DJ-1 upregulates the Nrf2/GPX4 signal pathway to inhibit trophoblast ferroptosis in the pathogenesis of preeclampsia

Ferroptosis is a newly discovered mode of cell death that involves disorders in iron metabolism and the accumulation of reactive oxygen species (ROS) in the plasma membrane. Preeclampsia (PE) is a gestational idiopathic disease that is characterized by hypertension and albuminuria, begins after 20 weeks of pregnancy. DJ-1 is a prerequisite for activating and stabilizing Nrf2 to allow translocation to the nucleus to carry out further functions. Detecting the expression levels of DJ-1, the Nrf2/GPX4 signaling pathway and ferroptosis markers in placental tissues of pregnant women with and without PE. Analyzing the effects of the ferroptosis inducer (RSL3) and the inhibitor (Fer-1) on the mortality rate of BeWo cells and DJ-1+/+, DJ-1−/− BeWo cells. Ferroptosis markers (MDA concentration and morphology of trophoblast cells) and DJ-1 and its downstream the Nrf2/GPX4 signaling pathway increased significantly in PE pathological state. The expression levels of DJ-1 protein in the control group and the PE group were positively correlated with the expression levels of Nrf2/GPX4 signaling pathway protein, and negatively correlated with the MDA concentration. BeWo cells were sensitive to the ferroptosis inducer (RSL3) and the inhibitor (Fer-1). The high expression levels of DJ-1 in BeWo cells can resist ferroptosis by regulating the Nrf2/GPX4 signaling pathway. Ferroptosis is involved in the pathogenesis of PE. DJ-1 can mediate the trophoblast cells ferroptosis and play a protective role in the pathogenesis of preeclampsia by regulating the Nrf2/GPX4 signaling pathway.

Observation of cell morphology. Placental   www.nature.com/scientificreports/ LTD)-100% acetone (Pharmaceutical Group Chemical Reagent Co., LTD) for dehydration, each time for 15 min. Then, the tissue was permeated for 2-4 h in acetone: 812 embedding agent (SPI, 90529-77-4) = 1:1, overnight in acetone: 812 embedding agent = 1:2, and in pure 812 embedding agent for 5-8 h. The pure 812 embedding agent was poured into the embedding plate, and the tissue samples were inserted into the embedding plate and then baked at 37 °C overnight. Put tissue in the oven at 60 °C and polymerize for 48 h and section at 70 nm on an ultra-thin slicer (Leica UC7, Leica, German). Sections were then double stained with 3% uranium acetatelead citrate. Next, the morphology of the syncytiotrophoblast cells were observed under a transmission electron microscope (HT7700, HITACHI, Japan) and determined by ImageJ software.

Quantitative real-time PCR (qRT-PCR).
For each sample, 50 mg of tissue were ground in a mortar and then mixed vigorously with 1 ml of Trizol (Thermo Fisher Scientific Co., Ltd.). Samples were then centrifuged at 12000 rpm for 5 min and the supernatant transferred into a new 1.5 ml microcentrifuge tube and mixed with 200 μl of chloroform. Samples were then centrifuged at 12,000 rpm for 15 min at 4 °C and the supernatant was transferred into a new 1.5 ml microcentrifuge. Samples were then mixed with the same volume of isopropanol and inverted several times to mix. Samples were then placed at − 20 °C for 30 min. Then, samples were centrifuged at 12000 rpm for 10 min at 4 °C; the supernatant was then discarded and 1 ml of 75% ethanol (anhydrous ethanol: dd H 2 O = 3:1) was added. The microcentrifuge tube was then oscillated to encourage precipitation and was then centrifuged at 8000 rpm at 4 °C. The supernatant was discarded and the pellet (total RNA) was dried at room temperature for 5-10 min. Then, the total RNA was resuspended in 30 µl of dd H 2 O water. The total RNA concentration and OD260/280 ratio were then determined with an ultra-micro spectrophotometer (Thermo Scientific, American). We used a TB Green0R Premix Ex TaqTM (Tli RNaseH Plus) and PrimeScriptTM RT Reagent Kit (Perfect Real Time) that was purchased from TaKaRa Company (Japan). Using total RNA as a template, we prepared the reverse transcription reaction solution and qRT-PCR reaction solution, in accordance with the manufacturer's instructions. Primer details are given in Table 2. qRT-PCR reaction system is given in Table 3. We used the levels of β-actin mRNA as a control; the relative expression levels of mRNA in placental tissues and cells were calculated by the 2 −∆∆CT method. The primers used in this experiment were synthesized by Fuzhou SunYa Biology Co., Ltd.
Western blotting. For western blotting, 50-100 mg of sample was mixed with 500 μl and 1000 μl of RIPA lysis buffer (Meilun Biotechnology Co., Ltd., Suzhou, China) and PMSF (1 mM) (Meilun Biotechnology Co., Ltd., Suzhou, China). Then, samples were centrifuged at 12000 rpm and 4 °C for 10 min. The protein concentration of each protein lysate (the supernatant) was then determined with a BCA Protein Assay Kit (Meilun Biotechnology Co., Ltd., Suzhou, China). Then, we took the same protein content from each sample (50-100 μg) and added the same volume of 5 × protein buffer (Meilun Biotechnology Co., Ltd., Suzhou, China). Then, protein samples were boiled for 10 min in boiling water.
SDS-PAGE (Beyotime Biotechnology Co., Ltd., Shanghai, China) gels were prepared in accordance with the manufacturer's guidelines and samples loaded (V = 50 μg/C/ 4 5 ) and separated. Separated proteins were then transferred to PVDF membranes and cut the membranes into suitable size according to protein ladder and the molecular weight of target protein. Incubated with antibodies raised against β-actin (Immunoway, 1:1000, 4 °C),  F ATC ATG TTT GAG ACC TTC AACA  R CAT CTC TTG CTC GAA GTC CA   DJ-1  F GTA GCC GTG ATG TGG TCA TTT  R CTG TGC GCC CAG ATT ACC T   Nrf2  F AAG TGG AAG AGC TAG ATA GTGCC  R CCT GAG ATG GTG ACA AGG GTT   GPX4  F CAG GAG CCAGGFAGT AAC GA  R GCC CTT GGT GGA TCT TCA   Table 3. qRT-PCR reaction system. Immunohistochemistry. Placenta tissue was sectioned into 3 mm blocks and put the bloks into xylene I (Shanghai Lingfeng Chemical Reagent Co., LTD, 1330-20-7) for 15 min-xylene II for 15 min-xylene III for 15 min-absolute ethanol I (China National Pharmaceutical Group Chemical Reagent Co., LTD, 100092683) for 5 min-absolute ethanol II for 5 min-85% alcohol for 5 min-75% alcohol for 5 min-rinse in distilled water. The blocks are placed in a repair box filled with citric acid (PH6.0) antigen retrieval buffer (Servicebio, G1202) for antigen retrieval in a microwave oven, heated on medium power for 8 min until boiling, then turned off the microwave oven, kept warm for 8 min and then transferred to medium-low power for heating 7 min. During this process, excessive evaporation of buffer should be prevented and the sections should not be allowed to dry. To cool to room temperature before proceeding, the sections are placed in PBS (PH7.4) (Servicebio, G0002) and shaken on the decolorization shaker 3 times for 5 min each. The blocks are placed in 3% hydrogen peroxide (Pharmaceutical Group Chemical Reagent Co., LTD, 10011208) and incubated at room temperature in dark- Complete medium was prepared by mixing 85% Ham's F-12K basic medium (Procell Life Science & Technology Co., Ltd.) with 13% fetal bovine serum standard (NEWZERUM) and 2% penicillin/streptomycin solution (Meilun Biotechnology Co., Ltd.). Ferroptosis inducer RSL3 stock solution was prepared by dissolving 5 mg of RSL3 powder (APExBio) in 11.3410 ml of DMSO (Suzhou Meilun Biotechnology Co., Ltd.); this prepared 1 mM of stock solution, which was then stored at -20 °C. Next, we prepared a working solution for the ferroptosis inducer RSL3 by diluting the RSL3 stock solution to concentrations of 0, 2, 4, 6, 8, 10 μM with complete medium. The working solution for the ferroptosis inhibitor Fer-1 was prepared by diluting Fer-1 stock solution (APExBio) to a concentration of 2.5 μM with complete medium. Puromycin solution (10 mg/ml) (Dalian Meilun Biotechnology Co., Ltd.) was diluted to a concentration of 0.5 μg/ml with complete medium.  Table 4. The absorbance value of 450 nm was measured in microwell plate spectrophotometer (BioTek Epoch). The standard curve is made according to the measured value and concentration of the standard tube, y (μmol/mL) is the standard concentration and x is the absorbance. LDH (U/10 4 cell) = 0.133y.

Quantitative Real-time PCR (qRT-PCR).
The specific operation was the same as 2.1.1. Primer details are given in Table 5.
Statistical analysis. Data were statistically analyzed by the SPSS version 19.0 software package. The measurement data were expressed as mean ± standard deviation (SD). Differences among groups were compared by single factor analysis of variance (ANOVA), the type of post hoc test used LSD and comparison between two groups was conducted by the independent sample t-test. α was 0.05 and P < 0.05 was statistically significant.
Ethical statement. The Ethics Committee of Fujian Maternity and Child Health Hospital approved all protocols (Reference: 2021KLRD631).

Baseline characteristics of the women included in this analysis.
There was no significant difference in maternal age (P = 0.252), gestational (G)(P = 0.073), parity (P)(P = 0.281) and diastolic pressure (DP) (P = 0.248) when compared between groups. The gestational age in the PE group was significantly lower than in the control group (P < 0.001). The prenatal BMI in the PE group was significantly higher than in the control group (P = 0.016). Systolic pressure (SP) in the PE group was significantly higher than in the control group (P = 0.025). (Table 6).  (Fig. 1).

Ferroptosis markers in preeclamptic placentae.
Morphology of placental trophoblasts. Figure 2A shows that in the control group, the syncytiotrophoblast cells in placental tissue showed slight evidence of pyknosis, the basement membrane (BM) was continuous and intact, thickness was uniform, the tight junctions between cells were reduced, and the local intercellular space was slightly widened. Cell membranes were intact with abundant microvilli (Mv) on their surface. Cells also exhibited a high electron density, abundant organelles, partial swelling, irregular shaped nuclei (N), local depression, a complete nuclear membrane, local widening of the perinuclear space, a large amount of heterochromatin, and  There was a high number of mitochondria (M) with an acceptable and intact structure, and a parallel crest. The rough endoplasmic reticulum (RER) was obviously dilated and degranulated; there was evidence of more floc accumulation and endoplasmic reticulum retention. The Golgi apparatus (Go) was slightly hypertrophic and the capsule was dilated; there was no typical autophagy structure in the visual field (magnification × 2500). As can be seen from Fig. 2B, the syncytiotrophoblast cells in placental tissue from the PE group showed obvious pyknosis. The tight junctions between cells were reduced, and the local intercellular space had widened slightly. The cell membranes were intact with abundant microvilli (Mv) on their surface. Intracellular electron density was high and there were many swollen organelles. Nuclei (N) showed irregular pyknosis; the nuclear membrane was intact, the perinuclear space was normal, and there was a greater amount of heterochromatin than the control group. There was an abundance of mitochondria (M); these were mostly swollen. The mitochondrial matrix was lighter in the membrane, and the crest had reduced and disappeared. The rough endoplasmic reticulum (RER) was obviously dilated, degranulated, and vacuolated. The membrane of the RER was damaged and disintegrated in severe cases; there was no typical autophagy evident in the visual field (magnification × 2500).
Images of syncytiotrophoblast cells in the two groups showed varying degrees of injury. The cells in the control group showed mild pyknosis, slight edema, a high intracellular electron density, abundant organelles, and partial swelling. Cells in the PE group showed relatively severe ferroptosis injuries, obvious pyknosis and cell shrinkage, high intracellular electron density, abundant organelles, and obvious swelling.  (Fig. 3A). The expression levels of DJ-1 protein in the PE group were significantly higher than those in the control group [control vs. PE = (1.14 ± 0.96) vs. (3.13 ± 1.83), P = 0.005] (Fig. 3B,C).
The proportion (%) of DJ-1 positive cells in the PE group was significantly higher than that in the control group [control vs. PE = (23.11 ± 23.28) vs. (39.80 ± 32.81), P < 0.001] (magnification × 200) (Fig. 3D,E). Table 6. Baseline characteristics of the women included in this analysis. a Compared with the control group P < 0.05. Comparison between two groups was conducted by the independent sample t-test.  (Fig. 3F). The expression levels of Nrf2 protein in the PE group were significantly higher than the control group[control vs. PE = (0.78 ± 0.57) vs. (3.98 ± 3.09), P < 0.001] (Fig. 3G,H).
Correlations of DJ-1 protein expression levels with the Nrf2/GPX4 signaling pathway protein expression levels and MDA concentrations in placental tissue. The expression levels of DJ-1 protein in the control group (Fig. 4A,B) and the PE group (Fig. 4C,D) were positively correlated with the expression levels of the Nrf2/GPX4 signaling pathway protein (P < 0.001), and negatively correlated with the concentrations of MDA (P < 0.001) (Fig. 4E,F) (Table7).  (Table 8 and Fig. 5).    Figure 7 show transfection efficiency data.

Discussion
In this study, we found that ferroptosis is involved in the pathogenesis of PE and that DJ-1 plays a protective role in the pathogenesis of PE by regulating the Nrf2/GPX4 signaling pathway to mediate ferroptosis in trophoblast cells.
According to the "six-stage model" theory for PE 15 , a series of PE clinical symptoms (such as hypertension and albuminuria) begin to appear after four stages: poor maternal immune tolerance during embryo implantation; abnormal trophoblast invasion in the critical period of placental formation; maternal-fetal interface stress response, and excessive concentrations of placental-derived damage factors that invade the maternal blood circulation. Ferroptosis is a new form of regulatory cell death that was first reported by Dixon in 2012 5 . The abnormal of iron mediates the excessive accumulation of lipid peroxidation, which involves a variety of physiological and  www.nature.com/scientificreports/ pathological processes, including cancer cell death, neurotoxicity, neurodegenerative diseases, acute renal failure, drug-induced hepatotoxicity, liver and heart ischemia/reperfusion injury and T cell immunity 4 . Placental oxidative stress (OS) and lipotoxicity have been confirmed to be characteristic manifestations of placental dysfunction; there is also evidence to suggest that trophoblast cell lines are sensitive to ferroptosis 16 . Barneo-Caragol et al. 17 reported that the ratio of lipid peroxidation to total antioxidant activity (TAA) in the placentas of pregnant women with early-onset PE was significantly higher than that in women with late-onset PE and healthy pregnant women. In addition, metabolic analyses of placental mitochondria showed that the levels of polyunsaturated fatty acid (PUFAs), and the degree of mitochondrial shrinkage, in pregnant women with severe PE were significantly higher than those in a control group with normal blood pressure 18 , thus confirming that ferroptosis plays a key role in the pathogenesis of PE. Zhang et al. 19 demonstrated that the concentration of MDA, and the level of total Fe 2+ , in the placenta of PE rats were also significantly increased coincident with a significant reduction in the levels of GSH and the activity of GPX. Furthermore, supplementation with a ferroptosis inhibitor was shown to alleviate the symptoms of PE in rats. There was also a significant increase in the concentration of MDA in the mitochondrial membrane, the extent of mitochondrial permeability transition pore (MPTP) opening, and the expression levels of mtDNA in the placentas of pregnant women with PE, thus confirmed the existence of ferroptosis; 20 these findings were consistent with our present experimental results.
Mitochondria are representative regulatory cell death (RCD) executive organelles, which function by releasing or recruiting specific cell death-promoting factors, including ferroptosis 21 . Different from the mitochondrial swelling of traditional apoptosis, necrosis and autophagy, electron microscopy showed that the mitochondrial membrane of ferroptosis cells was dense, the cristae decreased or even disappeared, but most of the nuclei were normal 22 . In this study, it was found that the mitochondria of PE trophoblast showed not only the characteristics of ferroptosis, including the weakening of intramembrane matrix, the decrease and disappearance of crest, but also the morphological manifestation opposite to that of typical ferroptosis, that is, volume swelling. Because PE is a complex pregnancy-related disease caused by many factors, many studies have shown that apoptosis and necrosis can also mediate the pathogenesis of PE 23 , so we reasonably speculate that the complex morphological changes of mitochondria are the result of different ways of cell death. This study also found that there were some depressions and heterochromatin damage in trophoblast nuclei in both PE group and control group. This seems to be inconsistent with the previous literature that the nuclear membrane of ferroptosis cells is normal 24 . However, it should be noted that even in normal pregnancy, placental tissue will experience varying degrees of OS damage caused by hypoxia/reperfusion 25 , which well explains the changes of nuclear membrane in this study.
Inhibitors of ferroptosis, apoptosis, autophagy, and necrosis, can alleviate the cell deaths of trophoblast under hypoxic conditions 19 . Kajiwara et al. 26 have found that human trophoblast cells are sensitive to the ferroptosis induced by the depletion or inhibition of GPX4 or lipase PLA2G6. This process is characterized by giant bubbles and vesicles in the plasma membrane; these observations are characteristic markers of ferroptosis. Beharier et al. 16 found that the mortality rate of the trophoblast cell line PHT and BeWo cells increased gradually under the action of different concentrations of the ferroptosis inducer RSL3. Furthermore, injury could only be alleviated by the administration of a ferroptosis inhibitor (Fer-1) rather than an inhibitor of apoptosis and necrosis. These results are consistent with those derived from present research and suggests that trophoblast ferroptosis occurs during the first four stages of PE.
The up-regulation of DJ-1 plays an important role in cell survival during the process of chronic OS. On the one hand, DJ-1 stimulates and maintains the normal production of ATP by directly increasing molecular chaperone activity and inhibiting p53 translocation to the mitochondria. On the other hand, DJ-1 mediates the downstream signaling pathway such as Nrf2/GPX4 to stimulate the production of superoxide dismutase (SOD) and GSH. The last three amino acids at the C-terminal of DJ-1 (Δ C3) play a key role in maintaining its dimer structure. The hydrophobic L187 residue can also alleviate the ferroptosis effect of erastin on DJ-1−/− mouse embryonic fibroblasts 27 . In a previous study, we found that the expression levels of DJ-1 mRNA and protein in the serum and placenta of patients with PE were significantly increased 28 . We also found that the up-regulated levels of Nrf2/GPX4 in the placenta of patients with severe PE enhanced the invasive ability of trophoblasts 29 . DJ-1 and the Nrf2/GPX4 signaling pathway plays a protective role in the process of ferroptosis. De Miguel C et al. 30 found that oxidative stress hypertension occurred in DJ-1−/− mice when the expression of DJ-1 was inhibited. This was accompanied by an increased renal MDA concentration and increased expression of the oxidative stress marker heat shock protein 60 mRNA. Yao et al. 31 constructed a rat model of sepsis-associated encephalopathy (SAE) and found that the levels of MDA and Fe 2+ in the hippocampus of SAE rats were significantly higher than controls, while the expression levels of Nrf2 and GPX4 proteins were significantly decreased. The activation of the Nrf2/ GPX4 signaling pathway by deferoxamine preconditioning could significantly reduce the levels of ferroptosis in hippocampal neurons and improve cognitive impairment. These previous research results are consistent with the conclusions of our current experiments. In the present study, we found that trophoblasts are sensitive to ferroptosis. We also confirmed that high expression levels of DJ-1 can inhibit ferroptosis in trophoblast cells via the Nrf2/GPX4 signaling pathway; we draw a conclusion that DJ-1 plays a protective role in the pathogenesis of PE.
Advantages and disadvantages of the current research. There are several advantages to the present research. First, we verified that ferroptosis was observed in the placentas of patients with PE, that trophoblast cell lines were highly sensitive to ferroptosis, and that DJ-1 played a protective role in the pathogenesis of PE by regulating the expression levels of the downstream Nrf2/GPX4 signaling pathway. Secondly, we investigated and demonstrated the process of ferroptosis in placenta from patients with PE and identified the roles of ferroptosis, DJ-1, and the Nrf2/GPX4 signaling pathway, in trophoblast cells. The major innovative aspect of our research was identifying the relative roles of DJ-1, the Nrf2/GPX4 signaling pathway, and ferroptosis, in the pathogenesis of PE at the level of the placenta.

Scientific Reports
| (2022) 12:2934 | https://doi.org/10.1038/s41598-022-07065-y www.nature.com/scientificreports/ Similarly, there are many areas that need to be improved in our experiment. For example, in order to better illustrate the role of trophoblast ferroptosis in the pathogenesis of PE, it is necessary to measure the levels of Fe 2+ and ferritin in placental tissue. In addition, in order to better demonstrate that DJ-1 can mediate ferroptosis in trophoblast can play a proactive role in the pathogenesis of PE through Nrf2/GPX4 signal pathway, future research should involve the comprehensive analysis of animal models of PE.

Conclusion
In the present research, we identified that ferroptosis is involved in the pathogenesis of PE and that DJ-1 can mediate ferroptosis in trophoblast cells and play a protective role in the pathogenesis of PE by regulating the Nrf2/GPX4 signaling pathway.

Data availability
The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.