Mouse anti-human FGL1 monoclonal antibody, goat anti-mouse FGL1 polyclonal antibody, purified recombinant human FGL1 (the similarity of FGL1 between humans and mouse is greater than 80%), and an FGL1 enzyme-linked immunosorbent assay (ELISA) kit were obtained from R&D Systems (Minneapolis, MN, United States). Antibodies recognizing actin, cytokeratin 7 (CK7), IL-6, and PlGF were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, United States), and 4′,6-diamidino-2-phenylindole hydrochloride (DAPI) was purchased from ATT Bioquest (CA, United States). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH, AM4300) was purchased from Thermo Fisher Scientific (Waltham, MA, United States). Rabbit anti-human FoxO3a monoclonal antibody was purchased from Cell Signaling Technology (Danvers, MA, United States). The FGL1 expression plasmid (pCMV6- FoxO3a) was purchased from Origene (Rockville, MD). All remaining reagents, unless otherwise specified, were obtained from Invitrogen (Carlsbad, CA, United States).

From February 3, 2017, to July 31, 2020, in the Department of Obstetrics and Gynecology of NCKUH, we prospectively collected 38 cases of PE meeting the ACOG criteria (2019) and 42 healthy controls. PE was defined as the development of hypertension in a previously normotensive pregnant woman after 20 weeks of gestation, accompanied by new-onset proteinuria. Hypertension was defined as systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg on at least two occasions and 4–6 h apart. Proteinuria was defined as ≥300 mg protein excretion in the 24-h urine collection or a protein concentration of ≥300 mg/L in urine (≥1 + on dipstick). Plasma samples of all participants were obtained at the time of admission for delivery. Human placentas used in this study were obtained at the end of gestation, immediately after vaginal delivery or during cesarean section. The tissues were harvested by penetrating the center of the placenta (umbilical cord insertion) and snap-frozen in liquid nitrogen, and others were fixed-embedded in paraffin for immunohistochemistry. The decidua regions were performed to RT-PCR and Western blotting. During blood pressure measurement, patients were resting in the supine position in a quiet environment, and measurements were obtained in a fasting state between 08:00 and 10:00 AM. None of the healthy controls had clinical signs of PE or other medical or pregnancy complications. The exclusion criteria were as follows: multiple pregnancies, premature rupture of membranes, chorioamnionitis, chronic hypertension, diabetes mellitus, autoimmune disorders, and fetal abnormalities.

The human trophoblast cell lines 3A-subE (ATCC CRL-1584) and HTR-8/SVneo (ATCC CRL-3271) were obtained from the American Type Culture Collection (ATCC, Manassas, VA, United States), cultured in growth media [high-glucose DMEM containing 10% heat-inactivated fetal bovine serum (FBS), 1% antibiotics (penicillin and streptomycin; Invitrogen), and L-glutamine (292 mg/L)] and maintained at 37°C with 5% CO₂. For transient transfections, Lipofectamine 2000 reagent (Invitrogen) was used according to the manufacturer’s instructions. The cells were collected for further experiments after transfection with plasmids (2 mg) for 48 h. Cells (2  ×  10⁵ cells/well) were seeded in 6-well plates, and after incubation with the indicated dose of L-NAME (0, 100, 1,000 μM) for 24 h, the cells were further analysed.

Adult pregnant mouse (body weight: 21–23 g), the C57BL/6 inbred strains, were purchased from the BioLASCO Animal Center (Taipei, Taiwan, ROC) and received in our animal facilities on day 7 of pregnancy (day 0 = day of positive sperm smear). All animals were given free access to food and water. For the PE-like mouse model, L-NAME (75 mg/kg/day) was subcutaneously administered with or without recombinant FGL1 (1 μg/kg/day) on days 8–14 of pregnancy (FGL1 and non-FGL1 groups, respectively) based on a previous report (Motta et al., 2015). Both the control and treatment groups of mouse were age-matched, and saline was administered to the control groups. Blood pressure was noninvasively measured during pregnancy on gestational days (GDs) 8–20. The mouse were placed in metabolic cages for urine collection on GDs 12–15 and 15–18. The entire placenta without yolk sac and uterine layers were collected at GDs 19 for RNA isolation and Western blotting. The animal care and experimental procedures conformed to the ARRIVE guidelines (Kilkenny et al., 2010) and were approved by the Institutional Animal Care and User Committee (150213) at NCKU, Tainan, Taiwan.

Total RNA from mouse placentas was extracted using the RBC Total RNA Kit (Life Biomedical Limited) according to the manufacturer’s instructions. Next, 2 μg of extracted RNA was reverse-transcribed to cDNA using MuLV reverse transcriptase, 1 mM dNTPs, and 0.5 mg/ml oligo (dT12–18) as previously described (Cheng et al., 2009). cDNA was amplified through qRT-PCR using Fast Quant Green Master Mix with Low ROX (Protech Technology Enterprise, WI, United States) with 0.5 μM primers according to the manufacturer’s protocol. The sequences of the primers are as follows: FGL1 (F: 5′-AGC​TGC​CTG​TGT​TAT​TTC​CTC​TCA-3′; R: 5′-CCA​GGG​AGC​CAT​TTT​ATT​TAT​CCA​A-3′) (Demchev et al., 2013), interleukin (IL)-1b (F: 5′-CAC​CTC​TCA​AGC​AGA​GCA​CAG -3′; R: 5′-GGG​TTC​CAT​GGT​GAA​GTC​AAC-3′), IL-6 (F: 5′-TCC​TAC​CCC​AAC​TTC​CAA​TGC​TC-3′; R: 5′-TTG​GAT​GGT​CTT​GGT​CCT​TAG​CC-3′) (Iwashita et al., 2014), placenta growth factor (PlGF) (F: 5′-TCT​GCT​GGG​AAC​AAC​TCA​ACA-3′; R: 5′-GTG​AGA​CAC​CTC​ATC​AGG​GTA-3′) (Fu et al., 2020), FMS-like tyrosine kinase 1 (Flt-1) (F: 5′-AGG​AGA​TGC​TCC​TCC​CAA​A-3′; R: 5′-GTG​CAG​GGA​TCC​TCC​AAA​T-3′) (Zhou et al., 2007), and GAPDH (F: 5′- TGC​ACC​ACC​AAC​TGC​TTA​GC-3′; R: 5′- GGC​ATG​GAC​TGT​GGT​CAT​GAG-3′) (Xie et al., 2012). GAPDH gene expression was used as an internal control, and the 2 − ΔΔCt method was used to analyze the relative gene expression (Livak and Schmittgen, 2001).

Histochemical analysis of 5-μm-thick sections was performed as described previously (Kang et al., 2020). Slides were counterstained using hematoxylin and eosin (Gill No. 1, Sigma-Aldrich), dehydrated in an ascending series of ethanol, cleared in xylene, and mounted with Cytoseal (Thermo Fisher Scientific). The center of the mouse placenta disc was marked with ink and sections were cut through the center. Photographs were obtained using an Olympus DP72 camera mounted on a BX61 Olympus microscope. The images were quantified using ImageJ software (National Institutes of Health) to analyze the spongiotrophoblast area.

Cells were fixed in 2% (w/v) formaldehyde and blocked with 10% (v/v) goat serum. Sections were dewaxed and rehydrated, followed by antigen retrieval. Peroxidase method: Sections exposed to the primary antibody for 30 min at room temperature, washed in phosphate-buffered saline (PBS), incubated for 30 min with rabbit-anti-mouse IgG conjugated with horseradish-peroxidase, then staining was achieved by incubating with 3,3′-diaminobenzidine (DAB; Sigma). Fluorescence method: One percent (w/v) Sudan black in 70% (v/v) ethanol was used for 1 min to block autofluorescence. Goat serum (10%, v/v) was applied for 1 h at room temperature for nonimmune blockade, followed by overnight incubation with mouse FGL1 polyclonal antibody (Sigma, diluted 1:100) at 4°C. Goat anti-mouse secondary antibody conjugated with Alexa Fluor-488 (Invitrogen) was applied at a 1:200 dilution with DAPI nuclear counterstaining for 1 h at room temperature. The sections were mounted using Shandon Immu-Mount (Thermo Fisher Scientific). For the negative control, the primary antibody was omitted to control for autofluorescence. Slides were stored at 4°C in the dark until photographed. ImageJ was used to assess the fluorescent images to quantify the mean fluorescence intensity or area.

FGL1 (pg/ml) was measured in diluted human and mouse serum using an ELISA kit (R&D Systems) per the manufacturer’s protocol.

Systolic and diastolic blood pressures were measured daily in conscious mouse by tail-cuff plethysmography using a BP-2000 blood pressure analysis system (Visitech Systems, Apex, NC, United States) in animals that had been prewarmed in a metal chamber maintained at 30°C. Each day, values obtained from three consecutive measurements were averaged. The mean arterial pressure [MAP; (2 × diastolic + systolic)/3] was calculated as previously described (Domanski et al., 1999).

Urinary protein concentrations were measured using Folin’s phenol reagent method (Lowry et al., 1951). Urine samples and bovine serum albumin standards were assayed in triplicate and read at 650 nm with an NIR filter in a spectrophotometer (model 6/20, Coleman). Sample urinary protein levels (mg/ml) were extrapolated from a standard curve derived from bovine serum albumin standards.

Tissues or cells were homogenized and lysed with a buffer containing 0.1% SDS, 0.5% sodium deoxycholate, 1% NP-40, and a protease inhibitor mixture (PMSF, aprotinin, and sodium orthovanadate). Total protein lysate (50 μg per lane) was separated on SDS-PAGE gels (10% running and 4% stacking), transferred to PVDF membranes and analyzed using the designated primary antibodies (1:1000 dilution). After the membranes were probed with an appropriate secondary antibody (1:2000 dilution), signals were detected using an enhanced chemiluminescence reagent (Amersham Pharmacia Biotech, United Kingdom) under a LAS3000 imaging system (Fujifilm, Tokyo, Japan). The band intensity was quantified using ImageJ software.

HTR-8/SVneo cells were seeded in 96-well plates (5 × 10³ cells/well) and cultured in DMEM containing 2% FBS for 12 h. Then, recombinant FGL1 (0, 50, 100, 200 ng/ml) was added to the culture medium, and the cells were cultured for another 48 h. Cell viability was subsequently analyzed by Cell Counting Kit-8 (CCK-8) colorimetric assays (Dojindo Molecular Technologies, Inc.). After treatment, 10 μl of CCK-8 was added to each well, and the incubation was continued at 37°C for 2 h in the dark. Then, the absorbance value at 450 nm was determined using a microplate reader (SpectraMAX 340; Molecular Devices, Sunnyvale, CA, United States).

All cellular experiments were performed using triplicate samples and repeated at least three times. Statistical analyses were conducted using Prism 7.0 (GraphPad Software). Data are expressed as the mean ± SEM. The D’Agostino-Pearson normality test was used to assess the normality of the distribution. Student’s t test was used for comparisons between the two groups. One-way analysis of variance (ANOVA), followed by post hoc analysis (Tukey’s test), was used for comparisons among multiple groups. For comparisons among multiple groups and different time points, two-way ANOVA was used. p < 0.05 was considered statistically significant.

First, FGL1 gene and protein expression levels in the placentas from L-NAME-induced PE-like mouse were measured to determine the correlation between FGL1 and PE. qPCR, Western blotting, and immunofluorescence staining indicated that the FGL1 mRNA and protein expression levels in the placentas of the mouse with PE were significantly higher than those in the normal placentas (mRNA: 1.71 ± 0.07 vs. 1.0 ± 0.04-fold, p = 0.0001; Figure 1A; protein: 2.28 ± 0.32 vs. 1.01 ± 0.04-fold, p = 0.03; Figure 1B; immunofluorescence intensity: 70.94 ± 5.88 vs. 43.86 ± 4.18 a.u., p = 0.0038; Figure 1C). FGL1 is a secreted protein; ELISA revealed that the FGL1 concentration was higher in the serum of the mouse with PE than in that of the control mouse (355.8 ± 41.69 vs. 203.8 ± 33.25 pg/ml, p = 0.0172; Figure 1D). For the analysis of FGL1 levels in humans, pregnant women with PE and age-matched healthy pregnant women without PE (normotensive) were enrolled. Clinical definitions and exclusion criteria are described in the Methods section. Clinical characteristics of study subjects are shown in Table 1. As expected, patients with PE had a higher blood pressure and percent proteinuria but reduced placental weight, birth weight, and 5-min APGAR score than the healthy subjects. In our study population, the PE group had a significantly higher BMI and lower gestational age of sample collection. Trophoblast FGL1 expression in human placental sections (Figure 1E) and placental FGL1 protein levels (Figure 1F) were both significantly higher in the women with PE than in the controls. Similar results were obtained for the serum FGL1 concentration (196 ± 24.25 pg/ml, n = 38 vs. 120.4 ± 15.69 pg/ml, n = 44, p = 0.0088; Figure 1G). Taken together, the results suggest that the placenta trophoblasts in decidua should be one of the sources of FGL-1 production in both mice and humans and FGL1 upregulation correlates with PE.

Our previous study indicated that FGL1 can be expressed in human trophoblasts (Kang et al., 2020). Here, we investigated whether L-NAME-enhanced placental FGL1 expression was correlated with trophoblasts. Two human trophoblast cell lines, 3A-sub E and HTR-8/SVneo, were stimulated with L-NAME at the indicated dose. As shown in Figure 2, Western blotting and immunofluorescence staining both indicated that FGL1 expression was dose-dependently upregulated under L-NAME stimulation, suggesting that L-NAME-induced progression of PE may also promote FGL1 expression in trophoblasts. We further found that exogenous FGL1 increased trophoblast proliferation (Figure 2E).

To determine whether the highly expressed FGL1 functioned as a protective factor against the progression of PE in mouse, we administered recombinant FGL1 to the mouse with PE. The body weight results revealed that the experimental group (PE + FGL1) had a minor improvement in body weight compared with the PE group during pregnancy (Figure 3A). Furthermore, FGL1 treatment significantly improved the MAP (Figure 3B), diastolic pressure and systolic pressure in the FGL1 group compared with the non-FGL1 group (diastolic: 58.61 ± 4.61 vs. 84.6 ± 3.95 mmHg, p < 0.0001; systolic: 112.28 ± 4.91 vs. 136.65 ± 1.73 mmHg, p = 0.0002; Figure 3C). The upregulated urinary protein expression under L-NAME stimulation at mid- and late pregnancy was significantly reduced by FGL1 treatment (mid: 14.93 ± 1.11 vs. 27.53 ± 1.57 mg/ml, p < 0.0001; late: 26.89 ± 1.47 vs. 38.93 ± 3.62 mg/ml, p = 0.0002; Figure 3D; n = 6/group). The results in this section suggest that upregulated FGL1 expression has critical antihypertensive and antiproteinuric effects in the mouse with PE, which may inhibit the progression of PE.

Next, the effects of FGL1 treatment on various parameters related to the fetus and placenta were investigated. Measurement of fetal weight and crown-to-rump length revealed that FGL1 treatment significantly reversed the L-NAME-reduced fetal growth (fetal weight: 0.57 ± 0.02 vs. 0.46 ± 0.02 g, p = 0.038; Figure 3E; crown-to-rump length: 1.43 ± 0.03 vs. 1.25 ± 0.04 cm, p = 0.033; Figure 3F; n = 12/group) but did not significantly affect placental weight or diameter (placental weight: 0.08 ± 0.002 vs. 0.08 ± 0.003 g, p = 0.564; Figure 3G; diameter: 6.0 ± 0.242 vs. 6.0 ± 0.229 cm, p > 0.99; Figure 3H; n = 12/group). In addition, hematoxylin-eosin staining performed to analyze the placental structure revealed that the L-NAME-increased spongiotrophoblast area was significantly reduced by FGL1 treatment (45,549 ± 4,585 vs. 61,232 ± 4,333 pixels, p = 0.046; Figure 3I; n = 6/group). These results indicate that FGL1 contributes to improving fetal growth and placental structure, thus retarding the progression of PE. Thus, upregulated FGL1 expression should function as a compensation factor to reduce the progression of PE.

Furthermore, qRT-PCR was used to elucidate how FGL1 treatment inhibited the progression of PE in mouse. The results revealed that FGL1 treatment significantly inhibited the expression of the proinflammatory cytokines IL-1b and IL-6 under L-NAME stimulation (IL-1b: 3.60 ± 0.75- vs. 6.34 ± 0.73-fold, p = 0.015, Figure 4A; IL-6: 3.26 ± 0.62- vs. 6.10 ± 0.87-fold, p = 0.014, Figure 4B; n = 6/group). The imbalanced expression of the angiogenic factors Flt-1 and PlGF was also improved in the FGL1 group compared with the non-FGL1 group (Flt-1: 0.93 ± 0.17- vs. 2.02 ± 0.19-fold, p = 0.0003, Figure 4C; PlGF: 0.96 ± 0.11- vs. 0.37 ± 0.08-fold, p = 0.0002, Figure 4D; Flt-1/PlGF ratio: 0.95 ± 0.09- vs. 7.29 ± 1.91-fold, p = 0.003, Figure 4E; n = 6/group). The protein levels of IL-6 and PlGF also showed consistent results (Figure 4F). These findings indicate that FGL1 inhibits the progression of PE by reducing the expression of proinflammatory cytokines and rescuing the dysregulated levels of angiogenic molecules.

Finally, the intracellular mechanisms by which L-NAME moderates placental FGL expression and the cell function of FGL1 were investigated. Digital prediction by Qiagen (Hilden, Germany) indicated that the transcription factor binding site of FoxO3a is contained in the FGL1 promoter. L-NAME-induced upregulation of FGL1 was accompanied by a decrease in FoxO3a, and the FGL1 increase was abolished under FoxO3a overexpression (Figures 5A,B). Here, the results suggest that the overexpression of FoxO3a inhibits FGL1 expression in L-NAME-induced trophoblast cells.