MicroRNA-548 regulates high mobility group box 1 expression in patients with preterm birth and chorioamnionitis

High mobility group box 1 (HMGB1) is a prototypic alarmin and plays an important role in the pathogenesis of inflammatory process in spontaneous preterm birth. This study was conducted to compare the levels of HMGB1 in amniotic fluid and amnion membranes in women with chorioamnionitis/intra-amniotic inflammation to the levels in healthy controls. We also aimed to elucidate the involvement of microRNA-548 (miR-548) in regulating HMGB1 expression and its function in human amniotic epithelial cells (hAECs). A bioinformatics analysis predicted the binding of HMGB1 by the miR-548 cluster. A repressed and forced expression assay in hAECs was performed to investigate the causal relationship between the miR-548 cluster and HMGB1. The levels of HMGB1 in amniotic fluid and amnion membranes were significantly higher in patients with intra-amniotic inflammation/chorioamnionitis than in those without inflammation. The miR-548 was significantly under-expressed in amnion membranes from patients with chorioamnionitis than in normal term controls. Repressed expression of miR-548 up-regulated HMGB1 expression in hAECs and increased its release from hAECs. Moreover, forced expression of miR-548 suppressed HMGB1 and inflammatory cytokines in hAECs, which increased when treated with lipopolysaccharide. These results suggest miR-548 can alter the inflammatory responses in hAECs, and might be involved in the pathogenesis of preterm birth by regulating HMGB1.


Results
HMGB1 expression in AF and amnion membrane. We analyzed HMGB1 levels in AF obtained from women with signs or symptoms of preterm birth who had amniocentesis to rule out infection (n = 38) or for either genetic indication (n = 10). Table 1 shows a comparison of clinical characteristics of patients with amniocentesis. Thirty four of 48 patients were diagnosed with IAI. The median [interquartile range] AF concentration of HMGB1 was significantly higher in patients with IAI than in those without IAI (563. 8  The expressions of HMGB1 mRNA and protein were increased in amnion membranes from women with preterm birth and acute chorioamnionitis as compared to normal controls ( Fig. 2A). To confirm these data, we measured the expression level of HMGB1 in primary human amnion epithelial cells (hAECs) isolated from amnion membrane obtained from patients who delivered preterm due to acute chorioamnionitis (patient-derived hAECs, PD-hAECs) and women who had full-term pregnancy as normal control (hAECs). As expected, HMGB1 expression in PD-hAECs was significantly higher than in hAECs (Fig. 2B). Additionally, HMGB1 release from PD-hAECs was highly augmented compared to hAECs (Fig. 2C). These date implicated HMGB1 as a novel candidate biomarker for the early prediction of preterm birth determined by chorioamnionitis.
MiR-548 cluster is under-expressed in amnion membrane from patients with preterm birth and acute chorioamnionitis. Several miRNAs, including the miR-548 cluster, are reportedly downregulated in spontaneous preterm birth cases compared to controls 39 . Thus, we investigated the correlation between HMGB1 and the miR-548 cluster. Several binding sites of the miR-548 cluster were computationally predicted in the 3′UTR of HMGB1 (Table 2). To explore whether the miR-548 cluster could be a negative regulator of HMGB1 in preterm birth with acute chorioamnionitis, the expression level of the miR-548 cluster, including miR-548aa, miR-548ai, miR-548-3p, and miR-548x-5p, was measured using qPCR in the amnion membrane. Also, we investigated miR-302b, which displays reduced expression in spontaneous preterm birth 39 . As shown in Fig. 3A, all members of the miR-548 cluster were significantly decreased in amnion membranes from preterm women with acute chorioamnionitis compared to those from normal term controls. Furthermore, the expression of the miR-548 cluster was reduced in PD-hAECs than in hAECs from normal healthy women (Fig. 3B). However, there was no significant difference in the expression level of miR-302b between preterm and normal controls (data not shown). These results suggested that miR-548 cluster could be an inhibitory regulator of HMGB1 expression in preterm birth with acute chorioamnionitis.
www.nature.com/scientificreports www.nature.com/scientificreports/ MiR-548 cluster can regulate expression of HMGB1 in hAECs. To validate the correlation between HMGB1 and the miR-548 cluster in patients with preterm birth and acute chorioamnionitis, we investigated whether the expression level of the miR-548 cluster is crucial to the HMGB1 expression level using the synthetic mimics and inhibitors for the miR-548 cluster (Table 3) in hAECs. To examine whether the downregulation of miR-548 cluster could induce the enhancement of HMGB1 expression, hAECs were transfected with miR-548aa inhibitor, miR-548ai inhibitor, miR-548a-3p inhibitor, or miR-548x-5p inhibitor followed by qPCR (Fig. 4A). Inhibitors of the miR-548 cluster obviously increased the expression of both mRNA and protein levels of HMGB1 in hAECs compared to control inhibitor (Fig. 4B). Moreover, HMGB1 levels increased in the culture medium of hAECs transfected with miR-548 inhibitor compared to those transfected with control inhibitor (Fig. 4C). These findings indicated that downregulation of the miR-548 cluster contributed to the overexpression of HMGB1 expression in preterm birth with acute chorioamnionitis. In addition, the negative correlation between the expression of the miR-548 cluster and HMGB1 level was implicated in preterm birth with acute chorioamnionitis.

Lipopolysaccharide-induced inflammation induces enhanced HMGB1 expression through downregulation of miR-548 cluster.
To evaluate the role of the negative correlation between the miR-548 cluster and HMGB1 in preterm birth with acute chorioamnionitis, hAECs were treated with lipopolysaccharide (LPS) (10 or 100 ng/mL) for 24 h to induce inflammation followed by measurement of the expression levels of the miR-548 cluster and HMGB1. Compared to control, LPS inhibited the expression of the miR-548aa and miR-548ai and led to increases of HMGB1 mRNA and protein in hAECs (Fig. 5A,B). A significant elevation of HMGB1 levels was also observed in the culture medium of LPS-stimulated hACEs as compared to normal control medium (Fig. 5C). These findings indicated that downregulation of the miR-548 cluster promoted HMGB1 expression in hAECs and its subsequent release from the cells upon LPS-mediated inflammation.

MiR-548 cluster can attenuate LPS-induced inflammation in hAECs.
To verify the effect of the miR-548 cluster on LPS-induced expression and release of HMGB1, hAECs were transfected with miR-548aa mimic, miR-548ai mimic, miR-548a-3p mimic, miR-548x-5p mimic, or control mimic. HMGB1 mRNA and protein expressions were pronounced in hAECs transfected with control mimic stimulated by LPS. In contrast, overexpression of the miR-548 cluster in cells transfected with miR-548 cluster mimics lowered HMGB1 induction upon LPS exposure (Fig. 6A). Moreover, HMGB1 levels were decreased in the culture medium of the miR-548 mimic transfected hAECs as compared to control mimic cells (Fig. 6B). Those findings indicated that the miR-548 mimic inhibited HMGB1 up-regulation in hAECs and the subsequent release of HMGB1 from hAECs exposed to an inflammatory stimulus. Furthermore, release of inflammatory cytokines (IL-1ß and IL-6) and chemokine (IL-8) was significantly reduced in the culture medium of hAECs transfected with miR-548 cluster mimics compared to control mimic cells. The data indicated that the overexpression of miR-548 cluster can suppress the inflammatory responses in hAECs.

Discussion
This study demonstrates that the HMGB1 levels in AF from the patients with IAI were significantly higher than from those without IAI, and that HMGB1 was highly expressed in amnion membranes of patients with preterm birth and acute chorioamnionitis as compared to those in normal term controls. In addition, we validated that LPS-mediated inflammation could be responsible of the observed HMGB1 up-regulation in hAECs and for the release of HMGB1 from hAECs in vitro. Our findings, confirmed both in vivo and in vitro, were consistent with the fact the inflammation induced the expression of HMGB1 in the amnion membrane and its release from the amnion epithelial cells. HMGB1 is actively released by stimulation of the innate immune system with exogenous pathogen-derived molecules. Neurons, astrocytes, erythroleukemia cells, neuroblastoma cells, and other tumor cells are also reportedly able to actively secrete HMGB1 12,[27][28][29]40 . Passive release of HMGB1 can be initiated when cellular integrity is damaged 32 . Considering that HMGB1 expression in hAECs increased when the cells were stimulated by LPS, HMGB1 could be released actively from amnion epithelial cells. Moreover, inflammation of www.nature.com/scientificreports www.nature.com/scientificreports/ the amnion membrane might lead to injury of amnion epithelial cells, which might cause the passive release of HMGB1. Regardless of the mechanism, the elevated levels of HMGB1 in AF could be considered a marker for amnion epithelial cell damage and injury, and representative of active signal transduction in amnion epithelial cells. It remains unclear how the released HMGB1 affects the amnion epithelial cells and plays a role in leading to preterm birth. We speculate that HMGB1 might induce cytokine, chemokine, and metalloproteinase synthesis that might be involved in preterm birth.
Little is known about the involvement of the miR-548 family in disease processes. Downstream targets of the miR-548 family are involved in breast cancer, inflammatory responses, and potential estrogen receptor sensitivity, indicating a potential biological role in the setting of preterm birth [41][42][43] . A recent study reported the suppression of endogenous miR-548 levels in the setting of viral infections 42 . Similarly, we observed that significant  www.nature.com/scientificreports www.nature.com/scientificreports/ suppression of the miR-548 cluster was observed in the amnion membrane from patients with preterm birth and acute chorioamnionitis compared to normal term controls. Moreover, when hAECs were treated with LPS, miR-548 was downregulated. A previous study reported that inflammation does alter miRNAs at intestinal epithelial barriers and regulates intestinal permeability by degradation of tight junction protein, which is the target protein of miRNAs 44 . Furthermore, in cervical cells, inflammation can significantly alter the expression of specific miRNAs; these alterations might be involved in cervical remodeling and preterm birth 45 . Although the mechanism of miR-548 downregulation following inflammatory stimuli has not been clarified, our results suggest that under-expressed miR-548 can play an important role in acute chorioamnionitis determined preterm birth.
This study demonstrates that LPS downregulated the expression of the miR-548 in hAECs, and the inhibition of miR-548 induced the up-regulation of HMGB1 mRNA and protein expression, and release of HMGB1 in the culture medium of hAECs transfected with miR-548 cluster inhibitors. These findings suggest that IAI can induce downregulation of the miR-548 in the amnion membrane, thus leading to the up-regulated expression of HMGB1 in amnion and its release from amnion epithelial cells. Moreover, the miR-548 cluster mimic could reverse the LPS-induced up-regulation of HMGB1 in hAECs compared to control mimic. The release of HMGB1 from hAECs was also significantly decreased in hAECs transfected with miR-548 cluster mimics. In addition, the release of inflammatory cytokines and chemokine was suppressed in the culture medium of hAECs transfected with miR-548 cluster mimics compared to control mimic cells. These findings suggest that miR-548 cluster mimics can inhibit the up-regulation of HMGB1 that occurs when inflammation is stimulated, and alters inflammatory responses in hAECs. Thus, we can speculate that miR-548 could effectively reverse and prevent the activation of innate immunity, and significantly attenuate damage in IAI.
A key strength of this study is our demonstration that amniotic epithelial cells can release HMGB1. HMGB1 is increased in the AF of women with preterm birth and IAI. Further, increased AF HMGB1 can cause preterm birth. However, the source of AF HMGB1 had not been determined. Our findings identified the important role of the amnion membrane in preterm birth by releasing HMGB1 and other proinflammatory cytokines during IAI. In addition, we elucidated that there is a microRNA/target protein network involved in preterm birth with IAI/chorioamnionitis. A few studies have reported that there are miRNAs differentially expressed in women with preterm birth; however, little is known about the role of miRNAs in preterm birth and how their involvement may induce preterm birth. We have demonstrated that inflammation can alter miR-548 expression in the amnion membrane and suppression of miR-548 can enhance inflammatory responses in the AF, leading to preterm birth.  Table 2. Predicted sequence alignment between miRNAs and HMGB1 3′UTR.

Figure 3. MicroRNA-548 cluster expression in amnion membrane and human amniotic epithelial cells (hAECs). (A)
Levels of miRNA-548aa, miRNA-548ai, miRNA-548a-3p, and miRNA-548x-5p in amnion membranes from women with preterm birth and chorioamnionitis and from normal term controls. (B) Levels of miRNA-548aa, miRNA-548ai, miRNA-548a-3p, and miRNA-548x-5p in patient-derived hAECs (PD-hAECs) and hAECs from normal healthy women. **Significant difference (P < 0.01) compared with control group, *Significant difference (P < 0.05) compared with control group. hAECs, human amniotic epithelial cells.  www.nature.com/scientificreports www.nature.com/scientificreports/ of proinflammatory cytokines and chemokines secreted during LPS treatment was directly caused by altered miRNA expression or secondary to the secretion of HMGB1. Moreover, the gestational age of obtaining amniotic fluid and amnion could not be matched between the patient and control groups. Therefore, the difference in HMGB1 concentration between the patient and control groups may have been affected by differences in gestational age. Although, one previous study has reported no significant difference in the median AF HMGB1 concentration between women in the mid-trimester and those at term not in labor 11 , further studies are required to examine the HMGB1 concentration in AF and amnion membrane throughout pregnancy.

Mimics or inhibitors
In conclusion, intra-amniotic inflammation induces the up-regulation of HMGB1 expression in the amnion membrane and HMGB1 release via suppression of miR-548. Moreover, miR-548 can attenuate inflammatory responses in amnion epithelial cells. Although the specific regulatory mechanism of miR-548 on HMGB1 expression remains unclear, these data demonstrate a potential molecular mechanism underlying the inhibitory effect of miR-548 on preterm birth, and implicate miR-548 as a potential therapeutic target in preterm birth with chorioamnionitis.

Methods
Patients and amniotic fluid collection. We investigated AF samples from 48 women with singleton pregnancies and clinically indicated amniocentesis. Samples were obtained by transabdominal amniocentesis for second trimester genetic karyotyping (n = 10) or to rule out AF infection in women who had preterm labor contractions refractory to tocolysis or advanced cervical dilatation (n = 38). Exclusion criteria were the presence of anhydramnios, human immunodeficiency or hepatitis viral infections, congenital anomalies, or abnormal karyotype. Gestational age was determined based on the last menstrual period confirmed by ultrasound examination www.nature.com/scientificreports www.nature.com/scientificreports/ prior to 14 weeks gestation. Preterm labor was defined as the presence of regular uterine contractions and documented cervical effacement and/or dilatation in patients before week 37 of gestation. IAI was defined by amniotic fluid interleukin (IL)-6 concentrations ≥2,600 pg/mL 46 . All women provided written informed consent before the collection of AF and placenta samples. The collection and utilization of the samples were approved by the Ethics Committee of Kangnam Sacred Heart Hospital (Ethics ref.: 2017-08-014). All experiments were performed in accordance with the relevant guidelines and regulations.
Chemical and microbiological studies of AF. Following retrieval under sterile conditions, AF was analyzed for glucose concentration, white blood cell (WBC) count, gram stain, and for standard culture of aerobic and anaerobic bacteria, including Ureaplasma and Mycoplasma species. The remaining AF was aliquoted into 2-mL vials and cryopreserved at −80 °C.
Enzyme-linked immunosorbent assay (ELISA) of AF HMGB1 and IL-6. The levels of IL-6 and HMGB1 in AF and cell culture medium (conditioned medium) were determined by an ELISA assay kit (Cusabio, Wuhan, China) according to the manufacturer's instructions. The levels of IL-1ß and IL-8 in culture medium were determined by an ELISA assay kit (Cusabio and RayBiotech, Norcross, GA, USA, respectively) according to the manufacturer's instructions.
Amnion collection and processing. Placentas (n = 6) were obtained from women with suspected chorioamnionitis caused by prolonged latency after preterm premature rupture of membranes and/or preterm labor (median gestational age of 29.4 weeks, range 25.3-35.0 weeks), and from healthy women (n = 4) undergoing elective cesarean delivery at term in the absence of labor (median gestational age 38.1 weeks, range 20.1-39.0 weeks). Amnion membrane was manually stripped from the chorion and divided into small pieces (approximately 1.0 cm 3 each). After sample collection, the remaining full-thickness placental biopsies were obtained and fixed in formalin. Biopsy specimens were interpreted by a placental pathologist who was blinded to the clinical and laboratory findings. Amnion samples histologically diagnosed with acute chorioamnionitis were used as patients' samples in the experiments. www.nature.com/scientificreports www.nature.com/scientificreports/ Isolation of amniotic epithelial cells. Amnion membranes (n = 5, median gestational age 38.3 weeks, range 37.4-39.0 weeks) were obtained from women who had completed full-term pregnancy during elective cesarean deliveries before the onset of labor and from those (n = 5, median gestational age 23.3 weeks, range 18.3-34.5 weeks) with preterm birth and acute chorioamnionitis. The primary hAECs were isolated from freshly obtained amnion as described previously [47][48][49][50] . Briefly, the amnion membrane was washed with ice-cold phosphate buffered saline (PBS, pH 7.2) to remove blood clots and cellular debris, and cut into pieces approximately 6 cm long. The first enzymatic digestion was performed by incubating the membrane pieces with 20 mL of pre-warmed 0.05% trypsin/EDTA (Thermo Fisher Scientific, Waltham, MA, USA) at 37 °C for 10 min with gentle shaking. The cells obtained at this step were discarded to remove blood clots and cellular debris. Second and third digestions were conducted the same way. The membrane pieces were transferred into new 50-mL conical tubes containing 20 mL 0.05% trypsin/EDTA and incubated at 37 °C for 30 min with gentle shaking. The digestion mixtures obtained from the second and third digestions were filtered through a 70-μm cell strainer and centrifuged at 500 × g for 10 min. Cell pellets were suspended and grown in DMEM/F12 (Gibco, Franklin Lakes, NJ, USA) supplemented with 10% fetal bovine serum, 100 U/mL penicillin, 100 μg/mL streptomycin (Sigma-Aldrich, St. Louis, MO, USA), and 10 ng/mL epidermal growth factor (R&D Systems, Madison, WI, USA).
Cell culture and reagents. The primary hAECs from amnion tissues were plated and maintained in culture at 37 °C in a humidified incubator in an atmosphere of 5% CO 2 . JetPrime transfection reagent was purchased from PolyPlus-Transfection (Strasbourg, France). Antibodies to HMGB1 (Cell Signaling Technology, Beverly, MA, USA) and Actin (Santa Cruz Biotechnology, Santa Cruz, CA, USA) were used. LPS from Escherichia coli strain 0111:B4 was obtained from Sigma-Aldrich.
Western blot analysis. Western blot analysis was performed according to standard procedures 51 . Equal amounts of proteins were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene fluoride membranes (Thermo Fisher Scientific). Target proteins were detected by enhanced chemiluminescence reagents (Thermo Fisher Scientific). www.nature.com/scientificreports www.nature.com/scientificreports/ MicroRNA prediction of target genes and functional and bioinformatics analysis. Binding sites and sequences of miR-548 cluster on the HMGB1 3′UTR were predicted by the target prediction programs TargetScan (http://www.targetscan.org), MiRDB (http://mirdb.org), and miRmap (https://mirmap.ezlab.org). Paired sequence alignment is marked by red color and continuous lines ( Table 2).
Forced and repressed expression of miRNA-548 cluster. To verify the effect of miRNAs, miR-548 inhibitor, mimic, and negative control groups were transfected using JetPrime (Polyplus-Transfection). Total RNA, including miRNA and proteins, were isolated at 36 h post-transfection and were used for western blot analysis and qPCR. The sequences of miR-inhibitor and mimic are listed in Table 2. hAECs were incubated in the presence of 10 and 100 ng/mL LPS to induce inflammation. Culture medium was collected after 24 h of incubation to measure the levels of HMGB1.

Statistical analyses.
All experiments were repeated at least three separate times. Statistical analyses were performed using SPSS version 18.0 (SPSS Inc., Chicago, IL, USA). Data were compared with Student's t test for parametric data or Mann-Whitney U test for nonparametric data. The Kruskal-Wallis test with Bonferroni corrections was used for multiple comparisons. Spearman correlations were used to measure co-linearity between the selected independent variables. Comparisons between proportions were done with Fisher's exact test. Statistical analysis of the immunoassays data was performed after logarithmic transformation of data. Statistical significance was indicated when p < 0.05.