All patients provided written informed consent, and this study was approved by the Human Research Ethics Committee of the Shanghai Changzheng Hospital at the Second Military Medical University. Thirty PE patients (mean age of 33.20 ± 0.46 years) and 30 control patients (mean age of 32.60 ± 0.47 years) who received a cesarean section were included in this study. Preeclampsia was clinically diagnosed before cesarean. The main criteria for PE diagnosis were as follows: systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg on two occasions at least 4 h apar, and proteinuria >0.3 g/24 h. The control pregnancy was defined as no medical complications and proteinuria, and maternal blood pressure <140/90 mm Hg. Clinical characteristics and detailed information of all patients enrolled in this study are summarized in Table 1. The original demographic data of enrolled patients can be found in Supplementary Table 1. The placenta specimen was resected from the middle of villous lobule, avoiding any visible blood clot or calcification. To minimize blood contamination, each piece of tissue was intensively washed with ice-cold phosphate-buffered saline and then immediately snap-frozen in liquid nitrogen after resection. Of the 30 pairs of placenta samples, four pairs were randomly selected for lncRNA and mRNA microarray analysis, and the other samples were used for the following verification.

Total RNA was extracted from snap-frozen placenta tissues obtained from PE placenta patients and control patients using TRIzol Reagent (Invitrogen, Carlsbad, CA, United States) according to the manufacturer’s instruction. The RNA integrity and concentration were evaluated using the NanoDrop ND-1000 spectrophotometer and 2100 RNA Nano 6000 Assay Kit (Agilent Technologies, Santa Clara, CA, United States).

RNA with RNA integrity number greater than 6.5 purified from total RNA following the removal of rRNA was amplified and transcribed into fluorescent cDNA along the entire length of the transcripts without 3′ bias utilizing random priming method, and cDNA was labeled and hybridized to the Human LncRNA Array V2.0 (8 × 60, Arraystar). In addition, 30,215 coding transcripts were detected using the microarray as well. The microarray was performed by KangCheng Bio-tech (Shanghai, China). The arrays were scanned using the Agilent Scanner G2505B (Agilent Technologies), and the acquired array images were analyzed using Agilent Feature Extraction software (version 11.5.1; Agilent Technologies). Quantile normalization and subsequent data processing were performed using GeneSpring GX v11.5.1 software package (Agilent Technologies).

To investigate the roles and associated pathways of differentially expressed lncRNAs and mRNAs, Gene Ontology (GO) pathway analysis was performed to annotate the transcripts with terms under the biological process, cellular component, and molecular function ontologies. GO annotations of microarray genes were downloaded from NCBI¹, UniProt², and the GO³. A Fisher exact test was performed in order to locate the significant enrichment pathway. The resulting P-values were adjusted using the Benjamini–Hochberg false discovery rate (BH FDR) algorithm. Pathway categories with a FDR < 0.05 were reported.

Total RNA (2 μg) was reverse transcribed with random hexamers using the Reverse Transcription System Kit (Promega Corporation, Madison, WI). LncRNA expression in placenta tissues was quantified using a standard quantitative reverse transcription real-time PCR (qRT-PCR) on the StepOne Plus system (Applied Biosystems, Foster City, CA, United States), in a total reaction volume of 20 μL, including 10 μL SYBR Premix Ex Taq (2x), 1 μL of PCR Forward Primer (10 uM), 1 μL of PCR Reverse Primer (10 uM), 2 μL of cDNA, and 6 μL of double-distilled water. Primer sets are listed in Supplementary Table 2. The qRT-PCR was performed with an initial denaturation step of 10 min at 95°C; 95°C (15 s), 60°C (30 s), 72°C (30 s) for a total 40 cycles; and a final extension step at 72°C for 5 min. All experiments were performed in triplicate. All samples were normalized to GAPDH. The geometric mean in each triplicate was used to calculate the relative lncRNAs concentrations (ΔCt = Ct of lncRNAs − Ct of GAPDH). The fold change of expression was calculated using the 2^{–Δ Δ Ct} method.

The immortalized EVT cell line HTR-8/SVneo was used. The cells were maintained in RPMI-1640 supplemented with 5% fetal calf serum, 1% L-glutamine 200 mM, and 1% penicillin–streptomycin (all from Invitrogen) under an atmosphere of 5% CO₂ at 37°C. To establish the hypoxia model, the cells were exposed to 94%N₂/1% O₂/5% CO₂ for 2, 24, 48, and 72 h. The control group was cultured under normal air condition for the corresponding period.

The coding sequence of human HIF-1α was amplified by PCR from human genomic DNA libraries (GENECHEM Company, Shanghai, China) and was subcloned into the pEZ-MO2 plasmid via homologous recombination according to manufacturer’s protocol (Hieff Clone^® Plus One Step Cloning Kit, YEASEN, Shanghai, China) to construct pEZ-MO2–HIF-1α overexpression. The sequence of the construct was confirmed by DNA sequencing at Sangon Biotech Company (Shanghai, China). The HTR-8/SVneo cells at 70% confluence were transiently transfected with pEZ-MO2-HIF-1α plasmid for 48 h using Lipofectamine 3000 (Invitrogen) according to manufacturer’s instructions. The empty pEZ-MO2 plasmid was also transfected as negative control (NC).

lncRNA-HEIPP in HTR-8/SVneo cells were knocked down by transfection of preannealed stealth small interfering RNA (siRNA) duplexes designed online by Thermo Fisher’s BLOCK-iT^TM RNAi Designer⁴, using Lipofectamine RNAiMAX transfection reagent (Invitrogen). Silencer^® Negative Control #1 siRNA (si-NC, AM4611, Ambion) was transfected as an NC. The efficiency of knockdown was determined by RT-PCR.

Cell viability was assessed using the Cell Counting Kit-8 (Dojindo Laboratories, Kumamoto, Japan) according to the manufacturer’s protocol. Cell proliferation curves were plotted using the absorbance at each time point (0, 12, 24, and 48 h). All experiments were performed in triplicate.

The invasion of HTR-8/SVneo cells across Matrigel was objectively evaluated in an invasion chamber according to the manufacturer’s protocol (Millipore). The cells transfected with si-NC or si-HEIPP (10 and 30 μL) were plated in the upper chambers of 24-well plates at a density of 1.0 × 10⁴ cells/well in RPMI-1640 medium containing 5% fetal bovine serum; the membrane pore of the Transwell chamber was 8 μm in diameter. The membranes were coated with Matrigel (BD, United States) to form matrix barriers at the concentration of 200 μg/mL. A 700-μL aliquot of RPMI-1640 medium containing 10% FBS was immediately placed in the lower well of the chamber as a chemoattractant. After incubation for 48 h at 37°C, cells were completely removed from the upper surface of the filter using a sterile cotton swab, and cells that had migrated to the lower surface were fixed and stained with 0.1% crystal violet. Finally, the cell migration ability was determined by counting the number of stained cells on the membranes in 10 randomly selected, non-overlapping fields at 20× magnification under microscope. The migration index was calculated as the ratio of the percentage of cell migration in the various treatments to that of the vehicle. Each experiment was performed in triplicate and repeated three times.

Bisulfite pyrosequencing was used to quantity DNA methylation at two sites in the DMR of the lncRNA-HEIPP gene. Two micrograms of DNA was subjected to bisulfate conversion using the EpiTect Bisulfite Kit (59104) (Qiagen Ltd., GmbH, Hilden, Germany). Pyrosequencing primers were designed using PyroMark Assay Design 2.0 (Qiagen Ltd., GmbH, Hilden, Germany) and listed in Supplementary Table 2. One hundred nanograms of bisulfite-converted DNA was PCR amplified in a 50-μL reaction volume containing 5 × PCR buffer (KAPA Ltd., United States), 10 mM of each dNTP (KAPA Ltd., United States), 50 mM forward and reverse primers (BGI Inc., Shanghai, China), and 1 U DNA polymerase (HotStart Taq, KAPA Ltd). The PCR cycling parameters were 95°C for 3 min; 40 cycles of 95°C for 30 s, a variable annealing temperature (Ta) for 30 s, and 72°C for 1 min; and extension at 72°C for 7 min. A 25-μL aliquot of each PCR product was subjected to pyrosequencing using the PyroMark Q96 ID Pyrosequencer (Qiagen Ltd.) following the manufacturer’s recommended protocols. The degree of methylation at each CpG site was determined using Pyro Q-CpG Quantification software. All samples were analyzed in triplicate.

The HTR-8 trophoblast cell line was used for ChIP-based analysis for enrichment of H3K4me3 on lncRNA-HEIPP promoter. ChIP assays were conducted according to the manufacturer’s instruction (Millipore, 17-10086). The chromatin extracted from cells was sheared, subjected to immunoprecipitation with H3K4me3 (Abcam) or immunoglobulin G (IgG) (Millipore) antibodies, reverse cross linked, and subjected to quantitative ChIP- PCR (qChIP). The qChIP was performed on sheared DNA with following primers: SP1: F: 5′ ATTTTGCTTCC TATCCCT 3′, R:5′ ACCGACAATCTCCTCAGT 3′, SP2: F: 5′ GA CTGAGGAGATTGTCGGT 3, R: 5′ CTTATGATGGTCTGGGT GA 3′, SP3: F: 5′ TGGAGGAGGGAGATGACA 3′, R: 5′ TGGA GGAGGGAGATGACA 3′, SP4: F: 5′ CATGCCCTCTCAGC CTAAC 3′, R: 5′ AGTCCCCCCAAGAAAAACA 3′. The four pairs of primers amplified −324 to −1,288 bp region upstream from the transcriptional starting site of lncRNA-HEIPP. The results were normalized to input and expressed as a percentage of the fold difference. IgG was used as an NC. Data were obtained from at least three independent experiments.

All data are expressed as mean ± SD (standard deviation), and all statistical analyses were performed using the SPSS statistical software package (SPSS, Inc., Chicago, IL, United States). All the data were tested for homogeneity of variance by Bartlett test before analyzing the significance. Comparisons were made using Student t-test and analysis of variance. P < 0.05 was considered statistically significant.

The demographic data of PE women and normal pregnant women are summarized in Table 1. There was no significant difference in maternal age between normal pregnancies and pregnancies with PE (P > 0.05). The systolic and diastolic blood pressure was significantly higher in pregnant women with PE (P < 0.001). The average proteinuria is 2.62 ± 0.20 g/24 h, which means that the cases enrolled this study were severe PE patients (proteinuria >2.0 g/24 h). Women from the PE group delivered earlier than normal pregnant women mostly due to the medical termination of the pregnancy, which is the major cause of the lower infant birth weight in the PE group, as there was no intrauterine growth restriction patient in the PE group enrolled in this study.

As the transcriptome of the placenta is dynamically changed during advancing gestation (Deyssenroth et al., 2017), to ensure the identical or similar baselines between the two disease groups, we selected four pairs of PE patients and controls with matched gestational age within 2 weeks (Supplementary Table 1). We aligned lncRNA array data into RefSeq_NR, UCSC, and Ensembl database and compared the lncRNA expression levels between four human PE placenta (P) and four normal samples (N). The expression profiles of lncRNAs were shown by calculating the log-fold change (P/N). Among these lncRNAs, 405 were consistently upregulated, and 344 lncRNAs were consistently downregulated (Figure 1A). The clustering analysis also showed the differential expression patterns of mRNAs between PE placentas and normal samples (Figure 1B). The complete dataset regarding the significantly upregulated and downregulated lncRNA or mRNA was deposit on the public repository⁵.

GO pathway analysis was performed to determine the lncRNA and mRNA enrichment in biological processes, cellular components, and molecular functions. HIF-1α pathway is known to play an important role in the pathogenesis of PE and identified as one of the most significant enrichment signaling pathways under our experimental settings (Figure 2A and Supplementary Table 3). We found that a number of lncRNAs and molecules are closely related with HIF-1α pathway (Figure 2B), and the top 10 lncRNAs (AF064860.7, AC027612.3, RP11-244N9.4, RP11-574M7.1, RP11-80I15.4, RP11-939C17.2, XLOC_013276, AC023085.1, AC009236.1, and RP11-184D12.1) showing mostly dramatic different expression between PE and normal placentas were selected for further validation (Supplementary Table 4).

To further validate the preliminary data from microarray, we performed qRT-PCR to determine the expression of the ten candidate lncRNAs related with HIF-1α pathway in 30 PE placenta tissues and 30 normal placenta tissues. We found that 6 of the 10 lncRNAs manifested significant upregulation in PE samples compared with normal samples (Figure 3A), which are consistent with the array results. Among the six lncRNAs, lncRNA-AF064860.7 showed the most significant differences between PE samples and normal samples, so that we named it lncRNA-HEIPP (high expression in PE placenta).

In the HTR-8/SVneo human trophoblast cell lines, we found that hypoxia could induce the expression of lncRNA-HEIPP, which reached the highest level at 24 h upon hypoxia treatment (Figure 3B). Although the other five candidate lncRNAs also showed time-dependent changes of expression upon hypoxia treatment, the increasing trend and fold change were less significant than that in lncRNA-HEIPP. Moreover, the overexpression of HIF-1α increased the expression of lncRNA-HEIPP, further suggesting the interaction between HIF-1α pathway and lncRNA-HEIPP (Figure 3C).

As the epigenetic regulatory elements such as DNA methylation and H3K4me3 occupancy were critical for the gene expression regulation in placental development and trophoblast invasion (Apicella et al., 2019; Kwak et al., 2019), we analyzed the epigenetic regulatory elements of lncRNA-HEIPP in the placenta in PE versus controls. Specifically, compared to the normal oxygen condition, the DNA methylation levels of lncRNA-HEIPP promoter in HTR-8/SVneo human trophoblast cells increased along with the extension of hypoxia time, which reached the highest level at 24 h upon hypoxia treatment (Figure 4A). However, the increase of methylation level was not significant at each time point (Figure 4B).

Enrichment of histone trimethylation at lysine 4 (H3K4me3) at lncRNA-HEIPP DMRs/promoters in HTR-8/SVneo human trophoblast cells that underwent hypoxia or normoxia was assessed using ChIP-qPCR. It was shown that the binding of H3K4me3 on lncRNA-HEIPP promoter was specifically enriched compared with the binding of IgG (Figure 5A). As shown in Figure 5B, the protein expression of H3K4me3 was not affected by hypoxia treatment. However, H3K4me3 manifested significantly higher promoter occupancy on lncRNA-HEIPP promoter in HTR-8/SVneo cells that underwent hypoxia for 24 h compared to those that underwent normoxia (Figure 5C), suggesting the histone modification, instead of the DNA methylation might be the epigenetic pathway that causes the increased expression of lncRNA-HEIPP in the placenta during PE undergone hypoxia.

Next, we assessed the biological functions of lncRNA-HEIPP in trophoblast cells. We knocked down the expression of lncRNA-HEIPP in HTR-8/SVneo human trophoblast cell line using HEIPP siRNA (si-HEIPP) and found that the proliferation of HTR8/SVneo cells was not affected (data not shown). However, the invasion of HTR8/SVneo cells was significantly promoted, compared to the cells treated with NC siRNA (Figure 6).