WT1 directs normal progesterone receptor-chromatin binding essential for uterine receptivity at peri-implantation.

Wt1 f/f and Pgr Cre/+ mice used in the present study were generated as previously described ( 44 , 45 ). Mice with uterine deletion of Wt1 ( Wt1 d/d ) were obtained by crossing Wt1 f/f mice with Pgr Cre/+ mice. All mice were housed in the animal care facility of Xiamen University according to the guidelines for the care and use of laboratory animals. All experimental procedures were approved by the Animal Care Committee of Xiamen University (XMULAC20170366). For fertility evaluation, 8-wk-old Wt1 f/f and Wt1 d/d females were mated with fertile wild-type males to induce pregnancy. On D5 of pregnancy, IS were visualized by i.v. injection of Chicago blue dye, and the number of IS was recorded. For the establishment of the innovative OVX model during pregnancy, pregnant mice were ovariectomized at 10:00 a.m. on D2 of pregnancy and subcutaneously injected with P4 (2 mg per mouse) dissolved in sesame oil at 10:00 p.m. on D2 of pregnancy. For the establishment of the conventional OVX model, adult female mice were ovariectomized, rested for 2 wk, and subcutaneously injected with E2 (100 ng per mouse) or P4 (2 mg per mouse) dissolved in sesame oil. For the establishment of the delayed implantation model, pregnant mice were ovariectomized at 8 a.m. on D4 of pregnancy and subcutaneously injected with P4 (2 mg per mouse) dissolved in sesame oil for 3 consecutive days starting on D5 of pregnancy. On D7 of pregnancy, an injection of E2 was performed to induce implantation. For the establishment of the P4 supplementation model, pregnant mice were ovariectomized at 8 a.m. on D4 of pregnancy and subcutaneously injected with P4 (2 mg per mouse) dissolved in sesame oil for 3 consecutive days starting on D5 of pregnancy. For the establishment of the artificial decidualization model, one uterine horn of pseudopregnant females, which were obtained by mating with vasectomized males, was infused with 20 μL sesame oil on D4. Decidual reaction was examined 96 h after oil infusion. Mouse blood samples were collected on D4. Serum E2 and P4 levels were measured by radioimmunoassay (RIA). Frozen sections were fixed in 4% paraformaldehyde (PFA) at 4 °C for 15 min. Following prehybridization, the slides were incubated with isotope-labeled probes at 45 °C for 4 h or digoxigenin-labeled probes at 65 °C overnight. Subsequently, the slides were subjected to RNase A treatment at 37 °C for 30 min. The results were detected by autoradiography or antibody against digoxigenin. Probes for Wt1 , Gli1 , Hoxa10 , Angptl7 , Ltf , Muc1 , Nr2f2 , Fst, and Hsd11b2 were used for hybridization. Primers used to synthesize probes have been listed in SI Appendix , Table S1 . Immunostaining was performed as previously described ( 46 ). For paraffin sections, the slides were subjected to heat-mediated antigen retrieval after deparaffinization and rehydration, followed by antigen blocking. For frozen sections, the slides were fixed in 4% PFA for 30 min at room temperature, followed by antigen blocking. The slides were blocked with 1% bovine serum albumin (BSA), and subjected to primary antibody incubation overnight at 4 °C, as well as secondary antibody incubation for 1 h at room temperature. Horseradish peroxidase (HRP)-labeled secondary antibodies (Zhongshan Golden Bridge Biotechnology, 1:200) were used for immunohistochemistry, while fluorescent dye-conjugated secondary antibodies (Jackson ImmunoResearch, 1:200) were used for immunofluorescence. Primary antibodies against WT1 (Abcam, Ab89901, 1:200), PR (CST, 8757, 1:200), ERα (Sigma, 06-935, 1:200), Ki67 (Abcam, Ab15580, 1:100), BrdU (Abcam, Ab6326, 1:200), HAND2 (Abcam, Ab200040 , 1:1,000), MUC1 (Abcam, Ab15481, 1:200), COX2 (CST, 12282, 1:200), 3β-HSDII (Santa Cruz, sc-30820, 1:200), P450scc (Santa Cruz, sc-18043, 1:200), CK (DAKO, Z062201, 1:200), α-SMA (BioGenex, MU128-UC, 1:200), and PDGFRα (CST, 3174, 1:200) were used in the present study, with detailed information in SI Appendix , Table S2 . qRT-PCR was performed as previously described ( 46 ). Total RNA was extracted from uterine tissues or isolated uterine stromal cells using TRIzol (Invitrogen) according to the manufacturer’s protocol. cDNA was synthesized using the PrimeScript RT reagent kit (TAKARA). qRT-PCR was performed using the TB Green Permix Ex Taq II kit (TAKARA). Primers used in the present study have been listed in SI Appendix , Table S1 . WB was performed as previously described ( 46 ). In brief, proteins extracted from uterine samples were subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) separation and polyvinylidene difluoride (PVDF) membrane transfer. The PVDF membrane was blocked with 5% skim milk, followed by primary antibody incubation overnight at 4 °C and secondary antibody incubation (Zhongshan Golden Bridge Biotechnology, 1:5,000) for 1 h at room temperature. The results were visualized by Supersignal West Pico (Thermo Scientific) according to the manufacturer’s protocol. Primary antibodies against WT1 (Abcam, Ab89901, 1:1,000), PR (CST, 8757, 1:500), ERα (Sigma, 06-935, 1:1,000), HA (Abmart, M20003 , 1:1,000), MYC (Abmart, M20002 , 1:1,000), and β-ACTIN (Bioworld, AP0060, 1:5,000) were used in the present study, with detailed information in SI Appendix , Table S2 . Total RNA was extracted from uterine tissue or isolated uterine stromal cells using TRIzol (Invitrogen) according to the manufacturer’s protocol. RNA sequencing was performed using DNBSEQ-T7 (China, BGI). RNA sequencing raw data were subject to quality control and filtration to obtain clean data using Trim Galore. High-quality clean data were aligned to the mouse reference genome (mm10) using STAR. The DESeq2 package was applied to identify DEGs (fold change > 1.2 and P value < 0.05). The ggplot2 and pheatmap packages were applied to generate volcano plots and heatmaps, respectively. Uterine samples were digested into single-cell suspension. The collected cells were crosslinked with 1% formaldehyde (CST) for 15 min at room temperature and terminated with 0.125 M glycine for 10 min on ice. The collected cells were lysed in Lysis Buffer 1 (50 mM HEPES pH7.5, 1 mM EDTA, 140 mM NaCl, 0.5% NP-40, 10% glycerol, 0.25% Triton X-100), Lysis Buffer 2 (10 mM Tris-HCl pH8.0, 1 mM EDTA, 0.5 mM EGTA, 200 mM NaCl), and Lysis Buffer 3 (10 mM Tris-HCl pH8.0, 1 mM EDTA, 0.5 mM EGTA, 100 mM NaCl, 0.1% Sodium Deoxycholate, 0.1% N-lauroylsarcosine). Chromatin DNA was sheared to an average of 200 to 500 bp using the BioRuptor sonicator (Diagenode). DNA fragments were incubated with primary antibody against WT1 (Proteintech, 12609-1-AP) or PR (CST, 8757) overnight at 4 °C, and then incubated with Protein A magnetic beads at 4 °C for 3 h. After washing with Low Salt Buffer (50 mM HEPES pH7.9, 2 mM EDTA, 150 mM NaCl, 1% Triton X-100, 0.5% Sodium Deoxycholate), High Salt Buffer (50 mM HEPES pH7.9, 2 mM EDTA, 500 mM NaCl, 1% Triton X-100, 0.5% Sodium Deoxycholate), LiCl Buffer (10 mM Tris-HCl pH8.0, 1 mM EDTA, 250 mM LiCl, 0.5% NP-40, 0.5% Sodium Deoxycholate), TE Buffer (10 mM Tris-HCl pH8.0, 1 mM EDTA), and elution with Elution Buffer (100 mM NaHCO 3 , 1% SDS), the protein–DNA complex was reversed overnight at 65 °C. Immunoprecipitated DNA was extracted by phenol: chloroform: isoamyl alcohol purification and ethanol precipitation. For sequencing, DNA libraries were constructed using the KAPA DNA Hyper Prep Kit (KK8502) and sequenced using Illumina Nova PE150. For qRT-PCR, average threshold cycle (Ct) values were used to calculate % enrichment compared to 1% input, and negative control Ct value was used to calculate fold enrichment. Detailed information of PCR primers and primary antibodies has been listed in SI Appendix , Tables S1 and S2 , respectively. Chromatin immunoprecipitation sequencing raw data were subject to quality control and filtration to obtain clean data using Trim Galore. High-quality clean data were aligned to the mouse reference genome (mm10) using Bowtie2. Peak calling was performed using MACS2 with default and recommended parameters (FDR threshold = 0.05). Peak annotation was performed using the ChIPseeker package. Peak overlapping was performed using the findOverlapsOfPeaks function of the ChIPpeakAnno package with default and recommended parameters (minoverlap = 1, maxgap = −1). Differential peaks were identified by MACS2 with a False Discovery Rate threshold of 0.05 and |log2FC| > 1. Motif enrichment analysis was performed using HOMER. Heatmaps and profile plots were generated by deepTools. The uterine samples were cut into 1 to 3 mm pieces, fixed in 3% glutaraldehyde and 2% formaldehyde in cacodylate buffer with 2 mM CaCl 2 , postfixed in 1% OsO 4 , and embedded in EMbed812. Ultrastructural analysis was performed using Hitachi HT-7800 transmission electron microscope. The uterine sections were pretreated as described in the immunostaining section. PLA was performed using antibodies against WT1 (Santa Cruz, sc-7385X, 1:200) and PR (CST, 8757, 1:200) with the Duolink In Situ Red Starter Kit (Merck, DUO92101) according to the manufacturer’s protocol. Antibody information is available in SI Appendix , Table S2 . Proteins were extracted from mouse uterine tissue or WT1/PR overexpressed 293T cells. 1 mg proteins were incubated with 1 μg antibodies against WT1 (Santa Cruz, sc-192), PR (CST, 8757), and HA (CST, 3724) overnight at 4 °C, and then incubated with Protein Agarose beads (Thermo) at 4 °C for 3 h. Subsequently, the beads were washed with lysis buffer, boiled at 98 °C for 10 min, and subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis and WB analysis. Antibodies used in this study are listed in SI Appendix , Table S2 . DNA sequences that contained PR and WT1 peaks in Nr2f2 , Fst, and Hsd11b2 genes were amplified by PCR from mouse genomic DNA and cloned into the firefly luciferase reporter vector pGL4.10. The motif mutant vectors were constructed by Quickchange site-directed mutagenesis using the complementary primer pair as shown in SI Appendix , Table S1 . For WT1 and PR overexpression, full-length PRA and WT1 cDNAs were inserted into P-CMV-HA and P-CMV-Myc vectors, respectively. 293T cells were transfected with the firefly luciferase reporter vector pGL4.10, the renilla luciferase control reporter vector pRL-TK, P-CMV-HA-PR, and P-CMV-Myc-WT1 using Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instructions. 1 μM P4 was supplemented 8 h after transfection, and cells were harvested 48 h after P4 treatment. Luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega, E1910). The human endometrial samples were collected by hysteroscopy–laparoscopy surgery in the Department of Obstetrics and Gynecology of the First Affiliated Hospital of Xiamen University. The sampling was approved by the Ethics Committee of Hospital of Xiamen University (XMYY-2021KYSB044), and all participants signed the Informed Consent. Endometriotic lesions were obtained from women (aged 25 to 35 y) with ovarian endometriosis during the secretory phase of the menstrual cycle. These patients showed regular menstrual cycles and had not received hormone treatment for at least 3 mo before the hysteroscopy–laparoscopy surgery. Statistical analyses were performed using the GraphPad Prism 8. The data were presented as mean ± SEM. All experiments included at least three independent replicates. Comparisons between two groups were performed by unpaired two-tailed Student’s t test, and one-way ANOVA followed by Bonferroni’s post hoc test was applied for multiple comparisons. P < 0.05 was considered statistically significant.

In order to gain a comprehensive understanding of P4-PR actions in the peri-implantation uterus, we first established an OVX mouse model during pregnancy in which pregnant mice were ovariectomized on the morning of D2. This time point was chosen because the uterus had been primed with preovulatory E2 before D2, and the P4 surge had not arrived yet. The ovariectomized females were supplemented with exogenous P4 (hereafter named as the P4 group) or sesame oil (hereafter named as the control group) on the evening of D2. Uteri from both the control group and the P4 group were collected 12 h after oil/P4 injection and subjected to bulk RNA sequencing (RNA-seq) and PR ChIP-seq ( Fig. 1 A ). WT1 was bioinformatically predicted as a cooperator of PR. ( A ) Diagram indicating the innovative OVX model established during the peri-implantation period. ( B ) Volcano plot showing the differentially expressed genes (DEGs) in the uterus after P4 injection. Significance is based on the negative binomial test with DESeq2. ( C ) Heatmap showing the expression of previously identified P4 responsive genes in the control and P4 groups. ( D ) Pie chart displaying the genomic distribution of PR. ( E ) Heatmap exhibiting the distribution of PR peaks in the control and P4 groups. ( F ) Motif enrichment analysis of the lost (control group only) and gained (P4 group only) PR peaks. ( G ) Venn diagram showing the overlap between the DEGs upon P4 treatment and genes with PR peaks upon P4 treatment. ( H ) Motif enrichment analysis of the PR-binding sites in the PR target genes identified in ( G ). ( I ) The top-ranked motifs in PR peaks without PRE. 2,449 up-regulated genes and 2,672 down-regulated genes were identified by RNA-seq (fold change ≥ 1.2 and P value < 0.05) upon P4 treatment ( Fig. 1 B ). The levels of previously reported P4 responsive genes were significantly increased after P4 injection ( Fig. 1 C ), demonstrating the efficiency of our OVX-P4 model. According to our ChIP-seq results, PR was predominantly distributed in gene promoters and intergenic regions ( Fig. 1 D ). After P4 supplementation, many PR peaks were up-regulated (gained), while only a few were down-regulated (lost) ( Fig. 1 E ). Motif enrichment analyses were performed on the gained PR peaks and the lost PR peaks, respectively. On the lost PR peaks, which represented peaks that only existed in the absence of P4, motifs of Krüppel-like transcription factors and ETS transcription factors were enriched. Meanwhile, motifs of Homeobox A11 (HOXA11) and chicken ovalbumin upstream promoter-transcription factor II (COUP-TFII) were observed on the gained PR peaks that emerged upon P4 treatment ( Fig. 1 F ). The DEGs in the P4 group were overlapped with genes with PR peaks upon P4 supplementation, and 1,246 genes that were directly regulated by the P4-PR signaling were identified ( Fig. 1 G ). When performing motif analysis on the PR-binding sites in these target genes, we unexpectedly found that the majority of the PR peaks did not possess its canonical motif PRE ( Fig. 1 H ), indicating the existence of other cooperative factors that mediated PR binding onto the chromatin. To explore such cooperators of PR, motif analysis was performed on PR peaks without PRE, and motifs of HOXA11, COUP-TFII, and WT1 were significantly enriched ( Fig. 1 I ). Since the significances of HOXA11 and COUP-TFII in uterine biology had been widely reported ( 19 – 23 ), we intended to explore the potential roles of WT1 in PR actions and uterine functions. We first examined the spatiotemporal expression of Wt1 in the peri-implantation uterus by in situ hybridization (ISH). Wt1 was undetectable on D1. Accompanied with the establishment of uterine receptivity on D4, the expression of Wt1 was significantly up-regulated in the subepithelial stromal cells. After embryo implantation, Wt1 was extensively expressed in decidual cells from D5 to D8 ( Fig. 2 A ). Consistent with its mRNA expression, WT1 protein was specifically located in the stromal cells underneath the luminal epithelium ( SI Appendix , Fig. S1 A ). However, unlike its dynamic mRNA level, the protein level of WT1 was relatively steady from D1 to D4, as revealed by immunohistochemistry (IHC) staining, which was probably attributed to the discrepant stability between mRNA and protein. Since the increase of WT1 mRNA synchronized with the rise of P4 level ( Fig. 2 A ), we surmised that the mRNA expression of WT1 was induced by P4. The conventional OVX model was employed, in which female mice were ovariectomized, rested for 2 wk to exhaust endogenous hormones, and then subjected to P4 injection to evaluate P4 responsiveness ( SI Appendix , Fig. S1 B ). We found that P4 treatment was not capable of up-regulating the mRNA level of WT1 ( SI Appendix , Fig. S1 C ), suggesting that Wt1 was not a P4 responsive gene, at least in this OVX model. Uterine WT1 deficiency resulted in defective uterine receptivity due to hampered P4 responsiveness in stromal cells. ( A ) ISH analysis of Wt1 in the peri-implantation uterus. (Scale bar, 100 μm.) ( B ) The percentage of vaginal plug-positive female mice with implantation sites (IS) in the Wt1 f/f and Wt1 d/d groups. ( C ) The average number of IS in the Wt1 f/f and Wt1 d/d uteri. *** P < 0.001. ( D ) Embryo implantation was examined by blue dye injection in the Wt1 f/f and Wt1 d/d uteri on D5. Morphologically normal blastocysts (Bl) were recovered from the Wt1 d/d uterus without IS. ( E ) IHC analysis of Ki67 and BrdU in the Wt1 f/f and Wt1 d/d uteri on D4. LE, luminal epithelium; S, stroma. (Scale bar, 100 μm.) ( F ) IHC analysis of HAND2 and ISH analysis of Gli1 in the Wt1 f/f and Wt1 d/d uteri on D4. LE, luminal epithelium; S, stroma. (Scale bar, 100 μm.) ( G ) QRT-PCR analysis of Hand2 and Gli1 in the Wt1 f/f and Wt1 d/d uteri on D4. The values were normalized to Gapdh level. Data are presented as mean ± SEM. Two-tailed unpaired Student’s t test. * P < 0.05. ( H ) Diagram indicating the conventional OVX model. ( I ) qRT-PCR analysis of Hand2 and Hoxa10 in the Wt1 f/f and Wt1 d/d uteri upon oil/P4 treatment. The values were normalized to Gapdh level. Data are presented as mean ± SEM. Two-tailed unpaired Student’s t test. * P < 0.05. ( J ) ISH analysis of Hoxa10 in the Wt1 f/f and Wt1 d/d uteri upon oil/P4 treatment. (Scale bar, 200 μm.) ( K ) IHC analysis of HAND2 in the Wt1 f/f and Wt1 d/d uteri upon oil/P4 treatment. (Scale bar, 200 μm.) ( L ) Diagram indicating the P4 supplementation model. ( M ) ISH analysis of Hoxa10 and Angptl7 in the Wt1 f/f and Wt1 d/d uteri after prolonged P4 treatment. (Scale bar, 100 μm.) ( N ) IHC analysis of HAND2 in the Wt1 f/f and Wt1 d/d uteri after prolonged P4 treatment. (Scale bar, 100 μm.) To elucidate the physiological significance of WT1 in uterine functions, mice with uterine ablation of Wt1 ( Wt1 d/d ) were generated by crossing Wt1 floxed mice ( Wt1 f/f ) with PR-Cre transgenic mice ( Pgr Cre/+ ). Specifically, the 8th and 9th exons of Wt1 that encode zinc finger domains were deleted ( SI Appendix , Fig. S2 A ). According to qRT-PCR and western blot (WB), both mRNA and protein of WT1 were effectively depleted in the Wt1 d/d uterus ( SI Appendix , Fig. S2 B and C ), whereas its expression in the ovary was unaffected ( SI Appendix , Fig. S2 C ). Wt1 d/d female mice failed to give birth to offspring when mated with wild-type fertile males ( SI Appendix , Fig. S2 D ). After excluding defects in ovulation ( SI Appendix , Fig. S2 E ) and uterine development ( SI Appendix , Fig. S2 F and G ), we next explored the specific pregnancy stage at which Wt1 d/d females displayed abnormalities. On D5, embryo implantation occurred normally in Wt1 f/f pregnant mice, while only 20% Wt1 d/d females showed sign of implantation ( Fig. 2 B ). Moreover, even in these 20% Wt1 d/d females that possessed IS, the average number of IS was significantly fewer compared to that in the Wt1 f/f group ( Fig. 2 C and SI Appendix , Fig. S3 A ). After sectioning these IS in Wt1 d/d females, we observed aberrant attachment reaction ( SI Appendix , Fig. S3 D and E ). In order to dissect when these pregnancies were lost in Wt1 d/d females, we further examined the pregnancy status on D6 and D8. The embryos showed retarded development on D6, and were ultimately absorbed by D8 ( SI Appendix , Fig. S3 B , C , F , and G ). In addition, the Wt1 d/d uterus failed to undergo oil-induced artificial decidualization ( SI Appendix , Fig. S3 H ). These findings clearly pointed to defective implantation in the absence of WT1, which led to subsequent pregnancy loss. Successful embryo implantation necessitates the synchronization between the acquisition of implantation competence by the blastocyst and the establishment of uterine receptivity. Since morphologically normal blastocysts were recovered from the Wt1 d/d uteri ( Fig. 2 D ), and the gene knockout strategy we used would not affect the genotype of the embryos, we then speculated that uterine WT1 loss compromised uterine receptivity. Under normal circumstances, uterine epithelial cells cease proliferation, while stromal cells undergo extensive proliferation on D4. However, aberrantly persistent epithelial proliferation and obviously reduced stromal proliferation were observed in the Wt1 d/d uterus, according to Ki67 immunostaining and BrdU incorporation assay ( Fig. 2 E ). In addition, the establishment of uterine receptivity requires luminal epithelial membrane transformation during which the microvilli on the apical surface of the luminal epithelial cells become flattening. Nevertheless, this process was impaired in the Wt1 d/d uterus, since sustained long microvilli were observed ( SI Appendix , Fig. S4 A ). Collectively, these results suggested that uterine ablation of Wt1 hampered uterine receptivity, thus leading to implantation failure. Considering that uterine receptivity is under the precise control of ovarian E2 and P4, we subsequently examined the expression of E2/P4 responsive genes in the uterus. Normally on D4, E2 responsive genes lactoferrin ( Ltf ) and mucin 1 ( Muc1 ) need to be down-regulated in the epithelium, while P4 responsive genes heart and neural crest derivatives expressed 2 ( Hand2 ) and GLI-Krüppel family member 1 ( Gli1 ) are induced in the stroma. However, the expressions of Ltf and Muc1 were not timely repressed in epithelial cells ( SI Appendix , Fig. S4 B and C ), and the expressions of Hand2 and Gli1 were not successfully up-regulated in stromal cells in the absence of WT1 ( Fig. 2 F and G ). Given the aberrant expressions of E2/P4 responsive genes upon WT1 loss, we employed the conventional OVX model to investigate whether uterine WT1 deficiency derailed E2/P4 responsiveness ( Fig. 2 H ). Since WT1 was specifically localized in uterine stromal cells, we mainly focused on stromal responsiveness to E2/P4. Exogenous P4 supplementation triggered the expressions of P4 responsive genes, including Hand2 and Hoxa10 , in the uterine stromal cells of Wt1 f/f mice. However, the expressions of Hand2 and Hoxa10 were not elicited by P4 treatment in the Wt1 d/d stroma ( Fig. 2 I – K ), implying hampered P4 responsiveness. Meanwhile, the loss of WT1 exerted no effects on the expression of E2 responsive genes in both stromal cells ( Pgr and insulin like growth factor 1 ( Igf1 )) and epithelial cells ( Muc1 ), as well as E2-dependent epithelial proliferation ( SI Appendix , Fig. S5 A and B ). These findings demonstrated that WT1 was indispensable for uterine responsiveness to P4 rather than E2. Since uterine WT1 depletion resulted in implantation failure due to P4 resistance, we wondered whether prolonged P4 supplementation could rescue the reduced P4 responsiveness and implantation defect in Wt1 d/d females. To address this, we used the delayed implantation model, as previously reported ( 24 ), in order to extend the duration of P4 treatment before embryo implantation ( SI Appendix , Fig. S5 C ). In Wt1 f/f females, the blastocysts successfully attached to the uterine luminal epithelium. Nevertheless, floating embryos were observed in the Wt1 d/d uterine cavity ( SI Appendix , Fig. S5 D ). The absence of cyclooxygenase 2 (COX2), a marker of embryo attachment reaction, in Wt1 d/d stromal cells further confirmed the implantation failure upon WT1 loss even with extended P4 exposure ( SI Appendix , Fig. S5 D ). In addition, prolonged P4 supplementation could not up-regulate the expression of P4 responsive genes (e.g., Hoxa10 , angiopoietin-like 7 ( Angptl7 ), and Hand2 ) in the Wt1 d/d stroma ( Fig. 2 L – N ). These results demonstrated that uterine WT1 deficiency diminished P4 responsiveness in uterine stromal cells. The aforementioned findings indicated that the loss of WT1 impaired P4 responsiveness in the uterus. We then screened for the specific step of the P4 signaling in which WT1 was potentially involved. We first assessed ovarian synthesis and secretion of E2/P4. Neither serum E2/P4 level nor ovarian expressions of 3β-hydroxysteroid dehydrogenase type II (3β-HSD II) and cytochrome P450 cholesterol side-chain cleavage enzyme (P450scc) were affected by Wt1 deletion on D4 ( SI Appendix , Fig. S6 A and B ). Furthermore, the expressions of ERα/PR in the uterus were checked. IHC and WB revealed comparable expression levels of ERα and PR in the Wt1 f/f and Wt1 d/d uteri on D4 ( SI Appendix , Fig. S6 C and D ). Subsequently, we investigated whether WT1 participated in the regulation of PR–chromatin binding and transcriptional activity. WT1 and PR ChIP-seq was performed using uterine cells on D4. 38,363 WT1 peaks and 32,122 PR peaks were identified with high confidence (FDR < 5%). Both WT1 and PR peaks were highly enriched near transcription start sites ( Fig. 3 A ), and showed similar genome-wide distributions, predominantly in promoters, introns, and intergenic regions ( Fig. 3 B ). Motif analysis revealed that PRE was significantly enriched in WT1 peaks, and vice versa ( Fig. 3 C and D ). Furthermore, we noticed that approximately 68% of the PR-binding sites were cobound by WT1 ( Fig. 3 E ), and WT1 peak summits coincided with PR peak summits on these overlapping sites ( Fig. 3 F ). Particularly, WT1 and PR simultaneously occupied on Hoxa10 , Nr2f2 (the gene coding for COUP-TFII), follistatin ( Fst ), and patched 1 ( Ptch1 ), which are uniquely expressed in the uterine stroma and have been reported involved in the regulation of uterine receptivity ( 20 , 25 – 27 ) ( Fig. 3 G and H and SI Appendix , Fig. S7 A ). ChIP-qRT further confirmed the co-occupancy of WT1 and PR on these genes in the uterus on D4 ( Fig. 3 I and J and SI Appendix , Fig. S7 B and C ). WT1 exhibited physical interaction and genomic co-occupancy with PR. ( A ) Heatmap showing the distributions of WT1 and PR near the transcription start site (TSS) in the uterus on D4. ( B ) Genome-wide distributions of WT1 and PR peaks in the receptive uterus. ( C ) The enrichment of PRE in WT1-binding sites. ( D ) The enrichment of WT1 motif in PR-binding sites. ( E ) Venn diagram showing the overlap between WT1 and PR peaks in the uterus on D4. ( F ) Normalized read density of WT1 and PR on PR-binding loci in the receptive uterus. ( G ) Genome browser view of normalized WT1 and PR ChIP-seq signals on the gene loci of Hoxa10 . ( H ) Genome browser view of normalized WT1 and PR ChIP-seq signals on the gene loci of Nr2f2 . ( I ) ChIP-qRT-PCR confirmation of WT1 and PR binding on Hoxa10 . Data are presented as mean ± SEM. Two-tailed unpaired Student’s t test. ** P < 0.01. ( J ) ChIP-qRT-PCR confirmation of WT1 and PR binding on Nr2f2 . Data are presented as mean ± SEM. Two-tailed unpaired Student’s t test. ** P < 0.01, *** P < 0.001. ( K ) PLA analysis indicating the interactions between WT1 and PR in the nuclei of uterine stromal cells on D4. The nuclei were labeled by DAPI. The negative control group was performed without primary antibody incubation. (Scale bar, 10 μm.) ( L and M ) Co-IP assays showing the interactions between WT1 and PR in the D4 uterus. ( N ) Co-IP assay showing the interactions between the two WT1 isoforms and PR. The co-occupancy of WT1 and PR on the chromatin implied a potential interaction between them. In order to prove the physical interaction between WT1 and PR, in situ proximity ligation assay (PLA) was performed. Spatially close proximity between WT1 and PR was observed in the nuclei of uterine stromal cells on D4 ( Fig. 3 K ). Coimmunoprecipitation (co-IP) assay further confirmed the protein–protein interaction between WT1 and PR in the D4 uterus ( Fig. 3 L and M ). WT1 possesses multiple isoforms arisen from alternative splicing, among which KTS+ and KTS− are the most prevalent and well-studied isoforms ( 14 , 15 ). We found that both KTS+ (long) and KTS− (short) isoforms of WT1 could interact with PR ( Fig. 3 N ). Collectively, these findings suggested that WT1 physically interacted with PR, and displayed co-occupancy with PR on target genes in uterine stromal cells. In order to reinforce the importance of WT1 in regulating PR–chromatin binding, PR ChIP-seq was performed using Wt1 f/f or Wt1 d/d uterine cells on D4. Consistent with the impaired P4 responsiveness, a large number of PR peaks were dramatically down-regulated upon Wt1 ablation ( Fig. 4 A – C ). It was noteworthy that the vast majority (>78%) of the decreased PR peaks were co-occupied by WT1 ( Fig. 4 D ). For instances, PR and WT1 simultaneously occupied on Nr2f2 , Fst, and 11β-hydroxysteroid dehydrogenase 2 ( Hsd11b2 ), and the binding of PR on these genes was significantly reduced in the absence of WT1 ( Fig. 4 E ), which was further confirmed by ChIP-qRT ( Fig. 4 F ). These observations suggested that WT1 was indispensable for the recruitment of PR onto target gene loci. WT1 mediated PR–chromatin binding via corecruitment and tethering. ( A ) Heatmap displaying the down-regulated PR peaks in the Wt1 f/f and Wt1 d/d uteri on D4. ( B and C ) Normalized PR tag density in the Wt1 f/f and Wt1 d/d uteri on D4. ( D ) Venn diagram showing the overlap between the down-regulated PR peaks and the WT1 peaks in the D4 uterus. ( E ) IGV visualization of the PR peaks in the Wt1 f/f and Wt1 d/d uteri on D4, as well as the WT1 peaks in wild-type uteri on D4 on the gene loci of Nr2f2 , Fst, and Hsd11b2 . ( F ) ChIP-qRT-PCR confirmation of reduced PR binding on Nr2f2 , Fst, and Hsd11b2 upon WT1 loss. Data are presented as mean ± SEM. Two-tailed unpaired Student’s t test. * P < 0.05, ** P < 0.01. ( G ) The overlap between the DEGs upon Wt1 deletion on D4, the P4 responsive genes identified by our OVX model, as well as the genes with WT1 peaks and down-regulated PR peaks upon Wt1 deletion. ( H ) Heatmap showing the expressions of WT1-PR target genes identified in ( G ) in the Wt1 f/f and Wt1 d/d uteri on D4. ( I ) Diagram indicating the potential mechanisms by which WT1 mediated PR–chromatin binding. ( J – L ) Luciferase report assays showing the transcriptional regulation of target genes by WT1 and PR. Data are presented as mean ± SEM. Two-tailed unpaired Student’s t test. ** P < 0.01, *** P < 0.001. Genes with WT1 peaks and down-regulated PR peaks upon WT1 loss were overlapped with DEGs in Wt1 d/d uterine stromal cells on D4, as well as P4 responsive genes found by our innovative OVX model, and 428 direct target genes of the WT1-PR transcription complex were identified ( Fig. 4 G ). Many of these target genes had been widely reported to participate in the establishment of uterine receptivity and the regulation of uterine functions, and their expressions were significantly decreased after Wt1 deletion ( Fig. 4 H ), which was further validated by ISH and qRT-PCR ( SI Appendix , Fig. S7 D and E ). These results demonstrated that WT1 mediated the recruitment of PR to chromatin to activate the transcription of target genes that were essential for uterine receptivity. To gain further insight into the potential mechanisms by which WT1 mediated PR–chromatin binding, genes with WT1 peaks and down-regulated PR peaks upon WT1 loss were subdivided into four categories according to motif enrichment analysis: 1) with WT1 motif but without PRE; 2) with both WT1 motif and PRE; 3) with PRE but without WT1 motif; 4) with neither WT1 motif nor PRE. We put our emphasis on genes with WT1 motif but without PRE, as well as genes with both WT1 motif and PRE, respectively representing the tethering and corecruitment manners through which WT1 mediated PR–chromatin binding ( Fig. 4 I ). We then performed the luciferase report assay to verify the role of WT1 in PR transcriptional activity. Sequences of PR and WT1 binding sites in Nr2f2 , Fst (with WT1 motif but without PRE, representing the tethering manner), and Hsd11b2 (with both WT1 motif and PRE, representing the corecruitment manner) were inserted upstream of the luciferase reporter gene. In both tethering and corecruitment manners, PR alone was incapable of promoting the transcription of the luciferase gene, while simultaneously overexpressing WT1 significantly activated luciferase activities. Site-directed mutations within the WT1 consensus motif in Nr2f2 , Fst, and Hsd11b2 promoters, as well as PRE in Hsd11b2 promoter, abrogated this effect ( Fig. 4 J – L ). These findings indicated that WT1 mediated PR–chromatin binding through both tethering and corecruitment manners, thus activating the transcription of target genes that were required for the establishment of uterine receptivity. In the human endometrium, PR was highly expressed in epithelial and stromal cells during the proliferative and early secretory phase. Its expression level in epithelial cells significantly declined in the middle secretory phase, and its expression in stromal cells was markedly decreased in the late secretory phase ( SI Appendix , Fig. S8 A ). Meanwhile, the expression of WT1 was restricted to the stromal cells, and its expression level was also decreased in the late secretory phase ( SI Appendix , Fig. S8 B ). These findings indicated that the expression of WT1 coincided with PR expression in the human endometrial stroma across the menstrual cycle. Given that endometriosis is often accompanied with P4 resistance, we analyzed the dataset ( GSE51981 ) of the endometrial tissues of patients with endometriosis and healthy women to explore the potential correlation between WT1 expression and P4 responsiveness. We observed a significant decrease in WT1 expression in the endometriosis group, as well as a slight reduction in the level of HOXA10 , a P4 responsive gene in HESCs, which indicated aberrant stromal P4 responsiveness ( SI Appendix , Fig. S8 C ). In addition, the expression level of HOXA10 was positively correlated with that of WT1 regardless of endometrial pathologic conditions ( SI Appendix , Fig. S8 D ). To confirm the findings above, we also collected the eutopic and ectopic endometrial samples from patients with endometriosis, and performed WB analysis of WT1. Both eutopic and ectopic endometrial tissues exhibited decreased expression level of WT1 compared with the endometrial tissues from healthy women ( SI Appendix , Fig. S8 E ), further demonstrating that decreased WT1 expression in the human endometrium was associated with P4 resistance-related disorders such as endometriosis.

The P4-activated PR signaling plays a dominant role in female reproduction, yet the detailed mechanisms underlying PR actions in the uterus remain poorly understood. Our study demonstrated that WT1 functioned as a cooperative factor of PR in uterine stromal cells by directly interacting with PR and mediating its chromatin binding via tethering and corecruitment manners, and thus enhanced PR responsiveness conducive to uterine receptivity and embryo implantation ( Fig. 5 ). Schematic diagram illustrating the role of WT1 in directing PR–chromatin binding in the uterus. In uterine stromal cells, WT1 physically interacted with PR, directed PR onto the chromatin via corecruitment and tethering manners, and facilitated PR to activate the transcription of target genes that were essential for uterine receptivity. Employing the conventional OVX mouse model and PR ChIP-seq, a previous study has uncovered a list of PR cooperators in the uterus, including GATA2 and SOX17 ( 10 ), which exhibit co-occupancy with PR on P4 responsive genes and modulate the PR signaling ( 11 , 12 ). However, most of the PR cooperators identified in this study are localized in uterine epithelial cells, owing to the fact that the ovariectomized females are deprived of hormones, so that the basal expression of PR in the uterus is restricted to epithelial cells. In order to seek for more PR cooperative factors in uterine stromal cells, we established an OVX model during the peri-implantation period to ensure that PR was properly expressed in stromal cells. An additional advantage of our model was that the uterus had been primed with preovulatory E2 before OVX, and could better recapitulate uterine P4 responsiveness under physiological conditions. Through PR ChIP-seq and motif enrichment analysis of PR-binding sites, transcription factors such as HOXA11, COUP-TFII, and WT1 were presumed as potential PR cooperators in the uterus. It was noteworthy that many of the cooperative factors we found were expressed in uterine stromal cells and had been documented to be essential for uterine receptivity, further confirming the efficiency of our model. Transcription factor WT1 recognizes and binds DNA through its four C2H2 zinc figures at the C-terminus to activate or repress gene expression ( 14 , 15 ). It has been reported that WT1 is extensively expressed in the murine urogenital system during embryogenesis, including the uterine wall ( 28 ). In the human endometrium, the localization of WT1 is restricted to the nuclei of stromal cells. Its expression increases during decidualization ( 29 ). In fact, the involvement of WT1 in endometrial stromal cell decidualization has been well documented. During the in vitro decidualization process in HESCs, WT1 is required to trigger the expression of insulin-like growth factor binding protein 1 (IGFBP1) and prolactin (PRL), two marker genes of decidualization ( 17 ). Meanwhile, WT1 up-regulates glucose transporter 1 (GLUT1) and very low-density lipoprotein receptor (VLDLR), which are responsible for glucose uptake and lipid accumulation, respectively, essential for proper decidualization ( 16 , 18 ). However, the physiological significance of WT1 in uterine receptivity and embryo implantation and its relation to the PR signaling remains ambiguous. In our study, we showed that WT1 was specifically expressed in the mouse uterine stroma, implying its conserved expression pattern in the uterus of different species. Furthermore, using a mouse model with uterine-selective depletion of Wt1 , we not only demonstrated that WT1 was essential for the establishment of uterine receptivity, but also revealed a role of WT1 as a PR cooperator facilitating its chromatin binding. Whether these functions of WT1 are conserved in the human endometrium warrants further exploration. We observed that 20% Wt1 d/d females showed signs of embryo implantation. It is unclear why embryo implantation still took place under such P4-resistant conditions. One explanation might be the individual differences in P4 responsiveness among Wt1 d/d females. Despite the occurrence of embryo implantation in the absence of uterine WT1, the average number of IS was significantly fewer, and aberrant attachment reaction was observed. Moreover, the embryos showed retarded development on D6 and were ultimately absorbed by D8. These results further reinforce the widely accepted notion that embryo implantation sets up subsequent developmental programming, and perturbations during this event exert ripple effects on pregnancy outcomes ( 30 – 32 ). Our findings revealed that WT1 mediated PR–chromatin binding and transcriptional activity in uterine stromal cells. Some earlier studies have reported several PR cofactors in the uterine stroma. It has been demonstrated that the transcriptional activity of PR is facilitated by the coordinate interactions with the steroid receptor coactivators (SRCs) ( 33 ). For example, SRC2, a member of the SRC family, is expressed in uterine stromal cells on D3-D4, and the loss of SRC2 results in severely compromised uterine functions, as well as failure in embryo implantation and decidualization due to hampered P4 responsiveness ( 34 , 35 ). Moreover, our previous study has shown that BMI1, a key component of polycomb repressive complex 1 (PRC1), localized in both epithelial and stromal cells of the receptive uterus, interacts with PR and the E3 ligase E6AP to regulate PR ubiquitination that is essential for its transcriptional activity and P4 responsiveness ( 36 ). Whether WT1 is associated with these PR coactivators remains to be verified. According to our current knowledge of PR actions, the P4-bound PR binds to the chromatin and then recruits coactivators like SRCs to regulate the transcription of target genes. Since previous studies have not compared the genome-wide chromatin occupancy of PR with/without the presence of these PR cofactors, it is difficult to determine whether they are required for PR–chromatin binding. Our study not only unraveled the co-occupancy of WT1 and PR on the chromatin but also delineated PR–chromatin binding status with/without WT1, which provided solid evidence for the indispensable function of WT1 in the recruitment of PR onto specific target gene loci. At the meantime, a previous study has performed an unbiased label-free endogenous immunoprecipitation followed by mass spectrometry in order to obtain the global proteome interaction map of WT1. WT1 has been shown to physically interact with many important transcription factors including p53 and ERα, as well as transcription coactivators and corepressors such as CBP/p300 ( 37 – 39 ). The potential involvement of these factors in the WT1-PR transcription complex needs further investigations. Combining RNA-seq and ChIP-seq analyses, we found that Nr2f2 was directly regulated by the PR-WT1 transcription complex, and its expression was aberrantly decreased in the Wt1 d/d uterine stromal cells on D4. COUP-TFII, encoded by Nr2f2 , is highly expressed in uterine stromal cells, and participates in the establishment of uterine receptivity ( 20 , 40 ). It has been reported that the expression of COUP-TFII is regulated by the PR-Indian hedgehog (IHH) signaling that emanates from the uterine epithelium ( 41 , 42 ). Here, we revealed a regulatory mechanism underlying COUP-TFII expression that could be directly induced by stromal WT1 and PR. Meanwhile, uterine COUP-TFII deficiency results in decreased PR level in stromal cells ( 20 ), implying a complex regulatory loop between PR and COUP-TFII. In the present study, we noticed that some previously well-accepted P4 responsive genes were actually not the direct targets of PR. HAND2 was one such example. Although it has been demonstrated that the expression of HAND2 in the uterus is up-regulated by P4 and repressed by the PR antagonist RU-486 ( 43 ), we did not detect PR peaks on its gene loci. However, it cannot be excluded that PR controls HAND2 transcription by binding to distal regulatory elements. In summary, our study provides a more comprehensive insight into the PR actions in the uterus during the receptive phase and identifies WT1 as a PR cooperator mediating its chromatin binding and transcriptional activity. These findings possess important clinical significance for the understanding of female endometrial disorders associated with P4 resistance.

Appendix 01 (PDF)