Promoter methylation changes and vascular dysfunction in pre-eclamptic umbilical vein

Background Hypertension is one of primary clinical presentations of pre-eclampsia. The occurrence and progress of hypertension are closely related to vascular dysfunction. However, information is limited regarding the pathological changes of vascular functions in pre-eclamptic fetuses. Human umbilical cord vein was used to investigate the influence of pre-eclampsia on fetal blood vessels in this study. Results The present study found that the vasoconstriction responses to arginine vasopressin (AVP) and oxytocin (OXT) were attenuated in the pre-eclamptic umbilical vein as compared to in normal pregnancy, which was related to the downregulated AVP receptor 1a (AVPR1a), OXT receptor (OXTR), and protein kinase C isoform β (PKCβ), owing to the deactivated gene transcription, respectively. The deactivated AVPR1a, OXTR, and PKCB gene transcription were respectively linked with an increased DNA methylation within the gene promoter. Conclusions To the best of our knowledge, this study first revealed that a hyper-methylation in gene promoter, leading to relatively reduced patterns of AVPR1a, OXTR, and PKCB expressions, which was responsible for the decreased sensitivity to AVP and OXT in the umbilical vein under conditions of pre-eclampsia. The data offered new and important information for further understanding the pathological features caused by pre-eclampsia in the fetal vascular system, as well as roles of epigenetic-mediated gene expression in umbilical vascular dysfunction.


Background
Pre-eclampsia (PE) is a leading cause of maternal morbidity, mortality, and premature birth in both developed and developing countries [1,2]. Although PE in women is a multi-systemic syndrome with unknown etiology, hypertension is a primary clinical presentation of PE. As a surrogate end point for vascular risk, vascular dysfunction is closely related to the occurrence and progress of hypertension. However, information regarding the pathological changes of vascular functions in pre-eclamptic fetuses is limited. The umbilical cord is a conduit between the developing fetus and placenta. Umbilical cord vessels are primary vascular structures that may reflect problems originated from inadequate changes in maternal and fetal vascular systems. Human umbilical cord normally contains two arteries and one vein. The umbilical vein supplies the fetus with oxygenated, nutrient-rich blood from the placenta. Vascular functions of the umbilical vein are so important for placental-fetal circulation and fetal development in utero. Therefore, the present study was conducted with umbilical veins from healthy and preeclamptic pregnancy to investigate whether and how vascular functions would be affected under conditions of PE.
Because umbilical vessels have no autonomic innervation [3,4], circulating and locally synthesized vasoactive substances are important in controlling vascular functions and blood flow in the placental-fetal circulation. As stress hormones, arginine vasopressin (AVP) and oxytocin (OXT) are mainly synthesized in the magnocellular neurons of the paraventricular and supraoptic nucleus of the hypothalamus. In most vascular beds, AVP and OXT are potent vasoconstrictors [5]. AVP has long been implicated in controlling blood pressure and vascular tone through binding of smooth muscle receptors (mainly classified into V1a (AVPR1a), V1b (AVPR1b), and V2 (AVPR2) subtypes) [6][7][8]. In normal delivery, high AVP concentrations in human umbilical cord blood have been reported [9,10]. Similarly, oxytocin (OXT), a nine amino acid neuropeptide, is also increased at late pregnancy and onset of labor [11,12]. The actions of both central and peripheral OXT are mediated through oxytocin receptor (OXTR) [13]. It has been reported that AVP-and OXT-induced vasocontractions are mainly regulated by protein kinase C (PKC) pathway [6,14,15].
In humans, high AVP and OXT concentrations are demonstrated in umbilical cord blood during normal delivery [9,10,12]. Do the high AVP and OXT in the circulation cause remarkable vasoconstrictions in umbilical vessels? Would AVP and OXT play the same physiological roles in pre-eclamptic umbilical vessels as they do in the normal ones? In fact, such is the paucity of knowledge of vascular reactivities of the umbilical vein, with very limited studies and information on umbilical vascular functions and none has compared umbilical vascular responses of AVP and OXT between PE and normal pregnancy. The present study, therefore, investigated the contractile responses of AVP and OXT in normal and pre-eclamptic umbilical vein, to reveal special features of umbilical vascular regulations and possible pathophysiological changes, as well as its underlying mechanisms under PE condition. The data gained in the present study provided new and critical information on regulations of umbilical vascular functions under preeclamptic conditions that in favor of further understanding the pathological features and mechanisms of PE as well as vascular diseases in fetal origins.

AVP or OXT-induced contractions in human umbilical vein
Both AVP and OXT could induce dose-dependent constrictions in human umbilical vein (HUV) (Fig. 1a, d). There were no significant differences in KCl-induced maximal contraction between NP and PE group (Fig. 1b,  e), whereas, the Emax (AVP-or OXT-induced contraction at 10 −4 mol/L) and pD2 (−log[50% effective concentration]) values for AVP and OXT were significantly decreased in pre-eclamptic umbilical vein (Fig. 1c, f ). These data indicated that pre-eclamptic umbilical vein was significantly insensitive to AVP and OXT.

Expression of AVP or OXT receptors in human umbilical vein
In the vasculature, AVP receptors include AVPR1a, AVPR1b, and AVPR2 [7]. Compared with NP, mRNA and protein levels of AVPR1a, not AVPR2, were decreased in the PE group (Fig. 2a, b). SR49059 (AVPR1aspecific antagonist) completely blocked AVP-mediated contractions in both NP and PE groups and without significant differences in AVP-induced vasoconstrictions between the two groups after pretreatment with SR49059 (Fig. 2c). Similarly, as shown in Fig. 2d and e, there was a significant decrease in mRNA and protein of OXTR in the PE group. Meanwhile, OXTR-specific antagonist (atosiban) could completely block OXTmediated contractions in the umbilical vein, without significant differences between NP and PE groups after pretreated with atosiban (Fig. 2f ). These data indicated that the decreased sensitivity of pre-eclamptic umbilical vein to AVP and OXT was related to the downregulated AVPR1a and OXTR due to the deactivated gene transcription, respectively.
The decreased sensitivity of AVP and OXT was also dependent on PKC pathway AVP-and OXT-induced vasocontractions are mainly regulated by PKC pathway [6,14,15]. As shown in Fig. 3a, PKC agonist (PDBu) caused weaker dosedependent contractions in pre-eclamptic HUV than that of NP group. In the vasculature, PKC mainly includes α, β, γ, δ, and ε isoforms [16]. There were no significant differences in PKCα, PKCγ, PKCε, and PKCδ mRNA expression between NP and PE group; however, mRNA levels of PKCβ were significantly decreased in PE compared with that in NP group (Fig. 3b). Protein levels of PKCβ were also significantly decreased in pre-eclamptic HUV (Fig. 3c). Meanwhile, PKC-specific antagonist (GF109203X) could restrain AVP-or OXT-induced vasoconstrictions in both NP and PE groups, without significant differences in AVP-or OXT-mediated vasoconstrictions between NP and PE group following pretreatment with GF109203X (Fig. 3d, e). Meanwhile, GF109203X could produce a weaker inhibitory effect on AVP-or OXT-mediated vasoconstrictions in NP group (Fig. 3d, e). These data indicated that the decreased sensitivity of pre-eclamptic umbilical vein to AVP and OXT was also related to the downregulated PKC pathway.

DNA methylation of CpG locus within AVPR1a gene promoter in human umbilical vein
AVPR1a is located on chromosome 12q14.2. To clarify whether the deactivated transcription of AVPR1a was associated with DNA methylation alterations, we assessed changes of AVPR1a transcription after adding 5-Aza-2′deoxycytidine (5-Aza, a specific DNA methylation transferase inhibitor) in human umbilical cord vein endothelial cells (HUVECs). In HUVECs, 5-Aza treatment significantly increased AVPR1a gene transcription (  Table 1 showed CpG labels. Next, we validated methylation levels of these 14 CpG sites by targeted bisulfite sequencing. The bisulfite conversion rate of each sample was higher than 99%, and no significant difference was observed between NP and PE group, indicating bisulfite conversion was efficient and reliable in the experiments (Fig. 4c). Compared with NP, the mean methylation percentage of these 14 CpG sites in pre-eclamptic umbilical vein was significantly increased with specific CpG site 5 and 6 ( Fig. 4de, Table 2). Correlation analysis between AVPR1a gene methylation and expression was also conducted. There was a significantly inverse correlation between the methylation statuses of CpG sites (5 and 6) in AVPR1a gene promoter and AVPR1a gene expression (Fig. 4f). In normal and pre-eclamptic HUVECs, after 5-Aza treatment, mRNA levels of AVPR1a were significantly increased and without significant differences between the two groups (Fig. 4g).
DNA methylation of CpG locus within OXTR gene promoter in human umbilical vein OXTR is located on chromosome 3p25. Sequence analysis identified one CpG island that contains 22 CpG sites within exons of the OXTR gene (Fig. 5a). Table 3 provided a key for CpG labels. In HUVECs, after 5-Aza treatment, OXTR mRNA level was significantly increased (Fig. 5b). Compared with NP, mean methylation percentage of the total 22 CpG sites in the PE group was remarkably increased (Fig. 5d), whereas, no significant differences were observed in each tested CpG site between NP and PE group (Fig. 5c, d). Table 4 showed the position and methylation levels of these 22 CpG sites. As shown in Fig. 5e, there was a significantly inverse correlation between the methylation status of 22 CpG sites in OXTR gene promoter and OXTR gene expression. In normal and pre-eclamptic HUVECs, mRNA levels of OXTR were significantly increased, and no significant differences were observed between the two groups after 5-Aza treatment (Fig. 5g). PKCΒ gene transcription in HUVECs (Fig. 6b). Targeted bisulfite sequencing showed that compared with NP, the mean methylation percentage of the total 44 CpG sites was remarkably increased with specific CpG sites (38)(39)(40)(41) within PKCΒ gene promoter in the PE group, whereas no significant difference was observed in other specific CpG sites between NP and PE group ( Fig. 6c-e). Position and methylation levels of the 44 CpG sites were listed in Table 5. After careful analysis of DNA methylation and expression data, it is concluded that there was . f Expression analysis of AVPR1a gene and its correlation with methylation levels of CpG sites (5 and 6). DNA methylation/mRNA correlation plots for AVPR1a gene identified by causal inference test (N = 30 each group). The Y-axis represents the relative expression level of AVPR1a gene which was detected with qRT-PCR method. The X-axis represents the relative mean methylation level of the CpG sites (5 and 6) at AVPR1a gene promoter in HUV. r: Pearson correlation coefficient. g mRNA levels of AVPR1a gene in normal and pre-eclamptic HUVECs after 5-Aza treatment (N = 8 each group). Error bars denote SEM. *P < 0.05; **P < 0.01; ***P < 0.001. N number of participants also a significantly inverse correlation between the methylation statuses of 38-41 CpG sites and PKCΒ expression (Fig. 6f ). After 5-Aza treatment, mRNA levels of PKCΒ were significantly increased, and no significant differences were observed between normal and preeclamptic HUVECs (Fig. 6g).

Discussion
This present study found a special feature of AVP-and OXT-mediated vascular contractions in pre-eclamptic umbilical vasculature. The main findings are as follows: (1) Compared with the normal control, the vasoconstriction responses to AVP and OXT were attenuated in preeclamptic umbilical vein, which was related to the downregulated AVPR1a, OXTR, and PKCB, owing to the deactivated gene transcription, respectively. (2) The deactivated AVPR1a, OXTR, and PKCB transcriptions were respectively linked with an increased DNA methylation within gene promoter. The data gained not only offered new information for further understanding the pathological features and mechanisms of pre-eclamptic umbilical cords, but also providing novel clues for roles of epigenetic-mediated gene expression in fetal vascular dysfunction.
Although it is well known that AVP and OXT can produce vascular contractions in adults [5,17], data regarding their functional effects on fetal blood vessels is limited. In human, umbilical vessels are only healthy fetal blood vessels that can be obtained ethically in medical studies. Both of AVP and OXT exhibited significant dose-dependent vasoconstrictions in human fetal umbilical vein, suggesting that the two peptides are critically involved in the regulating umbilical vascular tone and circulation. Notably, we found that compared with normal pregnancy, pre-eclamptic umbilical vein was significantly insensitive to AVP and OXT, which was not only associated with their respective receptors, but also correlated with PKC pathway. This finding was supported by the following data: (1) The mRNA and protein levels of AVPR1a and OXTR were remarkably decreased in preeclamptic umbilical vein. Meanwhile, AVPR1a-or OXTR-specific antagonist could completely block AVPand OXT-mediated contractions in umbilical vein, respectively. (2) PKC agonist caused weaker dosedependent contractions, and PKC antagonist produced a weaker inhibitory effect on AVP-and OXT-mediated vasoconstrictions in pre-eclamptic umbilical vein; furthermore, mRNA and protein levels of PKCβ were significantly decreased in pre-eclamptic umbilical vein. These data above indicated that the decreased sensitivity of pre-eclamptic umbilical vein to AVP and OXT was related to the downregulated AVPR1a, OXTR, and PKCB, particularly with their deactivated gene transcription.
The number of studies in humans and laboratory animals indicated promoter DNA methylation levels are important for transcriptional regulation of AVPR1a [18][19][20], OXTR [21][22][23], and PKCB [24][25][26][27]. In exploring the possible underlying mechanisms of the altered AVP-and OXT-mediated vascular functions, the present study also focused on epigenetic causes. Firstly, to clarify whether the deactivated transcriptions of AVPR1a, OXTR, and PKCB are owing to DNA methylation, we assessed the changes of these gene transcriptions after adding 5-Aza in HUVECs. 5-Aza treatment significantly increased AVPR1a, OXTR, and PKCB gene transcriptions,   PKCB gene promoter were obviously increased in umbilical vein under conditions of PE. Thirdly, we conducted correlation analysis between gene methylation and expression and found that there was a significantly inverse correlation between DNA methylation levels of gene promoter and gene transcription. Fourthly, we isolated and cultured HUVECs in vitro and evaluated expressions of these genes in both of normal and preeclamptic HUVECs after 5-Aza treatment. Compared with normal, mRNA levels of these genes were decreased in the pre-eclamptic HUVECs. After 5-Aza treatment, mRNA levels of these genes were significantly increased in both of normal and pre-eclamptic HUVECs, and no significant differences were observed in mRNA levels of these genes between the two groups. Together, the present study first indicated that transcriptions of AVPR1a, OXTR, and PKCB were regulated by DNA methylation in human umbilical vessels and revealed that hyper-methylation in AVPR1a, OXTR, and PKCB gene promoter, leading to a relatively low pattern of gene expressions, were responsible for the decreased sensitivity of AVP and OXT in pre-eclamptic umbilical vessels. Large amount of research showed that DNA methylation has been considered for contribution to the development for PE [28]. In the present study, we first demonstrated that DNA methylation-mediated gene    expression was also critically involved in the pathogenesis of vascular dysfunction in pre-eclamptic umbilical vasculature. The pathological and clinical importance of DNA methylation in pre-eclamptic vascular dysfunction deserves further investigation. Significance of our findings is also closely linked to "the development of chronic diseases in fetal origins." According to this theory, prenatal adverse factors have been demonstrated as major causes to induce diseases after birth [29][30][31]. PE could act as a stress for development fetuses. Thirty percent of newborns born from pre-eclamptic women experience some forms of adverse prenatal outcome, including prematurity and intrauterine growth retardation [1,2]. PE is a long-term disease risk factor for the mother and possibly the offspring too [32][33][34]. Evidence from clinical studies has proposed that children born from pre-eclamptic women have a higher risk of suffering neurological, psychological, and behavioral alterations, particularly cardiovascular diseases, including hypertension and stroke, compared to children born from normal pregnancies [32,33,[35][36][37][38][39][40]. However, to date, the mechanisms behind these vascular outcomes are poorly understood. In human, the umbilical cord is physiologically and genetically part of the fetus and may reflect problems originated from inadequate changes in the fetus with maternal history of PE. Interestingly, this study found that pre-eclamptic fetal umbilical vein showed a specific epigenetic-mediated vascular dysfunction, suggesting that pre-eclamptic fetal vascular system may undergo similar changes as it is represented in fetal umbilical vessels. It is rational that there may exist the same or similar abnormalities in preeclamptic fetal vascular systems as that observed in the fetal umbilical cord vein. Given this, due to epigenetic code that can be inherited, it put forward the hypothesis that the child with maternal history of PE are with a higher risk of diseases and disorders particularly in vascular problems. Recent cohort studies assessing whether maternal PE are associated with vascular problems in the offspring throughout childhood and early adolescence have provided supportive evidence for this concept [39][40][41][42]. Although this study did not investigate the offspring, the interesting finding in epigenetic-mediated umbilical vascular dysfunctions provides new important information for further studies on cardiovascular diseases in fetal origins.
In conclusion, this study firstly revealed a special feature of vascular regulations and pathophysiological changes in the umbilical vein under conditions of PE. Significances of the finding includes (1) offering new information for further understanding the pathological features of pre-eclamptic fetal umbilical vessels, and (2) underlining roles of epigenetic-mediated gene expression in pre-eclamptic umbilical vascular dysfunction, and (3) providing new insights into the underlying mechanisms of PE-related higher risks of vascular diseases and disorders in fetal origins. In addition, it is well known that an altered placenta-umbilical cord circulation plays an important role in the development of PE [2,43]. Whether and how the changed umbilical vascular dysfunction contributes or complicates to PE is another interesting and important direction for researching.  Fig. 6 DNA methylation of CpG locus at PKCΒ gene promoter in human umbilical vein. a One CpG island that contains 44 CpG sites, located at positions + 27 to + 321 from TSS at PKCΒ gene promoter. b mRNA levels of PKCΒ in HUVECs after treatment with 5-Aza. c-e The mean methylation status of CpG locus (the total and each tested) at PKCΒ gene promoter in HUV (N = 30 each group). f DNA methylation/mRNA correlation plots for PKCΒ gene identified by causal inference test (N = 30 each group). g mRNA levels of OXTR gene in normal and pre-eclamptic HUVECs after 5-Aza treatment (N = 8 each group). Error bars denote SEM.*P < 0.05; **P < 0.01; ***P < 0.001. N number of participants Table 5 The methylation status of CpG locus (the total and each tested) at PKCΒ gene promoter. The data was expressed as mean ± SEM. PE pre-eclampsia, NP normal pregnant. *P < 0.05; **P < 0.01; ***P < 0.001 given informed consent. Healthy pregnant participants were defined as blood pressure < 120/90 mmHg and no clinically significant complications. Pre-eclamptic pregnant participants were defined as blood pressure > 140/ 90 mmHg and significant proteinuria after the 20th weeks of pregnancy [1,44]. Women with essential hypertension or medical complications, such as diabetes and renal and cardiovascular diseases, were excluded from the study. The clinical characteristics of all participants were detailed in Table 6.

Measurement of vascular tension
Human umbilical cords were immediately acquired from normal and pre-eclamptic pregnant women after vaginal delivery and delivered within 1 h. The umbilical cords (10 cm in length) were kept in iced Krebs solution (containing in mmol/L NaCl 119, NaHCO 3  10 μmol/L), or GF109203X (PKC antagonist, 100 μmol/ L) were used for pretreating HUV rings for 30 min before application of AVP or OXT, and vessel responses were recorded [45]. All drugs were freshly prepared and purchased from Sigma-Aldrich (St. Louis, MO).

Quantitative real-time PCR (qRT-PCR) and Western blot analysis
Total RNA was isolated from HUV using Trizol reagent and was reversed transcribed using first-strand cDNA Synthesis Kit (Invitrogen). qRT-PCR was performed using SYBR Green Supermix Taq Kit (Takara Biotechnology Co., Ltd., Dalian, China) and analyzed on an iQ5 Real-Time PCR Detection System (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The primer sequences are listed in Table 7. ΔΔCt method was used to comparatively quantify the amount of mRNA levels. The protein abundance of AVPR1a, AVPR2, OXTR, and PKC (α and β) in HUV was measured with Western blot normalized to β-actin. The primary antibodies were the rabbit polyclonal antibody (Santa Cruz Biotechnology) against AVPR1a, AVPR2, OXTR, or PKCβ (all 1:1000). The secondary antibody was the goat anti-rabbit antibody (1:500; Beyotime Biotechnology, Jiangsu, China). Immuno-signals were visualized using UVP imaging system (EC3-Imaging-System, Upland, CA, USA). Imaging signals were calculated and analyzed, and then the ratio of band brightness to β-actin was acquired to measure relative protein expression levels as previously described [46,47].

DNA isolation and targeted bisulfite sequencing assay
To prepare genomic DNA, HUV rings were lysed with lysis buffer containing (10 mM Tris-Cl (pH 7.5), 10 mM NaCl, 10 mM EDTA, 0.5% sarcosyl, and 1 mg/mL proteinase K) and incubated overnight at 60°C. Genomic DNA was extracted from lysates by standard phenol/ chloroform technique and subjected to bisulfite conversion using EpiTect bisulfite kit (Qiagen) according to the manufacturer's protocols. DNA was quantified and then diluted to a working concentration of 20 ng/μL for Bi-Sulfite Amplicon Sequencing (BSAS) [48]. CpG islands located in the proximal promoter of AVPR1a, OXTR, and PKCB were selected according to the following criteria: (1) ≥ 200 bp length, (2) ≥ 50% GC content, (3) ≥ 60% ratio of observed/expected dinucleotides CpG. Based on the genomic coordinates of the candidate CpG sites (Table 3, 4,8), we carefully designed the BSAS primers in order to detect them in a panel (Table 7). After PCR amplification, products were sequenced by Illumina Hiseq 2000. Methylation level at each tested CpG site was calculated as the percentage of the methylated cytosines over the total tested cytosines. The average methylation levels were calculated using methylation

Isolation and culture HUVECs
Umbilical cords (about 15 cm in length) were excised from the placenta immediately after delivery and placed into cold sterile phosphate-buffered saline. Endothelial cells were isolated from umbilical veins as described previously [46]. HUVECs were cultured in DMEM containing 20% fetal bovine serum at 37°C with 5% CO 2 and 95% air humidified incubator. Cultures were passaged every 2-3 days and used in experiments 2 passages. In 5-Aza treatment studies, HUVECs were seeded and allowed to grow for 2 days with or without 5-Aza (Sigma-Aldrich, 10 −6 mol/L) and then mRNA were extracted for experiments.

Data analysis and statistics
All data were expressed as the mean ± SEM. Significance (P < 0.05) was ascertained by t test or two-way analysis of variance (ANOVA) followed by the Bonferroni test. Concentration-dependent response curves were performed by computer-assisted nonlinear regression. DNA methylation/mRNA correlation plots were identified by causal inference test (Graph Pad Prism software CA, USA).