This study was carried out in the department of Biochemistry in collaboration with the Department of Obstetrics and Gynecology at the Postgraduate Institute of Medical Education and Research (PGIMER), Chandigarh, India. The protocol of the study was approved by the Institute Ethics Committee (IEC) and written informed consent was obtained from all participants.

Pregnant women with the clinical diagnosis of normal pregnancy or pregnancy complicated with preeclampsia attending the Department of Obstetrics and Gynecology were included in this study. Pregnant women with normal pregnancy were divided into three groups based on the gestational age (first, second and third trimester, n = 30 in each group), while the preeclampsia group (n = 30) was characterized by clinical symptoms of systolic pressure 140 mmHg and diastolic pressure 90 mmHg, proteinuria 300 mg in 24 hr. Placental villi samples were collected after elective pregnancy terminations from pregnant women with normal first and second-trimester pregnancy, or after cesarean section or vaginal deliveries in case of normal third-trimester and pre-eclamptic pregnancies. Placental tissue obtained from each subject was processed in chilled PBS for the separation of clear placental villi free from fibroid tissue, blood clots, amnion, and basement membrane.

Two adherent cell lines, JEG-3, and HTR-8/SVneo [Source: American Type Culture Collection (ATCC), kindly provided by Dr S.K. Gupta, National Institute of Immunology, New Delhi, India] was used in this study. JEG-3 is placental choriocarcinoma-derived cell line, while HTR-8/SVneo is a cell line derived from first trimester transformed cells and acts as a close mimic of extra-villous trophoblasts. JEG-3 and HTR-8/SVneo cells were cultured in DMEMHG, with 4,500 mg/L glucose and RPMI-1640 medium, respectively, supplemented with L-glutamine and 25 mM HEPES, 10% FBS, sodium bicarbonate (3.7 and 2 g/L, respectively), penicillin (100 U/ml), streptomycin (100 mg/ml) and amphotericin B (0.25 mg/ml) under standard conditions (37°C, 5% CO₂, humidified atmosphere). Cell culture media and other reagents were purchased from Sigma Chemical Co.

The folate levels within placental villi samples and cell lines were estimated using the Lactobacillus casei assay, as previously explained (24). Briefly, 10% homogenate of villous tissue was prepared in folate extraction buffer containing 50 mM ascorbic acid. The homogenate was incubated for 10 min at 110 °C, followed by centrifugation and the treatment of separated supernatant with rat plasma conjugase, which hydrolyzes the polyglutamated folate to monoglutamated folate form. The free folate is utilized by L. casei for its growth; hence the levels of folic acid levels are determined by measuring bacterial growth. In each assay, the standards of known folic acid concentration were run along with the samples.

The overall degree of methylation of a genome is a useful measure of global methylation changes. Global DNA methylation was measured using an ELISA based kit Imprint^R Methylated DNA Quantification Kit (Sigma-Aldrich, Inc) according to the manufacturer's instructions. Briefly, the methylated DNA is detected using the capture and detection antibodies and then quantified colorimetrically. The amount of methylated DNA present in the sample is proportional to the absorbance measured. To calculate methylation levels relative to the methylated control DNA, we used the single point method, which gives the per cent methylation relative to the methylated control DNA.

An invasion assay was performed to study the invasive property of JEG-3 and HTR-8/SVneo in vitro and the effect of folic acid treatment on the invasion of these cell lines. It was done by using the BD Matrigel™ Matrix-GFR (Growth Factor Reduced), which is a reconstituted basement membrane. The Matrigel matrix layer occludes the pores of the PET membrane, blocking non-invasive cells from migrating through the membrane. Exogenous folic acid (10⁻⁷ M and 10⁻⁴ M) was added to the culture media, which originally contained no folic acid. 10⁻⁷ M concentration was chosen because it represents folic acid levels in the physiological ranges and 10⁻⁴ M represents a much higher concentration than the physiological range (17). 10⁻⁴ M was chosen to compare the effect of higher folate levels with physiological levels. Cells were incubated for 48 h at 37°C, 5% CO₂, humidified atmosphere. The invasive cells were able to detach themselves from and invade through the Matrigel matrix and the 8-micron membrane pores. The number of cells invading the Matrigel were counted, after being stained with Hoechst 33342 nuclear-fluorescent stain under Olympus BX61 epi automated fluorescence microscope under 460–90 nm and 40× magnification. The fold change was calculated with respect to the untreated control and the statistical analysis was done by comparing the means of the control and experimental sets by using paired Student's t-test.

Total RNA was isolated from placental villi, the JEG-3 and HTR-8/SVneo cell lines using TRIzol (Ambion, Life Technologies Corporation, CA, USA), and 1 mg was reverse-transcribed using RevertAidTM M-MuLV-RT kit (MBI Fermentas, Life Sciences, USA). The resulting cDNA was quantified using Applied Biosystems StepOnePlusTM Real-Time PCR System. The mRNA expression levels of glyceraldehyde-3- phosphate dehydrogenase (GAPDH) were used as the endogenous control. The comparative Ct method (DDCT) (25) was used for the quantification of the transcripts. Measurement of DCt was performed in triplicate.

DNA was extracted for methylation analysis from the placental villi and cultured JEG-3 and HTR-8/SVneo cell lines by using the genomic DNA isolation kit (Real Genomics, Real Biotech Corporation, Taipei, Taiwan). Promoter region CpG methylation analysis of the target genes was carried out by methylation-sensitive high-resolution melting (MSHRM) as previously explained using gene-specific primers targeting promoter region CpG islands (3–6). Briefly, MS-HRM can differentiate sequences on the basis of the GC content, which determines the temperature at which the dsDNA sequence denatured (26), using a standard curve comprising 0, 0.5, 5, 10, 20, 40, 60, 80, and 100% methylation standards. The methylation percentage values for each target sample were calculated according to the method of Migheli et al. (27) by imputing the average aligned fluorescence value of each sample in the Polyfit interpolation curve (MatLab program-The MathWorks, Inc., USA) (27).

All statistical analyses were performed using the SPSS software for Windows version 16.0 and GraphPad Prism software version 5.00.288. Between-group comparisons were made using one-way ANOVA; if found to be significant, the Fisher post-hoc test was applied. Student's t-test and the Mann–Whitney U-test were used to analyzing the data between two variables. Pearson's correlation analysis was used to estimate the correlations between different parameters. The effect of different epigenetic mechanisms on mRNA expression levels was assessed by multiple regression analysis. All statistical tests were two-sided, and differences were considered statistically significant at p < 0.05. Unless otherwise stated, all data are expressed as a mean (SEM).

At the genomic level, global transcription is regulated by the overall degree of methylation of a genome, which includes CpG methylation both at intergenic levels as well as gene-specific promoter region methylation. Comparing the relative percentage of global DNA methylation in normal placentas of three different trimesters, we found global DNA methylation to increase with gestation, from 25.5% in the first-trimester to 27.99% (p < 0.05) in second- trimester and finally to 36.07% (p < 0.001) at full term. There was an increase in percentage global DNA methylation to 39.6%, p < 0.01, in preeclamptic placenta compared to their gestation age-matched control placentas as shown in Figure 2A.

Global methylation was also found to be significantly different in different placental cells, with the least global methylation in HTR-8/SVneo (15.57%, p < 0.001) compared to JEG-3 cells (Figure 2B).

LINE1 constitutes almost 17–25% of the human genome with up to 600,000 copies per genome. CpG sites in LINEs comprise a big portion of the human genome, these are largely methylated in normal somatic tissue. Since these repeats make up a considerable portion of the human genome, the degree of methylation in these repeats comprises a significant portion of global methylation (28). However, other repeat elements and gene-specific promoter methylation are also important in determining total genomic methylation. In our study, MS-HRM analysis for the LINE1 promoter region in the preeclamptic placenta showed a higher methylation by 7% (p < 0.05) in the LINE1 promoter as compared to the normal third trimester. However, MS-HRM detected a nonsignificant change in methylation in normal gestational placental groups (Figure 2C).

LINE1 promoter was detected to be highly methylated in HTR-8/SVneo and JEG-3cells as compared to normal first-trimester placental villi (Figure 2D), the average methylation percentage being 34 and 28 % higher, respectively (p < 0.001).

We estimated the levels of expression for DNMT1, DNMT3a, and DNMT3B in different pregnancy groups. The mRNA expression of DNMT1 was found to vary significantly within the different placental groups. In normal placental groups, there was a 2-fold increase in the second trimester relative to the first trimester, followed by a further 7.5-fold increased expression in the third- trimester relative to the second trimester (p < 0.001). Preeclamptic placenta revealed a significant increase of 2.4-fold (p < 0.001) expression of DNMT1 relative to the third-trimester placenta (Figure 3A). The expression of DNMT1 was found to be 28- to 30-fold higher in HTR-8/SVneo cells (p < 0.001) in comparison to JEG-3 cells and normal first-trimester villi, as shown in Figure 3B.

Normal gestational pregnancy was found to be accompanied by a consistent increase in the DNMT3A level in placental villi groups (Figure 3C). As observed by a 2-fold (p < 0.01) increase in the second-trimester and 4.8-fold (p < 0.001) increase in the third-trimester, relative to first-trimester group. The preeclampsia group showed a further increase of 2.3-fold (p < 0.001) relative to the normal third-trimester group. A comparison of the DNMT3A expression in placental derived cell lines and first trimester villi (Figure 3D) showed a 2-fold and 16.8-fold lower expression in JEG-3 and normal first trimester villi, relative to HTR-8/SVneo cells (p < 0.05 and p < 0.001, respectively).

On accessing the mRNA expression of DNMT3B in different placental groups, a decreased expression of 3.2 and 2.4-fold was observed in term placenta (p < 0.001), relative to first- and second-trimester placental villi groups respectively (Figure 3E). The relative DNMT3B expression was found to vary non-significantly between the two placental cell lines, however, both the cell lines showed higher DNMT3B mRNA levels in comparison to normal first trimester villi, with 3.03-fold (p < 0.01) and 5.3-fold (p < 0.001) increase in JEG-3 and HTR-8/SVneo cells, respectively (Figure 3F).

To find the relation between various molecular aspects of our study, we have used Pearson's correlation analysis. Pearson's correlation analysis is a statistical relationship, or association, between two continuous variables. It was used to compare the mean data between two particular sets of experiments among groups. In our analysis, we have used Pearson correlation coefficient “r” value to define the relation as, if r <0.49: low relation, r = 0.5 to 0.7: moderate relation and r = 0.7 to 1: high relation.

In this study, on Pearson correlation analysis, among placental villi groups (normal first-, second- and third- trimester placenta and preeclampsia), revealed a significant positive and negative correlation between mRNA expression of DNMT1 vs. DNMT3A (r = 0.99; p < 0.001) and vs DNMT3B (r = −0.77), respectively. Global methylation also showed a significantly higher positive correlation with mRNA expression level of DNMT1 (r = 0.9; p < 0.05) and DNMT3A (r = 0.87; p < 0.05), while it was negatively correlated to mRNA expression of DNMT3B (r = 0.83, p < 0.05). Pearson correlation analysis showed a moderate correlation between LINE1 promoter methylation and DNMT3B expression (r = 0.5). Within placental villi groups, almost no correlation was observed between LINE1 promoter methylation and global DNA methylation. The folate level with different parameters showed a moderate negative correlation with DNMT1, DNMT3A, global DNA methylation, and LINE1 promoter methylation, as shown in Table 1.

Folic acid is a key source of the one-carbon group required to methylate DNA. DNA methylation is an epigenetic modification critical to the normal development of the placenta and regulation of its pseuduomalignant/invasive nature. In this context, we intended to find the role of folic acid in regulating the pseudomalignant/invasive nature of the placenta. Therefore, the effect of folic acid on the invasive potential of various trophoblasts was studied by supplementing folic acid to HTR-8/SVneo and JEG-3 trophoblast cell lines at two different concentrations i.e., 10⁻⁷M and 10⁻⁴M using matrigel based invasion assay. Among these two concentrations, 10⁻⁷M represents the concentrations in the physiological range (400–600 μg/day) and 10⁻⁴M represents a much higher concentration than the physiological range. The cells were treated for 48 h and allowed to invade across the Matrigel matrix coated PET membrane of cell culture inserts.

In vitro invasion assay showed a dose-dependent effect of folic acid supplementation on the invasiveness of both cell lines. In the JEG-3 cell line, there was a significant increase (p < 0.01) in invasion with 2.24-fold higher invasive potential at 10⁻⁷ M folic acid treatment with respect to untreated cells while treatment of folic acid at a higher concentration of 10⁻⁴M resulted in a significant (p < 0.01) decrease in invasion by 2.08-fold. Thus, showing a more significant difference in invasiveness at two different concentrations of folic acids (p < 0.001) (Figures 4A,B).

Studying the effect of folic acid treatment in the HTR-8/SVneo cell line revealed that the intrinsic invasive potential of HTR-8/SVneo was much higher (6.4-fold, p < 0.001) than the JEG-3 cell line. At 10⁻⁷M folic acid supplementation, there was a 2.4-fold (p < 0.001) increase in the invasive potential of the HTR-8/SVneo cell line with respect to control cells. There was a decrease in the invasion at 10⁻⁴M folic acid treatment by around 2-fold (p < 0.01) as compared to control (Figures 4C,D).

Folic acid levels were estimated in JEG-3 and HTR-8/SVneo cell lines and cytotrophoblasts with and without folic acid supplementation. It was estimated by a microbiological assay using Lactobacillus casei. Folic acid levels are expressed as mean ±SD in each group.

Folic acid levels were found to be equal in JEG-3 and HTR-8/SVneo cell lines before the supplementation of folic acid (Figure 5A). The folic acid supplementation resulted in a dose-dependent increase in folic acid levels in both JEG-3 and HTR-8/SVneo cells. At 10⁻⁴M concentration, the levels of folic acid in JEG-3 cells were 0.02 μg/10⁶ cells (p < 0.001), while that in HTR-8/SVneo cells, they were 0.019 μg/10⁶ (p < 0.001), as compared to control cells (0.014 μg/10⁶ cells).

Folic acid is a known mediator in DNA methylation. To elucidate the molecular mechanism of the changes induced in invasive potential of trophoblasts and regulation of pseudo-malignant nature of placenta, by folate supplementation, we analyzed the effect of folic acid supplementation on DNA methylation at the gene promoter level, global methylation, and LINE1 methylation levels.

Further to study the transcriptional profile of various genes under study, upon folic acid supplementation, we quantified the mRNA levels of various studied genes in JEG-3 and HTR-8/SVneo cell lines after giving folic acid treatment at 10⁻⁷ and 10⁻⁴ M.

Folic acid treatment was found to significantly decrease the mRNA levels of studied tumor suppressor genes, RASSF1A, RB1, PRKCDBP, and APC either in one or both cell lines in a dose-dependent manner. However, folic acid was found to induce a reverse trend in mRNA levels in JEG-3 and HTR-8/SVneo cells in the case of P16 as shown in Figures 6A–E.

Folate was found to act as an inducer of studied oncogenes and hTERT expression. The expression of c-jun, c-myc, EGFR, and hTERT increased in a dose-dependent manner in both cell lines, while only HTR-8/SVneo cells showed such dose-dependent increased expression for VEGF (Figures 6F–J).

Folic acid supplementation altered the expression levels of MMPs and their inhibitors in a reverse trend in both cell lines. The expression of MMP2 and MMP9 increased, while that of TIMP1 and TIMP2 decreased in a dose-dependent manner in JEG-3 and HTR-8/SVneo cells (Figures 6K–N).

DNMT1 mRNA level in placental cell lines showed a significant decrease upon the folic acid supplementation (Figure 7A). The decrease in DNMT1 expression at 10⁻⁷M folic acid treatment was 2.3- and 5-fold in JEG-3 and HTR-8/SVneo cells, relative to their respective control cells. It was further reduced by 11.8-and 7.8-fold at 10⁻⁴M folic acid treatment, with respect to their respective 10⁻⁷M folic acid-treated cells.

Similar to that observed for DNMT1, folic acid inhibited the DNMT3A mRNA expression in cells by 4.2-and 9.4-fold (p < 0.001) in JEG-3 with respect to untreated control cells at 10⁻⁷M and 10⁻⁴ folic acid. In HTR-8/SVneo cells there was an almost equal decrease in DNMT3A relative expression (1.5-fold) at both the folic acid concentrations used (Figure 7B).

DNMT3B mRNA expression was found to be increased in both cell lines due to folic acid supplementation. In JEG-3 cells it increased by 1.3-fold (p < 0.01) and 2-fold (p < 0.001) at 10⁻⁷ M and 10⁻⁴ M folic acid concentration while in HTR-8/SVneo cells it increased by almost 1.6-fold (p < 0.001), with folic acid treatment, relative to their untreated control cells (Figure 7C).

On Pearson correlation analysis, between various DNMTs, folate level, and global methylation level and LINE1 promoter methylation, in JEG-3 cells upon folic acid supplementation (Table 2), a strong negative correlation was observed between folate level vs DNMT1and 3A (r > 0.78), while it was positive with DNMT3B mRNA expression (r = 0.99, p < 0.05). Folate vs LINE1 promoter methylation was positively correlated (r = 0.86), while it was negatively correlated with global methylation (r = −0.87). A highly significant negative correlation was observed between global methylation and LINE1 promoter methylation in JEG-3 cells upon folic acid supplementation (r = −1, p < 0.01), global methylation also showed a significant positive correlation with DNMT3A (r = 0.99, p < 0.05) (Table 2).

In the case of HTR-8/SVneo, upon folic acid supplementation (Table 3), the Pearson correlation analysis between DNMTs mRNA expression, folate level, global methylation, and LINE1 promoter methylation showed a higher negative correlation in the case of folate level vs. expression of DNMT1 and DNMT3A (r = −0.77 and −0.83, respectively), while it was positive in case of DNMT3B (r = 0.98). However, global methylation was significantly correlated with mRNA expression of DNMT1 and DNMT3A (r = 0.99 and 1, respectively, p < 0.05). Furthermore, a significant positive correlation was observed between mRNA expression of DNMT1 vs DNMT3A (r = 0.99, p < 0.05) (Table 3).

The Pearson correlation analysis was also done between DNMTs mRNA expression, folate level, and global methylation level, separately in each group included in this study, which showed a significant correlation between DNMT3A vs. DNMT3B in first trimester placental villi (r = 0.811; p < 0.05), maternal blood leukocyte groups, viz, first trimester (r = 0.78; p < 0.05), second trimester (r = 0.8; p < 0.05), third trimester (r = 0.921, p < 0.01) and in JEG-3 cells (r = 1; p < 0.001). Furthermore, significant negative correlation was observed between folate level and global methylation in JEG-3 cells (r = −0.87; p < 0.05) and HTR-8/SVneo cells (r = −0.79; p < 0.05). Folate level vs. global methylation also showed higher Pearson coefficient value in preeclamptic placental group (r = 0.93) and in all maternal blood samples (r > 0.9), except that in first trimester.