Nrf2 signalling and autophagy are involved in diabetes mellitus-induced defects in the development of mouse placenta

It is widely accepted that diabetes mellitus impairs placental development, but the mechanism by which the disease operates to impair development remains controversial. In this study, we demonstrated that pregestational diabetes mellitus (PGDM)-induced defects in placental development in mice are mainly characterized by the changes of morphological structure of placenta. The alteration of differentiation-related gene expressions in trophoblast cells rather than cell proliferation/apoptosis is responsible for the phenotypes found in mouse placenta. Meanwhile, excess reactive oxygen species (ROS) production and activated nuclear factor erythroid2-related factor 2 (Nrf2) signalling were observed in the placenta of mice suffering from PGDM. Using BeWo cells, we also demonstrated that excess ROS was produced and Nrf2 signalling molecules were activated in settings characterized by a high concentration of glucose. More interestingly, differentiation-related gene expressions in trophoblast cells were altered when endogenous Nrf2 expression is manipulated by transfecting Nrf2-wt or Nrf2-shRNA. In addition, PGDM interferes with autophagy in both mouse placenta and BeWo cells, implying that autophagy is also involved, directly or indirectly, in PGDM-induced placental phenotypes. Therefore, we revealed that dysfunctional oxidative stress-activated Nrf2 signalling and autophagy are probably responsible for PGDM-induced defects in the placental development of mice. The mechanism was through the interference with differentiation-related gene expression in trophoblast cells.


Introduction
As shown in the literature, pregnant women with either pre-existing diabetes mellitus or gestational diabetes mellitus (GDM) are at higher risk of carrying a fetus with congenital anomalies including phocomelia, cardiac malformations, macrosomia and central nervous system malformations [1,2]. Among the diabetes mellitus-induced congenital malformations, congenital heart disease and anomalies of the nervous system are predominant, posing 3.4 times higher risk and 2.7 times higher risk, respectively [3]. Note that this may be because both the cardiovascular and nervous systems are formed in the early developmental stage, making them more vulnerable to external harmful factors. These congenital development defects comprise pulmonary atresia, double outlet right ventricle, tetralogy of Fallot and fetal cardiomyopathy [4]. However, the exact aetiology and pathogenesis for diabetes mellitus-induced congenital development defects remains controversial.
Hyperglycaemia, which promotes the generation of excess reactive oxygen species (ROS), is proposed to be the most important teratogen leading to congenital disease formation [5]. The well-substantiated excess production of ROS may be partially responsible for the development of congenital defects [6]. Moreover, the congenital defects of fetal organogenesis were caused by either direct damage or indirect impairment arising from hyperglycaemia, both of which affect placental development. Hyperglycaemia will lead to placental dysfunction and subsequent pregnancy complications [7], and GDM has been associated with alterations in placental anatomy and physiology. These alterations are mainly based on changes on the micro-anatomical and/or even molecular level [8].
In mammals, the regulation of autophagy by amino acids, and also by the hormone insulin, has been extensively investigated, but knowledge about the effects of glucose is more limited. Under certain conditions, autophagy can be activated by glucose [9]. Autophagy is critical to the process of development in mouse models [10], it is involved in development of the human placenta, and changes in oxygen and glucose levels participate in regulation of autophagic changes in cytotrophoblast cells [11]. Furthermore, autophagy plays important roles during starvation in the mouse fetus when the supply of amino acids through the placenta is suddenly cut off [12]. Therefore, in this study we focus on the impact of hyperglycaemia on placental development.
Placenta is the physiological and mechanical connection between fetal and maternal tissues for the respiration, nutrition and excretion of the fetus. The placenta also functions as an immunological barrier, fighting off internal infection. It transports pathogens/drugs and stores some fats, glycogen and iron. Finally, the placenta executes critical endocrine functions, producing hormones throughout pregnancy [13].
Mouse placenta is often used for placental study as it shares many similarities with human placenta, although certain unconformities in genetic regulation exist between them [14]. Histologically, the placenta is composed of maternal and fetal portions. The fetal portion of the placenta is derived from extraembryonic mesoderm (allantois). At E8.5, the allantois and the chorion join together in a process called chorioallantoic attachment. Soon afterwards, the chorion starts folding to form the villi, and forming a space into which the fetal blood vessels grow from the allantois [15]. The maternal portion of placenta derives from the maternal vasculature and the decidual cells of the uterus. The two portions are closely allied, functioning as a fetomaternal blend as villi are anchored to the decidua basalis to fulfil placental physiological functions. During placentation, the trophoblast lineage provides most of the structural components for the placenta and is indispensable to contact between fetal and maternal blood [14]. Therefore, a series of regulatory genes of the trophoblast lineage are seriously attended throughout the different developmental stages of development [16].
In this study, we demonstrate that hyperglycaemia could dramatically affect the development of mouse placenta. Furthermore, we demonstrated that the phenotypes in mouse placenta are closely related to the imbalance of oxidative stress-activated nuclear factor-like 2 (Nrf2) signalling and autophagy.

Experimental animals
The Kunming mice used in this study were obtained from the Laboratory Animal Centre of Sun Yat-Sen University (Guangzhou, China). Eight-week-old female mice were used to induce diabetes mellitus by injecting STZ (Sigma, MO, USA) dissolved in 0.01 M citrate buffer at a pH of 4.5 and a dose of 75 mg kg 21 body weight for three consecutive days. Blood glucose levels were measured 7 days after STZ injection by Roche Accu-Chek Aviva Blood Glucose System (Roche, USA). Diabetes mellitus was defined as a non-fasting blood glucose level greater than 288 mg dl 21 (16 mM) [17]. Control mice were maintained euglycaemic prior to and during pregnancy (4-8 mM). Two female mice were housed with one normal male mouse overnight in a cage. The day when vaginal plugs were observed was designated as embryonic day 0.5 (E0.5). During pregnancy, blood glucose levels were monitored every 6 days. At E13.5 and E18.5, the embryos were dissected by Caesarean section after the pregnant mice were anaesthetized by an intraperitoneal injecting pentobarbital (150 mg kg 21 body weight) [18]. The experiments were performed in triplicates for the three embryonic days, with 24 mice assigned to control and pregestational diabetes mellitus (PGDM) groups.

Placental morphological analysis
We examined whether PGDM altered the morphology of mouse placenta. The placentas were harvested at each assigned time. Placentas were photographed and then fixed in 4% paraformaldehyde, then dehydrated, embedded in paraffin wax and serially sectioned at 4 mm. For histology, the sections were de-waxed in xylene, rehydrated and stained with haematoxylin & eosin dyes (H&E) or periodic acid-Schiff dyes (PAS). The sections were photographed using a fluorescence microscope (Olympus IX50) linked to NIS-ELEMENTS F3.2 software. The average size (area) of the placental labyrinth and junctional zones were determined and evaluated through dividing the areas of labyrinth zone by the total area of placental transverse section, as previously described [19], a minimum of randomly five images from five samples were respectively assayed per group and per assigned time. The spongiotrophoblast layer was positive for PAS, which suggests that many of these spongiotrophoblasts were glycogen cells. The ratio of glycogen-positive area to total area was determined as previously described [20]. A minimum of randomly three images from five samples were respectively assayed per group and per assigned time.

Immunostaining
Immunostaining was performed on paraffin transverse sections against proliferating cell nuclear antigen (PCNA), light chain 3 (LC3B) and Nrf2 antibodies. Briefly, placental transverse sections were de-waxed in xylene, rehydrated and then heated in a microwave for antigen retrieval before exposure to the primary antibody with citrate buffer (pH ¼ 6.0). Then, sections were immersed in 3% hydrogen peroxide for 10 min to block rsob.royalsocietypublishing.org Open Biol. 6: 160064 endogenous peroxidase. Non-specific immunoreactions were blocked using 5% inactivated goat serum in PBS for 30 min at room temperature. The sections were washed in PBS and incubated with PCNA (1 : 500, Santa Cruz, sc-7907, CA, USA), LC3B (1 : 200, Cell Signaling Technology, D11, MA, USA) and Nrf2 (1 : 200, Santa Cruz, sc722) antibodies overnight with shaking at 48C. Following extensive washing, the sections were incubated in horseradish peroxidase (HRP) goat antirabbit IgG secondary antibody (1 : 400, EarthOx, 7074S, Millbrae, USA) for 3 h at room temperature in a dark box, and conjugated to DAB (Maixin, Fuzhou, China). After immunostaining, the sections were counterstained with haematoxylin. For intensity, analysis of PCNA expression was selected for semiquantitative analysis by H-SCORE [21].

TUNEL analyses
The extent of apoptosis in the placenta was established using an In Situ Cell Death Detection Kit (Roche, USA). Terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) staining, including the negative control, was performed according to instructions provided by the manufacturer, which we adapted for tissue labelling on the glass slides. The presence of TUNEL þ cells was established using image analysis software (Olympus, Japan). For intensity analysis of TUNEL expression was selected for semiquantitative analysis by H-SCORE [21], both control and experimental groups (n ¼ 8 placentas for each group).

Cell culture and gene transfection
BeWo, the trophoblast-derived choriocarcinoma cell line, was attained from ATCC (American Type Culture Collection, CCL-98, USA). The cells were cultured in a humidified incubator with 5% CO 2 at 378C in six-well plates (1 Â 10 6 cells ml 21 ) containing HAM'S/F-12 (Myclone, USA) supplemented with 10% fetal bovine serum (Gibco, Gaithersburg, MD, USA), and exposed to 7 mM, 17 mM, 30 mM D-glucose (Sigma); 7 mM D-glucose was used as a control, and 30 mM mannitol acting as an osmolarity control. The cells were photographed using an inverted fluorescence microscope (Nikon, Ti-u, Japan) linked to NIS-ELEMENTS F3.2 software. After 72 h incubation, the immunofluorescent staining against Nrf2 (1 : 200, Santa Cruz, sc722), F-actin (1 : 500, Life Technologies, A12379, USA) and LC3B (1 : 200, Cell Signaling Technology, D11) was performed in the incubated BeWo cells. A minimum of five images were assayed per treatment group. For gene transfection, the BeWo cells were transfected by GFP, Nrf2wt or Nrf2-shRNA with the help of lipofectamin 2000 (Invitrogen, CA, USA). Cells were plated to 50-70% confluence at the time of transfection and the preparation of plasmid DNAlipid complexes, which were subsequently added to the cells. Additionally, the HG&Nrf2-shRNA group would be treated with high glucose before transfection. The insertion sequence for the Nrf2 shRNA is GGGCAAGATATAGACCTTGGTCAA GAGCCAAGGTCTATATCTTGCCTTTTTTGA and GCAGT CTTCATTTCTGCTAATCAAGAGTTAGCAGAAATGAAGA CTGTTTTTTGA.

Cell counting kit-8 assay
Cell viability was assessed through a modified cell counting kit-8 (CCK8; Dojindo Molecular Technologies, Japan) assay.
Briefly, 10 ml of CCK8 (5 g l 21 ) was added into 96-well plates and incubated continually for 4 h at 378C. The absorbance values were measured at 450 nm using a BIO-RAD Model 450 Microplate Reader (BIO-RAD, CA, USA). Cell viability was indirectly established by the ratio of the absorbance value of 17 mM, 30 mM D-glucose or 30 mM mannitol-treated cells relative to the control (7 mM D-glucose). The final results were determined from analysing six independent experiments.

Measurement of intracellular reactive oxygen species
Intracellular ROS was determined using a non-fluorescent dye 2 0 7 0 -dichlorodihydrofluorescein diacetate (Sigma-Aldrich), which is oxidized by ROS generated by cells into a fluorescent dye 2 0 ,7 0 -dichlorofluorescin. The control and high glucosetreated BeWo were incubated in the presence of 10 mm 2 0 7 0 -dichlorodihydrofluorescein diacetate for 20 min. The fluorescence was measured using a BD FACSAria (Becton, Dickinson and Company, Franklin Lakes, NJ, USA).

RNA isolation and RT-PCR
Total RNA was isolated from placental tissues or BeWo cells using EZNA. Total RNA Kit (OMEGA, Georgia, USA) according to the manufacturer's instructions. Reverse transcription to synthesize cDNA was accomplished using PrimeScript RT Reagent Kit with gDNA Eraser (Takara, Shiga, Japan). PCR amplification of the cDNA was performed using specific mouse primers shown in electronic supplementary material, figure S1. PCR was performed in a BIO-RAD S1000 Thermal cycler (BIO-RAD). The cDNAs were amplified for 40 cycles. One round of amplification was performed at 958C for 5 s, at 568C for 30 s and at 728C for 30 s (TaKaRa, Japan). The PCR products (20 ml) were resolved on 2% agarose gels (Biowest, Spain) in a 1Â TAE buffer (0.04 M Trisacetate and 0.001 M EDTA) and with GeneGreen Nucleic Acid Dye (TIANGEN, China). The reaction products were visualized using a transilluminator (SYNGENE, UK) and a computer-assisted gel documentation system (SYNGENE). The sets of primers used for RT-PCR are provided in the electronic supplementary material, figure S1. The ratio between the intensity of the fluorescently stained bands corresponding to the genes and b-actin was calculated to quantify the level of the transcripts for those genes mRNAs [15,22]. The RT-PCR result was representative of three independent experiments.

Western blotting
Western blotting was performed in accordance with a standard procedure using a polyclonal antibody that specifically recognized P53, PCNA, Kelch-like ECH-associated protein 1 (Keap1), Nrf2, NAD(P)H Dehydrogenase Quinone 1 (NQO1), Beclin1 and LC3B. The collected placental tissues or BeWo cells were frozen in the liquid nitrogen and kept at 2808C. Protein from the placental tissues or BeWo cells was isolated from tissue homogenates or cell lysates using a radio-immuno-precipitation assay (RIPA, Sigma) buffer supplemented with protease and phosphatase inhibitors. Protein concentrations were quantified with the BCA assay.

Data analysis
Data analyses and construction of statistical charts were performed using a GraphPad PRISM 5 software package (Graphpad Software, CA, USA). The results are presented as the mean value ( x + s:e:m:). All data were analysed using ANOVA or t-test, which was employed to establish whether there was any difference between the control and experimental data. p , 0.05 was considered to be significantly different. The datasets supporting this article have been uploaded as part of the electronic supplementary material.

Developmental defects in mouse placenta occurred in the presence of high levels of maternal glucose
In our examination of the pregnant mice, we found that resorption or intermediate fetal death (indicated by arrows in figure 1b,   According to the characteristics of the morphological structure of mice, the placenta could be histologically divided into three layers under transverse section as shown in figure 2ad. In the representative H&E stained transverse sections of placenta, we could see that the ratio of placental junctional zone and labyrinth zone changed at both gestational ages E13.5 (control: 0.74 + 0.02, PGDM: 0.99 + 0.03, p , 0.001, n ¼ 25 for each group) and E18.5 (control: 0.52 + 0.02, PGDM: 0.66 + 0.04, p , 0.01, n ¼ 25 for each group) in the mice (figure 2a -e). It is suggested that in PGDM placenta, the relative area of the labyrinth zone was reduced while the relative area of the junctional zone was enhanced compared with the control group.

Cell proliferation and apoptosis in mouse placenta
were altered by exposure to high maternal glucose levels PCNA was employed to assess whether placental cell proliferation was altered in the presence of high glucose levels E18.5  PGDM: 186.40 + 3.71, p . 0.05, n ¼ 10) did not change significantly in presence of high glucose levels. Moreover, we observed significant reduction of TUNEL þ cells in both the labyrinth and junctional zones at gestational age E18.5 compared with control (figure 3e,f,kl ). Meanwhile, western blotting was used to examine the expressions of P53 and PCNA at the protein level. The results of this showed that P53 expression was downregulated in the PDGM groups at both gestational ages E13.5 ( p , 0.05, n ¼ 3) and E18.5 ( p , 0.01, n ¼ 3), but PCNA expression was downregulated at gestational age E13.5 ( p . 0.05, n ¼ 3) and upregulated at gestational age E18.5 ( p , 0.01, n ¼ 3) compared with control ( figure 3m-o). This implies that high levels of glucose interfere with normal placental cell proliferation and apoptosis.

Exposure to high levels of glucose leads to the excess production of ROS in mouse placenta
In order to determine if oxidative stress is involved in the phenotypes described above, we performed a series of experiments to affect the production of ROS in the presence of high levels of glucose. First, we found that SOD (superoxide dismutase) activities (U mg 21 protein) in the PDGM groups occurred at much higher rate than in the control groups at both gestational ages E13.  figure 5a). The immunostaining of Nrf2, a protein that plays an important role as an antioxidant, showed stronger expression in both the E13.5 and E18.5 PDGM groups than in the control groups (figure 5b -e). The diversity of Nrf2 expressions between the control and PDGM groups is distinct in the high-magnification images of the junctional (figure 5b 0e 0 ) and labyrinth zones (figure 5b 0 -e 0 ). Second, we assessed the mRNA expressions of oxidative stress-related genes including GPX1, Keap1, Nrf2, HO1 and NQO1 in both the presence and absence of high glucose   PDGM groups were also much higher than for the controls (figure 5jk). All of these data suggest that oxidative stress is indeed activated in the environment of high glucose levels in mice.

Exposure to high glucose leads to excess autophagy in mouse placenta and BeWo cells
In order to determine if autophagy is involved in the phenotypes described above, we performed a series of experiments to assess autophagy in presence of high glucose levels.

BeWo is employed to assess whether oxidation stress is involved in the differentiation of trophoblast cells
To further investigate the observations above, BeWo cells derived from human choriocarcinomas were employed to   7ce 0 ). RT-PCR data similarly showed that 30 mM Dglucose and mannitol environments could enhance the expressions of Nrf2 and NQO1, a Nrf2-regulated enzyme, compared with the 7 mM D-glucose control group (Nrf2, control and 30 mM: p , 0.01; control and mannitol: p , 0.05; NQO1, control and 30 mM: p , 0.01, control and mannitol: p . 0.05, n ¼ 3 for each group; figure 7f -g). The enhancement of Nrf2 expression in high glucose environments was confirmed using western blotting (Nrf2, control and 30 mM: p , 0.01, control and mannitol: p , 0.05, n ¼ 3 for each group; figure 7h-i). These results imply that exposure to high levels of glucose could activate oxidative stress to a significant degree in BeWo cells.   . These results showed that HAND1 was highly expressed when Nrf2 was over-expressed and minimally expressed when Nrf2 was knocked-down and exposed to high levels of glucose; that GCM1 expression was another way around as long as the Nrf2 gene manipulation (figure 8i-j).

Discussion
In the PGDM-induced mouse, we noted that fetal resorption occurred much more often than in control mice (figure 1a-e). Obviously, the predominant reason for fetal resorption is the presence of developmental defects in placenta; we confirmed this in mice with PGDM according to the previous reports [7]. Outwardly, placenta of mice with PGDM looks paler and weigh dramatically less than those in control mice, but there is not a significant difference in the placental diameters of the groups at gestational ages E13.5 and E18.5 (figure 1f-k). This implies that PGDM might affect the internal structures of placenta. Therefore, we carefully compared the histological structures of transverse sections of placenta from both the control and PGDM groups. We found that the ratio of spongiotrophoblast cells to labyrinth cells at E13.5 and E18.5 in placenta of mice with PGDM was significantly increased compared with those same ratios in the placenta of controls ( figure 2ae). This suggests that PGDM caused the relative enhancement of the junctional zone and the relative reduction of the labyrinth zone in presence of high levels of glucose.
The labyrinth zone is known as the fetal portion of the placenta, while the junctional zone acts as the maternal portion of the placenta. Since the labyrinth is rich with vascular networks, we determined the expressions of angiogenesis-related genes in these zones. The results showed that VEGF and HIF1a expressions rose and VEGFR1 and FGFR2 expressions decreased in mice with PGDM ( figure 2f -h), which may indicate that placental angiogenesis in the labyrinth zone is in a compensatory stage produced by impairment by high glucose levels.
Holzner et al. [23] studied the correlation between diabetes mellitus and histology of placenta in the past. After that, more evidence demonstrated that diabetes during pregnancy could cause structural and biochemical alterations in placental tissues, which in turn impact the normal physiological functions of placenta [24]. More recent literature even suggests that fetal hyperglycaemia leads to defects in placental angioarchitecture, which could initiate pathological responses such as added risk of cardiovascular diseases in fetal lateral life [25].
These pathological alterations in the placenta include dysfunctions of cell proliferation and apoptosis. Therefore, we used PCNA immunochemistry to determine the rate of cell proliferation. The results showed that cell proliferation decreased in both the labyrinth and junctional zones of the E13.5 PGDM placenta. By contrast, cell proliferation increased in the labyrinth zone and remained unchanged in the junctional zone in the E18.5 PGDM placenta ( figure 3a-d,i,j). Meanwhile, TUNEL staining results showed that PGDM altered cell apoptosis in neither the labyrinth zone nor the junctional zone of the E13.5 placenta and reduced cell apoptosis in both zones in the E18.5 placenta (figure 3e-h,k,l ). This tendency is generally confirmed by the western blotting results ( figure 3m-o). The potential mechanisms of decreased apoptosis in the placentas from the PGDM placenta are as follows. First, p53-induced apoptosis requires an accumulation of ROS, so a strong anti-oxidant intracellular environment that could hinder the induction of apoptosis [26]. In our study, PGDM have excessive ROS, and anti-oxidant of Nrf2 is higher expression, so that might be one of the potential mechanisms. Second, autophagy is important in cell death decisions and can protect cells by preventing them from undergoing apoptosis [27]. However, the experimental results on cell proliferation and apoptosis in the presence of PGDM cannot explain the phenotypes that we observed above. This means that there must be other possible mechanisms involved.
Interestingly, we found more PAS þ staining protrusions extended from the junctional zone or from glycogen islets in the labyrinth zone in the PGDM placenta at both gestational age E13.5 and gestational age E18.5, implying that PGDM promotes trophoblast cell differentiation. In the PGDM mouse placenta, the expressions of many trophoblast differentiation-related genes including Hand1, Mash2, PL1, MMP12 and IGF2 increased; the exception was that expression of GCM1 which was reduced in comparison with the control group (figure 4f -h). Hand1 expression is limited to placental trophoblast cells and is requisite for differentiation and/or maintenance of trophoblast cells [28]. Mash2 is able to suppress the differentiation of trophoblast giant cells (TGC) [29]. PL1 and Placental lactogen-2 (PL2) could act as the specific markers for TGC [30], and PLF works as a chemoattractant for endothelial cells in the maternal uterus [31].
For embryo implantation and trophoblast invasion, it is critical to maintain the close interaction between maternal decidual cells and fetal trophoblast cells. However, embryo implantation could not achieve this without the matrix metalloproteinases (MMPs) including MMP12 secreted by decidual cells [32]. The syncytiotrophoblast layer failed to fuse properly in Gcm1-deficient mouse placenta. Glial cells missing-1 (GCM1) plays the very important role in chorioallantoic development by identifying the folding sites of the chorionic plate and invagination of the allantoic mesoderm [33]. Placental IGF2 is deemed to modulate the development of mouse placenta as to diffusional exchange characteristics [34]. Thus, enhanced trophoblast cell differentiation in rsob.royalsocietypublishing.org Open Biol. 6: 160064 PDGM mice might be due to the alterations in gene expressions described above.
HIF1-a are involved in regulating the proliferation and differentiation of human trophoblast cells [31,35], implying that oxidative stress plays a role in placental development. Here, we found that the activity of SOD, one of the key antioxidant enzymes [36], increased in PDGM placenta compared with the controls (figure 5a). The Nrf2 is a transcription factor that acts as the crucial modulator of the redox homeostatic gene regulatory network. Immunostaining of Nrf2 showed that Nrf2 expressions in PDGM placentas at both gestational age E13.5 and gestational age E18.5 were stronger than those in the control placentas ( figure 5b -e).
Meanwhile, the reduction of Keap1 expressions and increases in expressions of Nrf2, HO1 and NQO1 (the downstream genes of Nrf2) seen in PDGM placenta were shown using RT-PCR and western assays (figure 5f-k). These results indicate that oxidative stress is activated by diabetes mellitus. Furthermore, the imbalance of oxidative stress reactions was responsible for interfering with the expressions of trophoblast cell differentiation-related genes in the presence of high levels of glucose. This hypothesis could be partially confirmed by the reduction of the BeWo cell viability rate and by excess ROS production while Nrf2 signalling is activated in BeWo cells caused by high concentrations of glucose ( figure 7).
Next, the correlation between the Nrf2 signalling and trophoblast cell differentiation could be revealed by the alteration of trophoblast cell differentiation-related gene expressions in BeWo cells in the presence of high levels of glucose ( figure 8ab). It is more easily conceivable that the expressions of HAND1 (which promotes chorion into ectoplacental cone or TGC) and GCM1 (which promotes chorion into labyrinth cells) changed as the expressions of Nrf2 genes in BeWo cells were manipulated (figure 8c -g).
Physiologically, autophagy is the process by which energy is supplied for embryonic development through the lysosomal degradation of cellular contents [37]. It has many functions in mammalian development including the development of placenta [38][39][40]. Under some pathological conditions including traumatic brain injury, ischaemia/reperfusion and tumour hypoxia, the relatively excessive accumulation of ROS could break cellular homeostasis, resulting in oxidative stress and mitochondrial dysfunction [41,42].
In this process, ROS could also promote the formation of autophagy [41]. Hung et al. [11] suggested that autophagy is important to protect trophoblast cells from injury caused by deficits in oxygen and glucose during pregnancy. Thus, we also assessed whether the autophagy was induced in the mouse placenta treated with high glucose and/or in the BeWo cells. Autophagy-related genes in the presence of high levels of glucose were detected in this study. People have found that autophagy-related genes (ATG) including ATG5, ATG7 and LC3 participate in the various stages of the autophagy process [43]. The importance of autophagy in embryo development has been confirmed by the mice with either Atg5 or Atg7 mutation could survive in the embryonic period but die soon after birth [44,45]. We found that the ratio of LC3B-II/LC3B-I was increased in groups treated with high levels of glucose compared with the control groups (figure 6e). We also discerned ATG5 and ATG7 are increased in PGDM groups compared with the control groups (figure 6e). When we transfected the Nrf2-wt, we discerned that the ratio of LC3B-II/LC3B-I was also increased compared with the GFP group. Meanwhile, the ratio of LC3B-II/LC3B-I in Nrf2-shRNA transfected BeWo cells at high protein levels was reduced, even after the BeWo cells were exposed to high levels of glucose (figure 8f,h). The same results appear in ATG5 and ATG7 (figure 8f,g). P62 decreased levels can be observed when autophagy is induced [46]. In our study, we have detected P62 was decreased in PGDM compared with the control (figure 6e,f 0 ), and when we transfected the Nrf2-wt, we discerned that the ratio of P62 was also decreased compared with the GFP group. Meanwhile, the ratio of P62 in Nrf2-shRNA transfected BeWo cells at low protein levels was increased, even after the BeWo cells were exposed to high levels of glucose (figure 8f,g). So the decreased levels of P62 in our in vivo and in vitro experiments confirmed autophagy is increased in high-glucose environments. All these data suggest that Nrf2 signalling-activated autophagy might also be involved in the differentiation of trophoblast cells.
In conclusion, we observed defects in placental development in mice with PGDM, which caused the relative reduction of the labyrinth zone and the enhancement of the junctional zone. Further experimental results imply that this might be correlated to the alteration of trophoblast cell differentiation-related gene expressions in high glucose settings rather than cell proliferation and apoptosis. Moreover, excess ROS production and dysfunction of Nrf2 signalling were confirmed as the causes of the alteration of trophoblast cell differentiationrelated genes, as schematically shown in figure 9. In addition, ROS may also promote the formation of autophagy, and the imbalance of cell autophagy might also contribute to the observed placental phenotypes directly or indirectly, although the precise molecular biological mechanism by which it has this effect must still be explored in the future.
Ethics. All processes involving animals in this study were performed in accordance with the procedures of Ethical Committee for Animal Experimentation, Jinan University.  Figure 9. A proposed model that depicts the potential mechanisms for hyperglycaemia-induced defects in placenta. rsob.royalsocietypublishing.org Open Biol. 6: 160064