The placental samples, from 25 women with severe preeclampsia and 25 with normal pregnancy, were obtained from the Department of Gynecology and Obstetrics, Yangzhou Women and Children Hospital (Yangzhou, China) between October 2014 and November 2016. The clinical data of the pregnant women are presented in Table I. Chorionic tissue blocks (~1 cm³) were collected from the mother surface of the placenta and near the root of the umbilical cord. Tissues were washed with sterile PBS, then immediately frozen in liquid nitrogen at the time of surgery and stored at −70°C until needed. This study was approved by the ethics board of Yangzhou Women and Children Hospital and informed consent was obtained from all 50 participants. The present study was approved by the ethics board of Yangzhou Women and Children Hospital and informed consent was obtained from all participants. All clinical investigations were conducted according to the principles of the Declaration of Helsinki.

Total RNA was extracted using TRIzol (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) and QIAGEN miRNeasy Mini kit (Qiagen GmbH, Hilden, Germany), quantitated on the NanoDrop 2000 (Thermo Fisher Scientific, Inc., Wilmington, DE, USA), labeled using the FlashTag Biotin HSR RNA Labeling kit (Affymetrix, cat no. 901911; Thermo Fisher Scientific, Inc.) and hybridized onto the GeneChip2 Hybridization (Affymetrix, cat no. 902413; Thermo Fisher Scientific, Inc.). Following the washing steps, the slides were scanned using the GeneChip2 Scanner 3000 7G (Affymetrix; Thermo Fisher Scientific, Inc.). All expressed data were normalized using the median normalization method, and significantly differentially expressed miRNAs were identified through volcano plot filtering. Data were analyzed by GCBI online software (GCBI, R3.3.1, http://www.gcbi.com.cn; Genminix Informatics Co., Ltd., Shanghai, China). RT-qPCR analysis for miR-337-3p, miR-187-3p, miR-122-5p, miR-26b-5p and miR-144-3p was performed with a miDETECT A TRACK™ miRNA qRT-PCR Starter kit (RiboBio Co., Ltd., Guangzhou, China). The miDETECT A TRACK™ Uni-RT primer was used for RT (Guangzhou RiboBio Co., Ltd., Guangzhou, China). The miDETECT A TRACK™ Forward Primer and Uni-Reverse Primer were used for qPCR (Guangzhou RiboBio Co., Ltd.). The reactions were incubated in a 96-well plate at 95°C for 10 min, followed by 40 cycles at 95°C for 2 sec, 60°C for 20 sec and 70°C for 10 sec. The miRNA Primers used in RT-qPCR were synthesized by RiboBio, Co., Ltd., and the primer sequences are commercially restricted. The expression of miRNA was normalized to the small nuclear RNA U6 expression as an endogenous control. The fold changes of mRNA expression were calculated by the 2^−ΔΔCq method.

The HTR-8/SVneo cells were maintained in the laboratory of Department of Gynecology and Obstetrics, the Affiliated Drum Tower Hospital of Nanjing University Medical School (Nanjing, China). The cells were cultured in RPMI-1640 (HyClone; GE Healthcare Life Sciences, Logan, UT, USA), supplemented with 10% heat-inactivated fetal bovine serum (Wisent, Inc., St. Bruno, Quebec, Canada), 100 U/ml penicillin and 100 mg/ml streptomycin, at 37°C in a humidified 5% CO₂ incubator.

The bioinformatics software TargetScan (version 7.0; http://www.targetscan.org/) (11) was used to predict the target genes of miR-144-3p. The Cox-2 3′untranslated region (3′UTR)-hRLuc reporters (PTGS2 WT) were created by ligation of the Cox-2 3′UTR PCR product into the XhoI and NotI site of the dual-luciferase reporter vector (pmiR-RB-REPORT™; RiboBio Co., Ltd.). The mutant reporters for the Cox-2 3′UTR (PTGS2 MT; CAUUUAAUGGGUACUGUAU) were generated by replacing the binding sites of miR-144-3p. Subsequently, HTR-8/SVneo cells were cultured in 96-well plates (1.5×10⁴) and transfected with the wild or mutant luciferase reporters (100 ng), and miR-144-3p mimics (50 nM) using Lipofectamine^® 2000 reagent (Invitrogen; Thermo Fisher Scientific, Inc.). All reactions were run in triplicate. After 48 h, luciferase activities were measured using the dual-luciferase reporter system (Promega, Madison, WI, USA). Normalized luciferase activity was reported as Renilla luciferase activity/luciferase activity.

To detect the expression levels of human miR-144-3p in placenta tissue, ISH was performed using a miRCURY locked nucleic acid (LNA) detection probe for miR-144-3p (hsa-miR-144-3p/DigN/AGTACATCATCTATACTGTA) (Exiqon; Qiagen GmbH; probe concentration 25 µM). The digoxigenin double-labeled LNA-modified probe, at a final concentration of 500 nmol/l, was added to the hybridization solution and hybridized according to the manufacturer's protocol. The sample images were captured using a light microscope at 100× magnification after the incubation of slides with 4-nitroblue-tetrazolium for 30 min at 25°C and with nuclear fast red for 5 min at room temperature (Roche Applied Science, Indianapolis, IN, USA). The staining intensity and the proportion of miR-144-3p-positive cells were detected in 50 placenta samples (25 from preeclampsia women and 25 from women without preeclampsia). A total of 10 fields of view in each slide was randomly selected for analysis. The 4 intensity grades of staining cells were negative, weak, intermediate and strong, with corresponding scores of 0, 1, 2 and 3 respectively. Another score was given for the proportion of positive cells with the scoring rule as follows: No positive cells (0 point), 1–24% positive cells (1 point), 25–49% positive cells (2 points), 50–74% positive cells (3 points), 75–100% positive cells (4 points). The staining index (SI) was calculated by multiplying the staining intensity and the proportion of positive cells. According to the definition of SI scores, low expression of miR-144-3p (1.0) was described as SI score ≤4, high expression of miR-144-3p (3.0) was described as SI score >8 and intermediate (2.0) >4 and ≤8.

Immunohistochemical staining was performed using an immunohistochemistry detection kit (cat no. PV-6000; OriGene Technologies, Inc., Beijing, China) according to the manufacturer's instructions. A rabbit polyclonal antibody for Cox-2 (1:1,000; ab15191, Abcam, Cambridge, UK) was used as primary antibody. A goat anti-rabbit secondary antibody (1:1,000; A0208; Beyotime Institute of Biotechnology, Beijing, China) was used as the secondary antibody. The expression of Cox-2 was evaluated using a scoring system by the combining the staining intensity score and proportion of positive cells. Quantification of the staining intensity of Cox-2 was performed using image analysis as with ISH. A total of 10 fields of view in each slide was randomly selected for analysis.

HTR-8/SVneo cells with stable miRNA-144-3p expression and parental cell lines were established using a lentiviral expression system. The lentiviruses containing MiR-144-3p (MIMAT0000436: GAGCAGGGAGCAGGAAGCTGTGTGTGTCCAGCCCTGACCTGTCCTGTTCTGCCCCCAGCCCCTCACAGTGCTTTTCAAGCCATGCTTCCTGTGCCCCCAGTGGGGCCCTGGCTGGGATATCATCATATACTGTAAGTTTGCGATGAGACACTACAGTATAGATGATGTACTAGTCCGGGCACCCCCAGCTCTGGAGCCTGACAAGGAGGACAGGAGAGATGCTGCAAGCCCAAGAAGCTCTCTGCTCAGCCTGTCACAACCTACTGACTGCCAGGGCACTTGGGA) or negative control were constructed by GeneChem Co., Ltd., (Shanghai, China). All the lentiviral vectors expressed enhanced green fluorescent protein and puromycin resistance gene, which allowed the measurement of the infection efficiency and to select for infected cells. Cells (5×10⁶/15 ml) were infected with Vector or MiR-144-3p lentivirus (15 µg), followed by culture in 5 µg/ml puromycin (Clontech Laboratories, Inc., Mountainview, CA, USA) to select stable lines. The resulting stable lines were used for further analysis.

Cells were collected and homogenized in lysis buffer containing 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.1% SDS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.02% sodium azide, 100 mg/l phenylmethylsulfonyl fluoride and 1 mg/l aprotinin. The supernatant was collected after centrifuging at 11,000 × g for 15 min at 4°C. Protein concentration was determined using the bicinchoninic acid method (Pierce; Thermo Fisher Scientific, Inc.). SDS-PAGE using 10% gels was performed to separate protein (50 µg/lane) and separated proteins were then transferred to polyvinylidene difluoride membranes. PVDF membranes were blocked with 5% non-fat milk for 1 h at room temperature and incubated with primary antibodies against Cox-2 (1:1,000; ab15191; Abcam), GAPDH as a loading control (1:1,000; sc-32233; Santa Cruz Biotechnology, Inc., Dallas, TX, USA) and Cox-1 as a specific control (1:1,000; ab109025; Abcam) overnight at 4°C. After washing, the membranes were incubated with goat anti-rabbit secondary antibody (1:5,000; cat no. sc-2004; Santa Cruz Biotechnology, Inc.) for 2 h at room temperature and visualized with enhanced chemiluminescence (Millipore, Billerica, MA, USA) by Quantity one software (version 62, Bio-Rad Laboratories, Inc., Hercules, CA, USA).

Results were analyzed using the SPSS 22.0 statistical software (IBM Corp., Armonk, NY, USA). Experimental data are presented as the mean ± standard deviation. The differences for multiple groups were estimated using one-way analysis of variance followed by Dunnett's T3 post hoc test, and the association between miR-144-3p expression and the Cox-2 level was analyzed by Spearman correlation analysis. The independent two-tailed Student's t-test was performed for comparisons between two groups. P<0.05 was considered to indicate a statistically significant difference.

Firstly, miRNA microarray analysis of samples from six pregnant women (three from normal and three from preeclampsia cases) and the expression of miRNAs in paired placentas were compared to identify differential expressed miRNAs. The inclusion criteria of differential miRNAs was differential expression of ≥2.0 fold changes. In total, 46 miRNAs were downregulated and 45 were upregulated in the six samples (Fig. 1). Subsequently, the five miRNAs with the highest fold changes (differential expression of >4.0 fold changes), including miR-337-3p, miR-187-3p, miR-122-5p, miR-26b-5p and miR-144-3p, were selected to verify the results of the microarrays. Notably, RT-qPCR validation demonstrated that the fold changes in the five miRNAs expression were 2.99, 15.37, 24.91, 0.87 and 0.016, respectively, in preeclampsia placentas compared with the paired normal placentas (Fig. 1). Ultimately, miR-144-3p was selected as the focus of further research as it exhibited the greatest fold change.

Cox-2, also known as prostaglandin-endoperoxide synthase 2 (PTGS2), was selected as a candidate miR-144-3p target based on bioinformatics analysis (TargetScan) and functional knowledge. The 3′UTR of Cox-2 has a complementary site for the miR-144-3p seed sequence at its 3′UTR (Fig. 2A and B), and it is highly conserved among species. To further determine whether Cox-2 is a direct target of miR-144-3p, a luciferase reporter construct harboring a fragment of the Cox-2 3′UTR that contained the miR-144-3p binding site was prepared (PTGS2 WT). A construct with mutated miR-144-3p binding site in the Cox-2 3′UTR (PTGS2 MT) was also generated. Luciferase activity was decreased by 46.5% in HTR-8/SVneo cells co-transfected with Cox-2 3′UTR (PTGS2 WT) and miR-144-3p (Fig. 2C and D). By contrast, the luciferase activity was from the Cox-2 3′UTR-mut (PTGS2 MT) construct was not significantly affected by transfection with miR-144-3p levels in HTR-8/SVneo cells (Fig. 2D). These results indicate that miR-144-3p targets the Cox-2 3′UTR directly and downregulates its expression.

The basic clinical characteristics of the patients in the preeclampsia and control group are presented in Table I. Patients in the preeclampsia group exhibited significant increase in the systolic pressure and diastolic pressure compared with the control group (P<0.0001). Notably, patients with preeclampsia exhibited proteinuria. Furthermore, the gestational age at delivery was significantly lower for patients with PE than for normal controls (P<0.0001). miRNA ISH analysis was used to compare the expression of miR-144-3p in placenta samples of the two groups (Fig. 3A and B). The results revealed that the expression of miR-144-3p was significantly decreased in placentas from the PE group compared with the normal group (P<0.05; Fig. 3C).

Tissue assays comprising 25 pairs of placentas with preeclampsia and normal placentas were examined for miR-144-3p expression using ISH. ISH revealed differential expression of miR-144-3p in preeclampsia and normal placentas (Fig. 3). Immunohistochemical staining for Cox-2 was performed using the same placenta tissues used for miR-144-3p ISH analysis (Fig. 4A and B). The expression of Cox-2 was negatively correlated with miR-144-3p, as demonstrated by Spearman correlation analysis (r=−0.509, P=0.001; Fig. 4C).

To further investigate the association between miR-144-3p and Cox-2, HTR-8/SVneo cells were transfected with miR-144-3p vector or Vector control. The lentiviral transduction efficiency was assessed (Fig. 5A-C). Western blot analysis was performed to investigate the effect of miR-144-3p on the expression of Cox-2 in HTR-8/SVneo cells in vitro. Overexpression of miR-144-3p decreased Cox-2 protein expression by 38.2% in HTR-8/SVneo cells (Fig. 5D and E).