Abstract

   Preeclampsia (PE) is a syndrome manifested by high blood pressure that
   could develop in the latter half of pregnancy; however, the underlying
   mechanisms are not understood. Recent evidence points to the function
   of noncoding RNAs (ncRNAs) as novel regulators of the invasion,
   migration, proliferation, and apoptosis of trophoblasts involved in the
   development of placental vasculature. Here, we investigated the role of
   long intergenic ncRNA 00473 (linc00473) in PE and the associated
   molecular mechanisms. The expression of linc00473 was downregulated in
   the placenta of patients with severe PE as revealed by qRT-PCR
   analysis. In vitro, linc00473 knockdown in trophoblast cell lines
   HTR-8/SVneo, JAR, and JEG3 significantly inhibited cell proliferation
   and promoted apoptosis, whereas linc00473 overexpression stimulated
   trophoblast proliferation. The mechanistic insights were provided by
   RNA-seq and qRT-PCR, which revealed that linc00473 could regulate the
   transcription of genes relevant to cell growth, migration, and
   apoptosis. In particular, linc00473 inhibited the expression of tissue
   factor pathway inhibitor 2 (TFPI2) through binding to lysine-specific
   demethylase 1 (LSD1). These results indicate that linc00473 could be
   involved in the pathogenesis and development of PE and may be a
   candidate biomarker as well as therapeutic target for this disease.

   Keywords: preeclampsia, linc00473, LSD1, proliferation

Introduction

   Preeclampsia (PE) is a pregnancy-related syndrome developed after
   20 weeks of gestation and characterized by high blood pressure and
   either proteinuria or pulmonary edema, liver dysfunction,
   thrombocytopenia, progressive renal insufficiency, and new-onset
   cerebral or visual disturbances.[43]^1 PE is associated with maternal
   multisystem organ damage, fetal growth restriction, and intrauterine
   distress[44]2, [45]3 and is a leading cause of perinatal complications
   and maternal mortality worldwide.[46]4, [47]5, [48]6 Currently, the
   most efficient method of PE treatment is delivery, which, however, is
   applied depending on the maturity of the fetus and may not be the
   optimal approach in certain conditions.[49]^7 Therefore, it is
   important to explore specific molecular mechanisms of PE to develop
   effective treatment as well as prevention measures for PE.

   PE has multifactorial etiopathogenesis, including inflammation,[50]^8
   oxidative stress,[51]^9 endothelial dysfunction,[52]^10 disturbed
   balance between proangiogenic and antiangiogenic factors,[53]^11 as
   well as inheritance[54]^12 and dietary habits.[55]^13 During the
   primary development of the placenta, extravillous trophoblasts (EVTs)
   of the fetal origin play an important role in the migration, invasion,
   and remodeling of the maternal spiral arteries in the myometrium and
   decidua.[56]14, [57]15 The process of uterine spiral artery remodeling
   is vital for decreasing maternal blood flow resistance and increasing
   uteroplacental perfusion.[58]^16 Previous studies documented excessive
   apoptosis of EVTs in patients with PE, resulting in poor cell
   proliferation and impaired capacity for migration and invasion of the
   spiral arteries in the myometrium, which negatively affected the
   conversion of the maternal arteries into large-diameter vessels
   necessary for fetus blood supply.[59]^17 However, despite extensive
   research, specific molecular mechanisms underlying EVT apoptosis and
   their decreased migration and invasion in PE patients are still poorly
   understood.

   Long noncoding RNAs (lncRNAs) are RNA molecules of more than 200
   nucleotides in length considered to play an important role in the
   regulation of gene expression both at the transcriptional and
   posttranscriptional level.[60]^18 Research of the past decade has
   implicated a number of lncRNAs in aging, neurological diseases, and
   cancer, thus providing a new dimension in our understanding of the
   molecular mechanisms involved in normal physiological processes and
   disease. The expression of lncRNAs is correlated with such cellular
   processes as apoptosis,[61]^19 cell proliferation,[62]^20
   migration,[63]^21 and stem cell pluripotency.[64]^22 The recent data
   indicate that lncRNAs may play a role in the pathogenesis of PE by
   regulating trophoblast invasion and apoptosis. Therefore, the
   identification of PE-associated lncRNAs and the related molecular
   mechanisms is critical for understanding PE development and improvement
   of treatment strategies.

   Long intergenic ncRNA 00473 (linc00473) is a 1,832-bp lncRNA located on
   chromosome 22q12.2, which is implicated in cancer cell growth,
   survival, and invasion. Thus, it has been reported that linc00473 is
   induced in mucoepidermoid carcinoma (MEC) associated with a unique
   chromosomal translocation CRTC1-MAML2 and that it promoted the
   proliferation and survival of human MEC cells, thus serving as a
   biomarker for CRTC1-MAML2-positive MEC.[65]^23 linc00473 was also shown
   to stimulate the growth and inhibit apoptosis in cervical cancer cells
   both in vitro and in vivo through suppression of interleukin
   enhancer-binding factor 2 (ILF-2) degradation.[66]^24 Moreover, a
   recent study indicated that linc00473 can play an important role in
   establishing pregnancy by mediating decidualization of human
   endometrial stromal cells.[67]^25

   Based on these data, we had a hypothesis that linc00473 may be involved
   in the pathogenesis of PE and tested it in this study by comparing the
   expression of linc00473 in placental tissues of women with normal
   pregnancies and PE and investigating the effects of linc00473 on
   trophoblast proliferation, apoptosis, migration, and invasion in vitro.
   Our findings may provide important insights into the mechanisms
   underlying the regulation of trophoblast function and the pathogenesis
   of PE.

Results

Linc00473 Is Downregulated in the Placenta of PE Patients

   We detected and verified the full poly(A)-positive sequence of
   linc00473 using rapid amplification of cDNA ends (RACE) ([68]Figures 1A
   and 1B). The results indicated that the full linc00473 sequence was
   consistent with the reference sequence ([69]NR_026860.1) in the 3′ end;
   however, there was an extra 39-bp stretch (red marker) in the 5′ end
   compared with the reference sequence.

Figure 1.

   [70]Figure 1
   [71]Open in a new tab

   Linc00473 Expression Is Decreased in PE

   (A) Agarose gel electrophoresis of PCR products generated in 3′ (left)
   and 5′ (right) RACE of linc00473 in HTR-8/SVneo cells. (B) The sequence
   of LINC00673 at the 3′ terminus is consistent with that published in
   the NCBI database ([72]NR_026860.1) but is elongated by 39 bp at the 5′
   terminus (red letters). The expression of linc00473 was analyzed by
   qRT-PCR and normalized to that of GAPDH. (C) linc00473 levels were
   lower in PE placentas (n = 50) compared to normal placentas (n = 50).
   (D) Results are presented as the fold-change in PE placental samples
   relative to the control, and linc00473 expression was classified into
   two groups. (E) linc00473 expression in trophoblast cell lines analyzed
   by qRT-PCR. The levels of linc00473 in HTR-8/SVneo, BeWo, JAR, and JEG3
   cells were normalized to that in HUVECs. (F and G) Relative linc00473
   expression in HTR-8/SVneo, JAR, and JEG3 cells transfected with
   linc00473 siRNAs (F) and pcDNA3.1-linc00473 (G). The data are presented
   as the mean ± SD of three independent experiments; **p < 0.01.

   Next, we evaluated the expression levels of linc00473 in 50 pairs of
   severe PE and normal placental tissues of pregnant women by qRT-PCR.
   Analysis of participants’ clinical characteristics ([73]Table 1)
   indicated that systolic and diastolic blood pressure and proteinuria
   were apparently higher in PE patients than in women with normal
   pregnancies (p < 0.01, p < 0.01, and p < 0.05, respectively), whereas
   the body weight of neonates from PE mothers was lower because of early
   pregnancy termination. As shown in [74]Figures 1C and 1D, linc00473
   expression level in the placenta was significantly lower (p < 0.05) for
   78.0% (39/50) of PE patients compared to normal pregnant women.

Table 1.

   Clinical Characteristics of Patients with PE and Normal Pregnancies
   Variable PE (n = 50) Control (n = 50) p Value (PE versus Control)
   Maternal age (year) 31.66 ± 4.103 33.28 ± 3.399 p > 0.05
   Maternal weight (kg) 74.8 ± 10.837 72.17 ± 9.448 p > 0.05
   Smoking 0 0 p > 0.05
   Systolic blood pressure (mm Hg) 162.4 ± 15.748 115.36 ± 7.889 p < 0.01
   Diastolic blood pressure (mm Hg) 106.6 ± 11.441 73.54 ± 8.581 p < 0.01
   Proteinuria (g/day) >0.3 <0.3 p < 0.05
   Body weight of infant (g) 2,413.2 ± 1,003.853 3,375 ± 379.749 p < 0.05
   [75]Open in a new tab

Linc00473 Influences Proliferation and Apoptosis of Trophoblasts

   We then assessed linc00473 expression in four trophoblast cell
   lines (HTR-8/SVneo, BeWo, JEG3, and JAR) after normalization to that in
   other cells relevant to pregnancy (human umbilical vein endothelial
   cells [HUVECs]) ([76]Figure 1E) and chose HTR-8/SVneo, JAR, and JEG3
   cells to investigate linc00473 functional activity. To assess linc00473
   effects in trophoblasts, linc00473 was overexpressed in the three cell
   lines and knocked down in HTR/SVneo cells, because linc00473 levels in
   JAR and JEG3 cells were relatively low and the cells were more suitable
   for linc00473 overexpression. qRT-PCR analysis revealed that linc00473
   expression was silenced in HTR/SVneo cells more efficiently by
   si-linc00473-1# and si-linc00473-3# than by si-linc00473-2#
   ([77]Figure 1F); therefore, si-linc00473-1# and/or si-linc00473-3# were
   used for subsequent experiments. The upregulation of linc00473
   expression after pcDNA-linc00473 transfection was confirmed in all
   three cell lines ([78]Figure 1G).

   Knockdown of linc00473 expression by si-linc00473 significantly
   suppressed cell proliferation in trophoblast cells as evidenced by the
   MTT assay, whereas linc00473 overexpression promote cell proliferation
   ([79]Figures 2A–2C). Knockdown of linc00473 expression by si-linc00473
   dramatically diminished clonogenic survival in HTR-8/SVneo cells as
   indicated by the colony-formation assay ([80]Figure 2D). At the same
   time, linc00473 overexpression significantly increased clonogenic
   survival of the cultured trophoblasts ([81]Figures 2D–2F). The results
   of ethynyl deoxyuridine (EDU) immunostaining confirmed these findings
   ([82]Figures 2G–2I). Overall, these data indicate that linc00473
   positively regulates the proliferation of HTR/SVneo, JAR, and JEG3
   trophoblast cells, suggesting that linc00473 deficiency can contribute
   to the development of PE.

Figure 2.

   [83]Figure 2
   [84]Open in a new tab

   Linc00473 Promotes Trophoblast Proliferation In Vitro

   (A–C) The viability of si-linc00473-transfected HTR-8/SVneo cells (A)
   or pCDNA3.1-linc00473-transfected JAR (B) and JEG3 (C) cells were
   analyzed by the MTT assay. (D–F) The proliferation of
   si-linc00473-transfected HTR-8/SVneo cells (D) and
   pcDNA3.1-linc00473-transfected HTR-8/SVneo, JAR (E), and JEG3 (F) cells
   was evaluated by the colony formation assay, and the number of colonies
   was counted. (G–I) Proliferating HTR-8/SVneo (G), JAR (H) and JEG3 (I)
   cells are marked with EDU (red) and cell nuclei are stained with DAPI
   (blue). The data are presented as the mean ± SD of three independent
   experiments; *p < 0.05 and **p < 0.01.

   Furthermore, we found that linc00473 was involved in the regulation of
   cell cycle progression and apoptosis in trophoblasts. Flow cytometry
   analysis revealed that linc00473 knockdown resulted in cell cycle
   arrest at the G0/G1 phase, whereas linc00473 overexpression reversed
   the effect ([85]Figures 3A–3C). Consistent with these results, the
   apoptosis rate in HTR/SVneo trophoblasts increased after linc00473
   knockdown by specific siRNAs ([86]Figure 3D).

Figure 3.

   [87]Figure 3
   [88]Open in a new tab

   Effects of linc00473 on Trophoblast Cell Cycle and Apoptosis In Vitro

   HTR-8/SVneo trophoblasts were transfected with linc00473-specific
   siRNAs (si-linc00473-1# and si- linc00473-3#) and JAR and JEG3
   trophoblasts were transfected with pcDNA3.1-linc00473. (A–C) Cell cycle
   was analyzed by flow cytometry in HTR-8/SVneo (A), JAR (B) and JEG3 (C)
   cells. Representative FACS images and related statistics are shown. (D)
   Cell apoptosis rates were analyzed by flow cytometry. LR, early
   apoptotic cells; UR, terminal apoptotic cells. The data are presented
   as the mean ± SD of three independent experiments; *p < 0.05 and **p <
   0.01.

Linc00473 Promotes the Migration and Invasion of Trophoblasts In Vitro

   There is little doubt cell metastasis is an important part of cancer
   progression and may worsen the outcomes of patients with trophoblast
   migration and invasion, which are important for the establishment of
   the blood flow between the mother and embryo, and the impairment of
   these processes can result in PE. Therefore, we evaluated the effects
   of linc00473 on the migration and invasion abilities of HTR-8/SVneo,
   JAR and JEG3 cells by Transwell assays. The results indicated that a
   decrease in linc00473 expression inhibited trophoblast migration and
   invasion compared to control cells, whereas linc00473 overexpression
   promoted them ([89]Figures 4A–4C), suggesting that linc00473 can
   influence the migration and invasion of trophoblasts, which are
   critical for the development of placental vasculature impaired in PE
   patients.

Figure 4.

   [90]Figure 4
   [91]Open in a new tab

   Effect of linc00473 on Trophoblast Migration, Invasion, and
   Vascularization In Vitro

   (A–C) The migration and invasion capacity of cultured trophoblasts
   transfected with si-linc00473 in HTR-8/SVneo cells (A) or
   pcDNA3.1-linc00473 in JAR (B) and JEG3 (C) cells were analyzed by the
   Transwell assays and were significantly lower or higher, respectively,
   than that of control cells. (D) The number of branch points and total
   branching length of the network were decreased in
   si-linc00473-transfected HTR-8/SVneo cells and HUVECs compared to
   control cells. The data are presented as the mean ± SD of three
   independent experiments; **p < 0.01.

Linc00473 Affects Network Formation Ability of Trophoblasts In Vitro

   It was reported that impaired spiral artery remodeling is associated
   with the pathogenesis of PE as it leads to hypoxia, placental ischemia,
   and the release of placental factors into the maternal blood
   circulation, resulting in the systemic inflammatory response and
   vascular endothelial injury. Therefore, we explored the role of
   linc00473 on the network formation ability in two types of
   pregnancy-related cells, HUVECs and HTR-8/SVneo trophoblasts. The
   knockdown of linc00473 weakened the network formation ability of both
   HTR-8/SVneo cells and HUVECs, as evidenced by a decrease in the number
   of branches and total branching length of the network compared to
   control cells ([92]Figure 4D).

Linc00473 Influences Gene Expression in Trophoblasts

   We next performed RNA transcriptome sequencing in scrambled negative
   control siRNA (si-NC) and si-linc00473-transfected HTR/SVneo cells to
   explore linc00473-associated changes in gene expression. A total of 249
   and 395 transcripts were increased or decreased, respectively, in cells
   with linc00473 deficiency ([93]Figure 5B; [94]Table S2). Evaluation of
   the biological pathways activated by linc00473 using the gene ontology
   (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases
   indicated that genes related to cell growth were deregulated in
   linc00473 knockdown cells ([95]Figure 5A). Expression changes in
   HTR-8/SVneo and JAR cells were further confirmed by qRT-PCR. Among the
   transcripts related to cell growth, apoptosis, and migration, the
   expression of tissue factor pathway inhibitor 2 (TFPI2) was
   significantly upregulated after linc00473 knockdown in HTR-8/SVneo
   cells ([96]Figure 5C) and downregulated by linc00473 overexpression in
   JAR cells ([97]Figure 5D). Consistent with these results, western
   blotting analysis indicated similar effects of linc00473 on TFPI2
   protein expression in HTR-8/SVneo and JAR cells ([98]Figures 5E and
   5F).

Figure 5.

   [99]Figure 5
   [100]Open in a new tab

   Linc00473 Knockdown Affected the Expression of Genes Involved in Cell
   Proliferation and Migration

   (A and B) Gene expression profiling was performed by RNA-seq in
   HTR-8/SVneo cells transfected with si-linc00473. Differentially
   expressed genes are shown (A). GO analysis of genes differentially
   expressed in si-NC- and si-linc00473-transfected cells. Cell growth was
   among significant biological processes affected by linc00473 depletion
   in trophoblasts as evidenced by the number of genes with altered
   expression (B). (C and D) Relative mRNA expression of genes involved in
   cell proliferation and migration in linc00473-deficient HTR-8/SVneo
   cells (C) and linc00473-overexpressing JAR cells (D) was analyzed by
   qRT-PCR. (E and F) TFPI2 protein expression in linc00473-deficient
   HTR-8/SVneo cells (E) and linc00473-overexpressing JAR cells (F) was
   assessed by western blotting. The data are presented as the mean ± SD
   of three independent experiments; *p < 0.05 and **p < 0.01.

Linc00473 Silences TFPI2 Epigenetically by Binding to LSD1

   To test linc00473 effects on gene expression, we first investigated the
   subcellular localization of linc00473 in HTR/SVneo, JAR, and JEG3 cells
   using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and small
   nuclear RNA U1 (RNU1) as cytoplasmic and nuclear markers, respectively
   ([101]Figures 6A–6C). The results indicate that linc00473 was mainly
   localized in the cell nucleus, supporting its role in transcriptional
   regulation.

Figure 6.

   [102]Figure 6
   [103]Open in a new tab

   Linc00473 Binding to LSD1 Inhibits TFPI2 Expression

   (A–C) Cell fractionation assay revealed that linc00473 was
   predominantly localized in the nucleus of HTR-8/SVneo (A), JAR (B), and
   JEG3 (C) cells. GAPDH and small nuclear RNA U1 were used as cytoplasmic
   and nuclear markers, respectively. (D–F) RIP assays showed that
   linc00473 was mainly bound to LSD1 in HTR-8/SVneo cells (D), JAR (E),
   and JEG3 (F) cells, respectively. (G–I) LSD1 silencing changed the
   expression of LSD1 and TFPI2 at the protein (G) and mRNA levels in
   HTR-8/SVneo (H) and JAR (I) cells. (J and K) Enrichment of LSD1 (J) and
   H3K4me2 (K) in the promoter region of TFPI2 detected by the ChIP assay.
   The data are presented as the mean ± SD of three independent
   experiments; *p < 0.05 and **p < 0.01.

   Lysine-specific demethylase 1 (LSD1) is a flavin adenine dinucleotide
   (FAD)-dependent amine oxidase that specifically performs demethylation
   of H3K4me2 and H3K9me2,[104]26, [105]27 thus affecting gene
   transcription. Recently, it was shown that by regulating target gene
   expression, LSD1 can influence tumorigenesis,[106]28, [107]29 embryonic
   differentiation,[108]30, [109]31 and the formation of
   euchromatin.[110]^32 Therefore, we hypothesized that linc00473 may
   control TFPI2 expression by recruiting LSD1 in trophoblasts. Indeed,
   the RNA immunoprecipitation (RIP) assay showed that linc00473 was
   precipitated with anti-LSD1 antibodies in HTR-8/SVneo, JAR, and JEG3
   cells ([111]Figures 6D–6F), thus confirming our hypothesis.

   We then investigated the potential functional relationship between
   linc00473 and LSD1. The knockdown of LSD1 by a specific siRNA resulted
   in TFPI2 upregulation both at the protein and mRNA levels ([112]Figures
   6G–6I), suggesting that linc00473 may regulate TFPI2 expression in
   trophoblasts through epigenetic mechanisms. The chromatin
   immunoprecipitation (ChIP) assay with antibodies against LSD1 and
   H3K4me2 revealed the enrichment of LSD1 and H3K4me2 in the promoter
   region of the TFPI2 gene, which was significantly decreased for LSD1
   but increased for H3K4me2 after transfection with si-linc00473
   ([113]Figures 6J and 6K).

TFPI2 Upregulated in PE Placental Tissues Influences Trophoblast Viability
In Vitro

   Evaluation of TFPI2 mRNA expression in 31 placental tissues of PE
   patients indicated that TFPI2 transcription was increased in PE
   compared to normal pregnancy ([114]Figure 7A), suggesting a TFPI2 role
   in PE pathogenesis. In HTR-8/SVneo and JAR trophoblasts, TFPI2
   overexpression significantly inhibited, whereas TFPI2 knockdown
   promoted = cell proliferation ([115]Figures 7B and 7C). Finally, we
   examined the biological effects of TFPI2 association with linc00473 by
   co-transfecting HTR-8/SVneo and JAR cells with linc00473 and TFPI2
   siRNAs and analyzing cell proliferation using colony-formation and EDU
   assays. The results indicated that si-TFPI2 could moderately rescue
   si-linc00473-inhibited proliferation of HTR-8/SVneo and JAR cells, and
   similar results were obtained by co-transfecting trophoblasts with
   pcDNA3.1-TFPI2 and pcDNA-linc00473 ([116]Figures 7D and 7E).

Figure 7.

   [117]Figure 7
   [118]Open in a new tab

   TFPI2 Inhibits Trophoblast Proliferation and Counterbalance the
   Activity of Linc00473

   (A) Relative expression of TFPI2 (fold change) in placental tissues of
   PE patients compared to those of healthy pregnant women (n = 31). (B
   and C) MTT analysis of cell viability for si-TFPI2- (B) or
   pcDNA3.1-TFPI2-transfected (C) trophoblasts. (D and E) Proliferation of
   HTR-8/SVneo and JAR cells analyzed by the colony-formation (D) and EDU
   (E) assays. The data are presented as the mean ± SD of three
   independent experiments; *p < 0.05 and **p < 0.01. (F) A schematic
   pathway illustrating the role of linc00473 in the proliferation and
   migration of trophoblasts in PE.

Discussion

   The human genome is estimated to contain approximately 16,000
   non-coding RNAs (ncRNAs), including microRNAs (miRNAs) and lncRNAs
   (>200 nucleotides). The lncRNAs are localized in the nucleus and are
   enriched in the chromatin or specific subnuclear compartments, where
   they can act as structural scaffolds of nuclear domains and influence
   chromatin organization,[119]^33 as well as transcriptional and
   posttranscriptional regulation of gene expression.[120]34, [121]35 In
   our previous studies, we have shown the involvement of several lncRNAs
   (MEG3, TUG1, PVT1, and MVIH)[122]36, [123]37, [124]38, [125]39 in the
   occurrence and development of PE. Thus, exploration and identification
   of PE-associated lncRNAs and their functions may provide a new angle in
   the understanding of PE pathogenesis and define novel therapeutic
   targets and diagnostic and prognostic markers for PE.

   In recent years, it has remained the focus of research to search for
   the pathogenesis of PE in placental trophoblastic cells about relating
   biological indicators. Currently, the methods of studying the mechanism
   of disease includes in vivo cell study, in vitro cell culture, animal
   models, etc. Among them, the study of cell culture in vitro only needs
   a simplified cell growth environment. It is convenient to obtain a
   uniform cell group and to facilitate the application of experimental
   intervention factors. We can observe the experimental results easily in
   a short time, and the cells have become a common method used by many
   researchers. At present, more cells are studied. There are
   HTR-8/SVneo,[126]^40 JEG-3,[127]^41 BeWo,[128]^42 JAR,[129]^43
   dNK,[130]^44 villous 3A,[131]^45 and so on. So, in our study, we
   explored the potential molecular mechanism of linc00473 mainly based on
   several different trophoblastic cell lines.

   Linc00473 has been mostly investigated for its role in cancer.[132]46,
   [133]47 In this study, we showed that linc00473 expression was
   dramatically decreased in the placental tissues of women with PE
   compared to that in women with normal pregnancies. In vitro, silencing
   of linc00473 expression suppressed the proliferation, cell cycle
   progression, invasion, and migration while promoting apoptosis of
   cultured trophoblasts, whereas linc00473 overexpression caused the
   opposite effects. Together, these findings characterize linc00473 as an
   important regulatory molecule involved in the control of the biological
   activity of trophoblasts, the main players in the development of PE,
   and suggest that linc00473 may be a promising biomarker for PE
   prediction.

   According to our results, linc00473 was predominantly localized in the
   nucleus, indicating its possible involvement in transcriptional
   regulation. Indeed, we found that linc00473 could bind LSD1, suggesting
   that it may affect LSD1-mediated epigenetic regulatory mechanisms and,
   through them, influence the onset and progression of PE.

   RNA sequencing (RNA-seq) and qRT-PCR analyses indicated that TFPI2 was
   markedly upregulated by linc00473 knockdown. TFPI2 is a
   matrix-associated Kunitz-type serine protease inhibitor that controls
   plasmin- and trypsin-mediated activation of zymogen matrix
   metalloproteinases involved in tumor progression and metastasis, and it
   was shown that TFPI2 expression inversely correlated with cancer cell
   invasion and migration.[134]48, [135]49, [136]50 Consistent with these
   findings, TFPI2 knockdown promoted proliferation of cultured
   trophoblasts and counterbalanced the inhibitory effects of linc00473
   deficiency. Our results also indicated that TFPI2 expression was
   silenced by LSD1 through epigenetic mechanisms. Based on these
   findings, we propose that linc00473 can inhibit TFPI2 expression by
   binding to LSD1 in trophoblasts, thus promoting their invasion and
   migration, the critical processes for proper uterine spiral artery
   remodeling in pregnancy, which are deregulated in PE.

   In summary, our study shows that linc00473 is downregulated in
   placental tissues from PE patients compared with normal pregnant women
   and that low expression of this lncRNA may potentially serve as a
   prognostic biomarker of PE. Moreover, our results strongly support the
   role of linc00473 as a scaffold and a member of the LSD1-mediated
   epigenetic regulatory pathway involved in the inhibition TFPI2
   expression during pregnancy. Together, the present findings suggest
   that linc00473 can be a novel molecular target for early diagnosis and
   treatment of PE ([137]Figure 7F). Further studies are needed to
   elucidate other potential mechanisms through which linc00473
   participates in the biological functions of trophoblasts in the context
   of PE.

Materials and Methods

Patients and Collection of Tissue Samples

   We obtained 50 paired placental samples from women with normal
   pregnancies and PE patients, who underwent cesarean deliveries in
   Jiangsu Province Hospital from August, 2016 to December, 2017. The
   placenta tissue samples (about 1 cm × 1 cm × 1 cm in size) were taken
   from the central area of the placenta maternal surface to avoid
   necrosis and calcification and were immediately frozen in liquid
   nitrogen and subsequently used for RNA and protein extraction.
   Clinicopathological characteristics of the participants are summarized
   in [138]Table 1. This research was authorized by the Ethnics Board of
   the First Affiliated Hospital of Nanjing Medical University, China, and
   all patients provided written informed consent.

Cell Culture

   Four human trophoblast cell lines (HTR/SVneo, JAR, JEG3, and BeVo) and
   HUVECs were purchased from the Type Culture Collection of the Chinese
   Academy of Sciences (Shanghai, China). HTR/SVneo and JAR cells were
   cultured in RPMI1640, JEG3 cells in MEM, BeVo cells in F12K, and HUVECs
   in endothelial cell medium (ECM) (all media from KeyGEN, Nanjing,
   China) supplemented with 10% fetal bovine serum (FBS) (Gibco-BRL,
   Invitrogen, Carlsbad, CA, USA), 100 U/mL penicillin, and 100 mg/mL
   streptomycin (Invitrogen). All cells were maintained in a humidified
   atmosphere with 5% CO[2] at 37°C.

RACE

   Both 5′ and 3′ RACE were conducted using the First Choice-RLM RACE kit
   (Life Technologies, Carlsbad, CA, USA) following the manufacturers’
   instructions.

Cell Transfection

   Plasmid vectors (pcDNA, pcDNA3.1-linc00473, and pcDNA3.1-TFPI2) were
   purified using the DNA Midiprep kit (QIAGEN, Hilden, Germany). Three
   different linc00473-specific siRNAs, LSD1 and TFPI2 siRNAs, and si-NC
   were purchased from Invitrogen; the sequences are shown in [139]Table
   S1. HTR/SVneo, JAR, and JEG3 cells were cultured in 6-well plates until
   80% confluence and then transfected with siRNAs using Lipofectamine
   2000 (Invitrogen, USA) or with plasmid vectors (4 g) using X-tremeGENE
   HP DNA Transfection Reagent (Roche, Pennsburg, Germany) following the
   manufacturers’ instructions. Cells were harvested 48 hr
   posttransfection and analyzed by qRT-PCR and western blotting.

RNA Preparation and qRT-PCR

   Total RNA was isolated from tissues or cultured cells with TRizol
   reagent (Invitrogen), and its quality and quantity assessed using
   NanoDrop2000c (Thermo Fisher Scientific, Waltham, MA, USA). RNA (1 μg)
   was reverse-transcribed to cDNA using Primer Script RT Master Mix
   (Takara, Dalian, China), and qRT-PCR was performed with SYBR Premix Ex
   Taq (Takara, Dalian, China) and specific primers ([140]Table S1) in an
   ABI 7500 system following the manufacturer’s protocol. The expression
   level of linc00473 was calculated by the 2^−ΔΔCT method, normalized to
   that of GAPDH, and converted to fold changes.

Subcellular Fractionation and RNA Isolation

   Cytoplasmic and nuclear RNA was isolated and purified using the PARIS
   Kit (Life Technologies) according to the manufacturers’ instructions.

Cell Proliferation and Colony Formation Assays

   For cell proliferation, HTR/SVneo, JAR, and JEG3 cells were seeded in
   96-well plates (3,000–4,000 cells/well) and transfected with
   si-linc00473 or pcDNA-linc00473; five replicates for each point were
   used. Cell viability was assessed every 24 hr using the Cell
   Proliferation Reagent Kit I (MTT) (Roche) based on the absorbance at
   490 nm measured in an ELx-800 University Microplate Reader (Biotech,
   Winooski, VT, USA).

   For the colony formation assay, cells were seeded in 6-well plates
   (800 cells/well), transfected with specific siRNA or plasmids, and
   cultured for 10–12 days with a change of medium at day 5. Cells were
   fixed with methanol and stained with 0.1% crystal violet
   (Sigma-Aldrich); colonies were observed under a microscope and counted
   using the Photoshop software. At least three independent experiments
   were performed.

EDU Assay

   The EDU assay was performed for accurate estimation of cell
   proliferation as a complementary method using the
   5-ethynyl-2-deoxyuridine labeling/detection kit (Ribobio, Guangzhou,
   China). HTR/SVneo, JAR, and JEG3 cells were seeded in 24-well plates at
   2∼3 × 10^4 cells per well and transfected with si-linc00473-1#,
   si-linc00473-3#, si-NC, pcDNA-linc00473, or pcDNA for 24 hr. Then
   200 mL/well of EDU labeling medium was added and cells were cultured
   for another 2 hr, fixed in 4% paraformaldehyde (pH 7.4) for 30 min,
   permeabilized with 0.5% Triton X-100 for 20 min at room temperature,
   washed in PBS (pH 7.4), and stained with anti-EDU solution for 30 min.
   After counterstaining with 250 mL DAPI (Invitrogen, Molecular Probes,
   Eugene, OR, USA) for 25 min, images of EDU-positive cells were captured
   under a fluorescence microscope (Nikon Corporation, Tokyo, Japan). The
   experiment was independently repeated three times.

Flow Cytometry Analysis of Cell Cycle and Apoptosis

   Cells transfected with si-linc00473 or pcDNA-linc00473 were harvested
   48 hr posttransfection by trypsinization, washed twice with PBS,
   stained with propidium iodide (PI) using the Cycle TEST PLUS DNA
   Reagent Kit (BD Biosciences, San Jose, CA, USA), and analyzed in a
   FACScan flow cytometer (BD Biosciences) equipped with a Cell Quest
   software (BD Biosciences) to estimate the percentage of cells in G0–G1,
   S, and G2–M phases. All samples were assayed in triplicate. Apoptosis
   was assessed by flow cytometry after double staining of the transfected
   cells with fluorescein isothiocyanate (FITC)-Annexin V and PI using the
   FITC-Annexin V Apoptosis Detection Kit (BD Biosciences), and the
   percentages of viable, dead, and early and late apoptotic cells were
   calculated and compared with those in controls.

Cell Migration and Invasion Assays

   Cell migration and invasion assays were performed using 24-well
   Transwell chambers with 8-μm pore size polycarbonate membranes
   (Corning, Corning, NY, USA). Cells were plated on the top side of the
   membrane pre-coated or not with Matrigel (BD, Franklin Lakes, NJ, USA)
   for invasion or migration assays, respectively. After incubation for
   36∼48 hr, cells inside the upper chamber were removed with cottons
   swabs, whereas those on the lower membrane surface were fixed by
   methanol and stained with 0.5% crystal violet solution. Five randomly
   selected fields were analyzed for each well.

Western Blotting Assay

   HTR/SVneo and JAR cells cultured in 6-well plates were transfected with
   si-linc00473 or pcDNA-linc00473 and total cellular proteins were
   extracted with radioimmunoprecipitation assay (RIPA) protein extraction
   reagent (Beyotime, Beijing, China) supplemented with protease inhibitor
   cocktail (Roche, Pleasanton, CA, USA) and phenylmethylsulfonyl fluoride
   (Roche). Proteins were separated by SDS-PAGE in 10% gels and
   transferred to 0.22-μm polyvinylidene difluoride membranes (Sigma),
   which were incubated with antibodies against LSD1 (1:1,000; CST,
   Danvers, MA, USA) or TFPI2 (1:1,000; Proteintech); anti-GAPDH antibody
   (1:3,000; Santa Cruz, CA, USA) was used as control. The intensity of
   protein bands was quantified by densitometry using the Quantity One
   software (Bio-Rad). At least three independent experiments were
   performed.

RNA-Seq Analysis

   The RNA-seq experiments were conducted in the Huada Genomics Institute
   (Wuhan, China), and the mRNA-seq library was obtained according to
   standard protocols (Illumina, San Diego, CA, USA). In brief, total RNA
   from si-NC- or si-linc00473-3#-transfected HTR-8/SVneo cells was
   isolated as described above and mRNA was purified using Dynabeads Oligo
   (dT) (Invitrogen Dynal) and reverse-transcribed into cDNA, which was
   then fragmented by nebulization to establish the mRNA-seq library.
   Then, the differentially expressed genes were screened by using the
   RSEM software and the differential gene expression patterns were
   analyzed by cluster analysis. GO annotation analysis of the
   distribution of gene functions was performed by using WEGO software
   ([141]http://wego.genomics.org.cn). Also, pathway enrichment analysis
   of the interaction between genes was analyzed by the KEGG database.

RIP Assay

   RIP assays were performed using the EZ-Magna RIP kit (Millipore,
   Billerica, MA, USA) according to the manufacturer’s instructions.
   HTR-8/SVneo, JAR, and JEG3 cells were grown in 15-cm plates, collected
   by centrifugation, and lysed in RIP lysis buffer so that 100 μL
   contained the lysate of 1∼2.0 × 10^7 cells. Cell lysates were used for
   immunoprecipitation with anti-LSD1, -EZH2, -AGO2, -DNMT1 and -DNMT3A
   antibodies and normal immunoglobulin G (IgG) (Millipore, Billerica, MA,
   USA) and the immunoprecipitated RNA was amplified by qRT-PCR using the
   Primer Script RT Master Mix kit and primers specific for linc00473 and
   IgG ([142]Table S1).

ChIP Assay

   ChIP assays were performed using the EZ-Magna ChIP A kit (Millipore)
   according to the manufacturer’s protocol. HTR-8/SVneo cells were
   cross-linked by 1% formaldehyde for 10 min at 25°C, lysed, and
   sonicated to obtain 200- to 500-bp DNA fragments. Primary LSD1 and
   H3K4me2-specific antibodies (Millipore) or control IgG were added to
   pre-cleared supernatants and the mixtures were incubated with rotation
   for 3 hr or overnight at 4°C. Then, Dynabeads Protein A/G were added to
   the mixtures and the samples were incubated with vortexing for 2 hr at
   4°C. The beads were then washed sequentially with low-salt and
   high-salt RIPA, LiCi, and TE and DNA was isolated by phenol/chloroform
   extraction. ChIP-qPCR was performed using SYBR Premix Ex Taq and
   specific primers ([143]Table S1), and the data were normalized to the
   input. At least three independent experiments were conducted.

Statistical Analysis

   The statistical analyses were performed with the SPSS 17.0 statistical
   software (IBM, Chicago, IL, USA). Each experiment was independently
   repeated at least three times and the data were expressed as the mean ±
   SD. Student’s t test or Mann-Whitney test were used for comparisons
   between two groups and ANOVA or Kreskas-Wallis tests were applied for
   multiple comparisons. p values less than 0.05 indicated statistical
   significance.

Author Contributions

   L.S., D.W. and Y.Z. conceived and designed the study. D.W. and Y.X.
   conducted the experiments and wrote the article. D.W., S.H., T.W. and
   S.W. collected clinical tissues and data. Q.Z., X.L., X.H. and J.W.
   coordinated and analyzed the data. All the authors contributed to, read
   and approved the final manuscript.

Conflicts of Interest

   No conflicts of interest were stated.

Acknowledgments