All mice were kept under controlled photoperiod conditions (lights on 07:00–19:00 h) and were supplied with food and sterilized H₂O ad libitum. Experiments were conformed to the regulations drafted by the Association for Assessment and Accreditation of Laboratory Animal Care in Shanghai. The ethics committee approval was obtained from the Institutional Ethics Committee of Shanghai Institute of Planned Parenthood Research Center (now Shanghai Institute for Biomedical and Pharmaceutical Technologies) for the commencement of the study.

The C57BL/6 Lgr4 gene-trap mouse strain was generated as previously described (Weng et al., 2008). Genomic DNA was extracted from tail biopsies and analyzed using the TaKaRa Taq^TM Hot version (Cat# R007A). Primers for genotyping are listed in Supplementary Table 1. The 23- to 25-day-old females were sacrificed to perform a series of in vivo analyses.

Oviducts were collected and fixed in Bouin’s solution for hematoxylin and eosin (H&E) staining and in 4% paraformaldehyde (PFA) for immunohistochemistry (IHC). For IHC assay, sections (4–5 μm) were deparaffinized in xylene and rehydrated in gradient alcohol concentrations. After antigen retrieval, the sections were blocked by 5% bovine serum albumin (BSA). The sections were then incubated with primary antibodies overnight at 4°C. The next day, the sections were incubated with secondary antibodies for 20 min and then developed with 3,3′-diaminobenzidine and counterstained with hematoxylin. Antibodies were diluted as follows: OVGP1 at 1:100 (Santa Cruz, sc-377267), PAX8 at 1:100 (Proteintech Cat# 60145-4-lg), CTNNB1 at 1:100 (BD Transduction Laboratories^TM, Cat# 610153), NR5A1 at 1:200 (Abcam Cat# ab217317), and NR5A2 at 1:500 (Abcam Cat# ab189876). Samples from at least three animals were evaluated to ensure reproducibility of the results.

Deparaffinized and hydrated slides were incubated in 3% acetic acid for 3 min, followed by staining in 1% Alcian blue solution (pH 2.5) for 30–60 min. The sections were washed in running tap water for 2 min and rinsed in diH₂O. After dehydration, tissue sections were mounted and observed under a microscope.

Deparaffinized and hydrated slides were incubated in 0.5% periodic acid solution for 5 min. After rinsing in three changes of distilled water, sections were placed in Schiff’s reagent for 15 min, followed by washing in tap water for 5 min. After dehydration, the sections were mounted and observed under a microscope.

The frozen sections were fixed in 0.2% glutaraldehyde in phosphate-buffered saline (PBS) for 15 min on ice then washed in PBS for 10 min. Then the sections were immersed in 1 mg/ml X-gal staining solution overnight at 37°C in the dark. A post-fixation in 4% PFA for 10 min was required. After rinsing in PBS, the sections were counterstained with nuclear fast red then rinsed in distilled water for 2 min. After dehydration, sections were mounted and observed under a microscope.

Briefly, the mouse oviduct was surgically isolated and placed into an enzymatic media [Dulbecco’s minimum essential medium DMEM/F12 supplemented with 0.25% collagenase type IV (Invitrogen, 17104019), and 10 U/ml DNase I (Sigma, 10104159001)] for 15 min at 37°C. The softened tissue was then washed two times with PBS and transferred into a dissection medium [PBS with 0.5% penicillin/streptomycin and 1% fetal bovine serum (FBS)]. The oviduct tissue was then mechanically dissected using micro-scissors. The oviductal cell suspension was diluted with an equal volume of medium DMEM/F12 supplemented with 5% fetal bovine and centrifuged twice at 300 × g for 5 min at 4°C. The supernatant was removed, and the pellet of oviductal cells was re-suspended in culture media (DMEM/F12 supplemented with 10% FBS, 1% insulin, transferrin, and selenium solution) and seeded in a 24-well plate at 37°C and 5% CO₂ in a humidified incubator. The culture medium was changed every 2 days. Once the adhered cells reached 70% confluence, we conducted the co-culture assay or E2 stimulated ELSA assays (OVGP1 kit: MyBioSource, Cat# abx576628, San Diego, CA, United States; CXCL8/IL-8 kit: R&D, Cat# D8000C). Before the E2 stimulation, cells were cultured in a medium without phenol red and FBS for at least 24 h.

On the day of oocyte collection, the medium of the epithelial cell culture was changed from the DMEM/F12 to M2 medium (Sigma, M7167). Germinal vesicle stage oocytes (GV oocytes) were collected from 6-week-old C57BL/6 female mice injected with 5 IU of pregnant mare’s serum gonadotropin (PMSG; Sigma) 40 h before. Then the cumulus–oocyte complexes were obtained by puncturing the antral follicles, denuded by repeat pipetting using a stripper in 0.3 mg/ml hyaluronidase (Sigma H4272), and classified by GV stage. GV oocytes were co-cultured with epithelial cells in M2 media. Oocytes were observed at 12 and 24 h for oocyte maturation. Grading of oocytes was performed by microscopic examination. Progression to the metaphase II (MII) stage was identified by extrusion of the first polar body into the perivitelline space, which was observed under a microscope.

Mouse oviductal epithelial cells (MOECs) were treated with or without proteins WNT4, RSPONDIN2, or compound CT99021 overnight, and chromatin immunoprecipitation (ChIP) analysis was performed as previously described (Wu et al., 2007). DNA was precipitated with a polyclonal NR5A2 (Abcam Cat# ab189876) antibody or mouse IgG as a negative control. Precipitated DNA was amplified with the primers designed to the Cyp11a1 and Hsd3b1 promoter regions. See the detailed information of the primers in Supplementary Table 1. PCR products were confirmed by sequencing.

Gene expression levels were estimated using fragments per kilobase of exon per million fragments mapped (FPKM) by StringTie v1.3.4d (Pertea et al., 2015). EdgeR v3.24.2, a R package, was used to measure differential gene expression (Robinson et al., 2010). The false discovery rate (FDR) control method was used to calculate the adjusted p-values in multiple testing to evaluate the significance of the differences. Genes with an adjusted p-value < 0.001, |log2FC| ≥ 2, and FPKM > 1 were used for subsequent analysis.

A gene annotation file was retrieved from ensembl genome browser 96 database.¹ To annotate genes with Gene Ontology (GO), clusterProfiler v3.4.4 (R package) was used.

The raw next-generation sequencing (NGS) data were deposited to the NCBI SRA database under project number (PRJNA691984). Reverse transcription polymerase chain reaction (RT-PCR) was performed to verify the mRNA-seq results using SYBR Premier EX Taq (TaKaRa). Genes were amplified with the indicated primers (Supplementary Table 1). Relative levels of mRNAs were calculated using MX3500pro software and normalized to the levels of endogenous actin in the same samples.

Statistical data were analyzed with GraphPad Prism version 5. The statistical data of oviduct weight, numbers of specific stained secretory cells in the oviduct epithelium, and percentages of different stages of oocyte and RT-PCRs were presented as means ± SEM. Analysis of variance or Student’s t-test was used for statistical comparison to determine significance. Statistical significance was set as NS, p > 0.05; ^∗p ≤ 0.05; ^∗∗p ≤ 0.01; and ^∗∗∗p ≤ 0.001. All presented results were from at least three independent experiments. The investigators were blinded to the group allocation during the experiment and when assessing the outcome.

To study the role of Lgr4 in the oviduct, we analyzed its gene expression pattern. The mRNA level of Lgr4 was significantly enhanced by 3.52-fold in the distal region compared with the proximal region of the mouse oviduct (Figure 1A). As LGR4, LGR5, and LGR6 are the three most closely related leucine-rich repeat family of G-protein-coupled receptors (de Lau et al., 2012), we also determined the mRNA levels of Lgr5 and Lgr6. Unlike Lgr4, the expressions of Lgr5 and Lgr6 showed no differences between the distal and proximal regions in the mouse oviduct (Supplementary Figures 1A,B). To further identify the distribution of LGR4 in the oviduct, we took advantage of the β-gal reporter gene, which was inserted into the first intron of the Lgr4 genomic locus to track endogenous Lgr4 mRNA expression (Weng et al., 2008). The β-gal signal was enhanced strongly in the distal region (dark blue) compared with the slightly stained proximal region (Figure 1B). From the tissue morphology, the β-gal positive cells were oviductal epithelial cells. Although a positive signal could be detected in the epithelial layer of all three regions (infundibulum, ampulla, and isthmus), it was clear that the signal in the ampulla was the strongest compared with that in the other two regions (infundibulum and isthmus) (Figure 1C). Under specific hormonal controls during different stages of the estrus cycle, the oviduct produces different levels of fluid and macromolecules that optimize the microenvironment for gamete maturation. Therefore, we examined mRNA levels of Lgr4 expression. RT-PCR assay revealed that the Lgr4 expression was enhanced by more than 2.85-fold at the pro-estrus stage and then reached its peak level at 4.35-fold at the estrus stage compared with the di-estrus stage. This result suggested the expression of Lgr4 fluctuated dynamically during the estrus cycle (Figure 1D). The data showed that Lgr4 was present in the oviductal epithelium and with a significantly enhanced expression level at the distal region, especially in the ampulla, suggesting a physiological function of Lgr4 in this specific region within the estrus cycle.

To determine the physiological role of the LGR4 protein in the murine oviduct, we compared the mutant and WT oviduct. The oviducts of KO female mice were significantly reduced in size (Figure 2A). In addition, the mean weight of isolated oviducts was 0.72 mg in the mutant mice, which was significantly lower than that in the WT group (1.32 mg) (Figure 2B). H&E staining was performed to determine the histological changes between the Lgr4 KO and WT oviducts. The cross-sectional area of the Lgr4 KO ampulla region was reduced in comparison with the WT control (Figure 2C, upper panel). The average number of the leaf-like folds in the WT ampulla was 21.00, while it was 12.60 in the mutant ampulla (Figure 2D). This result suggested that the loss of Lgr4 led to a reduced oviductal leaf-like folds. Furthermore, the columnar secretory cells with spot-like granules but no cilia were stained in dark blue in the WT (arrowed in bright green), while such type cells with less spot-like granules stained were reduced in the Lgr4-deficient oviduct (arrowed in red) (Figure 2C, lower panel). In the WT leaf-like fold, the average number of these secretory cells was 14.40, while in the mutant leaf-like fold, the number was reduced significantly to 5.60 (p < 0.001) (Figure 2E). As an ovarian dysfunction resulted from Lgr4 ablation (Pan et al., 2014), we removed the ovaries from both the WT and KO mice, and the ovariectomized female mice were then treated with daily injections of 100 ng per mouse of 17-β-estradiol or 1 mg per mouse of progesterone dissolved in 100 μl of sesame oil according to a previous report (Xiao et al., 2002). Histological defects with reduced numbers of leaf-like folds and secretory cells were still observed in the mutant ampulla (Supplementary Figures 2A–C). Together, these results suggested that the Lgr4 gene has an important role in the secretory cells in mouse oviduct.

One fundamental role of the oviduct is to secrete factors to support and promote oocyte maturation. We used a two-type cell co-culture assay as outlined in the methods to detect whether ablation of Lgr4 led to a deficiency in oviduct secretion (Figure 3A). Before the examination, we measured the cytokeratin (cytokeratin-18) and alpha-tubulin (Ac-Tub) levels in MOEC cultures. One day after seeding, free-floating columnar MOECs with apparent cilia on the external surface (left) and adhered MOECs with reduced cilia (right) were observed (Figure 3B, upper panel). Two days later, the main cell type in the well was the adhered MOECs with (left) or without reduced cilia (right) (Figure 3B, lower panel). Three days later, almost all the adherent MOECs lost their cilia. PAX8 is reported to be a typical secretory cell marker in the oviduct (Ghosh et al., 2017; Ito et al., 2020). Most of the adhered cells at this time expressed PAX8 (Figure 3C). The data suggested that the cultured oviductal epithelial cells retained their secretory cell characteristic without cilial cells. As the co-culture system is widely used to promote the maturation of oocytes (Lee et al., 2011, 2017; Dadashpour Davachi et al., 2016), it can also show the secretory function of the oviduct epithelial cell in promoting oocyte maturation. When both the WT and KO adhered MOECs reached 70% confluence, we co-cultured the MOECs with the mouse GV oocytes for 12–24 h. The percentage of the GV oocytes that developed to the MII stage (MII oocyte) in the MII medium only was 61.00%. When the GV oocytes were co-cultured with WT mouse oviduct epithelial cells, the percentage enhanced significantly to 76.26%. However, when the GV oocytes were co-cultured with the Lgr4 mutant MOECs, the percentage upregulated slightly but not significantly to 64.57% compared with the MII medium only group (Figure 3D). The oviduct epithelium is covered with a mucosal surface containing mucins (MUCs) at its apical site, which provides a physical barrier against invading pathogens. After pathogen recognition through Toll-like receptors, epithelial cells release several chemokines, e.g., interleukin 8 (IL-8), to attract the immune system (Lagow et al., 1999; Zerbe et al., 2003; Chaturvedi et al., 2008; Kowsar et al., 2013). Oviduct-specific glycoprotein 1 (OVGP1), also named MUC9, is one of the most important glycoproteins secreted directly by the oviduct epithelial cells (Aviles et al., 2010; Saint-Dizier et al., 2014; Algarra et al., 2016; Lamy et al., 2018; Zhao and Kan, 2019). Therefore, we examined the protein levels of released OVGP1 and IL-8 in the WT and KO cultured MOEC supernatants. The oviduct is highly sensitive to the fluctuating ovarian hormone levels, so we supplemented the cultured MOECs with 17-β-estradiol (E2) mimicking the estrus effect, as the E2 influences the secretion of MOECs positively (Chen et al., 2013, 2018). The average OVGP1 level of the WT MOEC supernatant was 60.88 ng/ml, while this protein level was induced significantly to 100.89 ng/ml at the E2-diestrus stage (E2 = 10.00 pg/ml). The level was further enhanced to 162.45 ng/ml in the WT MOEC supernatant at the E2-estrus stage (E2 = 50.00 pg/ml). However, the average OVGP1 levels of the KO MOEC supernatant were 24.27, 43.51, and 54.30 pg/ml at each specific stage without any significant changes (Figure 3E). Similarly, the average IL-8 levels were 33.70, 71.45, and 101.22 ng/ml specifically, which were enhanced significantly under the stimulation of different concentrations of E2 in the WT MOEC supernatant. The average IL-8 levels of the KO MOEC supernatant were 9.65, 12.32, and 23.82 pg/ml at each specific stage without significant changes (Figure 3F).

To further examine the role of Lgr4 in the secretion function of the mouse oviduct in vivo, we performed a set of histochemistry assays on the WT and Lgr4 KO oviduct in the estrus stage. The periodic acid–Schiff (PAS) staining was utilized to detect the presence of carbohydrates and carbohydrate compounds such as glycogen in the oviduct. The specific staining was observed in the WT group (arrowed in black), while in the Lgr4-deficient group only a low number of the specific stainings were detected (Figure 3G). In the Lgr4 KO group, the average number of the positive PAS staining signal was 2.00 per leaf-like fold, which was significantly lower than the wild type counterpart, 10.16 (Supplementary Figure 3A). A similar staining pattern was also found in the Alcian blue assay which was used to detect the acid mucosubstances and MUCs. Specifically, apparent signals were observed in the WT group (arrowed in black), while in the deficient group, the signals were reduced severely (Figure 3H). In the Lgr4 KO group, the average number of the positive Alcian blue staining signal was 2.33 per leaf-like fold, which was significantly lower than that in the WT, 23.00 (Supplementary Figure 3B). Together, all the in vitro and in vivo assays suggested that the secretion function of the MOECs was impaired by a deletion of the Lgr4 gene.

To reduce the influence of the estrus cycle on the transcriptom and to study the effect of Lgr4 deficiency on the transcriptom, we analyzed the wild type and mutant oviducts in the estrus stage. The RNA-seq revealed a downregulation of 2,597 genes and an upregulation of 762 genes with p_adj value < 0.001, |log₂fold change| > 2, and FPKM > 1 (Figure 4A and Supplementary Table 2). GO highlighted that differentially expressed genes (DEGs) were mainly involved in the regulation of epithelial tube morphogenesis; biosynthetic and secretory process, like sulfur amino acid, glycoprotein, and hormones; WNT signaling pathway; DNA-templated transcription, and others (Figure 4B). A set of secreted factor genes was checked between the WT and KO oviducts also at the estrus stage (Figures 4C,D). The mRNA level of Hspa4, the heat-shock protein family A (HSP70) member 4 (HSPA4), which is important for the sperm fertilizing ability, was downregulated significantly by 5.20-fold in the Lgr4 mutant oviduct compared with the WT oviduct. The mRNA levels of disintegrin and metalloproteinase domain 9 (Adam9), and CD46 molecule (CD46), which were reported to have functions in the fertilization process, respectively (Alminana et al., 2017), were both significantly reduced by 2.42- and 2.64-fold. Another downregulated gene was Vegfa. VEGFA is a critical factor in the oviduct throughout the animal estrous cycle regulating the dynamics of oviduct fluid secretion and tubal contractibility. The mRNA of Vegfa was also detected to be significantly attenuated by 6.60-fold in the Lgr4 mutant oviduct. Myosin-9, also known as MYH9, has been identified on human sperm as a binding partner of OVGP1 (Kadam et al., 2006). The mRNA level of Myh9 was decreased significantly by 3.27-fold in the Lgr4 mutant oviduct compared with the WT oviduct. Accordantly, we checked the protein level of OVGP1, observing that this protein was reduced significantly in the oviducts of Lgr4 KO mice compared with the WT (Figure 4D). The IHC assay on the PAX8 protein showed specific 3,3′-diaminobenzidine (DAB) signals in the WT oviductal epithelial nuclei, while no apparent positive signals were observed in the KO mouse epithelial nuclei (Figure 4E, arrowed in black) at the stage of estrus. Collectively, the data demonstrated that Lgr4 deficiency led to a hypofunction of oviducts via downregulating of genes linked to secretory processes.

Hormone synthesis is predominantly controlled by members of the cytochrome P-450 (CYP) enzyme family and hydroxysteroid dehydrogenases (HSDs) (Chatuphonprasert et al., 2018). CYP11A1 and 3β-HSD are responsible for the formation of almost all the steroid hormones (Luu-The, 2013). Since we observed deficits in the secretory function of the oviducts of the Lgr4 KO mice, we evaluated mRNA levels of the Cyp11a1 and 3β-Hsd1 genes. The relative mRNA levels of the two genes were downregulated significantly (Figure 5A and Supplementary Table 2) at the stage of estrus. To determine how LGR4 regulated these steroidogenesis genes, we identified the WNT response elements in the Cyp11a1 and 3β-Hsd1 gene promoters. By the method of in silico analysis,² we found that the two genes had two putative highly conserved NR5A1/2 binding sites in the proximal promoter regions (Figure 5B). Although the NR5A1 and NR5A2 share the same DNA binding sequences, these two transcription factors work in a non-overlapping way. These might suggest that NR5A1 and NR5A2 are presented in different systems. To demonstrate, we checked the presence of NR5A1 and NR5A2 in oviducts, respectively. By the IHC assay, we detected the apparent presence of NR5A2 but not NR5A1 in the WT mouse oviductal epithelium. Note that the corresponding signal was reduced severely in the mutant oviductal epithelial nuclei (Figure 5C, arrowed in black and Supplementary Figures 4A,B). The mis-location of NR5A2 protein observed in nuclei implied an impairment of the DNA binding ability with the transcriptional factor NR5A2.

LGR4 has been widely considered as a WNT effector amplifier. To study if LGR4 has a similar function in our model, we compared the protein level of CTNNB1 between the WT and KO mice. The accumulation of CTNNB1 in the cytoplasm of cells was detectable in oviduct epithelial cells of WT mice at the stage of estrus. In contrast, the accumulation of CTNNB1 in Lgr4 KO oviductal epithelial cells was significantly downregulated (Figure 5D, arrowed in red). In addition, a set of WNT/CTNNB1 signaling target genes like Enc1 (Fujita et al., 2001), Bcl2l2 (Lapham et al., 2009), Yap1 (Konsavage et al., 2012), and Axin2 (Yan et al., 2001; Jho et al., 2002; Lustig et al., 2002) were also downregulated significantly in the Lgr4 KO oviducts compared with the WT control at the estrus stage (Figure 5E). These data further confirmed the regulatory function of LGR4 on the canonical WNT signaling pathway in the oviduct epithelium.

We hypothesized that the activated CTNNB1 might increase NR5A2 binding to the Cyp11a1 and 3β-Hsd1 gene promoters in the mouse oviduct. To this end, we performed the ChIP-qPCR assay utilizing primary culture cells of the WT and Lgr4 mutant mouse oviducts. In the WT, the capacity of the transcription factor NR5A2 to bind with Cyp11a1 and 3β-Hsd1 promoter regions was improved by ∼3.26- and ∼2.25-fold, respectively, in response to the stimulation of factors WNT4/RSPO2. Such improved binding capacity of NR5A2 was not detectable in cultured cells of Lgr4 mutant mice. However, CT99021, a commercial GSK-3α/β inhibitor that activates the canonical WNT pathway, enhanced the DNA binding abilities in WT and mutant primary oviductal epithelial cells. Specifically, in the WT group, the capacity of DNA binding to the Cyp11a1 and 3β-Hsd1 promoters of the transcription factor NR5A2 was increased by ∼3.02- and ∼1.97-fold, respectively, while in the mutant group, such ability of NR5A2 to the Cyp11a1 and 3β-Hsd1 promoters was rescued by ∼5.11- and ∼5.54-fold, respectively (Figure 5F, upper panel). The relative mRNA fold change of the Cyp11a1 and 3β-Hsd1 genes after WNT4/RSPO2 or CT99021 application in the WT or Lgr4 mutant oviductal epithelial cells further suggested the activity of the transcription factor NR5A2 was regulated by the canonical WNT cascade (Figure 5F, lower panel).