Decidual Stromal Cell Ferroptosis Associated with Abnormal Iron Metabolism Is Implicated in the Pathogenesis of Recurrent Pregnancy Loss

Iron is necessary for various critical biological processes, but iron overload is also dangerous since labile iron is redox-active and toxic. We found that low serum iron and decidual local iron deposition existed simultaneously in recurrent pregnancy loss (RPL) patients. Mice fed with a low-iron diet (LID) also showed iron deposition in the decidua and adverse pregnancy outcomes. Decreased ferroportin (cellular iron exporter) expression that inhibited the iron export from decidual stromal cells (DSCs) might be the reason for local iron deposition in DSCs from low-serum-iron RPL patients and LID-fed mice. Iron supplementation reduced iron deposition in the decidua of spontaneous abortion models and improved pregnancy outcomes. Local iron overload caused ferroptosis of DSCs by downregulating glutathione (GSH) and glutathione peroxidase 4 levels. Both GSH and cystine (for the synthesis of GSH) supplementation reduced iron-induced lipid reactive oxygen species (ROS) and cell death in DSCs. Ferroptosis inhibitor, cysteine, and GSH supplementation all effectively attenuated DSC ferroptosis and reversed embryo loss in the spontaneous abortion model and LPS-induced abortion model, making ferroptosis mitigation a potential therapeutic target for RPL patients. Further study that improves our understanding of low-serum-iron-induced DSC ferroptosis is needed to inform further clinical evaluations of the safety and efficacy of iron supplementation in women during pregnancy.


Introduction
Miscarriage is the most common complication of pregnancy. Recurrent pregnancy loss (RPL), also known as recurrent spontaneous abortion, defined as two or more consecutive pregnancy losses, affects up to 5% of women trying to conceive. RPL may seriously compromise not only the physical but also the psychological well-being of women. Although genetic abnormalities, infection factors, endocrine disorders, uterine anatomic abnormalities, and antiphospholipid syndrome are known to be common causes of RPL, the etiology of approximately half of the cases is still unexplained [1]. The underlying cytological and molecular mechanisms for RPL remain largely enigmatic [1].
Human pregnancy is a complex process involving interactions between fetal-and maternal-derived components, as revealed by single-cell sequencing [2]. There is an increasing interest in the role of the decidua in orchestrating the homeostatic balance between the mother and fetus, though the role of the decidua in pregnancy maintenance has largely been neglected. Once conception occurs, the developing decidua undergoes dramatic changes in response to invading trophoblasts in such a way that supports further development or, alternatively, promotes active rejection of the embryo [3]. This quality control, which begins

Iron Deposition in DSCs of RPL Patients and Abortion Mice
To differentiate the iron status between normal pregnancy and RPL patients, we first compared their serum iron content. Serum iron in RPL patients was much lower than that in normal early pregnancies ( Figure 1A), although maternal hemoglobin did not differ between the two groups ( Figure 1B). Then we tested the intracellular iron level of DSCs by flow cytometry with calcein, whose fluorescence intensity negatively correlates with liable intracellular ferrous iron levels. Interestingly, the percent of calcein + DSCs was lower in RPL patients than in normal early pregnancies ( Figure 1C). This result was further confirmed by increased ferritin expression, a surrogate marker for the intracellular iron level, in DSCs from RPL patients ( Figure 1D). In addition, transferrin receptor 1 (TfR1), which induces cellular uptake of iron from transferrin (an iron-binding serum protein) by endocytosis, was higher in the RPL decidua, whereas ferroportin (Fpn), the sole membrane iron exporter, was decreased in the RPL decidua ( Figure 1E). The increased expression of TfR1 protein and decreased expression of Fpn protein were also shown in DSCs from RPL patients compared to those from normal pregnancies ( Figure 1F), suggesting that increased iron deposition in the decidua is associated with RPL.
( Figure 1G), accompanied by increased FTH (Ferritin heavy chain coding gene) and decreased SLC40A1 (Fpn coding gene) expression in the uterus of abortion-prone mice (Figure 1H). TFRC (TfR1 coding gene) expression did not differ between the two groups (Figure 1H). A similar phenomenon was also observed in the lipopolysaccharide (LPS)-induced abortion model ( Figure 1I, J). These data imply a correlation between iron deposition in DSCs and an adverse pregnancy outcome, though the serum iron level of RPL patients is low. Data represent the mean ± standard error of the mean (SEM) and are representative of three independent analyses. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

Low-Iron-Diet-Induced Iron Deposition in Decidua and Adverse Pregnancy Outcome in Mice
It was striking that low serum iron and decidual local iron deposition existed simultaneously in RPL patients. We fed female C57BL/6J mice with a low-iron diet (LID) for eight weeks before mating, while control mice were fed a normal diet (ND) (Figure 2A). The LID caused decreased growth in body weight ( Figure 2B), a higher rate of embryo , and TFRC expression in the uterus of pregnant C57BL/6J mice and LPS-treated pregnant C57BL/6J mice. Images are representative of three individual experiments. Data represent the mean ± standard error of the mean (SEM) and are representative of three independent analyses. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.
We established an abortion-prone model using female CBA/J × male DBA/2 mice and found decreased calcein but increased ferritin expression in DSCs of abortion-prone mice ( Figure 1G), accompanied by increased FTH (Ferritin heavy chain coding gene) and decreased SLC40A1 (Fpn coding gene) expression in the uterus of abortion-prone mice ( Figure 1H). TFRC (TfR1 coding gene) expression did not differ between the two groups ( Figure 1H). A similar phenomenon was also observed in the lipopolysaccharide (LPS)induced abortion model ( Figure 1I,J). These data imply a correlation between iron deposition in DSCs and an adverse pregnancy outcome, though the serum iron level of RPL patients is low.

Low-Iron-Diet-Induced Iron Deposition in Decidua and Adverse Pregnancy Outcome in Mice
It was striking that low serum iron and decidual local iron deposition existed simultaneously in RPL patients. We fed female C57BL/6J mice with a low-iron diet (LID) for eight weeks before mating, while control mice were fed a normal diet (ND) (Figure 2A). The LID caused decreased growth in body weight ( Figure 2B), a higher rate of embryo resorption ( Figure 2C,D), and a reduction in placental and fetal weights ( Figure 2E,F). Analysis of the DSCs from the pregnant mice revealed that the percent of calcein + DSCs was lower, while the percent of ferritin + DSCs was higher in the LID group ( Figure 2G), suggesting LID-induced iron deposition in the decidua. In addition, SLC40A1 expression was decreased, while FTH expression was increased in the uterus of the LID group ( Figure 2H). Thus, a LID might induce a reduction in the iron discharge of DSCs, resulting in local iron deposition in the decidua and leading to adverse pregnancy outcomes.  Figure 2C, D), and a reduction in placental and fetal weights ( Figure 2E, F). Analysis of the DSCs from the pregnant mice revealed that the percent of calcein + DSCs was lower, while the percent of ferritin + DSCs was higher in the LID group ( Figure 2G), suggesting LID-induced iron deposition in the decidua. In addition, SLC40A1 expression was decreased, while FTH expression was increased in the uterus of the LID group ( Figure  2H). Thus, a LID might induce a reduction in the iron discharge of DSCs, resulting in local iron deposition in the decidua and leading to adverse pregnancy outcomes.

Iron Supplementation Reduced Iron Deposition in Decidua of Abortion-Prone Mice and Improved Pregnancy Outcome
Both LID-fed mice and abortion-prone mice displayed local iron deposition in the decidua; thus, could iron supplementation ameliorate the pregnancy outcome of abortionprone mice? The World Health Organization recommends routine iron supplementation of 30-60 mg per day throughout pregnancy [15], which, converted into the dose by intraperitoneal injection in mice, is about 0.025-0.05 mg per day. The recommended dose of iron dextran injection for the treatment of human iron-deficiency anemia is 100 mg every one to three days, which, converted into the dose by intraperitoneal injection of mice, is about 0.26 mg per one to three days. As low serum iron existed in RPL patients but they were non-anemic, the spontaneous-abortion-prone mice (female CBA/J mice mated with male DBA/2 mice) were injected intraperitoneally with PBS or 0.1 mg iron dextran (the upper limit of iron supplement dose recommended by WHO but less than that in the treatment of iron-deficiency anemia) every two days. There was no difference in the maternal weight ( Figure 3A), fetal weight, or placental weight ( Figure 3B) of normal pregnant mice (female CBA/J mice mated with male BALB/c mice), spontaneous-abortion-prone mice and spontaneous-abortion-prone mice supplemented with iron. Spontaneous-abortion-prone mice models displayed a high embryo resorption rate compared with normal pregnancy models. Iron supplementation reduced embryo absorption ( Figure 3C,D) and relieved the local iron deposition in the decidua of spontaneous-abortion-prone mice ( Figure 3E). one to three days, which, converted into the dose by intraperitoneal injection of mice, is about 0.26 mg per one to three days. As low serum iron existed in RPL patients but they were non-anemic, the spontaneous-abortion-prone mice (female CBA/J mice mated with male DBA/2 mice) were injected intraperitoneally with PBS or 0.1 mg iron dextran (the upper limit of iron supplement dose recommended by WHO but less than that in the treatment of iron-deficiency anemia) every two days. There was no difference in the maternal weight ( Figure 3A), fetal weight, or placental weight ( Figure 3B) of normal pregnant mice (female CBA/J mice mated with male BALB/c mice), spontaneous-abortion-prone mice and spontaneous-abortion-prone mice supplemented with iron. Spontaneous-abortion-prone mice models displayed a high embryo resorption rate compared with normal pregnancy models. Iron supplementation reduced embryo absorption ( Figure 3C, D) and relieved the local iron deposition in the decidua of spontaneous-abortion-prone mice (Figure 3E).

Iron Overload Caused Ferroptosis of DSCs
As iron deposition in DSCs seemed to be associated with adverse pregnancy outcome, we explored the effect of iron overload on DSCs. Both FeSO 4 (ferrous iron) and ferric citrate (ferric iron) can enhance intracellular iron levels in DSCs, as calcein decreased after FeSO 4 and ferric citrate treatment ( Figures 4A and S1A). Cell damage was aggravated with an increase in FeSO 4 and ferric citrate concentrations ( Figures 4B and S1B). The mitochondrial function of DSCs was significantly impaired, as ATP production capacity decreased with an increase in FeSO 4 concentration ( Figure 4C). Mitochondrial function was further assessed by a Seahorse XF analyzer. As shown in Figures 4D and S1C, FeSO 4 and ferric citrate inhibited the maximum respiratory capacity and respiratory potential of DSCs in a dose-dependent manner. Furthermore, FeSO 4 and ferric citrate also increased the mitochondrial superoxide ( Figures 4E and S1D), intracellular reactive oxygen species (ROS) (Figures 4F,G and S1E), and lipid ROS (Figures 4H and S1F) of DSCs.

Iron Overload Caused Ferroptosis of DSCs
As iron deposition in DSCs seemed to be associated with adverse pregnancy outcome, we explored the effect of iron overload on DSCs. Both FeSO4 (ferrous iron) and ferric citrate (ferric iron) can enhance intracellular iron levels in DSCs, as calcein decreased after FeSO4 and ferric citrate treatment (Figures 4A and S1A). Cell damage was aggravated with an increase in FeSO4 and ferric citrate concentrations ( Figures 4B and S1B). The mitochondrial function of DSCs was significantly impaired, as ATP production capacity decreased with an increase in FeSO4 concentration ( Figure 4C). Mitochondrial function was further assessed by a Seahorse XF analyzer. As shown in Figures 4D and S1C, FeSO4 and ferric citrate inhibited the maximum respiratory capacity and respiratory potential of DSCs in a dose-dependent manner. Furthermore, FeSO4 and ferric citrate also increased the mitochondrial superoxide ( Figures 4E and S1D), intracellular reactive oxygen species (ROS) (Figures 4F, G and S1E), and lipid ROS (Figures 4H and S1F) of DSCs. The rising lipid ROS of DSCs attracted our attention, as lipid peroxidation was a functional marker for ferroptosis [16]. As expected, the normal morphology of the DSCs gradually decreased (Figures 5A and S2A) and cell death increased ( Figures 5B and S2B) under FeSO4 or ferric citrate treatment, and these effects were reversed by ferrostatin-1, an inhibitor of ferroptosis ( Figure 5C, D). Both TRFC overexpression (to induce cellular uptake of iron, Figure 5E) and SLC40A1 downregulation (to inhibit iron export, Figure 5H) enhanced FeSO4-induced lipid ROS ( Figure 5F, I) and cell death ( Figure 5G, J) in DSCs. These data indicate that iron overload causes ferroptosis of DSCs in vitro. The rising lipid ROS of DSCs attracted our attention, as lipid peroxidation was a functional marker for ferroptosis [16]. As expected, the normal morphology of the DSCs gradually decreased (Figures 5A and S2A) and cell death increased (Figures 5B and S2B) under FeSO 4 or ferric citrate treatment, and these effects were reversed by ferrostatin-1, an inhibitor of ferroptosis ( Figure 5C,D). Both TRFC overexpression (to induce cellular uptake of iron, Figure 5E) and SLC40A1 downregulation (to inhibit iron export, Figure 5H) enhanced FeSO 4 -induced lipid ROS ( Figure 5F,I) and cell death ( Figure 5G,J) in DSCs. These data indicate that iron overload causes ferroptosis of DSCs in vitro. (400 µM) with or without SLC40A1 knockdown. Data represent the mean ± standard error of the mean (SEM) and are representative of three independent analyses. * p < 0.05, ** p < 0.01, *** p < 0.001.

Ferroptosis Occurred in the DSCs of Abortion
Next, we sought to determine whether ferroptosis was enhanced in human DSCs from pregnancies complicated by RPL when compared to normal early pregnancies. As shown in Figure 6A, the lipid ROS in DSCs were elevated in these RPL patients compared to normal controls. Cell death of DSCs was also augmented in RPL patients ( Figure 6B). Similar results were observed with respect to the immune-response-mediated spontaneous abortion model (♀CBA/J × ♂DBA/2) and LPS-induced C57BL/6J abortion model (Figure 6C-F). Thus, ferroptosis was enhanced in the DSCs of pregnancies that resulted in abortion, in both the human and mouse model. (400 µM) with or without SLC40A1 knockdown. Data represent the mean ± standard error of the mean (SEM) and are representative of three independent analyses. * p < 0.05, ** p < 0.01, *** p < 0.001.

Ferroptosis Occurred in the DSCs of Abortion
Next, we sought to determine whether ferroptosis was enhanced in human DSCs from pregnancies complicated by RPL when compared to normal early pregnancies. As shown in Figure 6A, the lipid ROS in DSCs were elevated in these RPL patients compared to normal controls. Cell death of DSCs was also augmented in RPL patients ( Figure 6B). Similar results were observed with respect to the immune-response-mediated spontaneous abortion model (♀/J × ♂/2) and LPS-induced C57BL/6J abortion model ( Figure 6C-F). Thus, ferroptosis was enhanced in the DSCs of pregnancies that resulted in abortion, in both the human and mouse model.

Cystine-GSH-GPX4 Axis Participated in the Regulation of Ferroptosis of DSCs
Glutathione peroxidase 4 (GPX4) is an antioxidative enzyme that is one of the central regulators of ferroptosis [17], and it is also expressed in the mammalian uterus [18]. Western blot analysis revealed that treatment with FeSO4 downregulated GPX4 protein levels in the DSCs ( Figure 7A). GPX4 overexpression ( Figure 7B) inhibited FeSO4-induced lipid ROS ( Figure 7C) and cell death ( Figure 7D) in DSCs. Data represent the mean ± SEM and are representative of three independent analyses, *** p < 0.001.

Cystine-GSH-GPX4 Axis Participated in the Regulation of Ferroptosis of DSCs
Glutathione peroxidase 4 (GPX4) is an antioxidative enzyme that is one of the central regulators of ferroptosis [17], and it is also expressed in the mammalian uterus [18]. Western blot analysis revealed that treatment with FeSO 4 downregulated GPX4 protein levels in the DSCs ( Figure 7A). GPX4 overexpression ( Figure 7B) inhibited FeSO 4 -induced lipid ROS ( Figure 7C) and cell death ( Figure 7D) in DSCs.
Given that suppression of glutathione (GSH) biosynthesis with subsequent inhibition or degradation of GPX4 activity both contribute to regulating the initiation and execution of ferroptosis [19], we measured intracellular GSH levels in DSCs exposed to FeSO 4 . As shown in Figure 7B, FeSO 4 downregulated GSH levels in DSCs in a dose-dependent manner. Both GSH and cystine (a substrate for the synthesis of GSH) supplementation reduced FeSO 4 -induced lipid ROS and cell death in DSCs ( Figure 7F-I). These data suggest that the cystine-GSH-GPX4 axis participates in the regulation of ferroptosis in DSCs. Given that suppression of glutathione (GSH) biosynthesis with subsequent inhibition or degradation of GPX4 activity both contribute to regulating the initiation and execution of ferroptosis [19], we measured intracellular GSH levels in DSCs exposed to FeSO4. As shown in Figure 7B, FeSO4 downregulated GSH levels in DSCs in a dose-dependent manner. Both GSH and cystine (a substrate for the synthesis of GSH) supplementation reduced FeSO4-induced lipid ROS and cell death in DSCs ( Figure 7F-I). These data suggest that the cystine-GSH-GPX4 axis participates in the regulation of ferroptosis in DSCs.

Ferroptosis Inhibition Improved Pregnancy Outcomes in Two Abortion Models
To further determine whether ferroptosis modulation affected pregnancy outcomes, we invested the potential value of a ferroptosis inhibitor (liproxstatin-1) in preventing spontaneous abortion in vivo. Liproxstatin-1 effectively reversed embryo loss and reduced lipid ROS and cell death in DSCs from the spontaneous-abortion-prone mice ( Figure 8A-D). Owing to the regulatory effects of the cystine-GSH-GPX4 axis on ferroptosis in DSCs, we further treated spontaneous-abortion-prone mice with cysteine or GSH. As shown in Figure 8A-D, remarkably, both cysteine and GSH also reversed embryo loss and reduced lipid ROS and cell death in DSCs. In parallel experiments, we explored the effects of liproxstatin-1, cysteine, and GSH on LPS-induced abortion. Consistent with the results in the spontaneous-abortion-prone mice, the relatively higher resorption, lipid ROS, and cell death in DSCs of the LPS-induced abortion model was substantially reduced after treatment with liproxstatin-1, cysteine, or GSH ( Figure 8E-H). The results in Figure 8 demonstrate that ferroptosis inhibition has a therapeutic effect on embryo loss in both the spontaneous abortion model and LPS-induced abortion model.

Ferroptosis Inhibition Improved Pregnancy Outcomes in Two Abortion Models
To further determine whether ferroptosis modulation affected pregnancy outcomes, we invested the potential value of a ferroptosis inhibitor (liproxstatin-1) in preventing spontaneous abortion in vivo. Liproxstatin-1 effectively reversed embryo loss and reduced lipid ROS and cell death in DSCs from the spontaneous-abortion-prone mice ( Figure 8A-D). Owing to the regulatory effects of the cystine-GSH-GPX4 axis on ferroptosis in DSCs, we further treated spontaneous-abortion-prone mice with cysteine or GSH. As shown in Figure 8A-D, remarkably, both cysteine and GSH also reversed embryo loss and reduced lipid ROS and cell death in DSCs. In parallel experiments, we explored the effects of liproxstatin-1, cysteine, and GSH on LPS-induced abortion. Consistent with the results in the spontaneous-abortion-prone mice, the relatively higher resorption, lipid ROS, and cell death in DSCs of the LPS-induced abortion model was substantially reduced after treatment with liproxstatin-1, cysteine, or GSH ( Figure 8E-H). The results in Figure 8 demonstrate that ferroptosis inhibition has a therapeutic effect on embryo loss in both the spontaneous abortion model and LPS-induced abortion model.

Discussion
Despite the significant association between imbalances in iron metabolism and severe adverse pregnancy outcomes [9], little is known regarding the iron status of the decidua in the context of RPL. Here, we showed that the serum iron of RPL patients was much lower than that in normal early pregnancies, even though those RPL patients were nonanemic and maternal hemoglobin levels did not differ between the two groups. However, low serum iron and decidual local iron deposition existed simultaneously in RPL patients. This was further confirmed by mice fed with a LID, as the LID induced iron deposition in the decidua and caused an adverse pregnancy outcome in mice.
Several studies have revealed higher levels of circulating iron and local iron overload in preeclampsia [9,20]. It was striking, then, that low serum iron and decidual local iron deposition existed simultaneously in RPL patients. We further found that cellular ironexporter Fpn expression in DSCs was decreased in both RPL patients and LID-fed mice, thus decreasing iron export from the DSCs. This might be the reason for local iron deposition in DSCs from low-serum-iron RPL patients and LID-fed mice. Recent studies have also demonstrated that in iron-deficient mice, placental Fpn expression significantly decreased during the whole gestational period, which compromised iron delivery to the fetus [21,22]. Maternal iron availability in the circulation is mediated by hepcidin, which interacts with Fpn [23]; however, whether Fpn expression in DSCs of RPL individuals is regulated by hepcidin requires further research. Additional studies are also required to determine the exact mechanism(s) of maternal serum iron and Fpn suppression during RPL. In addition, the peripheral ferritin and the soluble transferrin receptor content

Discussion
Despite the significant association between imbalances in iron metabolism and severe adverse pregnancy outcomes [9], little is known regarding the iron status of the decidua in the context of RPL. Here, we showed that the serum iron of RPL patients was much lower than that in normal early pregnancies, even though those RPL patients were non-anemic and maternal hemoglobin levels did not differ between the two groups. However, low serum iron and decidual local iron deposition existed simultaneously in RPL patients. This was further confirmed by mice fed with a LID, as the LID induced iron deposition in the decidua and caused an adverse pregnancy outcome in mice.
Several studies have revealed higher levels of circulating iron and local iron overload in preeclampsia [9,20]. It was striking, then, that low serum iron and decidual local iron deposition existed simultaneously in RPL patients. We further found that cellular ironexporter Fpn expression in DSCs was decreased in both RPL patients and LID-fed mice, thus decreasing iron export from the DSCs. This might be the reason for local iron deposition in DSCs from low-serum-iron RPL patients and LID-fed mice. Recent studies have also demonstrated that in iron-deficient mice, placental Fpn expression significantly decreased during the whole gestational period, which compromised iron delivery to the fetus [21,22]. Maternal iron availability in the circulation is mediated by hepcidin, which interacts with Fpn [23]; however, whether Fpn expression in DSCs of RPL individuals is regulated by hepcidin requires further research. Additional studies are also required to determine the exact mechanism(s) of maternal serum iron and Fpn suppression during RPL. In addition, the peripheral ferritin and the soluble transferrin receptor content should be checked in the future to better explore whether low serum iron and decidual local iron deposition actually exist simultaneously in RPL patients.
Decidual local iron deposition also existed in the spontaneous abortion model and the LPS-induced abortion model. Iron supplementation reduced iron deposition in the decidua of the spontaneous abortion model and improved pregnancy outcome. Though it is common clinical practice to provide iron supplementation to pregnant women [15], it has also been shown that routine iron supplementation in iron-replete women does not translate into improved perinatal outcomes, but rather appears to be associated with significantly more adverse pregnancy events [20]. Iron supplementation beginning in midpregnancy had little to no effect on iron transfer and might even cause adverse effects, such as fetal growth restrictions [24]. However, most studies are based on healthy pregnancies, neglecting conditions with underlying placental or maternal abnormalities. Thus, more evidence-based medical research is needed to determine when and how much iron should be supplemented in pregnant women with a history of RPL, to prevent the occurrence of RPL. Anemia-cutoff should be regarded differently in normal versus pregnancies associated with iron deficiency (especially in RPL), where supplementation may be beneficial.
We found that iron overload can cause ferroptosis in DSCs, and ferroptosis was enhanced in the DSCs of pregnancies ending in abortion, both in the human and mouse models. Ferroptosis is a newly recognized mechanism of programmed cell death, characterized by iron accumulation and lipid-peroxidation-mediated cell membrane damage [16]. Investigating the role of ferroptosis in trophoblasts or placentas brings new insights into understanding the pathogenesis of preeclampsia and other placental-related diseases [12,13]. Uterine and placental ferroptosis also play a role in polycystic-ovary-syndrome-like pregnant rats with oxidative stress-related fetal loss [25]. Several genes are known to modulate the link between ferroptosis and the pathogenesis of miscarriage in humans [26]. DSCs affect the pregnancy microenvironment because of their role in the recruitment, differentiation, and function of immune cells as well as in tissue remodeling. Abnormal apoptosis, senescence, and arrested decidual growth can severely affect the biological role of the decidua, further resulting in tissue dysfunction in neighboring placental tissues [7,8,27]. In the present study, we provide evidence that enhanced ferroptosis in DSCs might also affect the pregnancy outcome and be associated with RPL. In addition, a ferroptosis inhibitor effectively reduced embryo loss in two mouse abortion models. At the molecular level, decreased GPX4 levels or decreased GPX4 activity induces ferroptosis through depletion of GSH and inhibition of lipid peroxidation in different tissues and cells [19]. In vivo knockout studies have shown that mice lacking the entire GPX4 gene experience early embryonic lethality [28]. In accordance with a role of GPX4 in ferroptosis in vivo, we showed that iron-overload-induced DSC ferroptosis was associated with decreased GPX4 and GSH levels. In parallel to mitigating ferroptosis, eliminating circulating ROS, and improving fetal survival [12,13], we found that treatment with cysteine (a substrate for the biosynthesis of GSH) and GSH can attenuate DSC ferroptosis and effectively reverse embryo resorption in the spontaneous abortion model and LPS-induced abortion model. However, further studies are still required to determine the exact mechanism(s) of FeSO 4 effects on cystine-GSH-GPX4 axis. Whether AcSL4-mediated lipid metabolism regulates ferroptosis [29] in DSCs also requires additional research.
Iron homeostasis is required for supporting maternal requirements, placental function, and fetal development, while dysregulated iron status is associated with the occurrences of several pathological conditions, as aberrant accumulation of intracellular iron leads to oxidative stress, which can subsequently promote or amplify ferroptosis [9], although iron supplementation reduced iron deposition in the decidua of the spontaneous abortion model and improved pregnancy outcome. The negative impact of improper iron supplementation on fetal survival in control pregnant rats has been previously reported [30]. Such information, together with a better understanding of low-serum-iron-induced DSC ferroptosis, is valuable in providing further clinical evaluation of the safety and efficacy of iron supplementation for women during pregnancy. Our findings also suggest that ferroptosis-inhibiting agents might broaden the therapeutic strategies for diseases that stem from lipotoxic tissue injury, such as decidua dysfunction that manifest as RPL. Whereas the clinical use of ferroptosis-mitigation strategies is still distant, future deployment of targeted ferroptosis therapeutics may serve to attenuate decidua dysfunction and its sequalae, including RPL.
In summary (Figure 9), we found that low serum iron and decidual local iron deposition existed simultaneously in RPL patients. Decreased ferroportin expression that inhibited the iron export from DSCs may be the reason that local iron deposition occurred in DSCs from low-serum-iron RPL patients and LID-fed mice. Iron supplementation reduced iron deposition in the decidua in the spontaneous abortion model and improved pregnancy outcome. Local iron overload caused ferroptosis in DSCs by downregulating GSH and GPX4 levels. Ferroptosis inhibitor, cysteine, or GSH supplementation effectively attenuated DSC ferroptosis and reversed embryo loss in the spontaneous abortion model and LPS-induced abortion model. Thus, decidual stromal cell ferroptosis associated with abnormal iron metabolism is implicated in the pathogenesis of recurrent pregnancy loss, and ferroptosis mitigation may be a potential therapeutic target for RPL patients. efficacy of iron supplementation for women during pregnancy. Our findings also suggest that ferroptosis-inhibiting agents might broaden the therapeutic strategies for diseases that stem from lipotoxic tissue injury, such as decidua dysfunction that manifest as RPL.
Whereas the clinical use of ferroptosis-mitigation strategies is still distant, future deployment of targeted ferroptosis therapeutics may serve to attenuate decidua dysfunction and its sequalae, including RPL. In summary (Figure 9), we found that low serum iron and decidual local iron deposition existed simultaneously in RPL patients. Decreased ferroportin expression that inhibited the iron export from DSCs may be the reason that local iron deposition occurred in DSCs from low-serum-iron RPL patients and LID-fed mice. Iron supplementation reduced iron deposition in the decidua in the spontaneous abortion model and improved pregnancy outcome. Local iron overload caused ferroptosis in DSCs by downregulating GSH and GPX4 levels. Ferroptosis inhibitor, cysteine, or GSH supplementation effectively attenuated DSC ferroptosis and reversed embryo loss in the spontaneous abortion model and LPS-induced abortion model. Thus, decidual stromal cell ferroptosis associated with abnormal iron metabolism is implicated in the pathogenesis of recurrent pregnancy loss, and ferroptosis mitigation may be a potential therapeutic target for RPL patients. Figure 9. Schematic diagram showing decidual stromal cell ferroptosis associated with abnormal iron metabolism implicated in RPL. Low serum iron and decidual local iron deposition exist simultaneously in RPL patients. Decreased Fpn expression that inhibits iron export from DSCs might be the reason that local iron deposition occurs in DSCs from low-serum-iron RPL patients and lowiron-diet-fed mice. Local iron overload caused ferroptosis in DSCs by downregulating GSH and GPX4 levels. Thus, ferroptosis in DSCs associated with abnormal iron metabolism might be implicated in the pathogenesis of RPL, and ferroptosis mitigation might be a potential therapeutic target for RPL patients.

Human Samples
This study recruited subjects aged from 20 to 35 years old from the Obstetrics and Gynecology Hospital of Fudan University, China, between July 2020 and December 2022 (the clinical characteristics of enrolled subjects are summarized in Table 1). Whole peripheral blood and decidual tissues of human first-trimester pregnancies were obtained under fasting conditions from clinically normal pregnancies (terminated for non-medical reasons, had at least one successful pregnancy and no history of spontaneous abortions, n = 62) and miscarriages (diagnosed as RPL, excluding those resulting from endocrine, anatomic, genetic abnormalities, infection, etc., n = 38).
DSCs were obtained from decidual tissue digesting in DMEM/F-12 supplemented with collagenase type IV (1.0 mg/mL, CLS-1; Worthington Biomedical, Lakewood, NJ, Figure 9. Schematic diagram showing decidual stromal cell ferroptosis associated with abnormal iron metabolism implicated in RPL. Low serum iron and decidual local iron deposition exist simultaneously in RPL patients. Decreased Fpn expression that inhibits iron export from DSCs might be the reason that local iron deposition occurs in DSCs from low-serum-iron RPL patients and low-irondiet-fed mice. Local iron overload caused ferroptosis in DSCs by downregulating GSH and GPX4 levels. Thus, ferroptosis in DSCs associated with abnormal iron metabolism might be implicated in the pathogenesis of RPL, and ferroptosis mitigation might be a potential therapeutic target for RPL patients.

Human Samples
This study recruited subjects aged from 20 to 35 years old from the Obstetrics and Gynecology Hospital of Fudan University, China, between July 2020 and December 2022 (the clinical characteristics of enrolled subjects are summarized in Table 1). Whole peripheral blood and decidual tissues of human first-trimester pregnancies were obtained under fasting conditions from clinically normal pregnancies (terminated for non-medical reasons, had at least one successful pregnancy and no history of spontaneous abortions, n = 62) and miscarriages (diagnosed as RPL, excluding those resulting from endocrine, anatomic, genetic abnormalities, infection, etc., n = 38).

Cell Treatment
Freshly isolated DSCs were cultured overnight in complete medium and further incubated in serum-free medium for 12 h, followed by stimulation with a range of concentrations of the FeSO 4 or ferric citrate for 48 h. In some experiments, DSCs were treated with SLC40A1-specific siRNA (si-SLC40A1: 5 -CCGAUCAAGGUUCAUUCAATT-3 ), or TFRC plasmid for 20 h using transfection reagent (L3000015, Invitrogen, Waltham, MA, USA) according to the manufacturer's instructions.

LDH Cytotoxicity Assay
The extracellular concentration of lactate dehydrogenase (LDH) was used to evaluate cell damage. LDH released by DSCs stimulated with FeSO 4 and ferric citrate was determined by the CyQUANT™ LDH Cytotoxicity Assay Kit (ThermoFisher Scientific, Waltham, MA, USA) following the manufacturer's instructions.

Analysis of ATP
The cellular level of ATP generated by DSCs stimulated by FeSO 4 and ferric citrate was determined using an ATP Assay Kit (S0026, Beyotime Biotechnology, Shanghai, China), following the manufacturer's protocol.

Oxygen Consumption Rate Analysis
The oxygen consumption rate (OCR) of DSCs was measured using a Seahorse XF96 analyzer [32]. For common tests, prepared cultured cells were washed with XF DMEM medium supplemented with 1 mM pyruvate, 2 mM glutamine, and 10 mM glucose, and preincubated at 37 • C for around 45 min in the absence of CO 2 . Proper concentration solutions of oligomycin, FCCP, and rotenone/antimycin A were loaded into the ports on the sensor cartridge, then the program was run according to the manufacturer's protocol for the Seahorse XF Cell Mito Stress Test Kit (103015-100, Agilent, Beijing, China).

7-Amino-Actinomycin D (7-AAD) Staining for Cell Death
Cell death was evaluated by 7-AAD staining [36] (BD Pharmingen, Franklin Lakes, IL, USA). DSCs were harvested and resuspended in 500 µL PBS containing 5 µM 7AAD working solution and then incubated in the dark for 30 min at 37 • C. The 7-AAD fluorescence of DSCs was determined immediately by flow cytometry (Beckman-Coulter, Bria, CA, USA) and analyzed with FlowJo software 7.6 (Tree Star, Ashland, OR, USA).

Analysis of GSH Content
The cellular level of glutathione was determined using a Reduced Glutathione Assay Kit [37] (A006-2-1, Jiancheng, Nanjing, China), following the manufacturer's protocol.

Mice
CBA/J female, DBA/2 male, BALB/c male, C57BL/6J female, and C57BL/6J male mice were purchased from Shanghai SLAC Laboratory Animal Co., Ltd. (Shanghai, China) and Beijing HFK Bioscience Co., Ltd. (Beijing, China) and bred in a room at 22-25 • C, with 40-60% relative humidity and 14 h light-10 h dark cycles. Eight-week-old CBA/J females were mated to BALB/c males to provide normal pregnancy (NP) models. All females were inspected the next morning for vaginal plugs. The day of visualization of a plug was designated as day 0.5 of pregnancy (GD 0.5). Eight-week-old CBA/J females were mated to DBA/2 males to establish spontaneous-abortion-prone (SA) models [38]. For the LPS-induced abortion model (LPS), C57BL/6J females were mated with C57BL/6J males and intraperitoneally injected with 0.25 mg/kg LPS at GD 7.5 [38]. For the low-iron-diet model [39], female C57BL/6J mice were fed with a low-iron diet (0.9 ppm Fe, D08080402, Research Diets, New Brunswick, NJ, USA) for eight weeks before mating. In some groups, pregnant CBA/J mice of SA or pregnant C57BL/6J mice of LPS were administrated with 200 mg/kg GSH (Sigma, Burlington, MA, USA) or 10 mg/kg Liproxstain-1 (Sigma, Burlington, MA, USA) [40] by intraperitoneal injection, or 200 mg/kg cystine intragastrically on GD 2.5, 4.5, and 6.5. In some groups, pregnant CBA/J mice of SA were administrated with 0.1 mg iron dextran or PBS intraperitoneally every 2 days for two weeks. All pregnant mice were monitored at GD 14.5. The percentage of fetal loss (embryo absorption rate) was calculated as: % of resorption = R/(R + V) × 100, where R represents the number of hemorrhagic implantation (sites of fetal loss) and V stands for the number of viable, surviving fetuses.

Intracellular Iron Content Measurement
Intracellular iron concentration was evaluated using flow cytometry [41].

Statistical Analysis
All variables were normally distributed in this study. Thus, variables are presented as means and standard error of the mean (SEM). One-way analysis of variance (ANOVA) was used to evaluate differences. A p-value of less than 0.05 was considered statistically significant. For variables with a p-value of less than 0.05 in ANOVA, a post hoc Dunnett t-test was performed to determine differences between each group. All analyses were carried out using GraphPad Prism 8 software (GraphPad, San Diego, CA, USA).

Institutional Review Board Statement:
This study was approved by the Research Ethics Committee of the Obstetrics and Gynecology Hospital, Fudan University (No. Kyy2018-65). All animal tests were conducted in accordance with the National Guidelines for Animal Care and Use in Research (China). The experimental methods in particular were carried out in accordance with the approved guidelines.
Informed Consent Statement: Patient samples were collected after the signing of consent forms. Data Availability Statement: All data presented in this study are included in this published article and its supplementary information files.

Conflicts of Interest:
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.