All procedures involving the use of animals were approved by the Experimental Animal Ethics Committee of the University of Nanjing Medical University (Nanjing, China) with the ethics number IACUC-1702002.

Eight-week-old specific pathogen-free (SPF) ICR mice were bred under controlled environmental conditions (12 h light/dark cycle, relative humidity of 40–70%, and temperature of 20–25°C).

The oestrous cycle phase was determined based on vaginal smears and staining with methylene blue (Shanghai Yuanye Biotechnology Co., Ltd., CAS: 7220-79-3). The oestrous cycle was divided into proestrus, oestrus, metestrus and dioestrus.

Female mice in proestrus or oestrus were injected intraperitoneally (ip) with 10 IU pregnant mare serum gonadotropin (PMSG, Sigma, USA) and then with 10 IU human chorionic gonadotropin (HCG, Sigma) 48 h later (the 10 IU group). Female mice in metestrus were injected ip with 5 IU PMSG and then with 5 IU HCG 48 h later (the 5 IU group).

Female mice in dioestrus were injected ip with physiological saline and then with physiological saline 48 h later (the CON group). Immediately after the HCG or physiological saline injection, female mice were mated with males at a ratio of 2:1. Mating was confirmed the next morning by the presence of a vaginal plug, and this day was considered 0.5 days postcoitum (dpc). If no vaginal plug was observed, vaginal smears were performed for one week.

We have 6 mice in each group. The mice were weighed at 8 a.m. At 3.5, 4.5 and 5.5 dpc, 0.7–0.8 ml of peripheral blood was taken from the inner canthal vein after anaesthetization with 0.2 ml/10 g tribromoethanol. Female mice were sacrificed after the injection of 0.1 ml/10 g trypan blue solution (Sigma, catalogue number: 93595) into their tail vein, and successful injection was confirmed by the mouths and ears of the mice turning blue. The uterus and ovaries were removed immediately and rinsed in precooled phosphate-buffered saline (PBS) (Beyotime, China, catalogue number: ST476). The uterus and ovaries were harvested and photographed immediately and then used for subsequent experiments. The number of implantation sites was determined by trypan blue staining. The ovaries were weighed; half of the uterus was fixed in 4% paraformaldehyde solution, and the other half was placed in a cryotube and stored at -80°C. The blood taken from the inner canthal vein was centrifuged at 3500 rpm for 5 min, and the supernatant was then stored at -80°C before the analysis of 17β-oestradiol levels. The next morning, the 4% paraformaldehyde was replaced with 75% ethanol, and 5 μm sections were used for haematoxylin and eosin (H&E) staining and immunohistochemistry (IHC).

The body weight, size and weight of the ovaries, serum E2 level, oestrous cycle, size and number of implantation sites in the uterus, and localization and expression of oestrogen receptors, TJ proteins, aquaporins and sodium channel proteins in endometrial epithelial cells were assessed.

RL95-2 cells (an endometrial adeno-carcinoma cell line with microvilli on the cell surface) were maintained in 25-cm² flasks using Dulbecco’s minimal essential medium (DMEM)/F12 supplemented with 10% (vol/vol) fetal bovine serum (FBS). The cells were maintained at 37°C in a humidified atmosphere and 5% CO2. RL95-2 cells were initially passaged using a standard trypsinization protocol, plated in 24-well culture dishes, and grown to 70% confluence. The cells were then grown in serum-free, phenol red-free medium for 12 hours before the experimental treatments. The cells were then treated for another 24 hours with either 10^-6M E2 (Sigma, USA, catalogue number: E8875), 10^-6M ERα agonist (PPT) (Tocris, UK, catalogue number: 1026), 10^-6M ERβ agonist (DPN) (Tocris, UK, catalogue number: 1494), or 10^-6M ER antagonist (ICI 182,780) (Tocris, UK, catalogue number: 1047).

A chemiluminescence detection kit (oestradiol determination kit, Beckman Coulter) and a chemiluminescence instrument (UniCel DxI 800, Beckman Coulter) were used for this analysis. Both the intra- and interassay variation were within the set range.

The morphological features of the implantation site were analysed by H&E staining. Uterine tissues isolated from mice were fixed with 4% paraformaldehyde overnight, embedded in paraffin, and cut into 5-μm-thick sections. The sections were deparaffinized and hydrated by brief incubations in xylene, ethanol, and water. Then, the sections were stained with haematoxylin, rinsed, and stained with eosin. The stained sections were dehydrated by brief incubations in water, alcohol and xylene. After mounting, the sections were observed with a bright-field microscope.

The expression of oestrogen receptors, TJ proteins, aquaporins and sodium channel proteins was assessed by IHC. The primary antibodies utilized are listed in Table 1 . PBS rather than the primary antibody served as the negative control.

Tissue was fixed overnight in 4% paraformaldehyde, embedded in paraffin, and cut into 5-μm-thick sections. The sections were deparaffinized, hydrated, and incubated in a histochemistry box containing 10 mM citrate buffer (pH 6.0) at 95°C for 15 min for antigen retrieval. The sections were pretreated with 3% H2O2 in 0.1 M Tris-buffered saline (TBS, pH 7.4) for 10 min to block endogenous peroxidase activity. The sections were washed with PBS three times for 3 min each and then treated with protein blocking solution at 37°C for 10 min and with a primary antibody (diluted 1:100 in 5% BSA or IHC primary antibody dilution buffer) overnight at 4°C. The sections were washed with PBS three times for 3 min each time, covered with enzyme-labelled anti-mouse/rabbit polymer or Solution C and incubated for 10 min at room temperature. They were then covered with horseradish peroxidase (HRP)-labelled streptavidin or Solution D and incubated at room temperature for 10 min. The slides were washed thoroughly with PBS (pH=7.4) between incubations.

After treatment with 3,3’-diaminobenzidine (DAB) chromogen substrate solution, peroxidase bound to the antibody complex was observed. The DAB reaction was monitored under a microscope to determine the optimal incubation time, and the reaction was stopped by washing with 0.1 M TBS multiple times. The immunolabelled sections were dehydrated via a graded ethanol series, cleared in xylene, and fixed. The slides were counterstained with haematoxylin before mounting. Brown deposits indicated positive signals and were evaluated under a Nikon Eclipse Ti microscope (Nikon, Japan).

RL95-2 cells were fixed in 4% paraformaldehyde for 20min. Sections were washed three times with phosphate buffered saline (PBS) for 5 min. Treat with 0.4% Triton X-100 for 10min, and then washed three times with PBS for 5min. Non-specific binding was blocked with 5% bovine serum albumin (BSA) for 30min. Sections were incubated with the following primary antibodies diluted (1:100) in blocking solution (5% BSA) overnight at 4°C. ERα (Abcam ab32063), CLDN3 (Invitrogen, 34-1700), AQP8 (bioworld, BS71279) and SCNN1β (Sigma, SAB5200106). Sections were then washed three times with PBS for 5 min. For the fluorescent detection Alexa Fluor™ 488 goat anti-rabbit (dilution 1:500, Thermo Fisher Scientific) secondary antibody was used and nuclear counterstaining was performed with 4,6-diamidino-2-phenylindole (DAPI, Life Technologies, 10236276001). Evaluation of the sections was performed using confocal microscopy (Nikon, Eclipse Ti, Japan).

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed routinely. Frozen uterine samples were homogenized in tissue lysis buffer (50 mM Tris-HCl pH 7.5, 100 mM NaCl, 2 mM EDTA, 0.1% SDS, 1% NP40) containing 1% protease and phosphatase inhibitors. The tissue homogenates were clarified by centrifugation at 12,000×g for 15 min at 4°C. A BCA protein concentration determination kit (Biyuntian, China, catalogue number: P0010) was used to determine the protein concentration, and the clarified supernatants were mixed with 6× SDS loading buffer (5:1) and transferred to a 70°C water bath for 10 min to denature the proteins. The samples were separated on NuPAGE™ 7% Tris-acetate protein gels (Invitrogen, catalogue number: EA0358BOX) and electrotransferred onto a PVDF Hybond nitrocellulose membrane (Millipore). The membrane was blocked in TBST (10 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 0.05% (w/v) Tween 20) containing 5% skim milk for 1 h at room temperature. Then, the membrane was incubated with a primary antibody (1:1000 dilution) at 4°C overnight, and a rabbit anti-mouse GAPDH polyclonal antibody (1:5000 dilution, Bioworld, catalogue number: AP0063) was used as a control. After washing with TBST, the membrane was incubated in 5% skim milk in TBST containing a secondary antibody (diluted 1:5000, HRP-conjugated goat anti-rabbit IgG, Invitrogen, catalogue number: Ab6721; HRP-conjugated goat anti-mouse IgG), Invitrogen, catalogue number: Ab6789) for one hour at room temperature. The membrane was washed 3 times with TBST, and an ECL kit (Thermo, catalogue number: 1863096; 1863097) was then used to visualize the bands. The membrane was scanned with a luminescence imaging analyser. The relative protein levels were evaluated with ImageJ analysis software (National Institutes of Health, Maryland, Baltimore, USA). The concentration in each sample is expressed as the grey value relative to that of GAPDH or ACTIN.

The data are presented as the means ± SDs of three or more independent experiments. For group comparisons, one-way ANOVA or Student’s t test was performed using Prism software version 5.0 for statistical data analysis (GraphPad Software, Inc.). Differences were considered significant at *p < 0.05, **p< 0.01, and ***p< 0.001.

To investigate the effects of superovulation treatments on female ​mice, 5 IU or 10 IU PMSG/HCG was injected ip into the mice. We assessed their body weight, ovarian morphology, ovarian weight, serum oestradiol levels and oestrous cycle. There was no significant difference in the mouse weight, which was measured every other day, among the different groups ( Supplemental Figure 1 ).

During the peri-implantation period (from 3.5 to 5.5 dpc), the ovaries of the mice in the 5 IU and 10 IU groups were larger than those of the mice in the CON group ( Figure 1A ), and the ovarian weight showed a dose-dependent increase, as the ovarian weights of the 5 IU and 10 IU groups were increased compared with that of the CON group ( Figure 1B ). However, there was no significant increase in the mouse weight ( Supplemental Figure 1 ). These changes were consistent with the clinical characteristics of ovaries after superovulation.

We measured the peripheral serum oestradiol and progesterone concentrations during the peri-implantation period. The serum oestradiol level in the 10 IU group was 2-fold higher than that in the CON group ( Figure 2A ), while the serum progesterone level in the 10 IU group increased by 100% at 3.5dpc and increased by 50% at 4.5dpc compared to the CON group, but the difference was not statistically significant ( Supplemental Figure 2 ). The mice in the CON group had a normal oestrus cycle, progressing from proestrus to oestrus, metestrus and dioestrus. In contrast, the oestrous cycle of the 10 IU group was disordered, and these mice remained in oestrus for a long time ( Figure 2B ). The above results indicate that our model sufficiently simulates the endocrine hormone alterations caused by clinical superovulation and a long-term hyperoestrogen state.

To study the effect of superovulation treatment on embryo implantation, we compared the uterine morphologies, implantation site numbers, and H&E staining of implantation sites among the three different groups.

At 3.5 dpc, no embryos were implanted in the uteri of mice in the CON group, but the uteri of mice in the 10 IU group showed obvious oedema and stubs. At 4.5 dpc, the uteri of mice in the CON group began to show implantation sites, but those of mice in the 5 IU group and 10 IU group did not yet show obvious implantation sites. At 5.5 dpc, the uteri of mice in the CON group showed complete implantation, while implantation in the uteri of mice in the 5 IU group and the 10 IU group occurred later than that in the CON group, and the spacing between the implantation sites was uneven ( Figure 3A ). There were significantly fewer implantation sites in the 10 IU group than in the CON group (p<0.001) and the 5 IU group (p<0.05) ( Figure 3B ).

To investigate the effect of hyperoestrogenism on implantation sites in mice at 5.5 dpc, we performed H&E staining. Morphological changes in the uteri were observed. Embryos in the CON group were evenly spaced and of the same size, and no gap existed between the embryo and the inner membrane. The implantation sites were mainly normal, the embryo spacing was even, and the gap between the embryo and the inner membrane was not obvious in the 5 IU group compared with the CON group; however, the embryo sizes differed between the two groups. In the 10 IU group, the gap between the embryo and the inner membrane was large, the embryo spacing was uneven, and the embryos were too small, resulting in abnormal implantation ( Figure 3C ). The probability of entry and miscarriage was higher in the 10 IU group than in the CON and 5 IU groups ( Figure 3D ).

To study the possible mechanism by which hyperoestrogenism disrupts embryo implantation during ovulation induction, IHC and western blotting were performed to assess the localization and expression of the oestrogen receptors, tight junction factors, the aquaporins and the sodium channel proteins in the endometria of 5.5 dpc mice and RL95-2cell line.

The results showed that ERα was localized in the luminal and glandular epithelia of the mouse endometrium ( Figure 4A ). The protein expression of ERα increased by 50% in the 10 IU group compared to the CON group ( Figure 4B ), and ERβ was not obviously expressed ( Figure 4A ). In RL95-2 cell line, ERα localized in cytoplasm. High oestrogen levels and oestrogen receptor agonists increased the expression of ERα, while oestrogen receptor inhibitors can inhibited the expression of ERα ( Figure 5 ). This result suggeststhat hyperoestrogenism regulates embryo implantation through ERα.

At 5.5 dpc, CLDN3 and OCLN were localized in the epithelium of the mouse endometrial cavity ( Figure 6A ). At 4.5 and 5.5 dpc, the CLDN3 protein level in the 10 IU group was lower than that in the CON group (p<0.05 and p<0.01, respectively, Figure 6B ). In RL95-2 cell line, CLDN3 was mainly localized in membrane. High oestrogen levels and oestrogen receptor agonists depressed the expression of CLDN3, while oestrogen receptor inhibitors can promote the expression of CLDN3 ( Figures 5 , 7 ). This result suggeststhat high oestrogen levels can reduce the expression of CLDN3 through ERα.

At 5.5 dpc, AQP3, AQP4, and AQP8 were not obviously expressed in the murine endometrial cavity epithelium or glandular epithelium ( Figure 8A ). At 4.5 dpc, the expression of AQP8 in the 10 IU group was significantly lower than that in the CON group (p<0.05). At 5.5 dpc, the expression of AQP8 in the 10 IU group was lower than that in the CON group ( Figure 8B ). In RL95-2 cell line, AQP8 was mainly localized in cytoplasm. High oestrogen levels and oestrogen receptor agonists depressed the expression of AQP8, while oestrogen receptor inhibitors can promote the expression of AQP8 ( Figures 5 , 7 ). This result suggeststhat high oestrogen levels can also reduce the expression of aquaporin AQP8 through ERα.

At 5.5 dpc, SCNN1α and SCNN1γ were not obviously expressed. SCNN1β was localized in the murine endometrial cavity epithelium and glandular epithelium ( Figure 9A ).

At 4.5 dpc, the expression of SCNN1β in the 5 IU group was significantly lower than that in the CON group (p<0.05); its expression in the 10 IU group was reduced, but the difference was not significant. At 5.5 dpc, the expression of SCNN1β did not significantly differ among the three groups ( Figure 9B ). In RL95-2 cell line, SCNN1β was mainly localized in cytoplasm and showed polarity. High oestrogen levels and oestrogen receptor agonists depressed the expression of SCNN1β, while oestrogen receptor inhibitors can promote the expression of SCNN1β ( Figures 5 , 7 ).