{"paper_id":"d526bffa-2255-476c-83ea-e9cab4f41aec","body_text":"Pregnancy is a tightly synchronized process, comprising implantation, decidualization, placentation and parturition  1 ,  2 . Successful embryo implantation requires cross-talk between the implantation-competent blastocyst and the receptive uterus  3 ,  4 . The uterus is composed of the endometrium and myometrium, with the endometrium consisting of stromal, glandular, and luminal epithelial compartments. Uterine glands play a critical role in establishing uterine receptivity  5 . In both mice and humans, these glands synthesize and secrete factors required for uterine receptivity, embryo implantation, and subsequent embryo development  6 ,  7 . In mice, leukemia inhibitory factor (LIF) is a key gland-derived cytokine secreted immediately before implantation on day 4 of pregnancy (Day 1 = morning of finding vaginal plug). LIF can substitute for nidatory estrogen to induce implantation in both normal and delayed implantation models  8 ,  9 , and loss of maternal LIF results in complete implantation failure  10 .\nIn both mouse and human uteri, the transcription factor Forkhead box A2 (FOXA2) is specifically expressed in glandular epithelial cells  11 ,  12 . Genetic ablation of  Foxa2  in the whole uterus or uterine epithelium results in implantation failure due to impaired LIF production  5 ,  13 ,  14 , underscoring a critical role for FOXA2 in glandular function and implantation success. Recent advances in tissue clearing and three-dimensional imaging have enabled detailed visualization of the uterine gland architecture  15 ,  16 . Using these approaches, we previously demonstrated that uterine glands undergo a FOXA2-dependent differentiation program characterized by extensive branching during the preimplantation period, a process that prepares glands for estrogen responsiveness and subsequent LIF production  17 . This defective gland branching phenotype was observed in mice with epithelial deletion of the estrogen receptor  Esr1  using  Pax2 -Cre or  Wnt7a -Cre  18 . To date, however, regulatory factors governing gland branching have been identified almost exclusively within glandular epithelial cells. Given that gland branching occurs within the stromal compartment, an important unanswered question is whether surrounding stromal cells contribute to this morphogenetic process. Moreover, although FOXA2 serves as a putative marker of uterine glands, glandular epithelium originates from FOXA2-negative luminal epithelium, raising the possibility that FOXA2 expression is not solely cell-autonomous but is instead influenced by cues from the neighboring stromal microenvironment.\nEpidermal growth factors (EGFs) and their receptors are critical for pregnancy  19 ,  20 ,  21 . EGF signaling is mediated through the ERBB family of receptor tyrosine kinases (ERBB1–4), which function as homo- or heterodimers  22 ,  23 . ERBB3 lacks intrinsic kinase activity distinguishing it from other ERBB receptors. Previous studies have shown that  Erbb3  is expressed in both epithelial and stromal cells in the peri-implantation mouse uterus  24 . Our previous study identified ERBB2 and ERBB3 as major receptors for heparin-binding EGF (HB-EGF) in early pregnancy  20 . In the study, we generated females with uterine-specific deletion of  Erbb3  ( Erbb3 f/f \n Pgr Cre/+ , hereafter  Erbb3 d/d ). Notably, around 70% of vaginal plug-positive  Erbb3 d/d  females failed to produce live offspring. A closer examination revealed a marked reduction in uterine gland branching in  Erbb3 d/d  uteri, leading us to hypothesize that uterine ERBB3 signaling regulates gland branching and function, thereby influencing female fertility and pregnancy success.\nIn this study, we demonstrate that  Erbb3 d/d  females exhibit implantation failure associated with reduced LIF secretion. Uterine deletion of  Erbb3  disrupts preimplantation gland branching and impairs glandular function required for the establishment of uterine receptivity. Interestingly, FOXA2 expression is markedly diminished in uterine glands of  Erbb3 d/d  females, suggesting that ERBB3 signaling is essential for maintaining glandular cell identity. Transcriptomic analysis of day 3 uteri further reveals a significant reduction in stromal  Igf1  expression in  Erbb3 d/d  mice, and uterine deletion of  Igf1r  recapitulates the glandular branching defects observed in  Erbb3 d/d  uteri, suggesting that stromal IGF1 signaling acts as a key downstream mediator of ERBB3-dependent regulation of uterine gland development and function.\n\nIn our previous study of uterine HB-EGF  20 , we generated  Erbb3 d/d  female mice with deletion of  Erbb3  in uterine stromal and epithelial cells by crossing  Erbb3 flox/flox  mice with  Pgr Cre/+  mice, and females with uterine epithelial-specific deletion of  Erbb3  by crossing  Erbb3 flox/flox  mice with  Ltf Cre/+  mice ( Erbb3 epi/epi ). Deletion efficiency was confirmed by Western blot analysis using day 4  Erbb3 d/d  uterine tissues (Supplementary Fig. 1a, b). Since  Erbb3  is expressed in epithelial and stromal cells in preimplantation uteri  24 , we further assessed compartment-specific deletion efficiency by isolating primary stromal and epithelial cells from  Erbb3 f/f  and  Erbb3 d/d  uteri. qPCR analyses demonstrated robust depletion of  Erbb3  in both compartments of  Erbb3 d/d  uteri (Supplementary Fig. 1c-e). Consistent with our previous findings, approximately 70% of plug-positive  Erbb3 d/d  females failed to deliver offspring  20 , whereas  Erbb3 epi/epi  females have pregnancy rates and litter sizes comparable to those of  Erbb3 f/f  females ( Fig. 1a  and Supplementary Table 1). We next investigated the developmental onset of this infertility. While  Erbb3 f/f  and  Erbb3 epi/epi  females exhibited normal embryo attachment, 57% (8/14) of  Erbb3 d/d  females showed implantation failure on day 5 of pregnancy ( Fig. 1b ,  c ). Recovery of blastocysts from these females indicates that the implantation failure is due to a maternal defect. This divergent reproductive outcome suggests that although epithelial  Erbb3  is dispensable for implantation, the combined loss of  Erbb3  in both epithelial and stromal compartments markedly compromises pregnancy success. Taken together, these findings implicate stromal  Erbb3  as a critical regulator of female fertility.\nOn day 5 of pregnancy, blastocysts attach to the luminal epithelium and nidate into a spear-shaped implantation chamber (crypt) that is connected with adjacent uterine glands  16 . Using 3D imaging, we assessed the chamber morphology of  Erbb3 f/f ,  Erbb3 epi/epi  and  Erbb3 d/d  implantation sites on days 5 and 6 of pregnancy. On day 5,  Erbb3 d/d  implantation sites exhibited shallower crypt invaginations compared with those of  Erbb3 f/f  and  Erbb3 epi/epi  implantation sites ( Fig. 1d ). However, development of the  Erbb3 d/d  crypts rapidly progressed, and by day 6 of pregnancy, no overt defects in crypt formation were observed across different genotypes ( Fig. 1f ). Interestingly, while glands in  Erbb3 f/f  and  Erbb3 epi/epi  uteri displayed multiple branches on day 5 of pregnancy, most  Erbb3 d/d  uterine glands retained a single, elongated structure lacking typical branching ( Fig. 1d ). Additional implantation chamber images are shown in Supplementary Figure 2. Glands were categorized as unbranched (stem only, 0 branches), containing one or more than one branch ( Fig. 1e ). Quantitative analysis revealed that the proportion of  Erbb3 d/d  glands with one or more branches was significantly reduced compared with control glands. By day 6,  Erbb3 f/f  and  Erbb3 epi/epi  crypts showed continued extension of glandular stems accompanied by coiled branches that were displaced away from the implantation chambers. In contrast, most  Erbb3 d/d  uterine glands remained as elongated main ducts with minimal or no branching ( Fig. 1f ,  g  and Supplementary Fig. 3). These results suggest that uterine gland branching is defective in  Erbb3 d/d  implantation sites.\nTo investigate the cause of impaired implantation in  Erbb3 d/d  females, we examined ovarian hormone levels and their putative receptor expression for uterine receptivity. The ovarian hormones, 17β-Estradiol (E 2 ) and progesterone (P 4 ), play key roles in pregnancy  25 ,  26 , and exert their functions primarily through estrogen receptor α (ESR1) and progesterone receptors (PR), respectively  27 . Serum levels of E 2  and P 4 , as well as uterine expression of their respective receptors, were comparable between  Erbb3 f/f  and  Erbb3 d/d  females on day 4 of pregnancy ( Fig. 2a - c ).  Erbb3 d/d  uteri showed expression levels of  Msx1 , a well-established marker of uterine receptivity  28 , comparable to those of control on day 4 of pregnancy ( Fig. 2d ), despite a reduced number of uterine glands observed in  Erbb3 d/d  sections. Under normal conditions on day 4 of pregnancy, luminal and glandular epithelial cells exit the cell cycle, whereas stromal cells remain highly proliferative  29 . While  Erbb3 f/f  uteri displayed this expected pattern,  Erbb3 d/d  uteri exhibited aberrant Ki67 staining in luminal and glandular epithelia. In contrast, no significant differences in stromal Ki67 staining were detected between  Erbb3 f/f  and  Erbb3 d/d  uteri ( Fig. 2e ).\nUterine glands undergo progressive branching morphogenesis and functional differentiation as the uterus acquires receptivity prior to implantation  17 . This differentiation is critical for the production of LIF in response to preimplantation E 2  secretion on the morning of day 4 of pregnancy  10 . Our observations of defective  Erbb3 d/d  glandular morphology on day 5 prompted us to examine uterine gland architecture on day 4 using 3D imaging. The majority of  Erbb3 f/f  and  Erbb3 epi/epi  uterine glands exhibited extensive branching, whereas  Erbb3 d/d  glands had significantly fewer glands with more than one branch, indicating defective glandular differentiation and potential functional defects ( Fig. 2f ,  g  and Supplementary Fig. 4). Indeed,  Erbb3 d/d  uteri displayed significantly reduced  Lif  expression with some inter-individual variability ( Fig. 2h ,  i ). Among the five  Erbb3 d/d  uteri examined, one exhibited  Lif  expression comparable to that of control uteri, whereas the remaining samples showed a pronounced reduction in glandular  Lif  expression. This decrease in  Lif  production was further quantitatively validated by qPCR analysis using RNA isolated from whole  Erbb3 f/f  and  Erbb3 d/d  uteri on day 4 of pregnancy ( Fig. 2j ). As introduced above, FOXA2 plays a key role in preimplantation gland differentiation and LIF production. We therefore examined FOXA2 expression by immunostaining. In  Erbb3 f/f  uteri, all glands were marked by FOXA2 staining, while  Lif  signals were detected in glandular epithelial cells located away from the uterine lumen. In contrast, FOXA2 signals were markedly reduced in  Erbb3 d/d  uteri ( Fig. 2h ,  i ). A quantitative Western blot analysis confirmed a significant reduction in FOXA2 protein levels in  Erbb3 d/d  uteri ( Fig. 2k ).\nTogether, these results demonstrate that  Erbb3 d/d  uteri have defective uterine gland branching and impaired function. The marked reduction in FOXA2 expression within  Erbb3 d/d  glands suggests that ERBB3 signaling plays a key role in regulating FOXA2-dependent gland identity and differentiation. Moreover, the wide inter-individual variability in  Lif  and FOXA2 expression observed in  Erbb3 d/d  uteri likely underlies the partial penetrance of the reproductive phenotype, providing an explanation for why approximately 30% of  Erbb3 d/d  females are able to carry pregnancies to term.\nWe next asked whether the glandular defects observed in adult  Erbb3 d/d  females arise from abnormalities during early postnatal uterine gland development. An important caveat when comparing the  Pgr -Cre and  Ltf -Cre models is the difference in the timing of Cre activation.  Pgr -Cre becomes active approximately 10 days after birth  30 , whereas  Ltf -Cre is activated by estrogen at puberty  31 . Thus, it is possible that defects in postnatal uterine adenogenesis contribute to the phenotypes observed in adult  Erbb3 d/d  females. In our  Erbb3  mouse colony, 6 weeks of age corresponds to early puberty before the activation of  Ltf -Cre. At this developmental stage,  Erbb3  transcripts were readily detected in both isolated epithelial and stromal cells from  Erbb3 f/f  uteri (Supplementary Fig. 5). The majority of uterine glands in both  Erbb3 f/f  and  Erbb3 d/d  uteri have little or no branching at this stage ( Fig. 3a  and Supplementary Fig. 6). Quantitative analysis revealed comparable numbers of glands across branching categories between the two genotypes ( Fig. 3b ). Importantly, glandular FOXA2 expression was unaltered in  Erbb3 d/d  uteri at this stage ( Fig. 3c ). Together, these results indicate that  Erbb3 d/d  uteri display normal uterine gland morphology and FOXA2 expression prior to puberty, suggesting that the gland branching defects observed in adult  Erbb3 d/d  females arise during or after puberty rather than from impaired postnatal adenogenesis. Because  Erbb3  deletion driven by  Pgr -Cre has little or no impact on prepubertal uterine gland development, and the principal distinction between the  Pgr -Cre and  Ltf -Cre models lies in stromal  Erbb3  deletion after puberty, we infer that stromal  Erbb3  plays a critical role in regulating uterine gland branching and function.\nSince  Lif  expression is reduced in  Erbb3 d/d  glands on day 4, we examined whether exogenous LIF could rescue embryo implantation. After confirming the bioactivity of recombinant LIF protein ( Fig. 3d )  12 , we administered LIF or a PBS vehicle control intraperitoneally to  Erbb3 f/f  and  Erbb3 d/d  females on day 4 and assessed implantation on day 6 ( Fig. 3e ). As expected, PBS-treated  Erbb3 f/f  females exhibited normal implantation, whereas a subset of PBS-treated  Erbb3 d/d  females showed implantation failure, confirming that the vehicle alone does not alter the genotype-specific phenotypes ( Fig. 3f ). Administration of LIF had no detectable effect on implantation outcomes in  Erbb3 f/f  females. In contrast, in the LIF-treated  Erbb3 d/d  females (n = 11), only a small fraction (2 of 11, ~18.2%) exhibited properly formed implantation sites (group #1). Around 1/3 of them (4 of 11) displayed overtly small, compromised implantation sites with faint blue bands (group #2). The remaining 5 females (5 of 11, ~45.4%) had implantation failure with little or no detectable blue bands (group #3) ( Fig. 3e - g ).\nTo further investigate the quality of implantation in these females, we examined the expression of COX2 and HAND2, two established markers of implantation progression and decidualization  32 . COX2 is initially expressed on the anti-mesometrial side of the implantation chamber on day 5 of pregnancy and subsequently shifts toward the mesometrial side as decidualization proceeds  33 . COX2 signals were detected in implantation sites from  Erbb3 f/f  controls and group #1  Erbb3 d/d  implantation sites ( Fig. 3h ). In contrast, little COX2 signals were detected in group #2  Erbb3 d/d  implantation sites. Additionally, the domain of HAND2 expression was reduced in group #2 implantation sites as compared to those of control and group #1 ( Fig. 3h ). COX2 expression patterns were also confirmed using 3D imaging. In group #2  Erbb3 d/d  implantation sites, embryos remained trapped within the luminal epithelium and displayed COX2 expression surrounding the embryo in a pattern resembling a typical day 5 implantation site ( Fig. 3i ), indicating a failure to progress beyond the initial attachment stage. Nonetheless, uterine glands in  Erbb3 d/d  females from both group #1 and group #2 continued to exhibit reduced branching. Together, these observations indicate that restoration of LIF alone is insufficient to rescue implantation progression and decidualization in  Erbb3 d/d  uteri.\nGiven the gland branching defects are observed in  Erbb3 d/d , but not  Erbb3 epi/epi  uteri, we hypothesize that the stromal environment lacking ERBB3 signaling impairs normal gland branching. Our previous study demonstrated that uterine glands undergo progressive branching during the preimplantation period  17 . By the evening of day 3 of pregnancy, a subset of glandular epithelial cells begins to express  Prss29  and becomes competent to respond to E 2  stimulation for subsequent LIF production. Therefore, we selected day 3 of pregnancy to examine the stromal contribution to gland branching at a stage when glands are actively undergoing branching morphogenesis. To globally investigate the cellular composition and transcriptional landscape of the uteri, we performed single-cell RNA sequencing (scRNA-seq) from  Erbb3 f/f  and  Erbb3 d/d  uteri on day 3 using the 10x Genomics Chromium Flex (Flex Gene Expression) platform ( Fig. 4a ).  Erbb3 f/f  and  Erbb3 d/d  uteri have comparable cell populations ( Fig. 4b ), and an unsupervised cell clustering identified 11 distinct cell populations, including luminal epithelial (Le_1, Le_2, Le_3, and Le_4), glandular epithelial (Ge), stromal (Str), mesothelial (Meso), endothelial (Endo), perivascular (Peri), immune (Immune_1 and Immune_2) subsets ( Fig. 4c ). Each cluster was annotated based on canonical marker gene expression ( Fig. 4d ), and transcriptional similarities among clusters were assessed by correlation analysis ( Fig. 4e ). Cell-type proportion analysis also showed fewer glandular epithelial and stromal cells, with a relative increase in luminal epithelial cells in  Erbb3 d/d  uteri compared to controls ( Fig. 4f ,  g ), suggesting a dysregulated stromal-epithelial balance in  Erbb3 d/d  uteri.\nDifferential expression analysis of stromal cells from  Erbb3 f/f  and  Erbb3 d/d  uteri identified 1,411 genes meeting the specified thresholds (|log2FC| > 0.25, min.pct > 0.1). Cross-referencing these genes with a curated growth factor gene set (GO:0008083; 150 unique factors) revealed 10 candidate growth factors that were altered in  Erbb3 d/d  stroma, including  Vegfa, Nrg2, Pdgfd, Bmp7, Gdf7, Hgf, Bmp1, Gmfb, Igf1,  and  Manf  ( Fig. 4h ). These scRNA-seq data indicate that uterine deletion of  Erbb3  disrupts stromal growth factor expression, potentially compromising stromal-epithelial communication during gland branching.\nTo confirm the candidate growth factors identified by scRNA-seq analysis ( Fig. 4h ), uterine stromal cells were isolated from  Erbb3 f/f  and  Erbb3 d/d  females on day 3 of pregnancy and subjected to bulk RNA sequencing. Principal component analysis showed that most  Erbb3 d/d  samples segregated from  Erbb3 f/f  controls, although one  Erbb3 d/d  sample clustered with the control group. This heterogeneity is consistent with our observation of incomplete phenotype penetrance in  Erbb3 d/d  females (Supplementary Fig. 7a) and was further supported by sample-to-sample correlation analysis (Supplementary Fig. 7b). Quality control assessment of expression distributions demonstrated well-normalized raw count data across all eight samples, supporting the robustness of downstream analyses (Supplementary Fig. 7c).\nGiven the gland branching defects observed in  Erbb3 d/d  uteri, we next examined whether established genes governing gland development were globally altered. Gene Set Enrichment Analysis (GSEA) revealed widespread changes in genes associated with “Gland Morphogenesis” following  Erbb3  deletion ( Fig. 5a ). Differential expression analysis identified 252 differentially expressed genes (DEGs), with 52 genes upregulated and 200 genes downregulated in  Erbb3 f/f  stromal cells compared with  Erbb3 d/d  stromal cells ( Fig. 5b ,  c ). Over-representation analysis of these DEGs showed strong enrichment for pathways related to reproductive system development, glandular and branching morphogenesis, and growth factor regulation ( Fig. 5d ), consistent with a requirement for stromal ERBB3 signaling in uterine gland development. We then examined individual genes within the enriched categories of “Morphogenesis of a branching epithelium” ( Fig. 5e ), “Gland morphogenesis” ( Fig. 5f ), and “Growth factors” ( Fig. 5g ).  Igf1  was identified as the top candidate gene among all three categories. Notably, among the 10 candidate growth factors identified by scRNA-seq ( Fig. 4h ),  Igf1  again emerged as the most significantly altered gene in the bulk RNA-seq dataset ( Fig. 5h ,  i ). To assess the spatial expression of  Igf1 , we performed fluorescence  in situ  hybridization on day 3 uterine sections, which revealed robust stromal  Igf1  expression in  Erbb3 f/f  uteri that was significantly reduced in  Erbb3 d/d  uteri ( Fig. 5j ). In addition, we confirmed reduced FOXA2 expression in  Erbb3 d/d  uteri ( Fig. 5j ), consistent with impaired glandular development.\nTo independently confirm the findings from our scRNA-seq and bulk RNA-seq analyses, we reanalyzed a publicly available scRNA-seq dataset of day 3 pregnant mouse uteri (GSA: CRA011266)  34 . Cells were reclustered by consolidating previously defined subpopulations into eight major cell types, including glandular epithelium (GE), luminal epithelium (LE), stromal cells (Str), and immune cells (Immune) (Supplementary Fig. 8a-c). We then assessed intercellular communication using ligand-receptor interaction analysis with CellChat. Collagen-mediated extracellular matrix signaling dominated the global communication network, with stromal cells serving as the primary signaling source (outgoing strength: 10.529), while LE and GE cells were the major signal recipients (incoming strengths: 7.503 and 6.307, respectively) ( Fig. 6a - c ).\nIGF signaling was again identified as one of the top stromal-to-epithelial communication pathways and exhibited a strictly unidirectional paracrine pattern ( Fig. 6b ,  d ). Specifically, stromal cells functioned as the exclusive source of IGF signals, transmitting exclusively to epithelial compartments ( Fig. 6e ). Gene expression mapping further supported this directionality, showing that the primary receptor  Igf1r  and its co-receptors  Itga6  and  Itgb4 \n 35  were highly enriched in epithelial populations on day 3 of pregnancy ( Fig. 6f ). Consistent with these findings, ligand receptor interaction analysis revealed significant communication probabilities for the  Igf1-Igf1r  axis directed from stromal cells to both luminal epithelium (0.016) and glandular epithelium (0.022) ( Fig. 6g ). Together, these analyses identified  Igf1-Igf1r  paracrine signaling as a strong candidate pathway through which stromal ERBB3 signaling influences uterine gland differentiation.\nTo validate the functional relevance of the IGF1-IGF1R paracrine pathway in gland branching, we examined uterine morphology in  Igf1r f/f \n Pgr Cre/+  females on day 4 of pregnancy. Three-dimensional imaging revealed that uterine-specific deletion of  Igf1r  resulted in a significant reduction in gland branching compared with  Igf1r f/f \n Pgr +/+  control mice ( Fig. 7a ,  b  and Supplementary Fig. 9). Quantitative analysis showed that whereas most glands in control uteri displayed complex higher-order branching (>1 branch), over half of the glands in  Igf1r f/f \n Pgr Cre/+  uteri lacked branches (0 branch), remaining as primitive, elongated stems ( Fig. 7c ). These findings are consistent with previous reports demonstrating a significant reduction in the number of implantation sites on day 5 of pregnancy in  Igf1r f/f \n Pgr Cre/+  females  36 . Despite this pronounced branching defect,  Igf1r  deletion did not affect expression of the glandular marker FOXA2 ( Fig. 7d ), indicating that the IGF1-IGF1R paracrine pathway regulates gland branching through a FOXA2-independent mechanism.\nWe next asked whether IGF1R activation is diminished in  Erbb3 d/d  uteri, which exhibit reduced  Igf1  expression. Previous studies have shown that IGF1 induces phosphorylation of IGF1R, leading to activation of AKT via phosphorylation at Ser473 as a canonical downstream signaling event  37 . Western blot analysis of whole uterine tissue collected on day 4 of pregnancy revealed reduced levels of p-AKT in  Erbb3 d/d  uteri compared with  Erbb3 f/f  controls (Supplementary Figs. 10a, b). To assess the cellular specificity of this defect, we performed immunostaining and found that p-AKT was broadly distributed across both stromal and epithelial compartments. Notably, p-AKT signals were markedly reduced in luminal and glandular epithelial cells of  Erbb3 d/d  uteri on days 3 and 4 of pregnancy (Supplementary Figs. 10c, d). Together, these findings indicate that uterine deletion of  Erbb3  impairs IGF1R activation, resulting in reduced epithelial AKT signaling, and support a model in which stromal ERBB3 maintains  Igf1  production to drive epithelial IGF1R-AKT signaling essential for preimplantation gland branching.\n\nOur current study demonstrates that EGF signaling is critical for uterine gland morphogenesis and differentiation. Uterine glands originate from the luminal epithelium  38 , invaginate into the underlying stroma, and subsequently undergo elongation and branching. These morphogenetic processes are essential for establishing a functional glandular network critical for embryo implantation during mouse pregnancy  38 . Multiple signaling pathways have been shown to play critical roles in early uterine adenogenesis, among which Wnt signaling is well characterized. Loss-of-function of  Wnt5a ,  Wnt7a , or  Ctnnb  results in a glandless uterine phenotype  39 ,  40 ,  41 ,  42 . Similarly, neonatal deletion of  Wnt4  in the uterus leads to a reduction in gland number  43 . Neonatal uterine deletion of  Foxa2  or an adherens junction protein  Cdh1  also abolish adenogenesis  5 ,  44 . However, detailed studies of uterine gland morphology were limited until the development of three-dimensional visualization methods for the uterus  15 ,  16 . Two recent studies showed that post-pubertal deletion of  Foxa2  and prenatal epithelial deletion of  Esr1  result in branchless glands and compromised gland function  17 ,  18 , suggesting that gland branching is essential for proper glandular function. Beyond these studies, our understanding of the signaling pathways that regulate uterine gland branching remains limited. Our current study identifies ERBB3-mediated EGF signaling as a key regulator of uterine gland branching and function, which advances our understanding of the molecular mechanisms underlying uterine gland branching.\nAn important caveat in comparing the  Pgr- Cre and  Ltf- Cre models is the difference in the timing of Cre activation.  Pgr- Cre becomes active approximately 10 days after birth  30 , whereas  Ltf- Cre is activated by estrogen at puberty  31 . Therefore, it is possible that defects in neonatal uterine adenogenesis in  Erbb3 d/d  mice substantially contribute to the phenotypes observed in adulthood. To address this possibility, we compared uterine gland morphology in  Erbb3 d/d  and  Erbb3 f/f  mice at 6 weeks of age, immediately prior to the initiation of  Ltf -Cre activity. In our  Erbb3  mouse colony, 6 weeks of age corresponds to early puberty, a developmental stage during which females have undergone minimal to no estrous cycles. As shown in  Figs 3a  and  3b , quantitative 3D imaging analysis revealed no significant differences in gland branch number between  Erbb3 d/d  and  Erbb3 f/f  uteri (P > 0.05 across all branching categories). In addition, expression of the glandular marker FOXA2 was comparable between genotypes ( Fig. 3c ). These findings suggest that ERBB3 signaling influences glandular FOXA2 expression only in the presence of ovarian hormones. Although the mechanism underlying the stage-specific regulation of FOXA2 by ERBB3 remains unclear, our data indicate that early postnatal uterine adenogenesis is not impaired in  Erbb3 d/d  mice. Therefore, the gland branching defects observed after puberty are unlikely to arise from a pre-existing developmental abnormality and instead reflect a requirement for ERBB3 signaling during the post-pubertal period.\nErbb3  expression in the stromal compartment plays a critical role in uterine gland branching. To date, the only mouse models reported to exhibit branchless uterine glands involve epithelial-specific deficiencies of  Foxa2  or  Esr1 \n 17 ,  18 . These findings raise the question of whether uterine gland branching is regulated intrinsically by the glands themselves during the estrous cycle and early pregnancy.  Erbb3  is expressed in both uterine epithelial and stromal cells.  Erbb3 f/f \n Pgr Cre/+  females with deletion of  Erbb3  in both epithelia and stroma display compromised gland branching, whereas  Erbb3 f/f \n Ltf Cre/+  females with epithelial specific deletion of  Erbb3  show normal gland morphogenesis. Notably, pre-pubertal adenogenesis is not impaired in  Erbb3 f/f \n Pgr Cre/+  uteri. Collectively, these results suggest that loss of stromal  Erbb3  is required to drive the gland branching defects. However, we cannot conclude that stromal  Erbb3  deletion alone is sufficient to drive the defects, as direct evidence from a stroma-specific knockout model is currently lacking.  Amhr2 -Cre has been used to delete genes in the uterine stroma, but a recent study showed that  Amhr2 -Cre has Cre activity in early embryos resulting in global genetic modification  45 ;  Pdgfra -Cre ERT2  has been used for uterine stromal lineage tracing experiments  46 , but  Pdgfra -Cre has activity in mesenchymal cells across multiple organs beginning at embryonic stages. Despite these limitations, our study highlights the essential role of stromal ERBB3 signaling in post-pubertal gland morphogenesis.\nThe stromal environment plays a critical role in shaping uterine glandular cell identity.  Foxa2  is expressed in uterine glands beginning at neonatal adenogenesis, and its gland-specific expression is maintained throughout adulthood  5 ,  13 ,  14 . FOXA2’s faithful presence in uterine glands makes it a putative marker of uterine glands. In humans, dysregulation of endometrial FOXA2 is associated with several uterine pathologies, including complex atypical endometrial hyperplasia, uterine carcinoma, and endometriosis  11 ,  47 . In mice, uterine deletion of  Foxa2  disrupts glandular function and compromises female fertility  5 ,  13 ,  14 . In addition, we recently demonstrated that uterine glands progressively acquire branches from day 1 to day 4 of pregnancy, and uterine epithelial deletion of  Foxa2  ( Foxa2 f/f \n Ltf Cre/+ ) impairs glandular branching  17 . These findings indicate that FOXA2 is essential for maintaining uterine gland identity and proper gland morphogenesis. However, because uterine glandular epithelium originates from FOXA2-negative luminal epithelium, it remains unclear whether FOXA2 expression in uterine glands is autonomously maintained by glandular epithelial cells or regulated by signals from the surrounding stroma. In this study, we show that FOXA2 levels are reduced in adult, but not prepubertal,  Erbb3 d/d  glands ( Figs. 2h  and  3c ), suggesting that uterine glandular cell identity is compromised in the absence of post-pubertal EGF signaling. Nevertheless, it is still unknown whether loss of FOXA2 causes glandular epithelial cells to revert to a luminal epithelial fate, or whether these FOXA2-negative glandular cells persist in an intermediate or transitional state between glandular and luminal epithelial identities.\nExogenous LIF administration is insufficient to functionally rescue implantation failure in  Erbb3 d/d  mice. In  Erbb3 d/d  uteri, defects in gland branching and reduced FOXA2 expression in uterine glands ultimately result in impaired glandular function and diminished LIF production ( Fig. 2 ). However, supplementation with exogenous LIF failed to rescue the implantation failure observed in  Erbb3 d/d  females. These findings indicate that uterine  Erbb3  deficiency leads to multiple defects in uterine receptivity, including glandular LIF insufficiency. Consistent with this interpretation, the reduced stromal IGF1 levels in  Erbb3 d/d  uteri further indicate that stromal cell function is also compromised. Because stromal, luminal epithelial, and glandular epithelial cells all play key roles in uterine receptivity, LIF supplementation alone is insufficient to overcome the implantation failure in  Erbb3 d/d  females.\nStromal IGF1 mediates the impact of ERBB3 signaling on uterine gland morphogenesis. Our finding that stromal ERBB3 is essential for gland branching prompted the hypothesis that stromal ERBB3 influences glandular morphology through one or more secreted mediators.  Igf1  was identified as a strong candidate mediator by our scRNA-seq and bulk RNA-seq analyses intersecting with a secreted protein library.  In vivo , endometrial cells can be influenced by two sources of IGF1: systemic circulating IGF1 and IGF made locally within the uterus. Circulating IGF1 is secreted predominantly by the liver. A previous study demonstrated that  Igf1 f/f \n Pgr Cre/+  females show decreased fertility  48 , indicating that systemic IGF1 is unable to compensate for reduced local IGF1 levels in the uterus during pregnancy.  Igf1 f/f \n Pgr Cre/+  females have oviductal embryo retention and compromised preimplantation embryo development  49 , which precludes their use for implantation studies. Gland branching and differentiation occur over several days, making functional rescue experiments particularly challenging. Moreover, IGF1 is a potent growth factor and systemic hormone  in vivo  (reviewed in  50 ). Repeated administration of exogenous IGF1 can induce marked physiological effects, including acute hypoglycemia, renal fluid retention, and increased capillary permeability. Given these broad mitogenic and metabolic actions, systemic IGF1 treatment would confound attempts to directly assess IGF1-mediated rescue of uterine gland defects.\nStromal IGF1 mediates ERBB3’s impact on gland morphogenesis via targeting epithelial IGF1R. Our cell-cell communication analysis using scRNA-seq data suggested that stromal IGF1 likely acts through its receptor IGF1R on epithelial cells. A recent study demonstrated that IGF1R is predominantly expressed in the uterine epithelium throughout early pregnancy  36 , while  Igf1  ligand expression is mainly expressed in the stroma. The spatial proximity strongly supports that stromal IGF1 signals directly through epithelial IGF1R. In fact, uterine deletion of  Igf1r  leads to reduced gland branching similar to the phenotype of  Erbb3 d/d  uteri. A comparable mechanism has been reported in the mammary gland, where IGF1 stimulation promotes epithelial expansion, while epithelial-specific  Igf1r  deficiency leads to branching defects  51 . These findings collectively support a model in which IGF1-IGF1R signaling mediates stromal-epithelial communication during gland development. IGF1 can also target the insulin receptor (INSR), which is expressed in the preimplantation uterus. However, recent studies showed that  Insr f/f \n Pgr Cre/+  females have numbers of implantation sites comparable to those of control females on day 6 of pregnancy, whereas  Igf1r f/f \n Pgr Cre/+  females display a significant reduction in implantation sites  36 ,  52 . These findings suggest that uterine IGF1R plays a more critical role than INSR during early pregnancy. Although we cannot exclude a contributory role for INSR in mediating IGF1 signaling, the glandular morphogenetic defects observed in  Igf1r f/f \n Pgr Cre/+  uteri demonstrate that IGF1R is a key mediator of uterine gland branching.\nSurprisingly, FOXA2 expression is comparable between  Igf1r f/f \n Pgr Cre /+  and control uterine glands, whereas it is markedly reduced in  Erbb3 d/d  glands. This finding suggests that ERBB3 regulates glandular branching through both a FOXA2-dependent and a FOXA2-independent mechanism mediated by the IGF1-IGF1R pathway ( Fig. 8 ). Nevertheless, several important questions remain unresolved. The mechanism by which ERBB3 regulates IGF1 expression is unclear. In addition, how ERBB3 controls glandular FOXA2 expression and how IGF1R activation promotes gland branching are not fully understood. Addressing these questions will require further investigation.\n\nErbb3 f/f \n 53  mice were crossed with  Pgr Cre/+ \n 30  or  Ltf Cre/+ \n 31  mice to generate  Erbb3 f/f \n Pgr Cre/+  ( Erbb3 d/d ) and  Erbb3 f/f \n Ltf Cre/+  ( Erbb3 epi/epi ) mouse lines.  Erbb3 d/d ,  Erbb3 epi/epi  and their control littermates were housed in the animal care facility at Cincinnati Children’s Hospital Medical Center according to the National Institute of Health and institutional guidelines for laboratory animals. All protocols were approved by the Cincinnati Children’s Animal Care and Use Committee. Mice were provided with irradiated Laboratory Rodent Diet 5R53 and autoclaved water ad libitum. These mice were housed under a 12:12 hr light:dark cycle.  Pgr Cre /+  mice were originally provided by Francesco DeMayo (National Institute of Environmental Health Science, Research Triangle Park, NC) and John B. Lydon (Baylor College of Medicine, Houston, TX).  Erbb3  floxed mice were provided by David Threadgill (Texas A&M University, College Station, TX).  Igf1r f/f \n 54  mice were obtained from Jackson Lab and crossed with  Pgr Cre/+  mice to generate  Igf1r f/f \n Pgr Cre/+  mouse line. All animal experiments involving  Igf1r f/f \n Pgr Cre/+  mice were performed in the animal facilities of Xiamen University in accordance with the approved guidelines of the Animal Welfare Committee of the Research Organization of Xiamen University. At least three mice from each genotype were used for each individual experiment.\nAdult females from each genotype were randomly chosen and housed with a fertile WT male of choice overnight; the morning of finding a vaginal plug was considered successful mating (day 1 of pregnancy). Pregnancy in plug-positive mice was confirmed on day 4 by flushing one uterine horn with saline to identify blastocysts. For pregnant mice on days 5 and 6, 100 μl of 1% Chicago blue in saline was injected intravenously for four minutes to visualize implantation sites as blue bands. In the absence of blue bands, uterine horns were flushed to assess the presence of embryos.\nFISH experiments were performed as previously described  12 . Briefly, frozen sections (12-μm) were processed on the same slide for each probe. Following fixation in 4% paraformaldehyde and acetylation, slides were hybridized at 55°C with digoxigenin-labeled  Lif ,  Igf1 , or  Msx1  probes. Anti-DIG-peroxidase was applied onto hybridized slides following washing and peroxide quenching. The color was developed by a fluorescein Tyramide Signal Amplification kit according to the manufacturer’s instructions (PerkinElmer). Counterstaining was performed using a CK8 antibody to label epithelial cells (TROMA-I, DSHB) and a FOXA2 antibody to label glandular cells (8186S, Cell Signaling Technology). Nuclei were counterstained with Hoechst 33342 (2 μg/ml, H1399, Thermo Scientific). Images were acquired using a confocal microscope (Nikon Eclipse TE2000).\nFor immunofluorescence, 12-μm frozen sections from each genotype were co-mounted on single slides to standardize staining conditions. Sections were incubated with primary antibodies against ESR1 (sc-542, Santa Cruz; 1:300), PR (8757, Cell Signaling Technology; 1:300), Ki67 (RM-9106-S, Invitrogen; 1:300), HAND2 (AF3876, R&D systems; 1:300), Vimentin (5741, Cell Signaling Technology; 1:300), CK8 (TROMA-1, Developmental Studies Hybridoma Bank; 1:300), E-cadherin (13–1900, 1:300, Invitrogen), FOXA2 (8186s, Cell Signaling Technology; 1:200), or Phospho-Akt (Ser473) (9277, Cell Signaling Technology; 1:300). Signals were visualized using Alexa Fluor 488-, 594-, or 647-conjugated secondary antibodies (Jackson ImmunoResearch; 1:200). Nuclei were counterstained with Hoechst. Confocal images were acquired using a Nikon Eclipse TE2000 microscope. The mesometrial pole is the dorsal side of the uterus where blood vessels enter and the placenta forms; the antimesometrial pole is the ventral side where a blastocyst initially attaches. Histological sections were oriented and labeled accordingly to ensure consistent regional comparisons.\nFluorescence intensity was quantified using ImageJ (NIH). Briefly, fluorescence images were opened and split into individual channels using the Split Channels function under the Image Color menu. The channel corresponding to the target signal was selected for analysis. For each image, regions of interest (ROIs) were manually delineated using a freehand selection tool, and the mean gray value within each ROI was quantified using the  Measure  function. All images within each comparison group were acquired using identical microscope settings, and fluorescence intensities are reported in arbitrary units (a.u.). The detailed step-by-step methods are illustrated in Supplementary Fig. 11. For quantification of  Lif  RNA and FOXA2 protein signals, fluorescence intensity was measured from each individual gland.\nSera were collected on day 4 of pregnancy (1100 h), and hormone levels were measured by enzyme immunoassay kits (Estradiol ELISA Kit, 501890 & Progesterone ELISA Kit, 582601, Cayman) as previously described  31 .\nThe whole-mount immunostaining with 3DISCO tissue clearing was performed as previously described  16 . Briefly, mouse uteri were fixed in Dent’s fixative (methanol:DMSO, 4:1) at −20 °C overnight and then washed with 100% methanol three times for 1 hour each. The samples were bleached with 3% H 2 O 2  in methanol at 4 °C overnight. After three washes in 1% PBST for 1 hour each, the samples were blocked in 5% BSA and then incubated with anti-E-cadherin (13–1900, 1:200, Invitrogen) and/or anti-COX2 (12282s, 1:100, Cell Signaling Technology) in the same buffer at 4 °C on a rotor for 7 days. After incubation, the samples were washed with 1% PBST six times for 1 hour each and incubated with Alexa-conjugated 488, 594 or 647 antibodies (1:200, Jackson Immuno Research) in a light-proof box for 4 days at 4 °C. After six washes in 1% PBST at room temperature, the samples were dehydrated in 100% methanol for 30 minutes and then cleared in benzyl alcohol/benzyl benzoate (BABB) solution for at least 1 hour in a light-proof box.\n3D pictures were acquired using a Nikon FN1 Upright Microscope. Samples were laid on slides, covered with BABB, and enclosed by cover slips for confocal imaging using a 10X objective with 8 μm Z-stack. Image files generated using Nikon NIS-Elements were imported into Imaris software (version 10.2.0; Bitplane) for visualization and three-dimensional (3D) reconstruction. Tissue 3D structures were generated using the  Surface  tool. Specific tissue regions were isolated by manual segmentation with the  Surface  tool following a published protocol (“Segmentation Using Imaris 10.2.0”  17 ), and the  Mask  option was applied for subsequent pseudocoloring. Three-dimensional images were generated using the  Snapshot  tool.\nEpithelial and stromal cells were isolated from mouse uteri by sequential enzymatic digestion as described previously  55 . Uteri were slit open longitudinally and cut into small fragments. For day 3 and day 4 pregnant uteri, tissue fragments were incubated with pancreatin (25 mg/mL; Sigma) and dispase (6 mg/mL; Gibco) for 30 min at 4 °C, followed by 20 min at room temperature and an additional 10 min at 37 °C. For uteri from 6-week-old adult mice, the same enzyme concentrations were used, but incubation was performed at 37 °C for 10 min. In both cases, luminal epithelial sheets were removed by gentle washing with HBSS and collected separately. The remaining tissue was then digested with type IV collagenase (0.5 mg/mL; Worthington) for 30 min at 37 °C to release stromal cells. Cell suspensions were filtered through a 70-μm nylon mesh to remove tissue clumps. The separated epithelial and stromal fractions were collected for RNA extraction and qPCR analysis. The purity of each fraction was confirmed by immunofluorescence staining for Vimentin (stromal cells), FOXA2 (glandular epithelium), and E-cadherin (all epithelium).\nTotal RNA was extracted from uterine tissues using TRIzol Reagent (Invitrogen, Cat# 15596026) according to the manufacturer's instructions. RNA concentration and purity were assessed by measuring absorbance at 260 and 280 nm; samples with A260/A280 ratios between 1.8 and 2.0 were used for downstream analysis. Five hundred nanograms of total RNA were reverse transcribed into cDNA using the PrimeScript RT Reagent Kit (TaKaRa Bio, Cat# RR037A) following the manufacturer's protocol. Quantitative PCR was performed on a StepOnePlus Real-Time PCR System (Applied Biosystems) using Fast SYBR Green Master Mix (Thermo Fisher Scientific). Reactions were run in triplicate with 1 μM primers. Relative gene expression was calculated using the 2 +ΔΔCt  method and normalized to  Gapdh . Primer sequences are listed in Supplementary Table 2.\nProtein extraction and Western blotting were performed as previously described  56 . The primary antibodies used in this study included rabbit anti-ERBB3 Antibody (1:1000; 12708s, Cell Signaling Technology), rabbit anti-FOXA2 antibody (1:1000; 8186s, Cell Signaling Technology), rabbit anti-AKT Antibody (1:1000; 9272s, Cell Signaling Technology), rabbit anti-p-AKT (Ser473) Antibody (1:1000; 4060s, Cell Signaling Technology), and rabbit anti-GAPDH antibody (1:1000; 2118s, Cell Signaling Technology). Bands were visualized using an ECL Prime Western blotting detection system (GE Healthcare). GAPDH served as loading controls.\nTotal RNA was extracted from day 3 mouse uterine stromal cells and subjected to paired-end RNA sequencing. Reads were aligned to the mouse reference genome (mm10) using HISAT2 with default settings. Alignments were converted to BAM and coordinate-sorted with samtools. To restrict analyses to uniquely mapped primary alignments, reads with NH:i:1 and without secondary/supplementary flags were retained. Gene-level read counts were obtained from the sorted BAM files using featureCounts (paired-end mode, chimeric/discordant pairs excluded) against a gene annotation (GTF) matching mm10. Differential expression analysis was performed in DESeq2 on raw counts; p values were adjusted by the Benjamini-Hochberg method, and genes with adjusted p < 0.05 were considered differentially expressed. Variance-stabilized counts were used for PCA and sample-to-sample correlation. Functional enrichment of DEGs was assessed using clusterProfiler for Gene Ontology biological processes.\nGene Set Enrichment Analysis (GSEA) was performed using the gseGO function in the clusterProfiler R package with the org.Mm.eg.db annotation database. Rather than restricting the analysis to differentially expressed genes, we used the full transcriptomic profile of all detected genes (~15,000). Genes were ranked by their log2 fold change values derived from the DESeq2 analysis ( Erbb3 f/f  versus  Erbb3 d/d ). Enrichment was assessed against Gene Ontology Biological Process (GO:BP) gene sets, with gene set sizes constrained between 10 and 500 genes. P values were adjusted using the Benjamini-Hochberg method, and terms with adjusted P < 0.05 were considered significantly enriched. A permutation seed of 123 was used to ensure reproducibility. Core enrichment (leading-edge) genes were extracted for downstream pathway analysis.\nSample preparation was performed using the GEM-X Flex Sample Preparation v2 Kit (1000781, 10x Genomics) according to the manufacturer's protocol. Endometrial tissues (~25 mg) from  Erbb3 f/f  and  Erbb3 d/d  females were isolated by scraping the luminal surface of the uterus with a scalpel on a pre-chilled glass slide maintained on ice, thereby excluding the myometrial layer, finely minced, and fixed in Fixation Buffer B (nuclease-free water, Conc. Fix & Perm Buffer B, and 37% formaldehyde) for 16–24 h at 4°C. Samples were subsequently washed with chilled PBS (850 ×  g , 5 min, 4°C), quenched in Quenching Buffer B, and enzymatically dissociated in Dissociation Solution containing collagenase I (0.1 g/10 mL HBSS; 17100-017, Gibco), collagenase IV (0.1 g/10 mL HBSS; LS004188, Worthington), and penicillin-streptomycin (P&S; 208330, Gibco) at 37°C for 40 min with intermittent pipetting every 10 min. The resulting cell suspensions were passed through a 40 μm strainer, washed with PBS, and either processed immediately for library construction or stabilized with Enhancer and 50% glycerol for storage at −80°C. Single-cell libraries were generated using the Chromium Next GEM Single Cell 3′ Gene Expression Flex v2 kit (10x Genomics) and sequenced on an Illumina platform.\nRaw sequencing data were processed using CellRanger (v9.0.0, 10x Genomics) and imported into R via the Seurat package (v5). Cells were retained if they contained 300–6,000 detected genes with mitochondrial content below 15%. Following normalization and principal component analysis, batch effects were corrected using Harmony, and cells were clustered at resolution 0.3 and visualized by UMAP. Eleven transcriptionally distinct populations were annotated based on canonical marker gene expression, encompassing luminal and glandular epithelium, stroma, mesothelium, endothelium, pericytes, and immune cells. Differential expressions between  Erbb3 f/f  and  Erbb3 d/d  stromal cells were assessed using FindMarkers (logFC > 0.25, min.pct > 0.1), and the resulting gene list was cross-referenced against a curated growth factor gene set (GO:0008083) to identify candidate paracrine factors dysregulated in  Erbb3 d/d  stroma.\nRaw data were obtained from Wang  et al.  (GSA: CRA011266)  34 , which profiled mouse uteri at 2.5, 3.5, 4, and 4.5 DPC. In that dataset, 0.5 DPC equals the day of vaginal-plug detection; in our study this corresponds to day 1, thus 2.5 DPC equals day 3 of pregnancy. FASTQ files from 16 sequencing runs (four runs per time point: runs 59–62 for DPC 2.5, 63–66 for DPC 3.5, 67–70 for DPC 4, and 71–74 for DPC 4.5) were processed with Cell Ranger v5.0.1 (10x Genomics) against the mm10 reference genome to generate filtered feature-barcode matrices. Individual Seurat objects were created for each run (Seurat v5.2.1; min.cells = 3, min.features = 200), then merged into a single object with run-specific cell-ID prefixes. Quality control retained cells expressing >500 and <6,000 genes with <15% mitochondrial reads. Data were log-normalized, the top 2,000 variable features were identified by VST, and expression values were scaled. PCA was performed on variable features, followed by Harmony batch correction across runs. UMAP embedding and graph-based clustering (resolution 0.5) were computed using 10 Harmony-corrected dimensions. Clusters were annotated on the basis of established markers into 19 cell types: luminal epithelium (LE), glandular epithelium (GE), stromal subsets (Str_1-Str_4), proliferating stroma (Prol_Str1-Prol_Str3), endothelial (Endo), mesothelial (Meso), pericyte (Peri), myofibroblast (Myo), macrophage subsets (Macro_1-Macro_3), NK cell subsets (NK_1, NK_2), T cells, and B cells. The Peri/Myo cluster was further sub-clustered at resolution 0.1 using Harmony dimensions 1–10 and separated into pericyte (sub-cluster 0;  Notch3 + /Rgs5 + ) and myofibroblast (sub-clusters 1 and 2, merged;  Acta2 + /Tagln + ) populations. For day 3-specific analyses, the DPC 2.5 subset was extracted. To focus on the stromal-epithelial signaling axis, stromal subclusters (Str_1-Str_4, Prol_Str1, Prol_Str3) were merged into a single “Str” class, immune populations were merged into “Immune,” and Prol_Str2 was removed. The resulting eight broad cell classes (LE, GE, Str, Meso, Endo, Peri, Myo, Immune) were validated by Spearman correlation heatmap and dot plot of canonical markers.\nCell-cell communication among LE, GE, and Str on day 3 of pregnancy was inferred using CellChat v1.6.1 with CellChatDB.mouse as the ligand-receptor reference. The normalized expression matrix was extracted from the RNA assay data layer, and a CellChat object was created with cell-type labels. Standard workflows were applied: overexpressed genes and interactions were identified, communication probabilities were computed, interactions with fewer than 10 cells in either sender or receiver group were filtered, pathway-level probabilities were calculated, and signaling networks were aggregated. Network centrality metrics were computed via netAnalysis_computeCentrality. Global signaling roles were visualized by scatter plot of outgoing versus incoming strength and by outgoing/incoming signaling pattern heatmaps. Information flow scores for all inferred pathways were ranked by summing interaction probabilities across all sender-receiver pairs per pathway. For the IGF pathway specifically, Str-LE and Str-GE signaling was visualized with circle plots (netVisual_aggregate), IGF-specific outgoing and incoming strengths were quantified by bar plot, receptor expression ( Igf1r, Itga6, Itgb4 ) across the three populations was shown by violin plot, and ligand-receptor pair communication probabilities were displayed with a bubble plot. All communication probabilities for inferred interactions were exported as source data.\nStatistical analyses were conducted using GraphPad Prism (v8.0) and R (v4.4.1) along with RStudio (2024.04.2). Each experiment was repeated at least three times. Data are shown as mean ± SEM. Statistical analyses were performed using a two-tailed Student’s  t -test, except for pregnancy outcomes presented in Supplementary Table 1, which were analyzed using a  Chi -square test. A  P  value less than 0.05 was considered statistically significant.","source_license":"CC-BY-4.0","license_restricted":false}