The aberrantly expressed miR-193b-3p contributes to preeclampsia through regulating transforming growth factor-β signaling

Preeclampsia (PE) is a leading cause of maternal mortality worldwide. Several studies have detected some differentially expressed microRNAs in the preeclamptic placenta, but few of the identified microRNAs demonstrated consistent findings among different research studies. In this study, high-throughput microRNA sequencing (HTS) of 9 preeclamptic and 9 normal placentas was performed. Seventeen microRNAs were identified to be up-regulated, and 8 down-regulated in preeclamptic placentas. Eight differentially expressed microRNAs except one identified in our study were determined to be consistent with at least one previous study, while sixteen were newly found. We performed qRT-PCR with independent 22 preeclamptic placentas and 20 control placentas to verify the differentially expressed microRNAs, and ten microRNAs were validated. The predicted target genes of the aberrantly expressed miR-193b-3p were enriched in the following gene ontology categories: cell motility and migration, cell proliferation and angiogenesis. We also found that miR-193b-3p significantly decreased the migration and invasion of trophoblast (HTR-8/SVneo) cells and that miR-193b-3p could regulate trophoblasts migration and invasion through binding onto the 3′UTR target site of TGF-β2. In conclusion, we identified a list of differentially expressed microRNAs in PE placentas by HTS and provided preliminary evidence for the role of miR-193b-3p in the pathogenesis of preeclampsia.

translational interference 8 . MiRNAs are involved in regulating trophoblast proliferation, apoptosis, migration and invasion, and have been suggested to play an important role in the regulation of placental development and function 9 . Several studies based on hybridization, RT-qPCR or sequencing analysis have detected a list of differentially expressed miRNAs in PE placenta. However, only limited number of these identified miRNAs demonstrated consistent findings among these studies. For instance, six miRNA profiling research analyses showed that miR-210 was aberrantly expressed [10][11][12][13][14][15] and three showed that miR-193b-3p consistently up-regulated in PE placentas [10][11][12] . Efforts have been made to identify how differentially expressed miRNAs contribute to the onset of PE. Evidence suggested that dysfunction of miRNAs could inhibit migration and invasion of human extravillous trophoblast-derived HTR-8/SVneo cells 16,17 . However, it is not yet the right time to use these findings from miRNA studies on preeclampsia to improve the management or early recognition of this condition. Further investigation is required to elucidate novel mechanisms underlying the molecular pathology of PE by specific miRNAs which have not been investigated, like miR-193b-3p.
High-throughput sequencing (HTS) platforms perform well in reproducibility, accuracy, specificity and sensitivity for quantitatively measuring RNA. In addition, the detection sensitivity has been shown to be increased along with increasing sequencing depth, which is superior to hybridization-based platforms, whereas RT-qPCR is limited by the number of microRNAs 18 21 also performed miRNA profiling of PE placentas using sequencing platforms. However, three 12,19,21 performed the research based on limited sample sizes, and one 20 conducted the research with limited sequencing depth.
In the present study, we performed HTS analysis on 9 preeclamptic and 9 normal placentas with enough sequencing depth to explore the differentially expressed miRNAs and further investigate their biological roles in the development of PE. Our study demonstrated that the expression of miR-193b-3p was significantly up-regulated in PE placentas, and it significantly decreased the migration and invasion of HTR-8/SVneo cells. Furthermore, we found that transforming growth factor-beta 2 (TGF-β2) was post-transcriptionally dysregulated by miR-193b-3p and confirmed that miR-193b-3p could regulate trophoblast cell invasion and migration by targeting TGF-β2.

Results
Characteristics of microRNA expression by HTS. We performed HTS on 9 preeclamptic placentas (P1-P9) and 9 normal placentas (control group, C1-C9). We analyzed miRNAs with ≥ 5 sequence reads and constructed a scatter plot of these 787 miRNAs with the normalized reads (per million). Comparing global miRNA expression profiles in placentas of preeclamptic patients with those of normotensive controls, we could intuitively observe the distribution of expression of the two groups (Fig. 1A). The majority of miRNAs were not significantly differentially expressed between control and preeclampsia subjects, and the expression level of the majority of  Table S1).
Differentially expressed microRNAs in placentas of patients with preeclampsia. A total of 25 miRNAs were significantly deregulated, 17 of which were up-regulated and 8 were down-regulated in preeclampsia (Table 1). Cluster analysis based on 46 selected microRNAs (p value ≤ 0.05; fold change ≥ 1.5 or ≤ 0.66) generated a tree with a clear distinction between the preeclampsia and control groups ( Fig. 2A).
Consistency of the identified differentially expressed microRNAs between the present study and previous reports. Using a combination of the differentially expressed microRNAs from our study and the results based on previous array analysis, we identified a complete list of 182 differentially expressed miRNAs. Forty-four differentially expressed miRNAs are present in Supplementary Table S2. These miRNAs were identified in at least two studies, and most of the studies on the corresponding miRNAs reported consistent findings. MiR-210, which was reported in 8 studies, was the most common recurring miRNA among the 44 differentially expressed miRNAs. Indeed, the function of miR-210 in preeclampsia has also been widely investigated compared to other miRNAs 12,22,23 . The second most common recurring miRNA was miR-193b-3p, which was presented in 4 studies and consistently showed up-regulation in preeclampsia. However, to our knowledge, no research on the function of miR-193b-3p in preeclampsia has been conducted to date. Although there are additional 2 miRNAs (miR-181a-5p and miR-363-3p) present in 4 studies, they were not found to be differentially expressed in our study. Focused on our own study, 8 differentially expressed miRNAs (miR-210, miR-193b-3p, miR-31-5p, miR-193b-5p, miR-10b-5p, miR-151-3p, miR-520-3p and miR-584-5p) were found to be consistent with at least one previous study. One miRNA, miR-192-5p, which demonstrated differential expression in our HTS analysis, was also present in two previous studies with the opposite findings. It should be noted that miR-192-5p showed no differential expression in our qRT-PCR validation in an independent cohort. The remaining differentially expressed miRNAs (16 miRNAs) from our study were newly found. Effect of miR-193b-3p on HTR-8/SVneo migration and invasion. We focused on microRNAs that demonstrated consistent findings among a number of independent studies. MiR-193b-3p was the second most common recurring miRNA after miR-210. We inferred that miR-193b-3p might have an important role in the progression of preeclampsia. We found that the genes predicted to be targets of miR-193b-3p by TargetScan were enriched in the following gene ontology categories: cell motility and migration, cell proliferation, positive regulation of nitrogen compound metabolic process and angiogenesis (Fig. 3A). In addition, we noticed that the most significantly enriched functional annotation cluster among these 221 targets of miR-193b-3p was cell motility, cell migration and cell motion (Fig. 3B). This result suggests that miR-193b-3p directly or indirectly regulates genes involved in cell migration.
We then evaluated the effects of miR-193b-3p on the migration of a trophoblast (HTR-8/SVneo) cell line with in vitro scratch assay (Fig. 4B) and transwell migration assay (Fig. 4C). We evaluated the efficiency of miR-193b-3p overexpression constructs (Pre-miR-193b-3p) and miR-193b-3p inhibitor (Fig. 4A). The fold-change of relative miR-193b-3p expression after transfecting Pre-miR-193b-3p in HTR-8/SVneo was 2.54. This fold-change and  46 differentially expressed microRNAs were selected using the following criteria: q value ≤ 0.05; fold change ≥ 1.5 or ≤ 0.66. The relative expression level was log transformed. Each column represents a placental sample from a pregnant control without preeclampsia (C1-C9) or with preeclampsia (P1-P10). The colour bar in the upper left indicates high expression (red) or low expression (green) (A). Validation of differentially expressed microRNAs by qRT-PCR. 2 down-regulated microRNAs (B) and 8 up-regulated microRNAs (C) are validated by qRT-PCR with independent control placentas (n = 20) and PE placentas (n = 22). The relative expression of each unique miRNA was normalized by the value of the U6 gene. We normalized the control as 1 by dividing each value by the average value of the control group. PE/Control-HTS indicates the fold change of PE placental microRNA expression level relative to the control as demonstrated by HTS. The mean raw Δ CT (CT ,miRNA − CT ,U6 ) values of each miRNA were shown on top of the bars. Data are shown as the mean ± S.E.M. Significant differences were determined by 2-tailed Student's t test or the Mann-Whitney test. ***P < 0.001, **P < 0.01, *P < 0.05. the relative miR-193b-3p expression in HTR-8/SVneo were similar to the result from our miRNA profiling of PE and control placentas by HTS. Transfecting miR-193b-3p inhibitor decreased the expression of miR-193b-3p in HTR-8/SVneo by 95%. As shown in our in vitro scratch assay, miR-193b-3p significantly decreased the migration of HTR-8/SVneo by 36% compared to the corresponding negative control (P < 0.001, Fig. 4B) and inhibition of miR-193b-3p significantly improved the migration of HTR-8/SVneo by 15% compared to the corresponding negative control (P < 0.05, Fig. 4B). As shown in transwell migration assay, miR-193b-3p significantly decreased the migration of HTR-8/SVneo by 37% compared to the corresponding negative control (P < 0.001, Fig. 4C) and inhibition of miR-193b-3p significantly improved the migration of HTR-8/SVneo by 19% compared to the corresponding negative control (P < 0.01, Fig. 4C).

TGF-β2 was identified as a direct target of miR-193b-3p.
We re-investigated the results of our previous study concerning protein-coding gene expression profiles in PE placentas 24 . We found that TGF-β2 was the only differentially expressed gene among the 16 predicted miR-193b-3p targets involved in the functional annotation cluster of cell motility, cell migration and cell motion (Fig. 3B).
We then measured TGF-β2 levels by qRT-PCR in all placental tissues and found that TGF-β2 mRNA was significantly down-regulated in PE placentas (P < 0.01, Fig. 5C). A significant negative correlation between the expression of miR-193b-3p and TGF-β2 mRNA was observed (P < 0.001, Fig. 5D). We also found that the TGF-β 2 protein level was approximately 65% lower in PE placentas ( Supplementary Fig. S1).

Discussion
In the present study, we detected a list of differentially expressed microRNAs in PE placentas by HTS. Eight differentially expressed miRNAs from our study were found to be consistent with at least one previous study and 16 microR-NAs were newfound. However, only limited number of miRNAs demonstrated consistent results among different research studies. The inconsistency might be caused by the different clinical characteristics of patients enrolled in these different studies. This also might be due to the application of different experiment platforms and different sample sizes 18,25 . Although our sample size is relatively larger than previously published studies on the same topic, the small cohort may still be a limitation of the present study. Different experiment platforms have apparent differences in performance in reproducibility, accuracy, detection sensitivity and specificity 18 . Our follow-up study was focused on miRNAs that demonstrated consistent results among a number of independent studies. For instance, miR-210 is the most commonly identified differentially expressed miRNA in placental. Indeed, miR-210 is one of the most investigated miRNAs in PE. It is a hypoxia-responsive miRNA involved in trophoblast migration and invasion, siderosis of interstitial trophoblasts and placental mitochondrial dysfunction 22,26,27 . Up-regulation of placental miR-210 was also detected in our research.
MiR-193b-3p is the second most common recurring miRNA after miR-210. The findings in 3 previous studies and ours are consistent, demonstrating that miR-193b-3p is up-regulated in preeclamptic placentas. Previous publications suggested that miR-193b-3p was a putative tumor suppressor. This gene was found to be greatly down-regulated in several types of malignant tumors 28,29 . Moreover, overexpression of miR-193b-3p repressed cell proliferation and inhibited cancer cells invasion, migration and growth 29,30 . Typically, this was reported in the tumour microenvironment, but what about the placental environment? During placental development, cytotrophoblasts proliferate to form anchoring villi and give rise to interstitial trophoblasts that invade the uterine decidua and endovascular trophoblasts that migrate into the maternal spiral arteries; accordingly, the trophoblasts' capacity for proliferation, migration and invasion plays a crucial role in successful placental development 31 . It appears that trophoblasts and cancer cells share similar biological characteristics in their proliferative, migratory and invasive properties 32,33 . Adequate invasion of tumour cells leads to cancer, but inadequate invasion of trophoblasts might result in poor placentation, leading to preeclampsia 34 . Consequently, the tumour suppressor role of miR-193b-3p is suggestive of its crucial role in the regulation of some principal trophoblasts events, leading to preeclampsia. In silico target prediction of miR-193b-3p by TargetScan and the following GO analysis showed that targets genes of miR-193b-3p were enriched in cell motility and cell migration. In addition, our in vitro scratch assay and transwell migration and invasion assays demonstrated that overexpression of miR-193b-3p significantly decreased the migration and invasion of HTR-8/SVneo cells, while inhibition of miR-193b-3p significantly improved the migration and invasion of HTR-8/SVneo cells. Our results indicated that the aberrant expression of miR-193b-3p in the preeclamptic placenta could contribute to the pathogenesis of preeclampsia through its inhibitory effect on the migration and invasion of trophoblasts.
We also showed that transforming growth factor-beta 2 (TGF-β2), a member of the transforming growth factor-β (TGFβ ) superfamily, was post-transcriptionally dysregulated by miR-193b-3p. And restoring TGF-β2 expression reversed the regulation of miR-193b-3p on trophoblast (HTR-8/SVneo) cells migration and invasion. Meanwhile, TGF-β2 mRNA and protein expression were significantly decreased in preeclamptic placentas. According to previous research, the TGFβ signalling pathway is crucial to placental biological processes 35 . It is involved in regulating a number of cellular processes, including growth, migration, invasion, epithelial to mesenchymal transition (EMT) and immune reactions 36,37 . For instance, the cells undergoing EMT manifest a migratory phenotype and acquire invasive properties 38 , and the differentiation of interstitial trophoblasts and endovascular trophoblasts is considered to be an EMT-like transformation process 39 . Previous studies suggest that TGFβ is potent inducer of EMT in various cells 40,41 and showed that the TGFβ family induced the invasive properties of a trophoblast cell line in vitro 42 . And one study showed that cell migration was improved by inducing TGF-β2 expression under hypoxia 43 . In addition, soluble endoglin (sEng), which is well known to be involved in preeclampsia, was also identified as a TGFβ inhibitor in previous studies 44,45 . Therefore, we inferred that aberrant TGFβ signalling in the placenta might play a key role in the pathogenesis of PE; aberrant expression of miR-193b-3p in preeclamptic placentas could exert an inhibitory effect on the migration and invasion of trophoblasts by directly targeting TGF-β2. The detailed mechanism remains to be elucidated.
In conclusion, we detected and verified a list of differentially expressed microRNAs in PE placentas by HTS and qRT-PCR, and provided preliminary evidence for the role of miR-193b-3p in the pathogenesis of preeclampsia by targeting TGF-β2.

Methods
Sample collection. All placenta samples used in this study were collected from normal pregnant woman who were Han Chinese in origin (control group; n = 29) and from preeclamptic pregnant patients (PE group; n = 31) by elective caesarean delivery in the absence of labour during the third trimester of gestation (Table 2, All results are presented as the mean ± S.D.; BMI, indicates body mass index; DBP, diastolic blood pressure; and SBP, systolic blood pressure). Detailed clinical characteristics of patients with 9 normal and 9 preeclamptic pregnancies in HTS study were shown in Supplementary Table S3. PE was defined as systolic blood pressure (SBP) ≥ 140 mmHg and/or diastolic blood pressure (DBP) ≥ 90 mmHg on 2 occasions at least 4 hours apart with proteinuria ≥ 300 mg/day from 24 h urine collection occurring after the 20th week of gestation but resolving by the 12th week postpartum. All clinical placentas were collected immediately after the caesarean section. The placental samples at the chorionic plate were separately taken from each quadrants along with central portion in the placenta disc. ~1 cm 3 fragments were dissected from the placenta, after removal of maternal blood by vigorous washing in phosphate buffered saline (PBS). The tissues were maintained in centrifuge tubes with RNAlater (Ambion Inc., Austin, TX), and then frozen at − 80 °C. All the patients did not accepte antihypertensive medication and other special medical treatment before termination of pregnancy. All women enrolled in the study gave their written informed consent to the sample collection and analyses. This research was approved by the Institutional Review Committee of Institutes of Biomedical Sciences at Fudan University. All experiments were performed in accordance with relevant guidelines and regulations.
MiRNA isolation and sequencing analysis. Total RNA was extracted using the Ambion mir-Vana ™ PARIS ™ Kit (Ambion , Carlsbad, California, USA) according to the manufacturer's instruction. RNA concentration and quality were determined with Agilent 2100 Bioanalyser (Agilent, Palo Alto, CA). 1 ug total RNA was obtained for the small RNA sequencing analysis and underwent further processing with the TruSeq Small RNA Sample Preparation Kit. Briefly, 1 ug total RNA underwent adapter ligation, reverse transcription, PCR amplification and pooled gel purification to generate a library product. Following this, we performed microRNA-sequencing using the Illumina HiSeq 2000 sequencing system with approximately 6 multiplexed samples in each sequencing lane. The raw Illumina reads were pre-processed and mapped. The number of mapped reads was an average of 25.8 million in the control group and 28.4 million in the PE group. Using R version 2.14.0 statistical software, we screened the differentially expressed microRNAs based on the following criteria: p value ≤ 0.01; fold change ≥ 1.5 or ≤ 0.66.

Validation of mRNA and miRNA by qRT-PCR.
A total of 500 ng RNA was isolated for qRT-PCR validation. We performed real-time quantification of miRNAs by stem-loop RT-PCR based on previous studies 46  A total of 500 ng RNA was isolated for mRNA qRT-PCR. The cDNA was synthesized using AMV Reverse Transcriptase system (Promega, Madison, Wisconsin, USA). TGF-β2 mRNA qRT-PCR was performed using the Applied Biosystems 7900HT with FastStart Universal SYBR Green Master (Rox) (Roche, Penzberg, Germany). The primers for TGF-β2 mRNA qRT-PCR are described in Supplementary Table S4.
The qRT-PCR data was analysed using the Δ Δ Ct method.

Literature Review.
We performed a search of literature published to October 2014 through PubMed to identify all articles reporting the differential expression of miRNAs in preeclampsia. In the 80 identified studies by using the the keywords "microrna" and "preeclampsia", 33 were studies of human preeclamptic placentas compared to normotensive controls, which included 12 array studies based on hybridization, RT-qPCR or sequencing analysis that presented the results as a list of differentially expressed miRNAs [10][11][12][13][14][15]19,21,[47][48][49][50][51][52][53] and 21 studies that focused on specific miRNAs.
In vitro scratch assay. An in vitro scratch assay was used as described previously 54 . The protocol of cell culture and transient transfection is described previously. When HTR-8/SVneo cells were grown to 90% confluence, the scratches were made. In order to minimize cell proliferation, the cells were allowed to grow in RPMI-1640 (Gibco, California, USA) without foetal bovine serum. The width of the scratch was monitored by LEICA DMI4000B within 24 h after transfection.
Transwell migration assay. The protocol of cell culture and transient transfection is described previously. Transwell compartments were prepared with 24-well format and 8 μ m pore size insert. For the lower compartment, we added 0.8 ml of RPMI-1640 (Gibco, California, USA) with 10% FBS; and for the upper compartment, we gently added 5 × 10 4 cells in 100 μ l serum-free RPMI-1640 (Gibco, California, USA). We incubated the cells in the transwell plate at 37 °C and 5% CO2 for 6 h. After 6 h, we carefully took the insert out. Cells on the upper surface of membranes were completely removed. and the cells migrated to the lower surface of membranes were fixed with 100% formaldehyde and stained with 0.05% crystal violet. The number of migrated cells was counted under LEICA DMI4000B. Cell migration level was presented as the percentage of migrated cell number compared with the corresponding control.
Transwell invasion assay. The protocol of cell culture and transient transfection was described previously. We prepared for the transwell 24-well fitted inserts (8 μ m pore size, Millipore Corp, Massachusetts, USA). The transwell inserts were pre-coated with 7 μ g matrigel (BD Biosciences, NJ, USA) in 50 μ l serum-free culture medium. For the lower compartment, we added 0.8 ml of RPMI-1640 (Gibco, California, USA) with 10% FBS. For the upper compartment, we gently added 5 × 10 4 cells in 100 μ l serum-free RPMI-1640 (Gibco, California, USA). We incubated the cells in the transwell plate at 37 °C and 5% CO2 for 24 h. After 24 h, we carefully took the insert out. Cells on the upper surface of membranes were completely removed, and the cells migrated to the lower surface of membranes were fixed with 100% formaldehyde and stained with 0.05% crystal violet. The number of invaded cells was counted under LEICA DMI4000B. Cell invasion level was presented as the percentage of invaded cell number compared with the corresponding control.
Western blotting. Placental tissues were homogenized in lysis buffer with added proteinase inhibitors (Roche, Penzberg, Germany). A BCA kit (Thermo Fisher Scientific, Waltham, Massachusetts, USA) was used to assess protein concentrations. 20 ug of proteins were resolved by 10% SDS-PAGE and transferred to a nitrocellulose membrane (Applied Biosystems, Waltham, Massachusetts, USA). The membranes were incubated with the primary antibodies against TGF-β 2 (Abcam, Cambridge, UK, 1:500) and GAPDH (Abmart, Shanghai, China, 1:1000) overnight at 4 °C. Following this, the membranes were incubated with horseradish peroxidase (HRP) conjugated secondary antibodies (Abmart, Shanghai, China, 1:3000). The signal intensity was quantitatively assessed with Fujifilm LAS-3000 Imager and was normalized to GAPDH. Statistical analysis. Analysis of microRNA-seq data was performed using R version 2.13.1. A 2-tailed t test was performed; and Mann-Whitney test was performed with SPSS 17 (SPSS Inc., Chicago, Illinois, USA) when the data pattern did not follow Gaussian distribution. The results are shown as the mean ± S.E.M. or mean ± S.D. A threshold of P value < 0.05 was considered to be significant. Correlation was calculated using Pearson correlation coefficient.