Hsa_circ_0002348 regulates trophoblast proliferation and apoptosis through miR-126-3p/BAK1 axis in preeclampsia

The patients were diagnosed of preeclampsia according to the American College of Obstetricians and Gynecologists (ACOG) as follows: systolic blood pressure (SBP) of ≥ 140 mmHg and/or diastolic blood pressure (DBP) of ≥ 90 mmHg with proteinuria of ≥ 300 mg/day (or a protein/creatinine ratio of ≥ 0.3 mg/dl or proteinuria of ≥ 1+) after 20 weeks of gestation. Severe preeclampsia was defined by one or more of the following criteria: maternal blood pressure ≥ 160/110 mmHg; severe clinical symptoms such as visual disturbances, pulmonary edema, epigastric or right upper quadrant pain, fetal growth restriction, oligohydramnios. The pregnant women were asked to take a rest of at least five-minutes with the relaxed limbs and an appropriate cuff size, before measuring blood pressure. We usually measured blood pressure in the right upper limb, requiring the cuff to be at the same level as the heart. Urine samples were required to be retained for clean catch midstream.

The placental tissues were obtained from 42 healthy pregnant women, 22 pregnant women with preeclampsia and 27 pregnant women with severe preeclampsia after delivery, a list made of their general characteristics (Additional file 1: Table S1). Among the three groups, no significant difference was observed in maternal age, body mass index (p > 0.05). The group of severe preeclampsia showed the earlier gestation of delivery and the lighter birth weight when compared with the control group (p < 0.001) and the group of mild preeclampsia (p < 0.05). Approved by the Ethics Committee of the Obstetrics and Gynecology Hospital of Fudan University (2021-198), the present study was permitted to collect all the samples from the participants with a written informed consent.

Total RNAs of the tissues and cells were extracted using TRIzol regent (Invitrogen, USA) to be purified following the manufacturer’s instructions. The circRNA and mRNA were reversely transcribed into cDNA using PrimeScript II 1stStrand cDNA Synthesis Kit (Takara, Japan), while the miRNAs were reversely transcribed into cDNA with TaqMan microRNA Reverse Transcription Kit (ThermoFisher, USA). Based on Roche480Real-Time PCR system, quantitative real-time PCR was performed using a SYBR®Premix Ex Taq II (Takara, Japan). β-actin served as internal control for circRNAs and mRNAs. The relative expression of RNAs was calculated via 2^−∆∆Ct method with a log2-transformed step to convert the fold-change values into a format that fitted the downstream analysis. A list was made of the primers sequences in qRT-PCR (Additional file 2: Table S2), all of which were synthesized by Genepharma Co., Ltd. (Shanghai, China).

In situ hybridization (ISH) was conducted with BaseScope™ Reagent Kit, V2-RED (#323900-USM, Advanced Cell Diagnostics (ACD) Hayward, CA). Hsa_circ_0002348 junction site-targeting/-non targeting labeled probes conjugated to horseradish peroxidase (HRP) from ACD (#701041 & #701021). According to the manufacturer’s protocol, 5 μm of formalin-fixed paraffin-embedded (FFPE) human placenta sections were baked to be deparaffinized, before pretreated with hydrogen peroxide, followed by target retrieval performed and hydrophobic barrier created using the Immedge™ hydrophobic barrier pen. The dried slides were placed on the slide rack to be incubated with Protease Plus at 40 °C for 30 min in the HybEZ^TMsystem (ACD). After that, the slides were hybridized with probes (hsa_circ_0002348) at 40 °C for 2hs in HybEZ^TMsystem, followed by signal amplification performed as follows: AMP 1 for 30 min at 40 °C; AMP 2 for 15 min at 40 °C; AMP 3 for 30 min at 40 °C; AMP 4 for 15 min at 40 °C; AMP 5 for 30 min at RT; and AMP 6 for 15 min at RT. Following counterstain with hematoxylin, chromogenic detection was performed using BaseScope™ Fast RED. According to RNAscope standards (score of 4+), BASEScope intensity was quantified by the numbers of positive cells that exhibited at least 10 red puncta.

The human first-trimester extravillous trophoblast (EVT)-derived cell line, HTR8/SVneo [15], which was purchased from American Type Culture Collection (ATCC, Manassas, VA, USA), was cultured in the RPMI 1640 media (Life Technologies, Shanghai, China) with 10% fetal bovine serum (FBS) (Life Technologies, Shanghai, China), with 100 μg/ml streptomycin and 100 U/ml penicillin as supplements. These cells were maintained under standard culture conditions of 5% CO₂ and 37 °C.

Small interfering RNAs (siRNAs) specific for the back-splice junction sequences of hsa_circ_0002348, miR-126-3p mimics, and their corresponding negative control oligonucleotides, were constructed by Genepharma Co. Ltd. (Shanghai, China), with a list made of the sequences of siRNA (Additional file 3: Table S3). Following the manufacturer’s instructions, the transfection of these oligonucleotides was performed using Lipofectamine 2000 (Invitrogen, USA).

To determine the expression levels of hsa_circ_0002348 and its host gene, cysteine rich transmembrane BMP regulator 1 (CRIM1) mRNA, respectively, HTR8/SVneo cells were treated with Actinomycin D and RNase R to effectively digest linear RNA. For Actinomycin D treatment, 2 mg/ml of actinomycin D or dimethylsulphoxide (Sigma-Aldrich, St. Louis, MO, USA) was added, as a negative control, to the cell culture medium. For ribonuclease R (RNase R) treatment, total RNA (4 μg) was mixed with 20 units of RNase R (Epicenter; USA, RNR07250) in the presence of 1× Reaction Buffer for 30 min at 37 °C. The nuclear and cytoplasmic fractions were isolated with the Nuclear and Cytoplasmic Extraction Kit (Cwbio, China) following the manufacture’s protocol. GAPDH, U1, MALAT1 and NEAT1 were employed as the positive controls for cytoplasmic and nuclear fractions, respectively.

Cell Counting Kit-8 (CCK8) assay (RiboBio, Shanghai, China) was performed to assess cell proliferation. The cells at 1 × 10³ were seeded into 96-well plates to be treated with 10ul of CCK-8 solution at the time intervals of 0, 24, 48 and 72 h after 2 h-incubation at 37 °C. The absorbance at 450 nm was read with an enzyme immunoassay analyzer (Thermo Fisher Scientific, Inc., Waltham, MA, USA). This experiment was repeated thrice.

Following 24 h-transfection, HTR8/SVneo cells were seeded into 6-well plates, with 500 per well. Incubated for 2 weeks, the cells were fixed in methanol and stained with 0.1% crystal violet.

With Annexin V-FITC/Propidium Iodide Apoptosis Detection Kit (BD Pharmingen, New York, USA, #556547), cell apoptosis was assessed according to the manufacturer’s instructions. Transfected for 48 h, the cells were collected to be stained with Annexin V-FITC and propidium iodide (PI), before analyzed by fluorescence-activated cell sorting using FACS Canto II (BD Biosciences, San Jose, CA, USA). The apoptosis data were calculated based on FlowJo V10 software (Tree Star, San Francisco, CA, USA). Those of quadrant 4 with FITC Annexin V and PI negative were considered to be viable; those of quadrant 3 with FITC Annexin V positive and PI negative, to be in early apoptosis; those of quadrant 2 with FITC Annexin V and PI positive, to be in late apoptosis; and those of quadrant 1 with FITC Annexin V negative and PI positive, to be necrotic. The rates of cell apoptosis were the sum of those which were derived from early apoptosis and late apoptosis.

When miR-126-3p mimics were transfected into HTR8/SVneo cells, RNA immunoprecipitation (RIP) was performed with the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (EMD Millipore Corp., Billerica MA, USA) containing an AGO2-specific antibody (cell signaling technology), and an IgG antibody as a negative control. Harvested, the cells were centrifuged and re-suspended with 100 μL RIP lysis buffer supplemented with protease and RNase inhibitor. Mixed with antibody coupled beads, the lysate was kept in rotation at 4 °C overnight. Following proteinase K treatment, the immunoprecipitated RNAs were extracted and reversely transcribed as described in the section of RNA extraction and qRT-PCR. Subsequently, the levels of hsa_circ_0002348 were detected by qRT-PCR.

The wide-type and mutant fragments of BAK1 3′UTR cDNA were amplified by PCR. As indicated in Additional file 2: Table S2, a list was made of the cloning primers. The amplified products were synthesized and inserted into psiCHECK-2 vectors (Promega, Shanghai, China). Afterwards, the psiCHECK-2 vectors, miR-126-3p mimics and NC mimics were transfected, as the negative controls, into T293 cells using Lipofectamine 2000 (Invitrogen, USA). After 48 h-incubation, the luciferase activity was quantified with the Dual Luciferase Reporter Assay kit according to the manufacturer’s instructions (Promega, Shanghai, China).

When the cells were lysed, total protein was extracted with Radio Immunoprecipitation Assay (RIPA) lysis buffer (Thermo Scientific, USA). The lysates were resolved with 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), before transferred to polyvinylidenefluoride (PVDF) membranes to be probed with antibodies directed against phosphorylated TOR (Ser2448) (p-mTOR), mTOR, phosphorylated ERK1/2 (T202/Y204) (p-ERK1/2), ERK1/2, phosphorylated p38 (T180/Y182) (p-p38), p38, phosphorylated JNK (T183/Y185) (p-JNK), JNK, BAK1 (Cell Signaling Technology, USA), and β-tubulin (Abcam, USA). Developed with chemiluminescent HRP substrate (Thermo Scientific, USA), the bands were visualized, the intensity of which was quantified based on ImageJ software. All results were normalized to β-tubulin, which was used as a loading control. A detailed description was made of the antibodies (Additional file 4: Table S4).

As previously reported [16], hsa_circ_0002348 vector was established with modified plasmid called 3D5 based on the pZW1 vector (Additional file 5: Fig. S1). The whole sequence of hsa_circ_0002348 (Additional file 6: Table S5) was inserted into the 3D5 vector following digestion with restriction enzymes XhoI and PacI using a seamless cloning assay, and the product was transformed into STBL3 competent cells. The recombinant vector sequences were validated by Sanger sequencing.

Institute of Cancer Research (ICR) pregnant mice aged 10 weeks and weighted around 40 g (Shanghai Slac Laboratory Animal Co. Ltd.), were recorded as embryonic day 0 (E 0) with a copulation plug. All animal experiments were performed in accordance with the guidelines issued by Fudan University for the care and use of laboratory animals. The mice were raised in a light-, temperature- and humidity-controlled environment with free access to standard rodent chow and tap water. As previously described [17, 18], we succeeded in establishing a mouse model by intraperitoneally (i.p.) injecting lipopolysaccharide (LPS) (Escherichia coli serotype 0111:B4, Sigma-Aldrich, USA) twice into the pregnant mice at a dose of 100 μg/kg bodyweight on GD 5 and GD 10, respectively. To evaluate the role of hsa_circ_0002348 in the in vivo development of preeclampsia, additionally, we injected LPS-induced preeclampsia mice with hsa_circ_0002348 plasmids (320 μg/kg bodyweight) or an equal amount of pZW1 as blank controls, using Entranster™-in vivo Transfection Reagent (Engreen, China) on GD 8, 12 and 14, respectively. Thereby, the mice were randomly divided into four experimental groups: the control, LPS, LPS+pZW1 and LPS+hsa_circ_0002348. The controls were injected with the same volume of saline on GD 5, 8, 10, 12 and 14 correspondingly. Besides the injections received on GD 5 and 10, LPS group mice were injected with the same dosage of saline on GD 8, 12 and 14. The animal experiments were independently repeated for three times.

To evaluate the specific symptoms of preeclampsia, blood pressure (BP) and urinary protein were examined. The indirect BP was measured in the conscious mice by tail cuff plethysmography of the Visitech System BP2000 (Apex, NC, USA). All mice were habituated for 3 days before the formal start of BP measurement, at least 5 consecutive measurements recorded on GD 5, 14 and 16, respectively. At the same time points, the mice in the metabolic cages underwent quantification of their urinary protein concentration of 24 h urine samples, which used a CBB protein test kit (SNM297, Baiaolaibo Technology, Beijing, China) according to the manufacturer’s instructions.

The mice’s kidneys were fixed and embedded in paraffin before sliced into 3-μm-thick serial sections, which were to be stained with hematoxylin and eosin (HE) or periodic acid-Schiff (PAS) for the examination of glomerular morphology. The mice’s placentas were stained with HE or Masson’s trichrome (MTC) staining for the examination of placental morphology. MTC was used to visualize collagen deposition, which was stained blue, while the muscle and cytoplasm appeared red to pink.

In the mice’s plasma, the protein levels of placental growth factor (PIGF), soluble fms-like tyrosine kinase-1 (sFlt-1), and tumor necrosis factor-α (TNF-α) were measured with enzyme-linked immunosorbent assay (ELISA) kits (ab197748; Abcam for PIGF, MBS9301224; Enzo Life Sciences, Inc. for sFlt-1, EMC102a.48; NeoBioscience for TNF-α) according to the manufacturer’s protocols.

Terminal deoxynucleotidyl transferase (TdT)-mediated-digoxigenin-11-dUTP nick end labeling (TUNEL) was used to detect DNA fragmentation in placenta tissues of pregnant mice from the four experimental groups according to the instructions of TUNEL kit (KGA7063, KeyGENBioTECH, Jiangsu, China).

The human and pregnant mouse derived placental tissues were fixed in 4% formaldehyde solution and embedded in paraffin before sectioned. Fluorescence in situ hybridization (FISH) was performed to examine the expression of miR-126-3p, as previously described [19]. The miR-126-3p probes (GenePharma, China) were generated based on the following sequence: 5′-FAM-CGCATTATTACTCACGGTACGA-FAM-3′ and the sequence was identical in humans and mice. The probes were stained green, and the nuclei were counterstained blue with DAPI. The slides were reviewed under Nikon Fluorescence Microscope. For immunohistochemistry, the slides were deparaffinized in xylene and rinsed with PBS, before 30 min-pretreatment in 3% H₂O₂ and 50% methanol to eliminate the endogenous peroxidase activity. After that, the slides were disposed with a heat-induced epitope retrieval (HIER) method for antigen repair, then washed with PBS thrice and incubated at room temperature for 30 min in PBS containing 1% BSA blocking buffer, followed by an addition of BAK1 antibody (Cell Signaling Technology, USA), before incubated overnight at 4 °C. The slides were then washed thrice in PBS before incubated for 1 h at RT with the goat anti-rabbit secondary antibody. When the slides were rinsed thrice in PBS again, the reaction was developed with peroxidase substrate diaminobenzidine. Under a microscope was observed hematoxylin counterstain.

We analyzed the dataset GSE100242 in the GEO datasets [20], which presented a circRNA expression resource of 20 human tissues and 63 circRNAs were identified to be expressed in placenta tissue. Given that endothelial cells were also the main component of the placenta and endothelial dysfunction played a role in the development of preeclampsia [21], we examined a circRNA expression profile of human umbilical vein endothelial cells (HUVECs) containing 7388 circRNAs, of which 1884 had high-throughput sequencing of RNA isolated by cross-linking and immunoprecipitation (HITS-CLIP) peaks that determined a potential function as miRNA sponge [22]. Through intersection analysis of the host gene sequences of the 63 placenta and 1884 endothelial cells-derived circRNAs, it was found that the gene sequence of 9 circRNAs completely overlapped, of which 8 were annotated in circBase [23] and selected for further analysis. Considering that circRNAs and their cognate mRNAs may have similar functions in the same disease such as preeclampsia [24‐26]. As a result, hsa_circ_0002348 became the research target for its host gene, CRIM1 expressed in both trophoblast and endothelial cells, and was essential for murine placenta development [27, 28] (Fig. 1A). Hsa_circ_0002348, located at chr2:36623756–36706837, consisted of 1041 nt, as the back-spliced circular product of the exon 2–7 of the CRIM1 transcript (Fig. 1B). The sequence of the splice junction of hsa_circ_0002348 was confirmed by Sanger sequencing, which was in concordance with its public sequence in circBase [23] (Fig. 1C). In order to further prove that hsa_circ_0002348 was a truly circRNA rather than a linear RNA, upon Act D treatment, the expression level of hsa_circ_0002348 was stable with a half-time > 24 h, while the linear transcript, CRIM1, was degraded with a half-time < 6 h (Fig. 1D) (p < 0.05). Additionally, hsa_circ_0002348 was resistant to the digestion of excessive RNase R with obviously degraded linear CRIM mRNAs, while the expression of the circular hsa_circ_0002348 remained unchanged (Fig. 1E) (p < 0.001). Furthermore, the separate qRT-PCR analysis of nuclear and cytoplasmic RNA exhibited that hsa_circ_0002348 was mainly present in the cytoplasm of HTR8/SVneo cells (Fig. 1F).

To identify whether hsa_circ_0002348 is associated with preeclampsia pathogenesis, we verified that it was significantly upregulated in placental tissues of 27 patients with severe preeclampsia, although not significant for those with mild clinical manifestation (n = 22), when compared with 22 normal controls (Fig. 2A) (p < 0.05). A significantly stronger positive signal of hsa_circ_0002348 was also observed in placental tissues of severe preeclampsia patients than in the controls of normal pregnancy (Fig. 2B) (p < 0.05). Furthermore, the expression of hsa_circ_0002348 was positively correlated with the systolic and diastolic BP and level of proteinuria in the patients with severe clinical manifestation (Fig. 2C) (p < 0.05), while no correlation was identified in the patients with mild clinical manifestation (Additional file 7: Fig. S2). These results showed that hsa_circ_0002348 was upregulated in preeclampsia placentas, indicating its involvement in the pathogenesis of preeclampsia.

To further figure out the cause-and-effect relationship between hsa_circ_0002348 and preeclampsia, we established a mouse model of preeclampsia using intraperitoneal (i.p.) injection of LPS and determined whether hsa_circ_0002348 overexpression vector could cause aggravation of preeclampsia (Fig. 3A). The successful induction of preeclampsia was confirmed, as manifested by the significantly higher systolic (Fig. 3C) (p < 0.0001) and diastolic BP (Fig. 3D) (p < 0.01) and proteinuria level (Fig. 3E) (p < 0.05), as well as by the reduced number of fetal mice and placentas (Fig. 3F), after LPS injections on GD 5 and GD 10. In line with results of human preeclampsia placentas, hsa_circ_0002348 was highly expressed in placentas of LPS-induced mouse model of preeclampsia (Fig. 3B). When concurrently injected with hsa_circ_0002348 vector on GD 8, GD 12 and GD 14, the mice displayed significantly enhanced expression of hsa_circ_0002348 in their placentas (Fig. 3B) (p < 0.05), afflicted with more severe preeclampsia-like symptoms such as fetal reabsorption besides higher BP, proteinuria, less fetal mice and placentas for LPS+circ_0002348 group when compared with LPS+pZW1-treated group (Fig. 3C–F) (p < 0.01). Compared with the controls, pregnant mice in LPS group displayed a significant decrease in PIGF expression (Fig. 3G) (p < 0.0001) and increase in sFlt-1 (Fig. 3H) (p < 0.0001), TNF-α (Fig. 3I) (p < 0.001), while more significant changing trends were observed in LPS+circ_0002348 group than in LPS+pZW1 group (Fig. 3G–I) (p < 0.01). Furthermore, the morphological changes in maternal kidney and placenta were assessed. The renal tissues showed relatively atrophic glomerular, and narrowed glomerular cystic cavities in pregnant mice after treated with LPS, LPS and hsa_circ_0002348 overexpression plasmid. Likewise, the labyrinth layer of placental tissues exhibited more blue staining in MTC staining, indicated a significant increase of stromal collagen deposition in LPS and LPS+circ_0002348 group when compared with the other three groups (Fig. 3J). Altogether, these findings revealed that hsa_circ_0002348 aggravated the typical clinical features and organ injury in the mouse model of LPS-induced preeclampsia.

Although trophoblast invasion is an initiating factor in the development of preeclampsia [29], apoptosis is particularly important to all stages of placental development and an activation of pro-apoptosis responses in the trophoblast cells could contribute to abnormal placentation in preeclampsia [6, 30]. Therefore, our focus was on trophoblast apoptosis. We investigated the effect of hsa_circ_0002348 on the apoptosis viability of placental trophoblast in the mouse model of LPS-induced preeclampsia. As indicated by EDU (red)/DAPI (blue) immunostaining assay, the apoptosis of trophoblast cells increased in the groups of LPS and LPS+pZW1, and increased more significantly in the group of LPS+circ_0002348 than in the control group (Fig. 4A).

Moreover, we explored the effect of hsa_circ_0002348 on the proliferation and apoptosis in vitro. We knocked down the expression of hsa_circ_0002348 using two small interfering RNAs (siRNAs) targeting the hsa_circ_0002348 junction site, and upregulated its expression by transfecting its overexpression vector. The knockdown and overexpression efficiency were confirmed by qRT-PCR (Fig. 4B, C) (p < 0.05). As shown by the CCK-8 assay, HTR8/SVneo proliferative ability was increased after transfection with si-circ_0002348-#1 and si-circ_0002348-#2 (Fig. 4D) (p < 0.001), but decreased with the hsa_circ_0002348 overexpression vector transfection (Fig. 4E) (p < 0.05). As verified by the colony formation assay, furthermore, HTR-8/SVneo cells proliferation was significantly promoted by the knockout of hsa_circ_0002348 (Fig. 4F) (p < 0.05), but inhibited by its overexpression (Fig. 4G) (p < 0.01). Additionally, when the cell apoptosis rate was tested by flow cytometry so as to evaluate the effect of hsa_circ_0002348 on trophoblast apoptosis, we found that the downregulation of hsa_circ_0002348 suppressed apoptosis significantly (Fig. 4H) (p < 0.001), whereas its upregulation promoted the total rate of apoptosis (Fig. 4I) (p < 0.01).

The mTOR and MAPK signaling pathways, which are both important intracellular signaling pathways that play a key role in regulating various cellular process such as cell proliferation and apoptosis [31, 32], are promising to be therapeutic targets [33]. Moreover, there is evidence that mTOR and MAPK pathways are involved in trophoblast biological activity in placenta-mediated pregnancy complications such as fetal growth restriction and preeclampsia [34, 35]. To investigate the possible pathways involved in hsa_circ_0002348 mediated trophoblast proliferation and apoptosis, we confirmed its role in mTOR and ERK1/2 pathways via western blot. Increased phosphorylation levels of mTOR and ERK1/2 were observed after hsa_circ_0002348 knockout in HTR8/SVneo cells. However, no apparent change was observed in the expression levels of p38 and JNK and their phosphorylation (Fig. 4J) (p < 0.05). Taken together, these results demonstrated that hsa_circ_0002348 regulated trophoblast proliferation and apoptosis through the repression of the mTOR and ERK1/2 pathway.

As indicated by the subcellular fractionation and localization assays, hsa_circ_0002348 was primarily localized to the cytoplasm, which indicated that it most probably played a regulatory role in a post-transcriptional manner. Given that circRNAs function as miRNA sponge that protects mRNAs by competing for their targeting miRNAs, we hypothesized that it could play a similar role. Therefore, we searched for putative miRNAs binding to hsa_circ_0002348 by making bioinformatics analyses based on Circnet, Interactome and RNA22 software. Consequently, we obtained five miRNAs including miR-126-3p, miR-145, miR-182, miR-377 and miR-3942, all of which had potential binding sites with hsa_circ_0002348 (Fig. 5A). Moreover, when HTR8/SVneo cells were transfected with the mimics of all the five miRNAs, trophoblast proliferation capacity was markedly augmented (Additional file 8: Fig. S3), especially the miR-126-3p presenting the most significantly changing trends (Fig. 5B) (p < 0.05), revealed by CCK-8 assays. Thus, we focused on the miR-126-3p and explored its interaction with hsa_circ_0002348 using anti-AGO2 RIP in HTR8/SVneo cells to transiently upregulate miR-126-3p. When pulled down by AGO2, the endogenous hsa_circ_0002348 was enriched in the cells treated with miR-126-3p (Fig. 5C) (p < 0.001), which indicated that miR-126-3p targeted hsa_circ_0002348 directly. As shown by the colony formation assays, furthermore, the ectopic overexpression of miR-126-3p prominently promoted trophoblast proliferation (Fig. 5D) (p < 0.01). Under the same treatment, an obviously decreased rate of apoptosis was observed by flow cytometry assay (Fig. 5E) (p < 0.01).

Next, to determine whether hsa_circ_0002348 regulated trophoblast proliferation and apoptosis by competing for miR-126-3p, we conducted rescue assays. Ectopic expression of hsa_circ_0002348 was found to attenuate remarkably the trophoblast proliferation, while the overexpression of miR-126-3p reduced this attenuation (Fig. 5F) (p < 0.05). Meanwhile, the overexpression of hsa_circ_0002348 induced cell apoptosis dramatically, and the re-expression of miR-126-3p was capable of approximately reversing the effect of hsa_circ_0002348 on the cell apoptosis in trophoblast cells (Fig. 5G) (p < 0.01). Furthermore, western blot showed that the overexpression of miR-126-3p resulted in a slightly increased level of phosphorylated mTOR and ERK1/2 in trophoblast cells (Fig. 5H) (p < 0.05).

Additionally, FISH analysis validated the expression alteration of miR-126-3p in the mice with LPS-induced preeclampsia as well as in the patients with preeclampsia, showing that miR-126-3p was sharply decreased in both (Fig. 5I, J). Particularly, after the injection of hsa_circ_0002348 overexpression plasmid, the lower expression level of miR-126-3p was detected in the placentas of LPS+circ_0002348 group than in those of other three groups (Fig. 5I). These findings illustrated that hsa_circ_0002348 exerted its effect on trophoblast proliferation and apoptosis when interacting with miR-126-3p.

In the identification of the hsa_circ_0002348/miR-126-3p-regulated downstream target genes which might play a role in trophoblast proliferation and apoptosis of preeclampsia, an integrative analysis of three public databases including Starbase, TargetScan and RNA22, predicted eight common target genes (Fig. 6A), one of which as BAK1 of proapoptotic protein (member of Bcl-2 family), was of particular interest to us for its extensive expression in essentially all organs and its role as a regulator of apoptosis in multiple cell types [36]. Thus, we selected BAK1 for luciferase reporter experiment, where HTR8/SVneo cells were transfected with BAK1 3′UTR or BAK1 mutant 3′UTR fragments, before with miR-126-3p mimics or NC mimics. The results showed that the luciferase activity of BAK1 3′UTR decreased in the group of miR-126-3p mimics when compared with the group of NC mimics, which was not observed in BAK1 mutant 3′UTR (Fig. 6B) (p < 0.05). Next, to verify that BAK1 expression is indeed regulated through hsa_circ_0002348 and miR-126-3p, we conducted qRT-PCR and western blot analysis in HTR8/SVneo cells. Hsa_circ_0002348 silencing significantly decreased BAK1 mRNA expression (Fig. 6C) (p < 0.01), so did miR-126-3p overexpression (Fig. 6D) (p < 0.01). Subsequent rescue experiments indicated that the overexpression of hsa_circ_0002348 significantly increased BAK1 expression, which was reversed by the co-expression of the miR-126-3p mimics (Fig. 5E) (p < 0.05). Similar results were validated at protein level by western blot analysis (Fig. 6F) (p < 0.05). However, miR-126-3p did not affect the expression of hsa_circ_0002348, even in the co-expression of hsa_circ_0002348 with miR-126-3p (Fig. 6G) (p < 0.01).

From IHC analysis conducted to further evaluate the effect of hsa_circ_0002348 on BAK1 protein expression in the placentas of the mice and patients, BAK1 protein expression was found to have increased in the placentas of the group of LPS and LPS+pZW1 when compared with the controls, with the highest level detected in the group of LPS+circ_0002348 (Fig. 6H) (p < 0.01). Similarly, BAK1 protein expression had a high level in the placentas of the patients (Fig. 6I) (p < 0.05). These data indicated that hsa_circ_0002348 played an important role in the placental pathogenesis of preeclampsia through miR-126-3p/BAK1 axis.