Between 1100 and 1800 miRNA-coding genes are currently known in humans and mice, although they are also found in numerous other species. These genes are involved in several physiological (control of the cell cycle and division and tissue growth and differentiation) and pathological processes (cancer and neurological diseases) [14].
The role of miRNAs in growth has been studied in various animal models. Analysis of altered miRNAs in human growth impairment suggests that miRNAs contribute to growth physiology by regulating the hypothalamic-pituitary-IGF axis, growth hormones (GH), insulin-like growth factors (IGF) and their related receptors [15]. miRNAs can cross the placental barrier and reach foetal blood. Their regulation occurs in maternal, foetal and placental tissue. Four miRNAs that are specifically expressed in the placenta-miRNA-141, miRNA-149, miRNA-229-5p and miRNA-135b-and can be detected in both maternal and foetal blood [16].
Furthermore, one of the factors known to cause placental dysfunction leading to the onset of IUGR is exposure to tobacco smoke, which is associated with an increase in neonatal mortality and morbidity and is also related to an increased risk of developing adult diseases in exposed children [17]. The methylation processes resulting from nicotine exposure can trigger a decrease in the expression of genes that are key to correct development of the placenta and uteroplacental arteries, as well as genes that are essential for adequate foetal growth. Nicotine also acts directly on placental acetylcholine receptors, causing vasoconstriction and increasing uterine vascular resistance, thereby resulting in decreased foetal blood flow [18].
Given the relationship between tobacco use during pregnancy, the epigenetic modifications caused by this exposure and IUGR, the objectives of this study were:

1.
To compare the difference in miRNA methylation in low-birth-weight newborns of smoking mothers with those in normal-weight newborns of non-smoking mothers; 2.
To establish a relationship between tobacco, altered miRNA expression patterns and IUGR.

Materials and Methods
This is a transversal study conducted in newborns at the Hospital Clínico Universitario Lozano Blesa (Zaragoza, Spain). Sample collection commenced on 1 October 2018 and was completed on 30 December 2019. The sample comprised ten umbilical cord blood from newborns of smoking and non-smoking mothers.

1.
Low-birth-weight newborns [19,20] exposed to tobacco smoke: children of mothers who smoked more than 10 cigarettes per day during the first trimester of pregnancy; 2.
Normal-weight newborns not exposed to tobacco smoke: children of mothers who did not smoke during pregnancy.

1.
Diseases in the mother that could cause IUGR; 2.
Not signing informed consent.
Study of differentially methylated miRNAs in umbilical cord blood

Procedure and Methodology
Patients who complied with the inclusion criteria were selected consecutively in the obstetrics clinic until obtaining the total sample size. After receiving explanations about the study, they were asked to sign the informed consent. The study was approved by the Ethical and Research Clinical Committee of the Health Research Institute of Aragon.
A cord blood sample was obtained during delivery and transferred to the Genetics laboratory at the School of Medicine of the University of Zaragoza for subsequent DNA extraction.
The genes to be studied had been selected previously by way of a MeDIP-seq (Methylated DNA immunoprecipitation sequencing) analysis.
For this analysis, a DNA sample was taken and cleaved into fragments of between 200 and 800 base pairs. These were then ligated to Illumina adapters and subsequently immunoprecipitated using the anti-5-methylcytosine antibody. The enriched DNA was amplified by PCR and purified. A subsequent quality control (QC) was performed.
The short fragments from the sequences generated using HiSeq4000 were then analysed from the 5 to the 3 end and aligned with the previously selected reference genome. A mapping was then performed thus obtaining the different hypermethylated sequences, the associated long non-coding RNA (ncRNA) fragments, the messenger RNA (mRNA) fragments and the small non-coding RNA (smallncRNA) fragments. The latter are part of the RNA silencing complex, which comprises the siRNA, piRNA and miRNA, with the latter being the subject of our study and of the analysis of results.
The percentage of cytosine methylation in the miRNA promoter sequences was quantified and the existence of significant differences between smoking habit and IUGR was analysed.
The selected methylated miRNA gene were analysed using the mirBASE and miRDB databases to determine in which regions of the genome they may act (TARGET) and induce gene silencing.
The mirBASE database is a platform that allows for searching published miRNA sequences and data annotation. Each entry contains the information for the hairpin precursor from which transcription of the different mature miRNAs is obtained, locating and classifying them on the basis of the region of the genome in which they act to induce post-transcriptional silencing. This database was created and is administered by the Griffiths-Jones laboratory in the Faculty of Biology, Medicine and Health at the University of Manchester.
After this initial search, another was conducted in the miRDB database. On this platform, each of the targets described according to the base complementarity of the different miRNAs studied was analysed, thereby predicting the functions of each of them.
After searching for possible targets for the miRNAs described in Tables 2 and 3 using the mirRDB database, the most relevant in terms of foetal growth were selected. The possible targets for each of the miRNAs described can be found in Tables 4-11. No targets were found for the sequence of miRNA3656.
With regard to foetal development, we have found targets for these miRNAs related to: 1. Placental angiogenesis, as Vascular Endothelial Growth Factor A (VEGFA) and Angiogenic Factor 1 (FA1); 2.

Discussion
Our study presents new findings concerning the relationship between miRNA regulation by methylation during pregnancy and foetal growth. We identified hypermethylation in seven miRNAs in cord blood from low-birth-weight newborns of smoking mothers with statistically significant differences with respect to normal-weight newborns of non-smoking mothers. These miRNAs are involved in foetal development by regulating processes such as placental angiogenesis, foetoplacental growth and foetal oxygenation mechanisms.
Exposure to tobacco during pregnancy affects the normal miRNA methylation patterns in foetuses, thereby inducing a decrease in gene expression in the regions affected, which in turn leads to deregulation of the synthesis of the resulting proteins [21,22]. miRNAs are essential regulatory elements in embryonic development. Alteration of the expression of the miRNAs found in our study (MIR7-1, MIR3918, MIR1244-1, MIR4721, MIR25, MIR93, MIR3656) may be associated with the decreased expression of proteins and key factors for correct placental development.
Firstly, we observed methylation-mediated silencing of miRNA-7-1 and miRNA-93 with targets in pro-angiogenic factors related to foetal and placental growth (VEGFA and PLGF). This should result in overexpression of these growth factors, which would in turn lead to normal foetal growth. However, we have observed this alteration of expression in foetuses with IUGR, which could be explained by an alteration of angiogenesis caused by the overexpression and related to the endothelial vascular dysfunction found in IUGR. Our findings are in agreement with those of Luo J et al., who found a statistically significant decrease in the expression of various families of miRNAs, especially miRNA-150, which exhibits pro-angiogenic activity in venous cord blood in foetuses with IUGR compared with the control group, in which the expression was normal [23]. A more recent study has also demonstrated the relationship between the expression of miRNA-206, the pathophysiology of IUGR and the vascular endothelial growth factor (VEGFA) in maternal blood [24].
Another possible target, in this case for the 3 end of miRNA-25, found with regards to vascular dysfunction is angiogenic factor 1, which is related to correct blood vessel development [25].
Secondly, miRNA-3918 targets placental growth factor (PLGF) and human chorionic somatomammotropin hormone (CSHL1). These proteins are essential for correct development of the placenta and adequate evolution of pregnancy. A decrease in PLGF is related to the risk of suffering pre-eclampsia, the pathophysiology of which is closely related to that of IUGR. Indeed, PLGF is currently used clinically for the early diagnosis of pre-eclampsia and IUGR [26][27][28].
Surprisingly, we found hypermethylated miRNA-3918 in foetuses with restricted growth, which means that PLGF expression should be increased, a finding that contradicts what would be expected for a foetus with IUGR as, in such cases, PLGF concentration should be reduced. However, previous studies have reported similar findings, with an increase in PLGF in IUGR, meaning that this factor may behave in different ways depending on the degree of placental oxygenation and the interaction with other pro-angiogenic factors such as VEGFA [29,30]. This situation is similar to our findings for angiogenic factors.
Thirdly, the 5 end of miRNA-93 and miRNA-3918 target insulin-like growth factorbinding protein 1 (IGF2BP-1) and 2 (IGF2BP-2), respectively, which are essential for correct foetal growth and have high expression in most foetal organs during embryogenesis. The proteins IGF2BP1 and IGF2BP2 interact with 5'-UTR of IGF2 and protect it from miRNAmediated silencing [31].
IUGR is a consequence of placental vascular dysfunction, and several other factors are involved in its pathophysiology. The IGF system is one of the most important endocrine and paracrine growth factor system that regulate foetal and placental growth and IGF-2 plays a major role in preservation of placental growth and function. IGFBP-2 concentration in placental lysates resulted to be an important variable in determining foetal growth [32].
Altered expression of IGFs and IGF binding proteins during intrauterine growth retardation has been shown in experimental animal models with elevated plasma IGFBP-2 and IGFBP-4 and reduce IGFBP-3 [33]. IGFBP-2 is the principal IGFBP in foetal plasma and after uterine artery ligation a strong increase in the concentration of IGFBP-2 has been observed and this mimics the rise in IGFBP-1 and IGFBP-2 levels seen in human IUGR [34]. IGF-2 cord serum concentration has a positive effect on both birth length and weight, whereas IGFBP-2 had a significant negative effect [35]. The proportional increase in IGFBP-2 in the circulation of foetal growth restricted foetuses without changes in IGFs indicates that there might be a causal relationship between the increase in IGFBP-2 and foetal growth. The increase observed in IGFBP-2 in growth-restricted foetuses would be a compensatory mechanism for a greater IGF tissue bioavailability as IGFBP-2, in contrast to IGFBP-3, does not form a long-lived ternary complex with the IGFs [36] and our observation with higher expression of IGFBP-2 gene due to hypermethylation of miRNA-3918 is consistent with this hypothesis.
It has been shown that in the female sex-selective increase in IGF-2 expression in the hypoxic guinea pig placenta of growth-restricted foetuses as a mechanism of adaptation [37].
The reduced expression of IGF-2 in children with IUGR and its relationship to the altered expression of miRNAs have been reported in previous studies, as has its involvement in placental dysfunction [38].
Finally, hypoxia-inducible factor-3 is a protein coded by the HIF3 gene, which is a possible target for miRNA-4721. HIF3 plays a key role in the adaptive tissue response to situations of chronic hypoxia that lead to alterations in tissues and organs. Hypoxiainducible factors (HIF1, 2 and 3) regulate oxygen homeostasis by controlling angiogenesis during periods of oxygen deficit. When HIF-1 levels decrease, HIF-2 and HIF-3 increase. This switch from HIF-1 to HIF-2 and HIF-3 signalling is required in order to adapt the endothelium to prolonged hypoxia. Recently, the miRNAs that endogenously regulate gene expression via the RNA interference (RNAi) pathway have been shown to play critical roles in the hypoxia response pathways [39].
miRNA-4721 methylation and the increased expression of HIF-3 may be an adaptive response to chronic oxygen deficit that occurs in IUGR as a result of altered tissue perfusion.
Our findings suggest a new perspective with regards to the role played by miRNAs in the regulation of angiogenic and placental growth factors and IUGR.
The results of our study agree with the foetal programming hypothesis proposed by Barker, who posits that foetuses with IUGR compensate for the hostile intrauterine environment by modifying their metabolism, which subsequently remains during postnatal development, thus leading to chronic diseases in adult life [40,41]. Consequently, these miRNAs may prove useful as biomarkers for the diagnosis and follow-up of these diseases. In light of our findings, we propose a new mechanism concerning the epigenetic regulation mechanisms of miRNAs and the foetoplacental angiogenic and growth factors essential for correct foetal development.
The main limitation with regards to the study of miRNAs lies in the lack of functional studies relating a specific miRNA to an associated pathology. Our study proposes a new explanation for the development of IUGR as well as a new line of research regarding the role of miRNAs in the regulation of the epigenetic mechanisms that modulate foetal growth.
Further studies are required to shed more light on the effect of tobacco use during pregnancy on silencing due to miRNA methylation and foetoplacental development.  Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of Aragon CEICA (protocol code PI16/0208 and date of approval, 14 September 2016).

Informed Consent Statement:
Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.