Novel lncRNA‐HZ04 promotes BPDE‐induced human trophoblast cell apoptosis and miscarriage by upregulating IP3R1/CaMKII/SGCB pathway by competitively binding with miR‐hz04

Normal pregnancy is essential for human reproduction. However, BaP (benzo(a)pyrene) and its metabolite BPDE (benzo(a)pyrene‐7,8‐dihydrodiol‐9,10‐epoxide) could cause dysfunctions of human trophoblast cells and might further induce miscarriage. Yet, the underlying mechanisms remain largely unknown. Herein, we identified a novel upregulated lnc‐HZ04 and a novel downregulated miR‐hz04 in villous tissues of unexplained recurrent miscarriage (RM) relative to those in healthy control tissues and also in BPDE‐treated human trophoblast cells. Lnc‐HZ04 directly and specifically bound with miR‐hz04, diminished the reduction effects of miR‐hz04 on IP3R1 mRNA expression level and on IP3R1 mRNA stability, and then activated the Ca2+‐mediated IP3R1/p‐CaMKII/SGCB pathway, which further promoted trophoblast cell apoptosis. The miR‐hz04 target site on lnc‐HZ04 played crucial roles in these regulations. In normal trophoblast, relatively less lnc‐HZ04 and more miR‐hz04 suppressed this apoptosis pathway and gave normal pregnancy. After exposure to BPDE or in RM tissues, p53 was upregulated, which might promote p53‐mediated lnc‐HZ04 transcription. Relatively more lnc‐HZ04 and less miR‐hz04 activated this apoptosis pathway and might further induce miscarriage. BaP could also induce mice miscarriage by upregulating its corresponding murine apoptosis pathway. Therefore, BPDE‐induced apoptosis of human trophoblast cells was associated with the occurrence of miscarriage. This work discovered the regulation roles of lnc‐HZ04 and miR‐hz04 and provided scientific and clinical understanding of the occurrence of unexplained miscarriage.


| INTRODUCTION
Approximately 15%-25% of total pregnant women experience miscarriage, 1 and roughly 3% suffer from recurrent miscarriage (RM, two or more consecutive miscarriages). 2 In general, chromosomal abnormalities, hormonal abnormalities, uterine deformation, psychological trauma, infection, stressful life events, and immune disorders are considered as the causes for miscarriage. 3,4 However, there are still approximately 50% miscarriage unexplained. 5 Thus, it is urgent to identify the unknown causes for miscarriage and explore their pathogenesis and regulation mechanisms.
Increasing epidemiological studies have indicated that exposure of environmental polycyclic aromatic hydrocarbons (PAHs) is closely associated with miscarriage. [6][7][8] High levels of BaP (benzo(a)pyrene), a representative of PAHs, have been detected out in fried or smoked food and on environmental particular matters. The total BaP intake was estimated to be 125 ng/person/day based on the daily food consumption and BaP levels. 9 The lipophilicity and long half-life of BaP further increase its accumulation in body. Female smokers have significantly higher levels of BaP (1.32 ± 0.68 ng/mL) in their follicular fluid compared with their nonsmoking counterparts (0.03 ± 0.01 ng/mL). 10 Higher levels of BaP have been detected out in the follicular fluid of women who were not pregnant (1.79 ± 0.03 ng/mL) compared with those that achieved a pregnancy (0.08 ± 0.03 ng/mL). 11 BaP can be metabolized to BPDE (benzo(a)pyrene-7,8-dihydrodiol-9,10-epoxide), which further form BPDE-DNA adducts. 12 A case-control study of miscarriage has shown that high level of BPDE-DNA adducts in human maternal blood is correlated with high risk of miscarriage during early pregnancy. 13 In animal studies, after treatment with BaP, the uterus morphology becomes abnormal, the number of germ cells and the fertility rates decrease, and the miscarriage rates increase. 14,15 All these lines of evidence show that BaP might induce miscarriage. However, the underlying mechanisms remain largely unknown.
In normal pregnancy, the trophoblast stem cells differentiate into villous and extravillous trophoblast cells immediately after implantation. Extravillous trophoblast cells take important roles in embryo implantation and the formation of functional placenta. 16 The apoptosis of extravillous trophoblast cell might result in abnormal placental development and miscarriage. 17,18 A higher degree of trophoblast cell apoptosis has been found in villous tissues of miscarriage patients. 19 On the other hand, human trophoblast cells are very sensitive to environmental carcinogens. Recently, we have found that BPDE could induce the apoptosis of human extravillous trophoblast Swan 71 cells due to the disorders of mitochondrial fission/fusion 20 or due to the upregulation of SAT1-mediated spermine metabolism. 21 Intracellular Ca 2+ plays important roles in the regulation of a variety of cellular processes, such as cell cycle, apoptosis, and migration. 22,23 Recent studies have shown that inositol trisphosphate receptor type 1 (IP 3 R 1 )-mediated Ca 2+ release upregulates the cytosolic-free Ca 2+ level, which induces rat miscarriage. 24 Ca 2+ binds to calmodulin, which activates the calmodulin-dependent protein kinase (CaMK), increases the CaMKII phosphorylation, 25 and then activates guanylyl cyclases (SGC), which subsequently triggers a wide range of signaling pathways, including mitochondrial damage and cell apoptosis. 26 Thus, IP 3 R/Ca 2+ /CaMKII/SGCB pathway might regulate cell apoptosis. However, whether this pathway might regulate apoptosis in human trophoblast cell is still unknown.
Long noncoding RNAs (lncRNAs) and microRNAs (miRNAs) are emerging to play significant roles in epigenetic regulation of trophoblast cell functions and miscarriage. [27][28][29][30][31][32] In general, these lncRNAs or miRNAs are differentially expressed in RM villous tissues relative to those in health control (HC) tissues, and they might regulate the signaling pathways that are related with the occurrence of miscarriage. 33,34 LncRNAs generally regulate miRNAs by taking the roles of competing endogenous RNAs (ceRNAs). 35 In this mechanism, a lncRNA and a specific mRNA, both of which harbor the same miRNA binding site, could mutually and negatively regulate each other's expression levels by competitively binding with this same miRNA. 36 Reversely, miRNAs could also bind to their target sites in lncRNAs to promote lncRNA degradation. 35 However, a large number of noncoding RNAs, as well as their interactions, are still unidentified. Moreover, lncRNA and miRNA that could simultaneously regulate the occurrence of miscarriage and the BPDE-induced apoptosis of human trophoblast cells have rarely been reported.
In our recent work, we have identified that a novel lnc-HZ01 and MXD1 (MAX dimerization protein 1) comprise a positive self-feedback loop, which is activated by BaP or BPDE exposure. 37 This loop is upregulated in both RM tissues and BPDE-exposed trophoblast cells, which further promotes EIF4E transcription, inhibits trophoblast cell proliferation, and induces miscarriage. We also found that, in BPDE-exposed human trophoblast cells and in RM | 3 of 20 HUANG et Al.
tissues, lnc-HZ03 and miR-hz03 upregulates p53/SAT1 axis and promote spermine metabolism, which further promote BPDE-induced human trophoblastic cell apoptosis and the occurrence of miscarriage. 21 Recently, we have found that lnc-HZ08 regulates BPDE-induced trophoblast cell dysfunctions by promoting PI3K ubiquitin degradation and is also associated with miscarriage. 38 In this work, we further explore novel lncRNA/miRNA and signaling pathway that are differentially expressed in RM tissues and in BPDE-exposed trophoblast cells to explore the roles of lncRNA/miRNA in regulation of this signaling pathway, to discover new mechanisms underlying the occurrence of miscarriage and BPDEinduced dysfunctions of human trophoblast cells, and to provide novel scientific and clinical understanding of the occurrence of miscarriage.

| Tissues collection and statement
Fifteen women who had undergone artificial miscarriages to terminate unwanted pregnancies (the HC group) and 15 patients with two or more consecutive unexplained miscarriage (the RM group) in the age between 25 and 30 were recruited, as described previously. 21,37,38 All HC groups have previous pregnancies. Any woman with one of the following features was excluded, including autoimmune abnormality, luteal phase defects, hyperprolactinemia, antiphospholipid antibody syndrome, or hyperandrogenemia; uterine abnormalities or cervical incompetence; abnormal karyotype of the parents or abortus; polycystic ovarian syndrome; the symptoms of endocrine or metabolic diseases; viral infectious diseases; tuberculosis, HBV, HCV, HIV, or with positive results from γ-interferon release tests; and eclampsia or preeclampsia. The villous tissue samples were collected from these two groups at 6-10 weeks of gestation to extract RNAs and proteins. The relative levels of two biomolecules in each RM or HC sample were linearly correlated using SPSS 24.0 software (SPSS Inc, Chicago, IL, USA). The research protocols have been approved by the Medical Ethics Committee of the West China Fourth Hospital, Sichuan University. Written informed consent was obtained from all the participants.

| 5′ and 3′ rapid amplification of cDNA ends assays
The 5′-and 3′-rapid amplification of cDNA ends (RACE) assays were performed using the SMARTer RACE 5′/3′ Kit (Takara Bio, Tokyo, Japan). Total RNAs were extracted using the RNeasy Micro Kit (Bio-Rad, Hercules, CA, USA) and treated with DNase. RNA integrity was assessed using the Agilent 4200 TapeStation system. After reverse transcription of total RNAs to cDNAs, 5′-and 3′-RACE PCRs were performed using the universal and gene-specific primers, which are listed in Table S3. The resulting PCR products were cloned into linearized RACE vector and were sequenced. The full-length nucleotide sequence of lnc-HZ04 is shown in Figure S1A.

| Cell apoptosis analysis
Swan 71 cells, which have been exposed to BPDE (0, 0.25, 0.5, 1.0, or 1.5 μM) for 24 hours, or transfected with various miRNA or plasmid for 24 hours, were harvested and treated with a commercial AnnexinV-FITC/propidium iodide (PI) Kit (Sigma-Aldrich, St. Louis, MO, USA) according to the manufacturer's instructions. The cells were immediately detected by flow cytometry (BD Accuri C6, Becton-Dickinson, Franklin Lakes, NJ, USA) and the data were analyzed by FlowJo software (TreeStar, San Carlos, CA, USA).

| Measurement of the intracellular Ca 2+
concentrations After exposure to BPDE (0, 0.25, 0.5, 1.0, or 1.5 μM) or transfection with various miRNA or plasmid for 24 hours, Swan 71 cells were harvested and their intracellular Ca 2+ concentrations were detected using Cell Ca 2+ Kit (Invitrogen) according to the manufacturer's instructions. The fluorescence intensity corresponding to Ca 2+ concentration in Swan 71 cells was detected on ECLIPSE Ti fluorescence microscope (Nikon), and the data were analyzed by Image J software. Ca 2+ concentrations were expressed as the change folds of the fluorescence intensity of the treated cells against that of the control cells.

| Quantitative real-time polymerase chain reaction
RT-qPCR analysis was performed as described previously. 21,37,38 Total RNAs were extracted from cells or tissues. Subsequently, mRNA and lncRNA were reversely transcribed into cDNA using the Prime Script RT Reagent Kit with gDNA Eraser (Takara). MiRNA was transcribed into cDNA using microRNA First-Strand cDNA Synthesis Kit (Thermo Fisher Scientific). Quantitative detection was performed using an iQ5 real-time detection system (Bio-Rad Laboratories) and SYBRVR Premix Ex TaqTM II (Bio-Rad Laboratories). The expression levels of RNA were calculated using the 2 −ΔΔCt method. GAPDH mRNA was used as the normalization internal control for lncRNA and mRNA; and U6 was used for miRNA. The primers used in these RT-qPCR assays are listed in Table S5.

| Fluorescence in situ hybridization
Fluorescence in situ hybridization (FISH) assays were performed using Invitrogen FISH Tag Detection Kits (Thermo Fisher Scientific) according to the manufacturer's instructions. Briefly, Swan 71 cells were rinsed with PBS and then fixed in 4% formaldehyde for 10 minutes at 25°C. The cells were permeabilized in PBS containing 0.5% Triton X-100 on ice for 10 minutes, and then washed with PBS. Subsequently, lnc-HZ04 was hybridized with Cy5-labeled ssDNA probe (5′-Cy 5-GGCTXTTGCCTGTGGTXTCTTTCTTC G-3′, X means locked nucleotide acid modification) (Empire Genomics, Williamsville, NK, USA) at 37°C for 12 hours. After staining nucleus with DAPI, images were digitally captured by confocal microscopy (N-STORM + A1R, Nikon, Japan) at 600× magnification. Nuclei with ≥5 foci were considered positive 42 and the relative levels of lnc-HZ04 in the nucleus and cytoplasm were calculated from 20 random cells.

| mRNA stability assays
The de novo RNA synthesis was blocked after addition of 5 μg/ mL ActD (Apexbio, Madison, WI, USA) to Swan 71 or HTR-8/ SVneo cells with overexpression or knockdown of lnc-HZ04 and/or miR-hz04. The total RNAs were extracted after 0, 1, 2, or 3 hours, and the RNA levels of lnc-HZ04 or IP 3 R 1 were detected by RT-qPCR. GAPDH was used as normalization internal standard. Expression levels were calculated using the ΔCt method by normalizing each value against the value at 0 hour.

| Dual-Luciferase reporter assays
Lnc-HZ04 and IP 3 R 1 mRNA containing wild-type (wt) or mutant (mut) binding site for miR-hz04 were generated and fused into the luciferase pGL3-basic reporter vector (Promega, Madison, USA). Swan 71 cells were seeded into 6-well plates and then co-transfected with 100 ng pmirGLO-HZ04-wt/-mut, or pmirGLO-IP 3 R 1 -wt/-mut and 200 nM miR-hz04 mimics or control in Lipofectamine 3000 Transfection Reagent (Invitrogen) according to the manufacturer's instructions. Then, the firefly luciferase activity in each well was measured using Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer's protocol.

| RIP assays
RIP experiments were carried out using the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore) according to the manufacturer's instructions. Briefly, cell lysates were incubated with protein A/G magnetic beads that had conjugated with antibody of IP 3 R 1 (dilution 1:100, ab264281, Abcam), CaMKII (dilution 1:100, sc-5306, SANTA Cruz), p-CaMKII (dilution 1:100, sc-32289, SANTA Cruz), SGCB (dilution 1:100, sc-393679, SANTA Cruz), or negative control IgG at 4°C overnight. The lysate was also used as preimmunoprecipitation input. After washing extensively, the immunoprecipitated RNA was extracted. Lnc-HZ04 and negative control GAPDH mRNA were quantified by RT-qPCR, with the pre-immunoprecipitation as input and IgG as control. The primers used in these assays are listed in Table S5.

| In vitro RNA pull-down assays
Lnc-HZ04 or IP 3 R 1 mRNA containing wild-type (wt) or mutant (mut) target site for miR-hz04 was transcribed in vitro from pGEM-T-HZ04-wt/-mut or pGEM-T-IP 3 R 1 -wt/-mut, respectively. The vector pGEM-T was used as negative control. All the transcripts were biotin-labeled by T7 RNA polymerase with the Biotin RNA Labeling Mix (Roche), treated with RNasefree DNase I (Roche), and then purified with an RNeasy Mini Kit (Qiagen, Valencia, CA, USA). These transcripts were then incubated with the lysates of Swan 71 cells or Swan 71 cells with overexpression or knockdown of wild-type lnc-HZ04 or IP 3 R 1 mRNA overnight at 4°C. The complexes were isolated by streptavidin agarose beads (Invitrogen), and RNA in the pull-down material was detected by RT-qPCR. The relative retention was calculated against the input RNA level.

| Chromatin immunoprecipitation assays
EZ-Magna Chromatin immunoprecipitation (ChIP) Kit (Millipore) was used in ChIP assays, as described previously. 37,38 After sonication of the chromatin, the anti-p53 monoclonal antibody (ab1101, Abcam) was used for immunoprecipitation at 4°C overnight on an inverse rotator, with IgG (ab172730, Abcam) as negative control. The complex was washed thrice and DNA was separated. Subsequently, the DNA fraction was analyzed by qPCR. Specific qPCR primers are listed in Table S5.

| Mice and experimental design
The mice mode was made according to the method described previously. 21,37,38 Briefly, C57BL/6 mice of 6-8 weeks (Charles River, Beijing, China) were raised in filter-top cages with a 12-h light/dark cycle, autoclaved bedding, food, and water. After mating overnight, the appearance of vaginal plug was defined as day 1 (D1). Then, the pregnant mice were randomly divided into three groups (each n = 9). Each group was dosed daily with corn oil (control) or 0.05 or 0.2 mg/kg BaP in corn oil by oral gavage from D1 to D13 and were euthanized on D14. The embryo resorption was identified by smaller or darker appearance than the viable and pink healthy embryos. 21,37,43 A random placenta was collected from each of the nine mice in each group, and RNA and proteins were extracted. The relative levels of two biomolecules in each mouse placenta tissue were linearly correlated using SPSS 20.0 software. The research protocols have been approved by the Medical Ethics Committee of the West China Fourth Hospital, Sichuan University.

| Statistical analysis
Each experiment was repeated independently thrice with similar results. The data were presented as mean ± SD (standard deviation). Statistical analysis was performed using SPSS 24.0 software (SPSS Inc, Chicago, IL, USA). Data were analyzed using independent samples Student's t-test between two groups and one-way analysis of variance for more groups with Dunnett's or LSD post hoc test. The correlation was analyzed by Pearson analysis. All graphs were prepared using GraphPad Prism version 8.0. Differences were considered significant when *P < .05, **P < .01, or ***P < .001.

CaMKII/SGCB pathway were upregulated and a novel miR-hz04 was downregulated in human trophoblast cells treated with BPDE and in human RM tissues
In our recent work, we have found that BPDE exposure could induce human trophoblast Swan 71 cell apoptosis. 20 Then, we next to explore the roles of novel noncoding RNAs in the regulation of trophoblast cell apoptosis. Based on the occupational exposure, diffusion of BaP from environment into cells, and high accumulation of BaP metabolite in follicular fluid and serum, we still used 0, 0.5, or 1.5 μM BPDE to treat human trophoblast Swan 71 cells 20,44 and then performed transcriptome sequencing to identify the differentially expressed noncoding RNAs. As described previously, 21,37,38 we have identified that 22 novel lncRNAs were increased and six were decreased with change fold >4 and P-values <.05 in both 0 versus 0.5 and 0.5 versus 1.5 μM BPDE-treated Swan 71 cells; and all these 28 novel lncRNAs have been validated by RT-qPCR analysis. In this work, we selected another novel lncRNA (lnc-14628) and investigated its roles in regulation of trophoblast cell functions in detail.
This lnc-14628 was further identified to be upregulated in BPDE-treated Swan 71 and HTR-8/SVneo cells by RT-qPCR analysis ( Figure 1A,B). It was verified as a sense transcript with a full length of 318 nucleotides (nt) by RACE assays ( Figure S1A, Table S1) and resided at chromosome 16 (chr 16:3,170,699 -3,171,016, NC_000016.10). Thus, this ln-cRNA was also termed as lnc-HZ04 and submitted to NCBI with accessory No. MT821846. The protein-coding potential of lnc-HZ04 was very weak, as analyzed by Coding Potential Assessment Tool (https://lilab.resea rch.bcm.edu/cpat/index. php) and NCBI ORF finder (https://www.ncbi.nlm.nih.gov/ orffi nder/). 45 Additionally, lnc-HZ04 was found to have no conserved domains using Conserved Domain Database (https://www.ncbi.nlm.nih.gov/cdd/) and Pfam (http://pfam. xfam.org/). 46,47 The data indicated that lnc-HZ04 could not encode a protein. Lnc-HZ04 was distributed to both the F I G U R E 1 Lnc-HZ04 and IP 3 R 1 /p-CaMKII/SGCB pathway were upregulated and miR-hz04 was downregulated in both BPDE-exposed human trophoblast cells and human RM tissues. A and B, RT-qPCR analysis of lnc-HZ04 level in BPDE-treated Swan 71 (A) and HTR-8/SVneo (B) cells. C and D, RT-qPCR analysis of the mRNA levels of IP 3 R 1 in BPDE-treated Swan 71 (C) and HTR-8/SVneo (D) cells. E, Western blot analysis of the protein levels of IP 3 R 1 , CaMKII, p-CaMKII, and SGCB in BPDE-treated Swan 71 and HTR-8/SVneo cells, with GAPDH as internal standard. The relative intensity of each band was quantified. F, The sequence and complementary sequences of miR-hz04 and its binding site on lnc-HZ04 and IP 3 R 1 mRNA. G and H, RT-qPCR analysis of miR-hz04 level in BPDE-treated Swan 71 (G) and HTR-8/SVneo (H) cells. I-M, RT-qPCR analysis of the RNA levels of lnc-HZ04 (I), IP 3 R 1 (J), CaMKII (K), SGCB (L), and miR-hz04 (M) in HC and RM tissues (each n = 15). N, Western blot analysis of the protein levels of IP 3 R 1 , CaMKII, p-CaMKII, and SGCB in HC and RM tissues (each n = 10). The relative intensity of each band was quantified. Representative data in (E, I-N) represent three independent experiments. Data in (A-D, G-H) show mean ± SD of three independent experiments. Two-tailed Student's t-test for (I-M); one-way ANOVA for (A-D, G-H); *P < .05, **P < .01, and ***P < .001. HC, healthy control group; n, the number of biologically independent samples; RM, recurrent miscarriage group cytoplasm and nucleus in Swan 71 cells, as determined by FISH ( Figure S1B,C).
BPDE could induce trophoblast cell apoptosis. 20 The apoptosis of Swan 71 cells was significantly increased after BPDE treatment ( Figure S1D), agreed with our previous results. 20 It has been reported that IP 3 R 1 is an important protein that regulates the apoptosis of ovarian carcinoma cells 48 and also regulates calcium homeostasis in Purkinje cells. 49 In our mRNA sequencing data, we also found that the mRNA level of IP 3 R 1 was significantly increased with BPDE concentrations, with expression-level differences >2fold and P-values <.05 in both 0 versus 0.5 and 0.5 versus 1.5 groups ( Figure S1E). This upregulation was further confirmed by RT-qPCR analysis in BPDE-treated Swan 71 and HTR-8/SVneo cells ( Figure 1C,D). In signaling pathway, the mRNA and protein levels of CaMKII, p-CaMKII, and SGCB, which were located downstream of IP 3 R 1 , were all upregulated in BPDE-treated Swan 71 and HTR-8/SVneo cells ( Figures S1F,G and 1E). Since Ca 2+ is required to activate this pathway, 25,26 the cellular Ca 2+ level in Swan 71 cells was determined. It was found that the Ca 2+ level was increased with BPDE concentrations ( Figure S1H). Therefore, BPDE induced the trophoblast cell apoptosis by activating the cellular Ca 2+ -mediated IP 3 R 1 /p-CaMKII/SGCB signaling pathway. Sequence alignment using NCBI Blast showed that no regions in lnc-HZ04 sequence were complementary with the sense or antisense strand mRNA of IP 3 R 1 , CaMKII, or SGCB. Additionally, lnc-HZ04 is located in chr16, distinct from that of IP 3 R 1 in chr3, CaMKII in chr10, and SGCB in chr4, ruling out the possible cis regulation of IP 3 R 1 , CaMKII, or SGCB expression by lnc-HZ04. Moreover, the interactions of lnc-HZ04 with mRNA or protein of IP 3 R 1 , CaMKII, or SGCB were very weak, as analyzed by NCBI Blast and catRAPID (http://servi ce.tarta glial ab.com/page/ catra pid_group). 50,51 Furthermore, RIP assays also showed that none of IP 3 R 1 , CaMKII, p-CaMKII, or SGCB protein could interact with lnc-HZ04 ( Figure S1I). Therefore, these results implied that lnc-HZ04 seemed not to directly interact with the mRNA or protein of IP 3 R 1 /p-CaMKII/SGCB signaling pathway. Since IP 3 R 1 was located upstream of this signaling pathway, we tried to seek for some novel miRNAs that might interact with both lnc-HZ04 and IP 3 R 1 mRNA.
In miRNA sequencing data, total 44 novel differentially expressed miRNAs were identified, all of which were not annotated in NCBI database. After selecting the miRNAs with expression-level differences >2-fold and P-values <.05 in both 0 versus 0.5 and 0.5 versus 1.5 groups, four were monotonously increased and nine were monotonously decreased with BPDE concentrations. Then, these 13 miR-NAs were validated by RT-qPCR analysis in untreated or 0.5 μM BPDE-treated Swan 71 cells ( Figure S1J). Among these candidates, a novel miR-hz04 might interact with both lnc-HZ04 and IP 3 R 1 mRNA, because both of which harbored the binding site for miR-hz04 ( Figure 1F), as predicted by RNA22 (https://cm.jeffe rson.edu/rna22/ Inter activ e/). 40 The sequence of miR-hz04 has been submitted to NCBI with No. MW192196. MiR-hz04 was further confirmed to be downregulated in BPDE-treated Swan 71 and HTR-8/SVneo cells ( Figure 1G,H). Thus, lnc-HZ04 and miR-hz04 might involve in the upregulation of IP 3 R 1 /p-CaMKII/SGCB apoptosis pathway in BPDE-treated trophoblast cells.
It has been reported that the apoptosis of human trophoblast cells may lead to miscarriage. 18 In order to correlate the occurrence of unexplained miscarriage with BPDE-induced apoptosis of human trophoblast cells, and also to verify whether the regulation mechanisms are consistent in both trophoblast cells and villous tissues, we collected villous tissue samples from unexplained RM and healthy control (HC) groups (each n = 15), as described previously. 21,37,38 The RM group had experienced two or more consecutive unexplained miscarriage, and the known causes, such as chromosomal abnormalities, hormonal abnormalities, and uterine deformation, have been excluded. Then, the expression levels of lnc-HZ04, miR-hz04, and this IP 3 R 1 /p-CaMKII/SGCB signaling pathway were all detected in the HC and RM tissues. RT-qPCR and western blot analysis showed that lnc-HZ04, IP 3 R 1 , CaMKII, p-CaMKII, and SGCB were all significantly upregulated but miR-hz04 was downregulated in RM tissues relative to HC tissues ( Figure 1I-N). These changes were agreed with those found in BPDE-treated trophoblast cells, suggesting that both cases may share the same regulation mechanisms: lnc-HZ04 and miR-hz04 might involve in the regulation of IP 3 R 1 /p-CaMKII/SGCB apoptosis pathway in both RM tissues and BPDE-treated trophoblast cells.

| Upregulated lnc-HZ04 activated the Ca 2+ -mediated IP 3 R 1 /p-CaMKII/SGCB pathway to induce cell apoptosis in BPDEtreated trophoblast cells and in RM tissues
It has been identified that lnc-HZ04 was upregulated in both RM tissues and BPDE-treated trophoblast cells ( Figure 1A,B,I). Then, the roles of lnc-HZ04 in regulation of trophoblast cell apoptosis were explored. Lnc-HZ04 was overexpressed by transfecting pcDNA3.1-HZ04 and silenced by transfecting siRNA-HZ04 (si-HZ04) in untreated or 0.5 μM BPDE-treated Swan 71 or HTR-8/SVneo cells, and the efficiencies were validated by RT-qPCR analysis ( Figure S2A,B). The overexpression of lnc-HZ04 increased, whereas knockdown of lnc-HZ04 suppressed, the apoptosis of untreated or BPDE-treated Swan 71 cells (Figures 2A,B and  S2C,D), showing that lnc-HZ04 promoted trophoblast cell apoptosis. In pathway, the overexpression of lnc-HZ04 upregulated, whereas knockdown of lnc-HZ04 downregulated, the mRNA and protein levels of IP 3 R 1 , CaMKII, p-CaMKII, and SGCB in untreated or BPDE-treated Swan 71 or HTR-8/ SVneo cells (Figures 2C-J and S2E,F). Moreover, the overexpression of lnc-HZ04 also increased, whereas knockdown of lnc-HZ04 decreased, the cellular Ca 2+ level in both untreated and BPDE-treated Swan 71 cells ( Figures 2K,L and S2G,H). Thus, the upregulated lnc-HZ04 activated the Ca 2+ -mediated F I G U R E 2 Upregulated lnc-HZ04 activated the Ca 2+ -mediated IP 3 R 1 /p-CaMKII/SGCB pathway and induced the cell apoptosis in both BPDE-exposed human trophoblast cells and human RM tissues. A and B, Flow cytometry analysis of the apoptosis rate of untreated or 0. IP 3 R 1 /p-CaMKII/SGCB signaling pathway and induced the apoptosis of untreated or BPDE-treated trophoblast cells. Subsequently, lnc-HZ04 and this IP 3 R 1 /p-CaMKII/SGCB pathway were correlated in villous tissues. After normalization of their relative expression levels, IP 3 R 1 , CaMKII, p-CaMKII, and SGCB were all positively and linearly correlated with lnc-HZ04 in RM tissues ( Figures 2M, N and S2I-M). The data points were highly dispersed in RM group but were relatively compressed in HC group. The locations of data points in both groups were obviously separated, manifesting that these molecules were differentially regulated in both groups. The correlations among lnc-HZ04, IP 3 R 1 , p-CaMKII, and SGCB in RM tissues were similar to those found in BPDE-treated trophoblast cells, suggesting that both cases might share the same regulation mechanisms, in which the upregulated lnc-HZ04 activated this IP 3 R 1 /p-CaMKII/ SGCB pathway.

| Downregulated miR-hz04 activated the Ca 2+ -mediated IP 3 R 1 /p-CaMKII/SGCB pathway to induce trophoblast cell apoptosis in BPDE-treated trophoblast cells and in human RM tissues
Since miR-hz04 was downregulated in both RM tissues and BPDE-treated trophoblast cells ( Figures 1G,H,M), the roles of miR-hz04 were explored. MiR-hz04 was overexpressed by transfecting its mimics and silenced by transfecting its inhibitor in both untreated and 0.5 μM BPDE-treated Swan 71 or HTR-8/SVneo cells, and their efficiencies were validated by RT-qPCR analysis ( Figure S3A,B). The overexpression of miR-hz04 inhibited, whereas knockdown of miR-hz04 increased, the cell apoptosis in either untreated or BPDEtreated Swan 71 cells ( Figures 3A,B and S3C,D). For signaling pathway, the overexpression of miR-hz04 reduced, whereas knockdown of miR-hz04 increased, the mRNA and protein levels of IP 3 R 1 , CaMKII, p-CaMKII, and SGCB in untreated or BPDE-treated Swan 71 or HTR-8/SVneo cells ( Figures 3C-F and S3E-H). The overexpression of miR-hz04 also decreased the cellular Ca 2+ level and vice versa ( Figures 3G,H and S3I,J). Taken together, the downregulated miR-hz04 activated the Ca 2+ -mediated IP 3 R 1 /p-CaMKII/ SGCB pathway and induced the cell apoptosis in untreated or BPDE-treated trophoblast cells.
The correlation of miR-hz04 with IP 3 R 1 /p-CaMKII/SGCB pathway was also analyzed in RM tissues. After normalization of their relative levels, IP 3 R 1 , CaMKII, p-CaMKII, and SGCB were all negatively and linearly correlated with miR-hz04 in RM tissues, but were almost constant and independent of miR-hz04 levels in HC tissues (Figures 3I,J and S3K-O). The locations of data points in both groups were relatively separated. These results were same as those found in BPDE-treated trophoblast cells. Therefore, both cases might share the same mechanisms, in which the downregulated miR-hz04 might activate the Ca 2+ -mediated IP 3 R 1 /p-CaMKII/SGCB pathway and induce the trophoblast cell apoptosis.

| Lnc-HZ04 acted as a ceRNA for miR-hz04 and upregulated IP 3 R 1 expression level to inhibit trophoblast cell apoptosis
Since miR-hz04 might base pair with lnc-HZ04 or IP 3 R 1 mRNA, we hypothesized that lnc-HZ04 in the cytoplasm might function as a ceRNA (competitive endogenous RNA) for miR-hz04 and might diminish the inhibition effect of miR-hz04 on IP 3 R 1 expression level. First, lnc-HZ04 overexpression upregulated the mRNA and protein levels of IP 3 R 1 , an effect which could be diminished by overexpressing miR-hz04 in either Swan 71 or HTR-8/SVneo cells ( Figure 4A-F). In contrast, knockdown of miR-hz04 diminished the reduction in the mRNA and protein levels of IP 3 R 1 caused by silencing lnc-HZ04 ( Figure 4A-F). All these results supported that lnc-HZ04 may function as a ceRNA for miR-hz04 and diminish the inhibition effects of miR-hz04 on IP 3 R 1 expression level.
To further explore how lnc-HZ04 and miR-hz04 regulated IP 3 R 1 mRNA level, the mRNA stability of IP 3 R 1 was determined after blockage of mRNA transcription by adding actinomycin D. The overexpression of lnc-HZ04 enhanced IP 3 R 1 mRNA stability in Swan 71 and HTR-8/SVneo cells while the overexpression of miR-hz04 diminished this enhancement effect ( Figure 4G,H). Knockdown of lnc-HZ04 reduced IP 3 R 1 mRNA stability, which could be diminished by silencing miR-hz04 in both cells ( Figure 4J,K). As control, the mRNA stability of GAPDH was not changed after alteration of lnc-HZ04, miR-hz04, or both ( Figure S4A-D). For cell phenotype, the ability of lnc-HZ04 to increase cell apoptosis was diminished by overexpressing miR-hz04 (Figures 4I and  S4E). Knockdown of lnc-HZ04 reduced cell apoptosis, which was diminished by silencing miR-hz04 ( Figures 4L and S4F). Taken together, lnc-HZ04 acted as a ceRNA for miR-hz04 and diminished the inhibitory effects of miR-hz04 on IP 3 R 1 expression level and mRNA stability, and finally induced the trophoblast cell apoptosis.

| Lnc-HZ04 upregulated IP 3 R 1 level by direct and specific binding of miR-hz04 on its binding site on lnc-HZ04
The potential interactions of miR-hz04 with lnc-HZ04 or IP 3 R 1 mRNA were directly determined using lnc-HZ04 or IP 3 R 1 mRNA containing wild-type (wt) or mutant (mut) | 11 of 20 HUANG et Al.
target site for miR-hz04 ( Figure 5A). Dual-luciferase assays showed that miR-hz04 could directly interact with its target site on lnc-HZ04 or IP 3 R 1 mRNA, but not with its mutant site ( Figure 5B,C), indicating that miR-hz04 might interact with lnc-HZ04 or IP 3 R 1 mRNA by direct complementary binding onto its target site.     Next, Ago2 RIP assays showed that both lnc-HZ04 and IP 3 R 1 mRNA could bind with Ago2 protein but could not bind with IgG control ( Figure S5A,B), showing that both RNAs might interact with miRNA. MS2-RIP assays further confirmed that the endogenous miR-hz04 was enriched by lnc-HZ04 or IP 3 R 1 mRNA containing wt target site for miR-hz04 but not by those containing the mutant site or by empty vector (Figure 5D-F). MiR-502-3p, 52 which lacks the sequence that might be complementary with lnc-HZ04 or IP 3 R 1 mRNA, could not be pulled down by either wt or mut lnc-HZ04 or IP 3 R 1 mRNA. Subsequently, RNA pulldown experiments provided more solid evidence. MiR-hz04 could be pulled down by biotin-labeled wt lnc-HZ04 or IP 3 R 1 mRNA, but not by their mutant or the beads with empty vector ( Figure 5G-I). MiR-502-3p, as negative control, could not be pulled down by wt or mut lnc-HZ04 or IP 3 R 1 mRNA. Taken together, these results indicated that miR-hz04 could directly interact with lnc-HZ04 or IP 3 R 1 mRNA by directly binding onto its target site. Furthermore, the competitive binding of miR-hz04 with lnc-HZ04 or IP 3 R 1 mRNA was determined. Ago2 RIP assays showed that the overexpression of lnc-HZ04 increased its enrichment by Ago2 protein but decreased the enrichment of IP 3 R 1 mRNA ( Figure 5J). In contrast, knockdown of lnc-HZ04 reduced its enrichment by Ago2 but increased the enrichment of IP 3 R 1 mRNA ( Figure 5K). Similarly, the overexpression of IP 3 R 1 decreased, where knockdown of IP 3 R 1 increased, the lnc-HZ04 enrichment by Ago2 protein (Figure 5L,M). Moreover, RNA pulldown assays further showed that miR-hz04, which was enriched by wt lnc-HZ04 but not by its mutant, was reduced after overexpression of IP 3 R 1 mRNA but increased after knockdown of IP 3 R 1 mRNA ( Figure 5N,O). Similarly, miR-hz04, enriched by wt IP 3 R 1 mRNA, was reduced with lnc-HZ04 overexpression but increased with lnc-HZ04 knockdown ( Figure 5P,Q). Taken together, all these results supported that lnc-HZ04 and IP 3 R 1 mRNA competitively bound with miR-hz04 by direct binding of miR-hz04 on its target site.
Finally, the effects of miR-hz04 binding site at lnc-HZ04 on the mRNA and protein levels of IP 3 R 1 were further explored. The overexpression of wt lnc-HZ04, but not its mutant or empty vector, could upregulate the mRNA and protein levels of IP 3 R 1 in both Swan 71 and HTR-8/SVneo cells ( Figure 5R-T). Moreover, co-transfection of mutant lnc-HZ04 and miR-hz04 abolished the promotion effects contributed by wt lnc-HZ04 and showed only the inhibitory effects of miR-hz04 on the expression levels of IP 3 R 1 (Figure 5R-T). Taken together, these results showed that miR-hz04 target site at lnc-HZ04 played crucial roles in regulation of IP 3 R 1 expression level in human trophoblast cells.

| Transcription factor p53 mediated the lnc-HZ04 transcription in human trophoblast cells
The level of lnc-HZ04 was upregulated in BPDE-treated trophoblast cells ( Figure 1A,B). Then, the mechanisms underlying this upregulation were explored. It has been reported that p53 (tumor protein p53) is a notable transcription factor that activates its downstream target genes to suppress cellular proliferation or to permanently eliminate damaged cells. 53 As identified by PROMO software, p53 might recognize the promoter sequence of lnc-HZ04. The mRNA and protein levels of p53 were upregulated in BPDE-treated trophoblast cells ( Figure 6A-C). p53 was overexpressed by transfecting pcDNA3.1-p53 and silenced by transfecting siRNA-p53 (si-p53) in Swan 71 or HTR-8/SVneo cells, and the efficiencies were validated by RT-qPCR analysis and western blot analysis ( Figures 6D and S6A,B). Experimentally, the overexpression of p53 increased, whereas knockdown of p53 reduced, the expression level of lnc-HZ04 in both Swan 71 and HTR-8/ SVneo cells (Figures 6E,F and S6C,D). Moreover, p53 ChIP assays showed that p53 could bind to the promoter region of lnc-HZ04 ( Figures 6G and S6E), confirming that p53 may act as a transcription factor to promote lnc-HZ04 transcription. ChIP assays further showed that the occupancy of p53 on the promoter region of lnc-HZ04 was increased in BPDE-treated Swan 71 or HTR-8/SVneo cells compared with that in the untreated cells (Figures 6G and S6E). Therefore, these results showed that the transcription factor p53 might promote lnc-HZ04 transcription in human trophoblast cells. BPDE F I G U R E 5 Lnc-HZ04 upregulated the IP 3 R 1 expression level by direct and specific binding of miR-hz04 onto its binding site at lnc-HZ04. A-C, PmirGLO vector containing lnc-HZ04 (B) or IP 3 R 1 mRNA (C) sequence with wild-type (wt) or mutant (mut) binding site for miR-hz04, together with miR-hz04, was co-transfected into Swan 71 cells for dual-luciferase detection. D-F, RT-qPCR analysis of the level of miR-hz04 that was associated with lnc-HZ04 or IP 3 R 1 mRNA containing wt or mut miR-hz04 binding site in MS2-RIP assays using Swan 71 cells, with miR-502-3p as negative control. G-I, RT-qPCR analysis of the level of miR-hz04 pulled down by biotinylated lnc-HZ04 or IP 3 R 1 mRNA containing wt or mut miR-hz04 binding site in RNA pull-down assays using Swan 71 cells, with miR-502-3p as negative control. (J to M) RT-qPCR analysis of the level of lnc-HZ04 or IP 3 R 1 mRNA enriched by Ago2 protein using Swan 71 cells with overexpression or knockdown of lnc-HZ04 (J and K) or IP 3 R 1 mRNA (L and M) in Ago2 RIP assays. N-Q, RT-qPCR analysis of the level of miR-hz04 enriched by wt or mut lnc-HZ04 (N and O) or IP 3 R 1 mRNA (P and Q) in Swan 71 cells with overexpression or knockdown of IP 3 R 1 mRNA (N and O) or lnc-HZ04 (P and Q), respectively. R-S, RT-qPCR analysis of IP 3 R 1 mRNA level in Swan 71 (R) or HTR-8/SVneo (H) cells with overexpression of wt or mut lnc-HZ04 and miR-hz04 mimics or mimics NC. T, Western blot analysis of the protein level of IP 3 R 1 in Swan 71 or HTR-8/SVneo cells with overexpression of wt or mut lnc-HZ04 and miR-hz04 mimics or mimics NC. Representative data in (T) represent three independent experiments. Data in (B-C, E-F, H-S) show mean ± SD of three independent experiments. Two-tailed Student's t-test for (B-C, J-M); one-way ANOVA for (E-F, H-I, R-S); two-way ANOVA for (N-Q); ns, non-significance; *P < .05, **P < .01, and ***P < .001. inhibitor NC, negative control of inhibitor; mimics NC, negative control of mimics; miR-hz04 mimics, overexpression of miR-hz04; miR-hz04 inhibitor, knockdown of miR-hz04; NC, negative control of siRNA; pcDNA3.1-HZ04, overexpression of lnc-HZ04; pcDNA3.1-IP 3 R 1 , overexpression of IP 3 R 1 mRNA; Vector, empty vector of pcDNA3.1; si-HZ04, knockdown of lnc-HZ04; si-IP 3 R 1 , knockdown of IP 3 R 1 mRNA exposure may further upregulate p53 expression level and promote lnc-HZ04 transcription.
The overexpression of lnc-HZ04 reduced, whereas knockdown of lnc-HZ04 increased, the expression level of miR-hz04 in untreated or BPDE-treated trophoblast cells ( Figures 6H,I and S6F,G), supporting that lnc-HZ04 negatively regulated miR-hz04 level in human trophoblast cells. The overexpression of p53 reduced, whereas knockdown of p53 increased, the expression level of miR-hz04 in untreated or BPDE-treated trophoblast cells ( Figures 6J,K  S6H,I), supporting that BPDE exposure may upregulate the p53 expression level, promote the lnc-HZ04 transcription, and further downregulate the miR-hz04 expression level in BPDE-exposed human trophoblast cells.
In tissues, the mRNA and protein levels of p53 were upregulated in RM tissues relative to those in HC tissues ( Figures 6L and S6J). p53 ChIP assays showed that the occupancy of p53 on the promoter region of lnc-HZ04 was increased in RM tissues relative to that in HC tissues ( Figure 6M). The levels of p53 were positively and linearly correlated with the levels of lnc-HZ04 in RM tissues ( Figures 6N and S6K). Lnc-HZ04 was negatively and linearly correlated with miR-hz04 in RM tissues ( Figure 6O). Thus, the upregulated p53 may promote lnc-HZ04 transcription and the upregulated lnc-HZ04 may subsequently suppress miR-hz04 expression in RM tissues. Taken together, combined with the results in cellular assays, p53 might promote lnc-HZ04 transcription and upregulate lnc-HZ04 expression level, which further downregulated miR-hz04 level in RM tissues relative to those in HC tissues.

| Downregulated miR-hz04 activated murine IP 3 R 1 /Camk2g/Sgcb pathway in pregnant mice with BaP-induced miscarriage
Environmental BaP exposure might be one cause for the unexplained miscarriage. Then, a mouse model was made in which pregnant mice were treated with 0, 0.05, or 0.2 mg/ kg BaP to induce miscarriage, as described in our recent work. 21,37,38 BaP could be metabolized to ultimately carcinogenic BPDE in mice. The uteri were collected from every mouse in each group. The obvious embryo resorption and increased embryo resorption rates in BaP-treated mice supported that exposure of BaP to pregnant mice induced miscarriage. 21,37,38 IP 3 R 1 , CaMKII, and SGCB are conserved in humans, rhesus, elephants, dogs, and mice ( Figure S7A-C, Table S2), showing that this apoptosis pathway is evolutionarily conserved among these species. However, lnc-HZ04 sequence is conserved in humans and rhesus, but not in mice ( Figure S7D), indicating its specific regulation roles in humans. The levels of murine miR-hz04 were reduced ( Figure 7A), whereas those of murine IP 3 R 1 , Camk2g, and Sgcb were all increased ( Figure 7B-E), in BaP-treated mice group. After normalization of their relative levels, IP 3 R 1 , Camk2g, p-Camk2g, and Sgcb were all negatively and linearly correlated with miR-hz04 in placenta tissues in BaPtreated mice ( Figures 7F, G and S7E-I). The locations of data points in both groups were obviously separated, implying that this pathway was differentially regulated in untreated and BaP-treated groups. All these correlations in BaP-treated mice were the same as those found in human trophoblast cells and in human RM tissues, indicating that three cases may share the same pathway and regulation mechanisms: the downregulated miR-hz04 may activate the IP 3 R 1 /p-CaMKII/ SGCB pathway, promote the trophoblast cell apoptosis, and finally induce the miscarriage after BPDE/BaP exposure.

| DISCUSSION
It has been reported that some lncRNAs or miRNAs regulate the occurrence of miscarriage, such as lncRNA MEG8, 28 HOTAIR, 27 lnc-SLC4A1-1, 30 and miR-98. 32 However, all these noncoding RNAs were identified from the tissues of RM patients who have already miscarried. The dysfunctions of BPDE-exposed human trophoblast cells and the occurrence of unexplained miscarriage have rarely been correlated and their underlying mechanisms have rarely been explored. In our recent studies, we have found that lnc-HZ01 inhibits human trophoblast cell proliferation and induces miscarriage. 37 Lnc-HZ03 and miR-hz03 promote trophoblastic cell apoptosis and the occurrence of miscarriage. 21 Lnc-HZ08 regulates BPDE-induced trophoblast cell dysfunctions by F I G U R E 6 Transcription factor p53 mediated the lnc-HZ04 transcription. A and B, RT-qPCR analysis of the mRNA level of p53 in BPDEtreated Swan 71 (A) and HTR-8/SVneo (B) cells. C, Western blot analysis of the protein level of p53 in BPDE-treated Swan 71 and HTR-8/ SVneo cells, with GAPDH as internal standard. The relative intensity of each band was quantified. D, Western blot analysis of p53 protein level in Swan 71 or HTR-8/SVneo cells with overexpression or knockdown of p53. The relative intensity of each band was quantified. E and F, RT-qPCR analysis of lnc-HZ04 level in untreated or 0.5 μM BPDE-treated Swan 71 cells with overexpression (E) or knockdown (F) of p53. G, p53 ChIP assay analysis of p53 occupancy on lnc-HZ04 promoter region in untreated or 0.5 μM BPDE-treated Swan 71 cells. H and I, RT-qPCR analysis of miR-hz04 expression level in untreated or 0.5 μM BPDE-treated Swan 71 cells with overexpression (H) or knockdown (I) of lnc-HZ04. J and K, RT-qPCR analysis of miR-hz04 level in untreated or 0.5 μM BPDE-treated Swan 71 cells with overexpression (J) or knockdown (K) of p53. L, Western blot analysis of the protein levels of p53 in HC and RM tissues (each n = 10). The relative intensity of each band was quantified. M, p53 ChIP assay analysis of p53 occupancy on lnc-HZ04 promoter region in HC and RM tissues. N, The correlation between the levels of lnc-HZ04 and p53 protein in HC (blue) and RM (red) tissues (each n = 10). O, The correlation between the levels of lnc-HZ04 and miR-hz04 in HC (blue) and RM (red) tissues (each n = 15). Representative data in (C, D, L) represent three independent experiments. Data in (A, B, E-K, M) show mean ± SD of three independent experiments. Two-tailed Student's t-test for (M); one-way ANOVA for (A, B, E, F, H-K); two-way ANOVA for (G); Pearson analysis for (N, O); *P < .05, **P < .01, and ***P < .001. HC, healthy control group; n, the number of biologically independent samples; RM, recurrent miscarriage group promoting PI3K ubiquitin degradation and is also associated with miscarriage. 38 In this work, we identified that a novel lnc-HZ04 is upregulated and a novel miR-hz04 is downregulated in BPDE-exposed human trophoblast cells and in RM tissues. Environmental BaP might induce miscarriage by promoting trophoblast cell apoptosis through upregulating the IP 3 R 1 /p-CaMKII/SGCB pathway, which is regulated by lnc-HZ04 and miR-hz04.

F I G U R E 7
Murine miR-hz04 was downregulated and murine IP 3 R 1 /p-Camk2g /Sgcb pathway was upregulated in pregnant mice model with BaP-induced miscarriage. A-D, RT-qPCR analysis of the RNA levels of miR-hz04 (A), IP 3 R 1 (B), Camk2g (C), and Sgcb (D) in each mouse group (n = 9). E, Western blot analysis of the protein levels of IP 3 R 1 , Camk2g, p-Camk2g, and Sgcb in each group (n = 6) and their relative intensities were quantified. F and G, The correlation between the mRNA (F, n = 9) and protein (G, n = 6) levels of IP 3 R 1 and the levels of miR-hz04 in the control mouse group (0, blue) and 0.2 mg/kg BaP-treated mouse group (0.2, red). H, The regulation mechanism of lnc-HZ04/miR-hz04. Lnc-HZ04 acts as a ceRNA for miR-hz04 and diminishes the inhibition effects of miR-hz04 on IP 3 R 1 mRNA expression, thus activates the Ca 2+ -mediated IP 3 R 1 /p-CaMKII/SGCB pathway to promote trophoblast cell apoptosis. In normal trophoblast cells, relatively less lnc-HZ04 and more miR-hz04 inhibit this apoptosis pathway and lead to normal pregnancy. After exposure to BPDE, relatively more lnc-HZ04 and less miR-hz04 activate this apoptosis pathway and may further induce miscarriage. Representative data in (A-E) represent three independent experiments. One-way ANOVA for (A-D); Pearson analysis for (F-G); *P < .05, **P < .01, and ***P < .001. n represents the number of biologically independent mice  The regulation mechanisms of lnc-HZ04 and miR-hz04 have been proposed ( Figure 7H). MiR-hz04 directly binds to its target site at IP 3 R 1 mRNA and promotes IP 3 R 1 mRNA degradation. Lnc-HZ04 directly and competitively binds with miR-hz04, diminishes miR-hz04-inhibited IP 3 R 1 RNA stability and increases IP 3 R 1 expression level, and thus upregulates the Ca 2+ -mediated IP 3 R 1 /p-CaMKII/SGCB pathway, which further promotes trophoblast cell apoptosis. In normal trophoblast cells, lnc-HZ04 is downregulated and miR-hz04 is upregulated, which inhibit IP 3 R 1 /p-CaMKII/SGCB pathway and trophoblast cell apoptosis, giving normal pregnancy. If trophoblast cells were exposed to BPDE or in RM villous tissues, p53 is upregulated; and it promotes lnc-HZ04 transcription and upregulates lnc-HZ04 expression level. Then, the upregulated lnc-HZ04 downregulates miR-hz04, activates the IP 3 R 1 /p-CaMKII/SGCB pathway, promotes the trophoblast apoptosis, and might further induce the miscarriage. In pregnant mice model with BaP-induced miscarriage, the regulation of murine IP 3 R 1 /p-Camk2g/Sgcb pathway by miR-hz04 is consistent with those found in BPDE-treated human trophoblast cells and in RM tissues, suggesting that environmental BaP induces miscarriage possibly by upregulating the Ca 2+ -mediated IP 3 R 1 /p-CaMKII/SGCB apoptosis pathway. In human system, an extra lnc-HZ04 is involved in the epigenetic regulation of this pathway, showing more complicated regulation network in human system than that in mice.
In general, lncRNAs act as ceRNAs to rescue miRNAsinhibited target mRNA levels and upregulate mRNA-related cellular functions. 54,55 For example, by acting as an endogenous "sponge," lncRNA-HULC downregulates miR-372 to reduce the translational repression of its target gene PRKACB, which induces phosphorylation of CREB and modulates selfregulation in hepatocellular cancer. 56 Additionally, lncRNA-ATB upregulates ZEB1 and ZEB2 by competitively binding to the miR-200 family and then induces epithelial-mesenchymal transition and invasion in hepatocellular carcinoma. 14,57 On the contrary, miRNAs can also regulate the expression levels of lncRNAs. 58 MiR-150 negatively modulates the expression of lncRNA-MIAT and suppresses the proliferation, migration, and tube formation of human microvascular endothelial cells. 59 In this work, we revealed that lnc-HZ04 functions as a ceRNA for miR-hz04, directly interacts with miR-hz04, and further reduces miR-hz04 level. At the same time, miR-hz04 also downregulates lnc-HZ04 level by promoting lnc-HZ04 degradation. Finally, lnc-HZ04 diminishes the inhibition effect of miR-hz04 on the mRNA expression level of IP 3 R 1 and promotes the trophoblast cell apoptosis. IP 3 /Ca 2+ /CaMK signaling pathway has been reported to modulate cell proliferation of cholangiocytes. 60 Ca 2+ / CaMK signaling cascade could modulate the expression and exon selection of IP 3 R 1 and contributes the migration and synaptic refinement in granule cells from rat cerebellar cortex. 61 The downstream CaMKII and sGC were also reported to inhibit cell cycle progression in the prefrontal cortex of mice 62 and to participate in spinal reflex potentiation in anesthetized rats. 63 In this work, we revealed that the upregulated IP 3 R 1 could activate cellular Ca 2+ -mediated CaMKII/SGCB pathway to induce trophoblast apoptosis in BPDE-treated trophoblast cells and in human RM villous tissues. The biomolecules in this pathway, as well as the intracellular-free Ca 2+ ion, might serve as miscarriage biomarkers.
It should be stressed that it is trophoblast cell apoptosis but not BPDE itself that directly induces miscarriage. Once the IP 3 R 1 /p-CaMKII/SGCB apoptosis pathway is activated, the upstream BPDE would become less important for the downstream miscarriage. More broadly, not only BaP or BPDE, but also other factors or pathways, such as cadmium 64 or hexavalent chromium (Cr VI ), 65 could also induce human trophoblast cell apoptosis, and might possibly induce miscarriage. Therefore, BPDE is only one of the upstream environmental factors that may induce trophoblast cell apoptosis; and it is trophoblast cell apoptosis that may directly induce miscarriage.
In summary, we have identified a novel lnc-HZ04. BPDE exposure promotes p53-mediated lnc-HZ04 transcription and upregulates lnc-HZ04 expression level. Lnc-HZ04 acts as a ceRNA for a novel miR-hz04, promotes trophoblast cell apoptosis by upregulating Ca 2+ -mediated IP 3 R 1 /p-CaMKII/SGCB signaling pathway, and might further induce miscarriage. This pathway serves as a bridge to connect BPDE-induced trophoblast cell apoptosis and the occurrence of miscarriage, providing novel scientific and clinical understanding of the occurrence of unexplained miscarriage.