Metformin Inhibits Trophoblast Ferroptosis by Suppressing ATF2 and Activating the PI3K/Akt/Nrf2 Pathway to Improve Gestational Diabetes Mellitus

ABSTRACT

Gestational diabetes mellitus (GDM) is a common metabolic disorder during pregnancy. Metformin (MET) has emerged as a promising alternative to insulin, but its mechanism of action in the placenta remains incompletely understood. In this study, we aimed to examine whether MET counteracts high‐glucose‐induced ferroptosis by regulating activating transcription factor 2 (ATF2) and to delineate the underlying molecular mechanism in GDM. We established in vitro GDM cell models using HTR‐8/SVneo trophoblast cells and an in vivo STZ‐induced hyperglycemic rat model, followed by treatments including ATF2 lentiviral manipulation and MET administration. We evaluated cell viability, ferroptosis markers (ROS, MDA, GSH, iron deposition), and PI3K/Akt signaling pathway activity. Our results demonstrated that in GDM patients and high‐glucose‐challenged HTR‐8/SVneo cells, MET promoted cell viability by suppressing ATF2, reactivating PI3K/Akt signaling, enhancing nuclear Nrf2 translocation, and upregulating GPX4, while concurrently lowering ROS and MDA levels, elevating GSH, and reducing iron deposition. These protective effects were reversed by ATF2 overexpression or pathway inhibitors. In STZ‐induced hyperglycemic rats, MET decreased placental ATF2 expression, restored the PI3K/Akt‐Nrf2‐GPX4 axis, alleviated ferroptosis, and reduced fetal weight. Collectively, our findings indicate that MET suppresses ATF2, activates PI3K/Akt, drives Nrf2 nuclear import, upregulates GPX4, and curbs high‐glucose‐induced trophoblast ferroptosis, offering new mechanistic insight and therapeutic avenues for GDM.

Metformin suppresses ATF2 expression, activating the PI3K/Akt pathway and promoting Nrf2 nuclear translocation. This upregulates GPX4, inhibiting ferroptosis in trophoblast cells. Erastin blocks system Xc⁻, depleting GSH and promoting ferroptosis. LY294002 and ML385 inhibit the pathway, validating the mechanism. The figure summarizes the protective role of metformin against trophoblast ferroptosis in GDM.

1. Introduction

Gestational diabetes mellitus (GDM) refers to abnormal blood glucose levels first detected or occurring during pregnancy [1]. As the most common endocrine disorder among pregnant women, its prevalence is increasing year by year. The International Diabetes Federation (IDF) epidemiological data shows that 16.2% of pregnant women have varying degrees of elevated blood glucose, among which 84% of pregnant women with high blood glucose have GDM. GDM significantly increases adverse outcomes for mothers during the perinatal period and after childbirth, including miscarriage, fetal malformations, preeclampsia, fetal growth restriction, fetal death, macrosomia, neonatal hypoglycemia, neonatal hyperbilirubinemia, and neonatal respiratory distress syndrome. It may also increase the risk of metabolic diseases such as obesity, hypertension, and Type 2 diabetes in future generations [2]. The placenta is an important organ in the process of maternal‐fetal exchange, responsible for nutrition, metabolism, exchange, and endocrine functions throughout the pregnancy. Placental trophoblast cells regulate maternal blood glucose stability by secreting various steroid hormones and cytokines, and placental dysfunction caused by trophoblast cell damage is an important pathogenesis of GDM [3].

Ferroptosis is a recently identified form of programmed cell death characterized by iron‐dependent lipid peroxidation, with distinct morphological, metabolic, and genetic features. Current evidence indicates that this death pathway plays a pivotal role in various pathological conditions, including malignancies, neurodegenerative disorders, ischemia–reperfusion injury, atherosclerosis, glucose metabolism disturbances, and acute or chronic kidney damage [4]. Trophoblast cells in the placenta are particularly susceptible to ferroptosis. Sustained hyperglycemia can trigger oxidative stress cascades, leading to excessive reactive oxygen species (ROS) accumulation, which impairs cell proliferation and invasion and increases the risk of GDM [5].

Glutathione peroxidase‐4 (GPX4) is a well‐established negative regulator of ferroptosis, blocking death signals by reducing lipid peroxides; loss of GPX4 activity results in ROS and lipid peroxide buildup within trophoblasts, ultimately inducing ferroptosis [6]. Nuclear factor‐E2‐related factor 2 (Nrf2) serves as a master transcription factor for antioxidant responses, capable of upregulating numerous antioxidant and iron‐homeostasis genes [7]. Under resting conditions, Nrf2 forms a cytoplasmic complex with Keap1 and is continuously degraded via the ubiquitin‐proteasome pathway. Upon oxidative challenge, conformational changes in Keap1 release Nrf2, which then translocates to the nucleus and initiates transcription of downstream antioxidant genes such as GPX4, HO‐1, and NQO1, thereby enhancing cellular resistance to lipid peroxidation and ferroptosis [8].

Currently, insulin remains the first‐line therapy for GDM, yet its high cost, subcutaneous administration requirement, and risk of hypoglycemia necessitate stringent glucose monitoring. Oral antidiabetic agents have limited data in pregnancy and among them, Metformin (MET) has gained attention due to accumulating evidence of its safety. Recent studies suggest that MET may exert metabolic protection by interfering with the ferroptosis pathway, and this mechanism is becoming a focal point in metabolic disease research.

Our previous study revealed the high expression of ATF2 in GDM and its role in promoting trophoblast ferroptosis, as well as the molecular mechanism by which it acts through the PI3K/Akt/Nrf2 pathway [9]. However, it remains to be clarified whether MET regulates ATF2 expression in GDM trophoblasts, whether this ATF2‐mediated regulation modulates ferroptosis through the PI3K/Akt/Nrf2 pathway, and whether this mechanism operates in vivo to improve pregnancy outcomes. In this study, we aimed to investigate these specific questions by examining the effects of MET on ATF2 expression and ferroptosis in trophoblast cells under high glucose conditions, with a particular focus on the regulatory role of MET on the PI3K/Akt/Nrf2 signaling pathway. By establishing in vitro trophoblast cell models and in vivo GDM rat models, we seek to gain a deeper understanding of the mechanism underlying placental dysfunction caused by gestational diabetes and explore the therapeutic potential of MET in improving pregnancy outcomes through regulating the ATF2‐PI3K/Akt/Nrf2‐GPX4 axis.

2. Materials and Methods

2.1. Patients and Clinical Samples

Before collecting placental tissues, informed consent was obtained from all participants to ensure compliance with ethical standards. Placental samples were obtained immediately after delivery. For histopathological and immunohistochemical analyses, tissues were fixed in 4% paraformaldehyde and embedded in paraffin. For protein assay, placental villous tissues were carefully dissected from the maternal side, washed with cold PBS, snap‐frozen in liquid nitrogen, and stored at −80°C until use. GDM was diagnosed by 75‐g OGTT at 24–28 weeks according to the International Association of Diabetes and Pregnancy Study Groups (IADPSG) criteria (fasting ≥ 5.1, 1‐h ≥ 10.0, or 2‐h ≥ 8.5 mmol/L). All patients received individualized diet and exercise guidance; those with fasting glucose > 5.3 mmol/L or 2‐h postprandial glucose > 6.7 mmol/L after 1–2 weeks of lifestyle intervention were considered poorly controlled and offered pharmacotherapy. Patients were assigned to Metformin (starting 0.5 g once daily, titrated up to 2.0 g/d), insulin (aspart and detemir, adjusted per glucose monitoring), or the group with poor blood glucose control without medication according to their informed choice (concerns about drug safety, preference for simple lifestyle adjustments, economic factors, or other personal inclinations). The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of Nanjing Medical University (IRB No. 2024[098], date of approval: February 28, 2024). Informed consent was obtained from all subjects involved in the study.

2.2. Cell Lines and Reagents

The human trophoblast cell line HTR8/SVneo was obtained from Pronosai (Wuhan, China) and maintained in RPMI 1640 medium (Gibco, Thermo Fisher Scientific, Waltham, Massachusetts, USA) supplemented with 10% fetal bovine serum (Gibco, USA). Cells were incubated at 37°C in a humidified atmosphere containing 5% CO₂. We used the following stimulus concentrations to treat cells: high glucose (Glu, 25 mM), MET (0.5 mM), Ferrostatin‐1 (Fer‐1, 2 μM), Erastin (Era, 5 μM), LY294002 (LY, 10 μM), and ML385 (ML, 5 μM). The standard RPMI 1640 medium (containing 11 mM glucose) served as the baseline culture control.

2.3. Lentiviral Transfection

To establish stable ATF2‐overexpressing cells, the human ATF2 gene (NM_001880) obtained from Genechem (Shanghai, China) was cloned into the GV492 lentiviral vector (Ubi‐MCS‐3FLAG‐CBh‐gcGFP‐IRES‐puromycin) using AgeI and NheI restriction enzymes and the In‐fusion recombination method, followed by sequence verification. Recombinant lentiviruses were produced by co‐transfecting 293T cells with the lentiviral vector and helper plasmids (psPAX2 and pMD2.G) using Lipofectamine 2000. After 72 h, viral supernatants were collected, centrifuged, filtered, and stored at −80°C. Cells were infected with lentivirus at an MOI of 20 in serum‐free medium for 16 h, followed by medium replacement.

2.4. Immunohistochemistry (IHC)

For the IHC analysis, tissue sections were deparaffinized in xylene and rehydrated through a series of decreasing ethanol concentrations. Antigen retrieval was performed using a 0.01 mol/L sodium citrate buffer (pH 6.0). The sections were treated with 3% hydrogen peroxide to block endogenous peroxidase activity, and then were incubated with the primary antibodies against ATF2 (11908‐1‐AP, 1:200, Proteintech, Chicago, Illinois, USA) and GPX4 (125066, 1:1000, Abcam, UK). Staining intensity was scored as 0 (no staining), 1 (mild), 2 (moderate), or 3 (intense), while the extent of staining was scored based on the percentage of stained cells: 0 (no staining), 1 (1%–25%), 2 (26%–50%), 3 (51%–75%), and 4 (76%–100%). The final staining score was calculated by product the intensity and extent scores, with 0–6 indicating low expression and ≥ 7 indicating high expression.

2.5. Immunofluorescence Staining (IF)

Cells were fixed with 4% paraformaldehyde and permeabilized with 0.2% Triton X‐100. After blocking with 5% serum in PBS‐Tween, primary antibodies against Nrf2 (16396‐1‐AP, 1:50, Pro teintech) were applied overnight at 4°C. Cells were then washed and incubated with secondary antibodies (Alexa Fluor 488, Abcam, Cambridge, UK) for 1 h at room temperature in the dark. Nuclei were stained with DAPI. Fluorescence was imaged using a confocal microscope and analyzed with image analysis software to quantify protein expression and localization. Nrf2 fluorescence intensity was measured separately in the nuclear and cytoplasmic regions of interests (ROIs), and the nuclear/cytoplasmic intensity ratio was calculated on a per‐cell basis. At least 50 cells per group from three independent experiments were analyzed.

2.6. Western Blot Analysis

Tissue or cell lysates were prepared in RIPA buffer with protease and phosphatase inhibitors. Protein concentration was measured using the bicinchoninic acid (BCA) kit (P0011, Beyotim, Shanghai, China). Equal amounts of protein were separated by SDS‐PAGE and transferred to PVDF membranes. Membranes were blocked with 5% nonfat milk in TBST for 2 h at room temperature, then incubated with primary antibodies overnight at 4°C. After washing, HRP‐conjugated secondary antibodies were applied for 1 h. Bands were visualized using an ECL system (Amersham Biosciences) and imaged. Protein expression was quantified by densitometry with ImageJ, normalized to β‐actin.

2.7. Cell Viability Assay

Cell viability was first assessed using the Cell Counting Kit‐8 (CCK‐8) assay (C0037, Beyotim, Shanghai, China). The treated cells were seeded at a density of 2000 cells per well in a 96‐well plate. 10 μL of CCK‐8 solution was added to each well and incubated at 37°C for 2 h at different testing time points. The absorbance was measured at 450 nm using an enzyme reader. The colony formation assay is used to evaluate the long‐term proliferation ability of cells. The treated cells are inoculated into 6‐well plates, with 500 cells per well. The cells are then cultured for 10–14 days, with the medium replaced every 3 days during this period. After the culture is completed, the colonies are fixed with 4% paraformaldehyde for 15 min and stained with 0.1% crystal violet for 30 min, and then counted and photographed.

2.8. Prussian Blue Staining

Following treatment, cells were gently washed three times with PBS, then fixed with 4% paraformaldehyde for 10–20 min. After fixation, cells were incubated with Prussian blue staining solution (BL1219A, Biosharp, Anhui, China) for 30 min. Subsequently, cells were rinsed twice with distilled water, each rinse lasting 2 min. Cells were then dehydrated, cleared, and mounted using neutral balsam mounting medium (BL704A). Finally, stained cells were observed and photographed under a light microscope.

2.9. Reactive Oxygen Species (ROS) Assays

ROS levels were assessed using a ROS Detection Kit (S0033, Beyotime, Shanghai, China) by flow cytometry. After the tissue is minced, the digestion and shaking suspension is filtered. The filtrate is collected and centrifuged to discard the supernatant, thereby obtaining a single‐cell suspension. Single‐cell suspension and cells from each treatment group were incubated with 10 μM DCFH‐DA in serum‐free medium at 37°C for 20 min, then washed three times to remove excess DCFH‐DA. DCFH‐DA is metabolized to DCFH by cellular esterases and oxidized to fluorescent DCF by ROS. After trypsinization, cells were suspended in serum‐free medium for analysis.

2.10. Malondialdehyde (MDA) Assay

MDA content was measured using a MDA Assay Kit (ADS‐W‐YH002, AIDISHENG, Jiangsu, China). After the tissues or cells are lysed with lysis buffer, the supernatant obtained by centrifugation is used for analysis. The microplate reader was preheated for 30 min while a water bath was heated to 90°C–95°C. The samples were incubated in the water bath at 90°C–95°C for 30 min, then cooled on ice. After a second centrifugation at 12 000 rpm for 10 min at 25°C, 200 μL of the supernatant was transferred to a 96‐well plate. The absorbance was measured at 532 and 600 nm, and ΔA was calculated as A ₅₃₂–A ₆₀₀.

2.11. Glutathione (GSH) Assay

To measure GSH levels, cells were collected, centrifuged, and lysed with 5% 5‐sulfosalicylic acid solution using ultrasound (300 W, 3 s on, 7 s off, 3 min total). After centrifugation at 4°C, 12 000 rpm for 15 min, the supernatant was stored on ice. The tissue was ground into powder, then an appropriate amount of protein was added to remove the reagents, and the mixture was homogenized. The supernatant was then collected. GSH concentration was then assessed using a Total Glutathione Assay Kit (ADS‐W‐G001, AIDISHENG, China) according to the manufacturer's instructions.

2.12. Animal Experiments

Animal studies were approved by the Ethics Committee of the hospital (IRB No. [2024] 098). 6–8 weeks old SPF SD rats (72 females, 36 males) were acclimated for 1 week and then mated at a 1:2 male‐to‐female ratio. The morning after mating, the presence of a vaginal plug was designated embryonic Day 0 (E0). Thirty‐six plug‐positive dams were randomly assigned to six groups (n = 6 each): ① Control; ② STZ Rats (STZ); ③ STZ Rats + ATF2‐overexpression lentivirus (STZ + LV); ④ STZ Rats + empty vector lentivirus (STZ + NC‐LV); ⑤ STZ Rats + Metformin (STZ + MET); ⑥ STZ Rats + ATF2‐overexpression lentivirus + Metformin (STZ + LV + MET). From E9 to E12, groups ②–⑥ received intraperitoneal streptozotocin (50 mg/kg per day) to represent the in vivo model of GDM. On E12, groups ③, ④, and ⑥ were injected intravenously with 1 × 10⁸ TU of the respective lentivirus by injecting through the tail vein (empty vector GV248 for ④; overexpression ‐ATF2 vector for ③ and ⑥). Groups ⑤ and ⑥ were gavaged with Metformin (300 mg/kg per day) from E12 to E18. Fasting blood glucose was measured on E12, and an oral glucose tolerance test (OGTT; 1 g/kg glucose, blood collected at 0, 60, 120 min) was performed on E14 to calculate the area under the curve (AUC). On E18, dams were euthanized; litter size and fetal weight were recorded, and placentae were harvested. Placental villous tissues were either fixed in 4% paraformaldehyde or snap‐frozen at −80°C for subsequent analyses.

2.13. Statistical Analysis

GraphPad Prism 9.5, SPSS 26.0, and ImageJ were used for statistical analysis. All cell experiments were repeated three times. Data were presented as mean ± standard deviation (SD) to indicate the variability among samples. Comparisons between groups were conducted using one‐way or two‐way analysis of variance and unpaired t‐tests. All statistical tests with p values < 0.05 were considered statistically significant. GraphPad Prism 9.5 was also used to generate graphical representations to ensure clear and standardized presentation of all data in the charts.

3. Results

3.1. ATF2 Was Decreased in the GDM Patients Treated With Metformin

Our previous Immunohistochemistry analysis of 45 cases of GDM placental tissues showed that ATF2 was highly expressed in GDM. When we grouped the GDM cases according to different treatment regimens, we found that in the GDM with Metformin treatment group, ATF2 was mostly expressed at a low level (Figure 1A,B). Clinical data analysis further showed that the Metformin‐treated group had lower neonatal birth weight and glycated hemoglobin (HbA1c) levels compared with insulin‐treated or poorly controlled untreated groups. Notably, serum ferritin and transferrin levels, indicators of iron metabolism, were lowest in the Metformin group and highest in the untreated group (Table 1). Western blot analysis also showed that in the GDM with MET treatment group, the expression of ATF2 was lower than that in the insulin treatment group and the group with poor blood glucose control without medication (Figure 1C). In contrast, the expression of GPX4 was the opposite. These findings suggest a statistical association between metformin treatment and reduced ATF2 expression, improved iron homeostasis, and elevated GPX4 levels in this patient cohort.

FIGURE 1.

ATF2 was decreased in the GDM patients treated with Metformin. (A) Immunohistochemical analysis of ATF2 expression in the placentas of patients with different treatment regimens for GDM (10×, Scale bar: 200 μm; 40×, Scale bar: 50 μm). (B) Comparison of immunohistochemical results (p < 0.001). (C) Western blot analysis of ATF2 and GPX4 in the placental villi of the normal blood glucose group and the different treatment groups of GDM (*p < 0.05; **p < 0.01).

TABLE 1.

Relationship between clinical data of different GDM treatment groups.

Characteristics	Met	Insulin	No‐control	p
Sample size (N)	12	15	18
Age (years)	33.25 ± 2.27	32.53 ± 3.93	32.17 ± 5.02	0.794
Days of delivery (d)	270.42 ± 7.98	269.67 ± 5.15	271.28 ± 6.69	0.784
Gravidity	3.75 ± 2.22	3.07 ± 1.94	2.50 ± 1.10	0.169
Parity	2.17 ± 0.94	1.60 ± 0.74	1.83 ± 0.79	0.210
Height (cm)	161.08 ± 6.96	161.60 ± 5.57	161.40 ± 5.21	0.978
Weight gain during pregnancy (kg)	11.25 ± 6.51	10.88 ± 4.06	12.52 ± 6.23	0.132
Prepregnancy weight (kg)	67.57 ± 14.45	67.10 ± 11.88	63.08 ± 12.71	0.564
Prepregnancy BMI (kg/m2)	25.95 ± 4.90	25.68 ± 4.32	24.20 ± 4.75	0.526
Weight at delivery (kg)	77.98 ± 15.10	77.91 ± 11.70	78.56 ± 11.72	0.987
Delivery BMI (kg/m2)	29.91 ± 4.56	29.80 ± 4.02	30.15 ± 4.48	0.972
Neonatal weight (g)	3470.83 ± 508.89***	3562.67 ± 456.08&&	4106.11 ± 525.16	0.002
OGTT (FBG, mmol/L)	5.25 ± 0.84	5.90 ± 0.91	5.37 ± 0.81	0.119
OGTT (1 h, mmol/L)	11.05 ± 1.80	11.75 ± 1.96	10.64 ± 1.86	0.249
OGTT (2 h, mmol/L)	10.19 ± 2.48	9.92 ± 2.40	9.40 ± 1.09	0.510
HbA1c (%)	5.44 ± 0.23**#	5.87 ± 0.65&&	6.31 ± 0.37	0.000
Ferroprotein (μg/L)	24.10 ± 9.09***###	46.20 ± 10.62&&&	64.5 ± 16.51	0.000
TG (mmol/L)	5.79 ± 0.75	5.54 ± 0.73	5.97 ± 0.60	0.217
TC (mmol/L)	4.17 ± 0.58	3.99 ± 0.43&	4.61 ± 0.90	0.041
TRF (g/L)	3.11 ± 0.49***#	3.54 ± 0.61	3.91 ± 0.53	0.001

Note: *Met versus no‐control, **p < 0.01, ***p < 0.001; ^#Met versus insulin, ^# p < 0.05, ^### p < 0.001; ^&Insulin versus no‐control, ^& p < 0.05,^&& p < 0.01, ^&&& p < 0.001.

Abbreviations: BMI, Body Mass Index, kg/m²; HbA1C, glycosylated hemoglobin; OGTT, oral glucose tolerance test; TC, total cholesterol; TG, triglycerides; TRF, transferrin.

3.2. Metformin Inhibits ATF2 and Promotes the Viability of HTR8/SVneo Cells Under High Glucose Stimulation

The HTR8/SVneo trophoblast cells were exposed to a 25 mM glucose to establish a GDM cell model. CCK‐8 assays showed that Metformin at 0.5 mM significantly enhanced cell viability under high‐glucose conditions compared with untreated HG controls (Figure 2A), and Western blot indicated that the expression of ATF2 decreased and GPX4 increased after MET stimulation (Figure 2B). Overexpressing ATF2 in cells under HG conditions reduced their proliferation and clone formation ability, but adding MET significantly increased cell viability, and further adding Erastin (a ferroptosis inducer that inhibits the cystine–glutamate antiporter system Xc⁻, depleting glutathione and promoting lipid peroxidation) led to a decrease in cell viability (Figure 2C,D). This suggests that MET inhibits ATF2 and promotes the viability of HTR8/SVneo cells under high glucose stimulation.

FIGURE 2.

MET inhibits ATF2 and promotes the viability of HTR8/SVneo cells under high glucose. (A) CCK‐8 results of HTR8/SVneo treated with various concentrations of MET (0, 0.25, 0.5, 0.75, 1.0 mM) under high‐glucose (25 mM glucose) (***p < 0.001, ****p < 0.0001). (B) Western blot analysis of ATF2 and GPX4 (ns, no difference; *p < 0.05; **p < 0.01). (C) CCK‐8 results in HTR8/SVneo overexpressing ATF2 and subsequently treated with 0.5 mM MET in combination with Era (5 μM) under high‐glucose (*HG vs. HG + MET, *p < 0.05, ***p < 0.001; ^#HG + NC‐LV vs. HG + LV, ^### p < 0.001, ^#### p < 0.0001; ^&HG + LV versus HG + LV + MET, ^&&&& p < 0.0001; ^%HG + LV + MET versus HG + LV + MET+Era, ^%%%% p < 0.0001). (D) Colony formation assay results under different treatments. ***p < 0.001, ****p < 0.0001.

3.3. Metformin Inhibits ATF2 and Reduces Ferroptosis in HTR8/SVneo Cells Under High Glucose Stimulation

We next assessed whether Metformin inhibits ferroptosis in trophoblast cells. Flow cytometry analysis showed that high glucose significantly increased intracellular ROS levels, which were further elevated by ATF2 overexpression and reduced by the ferroptosis inhibitor Fer‐1. Metformin treatment decreased ROS levels under HG conditions, and this effect was partially reversed by ATF2 overexpression. After adding ferroptosis activator Era on the basis of ATF2 overexpression combined with MET, the ROS level further increased (Figure 3A). MDA detection also showed that MET significantly inhibited the increase in MDA level caused by ATF2 overexpression under HG conditions (Figure 3B), while GSH showed the opposite result (Figure 3C), indicating that ATF2 overexpression aggravated lipid peroxidation induced by HG, and MET significantly alleviated lipid peroxidation induced by HG. Prussian blue staining revealed marked iron deposition in HG‐treated cells, which was exacerbated by ATF2 overexpression and ameliorated by Metformin (Figure 3D). These data collectively indicate that Metformin inhibits ATF2‐mediated ferroptosis‐associated oxidative stress, iron accumulation, and GPX4 downregulation in trophoblast cells exposed to high glucose, as supported by multiple convergent ferroptosis markers.

FIGURE 3.

Metformin inhibits ATF2 and reduces ferroptosis in HTR8/SVneo cells under high glucose. (A) Intracellular reactive oxygen species (ROS) levels measured by flow cytometry using the DCFH‐DA probe. (B) malondialdehyde (MDA) analysis. (C) Glutathione(GSH) analysis. (D) Prussian blue staining results. **p < 0.01, ***p < 0.001.

3.4. Metformin Inhibits ATF2 From Activating the PI3K/Akt Pathway to Promote the Nuclear Translocation of Nrf2 and Enhance the Expression of GPX4 in HTR8/SVneo Cells Under High Glucose Stimulation

We next examined the molecular mechanism underlying Metformin's anti‐ferroptotic effects. Western blot analysis revealed that after HG stimulation of trophoblast cells, MET could inhibit ATF2 and promote the expression of p‐PI3K, p‐Akt, and GPX4 proteins. The overexpression of ATF2 combined with MET weakened this promoting effect, and the addition of Erastin further attenuated the PI3K/Akt/GPX4 pathway (Figure 4A). Both immunofluorescence and nuclear‐cytoplasmic separation Western blot analysis showed that the overexpression of ATF2 inhibited Nrf2 nuclear entry, while MET could promote Nrf2 entry into the nucleus (Figure 4B,C). These results suggest that Metformin suppresses ATF2 to activate the PI3K/Akt pathway, thereby facilitating Nrf2 nuclear entry and subsequent GPX4 upregulation. To confirm the necessity of the PI3K/Akt/Nrf2 pathway in Metformin's action, we used the PI3K inhibitor LY294002 and the Nrf2 inhibitor ML385. Both inhibitors significantly reduced Metformin‐induced cell viability under HG conditions, as assessed by CCK‐8 and colony formation assays, regardless of ATF2 overexpression status (Figure 5A,B). ROS and MDA levels were significantly higher after both inhibitors' treatment, while GSH levels showed the opposite trend (Figure 5C–E). Prussian blue staining confirmed increased iron deposition after inhibitor treatment (Figure 5F). Immunofluorescence and nuclear‐cytoplasmic separation Western blot analysis also both showed that the expression of Nrf2 in the nucleus decreased after the addition of the inhibitors (Figure 5G,H). Western blot analysis also showed that the expression of p‐PI3K, p‐Akt, and GPX4 decreased after the addition of the inhibitors (Figure 5I). These results further confirm that even under high expression of ATF2, MET's regulation of trophoblast cell ferroptosis and cell viability under high glucose stimulation still depends on the activation of the PI3K/Akt pathway and the entry of Nrf2 into the cell nucleus.

FIGURE 4.

Metformin inhibits ATF2 from activating the PI3K/Akt pathway to promote the nuclear translocation of Nrf2 and enhance the expression of GPX4 in HTR8/SVneo cells under high glucose. (A) Western blot analysis of expression of ATF2/PI3K/Akt/GPX4 signaling pathway. (B) Quantification of nuclear/cytoplasmic Nrf2 fluorescence intensity ratio. (C) Western blot analysis of Nrf2 after nucleoplasmic separation. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

FIGURE 5.

Add inhibitors to verify the effect of MET on ferroptosis of trophoblast cells in HTR8/SVneo cells under high glucose. (A) CCK8 assay of the proliferation ability of trophoblast cells after adding inhibitors. (B) Colony formation ability analysis. (C) ROS analysis by flow cytometry using the DCFH‐DA probe. (D) MDA analysis. (E) GSH analysis. (F) Prussian blue staining results. (G) Quantification of nuclear/cytoplasmic Nrf2 fluorescence intensity ratio. (H) Western blot analysis of Nrf2 after nucleoplasmic separation. (I) Western blot analysis of expression of ATF2/PI3K/Akt/GPX4 signaling pathway. *p < 0.05, **p < 0.01, ***p < 0.001.

3.5. Metformin Down‐Regulates ATF2 Expression to Inhibit Ferroptosis in the STZ‐Induced Hyperglycemic Rat Model

Finally, in order to evaluate the effect of MET regulating ATF2 in vivo on iron death of trophoblast cells in GDM, we established a GDM rat model (Figure 6A). Compared with the control group, the OGTT in the model group was significantly increased. Metformin treatment significantly improved glucose tolerance, as reflected by reduced OGTT AUC values (Table 2). There was no significant difference in the number of fetal rats among the groups (p > 0.05). Fetal weight was significantly increased in the STZ Rats compared with controls, and Metformin treatment significantly reduced fetal weight (p < 0.01) (Figure 6B). Immunohistochemical staining showed that the expression of ATF2 increased and the expression of GPX4 decreased in the STZ Rats, while the expression of ATF2 decreased and the expression of GPX4 increased after MET treatment. Compared with the STZ + ATF2‐over group, the expression of ATF2 decreased and the expression of GPX4 increased in the STZ + ATF2‐over + MET group (Figure 6C). ROS and MDA levels were elevated in the STZ Rats and further increased by ATF2 overexpression, but significantly reduced by Metformin; GSH levels showed the opposite pattern(Figure 6D,E). Western blot analysis of placental tissues confirmed that Metformin downregulated ATF2 and upregulated p‐PI3K, p‐Akt, and GPX4, effects that were attenuated by ATF2 overexpression (Figure 6F). Nuclear‐cytoplasmic separation Western blot further demonstrated that Metformin promoted Nrf2 nuclear translocation in vivo, which was suppressed by ATF2 overexpression (Figure 6G). Those further confirmed that MET can inhibit ATF2 and activate the PI3K/Akt signaling pathway to reduce ferroptosis of trophoblast cells in STZ Rats and exert a promoting cell viability effect.

FIGURE 6.

Metformin inhibits trophoblast ferroptosis via ATF2 suppression and PI3K/Akt/Nrf2 activation in the STZ‐induced hyperglycemic Rat Model. (A) Schematic description of the animal experimental design. (B) Comparison of the number and weight of fetal rats. (C) The immunohistochemical results of ATF2 and GPX4 in rat placental tissues (10×, Scale bar: 200 μm; 40×, Scale bar: 50 μm). (D) ROS analysis by flow cytometry using the DCFH‐DA probe in rat placental tissues. (E) Malondialdehyde (MDA) and glutathione (GSH) analysis in rat placental tissues. (F) Western blot analysis of expression of ATF2/PI3K/Akt/GPX4 signaling pathway in rat placental tissues. (G) Western blot analysis of nuclear and cytoplasmic protein expression of Nrf2 in rat placental tissues. (H) The schematic diagram of the model presented in this study. *p < 0.05, **p < 0.01, ***p < 0.001.

TABLE 2.

Rat OGTT results (x ± sd, mmol/L).

Group	0 h	1 h	2 h	AUC
Control	4.92 ± 0.44	6.15 ± 0.51	5.67 ± 0.39	11.44 ± 0.88
STZ	19.73 ± 2.15	29.53 ± 2.09	26.92 ± 2.55	52.86 ± 3.57**
STZ + LV	21.42 ± 3.80	28.28 ± 4.13	27.37 ± 4.95	52.68 ± 6.76
STZ + NC‐LV	21.13 ± 2.85	27.33 ± 3.00	26.60 ± 3.99	51.20 ± 5.98
STZ + MET	10.08 ± 1.99	19.32 ± 2.70	12.98 ± 3.66	30.85 ± 3.82&&
STZ + LV + MET	10.38 ± 2.69	15.52 ± 3.17	12.02 ± 2.37	26.72 ± 3.98

Note: *Control versus STZ, **p < 0.01; ^&STZ versus STZ + MET, ^&& p < 0.01.

Abbreviations: OGTT, oral glucose tolerance test; STZ, intraperitoneal streptozotocin (50 mg/kg per day) to represent the in vivo model of GDM; STZ + LV, STZ+ 1 × 10⁸ TU of LV‐ATF2‐ overexpression by injecting through the tail vein; STZ + NC‐LV, STZ + 1 × 10⁸ TU of empty vector GV248 through the tail vein; STZ + MET, STZ+ Metformin (300 mg/kg per day); STZ + LV + MET, GDM+ 1 × 10⁸ TU of LV‐ATF2 + Metformin.

4. Discussion

Due to the high prevalence of risk factors for metabolic diseases, the incidence of gestational diabetes mellitus (GDM) is on the rise globally [10]. Its essence is a metabolic disorder centered around insulin resistance and compensatory hyperinsulinemia, which can cause short‐term and long‐term complications for both the mother and the newborn, including preeclampsia, hypertension in the mother, large‐for‐gestational‐age infants, hypoglycemia, and metabolic diseases in adolescents [11].

Multiple related clinical, in vitro, and in vivo studies have confirmed the close relationship between iron overload and the onset of GDM. Ferroptosis is an iron‐dependent cell death process characterized by iron dysregulation leading to iron‐dependent lipid peroxidation‐mediated programmed cell death. Iron overload during pregnancy can trigger ferroptosis, thereby causing GDM [12]. Previous studies have suggested that a high blood glucose environment, decreased glutathione levels, iron transport disorders, and increased lipid peroxidation in trophoblast cells thus all indicate that high blood glucose can lead to ferroptosis. Accumulation of iron may cause exacerbation of lipid peroxidation and protein carbonation within the cells, thereby triggering ferroptosis [13]. Our previous research found that placental trophoblast cells of GDM patients showed mitochondrial swelling, cristae rupture or disappearance, which is consistent with the ultrastructural features of ferroptosis. Compared with normal pregnant women, GPX4 was significantly downregulated in the placenta of GDM patients, and the high expression of ATF2 in GDM and its role in promoting trophoblast ferroptosis by inhibiting the PI3K/Akt/Nrf2 signaling pathway [9]. This suggests that ferroptosis is a key link connecting the high glucose metabolic disorder of GDM with trophoblast cell dysfunction. Systemic intervention in iron homeostasis and lipid peroxidation pathways is expected to improve maternal and fetal outcomes [14].

We have previously identified that the dysregulation of iron homeostasis genes in placental trophoblast cells and ferroptosis may be potential mechanisms underlying the development of gestational diabetes. This study aims to determine whether Metformin treatment for GDM is correlated with the regulation of trophoblast cell ferroptosis through clinical research, molecular biological data, and the construction of animal models, in order to explore the therapeutic mechanism of Metformin in GDM.

Although the insulin resistance of GDM patients increases throughout the entire pregnancy process, insulin remains a widely accepted and safe option due to its safety in not crossing the placental barrier [15]. However, during pregnancy, insulin treatment may have reasons such as cost, availability, fear of needles and hypoglycemia, as well as patient preferences [16]. Metformin can serve as an alternative solution. The traditional view holds that Metformin mainly reduces blood glucose by inhibiting hepatic gluconeogenesis and enhancing peripheral insulin sensitivity [17]. However, several recent clinical cohorts have found that GDM patients receiving Metformin treatment, even with comparable blood glucose control to the insulin group, have significantly lower rates of macrosomia in newborns, higher levels of lipid peroxidation in umbilical cord blood, and less placental iron deposition, suggesting the existence of “hypoglycemic extrinsic effects” [18]. Many articles have compared the safety and efficacy of using Metformin and insulin for the treatment of GDM, and have compared various factors such as fasting blood glucose levels, gestational duration, and neonatal body weight of the patients. Compared with insulin treatment, pregnant women using Metformin had a significantly shorter gestational duration, and the increase in maternal body mass was also lower than that in the insulin group. There was no statistically significant difference in the incidence of cesarean section, preeclampsia, and premature birth between the Metformin group and the insulin group. Compared with insulin treatment, women with GDM who took Metformin had less weight gain, and the occurrence of severe hypoglycemia and gestational hypertension syndrome was also reduced. At the same time, Metformin can reduce the possibility of GDM patients developing Type 2 diabetes in the long term [19]. At the same time, a study observed GDM pregnant women who were treated with Metformin for two consecutive years. The follow‐up of their offspring revealed that, through bioimpedance assessment, children exposed to Metformin in the womb had slightly larger mid‐upper arm circumference, subscapular skinfold thickness, and biceps skinfold thickness. However, there were no differences in total fat mass and body fat percentage, abdominal circumference, and no differences in physical, social, and language development assessments compared to the insulin group. Therefore, the safety, efficacy and mechanism of action of Metformin in the treatment of GDM warrant further investigation [20].

Recently, it has been discovered that Metformin and ferroptosis are also related. Many studies have focused on metabolic‐related diseases. In nonalcoholic fatty liver disease rats, Metformin activates Nrf2, increases the expression of GPX4 and Solute Carrier Family 7 Member 11 (SLC7A11), and exerts an anti‐ferroptosis effect on vascular calcification [21]. Another study found that Metformin can inhibit the formation of foam cells and the occurrence of ferroptosis through the AMPK/ERK signaling pathway, slow down the development of atherosclerosis [22], and also inhibit ferroptosis through the AMPK/Nrf2 pathway to slow down the development of diabetic cardiomyopathy [23]. There have also been experimental studies on the role of ferroptosis in high glucose and high fat‐induced damage to pancreatic β cells, and it has been pointed out that Metformin may exert an anti‐ferroptosis effect by upregulating GPX4, thereby playing a role in protecting pancreatic β cells in Type 2 diabetes [24]. Some studies have confirmed that Metformin inhibits the activity of the p38MAPK/ATF2 signaling pathway to improve cognitive dysfunction in diabetic rats, alleviate neuro‐pathological damage in the hippocampus, inhibit hippocampal neuron apoptosis, and the expression of apoptosis‐related proteins. Some studies have also found that Metformin exerts anti‐tumor effects by inducing ferroptosis in lung cancer cells through the Nrf2/HO‐1 signaling pathway, providing a theoretical basis for the drug treatment of lung cancer patients [25].

The results of this study show that under high glucose stimulation, the expression of ATF2 significantly increases, accompanied by a decrease in PI3K/Akt phosphorylation levels, a reduction in Nrf2 nuclear translocation, and a downregulation of GPX4; while treatment with Metformin or knockdown of ATF2 can reverse these changes and inhibit lipid ROS and iron deposition (Figure 6H). PI3K/Akt is a classic antioxidant stress pathway [26], and its activation can enhance the stability of Nrf2 and promote its nuclear translocation [27]. Nrf2, as the master switch of antioxidant genes, can bind to ARE sequences once it enters the cell nucleus and activate the transcription of downstream genes such as GPX4 and HO‐1, thereby eliminating lipid peroxides and inhibiting ferroptosis [28]. This experiment observed that after Metformin restored the levels of p‐PI3K and p‐Akt, the nuclear localization signal of Nrf2 significantly increased, and the expression of GPX4 subsequently rebounded, indicating that the activity of PI3K/Akt is a necessary prerequisite for the nuclear translocation of Nrf2 and the upregulation of GPX4. This sequential relationship was further verified in the intervention experiments with LY294002 (PI3K inhibitor) and ML385 (Nrf2 inhibitor). Both inhibitors significantly weakened the induction effect of Metformin on GPX4 and re‐elevated reactive oxygen species (ROS), malondialdehyde (MDA), and iron deposition levels, indicating that the anti‐ferroptosis effect of Metformin depends on the integrity of the PI3K/Akt‐Nrf2‐GPX4 axis. This study combined functional gain and loss strategies to confirm that ATF2 is the upstream negative regulatory node of the PI3K/Akt/Nrf2/GPX4 axis and ATF2 may serve as a potential upstream mediator through which metformin exerts its protective effects on placental trophoblasts. Recent studies have also revealed multiple pathways that regulate ferroptosis in trophoblast cells during gestational diabetes mellitus (GDM). The research shows that GLUT1 exacerbates ferroptosis in trophoblast cells by modulating AMPK/ACC‐mediated lipid metabolism, thereby promoting fetal growth restriction associated with gestational diabetes [29]. Adiponectin corrects the imbalance of fatty acid oxidation/oxidation by restoring the activity of CPT‐1, thereby improving placental damage in patients with gestational diabetes [5]. These studies emphasize the diversity of the multiple pathways that jointly act on ferroptosis regulation, including antioxidant responses, glucose transport and metabolic remodeling, as well as fatty acid oxidation. However, the upstream transcriptional regulators that link metformin to these ferroptosis pathways have not been fully explored. Our research results reveal a transcription factor‐centered signaling cascade that integrates oxidative stress response with cell survival signal transduction, providing a new mechanism perspective for the protective effect of metformin on the placenta in gestational diabetes. Overall, our research suggests that the protective logic of Metformin for GDM is no longer simply “hypoglycemic”, but may target the “iron overload‐lipid peroxidation‐ferroptosis” axis to reshape the placental antioxidant program, thereby improving placental function for the treatment of gestational diabetes.

However, our research has several limitations. First, the process by which Metformin acts on ATF2 deserves further in‐depth study. Only the overexpression/knockdown + inhibitor model has been adopted. Further experiments related to genomics and drug‐protein binding interactions are needed to confirm whether Metformin directly interferes with the transcriptional activity or phosphorylation state of ATF2. Furthermore, both the 0.5 mM concentration used in our in vitro experiments and the 300 mg/kg/day dose administered by gavage in the STZ‐induced rat model exceed the corresponding clinically relevant ranges. Future studies incorporating more physiologically relevant concentrations—particularly in longer‐term culture systems and refined animal models—are warranted to further substantiate the translational potential of this pathway. Second, in vitro experiments relied solely on the immortalized HTR‐8/SVneo line, whose metabolic and oxidative‐stress responses may differ from those of primary trophoblasts, so expanding the work to include additional cell lines (e.g., BeWo, JEG‐3, primary term trophoblasts) would strengthen the conclusions. Third, it is only applicable to animal experiments currently and the construction of animal models requires further research. STZ is a β‐cell toxin that leads to insulin deficiency and severe hyperglycemia. This model cannot fully replicate the metabolic and hormonal environment of human GDM. Our findings should be regarded as evidence obtained in this experimental hyperglycemic model rather than direct proof of the therapeutic effect on GDM patients. Furthermore, while we observed a significant reduction in fetal weight following metformin treatment in the STZ‐induced hyperglycemic rats, these findings should be interpreted cautiously. Without comprehensive characterization of fetal and placental growth parameters—such as fetal/placental weight ratio, placental efficiency, organ‐specific weights, and histomorphometric analysis—the observed changes in fetal weight alone cannot be conclusively attributed to improved placental function. We did not follow up adult glucose tolerance, neurobehavioral performance, or placental epigenetic changes in the offspring, so the potential long‐term side effects of Metformin and sustained ATF2 suppression on postnatal development remain unexplored. Given that Metformin can penetrate the placental barrier, its safety during pregnancy in patients with gestational diabetes and its long‐term effects on the offspring require further research. We need to establish a more clinically relevant model, such as a diet rich in fat and glucose combined with a low dose of STZ, and multi‐center randomized controlled trials or long‐term follow‐up cohorts of offspring need to be conducted to evaluate the effects of MET exposure on the metabolism and neurodevelopment of the offspring.

5. Conclusion

In conclusion, our in vitro and in vivo experiments demonstrate that, in HTR‐8/SVneo cells and a streptozotocin‐induced GDM rat model, Metformin treatment correlates with reduced ATF2 expression, activates the PI3K/Akt pathway, promotes Nrf2 nuclear translocation, upregulates GPX4, and thereby suppresses ferroptosis and enhances cell viability. These findings suggest that Metformin may exert its therapeutic effects in GDM, at least in part, associated with modulation of this ATF2–PI3K/Akt–Nrf2–GPX4 axis, although further clinical validation is warranted.

Author Contributions

Dandan Xia: conceptualization, validation, writing – original draft, visualization. Yuhui Zhang: methodology, investigation, validation, visualization. Wenjie Liu: methodology, validation. Siyu Li: methodology, investigation, validation. Chenying Zhang: methodology. Guangtong She: conceptualization, methodology, writing – review and editing, supervision. Huiyan Wang: conceptualization, methodology, writing – review and editing, supervision, funding acquisition. All authors have reviewed and approved the final version of this manuscript for publication.

Funding

This work was supported by High‐Level Leading Talents Project of Changzhou Health Commission (2022CZLJ023), Changzhou Sci&Tech Program (CJ20260776, CJ20245042, CJ20220117), and the Changzhou Municipal Health Commission Youth Project (QN202374).

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author.