http://orcid.org/0000-0002-4480-9448
Figures
Abstract
Background
In Fetal Growth Restriction ‘fetal programming’ may take place via DNA methylation, which has implications for short-term and long-term health outcomes. Small-for-gestational age fetuses are considered fetal growth restricted, characterized by brain-sparing when fetal Doppler hemodynamics are abnormal, expressed as a cerebroplacental ratio (CPR) <1. We aimed to determine whether brain-sparing is associated with altered DNA methylation of selected genes.
Methods
We compared DNA methylation of six genes in 41 small-for-gestational age placentas with a normal or abnormal CPR. We selected EPO, HIF1A, VEGFA, LEP, PHLDA2, and DHCR24 for their role in angiogenesis, immunomodulation, and placental and fetal growth. DNA methylation was analyzed by pyrosequencing.
Results
Growth restricted fetuses with an abnormal CPR showed hypermethylation of the VEGFA gene at one CpG (VEGFA-309, p = .001) and an overall hypomethylation of the LEP gene, being significant at two CpGs (LEP-123, p = .049; LEP-51, p = .020). No differences in methylation were observed for the other genes.
Conclusions
VEGFA and LEP genes are differentially methylated in placentas of small-for-gestational age fetuses with brain-sparing. Hypermethylation of VEGFA-309 in abnormal CPR-placentas could indicate successful compensatory mechanisms. Methylation of LEP-51 is known to suppress LEP expression. Hypomethylation in small-for-gestational age placentas with abnormal CPR may result in hyperleptinemia and predispose to leptin-resistance later in life.
Citation: Bekkering I, Leeuwerke M, Tanis JC, Schoots MH, Verkaik-Schakel RN, Plösch T, et al. (2019) Differential placental DNA methylation of VEGFA and LEP in small-for-gestational age fetuses with an abnormal cerebroplacental ratio. PLoS ONE 14(8): e0221972. https://doi.org/10.1371/journal.pone.0221972
Editor: Linglin Xie, Texas A&M University College Station, UNITED STATES
Received: May 9, 2019; Accepted: August 15, 2019; Published: August 30, 2019
Copyright: © 2019 Bekkering et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Fetal Growth Restriction (FGR) complicates 5 to 10% of all pregnancies and is a major risk factor for fetal and neonatal morbidity and mortality. Fetuses with an estimated fetal weight <10th percentile for gestational age are commonly classified as small-for-gestational age (SGA), however, many of these fetuses do reach their biological potential and are constitutionally small. Doppler assessment of fetal hemodynamics discerns between SGA and FGR. The fetal vascular response to placental insufficiency, resulting in FGR is called “brain-sparing”, a circulatory redistribution in which blood flow is preferentially shunted to the brain [1]. Brain-sparing can be quantified with the cerebroplacental ratio (CPR), calculated as Middle Cerebral Artery (MCA) Pulsatility Index (PI) divided by Umbilical Artery (UmbA) PI. As the name suggests, brain-sparing is thought to be a beneficial adaptive mechanism to prevent brain damage [2,3]. However, recent studies suggest that FGR fetuses with brain-sparing might be at an increased risk for neurodevelopmental and behavioral deficits, such as reduced gross and fine motor skills, and a reduced intelligence quotient [4].
Brain-sparing is indicating an adverse intrauterine condition that can lead to ‘fetal programming’, which has implications for short-term as well as long-term health. This programming can occur via alterations in DNA methylation, an epigenetic mechanism, which regulates gene expression. Differential methylation and/or gene expression has been reported in FGR for various genes, including those encoding for erythropoietin (EPO), hypoxia inducible factor 1 alpha subunit (HIF1A), vascular endothelial growth factor A (VEGFA), leptin (LEP), pleckstrin homology-like domain family A member 2 (PHLDA2) and 3-β-hydroxysterol delta-24-reductase (DHCR24) [5–8]. In this study, we focused on methylation of these genes in the placenta as they have profound roles in angiogenesis, neurodevelopment, immunomodulation and placental and fetal growth.
The methylation status of angiogenic genes in the placenta is of interest because of the possible hypoxic state that the FGR fetus can encounter. This hypoxic stress drives activation of angiogenic compensatory mechanisms such as EPO, HIF1A and VEGFA. EPO is expressed in the placenta and supports the fetal condition by e.g. regulating fetal red blood cell production, promoting erythroid differentiation and initiating hemoglobin syntheses [9]. HIF1A is a pro-angiogenic regulator in the anti-hypoxic response [10]. In response to hypoxia, HIF1A binds to the enhancer region of EPO and drives its expression [11]. The effect of HIF1A is favorable in the acute phase but can become pathological if prolonged, e.g. in placenta dysfunction in preterm preeclampsia (PE) [12]. VEGFA is an important regulator of angiogenesis which stimulates placental endothelial cell proliferation [8]. Increased VEGFA expression has been found in FGR placentas [8]. In preterm PE placentas, hypomethylation of VEGFA has been reported, resulting in upregulated expression [13]. We hypothesized that methylation of EPO, HIF1A and VEGFA in the placenta may play a role in the response to hypoxia and differs for fetuses that show brain-sparing.
With respect to the presence of brain-sparing, we expect to see methylation changes in LEP, PHLDA2 and DHCR24, as these genes are associated with neurodevelopment. LEP, commonly known as the “hormone of energy expenditure”, is an important metabolic hormone which is highly expressed in the placenta and regulates placental and fetal growth [14]. In the fetus, LEP is suggested to play a role in fetal growth, angiogenesis, hematopoiesis, brain development, and immunity [14]. Recent studies have associated changes in LEP methylation and expression with fetal growth, SGA, FGR, and neurodevelopmental outcome [6,7,15–19]. So far, methylation of LEP and its relation to the CPR has not been investigated. LEP is a possibly important hormone to investigate in FGR since short- and long-term adverse effects of FGR include obesity, metabolic syndrome, and diabetes. Altered methylation and/or expression of LEP may contribute to the susceptibility of FGR infants to these diseases. PHLDA2 is a well-known maternally expressed imprinted gene, which represses placental growth. This highlights the important balance of PHLDA2 needed for optimal fetal growth. Various studies have reported overexpression of this gene in SGA placentas and association with (low) birth weight, while others did not report a significant difference [20,21]. Overexpression has also been associated with reduced fetal movements followed by SGA, further emphasizing the clinical relevance of PHLDA2 [22].
DHCR24 encodes for an enzyme that catalyzes the final step in the synthesis of cholesterol. Cholesterol is essential for cell membranes and critical for fetal development, particularly brain development [23]. Furthermore, DHCR24 has an anti-apoptotic effect on neurons, is thought to be neuroprotective, and its expression is down-regulated in SGA placentas [24].
In this study, our aim was to investigate whether methylation of selected angiogenic and neurodevelopmental genes differs between SGA placentas with normal or abnormal Doppler fetal hemodynamics, expressed as a normal/abnormal CPR. This seems to be the case for VEGFA and LEP methylation. It is of clinical importance to discern the presence of brain-sparing with the CPR in these SGA placentas as this indicates a compensatory mechanisms to placental insufficiency which is considered to give rise to fetal programming.
Materials and methods
Study population
We performed a prospective observational cohort study from June 2012 until September 2014 at the Obstetrics and Neonatology Departments of the University Medical Center Groningen (UMCG). The study was approved by the Institutional Review Board of the UMCG and written informed consent was obtained from the mothers of the 48 SGA fetuses that were included [25]. SGA was defined as an estimated fetal weight below the 10th percentile or deflecting fetal growth of at least 30 percentiles. Exclusion criteria adopted were: multiple pregnancy; structural and chromosomal abnormalities; and intrauterine infection. All deliveries and admissions to the neonatal ward/neonatal intensive care unit took place in the UMCG. In our study 41 of the SGA cases were included, based on availability of both placental tissue and CPR data. Embedded paraffin slices of placental tissue from this cohort were stored at the UMCG. Usage of the samples was approved by the UMCG Medical Ethical Committee in line with the code of conduct for responsible use.
Selection of genes and primer design
We selected biologically relevant genes involved in placental/fetal growth, angiogenesis or neurodevelopment. The genomic target region (Table 1) was chosen based on existing literature regarding regulation by methylation and/or association with SGA, FGR, or PE.
The analyzed EPO promoter region has been reported to repress transcription via hypermethylation in cancer [26] and contains a hypoxia-inducible factor-1B binding site [27]. The selected HIF1A promoter region contains a hypoxia response element that is thought to facilitate positive HIF1A autoregulation via hypomethylation [28], and a kaiso binding site-3 suggested to mediate methylation-dependent HIF1A repression [27]. The VEGFA region has been studied by Sundrani et al. who reported hypomethylation in PE placentas [13]. Our LEP region overlaps the promoter region investigated by Lesseur and colleagues which is regulated by methylation [29–31]. Hypermethylation of the selected PHLDA2 promoter region silences expression [32]. For DHCR24 we selected a cytosine-phosphate-guanosine (CpG) dinucleotides-rich promoter region reported to regulate transcription via methylation [33].
All PCR and sequencing primers (Table 1) were designed with the PyroMark Assay Design software (Qiagen, Hilden, Germany) and a gene specific biotinylated reversed/forward primer approach was used to label the PCR product with biotin. For bisulfite amplification we prepared a mastermix containing 12.5 μL HotStarTaq DNA Polymerase, 10.5 μL sterile water and a 1 μL mix of forward and reverse primer (10:10 μM) per 1 μL bisulfite template. We included a mastermix negative control to check for contamination. Cycling conditions on the T100 Thermal Cycler (Bio-Rad, Hercules, CA) were the same for all assays except for HIF1A: 95°C for 15 min, 45 cycles of 94°C for 30 s, 58°C (HIF1A 56°C) for 30 s, 72°C for 30 s, followed by a final step of 72°C for 7 min and stored at 4°C. A DNA ladder and 3–5μL of each PCR product was loaded and run on 2% agarose gel with ethidium bromide staining to visualize presence or absence of PCR products and contamination.
Placenta was processed by investigators unaware of the clinical data. First, the fetal side was selected by our pathologist on a hematoxylin-eosin stained section. Cores with a diameter of 1mm from the fetal side were used for DNA isolation with the QIAamp DNA FFPE (Formalin-Fixed, Paraffin-Embedded) Tissue Kit (Qiagen). The FFPE tissue protocol is designed for sections of 5–10μm and was adapted for 1mm cores. This entailed vortexing the samples at step 3 for 1 minute and crushing the cores manually with a pipet tip. Quality and concentration of the isolated DNA was checked with the NanoDrop®ND-1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA). Samples were stored at -20°C.
For bisulfite conversion, we used the EZ methylation Gold-Kit (Zymo Research, Irvine, CA) according to the supplier’s protocol with an additional 30s centrifuge round at full speed after step 8. Samples were divided in two separate batches for adequate desulphonation time, random and blinded. In total, 500 nanograms of each DNA sample was bisulfite converted for subsequent analysis by pyrosequencing.
Quantification of methylation via pyrosequencing
We used the PyroMarkQ48 Autoprep (Qiagen) for pyrosequencing, according to the supplier’s protocol. Methylation levels were analyzed with PyroMark Q48 Advanced Software (Qiagen). Each of the CpG sites were quality control checked and the percentage of DNA methylation at a given CpG calculated. CpGs that passed the quality check were included in data-analysis. Excluded samples were repeated.
Doppler ultrasound
Doppler parameters of the UmbA and MCA were measured once a week prior to maternal admission to the hospital and twice a week for the duration of admission by two experienced operators, as described elsewhere [25]. The CPR was calculated as MCA-PI divided by UmbA-PI. CPR data was transformed to z-scores as appropriate for the gestational age-specific mean [25]. A CPR of <1 was considered abnormal. For data analysis, we used the last measured CPR before delivery, with a maximum of seven days prior to birth.
Statistical analysis
We assessed normality with the Shapiro-Wilk test, data with a Shapiro-Wilk p-value of < .05 was considered to follow a non-normal distribution. A parametric or non-parametric approach was chosen as appropriate for the data. Continuous variables were compared using Student’s T-test or Mann-Whitney U test, categorical data with Chi-square or Fisher’s Exact test, and correlations with Pearson’s Test or Spearman’s Rho. We used simple and multiple linear regression analyses to analyze independent associations between variables. For this, non-normal distributed data were logarithmically transformed. A p-value of < .05 (two-tailed) was considered statistically significant. Descriptive statistics are presented as mean (SD), median (range), or frequency (%). We used the software Statistical Package for the Social Sciences (SPSS) 23.0 for statistical analyses.
Results
Characteristics of the study population
The study cohort consisted of 48 SGA patients of whom we selected 41 based on availability of both placenta material and CPR data. Patients were divided in two groups based on a normal (>1) or abnormal (<1) CPR (Table 2). The groups significantly differed in gestational age at the time of Doppler scan as well as at birth. Birth weight, birth weight percentile and head circumference were smaller in the abnormal CPR group. The use of antenatal steroids was significantly higher in the abnormal CPR group (82%) compared with cases with normal CPR (29%). All infants in the abnormal CPR group were delivered via caesarean section compared to 42% of infants with a normal CPR. Apgar scores and venous pH were also lower for infants with an abnormal CPR. Duration of admission to the neonatal ward/neonatal intensive care unit was significantly longer, and more often complicated by Respiratory Distress Syndrome or sepsis in the abnormal CPR group.
Average methylation of selected genes
We calculated the average methylation for each of the six genes (Table 3) after assessing the correlation coefficients (S2 Table) between individual CpGs within one gene. We observed a marginal non-significant trend toward hypermethylation on average (p = .095) for VEGFA in the abnormal CPR group. For LEP we found an average difference between the groups of 2.5% (95% CI: 0.5 to 5), (U = 118, z = -2.28, p = .023). Average methylation of EPO, HIF1A, PHLDA2 and DHCR24 did not differ between groups.
Methylation of individual VEGFA and LEP CpGs
For VEGFA and LEP we compared the median or mean differences, depending on the normality of the distribution, at the individual CpGs between the CPR groups. Because of the marginal trend toward significance for the average methylation of VEGFA we analyzed all 3 CpGs individually (Fig 1, S1 Table). We intended to also analyze another CpG locus on VEGFA in close proximity to the transcription site, but were unable to design primers of high enough quality for analysis. Of these sites, only CpG-1 (-309) methylation showed a statistically significant increase of 0.74% (95% CI: 0.328% to 1.15%) in the abnormal CPR group, t(39) = 3.635, p = .001 (Fig 1B). This is a medium effect size (Cohen’s d = 0.642). LEP showed a trend towards higher methylation in the normal CPR group at all CpGs (Fig 2, S1 Table). Of these, CpG-2 and CpG-11 were statistically significantly different between CPR groups. Mean methylation at CpG-2 was 5.96% (95% CI: 0.014 to 11.91) higher for normal CPR than abnormal CPR, t(32) = -2.042, p = .049 (Fig 2B) and median methylation at CpG-11 was 7.99% (95% CI: 1.18 to 14.52) higher (U = 116, z = -2.33 p = .020; Fig 2C).
Mean shown for CpG-1 (-309) & CpG-2 (-313) and median for CpG-3 (-331). And C; boxplot with 5–95 percentile whiskers mean methylation of VEGFA CpG-1 (-309). Double asterisk (**) indicates significance of p <0.01.
A; Methylation levels of LEP represented relative to the Transcription Start Site (TSS). CpG-1 (-127), CpG-2 (-123), CpG-3 (-118), CpG-4 (-115), CpG-5 (-100), CpG-6 (-95), CpG-7 (-85), CpG-8 (-74), CpG-9 (-71), CpG-10 (-62), CpG-11 (-51), CpG-12 (-38), CpG-13 (-33). Values are mean or median methylation as given in S1 Table. BC; boxplots with 5–95 percentile whiskers. B; Mean methylation of LEP CpG-2 (-123). Normal CPR (N = 20) Abnormal CPR (N = 14). C; Median methylation of LEP CpG-11 (-51). Asterisk (*) indicates significance of p <0.05.
Association between CPR z-score and VEGFA/LEP methylation
To determine whether DNA methylation gradually changed as the CPR-z score deviated more from zero, we analyzed the association between the CPR z-score and methylation of VEGFA or LEP with Spearman's rank-order correlation. Because the CPR is not constant throughout gestation we used the CPR-z score. The average methylation of VEGFA, CpG-2, or CpG-3 did not correlate with the CPR z-score, however, VEGFA CpG-1 correlated negatively with CPR z-score, rho = -0.360, p = .021 (Fig 3A). The average methylation of LEP, and LEP CpG-2 did not correlate statistically significant (rho = .253, p = .11 and rho = .07, p = .693, respectively). LEP CpG-11 revealed a statistically significant positive correlation with the CPR z-score, rho (39) = .350, p = .025 (Fig 3B).
VEGFA CpG-1 (A), LEP CpG-11 (B). Asterisk (*) indicates significance of p <0.05.
Relationship between placental methylation CPR z-score, birth weight, and antenatal steroids
Because of the differences between the CPR groups in gestational age, birth weight z-score, and antenatal steroid usage we further investigated the influence of these parameters on the significantly altered CpGs: LEP CpG-2, LEP CpG-11 and VEGFA CpG-1.
Simple linear regression analyses did not reveal a relationship between methylation of LEP CpG-2 or VEGFA CpG-1 and gestational age, birth weight z-score, or antenatal steroid use. The CPR z-score predicted LEP CpG-11 methylation statistically significant, F(1, 39) = 4.699, p = .036, which explained 10.8% of the variance in methylation. Antenatal steroid use was associated with LEP CpG-11 methylation, F(1, 39) = 4.995, p = .031, and explained 11% of the variance in methylation. Gestational age was not associated with LEP CpG-11 methylation, F(1,39) = 2.498, p = .122. As gestational age was closely related to antenatal steroid use we decided not to enter gestational age in the multivariate model to avoid multicollinearity.
Multiple linear regression analyses
Simple linear regression analysis showed a statistically significant relation between LEP CpG-11 methylation, CPR z-score and the use of antenatal steroids. Therefore, we entered CPR z-score and antenatal steroid use in a multiple regression model for the prediction of methylation. With both variables included, the predictive value of the model was statistically significant F(2,38) = 3.310, p = .047. This combined model explained 14.8% of the variance in LEP CpG-11 methylation.
Discussion
Fetal programming might take place in response to the adverse in-utero environment such as in FGR, which has implications for the short-term as well as long-term health outcome. In this study, we investigated fetal programming of selected angiogenic and neurological genes. We analyzed methylation of EPO, HIF1A, VEGFA, LEP, PHLDA2 and DCHR24 in SGA placentas and found differential methylation of VEGFA and LEP depending on the presence of brain-sparing, defined as an abnormal CPR (<1). Average methylation levels of EPO, HIF1A, PHLDA2 and DCHR24 did not differ between SGA with or without an abnormal CPR. Comparison of average VEGFA methylation showed a marginal tendency towards significance between groups and CpG-1 (VEGFA-309) was significantly hypermethylated in the abnormal CPR group. For LEP, we observed an overall hypomethylation, and at the specific CpGs: CpG-2 (LEP-123) and CpG-11 (LEP-51) in the abnormal CPR group.
We found a decrease in methylation of 2.5% on average, 6% at CpG-2, and 8% at LEP CpG-11 in the abnormal CPR group. Methylation of CpG-11 is of special interest because this CpG is located at a binding site for CCAAT/enhancer binding protein-ɑ (CEBPA), a principal regulator of adipocyte differentiation and function [30]. Binding of CEBPA at this site highly activates transcription of LEP [31]. Methylation at this CpG results in transcriptional repression [29–31]. SGA infants with an abnormal CPR might thus have increased levels of Leptin in their placentas due to hypomethylation in the LEP region, specifically at CpG-11.
Recent studies have associated differential LEP expression and methylation with fetal growth, SGA, and neurodevelopmental outcome [6,7,15–19]. Lea et al. demonstrated less intense presence of Leptin in placentas from FGR singletons and growth-restricted infants in twin pregnancies [6]. Another study reported increased LEP expression in FGR [17]. Lazo-de-la-Vega-Monroy et al. support this by reporting elevated placental Leptin concentrations in SGA which was inversely correlated with placental weight [7]. In contrast to the elevated placenta levels, they showed that cord blood levels of Leptin and placental LEP receptor expression was lower in SGA, which both correlated positively with placental weight. Our findings are in line with the reported elevation of Leptin concentration [7,17], and is contradicting with decreased levels in SGA (9). Lesseur et al. investigated methylation of CpG-85 until CpG-33, as in the present study (CpG-7 to CPG-13), and did not observe an association between placental LEP methylation and birth weight [15]. They did find, however, an association between DNA methylation and LEP expression in their cohort. Disagreements between Lesseur et al. and our findings could result from our study-design. Our cohort consisted only of SGA cases and we compared methylation based on their CPR while Lesseur and colleagues compared placental methylation of SGA with appropriate-for-gestational age (AGA) fetuses.
Several potential confounders might have influenced our findings, as these were different between cases with abnormal and normal CPR, including antenatal steroids and gestational age at birth. Antenatal steroids were administered more often in the abnormal CPR group. Linear regression analyses revealed that both the CPR z-score and antenatal steroids were associated with LEP CpG-11 methylation. Both remained independent and significant predictors in the combined multivariate regression analysis. We were not able to discern exactly how much methylation was explained by which parameter in this small cohort because a large majority of the abnormal CPR group had received antenatal steroids. The observed hypomethylation and hypothesized elevated LEP expression could also be in line with the effects of antenatal betamethasone, as this is known to increase placental LEP concentrations [34]. Differences in methylation might also partly be explained by differences in gestational age between groups as it is known that gestational age is an important determinant of expression of leptin in placenta and of maternal leptin concentrations [35]. Placental LEP expression is higher and LEP receptors are less abundant in early pregnancy which could fit with the observed hypomethylation of LEP in the abnormal CPR group having a lower gestational age [36]. Therefore, we examined gestational age in our single and multiple regression analysis and found that this was not predictive of LEP CpG-11 methylation.
VEGFA was hypermethylated at CpG-1 in the abnormal CPR group. This specific CpG is a known binding site for HIF1A, an important pro-angiogenic factor which is beneficial in the acute response to hypoxia [13]. VEGFA is also a key-player in the anti-hypoxic response as it stimulates endothelial cell proliferation. This suggests that VEGFA is less expressed in the placenta of fetuses with an abnormal CPR because methylation can repress VEGFA expression [37]. The observed hypomethylation can be a sign of compensation to hypoxic stress in fetuses with a normal CPR, or that these fetuses were constitutionally small and not in a distressed and hypoxic state.
Differential methylation of VEGFA CpG-1 is of importance with regard to emerging preclinical research in placenta-directed gene therapy for FGR [38]. If this gene therapy would reach the clinical stage, assessment of VEGFA methylation in fetal cells derived from the maternal blood might be used to identify fetuses that could potentially benefit from placenta-directed gene therapy such as adenovirus vectors containing VEGFA. Again, the altered methylation status of VEGFA is the set-point of methylation throughout life which might predispose to the development of cardiovascular disease.
Our results suggest that adverse intrauterine conditions due to FGR and an abnormal CPR alter methylation of LEP and VEGFA. These altered methylation levels are a reflection of intrauterine conditions and represent the level of placental LEP and VEGFA during gestation. The placenta regulates its own growth and that of the fetus via LEP. Therefore, one might speculate that altered LEP levels are part of FGR pathophysiology. Altered LEP levels may dysregulate the development of important fetal endocrine feedback systems that regulate metabolism and adipose tissue differentiation. Furthermore, LEP may play a role in neuronal functional differentiation, glial development and development of the hippocampus and the cingulate cortex which might affect memory and cognition [39]. Altered LEP levels during fetal development might thus have an effect on memory, motor, and cognitive abilities.
Another function of placental LEP is immunomodulation which is essential for maintaining placental function at the interface between mother and her allogeneic fetus. With regard to immunomodulation, both hypomethylated LEP and hypermethylated VEGFA could contribute to low-grade inflammation. Although mostly known for its pro-angiogenic activity, VEGFA has an immunosuppressive role by inhibiting T-cell and dendritic cell development, and stimulating regulatory T-cells (Tregs) and myeloid-derived suppressor cells [40]. In contrast, LEP is a powerful pro-inflammatory hormone which can activate the innate and adaptive immune response [41]. LEP inhibits differentiation of Tregs and drives macrophage polarization to the inflammatory M1 phenotype. Extrapolating the hypo- and hypermethylation of LEP and VEGFA, respectively, this would suggest an increased set-point for LEP and decreased VEGFA levels in SGA fetuses with brain-sparing. These changes combined may predispose the infant and later adult, to a pro-inflammatory epigenetic profile that cannot easily be altered. Fetal programming may make ex-FGR infants susceptible to impaired neurodevelopment and a state of chronic low-grade inflammation, associated with metabolic syndrome, diabetes mellitus type 2 and cardiovascular disease [42].
Our study has several strengths. For the first time, we investigated methylation in an SGA cohort in relation to the CPR. This study is an extension of a well-designed prospective cohort study for which numerous clinical parameters were meticulously recorded [25]. Categorizing the SGA group depending on CPR allowed us to investigate methylation changes depending on severity of FGR and brain-sparing instead of the association between SGA and methylation in general. We adopted a gene specific approach and analyzed methylation levels of six gene promoter areas based on existing literature. Although this is a limited selection in comparison with genome-wide methylation profiling, this gave us the advantage of focusing on biologically relevant genes.
We also recognize some limitations. First of all, because the placentas were embedded in paraffin, we were not able to connect our methylation data with quantitative gene expression outcomes. Other researchers, however, have previously studied the association between DNA methylation and gene expression. The VEGFA region studied by us is a CG-rich promotor region in close proximity to the transcription start site which is associated with gene repression [43]. Induction of targeted DNA methylation of the VEGFA promoter by Siddique et al. demonstrates that VEGFA expression is repressed by increasing amounts of DNA methylation [37].
The association between LEP methylation and expression has also been thoroughly studied [31]. Negative associations with gene expression have been found in the maternal and fetal side of the placenta [18,44]. In line with this, methylation of a larger LEP promoter region in which our 13 investigated CpGs are located was also correlated with gene expression in the white adipose tissue visceral adipocyte fraction [45]. This relationship was explored further in a functional experiment where DNA methylation was inhibited in cultured human primary skin fibroblasts and HeLa cells which resulted in increased LEP expression [45].
The negative associations described here fit with the theory that increased gene promotor DNA methylation of LEP and VEGFA results in decreased expression. We expect that this relationship also exists within our cohort and assume that this could have an effect on gene expression.
Second, we studied a small sample size and the variation in gestational age at birth was great. Additionally, with this study design we cannot conclude whether the observed methylation changes are adaptive or maladaptive in the fetal period or thereafter, but the presence of a difference suggests that altered programming occurs in the placenta of FGR fetuses with an abnormally reduced CPR antenatally. Whether the observed differences in DNA methylation are sufficient to alter gene expression enough to cause a physiological effect could be explored further in an animal model of fetal brain-sparing.
Conclusion
In conclusion, this study demonstrates that LEP and VEGFA are differentially methylated in SGA placentas, depending on the CPR. This suggests that the presence of FGR as defined by the CPR influences the epigenetic profile of the placenta differently. The increased risk of cardiovascular disease and metabolic syndrome are known consequences of FGR. Clearly, more research is needed to investigate the clinical significance of altered placental LEP, VEGFA, and other genes in relation to clinical care of growth-restricted children who will develop into adulthood. This study adds to the growing body of evidence regarding epigenetic programming in FGR. With increasing knowledge of the placenta and its epigenetic profile, the placenta can care for and protect the fetus throughout pregnancy and beyond birth by providing physicians with the methylation status at birth. This methylation profile might be used as a prognostic tool for targeted preventive interventions in the future.
Supporting information
S1 Table. Methylation of VEGFA and LEP per CPR group.
TSS: location relative to the Transcription Start Site, Values shown as median (range) or mean (SE), a Normal CPR n = 24, Abnormal CPR n = 17, b Normal CPR n = 20, Abnormal CPR n = 14, * significant at the .05 level (2-tailed), ** significant at the .01 level (2-tailed), c Student’s T-test, d Mann-Whitney U Test.
https://doi.org/10.1371/journal.pone.0221972.s001
(DOCX)
S2 Table. Spearman’s rho correlation coefficients between individual CpGs within EPO, HIF1A, VEGFA, LEP, PHLDA2 and DHCR24.
* Correlation is significant at the .05 level (2-tailed). ** Correlation is significant at the .01 level (2-tailed).
https://doi.org/10.1371/journal.pone.0221972.s002
(DOCX)
Acknowledgments
The authors would like to thank the women and infants who participated in this study.
References
- 1. Wladimiroff JW, Tonge HM, Stewart B, Reuss A. Severe intrauterine growth retardation; assessment of its origin from fetal arterial flow velocity waveforms. Eur J Obstet Gynecol Reprod Biol. 1986;22: 23–28. pmid:3522306
- 2. Broek AJM Van Den, Kok JH, Houtzager BA, Scherjon SA. Behavioural problems at the age of eleven years in preterm-born children with or without fetal brain sparing: A prospective cohort study. Early Hum Dev. 2010;86: 379–384. pmid:20554130
- 3. Beukers F, Aarnoudse-Moens CSH, van Weissenbruch MM, Ganzevoort W, van Goudoever JB, van Wassenaer-Leemhuis AG. Fetal Growth Restriction with Brain Sparing: Neurocognitive and Behavioral Outcomes at 12 Years of Age. J Pediatr. 2017;188: 103–109. pmid:28693788
- 4. Miller SL, Huppi PS, Mallard C. The consequences of fetal growth restriction on brain structure and neurodevelopmental outcome. J Physiol. 2016;594: 807–823. pmid:26607046
- 5. Koukoura O, Sifakis S, Spandidos DA. DNA methylation in the human placenta and fetal growth (review). Mol Med Rep. 2012;5: 883–889. pmid:22294146
- 6. Lea RG, Howe D, Hannah LT, Bonneau O, Hunter L, Hoggard N. Placental leptin in normal, diabetic and fetal growth-retarded pregnancies. Mol Hum Reprod. 2000;6: 763–769. pmid:10908288
- 7. Lazo-De-La-Vega-Monroy ML, González-Domínguez MI, Zaina S, Sabanero M, Daza-Benítez L, Malacara JM, et al. Leptin and its Receptors in Human Placenta of Small, Adequate, and Large for Gestational Age Newborns. Horm Metab Res. 2017;49: 350–358. pmid:28351089
- 8. Szentpéteri I, Rab A, Kornya L, Kovács P, Joó JG. Gene expression patterns of vascular endothelial growth factor (VEGF-A) in human placenta from pregnancies with intrauterine growth restriction. J Matern Fetal Neonatal Med. 2013;26: 984–9. pmid:23350655
- 9. Conrad KP, Benyo DF, Westerhausen-Larsen A, Miles TM. Expression of erythropoietin by the human placenta. FASEB J. 1996;10: 760–8. pmid:8635693
- 10. Zimna A, Kurpisz M. Hypoxia-Inducible factor-1 in physiological and pathophysiological angiogenesis: Applications and therapies. BioMed Research International. 2015. pmid:26146622
- 11. Varma S, Cohen HJ. Co-transactivation of the 3’ erythropoietin hypoxia inducible enhancer by the HIF-1 protein. Blood Cells Mol Dis. 1997;23: 169–176. pmid:9236155
- 12. Iriyama T, Wang W, Parchim NF, Song A, Blackwell SC, Sibai BM, et al. Hypoxia-Independent Upregulation of Placental Hypoxia Inducible Factor-1α Gene Expression Contributes to the Pathogenesis of Preeclampsia. Hypertension. 2015;65: 1307–1315. pmid:25847948
- 13. Sundrani DP, Reddy US, Joshi A a, Mehendale SS, Chavan-Gautam PM, Hardikar A A, et al. Differential placental methylation and expression of VEGF, FLT-1 and KDR genes in human term and preterm preeclampsia. Clin Epigenetics. 2013;5: 6. pmid:23621880
- 14. Ashworth C. Placental leptin. Rev Reprod. 2000;5: 18–24. pmid:10711732
- 15. Lesseur C, Armstrong DA, Paquette AG, Koestler DC, Padbury JF, Marsit CJ. Tissue-specific Leptin promoter DNA methylation is associated with maternal and infant perinatal factors. Mol Cell Endocrinol. Elsevier Ireland Ltd; 2013;381: 160–167. pmid:23911897
- 16. Hillman SL, Finer S, Smart MC, Mathews C, Lowe R, Rakyan VK, et al. Novel DNA methylation profiles associated with key gene regulation and transcription pathways in blood and placenta of growth-restricted neonates. Epigenetics. 2015;10: 50–61. pmid:25496377
- 17. McMinn J, Wei M, Schupf N, Cusmai J, Johnson EB, Smith AC, et al. Unbalanced Placental Expression of Imprinted Genes in Human Intrauterine Growth Restriction. Placenta. 2006;27: 540–549. pmid:16125225
- 18. Lesseur C, Armstrong DA, Murphy MA, Appleton AA, Koestler DC, Paquette AG, et al. Sex-specific associations between placental leptin promoter DNA methylation and infant neurobehavior. Psychoneuroendocrinology. 2014;40: 1–9. pmid:24485470
- 19. Ciesielski TH, Marsit CJ, Williams SM. Maternal psychiatric disease and epigenetic evidence suggest a common biology for poor fetal growth. BMC Pregnancy Childbirth. England; 2015;15: 192. pmid:26303856
- 20. Jensen AB, Tunster SJ, John RM. The significance of elevated placental PHLDA2 in human growth restricted pregnancies. Placenta. Elsevier Ltd; 2014;35: 528–532. pmid:24953163
- 21. Green BB, Armstrong DA, Lesseur C, Paquette AG, Guerin DJ, Kwan LE, et al. The Role of Placental 11-Beta Hydroxysteroid Dehydrogenase Type 1 and Type 2 Methylation on Gene Expression and Infant Birth Weight. Biol Reprod. 2015;92: 1–8. pmid:25788665
- 22. Janssen AB, Tunster SJ, Heazell AEP, John RM. Placental PHLDA2 expression is increased in cases of fetal growth restriction following reduced fetal movements. BMC Med Genet. BMC Medical Genetics; 2016;17: 17. pmid:26944942
- 23. Woollett LA. The origins and roles of cholesterol and fatty acids in the fetus. Curr Opin Lipidol. 2001;12: 305–12. pmid:11353334
- 24. Diplas AI, Lambertini L, Lee MJ, Sperling R, Lee YL, Wetmur J, et al. Differential expression of imprinted genes in normal and IUGR human placentas. Epigenetics. 2009;4: 235–240. pmid:19483473
- 25. Tanis JC, Boelen MR, Schmitz DM, Casarella L, van der Laan ME, Bos AF, et al. Correlation between Doppler flow patterns in growth-restricted fetuses and neonatal circulation. Ultrasound Obstet Gynecol. 2016;48: 210–216. pmid:26358663
- 26. Steinmann K, Richter AM, Dammann RH. Epigenetic silencing of erythropoietin in human cancers. Genes and Cancer. 2011;2: 65–73. pmid:21779481
- 27. Pierre CC, Longo J, Bassey-Archibong BI, Hallett RM, Milosavljevic S, Beatty L, et al. Methylation-dependent regulation of hypoxia inducible factor-1 alpha gene expression by the transcription factor Kaiso. Biochim Biophys Acta—Gene Regul Mech. Elsevier B.V.; 2015;1849: 1432–1441. pmid:26514431
- 28. Koslowski M, Luxemburger U, Türeci Ö, Sahin U. Tumor-associated CpG demethylation augments hypoxia-induced effects by positive autoregulation of HIF-1α. Oncogene. 2011;30: 876–882. pmid:21042279
- 29. Stöger R. In vivo methylation patterns of the leptin promoter in human and mouse. Epigenetics. 2006;1: 155–162. pmid:17965621
- 30. Krempler F, Breban D, Oberkofler H, Esterbauer H, Hell E, Paulweber B, et al. Leptin, peroxisome proliferator-activated receptor-gamma, and CCAAT/enhancer binding protein-alpha mRNA expression in adipose tissue of humans and their relation to cardiovascular risk factors. Arter Thromb Vasc Biol. 2000;20: 443–449.
- 31. Melzner I, Scott V, Dorsch K, Fischer P, Wabitsch M, Brüderlein S, et al. Leptin gene expression in human preadipocytes is switched on by maturation-induced demethylation of distinct CpGs in its proximal promoter. J Biol Chem. 2002;277: 45420–45427. pmid:12213831
- 32. Li Y, Huang Y, Lv Y, Meng G, Guo QN. Epigenetic regulation of the pro-apoptosis gene TSSC3 in human osteosarcoma cells. Biomed Pharmacother. 2014;68: 45–50. pmid:24268429
- 33. Drzewinska J, Walczak-Drzewiecka A, Ratajewski M. Identification and analysis of the promoter region of the human DHCR24 gene: Involvement of DNA methylation and histone acetylation. Mol Biol Rep. 2011;38: 1091–1101. pmid:20568014
- 34. Marinoni E, Letizia C, Ciardo F, Corona G, Moscarini M, Di Iorio R. Effects of prenatal betamethasone administration on leptin and adiponectin concentrations in maternal and fetal circulation. Am J Obstet Gynecol. 2008;199. pmid:18456235
- 35. Henson MC, Castracane VD. Leptin in Pregnancy. Biol Reprod. 2000;63: 1219–1228. pmid:11058523
- 36. Henson MC, Castracane VD. Leptin in Pregnancy: An Update. Biol Reprod. 2006;74: 218–229. pmid:16267210
- 37. Siddique AN, Nunna S, Rajavelu A, Zhang Y, Jurkowska RZ, Reinhardt R, et al. Targeted methylation and gene silencing of VEGF-A in human cells by using a designed Dnmt3a-Dnmt3L single-chain fusion protein with increased DNA methylation activity. J Mol Biol. Elsevier Ltd; 2013;425: 479–491. pmid:23220192
- 38. Krishnan T, David AL. Placenta-directed gene therapy for fetal growth restriction. Semin Fetal Neonatal Med. Elsevier Ltd; 2017;22: 415–422. pmid:28522033
- 39. Udagawa J, Hashimoto R, Suzuki H, Hatta T, Sotomaru Y, Hioki K. The Role of Leptin in the Development of the Cerebral. Endocrinology. 2007;147: 647–658. pmid:16282354
- 40. Voron T, Marcheteau E, Pernot S, Colussi O, Tartour E, Taieb J, et al. Control of the immune response by pro-angiogenic factors. Front Oncol. 2014;4: 1–9. pmid:24765614
- 41. Naylor C, Petri WA. Leptin regulation of immune responses. Trends Mol Med. Elsevier Ltd; 2016;22: 88–98. pmid:26776093
- 42. Minihane AM, Vinoy S, Russell WR, Baka A, Roche HM, Tuohy KM, et al. Low-grade inflammation, diet composition and health: Current research evidence and its translation. Br J Nutr. 2015;114: 999–1012. pmid:26228057
- 43. Jones PA. Functions of DNA methylation: Islands, start sites, gene bodies and beyond. Nat Rev Genet. Nature Publishing Group; 2012;13: 484–492. pmid:22641018
- 44. Bouchard L, Thibault S, Guay SP, Santure M, Monpetit A, St-Pierre J, et al. Leptin gene epigenetic adaptation to impaired glucose metabolism during pregnancy. Diabetes Care. 2010;33: 2436–2441. pmid:20724651
- 45. Marchi M, Lisi S, Curcio M, Barbuti S, Piaggi P, Ceccarini G, et al. Human leptin tissue distribution, but not weight loss-dependent change in expression, is associated with methylation of its promoter. Epigenetics. 2011; pmid:21931275