Loss of Function of the Zxdb Gene Leads to a Decrease in the Decidualization Rate and Number of Pups Born in Mice by Affecting the Expression of the Cell Adhesion Molecules

TL;DR

This study shows that the Zxdb gene in mice is important for female fertility, as its loss reduces embryo implantation and litter size.

Contribution

The study is the first to link Zxdb gene loss to reduced female fertility in mice through disrupted cell adhesion.

Findings

Loss of Zxdb in mice reduces decidualization rates and litter size.

Zxdb loss affects adhesion molecule expression in uterine tissue.

Disordered adhesion molecules may cause lower embryo implantation rates.

Abstract

The Zinc Finger X-Linked Duplicate B (ZXDB) gene is one of a pair of replicated zinc finger genes on chromosome Xp11.21. The homologous gene of ZXDB in mice is Zxdb. Recent studies have found that Zxdb plays a role in the spermatogenic process of mice; however, its impact on the female reproductive system has not yet been explored. In our study, we found, for the first time, that the loss of function of Zxdb leads to reduced decidualization rates and a decrease in litter size in female mice. Secondly, we found that maternal loss of Zxdb is the determinant of these phenotypes. Thirdly, the transcriptional and proteomic differential expression genes in the uterine tissues of wild-type (WT) and Zxdb knockout (Zxdb-KO) mice were significantly enriched in signaling pathways such as adhesion molecules. Finally, we demonstrated that the disorder of expression and uneven distribution of adhesion molecules in mouse uterine tissue may be the main reason for the decline in embryo implantation rate. In conclusion, we have established for the first time a link between the Zxdb gene and reduced female fertility. This study will help provide guidance and genetic counseling for future common clinical complications such as Recurrent Spontaneous Abortion (RSA) or Recurrent Implantation Failure (RIF).

Figures

Full text

1. Introduction

Embryo implantation is an early pregnancy event and a pivotal step in the mammalian reproductive process [1,2]. It refers to the process by which the blastocyst and the mother recognize each other, locate each other in the uterus, and adhere to and invade specific parts of the mother’s uterus [3,4]. It is estimated that 50% of pregnancies result in miscarriage due to biochemical loss or implantation failure in the preclinical stage [5,6], and another 20% are clinically confirmed as pregnancy failure, mainly occurring in the first three months of pregnancy [7]. Successful implantation needs to occur during the WOI (window of implantation), so that the normally developing blastocyst can interact with the endometrium in a receptive state for a short period of time [8]. At this time, endometrial epithelial cells express specific adhesion molecules to provide support for embryo attachment. If this WOI is missed, the embryo will not be able to implant.

Decidualization occurs during the secretory stage of the menstrual cycle and is controlled by progesterone and other ovarian hormones [9]. It is a prerequisite for successful embryo implantation [10,11] and results from complex interactions between hormones, transcription factors, morphogenetic factors, cytokines, cell cycle regulators, and signaling pathways [12]. In mice, there is a slight difference from humans; the decidualization process begins only after the embryo enters the uterus [9,13]. Decidualized cells change in cell morphology from slender fibroblast-like endometrial stromal cells to round or polygonal decidua cells. Decidua cells produce a large number of secreted proteins, including some marker proteins [14,15,16]. Among them, CDH1 (E-cadherin) and CDH2 (N-cadherin) are members of the cadherin family and play a crucial role in cell adhesion and signal transduction, especially during epithelial–mesenchymal transition (EMT) and mesenchymal–epithelial transition (MET) [15,17]. The ratio of CDH1/CDH2 is often used as a marker for MET/EMT, and the dynamic balance of MET/EMT is crucial for the success of pregnancy. Abnormal ratios may lead to implantation failure and recurrent miscarriage [18,19]. The localization and subsequent decidualization of endometrial vascular permeability responses suggest that they occur in response to interactions between the embryo and the endometrium [20,21]. Subsequently, the embryos begin to invade the endometrial tissue and are successfully implanted.

The ZXDB gene is one of a pair of replicated zinc finger genes on chromosome Xp11.21, the other being ZXDA. ZXDB has only one exon, with a GC content of approximately 70%, and encodes 803 amino acids. It is highly conserved in many species [22]. The ZXDB refers to the human gene; Zxdb is the murine ortholog. The predicted ZXDB protein contains 10 zinc finger tandem motifs that bind to DNA, and the protein’s 576-703 amino acid sequence also has a transcriptional activation domain for gene expression. RT-PCR and transcriptome sequencing detected the expression of ZXDB in most human tissues, suggesting that it may play an important fundamental role in the development of organisms. Previous studies have found that ZXDB can affect the expression of MHCI and MHCII molecules [23,24].

In the mouse genome, there is a frameshift mutation in the Zxda gene, so the Zxdb gene is the only functional mouse homolog. A recent study has found that the FBXO38-dependent degradation of ZXDB is a key factor for achieving chromatin dynamic regulation, especially the centromere protein CENP-B. The ZXDA/B protein binds to the centromere protein CENP-B. Functionally, the ZXDA/B factor controls the level of CENP-B proteins associated with the chromatin. In testes, the Fbxo38 gene is specifically expressed in Sertoli cells, while Sertoli cells lacking Fbxo38 exhibit stable ZXDB protein and increased chromatin in the centromere region. After KO of the Fbxo38 gene, the mice grow slowly and affect multiple organs, including the male reproductive system. Detailed analysis of the mutant testes revealed pathological changes in the spermatogenic tubules, along with a significant reduction in sperm production and decreased fertility [25].

Research in human cell lines or samples is currently very limited. Previous studies regarded ZXDB and ZXDA as similar proteins for investigation. Currently, there is insufficient research specifically targeting the ZXDB gene, and its important role as a transcription factor, especially in the reproductive system, remains to be explored.

Here we have, for the first time, discovered that knockdown or knockout of the Zxdb gene can lead to decidualization and a decrease in the litter size. Sequencing analysis of pre-implantation uterine tissues, conducted through transcriptomics and proteomics, revealed that significantly enriched pathways included those related to adhesion molecules, immune responses, and other relevant pathways. WB experiments confirmed that the expression of adhesion molecules such as CDH1(E-cadherin), CDH2(N-cadherin), and BMP2 was disrupted. Immunofluorescence experiments revealed abnormal expression and distribution of CDH1 in the uterus, resulting in a decreased implantation rate. This study is conducive to deepening the analysis of RSA factors and genetic counseling by scientific researchers and clinicians.

2. Materials and Methods

2.1. Cell Line and Cell Culture

The Neuro-2a(N2a) cell line was purchased from the Cell Bank of the Chinese Academy of Sciences. Cell culture was performed using DMEM (GIBCO, 10566016, Waltham, MA, USA) medium supplemented with 10% high-quality fetal bovine serum and 1% penicillin–streptomycin solution (Beyotime, C0222, Shanghai, China). Cells were cultured in an incubator at 5% CO_2_ and 37 °C. The siRNA transfection process was carried out according to the instructions provided in the Lipofectamine RNAi MAX kit (Invitrogen, 13778075, Thermo Fisher Scientific, Waltham, MA, USA). siRNA (Zxdb) sense sequence: 5′-GAGCUUCACCACCGUCUACAA-3′; siRNA (Zxdb) antisense sequence: 5′-GUAGACGGUGGUGAAGCUCUU-3′. The use of 2′-O-methyl (2′-OMe) modification for siRNA is to enhance its stability. All siRNAs were ordered from Shanghai GENEray Bioengineering Co., LTD. (Shanghai, China).

2.2. Mouse Breeding and Ethics

Both C57BL/6J and ICR mice involved in this laboratory were purchased from Shanghai JieSiJie Laboratory Animal Co., Ltd. (Shanghai, China). The female and male ICR mice were 6 to 8 weeks old, the C57BL/6J female mice were 4 to 6 weeks old, and the C57BL/6J male mice were over 8 weeks old. All the mice in this experiment were of SPF grade. The experimental process strictly adhered to the regulations of the Animal Ethics Committee of the School of Life Sciences, Fudan University. The content of the animal experiments has obtained ethical approval, with the approval number: No. 2022JS071.

2.3. Embryo Acquisition and Embryo Culture

In this study, fertilized eggs were obtained through in vivo/in vitro fertilization. Each female mouse was injected with 5 to 10 IU of PMSG (Nanjing Aibei Biotechnology Co., Ltd., M2620, Nanjing, China). After 48 h, 5 to 10 IU of HCG (Nanjing Aibei Biotechnology Co., Ltd., M2530, Nanjing, China) was injected. The mice were mated in one cage, and the vaginal plugs were checked the next morning. For mice with vaginal plugs, 14 h after HCG injection, they were sacrificed. The fallopian tubes at the junction of the uterus and ovary were removed through an opening on the back. Under a stereomicroscope, oocytes were drawn out in M2 buffer (Nanjing Aibei Biotechnology Co., Ltd., M1250, Nanjing, China) with a 1 mL syringe, and the zona pellucida was removed using hyaluronidase (Nanjing Aibei Biotechnology Co., Ltd., M2215, Nanjing, China). Transfer clean eggs or embryos to KSOM medium (MERCK, MR-101, Darmstadt, Germany). The culture medium needs to be prepared in advance and balanced overnight. The morphological observation of the embryo is carried out according to the changes in the division time. Generally, it is checked once a day until it reaches the blastocyst stage.

2.4. siRNA Injection Embryo and Uterine Horn Interference Experiment

siRNA embryo injection was carried out by microinjection. When operating, first place the obtained Cumulus Oocyte Complexes (COCs) in hyaluronidase (Nanjing Aibei Biotechnology Co., Ltd., M2215, Nanjing, China) for digestion to remove the outer granular cells. Embryos with granulosa cells removed were washed and transferred to KSOM medium for further culture. Microinjection was performed when the embryos were cultured to the prokaryotic stage. Subsequently, a 20 µM siRNA solution that had been pre-centrifuged was directly injected into the cytoplasm or prokaryote of the fertilized egg using a microscopic operating system and a glass capillary needle. The injection volume was approximately 2 to 5 pl per fertilized egg. After the injection was completed, the embryos were immediately transferred into KSOM culture medium and cultured at 37 °C and 5% CO_2_. For the RNA interference experiments on uterine tissue, the abdominal incision exposed the uterine horn after the female mice were anesthetized. A fine needle was used to directly inject the siRNA solution into the uterine cavity, avoiding penetration of the uterine wall. After injection, the abdominal cavity was sutured, and samples were collected 7~8 days after recovery.

2.5. Construction of the Zxdb-KO Mouse Model

The Zxdb-KO mouse model was constructed by our research team using a dual sgRNA protocol. Guide RNA sequence1: 5′-ACGCCGAGAUGGAAAUCCCG-3′; Guide RNA sequence2: 5′-GGGACCGCAGACAGACAGCG-3′; sgRNAs were prepared by in vitro transcription using the T7 High Yield RNA Transcription kit (Novoprotein, E131, Suzhou, China). Then, superovulation was performed on the female mice, and they mated with the male mice to obtain fertilized eggs. The Ribonucleoprotein complex (200 ng/μLCas9 protein + 100 ng/μL sgRNA) was injected into the prokaryote or cytoplasm of the fertilized egg by microinjection. After injection, the embryo was transplanted into the uterus of the surrogate mother mouse. After the birth of the pups, genotypes were identified through PCR, sequencing, and other methods to screen for positive F0 generation (Founder mouse). Finally, the positive mice were expanded, and a stable genetic model was obtained.

2.6. Mouse Genotype Identification

Seven days after the birth of the mice, their nails were cut to extract genomic DNA for PCR genotype identification. The first genotype identification experiment identified the specific genotype of the mutant through PCR. There is a 528 bp deletion between sgRNA1 and sgRNA2 target sites. Subsequently, primers were designed across breakpoints, and the genotype was directly determined based on the size of the electrophoresis bands. Primer 1F: 5′-GCCTCACCGACGCCGAGATGGAAATCTGTCT-3′; Primer 2F: 5′-CTTGACCCGGTGGGTGGCGATGTAGAAACC-3′. Universal primers Primer1 and 2R: 5′-GTGGTCAGGGTGAGGACGCCGTTCTCGAA-3′. The breakpoint product has a KO band of 187 bp and a WT band size of 308 bp. Both the homozygous and WT mice have only one band, while the heterozygous has two bands.

2.7. Transcriptome and Proteomics Sequencing Analysis

Uterine tissue of WT and Zxdb-KO mice was collected with vaginal plugs on the third day. Transcriptome and proteome sequencing were carried out in Tsingke Biotechnology Co., Ltd. (Beijing, China). For the sequencing results of transcriptome data, firstly, conduct quality control and filtering on the original sequencing data; subsequently, high-quality sequences are aligned to the reference genome or transcriptome. Next, quantitative analysis of gene expression levels (such as counting reads) is carried out, and on this basis, differential expression analysis is conducted to screen for genes with significant changes among different groups. Finally, enrichment analysis of the differentially expressed genes was conducted in combination with functional annotation databases to reveal the biological functions, pathways, or regulatory networks they are involved in. The DESeq2 (1.45.0) was used for significance analysis of differences in the project. A p-value of less than 0.05 and a ratio of expression multiples greater than or equal to 2 (log2FC ≥ 1) were taken as the significance criteria for differences. Proteomics employs Ultra-Fast technology. This analytical method is based on high-throughput mass spectrometry technology and combines ultrafast liquid chromatography with high-resolution mass spectrometry platforms (such as Orbitrap Astral) to achieve rapid separation and detection of complex protein samples. Through efficient database search (such as MaxQuant, FragPipe) and bioinformatics analysis, the system systematically analyzes differentially expressed proteins and their functions.

2.8. Trypan Blue Dyeing

One percent trypan blue dye (Beyotime, C1313M, Shanghai, China) was injected into the tail vein of WT and Zxdb-KO mice on the 7th day of pregnancy for in vivo assessment of the rate of decidualization.

2.9. Hematoxylin–Eosin Staining

Hematoxylin–Eosin (H&E) staining was carried out in accordance with the H&E staining kit instructions (Beyotime, C0105M, Shanghai, China). Paraffin sections of uterine tissue were subjected to gradient dewaxing with xylene and anhydrous ethanol, followed by water washing. The nuclei were stained with hematoxylin for 4 to 10 min and differentiated to turn blue, and the cytoplasm was stained with eosin for 3 to 5 min. Then, the sections were dehydrated with gradient ethanol, made transparent with xylene, sealed with neutral gum, and finally photographed under the microscope (Olympus, IX73, Tokyo, Japan) for observation and analysis.

2.10. Immunofluorescence Staining

Mouse uterine tissue was embedded in OCT (SAKURA, 4583, Torrance, CA, USA), rapidly frozen in liquid nitrogen, and then frozen into sections with a thickness of 7~10 μm. The sections were washed with PBS, dried at 37 °C, and sealed with 5% BSA and goat serum (Solarbio, SL038, Beijing, China). Subsequently, the primary antibody was incubated overnight in a wet box at 4 °C. On the second day, the corresponding fluorescent secondary antibodies (Abcam Goat Anti-Rabbit IgG H&L Alexa Fluor^®^ 594, ab150080, Shanghai, China) and DAPI (Beyotime, P0131, Shanghai, China) were incubated after the runs on the slides. After PBS washing, the plates were encapsulated with anti-quenching mounting medium and finally observed, and images were collected under a fluorescence microscope.

2.11. Quantitative PCR

RNA was extracted using the traditional TRIzol (Invitrogen, 15596018CN, Carlsbad, CA, USA) method. Reverse transcription of 1 μg of RNA. Reverse transcription and genome removal were performed using Novozyme HiScript III RT SuperMix for qPCR Kit (Vazyme, R323-01, Nanjing, China). Dilute the cDNA and perform quantitative PCR using tSYBR^®^ Green Realtime PCR Master Mix (TOYOBO, QPK-201, Osaka, Japan).

2.12. Western Blot (WB)

The WB experiment was conducted according to the classic wet-film transfer method. The antibodies involved in this study include β-actin (Servicebio, GB11001-100, Wuhan, China), CDH1 (Zenbio, 340341, Chengdu, China), CDH2 (Shanbenbio, 1206, Hefei, China), and BMP2 (Shanbenbio, 52878, Hefei, China).

2.13. Statistical Analysis

All statistical analysis and data visualization were performed using Prism 9 software. Data are expressed as the mean ± standard deviation (mean ± SD). Analysis of variance (ANOVA) was used for inter-group comparisons. The t-test is used to compare the differences in the means between two groups. The drawing software uses Biorender and CorelDRAW 2025. ns, no significant difference; statistical significance was expressed as * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.

3. Results

3.1. Expression Analysis of Zxdb in the Reproductive System of Female Mice

To explore the expression of Zxdb in the early stage of the embryo, we re-analyzed the transcriptome and proteome data of mouse embryos from Shaorong Gao’s Data [26]. The results from Figure 1A show that the mRNA level of Zxdb was not present in all embryos during the single-cell zygote stage. No mRNA expression of Zxdb was detected at the two-cell, four-cell, or eight-cell stage. At the morula and blastocyst stages, only some embryos expressed Zxdb mRNA. Overall, Zxdb mRNA was at a very low or absent expression level. As an internal reference gene for the early embryonic development, the Pipa gene is often used to standardize embryonic RT-qPCR or RNA-seq data to ensure the accuracy and comparability of gene expression analysis at different development stages [27,28,29]. Figure 1B shows that as the embryo develops and the number of cells increases, the mRNA of the Pipa gene shows a good, gradual increase. After homogenization, it was found that Zxdb was barely detectable in the two-cell, four-cell, eight-cell, morula, and blastocyst stages. To better understand the expression of Zxdb, we re-analyzed the proteomic data of embryos. Figure 1D shows that PIPA protein expression level steadily increased as the embryo developed from zygote to blastocyst, while the ZXDB protein was not detected at any stage of embryonic development, which also suggests that Zxdb may not play a key role in early embryonic development.

In addition to observing the expression of the Zxdb gene in embryos, we also investigated the expression of Zxdb in the uterine tissue of mice. The uterine tissues of E1.5, E3.5, E5.5, and E7.5 mice were selected. As can be seen from the H&E staining results (Figure 1E), with the secretion of HCG hormone and others, the uterine cavity of the mice began to close from E1.5 to E7.5. When the embryo entered the uterine cavity on day 4.5, there was still some space in the uterus. Once the embryo is implanted, the uterine cavity will close rapidly. The fluid is discharged, and then the embryo invades the mother’s uterine tissue. Around the 7th day, placental blood vessels begin to form (Figure 1E). The quantitative PCR results (Figure 1F) show that the expression of Zxdb gradually increased with the increase in gestational age. In both humans and mice, the majority of embryo loss occurs in the early stage of pregnancy [30,31]. The fertilized egg of a mouse usually begins to enter the uterus and attach to the endometrium after fertilization. A receptive endometrium plays a crucial role in the embryonic development of a mouse. The continuously enhanced expression of Zxdb in the uterine tissue of mice after pregnancy may have a certain effect on the receptivity of the endometrium and maintain the stable implantation of the embryo.

3.2. Decreased Expression of Zxdb by Maternal Factors Leads to Reduced Mouse Embryo Implantation

To further study whether knockdown of Zxdb in uterine tissue in pregnant mice affected embryo implantation, we injected stably modified siRNA directly into the mouse uterine tissue before embryo implantation on day 3 of pregnancy, and the validation of knockdown efficiency was conducted in N2a cells, with an efficiency of approximately 70%. As shown in Figure 2A, mice aged about 6 weeks were treated with PMSG, injected with HCG 48 h later, and then they mated with male mice in cages. The next morning, the vaginal plug was checked. On the third day of embryonic development, the uterus of mice was pulled out, and the uterine horns of the mice were subjected to injection experiments. In the negative control group, the left side served as a self-control, and the right side was injected with normal saline; in the si-Zxdb group, a negative control was injected on the left side as a self-control, and si-Zxdb was injected on the right side (Figure 2B). When the embryo developed to about 7.5 days, the mice were sacrificed. As shown in Figure 2B,C, there was no significant difference in the number of embryos implanted in the left uterus and the right uterus in the negative control group. In the si-Zxdb group, there was a significant difference in the number of embryos implanted in the left uterus and the right uterus. This result shows that knocking down the expression of Zxdb in mouse endometrial tissue reduces the decidualization rate of the mice.

Although from the perspective of transcriptomics and proteomics, the expression of Zxdb in early embryos is very little or non-existent, the effect of this minimal Zxdb expression on embryonic development in mouse embryos remains unclear. To this end, we conducted siRNA knockdown Zxdb experiments on mouse embryos of two strains: C57BL/6J and ICR. Figure 2D shows the results of embryonic development in ICR mice. The results showed that compared with the mock group and negative control group, there were no significant differences in the two-cell, four-cell, eight-cell, morula, and blastocyst formation rates of the injected embryos in the si-Zxdb group (Figure 2E). The corresponding mean values of two-cell, four-cell, right-cell, morula, and blastocyst formation rates in the mock group were 98 ± 4.47% (n = 5), 86.8 ± 9.57% (n = 5), 82.8 ± 13.8% (n = 5), 84.8 ± 12.07% (n = 5), and 81.3 ± 11.23% (n = 5), respectively. The corresponding mean values for two-cell (n = 5), four-cell (n = 5), eight-cell (n = 5), morula (n = 5) and blastocyst (n = 5) formation rates in the negative group were 98.18 ± 2.58% (n = 5), 92.15 ± 7.72% (n = 5), 89.65 ± 6.41% (n = 5), 86.89 ± 7.12% (n = 5) and 79.96 ± 8.04% (n = 5), respectively. The corresponding mean values for two-cell, four-cell, eight-cell, morula, and blastocyst formation rates in the si-Zxdb group were 96.54 ± 5.23% (n = 5), 92.09 ± 6.81% (n = 5), 85.97 ± 4.60% (n = 5), 84.86 ± 5.86% (n = 5), and 82.35 ± 5.04% (n = 5), respectively. For pictures of different developmental stages, please refer to the Supplement Figure S1. The results of interference in C57BL/6J mouse embryos were consistent with those in ICR mouse embryos.

3.3. Construction and Identification of Zxdb-KO Mice

To further investigate the effect of Zxdb, our team constructed the Zxdb-KO mouse genetic model. Figure 3A shows the process of gene-edited mice. Figure 3B shows a schematic diagram of two sgRNA targeted editing targeting the Zxdb gene. In addition to retaining some embryos to develop into blastocysts for editing efficiency testing, all the remaining two-cell embryos were transplanted into surrogate mother mice through the opening at the front end of the fallopian tube. After multiple experiments, a total of four mice were produced (Figure 3C). After genotype identification, a 528 bp deletion Zxdb-KO male mouse was screened out (Figure 3D). The first genotype identification experiment identified the specific genotype of the mutant through PCR. Subsequently, as shown in Figure 3E, the design diagram of the primers was presented. Primers were designed across breakpoints, and the genotype was directly determined based on the size of the electrophoresis bands. The breakpoint product has a KO band of 187 bp and a WT band size of 308 bp (Figure 3F). As CRISPR/Cas9 technology has a certain possibility of being off-target, in order to detect the off-target efficiency, PCR was performed on the sites that the software predicted might be off-target, and first-generation sequencing verification was carried out. The verification results show that none of the possible off-target sites were edited, and there was no off-target effect.

3.4. The Loss-Function of the Zxdb Gene Influences Female Reproductive Phenotype in Mice

To examine whether the phenotypes of genotype knockout mice were consistent with the results of previous RNA interference models, we conducted experiments on homozygous Zxdb-KO mice and WT mice. Figure 4A shows the number of embryos implanted in the uterus of WT and Zxdb-KO mice. The results showed (Figure 4B) that the average number of Zxdb-KO embryos was 9 (n = 9), which was significantly lower than the number of embryo implants in WT mice, which was 15 (n = 9), p < 0.0001. In addition, we conducted a statistic on the number of newborns (Figure 4C). The statistics showed that the average number of offspring born in WT mice was 14 (n = 7), and that in the KO mouse group was 10 (n = 7), p < 0.01. In conclusion, both the number of embryo implants and the number of offspring were significantly reduced after Zxdb knockout.

3.5. Characterizing Differentially Expressed Genes (DEGs) by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway Analysis

To study the effects of Zxdb knockout on the reproduction of female mice, especially the overall impact on embryo implantation. We performed transcriptomic sequencing on WT and Zxdb-KO mice. The results of principal component analysis (PCA) showed that the Zxdb-KO group and the WT group clustered well, indicating that the expression patterns of genes in the samples of the same group were similar (Figure 5A). The volcano plot (Figure 5B) shows that compared with WT, the Zxdb-KO group had a total of 3305 DEGs, among which 1951 DEGs were upregulated and 1354 DEGs were downregulated. The knockout sequencing results showed that after knocking out Zxdb, the expression of Zxdb was lost (Figure 5C), which was consistent with the genotype identification results.

According to gene ontology (GO) analysis, DEGs in Zxdb-KO vs WT mouse transcriptome can be categorized into 14 major biological processes and cellular components (Figure 5D): extracellular region (18.2%), extracellular exosome (14.5%), cell adhesion (9.8%), extracellular space (8.7%), cell surface (6.2%), plasma membrane (5.9%), extracellular matrix (5.4%), collagen-containing extracellular matrix (4.8%), basement membrane (4.1%), apical plasma membrane (3.7%), extracellular matrix structural constituent (3.5%), heparin binding (3.2%), calcium ion binding (2.9%), signaling receptor binding (2.6%), and other processes (approximately 26% combined for remaining terms) (Figure 5D). Based on the hypergeometric test, 12 of these terms were significantly enriched for differentially expressed genes, including extracellular region (p = 1.33 × 10^−42^), extracellular space (p = 1.33 × 10^−42^), extracellular exosome (p = 2.52 × 10^−32^), cell adhesion (p = 1.84 × 10^−35^), heparin binding (p = 1.80 × 10^−32^), extracellular matrix (p ≈ 10^−20^), collagen-containing extracellular matrix (p ≈ 10^−19^), cell surface (p ≈ 10^−16^), plasma membrane (p ≈ 10^−13^), basement membrane (p ≈ 10^−12^), apical plasma membrane (p ≈ 10^−10^), and extracellular matrix structural constituent (p ≈ 10^−8^). Similar to the results of the GO analysis, the KEGG plot of Zxdb-KO mice showed that the adhesive-related pathways were also disordered (Figure 5E): the focal adhesion and cell adhesion molecules were downregulated, weakening the adhesion between the embryo and the uterus. Upregulation of ECM–receptor interaction may lead to abnormal ECM remodeling and hinder trophoblast invasion. Although the upregulation of PI3K-Akt, Rap1, and calcium signaling promotes proliferation, the out-of-control timing may cause inflammation or decidual differentiation disorders, disrupt the receptivity of the implantation window, and ultimately lead to insufficient decidual support, increased embryo absorption, and decreased litter size.

The heat map of Zxdb-KO mice shows dysregulation of key genes for implantation (Figure 5F): downregulation of Cdh1, Itga9, Cdh2, Col26a1, and Cxcl14 weakens embryonic adhesion and ECM remodeling. Downregulation of Mmp2 inhibits invasion. Downregulation of Notch1, Ihh, and Hand2 blocks decidualization and differentiation. Although the upregulation of Bmp2, Ptgs2, and Igf1 compensates, the disordered sequence causes inflammation. Overall disruption of uterine receptivity leads to implantation failure, insufficient decidual support, increased embryo absorption, and a decrease in the birth rate. These data indicate that Zxdb knockout might invoke a variety of genes participating in a wide range of extracellular matrix and cell adhesion-related biological processes.

3.6. Integrative Transcriptome and Proteome Analysis of Zxdb-KO Mice

To gain a clearer understanding of the impact of Zxdb being knocked out on the reproduction of female mice, we also conducted proteomic analysis. The proteomics results of PCA analysis showed that the Zxdb-KO group and the WT group clustered well, indicating that the expression patterns of genes in the samples of the same group were similar (Figure 6A). The volcano diagram (Figure 6B) shows that there are 253 upregulated proteins (red dots), 297 downregulated proteins (green dots), and 11,110 proteins with no significant change (gray dots). It was indicated that after Zxdb knockout, the expression levels of approximately 550 proteins changed significantly, suggesting that Zxdb knockout has a considerable impact on protein expression in mouse cells/tissues. Figure 6C shows the HeatScatter Plot of the expression levels of genes jointly identified by each sample group at the transcriptional level (genes) and the translation level (proteins), which can visually reflect the density distribution and correlation results of protein and gene expression levels. Control group (ρ = 0.33): There was a moderate positive correlation between gene expression and protein expression. KO group (ρ = 0.322): After knocking out Zxdb, the correlation coefficient slightly decreased but did not change much, still showing a moderate positive correlation. The correlation coefficients were close and relatively low, indicating that the overall coupling of gene–protein expression was weak in both groups. After the Zxdb knockout, there was no significant change in the gene–protein correlation, indicating that the knockout did not significantly disrupt the overall correlation, but it might affect the transcription–translation coupling on specific genes. The Venn diagram (Figure 6D) shows the overlapping relationships of all genes, diff genes, diff proteins, and all proteins. The number of differentially expressed genes was far greater than that of differentially expressed proteins (5194 genes vs 1206 genes), indicating that knockout mainly affects the transcriptional level, and transcriptional changes are not completely conducted to the protein level (possibly due to translation regulation or protein stability). The small overlap (105 genes) indicates that many differentially expressed genes did not cause protein changes, suggesting that transcriptional changes were not fully conducted to the proteins, which may involve post-transcriptional regulation. The KEGG enrichment bubble plot (Figure 6E) shows that in Zxdb-KO mice, the proteomes of “focal adhesion” and “ECM–receptor interaction” are significantly enriched (low p-value, Rich Factor 0.2~0.3, and high count); “cell adhesion molecules” are mainly characterized by transcriptome changes. The proteome shows greater significance in the focal adhesion and ECM interaction pathways, suggesting that post-translational regulation after Zxdb knockout enhances the alteration of adhesion signals, which may change the proliferation and invasion of uterine stromal cells.

3.7. Expression Changes of Decidualization Marker Genes and Adhesion Factor-Related Genes

To observe differences in uterine tissue before embryo implantation, uterine tissue was obtained from WT and Zxdb-KO mice on day 3 of gestation. Comparison of H&E staining of uterine implantation sites showed that the decidua layer in the WT group was thick, uniform, and dense, with rich cells, forming a full and regular ring structure surrounding the embryonic cavity. In the Zxdb-KO group, the decidua became significantly thinner, sparse, and uneven, with loose structure, increased edge folds, and flat implantation sites (Figure 7A). These histological differences reflect that Zxdb knockout leads to decidualization defects, insufficient differentiation of maternal decidual cells, an inability to fully support embryo implantation and development, and may ultimately show a significant reduction in litter rate. Based on the enrichment results of differentially expressed genes in the transcriptome and proteome, we conducted WB verification on some crucial adhesion molecules (CDH1, CDH2, and BMP2) of decidualization and embryo implantation. Compared to the WT group, the expression of CDH1 and CDH2 proteins decreased, and the expression of BMP2 protein was upregulated. The expressions of BMP2, CDH1, and CDH2 were disordered compared to the WT, which might lead to abnormalities during embryo implantation and decidualization, resulting in phenotypes such as miscarriage or low fertility. During embryo implantation, normal “cadherin switching” on the uterine side requires downregulation of CDH1 and upregulation of CDH2. The CDH1/CDH2 ratio increases in our result (Figure 7F); this means that the endometrial receptivity window was not open normally, or the embryo develops abnormally, which would destroy the key steps of embryo adhesion and invasion, thus having a significant negative impact on embryo implantation. Immunofluorescence showed that the expression intensity of CDH1 was lower compared with that in the WT (Figure 8). The enlarged image confirmed that the expression network was damaged and its continuity was disrupted. This spatiotemporal disorder eventually leads to the inability of embryos to attach and implant normally, resulting in a significant decrease in the implantation rate.

4. Discussion

The process of embryo implantation and decidualization involves the synergistic effect of a large number of genes. Previous studies, through genetic screening of families with recurrent spontaneous abortion, have identified some key genes. These genes related to recurrent miscarriage include HOXA10 [32,33], FOXO1 [34,35], CDH1 [32,36], LIF [37], MUC1 [38], and BMP2 [39] genes. However, even so, a large number of unexplained recurrent spontaneous abortion (URSA) cases still have no identifiable cause [40]. Perhaps there are more genes involved, especially the transcription factors that regulate differentiation and development.

Progesterone receptor (PR) mediates progesterone signaling, promoting the differentiation of endometrial stromal cells into decidua cells. Transcription factors are crucial in this process [12]. The key transcription factors include CEBPB, HOXA10, FOXO1, and ZEB1, which regulate cell proliferation, differentiation, and the expression of markers, ensuring the establishment of uterine receptivity [14]. CEBPB promotes the differentiation of human embryonic stem cells (hESCs), and decidualization does not occur in the mouse uterus without CEBPB [41,42]. HCG improves endometrial receptivity by increasing the expression of HOXA10 [43]. FOXO1 competes with PGR to control the expression of a subset of genes involved in decidualization [44]. ZEB1 affects embryo implantation by regulating decidualization [45]. Abnormalities in these genes can lead to decidualization defects, implantation failure, or pregnancy complications. ZXDB is a zinc finger transcription factor. Here, we explored the siRNA uterine injection model and the Zxdb-KO genetic model to study the role of Zxdb in the female reproductive system of mice. We found that it plays an important role in decidualization and embryo implantation.

The expression of the Zxdb gene in mouse embryos and mouse uterus tissues was observed in our study. siRNA interference experiments on embryos confirmed that knockdown of Zxdb did not affect embryo development in the early stages of embryonic development (zygotic stage to blastocyst stage). By knocking down Zxdb in uterine tissue, we confirmed that the reduction in expression of Zxdb, which is a maternal factor, affects embryo implantation. However, there may be a post-implantation time period, such as from post-implantation to the fetal heart beating stage. Whether it has an impact on the development of the embryo still requires further research. In this study, the use of inbred strain ICR mice to construct the genetic model was based on their strong reproductive ability, which is highly suitable for conducting research on reproductive-related topics. However, in order to eliminate the influence of different strains, we also performed RNA interference of the Zxdb gene on the embryos of C57BL/6J mice. The results of early embryonic development of C57BL/6J embryos were consistent with those of ICR embryos, indicating that Zxdb has a very limited role in early embryonic development. At present, the majority of KO mouse models available on the market are based on the C57BL/6J strain. Both the disadvantages and the advantages are quite obvious. The ICR mouse model might be a more suitable choice, especially for research topics related to embryo implantation and reproductive development, and it is worth considering.

Transcriptome and proteome are good tools for studying the effect of KO mouse genes on mice as a whole and have been widely used [46,47]. In our study, we used the transcriptome and proteome together to analyze the overall effect of Zxdb-KO mice. The heat map of the Zxdb-KO mouse transcriptome shows that the expression of key genes related to cell adhesion, embryo implantation, and decidualization is severely dysregulated. The primary cause may be the defect of adhesion molecules: Cdh1 and Cdh2 are significantly downregulated, weakening the initial adhesion between the embryo and the uterine epithelium. The CDH1/CDH2 ratio is often used as a marker for MET/EMT, and the dynamic balance of MET/EMT is crucial for successful pregnancy [17,18,19]. The mouse uterine tissues selected in this study were from the pre-implantation stage, reflecting the changes in the endometrium before implantation, which can fully and directly reflect the environment before embryo implantation and provide a complete picture of the environment before implantation. However, it provides limited information on the decidualization state after implantation. Changes in the uterine tissues after implantation may require further research. In addition, other adhesion molecules were also affected. The downregulation of Col26a1 and Cxcl14 may further disrupt the structure and chemotactic balance of the extracellular matrix (ECM), hindering blastocyst attachment and trophoblast invasion. Downregulation of Mmp2 may inhibit ECM remodeling, while upregulation of Adamtsl5 may cause excessive proteolysis, leading to uncontrolled invasion or fibrosis. Although the upregulation of Bmp2, Ptgs2, Igf1, and Fgf2 attempts to compensatorily promote proliferation, the disordered sequence is prone to cause inflammation or tissue disorders. The upregulation of Tfrc reflects metabolic stress and further aggravates the abnormality of the implantation microenvironment. In conclusion, Zxdb knockout may disrupt the adhesion receptivity and decidualization ability of the uterine “implantation window”, leading to unstable embryo adhesion, invasion obstruction, and insufficient decidual support. Ultimately, it is manifested as increased embryo absorption and a significant decrease in the birth rate.

Furthermore, the differences between the transcriptome and the proteome are also worthy of attention. We found that after Zxdb was knocked out, it was more affected by the transcriptome and rarely transmitted to the proteome. On the one hand, it might be because Zxdb regulates the transcriptome level more, and on the other hand, it might be limited by the technical limitations of proteomics. Many trace amounts of protein will be submerged by high-abundance proteins during detection, making them undetectable. ZXDB protein is a good example. Here, ZXDB is a trace protein. In our proteomics data, the ZXDB protein was not detected in either the WT or Zxdb-KO groups. However, this problem does not exist in the transcriptome. By comparing the correlations between the expression of genes and proteins in the two omics, the potential regulatory relationship between genes and proteins can be quickly understood. Of course, a comprehensive analysis of transcriptomics and proteomics can better assess the overall impact on mice after gene knockout.

Some limitations of this study should also be noted in future research. Firstly, one point worth discussing is that Zxdb is a transcription factor. Generally speaking, it may mainly regulate the direction of differentiation and development by regulating gene expression or coordinating with other transcription factors. This research focuses on the relationship between genes and phenotypes and explores the main pathways and DEGs affected by the deletion of Zxdb. However, the motifs or molecular interactions of its action require more in-depth research by subsequent investigators. Secondly, Zxdb is an X-linked gene. The impact of the loss of function of this gene on the phenotype of mice is influenced by X chromosome inactivation. Perhaps there will be some differences in the phenotypes of heterozygous mutant and homozygous KO mice, and more mice are needed for further research. Additionally, the lack of clinical family data is a regrettable aspect. If some loss-of-function mutations of the ZXDB gene were found in RSA or RIF patients, they would undoubtedly have more significant contributions. Finally, the homologous gene of Zxdb in mice, Zxda, is a long non-coding gene. The relationship between Zxda and Zxdb and their impact on the mouse phenotype have not been explored and discussed in this study. More experiments in the future may be needed for further investigation.

5. Conclusions

This study is the first to confirm that the loss of Zxdb function is associated with decidualization of the uterus and the reduction in the number of mouse offspring. Secondly, we demonstrated that the loss of Zxdb function caused by maternal factors is the cause of these phenotypes. Thirdly, comprehensive experiments including transcriptomics, proteomics, Western blotting, and immunofluorescence confirmed that the expression and distribution of adhesion molecules such as CDH1, CDH2, and BMP2 were abnormal in the endometrial tissue of the uterus before embryo implantation. In conclusion, this study reveals the association between Zxdb and the decline in female reproductive capacity in mice, providing a good model for studying embryo implantation and decidualization of the uterus and offering new insights and genetic counseling guidance for researchers and clinicians studying RSA or RIF.