Potential regulatory network in the PSG10P/miR‐19a‐3p/IL1RAP pathway is possibly involved in preeclampsia pathogenesis

Abstract Preeclampsia (PE), a pregnancy‐specific disorder, is a leading cause of perinatal maternal‐fetal mortality and morbidity. Impaired cell migration and invasion of trophoblastic cells and an imbalanced systemic maternal inflammatory response have been proposed as potential mechanisms of PE pathogenesis. Comparative analysis between PE placentas and normal placentas profiled differentially expressed miRNAs, lncRNAs, and mRNAs, including miR‐19a‐3p (miRNA), PSG10P (lncRNA), and IL1RAP (mRNA). This study was conducted to investigate their potential roles in PE pathogenesis. The expression of miR‐19a‐3p, PSG10P, and IL1RAP was examined in PE and normal placentas using RT‐qPCR. An in vitro experiment was performed in human trophoblast HET8/SVneo and TEV‐1 cells cultured in normoxic and hypoxic conditions. MiR‐19a‐3p targets were identified using Targetscan, miRanda, and PicTar analysis as well as luciferase reporter assays. The mouse model of PE was conducted using sFlt‐1 for in vivo tests. Lower levels of miR‐19a‐3p, but higher levels of PSG10P and IL1RAP were observed in PE placentas and the trophoblast cells in hypoxia. Luciferase reporter assays confirmed that PSG10P and IL1RAP were both direct targets of miR‐19a‐3p. Exposure to hypoxia inhibited cell viability, migration, and invasion of HET8/SVneo and TEV‐1 cells. Knocking out PSG10P and IL1RAP or overexpressing miR‐19a‐3p rescued the inhibition caused by hypoxia. In vivo experiments showed that IL1RAP promoted the expression of caspase‐3, a key apoptosis enzyme, but inhibited MMP9, which is responsible for degrading the extracellular matrix, suggesting a significant role of IL1RAP in cell proliferation, migration, and invasion. miR‐19a‐3p, PSG10P, and IL1RAP were all found to be involved in PE pathogenesis. With a common targeting region in their sequences, a regulatory network in the PSG10P/miR‐19a‐3p/IL1RAP pathway may contribute to PE pathogenesis during pregnancy.


| INTRODUCTION
Preeclampsia (PE) is a pregnancy-specific disease characterized by hypertension (defined as systolic blood pressure P140 mm Hg or diastolic blood pressure P90 mm Hg) and proteinuria (>300 mg protein in a 24-hour urine specimen and/or protein to creatinine ratio of >0. 30) that occurs after 20 weeks of pregnancy. 1,2 PE affects 3%-8% of pregnancies worldwide and triggers problems in the liver, kidney, brain, and the clotting system during pregnancy. 3 As a result, PE is still one of the leading causes of maternal and fetal morbidity and mortality. 4 Growing evidence supports that PE results from abnormal placental development and function, 3,4 however, the pathophysiological mechanisms of PE remain poorly understood. In normal pregnancies, cytotrophoblasts (CTBs) migrate through the decidua and myometrium in order to invade the maternal spiral arteries of the endothelium and tunica media. 5 Inadequate invasion possibly results in reduced uteroplacental perfusion pressure (RUPP) and placental ischaemia/hypoxia. 6 Ischaemia-induced oxidative stress also leads to the apoptosis and necrosis of the placenta as well as the release of several placental factors, resulting in endothelial dysfunction, and systemic inflammation. 7 In normal pregnancies, placentas promote maternal tolerance to fetal antigens partly through producing immunosuppressive molecules, inhibiting chemokine expression and complement activation, entrapping professional antigen-presenting cells, and excluding immune cells. [8][9][10] However, in patients with PE, abnormal placentation is associated with an exaggerated systemic maternal inflammatory response, which contributes to adverse pregnancy outcomes. [8][9][10] Upon deciphering the human genome, large-scale investigations of non-coding RNAs (ncRNAs), including miRNAs and lncRNAs, have been conducted in many fields. MicroRNAs (miRNAs) are small (18-24 nt) endogenous non-coding single-stranded RNAs. MiRNAs regulate mRNA, which encode proteins that modulate cellular functions, by binding to the untranslated region (3′UTR) of target mRNAs. As a result, target gene expression can be regulated post-transcriptionally through degradation of target mRNAs. 11 Therefore, miRNAs play important roles in physiological homoeostasis in health and pathophysiological dysregulation in disease. 11,12 To date, more than 1000 miRNAs have been validated in humans, among which more than 600 miRNAs are expressed in the human placenta. 13 Abnormal expression of specific miRNAs has been reported in various studies of human diseases, including PE. 11 Comparative analysis between PE placentas and normal placentas identified 11 up-regulated and 23 down-regulated miRNAs. 14 Among them, Has-miR-19a was downregulated by 0.488-fold. Transcriptional downregulation of miR-19a was also observed under oxidative stress. 15 Overexpression of miR-19a protects endothelial cells from lipopolysaccharide-induced apoptosis 16 and enhanced cell invasion and metastasis. 17 These findings imply a putative implication of miR-19a in the physiological pathological mechanisms of PE.
LncRNAs (long non-coding RNAs) are non-coding RNAs that are longer than 200 nt in length. LncRNAs can regulate gene expression potentially through complementary binding or as coregulators. 18 Dysregulation of lncRNAs has been associated with many human diseases. 18 To better understand the expression of lncRNAs and their biological functions in PE, an expression profile of lncRNAs of PE placenta was sequenced by He et al. 19 In the GSE50783 dataset, lncRNA-PSG10P was significantly up-regulated in PE placentas compared to normal placentas. 19 Pregnancy-specific glycoproteins (PSGs) are the most abundant trophoblastic proteins in maternal blood during pregnancy. 20 PSG10P is a noncoding pseudogene.
Additionally, significant overexpression of IL-1 receptor accessory protein (IL1RAP) mRNA was also detected in the GSE50783 dataset.
Interleukin-1 (IL-1) family cytokines are potent components of the innate immune system. 19 Some IL-1 family members bind the IL-1 receptor 1 (IL-1R1), which then complexes with the IL-1RAP, triggering inflammatory signaling. 21,22 The interplay between IL1RAP and IL-1R1 is indispensable for transmission nb IL-1 signaling. Degradation of IL1RAP induced by ubiquitination interferes with the NF-κB activation induced by IL-1β and the inflammatory and immune responses. 23 Therefore, the critical role of IL1RAP in IL-1 signal transduction suggests the involvement of IL1RAP in various biological and pathological functions affected by IL-1 that represents an important pro-inflammatory cytokine.
Bioinformatic analysis revealed that PSG10P can promote the expression of IL1RAP by sponging miR-19a-3p. Many studies have been performed on the interactions among RNA molecules, such as interactions between lncRNA-miRNA and miRNA-mRNA. 24    3 days for both conditions. Cell viabilities were examined at 0, 12, 24, and 48 hours using a cell counting kit-8 (CCK-8, Dojindo Laboratories, Mashiki-machi, Japan). The relative cell viability was calculated by comparing cell number at later time points (12, 24, and 48 hours) to that at the starting point (0 hour).  | 855 culture medium was used in the subsequent cell culture. At 48 hours after scratching, microscopic images of the same area were captured.

| Transwell cell invasion assay
Cell migration was calculated based on the distance between the initial and final positions.

| RNA pull-down assay
An RNA pull-down assay was performed to selectively extract the The enrichment of PSG10P was quantified using qPCR as described above.

| Animal treatment
The animal experiment was approved by the Ethical Committee for

| Data and tissues collection
Non-invasive tail-cuff blood pressure (BP) was measured in conscious mice using a non-invasive automated sphygmomanometer (BP-98A, collected for RT-qPCR and western blot assays as described above and/or the following immunohistochemical analysis.

| Immunohistochemical staining
The placenta tissues were embedded in paraffin and sectioned into

| Statistical analysis
Data were analysed using SPSS 12.0 software (SPSS, Inc., Chicago, IL, USA). One-way analysis of variance followed by Scheffe's posthoc test was performed to evaluate statistical significance (P < 0.05).

| RESULTS
3.1 | Significantly higher PSG10P and IL1RAP levels but lower miR-19a-3P levels are present in PE placentas compared to levels in normal placentas GSE50783 datasets revealed a significantly higher expression level of PSG10 in PE placentas than in normal placentas (P < 0.05, Figure 1A1), 19 which was also identified in the present study using more tissue samples (P < 0.01, Figure 1A2), but PSG10 was not detectable in the serum from either patients with PE or healthy controls (data not shown). miR-19a-3p, which was previously reported to be down-regulated in PE placentas, 14 was predicted to bind PSG10. The present study not only identified the downregulation of miR-19a-3p in PE placentas (P < 0.01, Figure 1B1), but also found a dramatic reduction in the serum of patients with PE (P < 0.01, Figure 1B2). From GSE50783 datasets, we found that five up-regulated mRNAs (INHBA, IL1RAP, TREM1, PSG11, and RASEF) in PE placentas were targeted by miR-19a-3p. However, our PCR measurement only showed the up-regulation of INHBA (P < 0.01) and IL1RAP (P < 0.001, Figure 1C), with the greatest increase in IL1RAP. Figure 1D1 shows the expression profile of IL1RAP in GSE50783 datasets. IL1RAP protein level was further assessed using western blotting. In agreement with the mRNA expression analysis, IL1RAP protein levels were increased in the PE placentas compared to that in normal placentas (P < 0.01, Figure 1D2).  were evaluated in this study. MAP showed a significant correlation with all three genes/proteins (all P < 0.05) in patients with PE (Figure 1F), but the correlation in healthy controls was poor ( Figure 1E).
RT-qPCR analysis showed that transfection with shRNA-PSG10P reduced PSG10P expression in both HTR-8/SVneo and TEV-1 cells compared to that in the negative control (both P < 0.01, Figure 3A).
However, no change in the expression level was observed upon scramble treatment, confirming the successful knockout of PSG10P.
The bioinformatic tools (Targetscan, miRanda, and PicTar) predicted that PSG10P is a direct target of miR-19a-3p and showed a putative complementary region. In luciferase reporter assays, both wild-type and mutant-type of PSG10P constructs showed increased luciferase activity compared to that of the negative control (P < 0.001, Figure 4C). However, transfection with miR-19a-3p mimics significantly reduced the luciferase activity of the wild-type of PSG10P constructs (P < 0.01), but not that of the mutant type of PSG10P constructs, which confirmed the prediction of the target region. In addition, the interaction between miR-19a-3p and PSG10P was also identified by RNA pull-down assay: significantly elevated In a luciferase reporter assay, transfection with miR-19a-3p mimics significantly reduced the luciferase activity of the wild-type of PSG10P constructs, but not that of the mutant PSG10P constructs (C). RNA pulldown analysis indicated that PSG10P interacted with WT-hsa-miR-19a-3p, but not with MT-hsa-miR-19a-3p (D). *P < 0.05, **P < 0.01, and ***P < 0.001 relative PSG10P enrichment was only observed in wild-type miR-19a-3p, but not in the mutant ( Figure 4D).

| IL1RAP is a direct target of miR-19a-3p
The mimics of miR-19a-3p significantly increased its relative RNA expression, while the inhibitors significantly inhibited its expression ( Figure 5A). However, opposite expression behaviours were detected for IL1RAP; significantly lower relative RNA expression of IL1RAP was observed upon miR-19a-3p mimics treatment, but higher expression was observed in inhibitor treatment ( Figure 5B). The putative binding site of miR-19a-3p in the 3′-UTR region of IL1RAP is shown in Figure 5C. The transfection with miR-19a-3p mimics significantly reduced the luciferase activity of the wild-type of IL1RAP 3′UTR constructs (P < 0.01, Figure 5D), but not that of the mutant-type.

| Increased miR-19a-3p expression rescued PE symptoms in vivo
The mouse model of PE was established using sFlt-1. From day 13, MAP significantly increased and reached as high as 140 at day 18 in the PE pregnant mice (P < 0.01, Figure 7D), urinary protein concentrations in the PE pregnant mice were dramatically increased (P < 0.05, Figure 7E), and PE pregnant mice showed reduced fetal survival number (P < 0.05, Figure 7F). miR-19a-3p expression was decreased in the placenta of PE pregnant mice (P < 0.05, Figure 7A).
In the luciferase reporter assay, transfection with miR-19a-3p mimics significantly reduced the luciferase activity of the wild-type of IL1RAP constructs, but not that of the mutant IL1RAP constructs (B). **P < 0.01 known that sFlt-1 can cause vascular endothelial dysfunction and further induce many abnormal symptoms commonly observed in PE, such as oxidative stress, imbalanced inflammatory response, hypotension, and proteinuria. 26,27 The relationship between PSG10P and miR-19a-3p was identified in this study by a series of analyses. MiR-19a was reported to significantly regulate cell proliferation and migration. miR-19a-3p inhibits invasion, migration, and bone metastasis in prostate cancer, 28  IL-1, a pleiotropic cytokine, can activate the transcription of numerous genes and synthesis of a range of proteins involved in immunological responses. 30 The critical role of IL1RAP in IL-1 signal transduction suggests the involvement of IL1RAP in various inflammatory and immune responses affected by IL-1. A previous report showed that a higher expression level of IL1RAP was associated with increased neutrophil counts in the airways of asthmatic patients, suggesting enhanced neutrophilic inflammation. 31 The expression level of IL1RAP is likely associated with uterine sensitivity to IL-1 during late gestation and delivery. 32 IL1RAP is abnormally upregulated in the placentas of pregnant women with sickle cell anaemia [33]. Sickle cell disease is often accompanied with acute and chronic inflammation [33]. Our results also showed significantly higher IL1RAP levels in PE placentas than in normal placentas. We found that both PSG10P and IL1RAP were targets of miR-19a-3p.
The presence of PSG10P could potentially sequester miR-19a-3p, resulting in accumulated free IL1RAP in patients with PE or associated transfected cell lines. Therefore, the higher PSG10P level could be a potential contributor to the lower level of miR-19a-3p. As a result, IL1RAP was up-regulated, promoting the systemic maternal inflammatory response, which is characteristic in patients with PE.
Additionally, IL1RAP also involved in regulating cell growth and invasion properties, as demonstrated by our in vitro study. Caspase-3 is a caspase protein with a significant role in cell apoptosis, while MMP9 is involved in degrading the extracellular matrix, resulting in altered migration and invasion ability. A higher level of IL1RAP was always associated with more active caspase-3 and lower MMP9, indicating that cell death was promoted and cell migration was F I G U R E 7 Increased miR-19a-3p expression rescued pre-eclampsia (PE) symptoms in a mouse model. A mouse model of PE was established using sFlt-1. Among 36 PE mice, 12 mice were randomly selected for mmu-miR-agomir treatment as a negative control (PE+NC), 12 for mmu-miR-19a-3p agomir treatment (PE+Agomir), and 12 as a PE control (PE). In addition, another 12 normal pregnant mice were orally treated with saline solution as a control (Nor). (A) miR-19a-3p expression in placentas was assessed via RT-PCR. (B) Protein levels of IL1RAP, caspase-3, and MMP-9 in placentas were assessed via western blot assay. (C) Ki67 expression in placentas was assessed via immunohistochemical staining. (D) The blood pressure, (E) urine protein concentrations, and (F) fetal survival number of the mice were recorded. *P < 0.05, **P < 0.01 vs Normal group; # P < 0.05, ## P < 0.01 vs PE group inhibited. As mentioned, the capabilities of trophoblasts, including proliferation, migratory, and apoptosis properties, are essential for adequate placental formation and facilitating normal pregnancy. Our results strongly demonstrated that lncRNA-PSG10P/miR-19a-3p/ IL1RAP cooperate with each other as a regulatory network to manipulate CTBs during pregnancy.
There are some shortcomings of the present study. The potential effects of up-regulated IL1RAP on the inflammatory response in patients with PE have not been discussed in the present study, although the pro-inflammatory role of IL1RAP has been shown in other pathological models. In addition, the underlying mechanisms by which IL1RAP modulates the viability, migration, and invasion of trophoblasts under hypoxemia remain unclear. Therefore, all these questions should be answered in a further study.
In summary, miR-19a-3p, lncRNA-PSG10P, and IL1RAP are involved in PE pathogenesis. With a common targeting region in their sequences, a regulatory network in the lncRNA-PSG10P/miR-19a-3p/IL1RAP pathway may contribute to PE pathogenesis during pregnancy.

DISCLOSURE OF INTERESTS
The authors declare that they have no conflict of interest.