EpCAM supports exit from pluripotency of embryonic stem cells via Eomes.

Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Cell Death Dis . 2026 Apr 11;17(1):389. doi: 10.1038/s41419-026-08734-w Search in PMC Search in PubMed View in NLM Catalog Add to search EpCAM supports exit from pluripotency of embryonic stem cells via Eomes Ningyue Gong Ningyue Gong 1 Department of Otorhinolaryngology, LMU University Hospital, LMU, Munich, Germany Find articles by Ningyue Gong 1, # , Mahesh Gouda Mahesh Gouda 1 Department of Otorhinolaryngology, LMU University Hospital, LMU, Munich, Germany 2 Department of Psychiatry and Psychotherapy University Hospital Bonn, University of Bonn Venusberg-Campus 1, Bonn, Germany Find articles by Mahesh Gouda 1, 2, # , Ana Marija Balaz Ana Marija Balaz 1 Department of Otorhinolaryngology, LMU University Hospital, LMU, Munich, Germany Find articles by Ana Marija Balaz 1 , Jiahang Song Jiahang Song 1 Department of Otorhinolaryngology, LMU University Hospital, LMU, Munich, Germany Find articles by Jiahang Song 1 , Gisela Kranz Gisela Kranz 1 Department of Otorhinolaryngology, LMU University Hospital, LMU, Munich, Germany Find articles by Gisela Kranz 1 , Julia Hess Julia Hess 3 Research Unit Translational Metabolic Oncology (TMO), Institute for Diabetes and Cancer (IDC), Helmholtz Diabetes Center, Helmholtz Munich, Neuherberg, Munich, Germany 4 Department of Radiation Oncology, LMU University Hospital, Ludwig Maximilians University Munich, Munich, Germany 5 Joint Heidelberg-IDC Translational Diabetes Program, Department of Inner Medicine I, Heidelberg University Hospital, Heidelberg, Germany 6 German Center for Diabetes Research (DZD), Neuherberg, Germany Find articles by Julia Hess 3, 4, 5, 6 , Philipp Baumeister Philipp Baumeister 1 Department of Otorhinolaryngology, LMU University Hospital, LMU, Munich, Germany 7 Bavarian Cancer Research Center (BZKF), Munich, Germany Find articles by Philipp Baumeister 1, 7 , Kristian Unger Kristian Unger 4 Department of Radiation Oncology, LMU University Hospital, Ludwig Maximilians University Munich, Munich, Germany 7 Bavarian Cancer Research Center (BZKF), Munich, Germany 8 German Cancer Consortium (DKTK), Partner Site, Munich, Germany 9 Comprehensive Cancer Center (CCC), Munich, Germany Find articles by Kristian Unger 4, 7, 8, 9 , Vera Katalina Vera Katalina 1 Department of Otorhinolaryngology, LMU University Hospital, LMU, Munich, Germany Find articles by Vera Katalina 1 , Martin Canis Martin Canis 1 Department of Otorhinolaryngology, LMU University Hospital, LMU, Munich, Germany Find articles by Martin Canis 1 , Olivier Gires Olivier Gires 1 Department of Otorhinolaryngology, LMU University Hospital, LMU, Munich, Germany Find articles by Olivier Gires 1, ✉ Author information Article notes Copyright and License information 1 Department of Otorhinolaryngology, LMU University Hospital, LMU, Munich, Germany 2 Department of Psychiatry and Psychotherapy University Hospital Bonn, University of Bonn Venusberg-Campus 1, Bonn, Germany 3 Research Unit Translational Metabolic Oncology (TMO), Institute for Diabetes and Cancer (IDC), Helmholtz Diabetes Center, Helmholtz Munich, Neuherberg, Munich, Germany 4 Department of Radiation Oncology, LMU University Hospital, Ludwig Maximilians University Munich, Munich, Germany 5 Joint Heidelberg-IDC Translational Diabetes Program, Department of Inner Medicine I, Heidelberg University Hospital, Heidelberg, Germany 6 German Center for Diabetes Research (DZD), Neuherberg, Germany 7 Bavarian Cancer Research Center (BZKF), Munich, Germany 8 German Cancer Consortium (DKTK), Partner Site, Munich, Germany 9 Comprehensive Cancer Center (CCC), Munich, Germany ✉ Corresponding author. # Contributed equally. Received 2025 Aug 8; Revised 2026 Mar 12; Accepted 2026 Mar 27; Collection date 2026 Dec. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . PMC Copyright notice PMCID: PMC13076738  PMID: 41963283 Abstract Epithelial cell adhesion molecule (EpCAM) is a tumor-associated antigen that marks pluripotent embryonic stem cells (ESCs). Regulation of Epcam expression yields a spatiotemporal patterning during embryogenesis that is thoroughly mimicked in a 3D model of spontaneous differentiation of embryoid bodies (EBs). Here, we present a role of EpCAM in exit from pluripotency of murine ESCs (mESCs) to establish cardiomyocytes in EBs. Comparative transcriptomic analysis of wildtype and Epcam -knockout mESCs at strategic time points of spontaneous differentiation uncovered molecular deficiencies of Epcam -knockout ESCs in “Wnt signaling” and “Heart development”. Multi-level bioinformatic analyses revealed central lineage-defining transcription factors Eomes , Foxa2 , and Gata6 as differentially expressed genes (DEGs) that are misregulated in Epcam -knockout mESCs. Gene expression association of Epcam with Eomes , Foxa2 , and Gata6 was prominent at day three of spontaneous differentiation, representing primitive streak formation in EBs. Interrogation of public single-cell RNA sequencing (scRNAseq) datasets supported a co-expression of Epcam and Eomes at early stages of murine embryogenesis in epiblast, primitive streak, nascent mesoderm, extraembryonic ectoderm and endoderm. Newly generated scRNAseq of wildtype mESCs in spontaneous differentiation delineated the formation of epiblast, primitive streak, endo- and mesoderm cells, and cardiomyocytes. Expression and pseudotime analysis positioned Epcam expression slightly ahead of Eomes at the transition of early to late primitive streak, along with rising Wnt signaling. Accordingly, conditional re-expression of Epcam or Eomes but not of Foxa2 or Gata6 complemented differentiation defects of Epcam -knockouts and confirmed an involvement of Wnt signaling in the EpCAM-dependent activation of Eomes . Hence, defective exit of pluripotency in Epcam -deficient ESCs is linked to Eomes regulation via Wnt signaling. Subject terms: Cell signalling, Differentiation Introduction Murine embryonic stem cells (mESCs) are self-renewing and pluripotent cells that give rise to undifferentiated progeny upon symmetric division and to ecto-, meso-, and endodermal lineages following asymmetrical division [ 1 – 3 ]. Central aspects of embryonic development can be thoroughly recapitulated in a manageable in vitro 3D model of embryoid bodies (EBs) [ 4 ]. Murine EBs have facilitated the identification of transcriptomic changes associated with early and late pluripotency, exit from pluripotency, primitive streak formation, and cell specification within a time frame of five days [ 5 ]. Spontaneous differentiation of EBs in the absence of any stimulant gives rise to various cell lineages, prominently mesoderm-derived cardiomyocytes [ 6 , 7 ]. Epithelial cell adhesion molecule EpCAM is a pan-carcinoma antigen [ 8 ]. that is also expressed on the cell surface of human and murine pluripotent stem cells, in tumor-initiating cells of various carcinoma entities [ 9 – 11 ]. and in liver progenitors including hepatic endoderm [ 12 – 14 ]. Although EpCAM is commonly regarded as an epithelial marker, EpCAM´s expression is substantially more intricate, and programmed patterning occurs along different cell lineages. Single-cell level analysis of the expression of Epcam during murine gastrulation uncovered a complex patterning during differentiation. Strict loss of Epcam gene expression was observed in nascent mesoderm progenitors whereas endodermal lineages maintained high levels of Epcam . Patterning of EpCAM was confirmed in perinatal mouse embryos, with a selective expression in cells of the endodermal lineage [ 14 , 15 ]. Loss- and gain-of-function of Epcam in mESCs demonstrated a requirement for a strictly regulated expression across lineages, as over-expression and knockout of Epcam resulted in faulty differentiation [ 15 ]. An involvement of Epcam in the regulation of pluripotency and differentiation was likewise reported for human ESC (hESC) and induced pluripotent stem cell (iPSC) [ 16 – 19 ]. ScRNAseq analysis of differentiating human iPSC revealed a predominant expression of Epcam in pluripotent iPSC, early and late primitive streak, and in endodermal derivatives. Pseudotime analysis depicted a fluctuating expression of Epcam starting with a pronounced presence in pluripotent cells and primitive streak. In the pseudotime stream, Epcam expression was substantially reduced in mesodermal progenitors, early cardiac progenitors, and definitive endoderm. Further progression into mesodermal, cardiac, and endothelial lineages was characterized by complete loss of Epcam , whereas definitive and hepatic endoderm cells experienced a strong enhancement of Epcam expression to levels superior to pluripotent iPSC [ 14 ]. In hESC, the intracellular domain EpICD of EpCAM, which is generated via regulated intramembrane proteolysis (RIP) of full-length EpCAM [ 19 – 21 ]. binds and regulates promoters of pluripotency factors Oct3/4, Nanog, Sox, KLF4, and cMyc [ 19 ]. The initial step of RIP of EpCAM sheds a soluble extracellular domain EpEX that promotes the transcription of pluripotency genes upon activation of the epidermal growth factor receptor (EGFR), signal transducer and activator of transcription 3 (STAT3), and LIN28 [ 17 ]. In colorectal [ 22 ]. and head and neck squamous cell carcinoma (HNSCC) [ 23 , 24 ]. EpEX functions as an EGFR ligand modulating epithelial-to-mesenchymal transition (EMT). During mesodermal differentiation of mESC and after EMT induction in carcinoma cells, endocytosis and degradation of EpCAM support its gradual loss at the plasma membrane [ 25 ]. Thus, evidence suggests regulatory functions of EpCAM in stem and carcinoma cell differentiation, but molecular networks and key regulatory genes remain insufficiently characterized [ 26 ]. Here, EpCAM´s role in mESC differentiation was interrogated through scRNAseq analysis and a comparative, temporal transcriptomic analysis of wild-type and knockout cell lines. We reveal an association of EpCAM with cardinal lineage-regulatory transcription factors Eomesodermin ( Eomes ), Forkhead box A2 ( Foxa2 ), and GATA-binding factor 6 ( Gata6 ), and identify Eomes as important regulator of EpCAM-dependent cardiomyocyte formation. Results Kinetic changes associated with loss of EpCAM in mESC Cell surface expression of EpCAM was analyzed over time during spontaneous differentiation of mESC line E14TG2α in EBs. Under pluripotency conditions, > 98% of cells expressed EpCAM to high levels. At D3 of differentiation, EpCAM expression was homogeneously enhanced and gradually declined at D5 and D7 to less than 10% positive cells (Fig. 1A and Supplementary Fig. 1A, B ). EpCAM was selectively expressed in EBs upon spontaneous differentiation and was mutually exclusive to cardiomyocyte marker alpha cardiac actin (α-CAA) (Supplementary Fig. 1C ). Fig. 1. EpCAM expression and impact during spontaneous differentiation of ESCs. Open in a new tab A EpCAM expression was analyzed in WT E14TG2α mESCs by flow cytometry during spontaneous differentiation in embryoid bodies (EBs). Percentages of EpCAM-positive cells (left) and normalized EpCAM mean fluorescence intensity (EpCAM MFI, right) with SD from n = 3 independent measurements are shown. ** p -value < 0.01; *** < 0.001; **** < 0.0001. B WT and EpCAM −/− cells were subjected to spontaneous differentiation in EBs in the absence of LIF. Contraction rates were quantified from D10 to D14 as mean with SD of n = 3 independent experiments. **** p -value < 0.0001; ns: not significant. C Cardiomyocyte marker α-CAA was quantified by qRT-PCR in WT and EpCAM −/− E14TG2α over time. Shown are mean with SD of n = 3 independent experiments. **** p -value < 0.0001. D E14TG2α EpCAM −/− ESCs (clones #56 and #114) were stably transfected with conditional, doxycyclin-responsive vectors expressing EpCAM in fusion with enhanced green fluorescence protein (EGFP). Transgene expression was induced via doxycycline in 2D-culture and EBs were generated. Upon doxycycline removal, transgene expression was assessed upon flow cytometry quantification of EpCAM-EGFP. Shown are mean with SD from n = 3 independent experiments. * p -value < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001. Contraction rates of EBs from WT, EpCAM −/− ESCs (clones #56 and #114), and EpCAM-EGFP re-expression clones at D14 of spontaneous differentiation are shown as mean with SD from n = 3 independent experiments including at least six EBs per experiment and treatment. Where indicated, EpCAM −/− ESCs (#56 and #114) re-expression clones were pre-treated with doxycycline (dox). Expression of α- CAA was quantified by qRT-PCR at day 14. Shown are mean with SD from n = 3 independent experiments performed in triplicates. * p -value < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001. E Scheme of experimental procedure. Wild-type (WT) and Epcam knockout (EpCAM −/− ) murine E14TG2α were cultured in hanging drops in the absence of LIF. Embryoid bodies (EBs) were cultured for 14 days, and contraction was quantified. At D0, 3, 7, and 10, 3´-bulk sequencing was performed on quadruplicates. F Principal component analysis (PCA) of WT and EpCAM −/− cells (clones #56 and #114) at D0, D3, D7, and D10. G Top ten differential gene sets identified upon GSEA using within GSEA-MSigDB are depicted with normalized enrichment scores (NES) and normalized p-values. Shown are the GSEA from WT versus EpCAM −/− cell clones #56 and #114 at D3. H Contraction rates of EBs from WT and EpCAM −/− ESCs (clones #56 and #114) at D10 and D14 of spontaneous differentiation are shown as mean with SD from n = 3 independent experiments including at least six EBs per experiment and treatment. Where indicated, EpCAM −/− ESCs (clones #56 and #114) were treated with MEK inhibitor (MEKi 1 µM), GSK3β (25 nM), or WAY (50 nM) ( C ). * p -value < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001; ns: not significant. Expression of α-Caa was quantified by qRT-PCR at day 14 of spontaneous differentiation in the presence or absence of MEK inhibitor in EpCAM −/− ESCs (clones #56 and #114). Shown are mean with SD from n = 3 independent experiments performed in triplicates. * p -value < 0.05, ** < 0.01, *** < 0.001. CRISPR-Cas9-generated EpCAM −/− single cell clones [ 15 ]. Expressed no detectable levels of EpCAM (Supplementary Fig. 1D, E ) and were impaired in spontaneously forming contracting cardiomyocytes (Fig. 1B ). This impairment was accompanied by a loss of α-CAA mRNA induction in knockout clones compared to WT cells (Fig. 1C ) and reduced levels of α-CAA protein without recognizable patterning (Supplementary Fig. 1F ). The phenotype specificity was confirmed upon conditional re-expression of EpCAM-EGFP from a doxycycline-inducible promoter in Epcam −/− clones #56 and #114. Expression of EpCAM-EGFP was inducible and resulted in correct sub-cellular localization at the plasma membrane (Supplementary Fig. 2A, B ). Expression analysis further revealed leakage in clone #56 in the absence of doxycycline. Leakage in clone #56 led to a 2.26-fold enhanced EpCAM-EGFP expression compared to clone #114. Doxycycline treatment induced a strong EpCAM-EGFP expression in 91.2% and 88.7% of #56 and #114 cells, respectively (Supplementary Fig. 2C, D ). Transgene expression was induced before EB formation and relieved during spontaneous differentiation to mimic EpCAM´s gradual loss along differentiation in WT ESCs (Fig. 1D ). In knockout clone #56, EpCAM leakage in absence of doxycycline at the time point of EB formation complemented cardiomyocyte formation to 50% of WT. Enhanced expression of EpCAM upon doxycycline addition restored WT contraction rates in clone #56 at D14 ( > 90% of WT). Weak EpCAM leakage in clone #114 had no influence on contraction rates in absence of doxycycline, whereas EpCAM re-expression restored contraction to 50% of WT after doxycycline addition. α-CAA expression levels confirmed contraction rates observed after EpCAM re-expression in EpCAM −/− clones (Fig. 1D ). Thus, enhanced expression of EpCAM at early stages of spontaneous differentiation revealed essential to ensure cardiomyocyte formation in EBs. Next, EpCAM´s impact on the transcriptome was addressed using bulk 3´-RNA sequencing (3´-RNAseq) across cardinal stages of spontaneous differentiation (GSE293121). Pluripotent mESCs (D0; in the presence of LIF) and spontaneously differentiating EBs at D3 of LIF withdrawal corresponding to genetic changes in association with primitive streak, and D7 and D10 representing early and late time points of differentiation were assessed for WT and knockout mESC lines (Fig. 1E ). Principal component analysis (PCA; n = 10,000 most variant genes) showed clustering of biological replicates for each cell lines and time points (Fig. 1F ). Differentially expressed genes (DEGs; log2FC > 1, FDR < 0.05) consistently observed between WT and both EpCAM −/− clones revealed minor transcriptomic differences under pluripotency (D0, n = 36 DEGs). At D3, D7, and D10, n = 58, n = 448, and n = 309 DEGs were identified, respectively (Supplementary Fig. 3A and Supplementary Table 1 ). At D3, most DEGs were down-regulated in EpCAM −/− cells versus WT, whereas up- and down-regulated DEGs were distributed more evenly at D7 and D10. Hence, kinetic transcriptomic differences between WT and EpCAM −/− ESCs suggested early dysregulation starting at D3 and amplifying at D7 and D10. Gene set enrichment analysis (GSEA) using the Molecular Signature Database (MSigDB) delivered insights in murine canonical pathways affected by EpCAM. Both EpCAM −/− clones used in this study were highly similar regarding transcriptional differences compared to WT cells with n = 117 common gene sets (Supplementary Fig. 3B, C ). Starting from D3 and ongoing, WT cells were characterized by an up-regulation of the canonical pathways (CP) “Wnt signaling”, “Heart development”, and “NCAM1 interactions” compared to EpCAM −/− cell clones (Fig. 1G and Supplementary Fig. 3D ). Common down-regulated hallmarks in WT versus EpCAM −/− were “Golgi associated vesicle biogenesis” (D0 and D7), “SRP-dependent cotranslational protein targeting to membrane” (D3), and “Mitochondrial translation” (D10). Hence, differences in regulated gene sets between WT and EpCAM −/− cells mirrored functional deficiencies of knockout cells in “Heart development” and indicated “Wnt signaling” as potentially involved. Treatment with Wnt-activating compounds AZD6244 (MEK inhibitor [ 27 , 28 ]. CHIR99021 (GSK3β inhibitor), and WAY262611 restored contraction rates in EpCAM −/− clones. AZD6244 had the strongest restoring effect and rescued contraction rates to 50.9% and 56.4% of WT cells in EpCAM −/− clones #56 and #114, respectively. CHIR99021 and WAY262611 recovered 36.6%/33.8% and 26%/31.9% of the WT contraction rate in EpCAM −/− clones (Fig. 1H ). Based on the superior complementation of the EpCAM knockout phenotype by MEKi, we interrogated the levels of α-CAA following treatment and show a significant up-regulation (Fig. 1H ). Notably, Wnt signaling was required in the early phase of EB differentiation from D0 to D6, whereas a permanent activation throughout the differentiation period was ineffective. Thus, EpCAM-related defects in exit from pluripotency and differentiation are associated with disturbance of Wnt activation in the early phase of ESC differentiation. Identification of functional DEGs Next, we sought to identify genes involved in EpCAM-dependent processes of mESC differentiation with particular focus on cardiomyocyte development, which we termed functional DEGs (fDEGs). For this purpose, a three-tiered selection was conducted on our bulk transcriptomic dataset. Tier 1: DEGs with a Log 2FC > 1.0 and FDR < 0.05 were identified between WT and both KOs. Tier 2: Significantly activated or suppressed hallmarks of the molecular signature database (MSigDB) were defined by GSEA comparing WT cells with both KO clones. Tier 3: DEGs from Tier 1 that contribute to selected hallmarks relevant to cardiomyocyte formation were extracted independently of significant changes in the corresponding hallmarks. Here we concentrated on “Pluripotency”, “Stem cell differentiation”, “Mesoendodermal differentiation”, “Mesoderm differentiation”, “Endoderm differentiation”, “Ectoderm differentiation”, “Heart development”, “Cardiac contraction”, and “WNT signaling”. Functional DEGs are the intersection of genes across tiers, ensuring they represent DEGs contributing to gene networks relevant to the phenotype of EpCAM knockouts (see scheme in Fig. 2A ). Fig. 2. Definition of EpCAM-associated functional DEGs in mESC exit from pluripotency. Open in a new tab A Schematic representation of the workflow and three-tiered selection applied to identify functional DEGs (fDEGs) between WT and EpCAM −/− cells. B – D Heatmaps of functional DEGs at D3, D7, D10 of spontaneous differentiation identified between WT and EpCAM −/− cells (clones #56 and #114). E Expression of Epcam and n = 18 fDEGs is shown as dot plot graph in the indicated murine lineages. F UMAP representation of co-expression patterns of Epcam with Eomes , Foxa2 , and Gata6 , respectively, in lineages of murine development. APS anterior primitive streak, DE definitive endoderm, EXE extraembryonic endoderm, EXEct: extraembryonic ectoderm, Epi epiblast, G gut, IM intermediate mesoderm, MM mixed mesoderm, NM nascent mesoderm, NC notochord, PE parietal endoderm, PS primitive streak, VE visceral endoderm. G WT and EpCAM −/− ESCs (clones #56 and #114) were subjected to spontaneous differentiation in EBs and were analyzed by qRT-PCR at indicated time points for the expression of Eomes , Foxa2 , and Gata6 in independent experiments ( n = 3). Shown are mean with SD. * p -value < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001; ns: not significant. Expression levels are displayed for WT (blue), EpCAM −/− clones #56 (red) and #114 (green). H Immunofluorescence staining of EpCAM in combination with Eomes, Foxa2, and Gata6 is shown in wildtype E14TG2α cells at the indicated time points of spontaneous differentiation in EBs. Shown are representative sections of n = 3 independent experiments performed with multiple EBs. Under pluripotency conditions, central regulators of “Stem cell differentiation” and germ layer definition such as Gata6 , Nanog , Stat3 , Esrrb , and Foxa2 were determined in the GSEA/hallmark intersect (Supplementary Fig. 4A ). Expression of coding and non-coding transcripts related to primitive streak formation have been reported at D3 of spontaneous differentiation in EBs [ 5 ]. Shortly thereafter (D3.5) initiation of patterning of EpCAM expression occurs, which was associated with a complete loss of expression in mesodermal cells and retention in endodermal cells [ 15 ]. At D3, transcription factor (TF) Foxa2 was determined as fDEG enriched in “Mesoendodermal differentiation” and “Heart development” (triple-intersected). TF Gata6 was a triple-intersected fDEG identified in “Pluripotency”, “Stem cell differentiation”, “Endoderm differentiation”, and “Heart development” (Fig. 2B and Supplementary Fig. 4B ). Foxa2 , Gata6 , and Wnt11 were DEGs identified in the GSEA, and Mesp1 , a central regulator of cardiomyocyte formation, and DKK1 , a negative regulator of Wnt signaling, were identified as DEGs involved in “Mesoderm differentiation” and “Endoderm differentiation” (double-intersected). Further prominent regulators of differentiation included T ( Brachyury ), Eomes , and Mixl1 (Fig. 2B and Supplementary Fig. 4B ). Specifically assessing “Heart development” and “Cardiac contraction” hallmarks at this early time point of spontaneous differentiation, fibroblast growth factor 8 and 10 ( Fgf8 , Fgf10 ) and the ubiquitously transcribed tetratricopeptide repeat containing, Y-linked protein Uty were identified as DEGs. All DEGs at D3 were repressed in EpCAM −/− clones except cyclin-dependent kinase inhibitor Cdkn2a , suggesting a lack of induction of central lineage regulators in combination with an induction of cell cycle inhibitors in EpCAM KOs (Fig. 2B and Supplementary Fig. 4B ). At D7 and D10, a total of n = 15 and n = 7 triple-intersected genes, respectively, were identified; four of which were shared between both differentiation time points, i.e ., Esrrb , Fbxo15 , Fgf4 , and Nanog . fDEGs included pluripotency genes and central regulators of differentiation such as Nanog , Twist1 , Foxa2 , and Klf5 . Collagen V alpha sub-units Col5a1 , Col5a2 , the endodermal switch Foxa2 at D7, caudal-type homeobox protein 2 ( Cdx2 ) and heart- and neural crest derivatives-expressed protein 1 ( Hand1 ) at D10, were enhanced in WT cells, whereas all other functional DEGs showed an enhanced expression in both EpCAM −/− clones (Fig. 2C, D and Supplementary Fig. 4C, D ). Cdx2 is a TF expressed in intestinal epithelial cells that is relevant to the differentiation and maintenance of intestinal lining. Hand1 is a TF with an ill-understood role in cardiac morphogenesis. Thus, late time points of spontaneous differentiation in the absence of EpCAM were characterized by a reduced expression of TFs Foxa2 , Cdx2 , and Hand1 , and the retention of pluripotency factors ( Nanog , Klf5 ), indicating a defect in pluripotency exit. Single cell resolution of EpCAM-associated DEG expression in embryogenesis Temporal and cell-specific expression of Epcam and associated fDEGs was investigated in murine embryogenesis in public scRNA-seq datasets covering post-gestation days E3.5-E6.75 and E6.5-E8.5, respectively [ 29 , 30 ]. This approach was chosen to interrogate physiological relevance and co-expression patterns of EpCAM and fDEGs at single-cell resolution during murine in vivo embryogenesis. At early pre-gastrulation time point E3.5, Epcam was strongly expressed, showed transient decrease at E4.5, and reinforced expression at E5.5, E6.5, and E6.75 (Supplementary Fig. 5A ). Eomes was expressed in few cells at early time points E3.5-5.5 and was more strongly and frequently expressed in cells at days E6.5 and E6.75. Foxa2 was not expressed at E3.5 and E5.5, showed a weak expression at E4.5, and strongest expression at E6.5/E6.75. Gata6 was expressed in selected cells at E3.5 and E4.5, and very weakly at E6.5 (Supplementary Fig. 5A ). Correlations of expression of fDEGs with Epcam are depicted in Supplementary Fig. 5B . Pluripotent and epiblast cells were mapped to gestation days and the expression of Epcam and functional DEGs was analyzed. Epcam was expressed in pluripotent cells, most strongly at E3.5, and in the epiblast at E5.5 and 6.5, and showed considerable overlap with Eomes , which was additionally expressed at E4.5 in pluripotent cells. Foxa2 was primarily expressed in pluripotent cells at E4.5, in pluripotent and epiblast cells E6.5, and showed overlap with Epcam . Gata6 was expressed in pluripotent cells at E3.5, E4.5, and E6.5, showing overlap with Epcam (Supplementary Fig. 5C ). The expression of Epcam and fDEGs was analyzed at later gestation days E6.5–E8.5 [ 29 ] (Supplementary Fig. 6A for detailed expression profiles and kinetics). Numbers of cells expressing Epcam , Eomes , Fgf8 , and T showed a sharp increase at gestation day E7.5, starting as co-expression of Epcam with Eomes and Fgf8 at E7.0–7.25 (Supplementary Fig. 6B-D ). Epcam was most frequently expressed in the epiblast, extraembyonic ecto- and endoderm, primitive streak, rostral neuroectoderm, and in fewer cells of the gut, nascent mesoderm, surface ectodem, and visceral endoderm (Fig. 2E ). Co-expression of Epcam with Eomes , Foxa2 , and/or Gata6 was predominant in (anterior) primitive streak, and endodermal tissue. Specifically, co-expression of Epcam with Eomes occurred in the epiblast, (anterior) primitive streak, extraembryonic ectoderm, nascent mesoderm, extraembryonic ectoderm, and visceral endoderm. (Fig. 2F ). This timely association of Epcam expression with Eomes, Foxa2, and Gata6 (Fig. 2G ), and with Dkk1, Fgf8, Fgf10, Mesp1, Mixl1, T, and Wnt11 was confirmed at D3 of spontaneous differentiation in EBs of WT but not of EpCAM −/− clones (Supplementary Fig. 7 ). Immunofluorescence staining of WT cells demonstrated a co-expression of EpCAM with Eomes and Foxa2 in single cells at D3 and a selective Gata6 expression in the absence of EpCAM at D7 (Fig. 2H ). To decipher cross-species associations of Epcam with central genes in embryonic differentiation, we assessed expression patterns in the landscape of human gastrulation in the public dataset GSE155121 [ 31 ]. Carnegie stages 12-16 of human embryonic development, representing post-fertilization age 29-31 days to day 39, were examined for the expression of fDEGs (Supplementary Fig. 8A ). Epcam expression was observed at CS12, CS13-14, and CS15-16, and primarily in epithelium, epidermis, and endoderm. Except for Fgf4 , ESRRB , Nanog , and TBXT , which were expressed in only very few cells, all other fDEGs showed spatiotemporally defined expression patterns (Supplementary Fig. 8A ). Epcam was co-expressed with Foxa2 , Gata6 , and Dkk1 in single endodermal cells. In the epidermis, co-expression was seen with Stat3 , Fgf8 , and Twist1 , whereas neural progenitors co-expressed Epcam and Foxa2 , Sall3 , Fgf8 , and Twist1 . In somites and spanchnic lateral plate mesoderm (LPM), Epcam was co-expressed with Gata6 , Dkk1 , Sall3 , and Twist1 (Supplementary Fig. 8B ). Similarly to murine embryonic development, Epcam was expressed in high numbers of single cells (Supplementary Fig. 8C ), and strongest co-expression at the level of single cell numbers was seen with Gata6 , Foxa2 , Dkk1 , and Fgf8 (Supplementary Fig. 8D, E ). Co-expression with Eomes could not be addressed based on low numbers of positive cells at these developmental stages. Co-expression of Epcam and Foxa2 occurred in epidermis, epithelium, endoderm, and in few neural (progenitor) single cells. Co-expression with Gata6 was less confined and observed in splanchnic LPM, endoderm, and endothelium. The expression of Epcam , Foxa2 , and Gata6 occurred specifically in epidermis, epithelium, and endoderm. Furthermore, co-expression of Epcam with Dkk1 , Stat3 , and Fgf8 was seen in endoderm, epidermis, epithelium, and neural cells (Supplementary Fig. 8F ). scRNAseq analysis of mESC in spontaneous differentiation Next, we resolved gene (co)-expression patterns at single cell-level during spontaneous differentiation of E14TG2α EBs. scRNAseq analysis yielded transcriptomes from n = 8,594, n = 11,243, and n = 8,248 cells at D3, 5, and 7, respectively (GSE318465). Gene expression-based clustering of single cells revealed segregated clusters at D3, D5, and D7, with increasing sub-clustering over time (Fig. 3A ). Marker gene-based cell annotation defined epiblast, formative epiblast, and early primitive streak at D3, late primitive streak, extraembryonic mesoderm, early paraxial mesoderm, and endoderm at D5, and mesenchyme, endoderm, and cardiomyocytes at D7 (Fig. 3A and Supplementary Fig. 9A ). Fig. 3. scRNAseq analysis of mESC in spontaneous differentiation. Open in a new tab A WT E14TG2α mESCs were subjected to spontaneous differentiation in EBs for three, five, and seven days and were analyzed by scRNAseq. Shown are uniform manifold approximation and projections (UMAP) of the differentiation days (upper left) and cell types (right) with proportions of cell types over time (lowe left). B UMAPs of Epcam and the indicated functional DEGs expressed over time in WT E14TG2α mESCs. C , D Co-expression of Epcam with Nanog, Eomes, Gata6, and Foxa2 is depicted at single cell-level in UMAPs ( C ) and in correlation plots with Pearson correlation and p -values ( D ). E Expression levels of Epcam and fDEGs ( n = 18) is shown in violin plots over time (D3, 5, and 7). Epcam was expressed in the epiblast, formative epiblast, early primitive streak, late primitive streak, and endoderm. fDEG estrogen-related receptor beta ( Esrrb ) encodes a self-renewal and pluripotency-associated nuclear receptor co-expressed with Epcam in epiblast cells. Epcam was further co-expressed with pluripotency gene Nanog in epiblast, formative epiblast, early primitive streak and selected late primitive streak cells. In the latter, Epcam and Nanog were co-detected with Brachyury , Mesp1 , and Gata6 . Upon further differentiation to mesodermal cells, Brachyury and Mesp1 were silenced, while Hand1 was strongly induced. The expression of Epcam in endodermal cells was accompanied by Wnt signaling inhibitor Dkk1 in selected cells and by central endodermal reprogramming transcription factor Foxa2 (Fig. 3B ). Co-expression at single cell level was consistently observed between Epcam , Nanog , Eomes , and Foxa2 , while Epcam and Gata6 showed exclusive expression patterns (Fig. 3C, D ). The expression of Epcam and all fDEGs across D3, 5, and 7 demonstrated further association of Epcam with Fgf8 , Fgf4 , Nanog , Uty , Eomes , and Sall3 at D3. Later stages of differentiation (i.e. D5 and 7) were characterized by loss of Epcam and expression of Gata6 , Twist1 , and Hand1 (Fig. 3E and Supplementary Fig. 9B, C ). Insights in signaling functions associated with fDEGs and cellular lineages were obtained upon correlation with PROGENy activity scores. Epcam and fDEGs expressed at early stage D3 were generally associated with stronger pathway activity scores than fDEGs expressed at later time points. Epcam , Nanog , and Eomes particularly strongly correlated with EGFR, MAPK, and VEGF signaling (Fig. 4A , left panel). These associations were corroborated by high positive correlations of Epcam / Nanog / Eomes high cells of the epiblast and primitive streak with EGFR, MAPK, VEGF activities (Fig. 4A , right panel). Further differentiated cells of mesodermal and endodermal lineage were characterized by a sharp downregulation of these signaling pathways, and the induction of alternative pathways including NFκB, TNFα, TGFβ, p53, hypoxia, and Trail signaling. Increasing Wnt signaling was observed starting in late primitive streak cells and increased towards cardiomyocytes (Fig. 4A , right panel). Hence, Epcam , Nanog , and Eomes associated with elevated EGFR, MAPK, and VEGF activities in early mESC differentiation, and with the onset of Wnt signaling in late primitive streak. Fig. 4. PROGENy and pseudotime analysis of mESC in spontaneous differentiation. Open in a new tab A Correlation analysis of Epcam and fDEGs ( n = 18) with pathway activities scores inferred using PROGENy at single cell-level are shown (left heatmap). Correlation analysis of cell types identified in spontaneously differentiating WT E14TG2α mESCs with pathway activities scores inferred using PROGENy at single cell-level are shown (right heatmap). B , C scTour- ( B ) and Diffusion pseudotime-based pseudotime analyses ( C ) of spontaneously differentiating WT E14TG2α mESCs are depicted over time and cell types (root cells: Epiblast). D Expression of Epcam, Nanog, Eomes, Gata6, and Foxa2, and Wnt signaling activity are plotted within the DPT-pseudotime stream of spontaneously differentiating WT E14TG2α mESCs. Color coding marks cell types as in ( C ). Pseudotime trajectories of differentiating mESCs were inferred using scTour and Diffusion Pseudotime (DPT) with epiblast as root cells. While scTour infers time as a latent variable, DPT models diffusion-based cellular transitions from root cells [ 32 , 33 ]. ScTour demonstrated the derivation of Epcam high formative epiblast from epiblast cells, which further progressed into early and late primitive streak. Primitive streak cells separated to form Epcam low/negative early paraxial and extraembryonic mesoderm, and Epcam high endoderm. Mesodermal lineages progressed into mesenchyme and, ultimately, cardiomyocytes (Fig. 4B ). DPT analysis confirmed the sequence of lineage formation and segregation of meso- and endodermal progeny at the primitive streak stage (Fig. 4C ). Gene expression profiles along the DPT-based pseudotime stream revealed strong expression of Epcam , Nanog , and Eomes in epiblast and primitive streak, with elevated Eomes in late primitive streak that was accompanied by consistently increasing Wnt signaling. The latter signal was further enhanced in early mesodermal progenitors and decreased in more differentiated mesenchymal cells and cardiomyocytes, in which the expression of Gata6 was prominent (Fig. 4D ). Hence, scRNAseq analysis confirmed Epcam expression in early cells of spontaneously differentiating mESC and an overlapping pattern with Eomes and Nanog in epiblast and primitive streak. The data further suggest a critical function of Epcam during segregation of meso- and endodermal cells in the primitive streak, which is accompanied by increased Wnt signaling and sharp Eomes induction. Reconstitution of Eomes expression complements EpCAM deficiency Defective cardiomyocyte formation in Epcam knockout mESCs may originate from a malformation of early primitive streak cells or a reduction of endodermal cells required for instructing cardiomyocytic mesodermal precursors [ 34 ]. Early co-expression of Epcam with Eomes and Gata6 in late primitive streak and with Foxa2 in endoderm prompted us to interrogate the capacity of these selected genes to rescue the phenotype of EpCAM −/− clones. Eomes , Gata6 , and Foxa2 were individually re-expressed as EGFP fusions in EpCAM −/− clones #56 and #114 using doxycycline-responsive expression vectors. Gene expression was induced before EB formation and relieved during spontaneous differentiation to mimic expression patterns observed in WT ESCs. Expression of all transgenes was inducible and resulted in correct sub-cellular localizations (Supplementary Fig. 10A, B ). Quantification of EGFP-fusions by flow cytometry confirmed the inducibility of transgenes and revealed leakage of Eomes , and Gata6 to varying degree, with clone #56 showing stronger Eomes leakage than clone #114 (Supplementary Fig. 11 ). In both EpCAM −/− clones, Eomes-EGFP, Foxa2-EGFP, and Gata6-EGFP expression was gradually lost with differentiation (Fig. 5A ). In knockout clone #56, Eomes-EGFP leakage in the absence of doxycycline at the time point of EB formation complemented cardiomyocyte formation to 47% of WT. Enhanced expression of Eomes-EGFP upon doxycycline addition greatly restored WT concentration rates in clone #56 at D14 (77% of WT). Weak Eomes-EGFP leakage in clone #114 did not influence contraction rates in the absence of doxycycline but was restorable to 62% of WT after doxycycline-mediated induction (Fig. 5B ). α-CAA expression levels were used as a molecular surrogate for the observed functional complementation and confirmed contraction rates observed after Eomes-EGFP re-expression in EpCAM −/− clones (Fig. 5C ). Single Foxa2-EGFP re-expression complemented cardiomyocyte formation only in EpCAM −/− clone #56 but not #114, and single Gata6 re-expression had no measurable effects on the contraction of any EpCAM −/− clone (Fig. 5D, E ). Consequently, selective re-expression of Eomes was observed after MEK inhibition and EpCAM re-expression (Fig. 5F, G ). Thus, Epcam and Eomes re-expression in EpCAM −/− clones complemented the deficiency in exit of pluripotency and spontaneous differentiation to cardiomyocytes. Fig. 5. Temporal re-expression of Eomes complements knockout defects in 3D spontaneous differentiation. Open in a new tab A E14TG2α EpCAM −/− mESCs (clones #56 and #114) were stably transfected with conditional, doxycyclin-responsive vectors for Eomes , Foxa2 , and Gata6 in fusion with enhanced green fluorescence protein (EGFP). Transgene expression was induced via doxycyclin in 2D-culture and EBs were generated. Upon doxycycline removal, transgene expression was assessed upon flow cytometry quantification of EGFP. Shown are mean with SD from n = 3 independent experiments. * p -value < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001. B , D , E Contraction rates of EBs from WT, EpCAM −/− ESCs (clones #56 and #114), and Eomes ( B ), and Foxa2 ( D ) and Gata6 ( E ) re-expression clones at D14 of spontaneous differentiation are shown as mean with SD from n = 3 independent experiments including at least six EBs per experiment and treatment. Where indicated, EpCAM −/− ESCs (#56 and #114) re-expression clones were pre-treated with doxycycline (dox). C Expression of α-CAA was quantified by qRT-PCR at day 14 of spontaneous differentiation in EpCAM −/− ESCs upon conditional re-expression of Eomes . Shown are mean with SD from n = 3 independent experiments performed in triplicates. * p -value < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001. F , G Eomes expression was quantified by qRT-PCR at day seven of spontaneous differentiation in EpCAM −/− ESCs (clones #56 and #114) treated with Mek inhibitor (Meki) ( F ) and in sub-clones re-expressing Epcam ( G ). Shown are mean with SD from n = 3 independent experiments performed in triplicates. * p -value < 0.05, ** < 0.01, *** < 0.001, n.s. not significant. Epcam induces Eomes transcription during differentiation Eomes is a T-box TF orchestrating exit from pluripotency and lineage specification towards mesoderm and definitive endoderm. Thereby, Eomes represents a central regulator of early embryonic development that links early developmental signaling to germ layer specification [ 35 ]. We aimed at understanding underlying regulatory mechanisms of Eomes as a key mediator of Epcam -dependent effects on differentiation. We focused on Wnt signaling that was affected by Epcam knockout and sharply increased at the transition from early to late primitive streak cells in scRNAseq data. To interrogate regulatory effects of Wnt signaling and EpCAM on Eomes transcription, EpCAM −/− clone #114 bearing a conditional Epcam-EGFP expression plasmid was stably transfected with a reporter plasmid consisting of the Eomes promoter driving the expression of mCherry as detectable marker. The resulting cell line was subjected to 2D spontaneous differentiation via LIF withdrawal and was monitored by flow cytometry using forward scatters of area and height (FSC-A vs. FSC-H). Differentiation resulted in characteristic morphological changes including elongated, spindle-shaped cells and in the reduction of pluripotency markers Nanog and Oct3/4 (Fig. 6A, B ). Over time, proportions of differentiated cells increased along with mCherry-positive cells reflecting Eomes promoter activity (Fig. 6C, D ). mCherry-positive cells represented 16.99% of all cells at day 4 and 85% were allocated to differentiated cells (Fig. 6D ). The influence of Wnt signaling on differentiation and Eomes transcription was tested using Wnt activator CHIR99021. Wnt activation promoted mESC differentiation, affecting 40.61% of cells versus 3.96% in the absence of CHIR99021 (Fig. 6E ). Eomes promoter activity was induced in parallel to cell differentiation with 10.24% of total positive cells and 78.9% of expression in differentiated cells (Fig. 6F , Supplementary Fig. 12 ). Fig. 6. Influence of Wnt signaling on differentiation and Eomes promoter activity. Open in a new tab A E14TG2α EpCAM−/− ESC clone #114 was stably transfected with a conditional Epcam expression vector and an Eomes promoter-reporter plasmid expressing mCherry fluorescence protein. ESCs were subjected to 2D spontaneous differentiation via LIF withdrawal. Micrographs of E14TG2α EpCAM−/− ESC morphology) in the presence (pluripotent) and absence (differentiated) of LIF in 2D culture are shown. B Pluripotent (Pluri) and differentiated (Diff) mESC were enriched by FACS and the expression of Nanog, Oct3/4, and Sox2 was analyzed by qRT-PCR. Shown are mean and SD of n = 3 independent experiments performed in triplicates (** < 0.01). C Proportions of pluripotent and differentiated cells are shown as pie charts with mean percentages of cells in pluripotency, differentiation, and in an undefined state from n = 3 independent experiments at D0, 1, and 4 of differentiation in 2D. D Quantification of Eomes promoter activity as mCherry-positive cells and mean fluorescence intensity in pluripotent and differentiated E14TG2α EpCAM−/− cells at D4 from n = 3 independent experiments. Shown are mean with SD. * p -value < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001. E E14TG2α EpCAM−/− ESCs harboring a conditional Epcam expression vector and an Eomes promoter-reporter expressing mCherry were subjected to 2D spontaneous differentiation via LIF withdrawal in the presence or absence of Wnt activator CHIR99021 (CHIR). Differentiation was assessed by flow cytometry using forward scatters of area and height (FSC-A vs. FSC-H). and proportions of cells are given as pie charts ( n = 3 independent experiments. F Quantification of Eomes promoter activity as percentage of mCherry-positive cells and mean fluorescence intensity in control- and CHIR99021-treated cells (CHIR- and CHIR + ) from n = 3 independent experiments. Shown are mean with SD. One-way ANOVA with post-hoc multiple comparisons: p -value ** < 0.01, *** < 0.001, **** < 0.0001. Next, EpCAM-EGFP expression was induced via doxycycline in EpCAM −/− clone #114 equipped with an Eomes -mCherry reporter without LIF for 24 h in the absence and presence of Wnt inhibitor XAV939 (Fig. 7A ). Following a measurable induction of EpCAM-EGFP expression at day 1, EpCAM-EGFP was lost over time and was not influenced by XAV939 (Fig. 7B ). No significant differences in pluripotent versus differentiated cell percentages were observed at day 1 based on treatment (average range of 6.30-6.62% of differentiated cells) (Fig. 7C ) and the Eomes promoter was inactive (Supplementary Fig. 13A, B ). Spontaneous differentiation of EpCAM −/− clone #114 in any cell lineage was increased at D4 to an average 15.4% and was further supported by EpCAM-EGFP re-expression (20.33%). XAV939 had no additional effect, yielding proportions of differentiated cells comparable to doxycycline-treated samples (20.86%) (Fig. 7C ). Specifically monitoring Eomes promoter activity revealed a significant up-regulation upon re-expression of EpCAM-EGFP in pluripotent and, more pronouncedly, in differentiated cells (33.4% positive cells). EpCAM-induced upregulation of the Eomes promoter was partially counteracted upon Wnt inhibition and was reduced to 25% (Fig. 7D , Supplementary Fig. 13C, D ). We conclude that re-expression of EpCAM in ESC re-activated Eomes transcription, partially via Wnt-dependent pathways. Fig. 7. Epcam influence on Eomes expression. Open in a new tab A E14TG2α Epcam −/− ESCs (clone #114) harboring a conditional EpCAM-EGFP expression vector and an Eomes promoter-reporter expressing mCherry were subjected to 2D spontaneous differentiation via LIF withdrawal in the presence or absence of doxycycline for 24 h to induce EpCAM-EGFP expression and of Wnt inhibitor XAV939. B EpCAM-EGFP expression was monitored by flow cytometry at D1 and 4. Shown are representative histograms from n = 3 independent experiments with cells without doxycycline (gray line), with doxycycline for 24 h (light green), and with doxycycline and XAV939 (dark green). C E14TG2α Epcam −/− ESCs differentiation was assessed by flow cytometry using forward scatters of area and height (FSC-A vs. FSC-H) at D1 and 4 from n = 3 independent experiments. Proportions of pluripotent and differentiated cells are given as pie charts. D mCherry expression was assessed by flow cytometry in pluripotent and differentiated ESCs. Quantification of Eomes promoter activity as mCherry-positive cells (left) and mean fluorescence intensity (right) in pluripotent and differentiated E14TG2α EpCAM −/− ESCs (clone #114) from n = 3 independent experiments. Dox-: control without doxycycline, Dox + : doxycycline treatment (24 h), Dox+ XAV939: doxycycline and XAV939 treatment (24 h). Shown are mean with SD. One-way ANOVA with post-hoc multiple comparisons: * p -value < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001. Discussion During embryonic development, mesodermal lineages efficiently suppress Epcam expression at the onset of gastrulation in the primitive streak, while endodermal lineages express EpCAM [ 14 , 15 , 36 ]. Mechanisms involved in this selective expression were partly uncovered in hESCs, where EpCAM − /CD56 + early mesodermal progenitors are the foundation of all subsequent mesodermal lineages [ 36 ]. Oppositely, B-cell receptor-associated protein 31 BAP31 colocalized with and up-held EpCAM expression along with a state of proliferation and repression of differentiation of hESCs [ 37 ]. In concordance with our present findings, human iPSC down-regulate EpCAM expression after primitive streak induction at the bifurcation between endo- and mesodermal lineages, resulting in EpCAM-positive endoderm and EpCAM-negative mesoderm [ 14 ]. Mechanistically, Epcam gene repression was attributed to histone modifications, primarily trimethylation of lysine 27 in the promoter region by chromatin remodelers SUZ12 and JMJD3 [ 19 ]. The reduction of EpCAM protein during murine guided mesodermal ESC differentiation and in human carcinoma cells undergoing EMT is mediated by enhanced endocytosis and reduced membrane recycling via binding to differential Rab proteins [ 25 ]. Hence, regulatory mechanisms are implemented at various levels governing EpCAM expression and in multiple species including human, murine, porcine, xenopus, and zebrafish cells [ 9 , 10 , 15 , 21 , 26 , 38 – 43 ]. Downstream effects of EpCAM in differentiation are less well understood. Early studies in mESCs suggested a linkage of EpCAM to the expression of c-Myc, Sox2, Oct-4, and Stat3, a reduction of proliferation and the induction of endo- and mesodermal markers upon silencing of EpCAM [ 9 , 10 ]. The binding of the intracellular signaling domain of EpCAM (EpICD) to regulatory DNA elements of c-Myc, Oct-4, Nanog, Sox2, and Klf4 is enhanced under pluripotency and is gradually lost during differentiation of human ESCs [ 19 ]. In the present study, effects of a genetic knockout of EpCAM under pluripotency conditions were minor and did not support a major function in maintenance of pluripotency, except for a down-regulation of c-Myc . Activation of LIF receptor under pluripotency conditions induces Stat3, which plays a central role in maintenance of pluripotency by activating Nanog expression and by transcriptional cooperation with Nanog, Sox2, and Oct4 [ 44 – 46 ]. It is conceivable that such regulatory circuit overpowers EpCAM-mediated effects under pluripotency conditions and therefore masks effects of the genetic loss of Epcam (Fig. 8 ). Fig. 8. Schematic representation of molecular interactions of EpCAM and Wnt signaling Components in mESC under pluripotency and differentiation. Open in a new tab Shown is a schematic representation of the signaling interplay of EpCAM and Wnt signaling in pluripotency and differentiation of murine embryonic stem cells. Interactions include ligand-receptor pairs invovlved in downstream cascades leading to Stat, β-Catenin, Nanog, Sox2, Oct4 and the regulation of a pluripotent state or, alternatively, exit from pluripotency and induction of lineage specificity. Interactions are visualized in the presence and absence of EpCAM (wildtype: WT; CRISPR-Cas9-medicated Epcam knockout: EpCAM-KO). Augmented temporal resolution of transcriptomic changes in mESC differentiation revealed three major phases. Early and late pluripotency occur within the initial 24-48 h of differentiation in EBs, respectively, and are followed by molecular features of primitive streak formation (72 h) and subsequent cell specialization (96 h and beyond) [ 5 ]. Our time-resolved bulk and scRNA sequencing analyses demonstrated a role of Epcam in exit from pluripotency and progression of cell differentiation via linkage to several key transcription factors, receptors, and ligands. Exit from pluripotency was accompanied by an initial increase in Epcam expression, followed by a sharp loss, reminiscent of human iPSCs [ 14 ]. Primitive streak formation at 72 h is characterized by an upregulation of Eomes , Brachyury ( T ), Mixl1 , and Evx1 , coinciding with peaking co-expression of Epcam . Upon LIF withdrawal, effects of increased EpCAM on Wnt signaling may become relevant in the repression of key stemness genes such as Nanog and the induction of differentiation-regulating transcription factors (Fig. 8 ). In Zebrafish EpCAM serves as a scavenger for Kremen1-DKK2 to de-repress Wnt signaling in liver progenitors [ 47 ]. thus supporting its role in the hepatic endodermal lineage [ 14 ] via Wnt regulation. EpCAM interaction with Kremen1 and DKKs may further impact on ESC survival via Wnt-independent mechanisms, since Kremen1 reportedly acts as apoptosis-inducing receptor [ 48 ]. Expression of key primitive streak formation and gastrulation transcription factors Brachyury , Eomes , Mixl1 , and Twist1 , as well as of central regulators of cell fate ( Nanog , Dkk1 , Foxa2 , Gata6 , Hand1 , Mesp1 , and Sall3 ) was significantly reduced in EpCAM −/− ESC. Thus, loss of EpCAM was prominently linked to a concurrent lack of induction of numerous transcriptional regulators of primitive streak formation such as Eomes , Brachyury , Mixl1 [ 49 – 53 ]. endo- and mesoderm ( Foxa2 , Gata6 ) [ 54 , 55 ]. and cardiomyocytic differentiation ( Mesp1 ) [ 56 , 57 ]. scRNAseq analysis of wildtype mESC corroborated such a role of Epcam in the regulation of Eomes at the transition of early to late primitive streak via Wnt activation. Pseudotime analysis of differentiating mESCs validated the postulated Epcam /Wnt/ Eomes axis in cardiomyocyte formation at single cell-level and its implication in gastrulation. The reduction of markers for cardiomyocyte formation such as α-Caa in Epcam −/− cells most likely represents a downstream consequence of the impairment of this Epcam /Wnt/ Eomes axis. Eomes is a pivotal regulator of early mesoderm and cardiac progenitor formation and operates upstream of the cardiogenic cascade comprising Eomes , Mesp1 , Mixl1 , Nkx2-5 , and sarcomeric genes including α-Ca a [ 35 , 53 , 58 ]. Consistent with this framework, loss of Epcam lowers Eomes , which attenuates the activation of the Mesp1 / Mixl1 module and the subsequent Nkx2-5 -dependent program, ultimately resulting in decreased expression of cardiac structural genes such as α-Caa . The lack of TF induction in Epcam −/− clones was paralleled by increased expression of cycline-dependent kinase inhibitor Cdkn2a that encodes cell cycle inhibitors p16(INK4A) and p14(ARF). p16-dependent senescence was demonstrated to limit cellular plasticity during the generation of induced pluripotent stem cells and totipotent cells [ 59 , 60 ]. Hence, Cdkn2a upregulation may further reduce exit of pluripotency and limit differentiation observed in EpCAM −/− mESC. Conditional expression of EpCAM complemented defects in cardiomyocyte formation, provided it mimicked its natural expression dynamic of a strong initial expression followed by sharp loss. Re-expression of EpCAM restored Eomes expression and the contraction capacity of EBs close to wildtype levels. Co-expression of Epcam and Eomes in early stages of murine embryogenesis and at early stages of spontaneous differentiation in EBs corroborated an interdependency of both genes at single-cell level and in an inferred pseudotime stream of differentiation. Wnt activators and inhibitors provided further evidence for a role of Wnt signaling in the induction of differentiation, EpCAM-mediated Eomes transactivation, and cardiomyocyte formation. These findings are in line with a reported function of Wnt-mediated Eomes regulation during cardiac induction in hESCs, which follows a dynamic expression during ESC differentiation comparable to Epcam . Wnt-driven Eomes expression is initially required to suppress the mesoderm repressor Sox2 and initiate mesoderm formation. Thereafter, inhibition of Wnt activity is primordial to allow further lineage specification to cardiomyocytes through the restriction of cardiac repressors Msx1 and Cdx2 [ 61 ]. Consequently, Epcam −/− ESC showed enhanced Cdx2 expression at later time points of spontaneous differentiation (D7), supporting the model of an Epcam / Wnt / Eomes -axis promoting mESC differentiation to cardiomyocytes. Accordingly, the knockout of Epcam recapitulates an Eomes deletion. ESCs harboring an Eomes deletion retain their characteristic morphology, express pluripotency markers Nanog, Oct4/4, and Sox2, but fail to form mesodermal progeny and are incapable of generating cardiac mesoderm and definitive endoderm [ 53 , 58 ]. Hence, lack of Epcam and Eomes , both, negatively impact on the ability of ESCs to exit pluripotency. Effects of EpCAM on differentiation via additional signaling pathways is likely, given the broad range of reported activities of EpCAM. Particularly, a regulation of the MAPK/Erk1/2 pathway by EpCAM has been demonstrated in tumors and in stem cells [ 18 , 22 , 23 , 62 , 63 ]. It is therefore worth mentioning that, despite clear reports of Wnt activation via MEK inhibitors [ 27 , 28 ], a contribution of Erk1/2 to the complementation of EpCAM −/− ESC cannot be excluded. Such scenario suggests potential relationships between the newly identified Epcam /Wnt/ Eomes axis and additional EpCAM-dependent cellular signals (reviewed in [ 26 ]). Association of Epcam and fDEGs with elevated EGFR and MAPK signaling at stages of primitive streak formation is in line with the proposed role of EpEX in inducing EGFR signaling in iPSCs [ 17 , 18 ]. Further EpCAM-dependent regulatory mechanisms comprise signaling pathways involving RIP of EpCAM, interaction with embryonic Ras (ERas), E-cadherin and claudins [ 15 , 17 – 19 , 21 , 42 , 47 ]. Both RIP products and ERas-mediated signaling may therefore be instrumental in the Epcam /Wnt/ Eomes axis and correlated PROGENy pathways. In summary, we provide to our knowledge the first evidence for a function of EpCAM in exit of pluripotency via Wnt-mediated regulation of Eomes . Limitations of the study Functional and transcriptomic analyses of the genetic knockout of Epcam presented in this study are limited to a 3D model of differentiation. Differences to murine embryogenesis in vivo can therefore not be excluded. This limitation has been partly overcome using public and in-house scRNAseq datasets of murine embryogenesis and differentiation, demonstrating correlations of Epcam expression with relevant TFs at the single-cell level. Future work should therefore aim at a validation of the presented Epcam/Wnt/Eomes axis in vivo. Materials and Methods Biological and technical replicates, statistical analysis Throughout the manuscript, biological replicate is referred to as a fully independent experiment performed with newly generated materials, whereas a technical replicate is a repeated measurement with identical materials. Samples sizes are indicated for each experiment in the cognate figure legends and comprise a minimum of n = 3 biological replicates. Bulk and scRNAseq experiments were generated in quadruplicates. Sample exclusion was performed for RNA sequencing samples based on obvious divergence in transcriptomic profiles ( n = 1000 most differential genes) and was applied for one sample of bulk seq data of wild-type cells at D3. Statistical analysis of comparisons of two samples with similar variances within samples was performed with an unparied student t-test where a p -value ≤ 0.05 was considered significant. Statistical analysis of comparisons of multiple samples was performed with a one-way ANOVA and post-hoc multiple testing with Tuckey or Bonferroni corrections. P -values ≤ 0.05 were considered significant. Cell lines, cell culture and stable cell transfection Mouse embryonic stem cells (ESC) E14TG2 α were cultured in Mouse ES Cell Basal Medium (ATCC, Manassas, Virginia, USA), 2-mercaptoethanol (Gibco™; 10 nM; Paisley, Scotland) and leukemia inhibitory factor (ESGRO®LIF; 1000 U/mL; Becton Dickinson, Heidelberg, Germany) supplemented with 10% FBS (Bio&SELL, Feucht, Germany) on 0.1% gelatin (InSCREENeX, Braunschweig, Germany) coated T25 flasks. E14TG2 α EpCAM knockout #56 and #114 are independent single-cell clones generated using guide-RNAs targeting exon 2 and exon 4, respectively [ 15 ]. Clone #56 has a 13 bp deletion leading to a truncated putative protein of 35 aa. Clone #114 carries a 13 bp deletion putatively generating a 134 aa product. Both putative protein products miss the transmembrane domain and would therefore not be anchored in the plasma membrane to deploy their function. EpCAM-specific antibodies target the very N-terminus and failed to detect any protein product. EpCAM knockout clones were transfected with mEpCAM-EGFP, mEomes-EGFP, mFoxa2-EGFP, mGata6-EGFP in a PiggyBac expression backbone driven by a doxycycline-responsive promoter with a blasticidin selectable marker. This vector was derived from pFH2.94_PB-TRE3GS-Gag-MCP-T2A-eGFP-PGK-Tet3G-Blast (Addgene #205540) by replacing the Gag-MCP-T2A-eGFP cassette with each gene of interest. The vector ensures stable genomic integration and long-term maintenance of inducible transgene expression [ 64 ]. Consistent results from Western blot, flow cytometry, and immunofluorescence analyses performed at multiple time points confirmed stable transgene expression without variation in protein size, localization, or induction efficiency. Cells were selected and cultivated with 12.5 ng/μl blasticidin (Invivogen, San Diego, USA). Cond.EpCAM-#114 cells were transfected in a 6-well plate with mEomes promoter reporter-mCherry in the pEZX-PM02 backbone containing a puromycin selectable marker and then continuously selected with 2.5 ug/ml puromycin (Invivogen, San Diego, USA) and 12.5 ng/ul blasticidin. Spontaneous and directed 2D differentiation ESC (100,000 cells/well) were seeded in 6-well plates in medium w/o LIF, 37 °C, 5% CO 2 . For Wnt activation and inhibition, 30 μM CHIR 99021 (Sigma, St.Louis, MO, USA) and 10 μM XAV-939 (Sigma, St.Louis, MO, USA) were supplemented for 24 h, respectively. Cells were washed with PBS and continuously differentiated in medium w/o LIF. At time points day 1, 4, and 7, cells were harvested for further analysis. Embryoid bodies (EBs) formation and contraction To generate EBs, 500 cells in 20 µl of LIF-free medium were plated on the lid of a 150 mm tissue culture dish and cultured as hanging drops by inverting the lid. After three days, EBs were transferred to ultra-low attachment plates (Nunc, Wiesbaden, Germany) for four days before transfer to standard 96-well plates for further differentiation up to 14 days. During this period, 100 ul of fresh culture medium was replaced every two days. Contraction of EBs was analyzed after 10–14 days by counting under a microscope in 96-well plates. Contraction rate represents the percentage of contracting EBs standardized to the number of EBs formed. Treatment with Wnt11 and Wnt-regulating compounds, cryo-preservation and sectioning EpCAM-KO EBs were allowed to differentiate spontaneously for 14 days and were treated with chemicals and proteins for the initial six days, including 1 µM MEK inhibitor AZD6244 (Selleckchem, Munich, Germany), 25 nM GSK3β inhibitor CHIR99021, 50 nM DKK1 inhibitor WAY (Sigma-Aldrich, Steinheim, Germany), and 0.75 ug/ml Wnt11 (Bio-techne, Minnesota, USA) before assessing contraction rates at D14. EBs from D3, D5, D7, D14 were embedded in tissue-tek (Sakura Finetek, Germany), snap-frozen in liquid nitrogen, cut into 4 μm thick sections, and stored at −20 °C. Immunohistochemistry (IHC) and Immunofluorescence (IF) Immunohistochemistry and immunofluorescence staining of EpCAM (BD552370, 1:500, Abcam, Waltham, USA), α-CAA (AC1-2042, 1:200, Sigma, Germany), Foxa2 (#8186, 1:400, Cell Signaling, Leiden, The Netherlands), Eomes (ab216870, 1:200, Abcam, Waltham, USA), Gata6 (55435-1-AP, 1:500, Proteintech®, Planneg-Martinsried, Germany) was performed using avidin-biotin-peroxidase complex method (Vectastain, Vector laboratories, Burlingma, CA, USA) and Alexa Fluor-488- and Alexa Fluor-647-conjugated secondary antibodies (Invitrogen, Thermo Fisher, Darmstadt, Germany) respectively. Confocal microscopy imaging was conducted with a TCS-SP8 scanning system and a DM-IRB inverted microscope (Leica, Nussloch, Germany). Flow cytometry (FACS) Single cell suspension of EBs was generated by treatment with 1 mL Accutase (Capricorn-scientific, Hesse, Germany) (1 min, RT), neutralization with culture medium, washing with PBS, and filtering with a 40 µm filter (Pluriselect, Leipzig, Germany). Cells were stained with EpCAM-specific antibody (#14-5791-81, Invitrogen, Waltham, USA) for 15 min on ice, washed three times in PBS-3%FBS, and stained with fluorescein isothiocyanate-conjugated rabbit anti-mouse secondary antibody. Measurement of cell surface expression of EpCAM was performed in a CytoFLEX device (Beckman Coulter, Krefeld, Germany). Immunoblotting Immunoblotting of EpCAM, Eomes, Foxa2, Gata6 and β-actin (Santa cruz biotechnology, Dallas, USA) was performed with 10–50 µg whole cell lysate (PBS, 1% triton X-100, Roche complete protease inhibitors) separated in a 10-15% SDS-PAGE, and transferred on PVDF membranes (Millipore, Germany). Primary antibodies were detected with HRP-conjugated secondary antibodies and ECL reagent (Merk Millipore, Darmstadt, Germany). All uncropped immunoblot are available as Supplementary Figure_immunoblots . Quantitative reverse transcription polymerase chain reaction (RT-qPCR) Total mRNA was prepared using RNeasy Mini Kit (Qiagen, Hilden, Germany) and reverse transcribed with QuantiTect Reverse Transcription-Kit (Qiagen, Hilden, Germany). cDNA was amplified using SYBR-Green PCR mastermix (Qiagen, Hilden, Germany) and specific primers. Normalization across samples were performed using glucuronidase beta (GUSB) expression. Gene expression levels were calculated according to the equation 2-ΔΔCT, where ΔCT was defined as the difference between CT of interest genes and CT of control genes. The following primers were used for qRT-PCR (forward/backward): GUSB: FW CAACCTCTGGTGGCCTTACC; GGGTGTAGTAGTCAGTCACAGAC Eomes: GGCCTACCAAAACACGGATATC; TTTCTGAAGCCGTGTACATGGA Foxa2: CCCTACGCCAACATGAACTCG; GTTCTGCCGGTAGAAAGGGA Gata6: GGTCTCTACAGCAAGATGAATGG; TGGCACAGGACAGTCCAAG Twist1: CCCACCCCACTTTTTGACGA; CAGTGGCTGATTGGCAAGAC T: GCTTCAAGGAGCTAACTAACGAG; CGTCACGAAGTCCAGCAAGA Fgf10: TTTGGTGTCTTCGTTCCCTGT; TAGCTCCGCACATGCCTTC Mixl1: CTACCCGAGTCCAGGATCCA; ACTCCCCGCCTTGAGGATAA DKK1: TGAGGGCGGGAACAAGTA; TTCGGCAAGCCAGACAGA Mesp1: GCTCGGTCCCCGTTTAAGC; ACGATGGGTCCCACGATTCT α-CAA:CTGGATTCTGGCGATGGTGTA; CGGACAATTTCACGTTCAGCA Wnt11: GGATATCCGGCCTGTGAAGG; TCCACCACTCTGTCCGTGTA Fgf8: TGTTGCACTTGCTGGTTCTC; CGGCTGTAGAGCTGGTAGG Hand1: GTGAGTGCATCCCCAATGTG; GCCAGCACGTCCATCAAGTA Cdx 2 : CTGCTGTAGGCGGAATGTATGTCT; AAGGCTTGTTTGGCTCGTTACAC Esrrb : CGATTCATGAAATGCCTCAA; CCTCCTCGAACTCGGTCA Fgf4 : ACTACCTGCTGGGCCTCAA; ACTCCGAAGATGCTCACCAC Uty : AGATGAAGACGCTGTTGAAC; CTAATTGCCCACTGAAATGC Sall3 : CCTGATTCTTCCTGGTGGAGT; CTCTGGAAAACGCCACAGAC Nano g: TCTTCCTGGTCCCCACAGTTT; GCAAGAATAGTTCTCGGGATGAA Cloning procedures PCR amplification of the inserted gene was conducted using Q5® High-Fidelity DNA Polymerase (New England Biolabs, Frankfurt, Germany) in thermal cycler (BioER, Bremgarten, Switzerland). DNA template was purchased from Addgene as bacterial stabs including Tet3G-on-vector (#205540), mEpCAM, mEomes (#200888), mFoxa2(#33014) and mGata6 (#72694) inserts. NEBuilder® HiFi DNA Assembly (New England Biolabs, Frankfurt, Germany) was used to ligate inserted coding sequences and vectors. The following primers were used for cloning (forward/backward): mEpCAM: TCGTAAAGCTAGCGGATCCGCCACCATGGCGGGTCCCCAGGCCCT GCCGCCGCCGCCGCCGCCGGCATTAAGCTCTCTGT mEomes: TCGTAAAGCTAGCGGATCCGCCACCATGCAGTTGGGAGAGCAGCT GCCGCCGCCGCCGCCGCCGGGACTTGTGTAAAAAGCA mFoxa2: TCGTAAAGCTAGCGGATCCGCCACCATGCTGGGAGCCGTGAAGAT GCCGCCGCCGCCGCCGCCGGATGAGTTCATAATAGGCC mGata6: TCGTAAAGCTAGCGGATCCGCCACCATGGCCTTGACTGAG GCCGCCGCCGCCGCCGCCCTGTTCTCGGGGTTGGCG RNAseq and differential expression (DE) analysis RNA was extracted using the RNeasy Mini Kit (Qiagen, Germany) and quantified using Qubit™ RNA BR Assay Kit (# Q10210 ) with a Qubit Fluorometer (Thermo Fisher Scientific, MA, USA). RNA sequencing libraries were prepared with 100 ng total RNA input using the QuantSeq 3′ mRNA-Seq Library Prep Kit FWD for Illumina (#SKU:015.96; Lexogen, Austria). For library amplification, PCR cycles were determined with the PCR Add-on Kit for Illumina (#SKU:020.96, Lexogen) and individual libraries were amplified with 18 PCR cycles. Quality and quantity of the libraries were evaluated using Quanti-iT PicoGreen dsDNA Assay Kit (P7589, Thermo Fisher) and Bioanalyzer High Sensitivity DNA Analysis Kit (#5067-4626, Agilent Technologies). Libraries were sequenced with 150 bp paired-end mode on a HiSeq400 sequencer (Illumina, Germany). 3´-Bulk RNAseq data are available as GSE 293231 at GEO ( https://www.ncbi.nlm.nih.gov/geo/ ). Differential expression (DE) analysis of bulk RNA-seq data was performed using DESeq2 [ 65 ]. Wild-type (WT) cells served as controls and were compared to Epcam knockout lines (#114 and #56) at D0, D3, D7, and D10. DE genes (DEGs) were identified based on thresholds of |log2 fold change | > 1.0 and FDR ≤ 0.05. Volcano plots represents the up and down-regulated DEGs of differentiation stages between Epcam knockouts vs. WT. Gene set enrichment analysis (GSEA) was conducted on genes ranked by fold change and visualized with clusterProfiler package (Bioconductor) in R. Enriched gene sets were cross-referenced with canonical pathways (CP) from the Molecular Signatures Database (MSigDB; https://www.gsea-msigdb.org/gsea/msigdb ), focusing on pluripotency, mesoendoderm differentiation, mesoderm and ectoderm differentiation, heart development, cardiac contraction, and Wnt signaling pathways. Signature genes were identified by intersecting the gene list from DEGs, GSEA, and MSigDB hallmark pathways of differentiation stages. Venn diagrams were used to visualize common genes across comparisons. Public single cell RNA-sequencing dataset analysis Single-cell RNA-sequencing datasets are available at GSE100597 (GEO) and https://github.com/MarioniLab/EmbryoTimecourse2018 , respectively. GSE100597 contains scRNA-seq data from mouse embryos at gestation days E3.5 ( n = 10 embryos, n = 99 cells), E4.5 ( n = 5 embryos, n = 105 cells), E5.5 ( n = 9 embryos, n = 267 cells), and E6.5 ( n = 11 embryos, n = 250 cells) [ 29 ]. The dataset from Pijuan-Sala et al. contains scRNA-seq data 116,312 cells from gestation day E6.5-E8.5 [ 30 ]. Data was visualized with uniform manifold approximation and projection (UMAP) in R ( https://www.r-project.org/ , version R 4.3.0). Publicly available preprocessed single-cell RNA sequencing data were integrated to explore lineage and regulatory processes across species and developmental stages. Single-cell data from C57Bl/6Babr mouse embryos ( GSE100597 ) at stages E3.5, E4.5, E5.5, and E6.5 were processed with Seurat package in R for data normalization, log2 transformation, and scaling. Embryonic developmental cell lineages score was calculated using the AddModuleScore function in Seurat [ 66 ]. Additionally, single cell transcriptional profiles of 116,312 cells from C57BL/6 mice embryos (E6.5 to E9.5) [ 30 ].were analyzed using the Scanpy package in Python [ 67 ]. Preprocessed human embryogenesis ( GSE157329 ) single-cell data from embryos at Carnegie stages (CS) 12–16 (4–6 weeks) and across 18 developmental systems based on over 180,000 transcriptomes [ 68 ]. were analyzed using Scanpy. Single cell datasets are visualized for embryonic stages and identified cell lineages. Expression patterns of signature genes were mapped across developmental stages and cell lineages, followed by analysis of co-expressed signature genes. Single cell RNA capture, library prep, and sequencing Single cell RNA sequencing was performed with the BD Rhapsody HT single cell analysis system (BD Biosciences, Heidelberg, Germany) using adapted protocols. Wildtype murine embryonic stem cells E14TG2α were harvested in four individual consecutive passages to form embryoid bodies for spontaneous differentiation in the absence of LIF for three, five, and seven days (D3, D5, D7) generating a total of n = 12 samples. EBs of each biological replicate were collected in separate reservoir tanks, transferred to 50 mL Falcon tubes, and sedimented via gravity before carefully removing supernatant. EBs were resuspended in 2 mL of TrypLE solution and incubated with constant mixing at 37 °C for 1–2 min to generate single cell suspensions. Cells from each replicate and time point were separately centrifuged 5 min at 400 g, washed once with PBS with 1% FBS (FACS buffer). From here on, all steps were conducted on ice. Custom BD TM Single Cell Multiplexing Kit (#626545; anti-mouse MHC-calls I antibody, direct labelling) was used for sample tagging and pooled processing according to the manufacturer´s protocol. Labelled cells from four replicates (each 2,5 × 10 5 cells) were pooled for each individual time point to three samples (D3, D5, D7) and washed twice in 2 mL of FACS buffer, centrifuged at 400 g for 5 min, and finally resuspended in 620 μL cold sample buffer. Staining cells with viability markers was achieved by adding 3.1 μL of 2 mM Calcein AM and 3.1 μL of 0.3 mM DRAQ7™ to the cell suspension in cold sample buffer (37 °C in the dark, 5 min). 10 µL cells were counted using BD Rhapsody scanner 10 μL in an INCYTO™ disposable hemocytometer with the remaining cells kept on ice and protected from light. Then, BD Rhapsody 8 lane cartridge were primed, loaded with labelled cells and, thereafter, with BD Rhapsody™ Enhanced Cell Capture Beads. After cell lysis and release of polyadenylated mRNAs, BD Rhapsody™ Enhanced Cell Capture Beads were retrieved, washed, and processed for library preparation. Reverse transcription was performed with the BD Rhapsody TM cDNA kit (#633773) and library preparation with the BD Rhapsody TM Whole Transcriptome Analysis Amplification kit (#633801), generating Illumina-compatible libraries. cDNA library preparation and sequencing as paired-end reads were performed by the group of Dr. Inti Alberto De La Rosa Velásquez of the Core Facility Genomics (Helmholtz Center Munich, Germany) on an Illumina NovaSeq X Plus sequencer in a 10B200 flow cell. The scRNAseq dataset is available under GSE318465 at the Gene Expression Omnibus ( https://www.ncbi.nlm.nih.gov/geo/ ). In-house scRNAseq data analysis In-house scRNAseq data were processed using the Seurat package in R. Low-quality cells and doublets were removed through standard quality control process. Subsequently, 2000 highly variable genes were identified for downstream analysis. Harmony analysis was employed to remove batch effects. Cell clustering analysis was performed using FindClusters and FindNeighbors functions. Pearson correlation analysis was conducted to estimate the association between selected genes. PROGENy [ 69 ]. was used to assess pathway activities based on perturbation-derived gene signatures at the single-cell level and subsequently compared across cell populations or between different genes. Differential pathway activities were analyzed to depict signaling dynamics during cellular state transitions and the associations of individual genes with signaling pathways. Diffusion pseudotime (DPT) analysis was performed to infer cellular trajectories based on diffusion maps constructed from the normalized single-cell expression data [ 32 ]. Epiblast cell population were specified as the root for pseudotime calculation, and cells were ordered along the inferred developmental trajectory. scTour was performed to construct cellular trajectories by learning latent temporal representations from the normalized single-cell expression data [ 33 ]. Supplementary information Original Data (1.3MB, pdf) Supplementary Figures (9.9MB, pdf) Supplementary Table 1 (24.7KB, xlsx) 41419_2026_8734_MOESM4_ESM (17.8KB, docx) Acknowledgements Library preparation and single cell sequencing of in-house mouse embryonic stem cell samples were performed in collaboration with Dr. Inti Alberto De La Rosa Velásquez at the Core Facility Genomics (CF-GEN; Helmholtz Center Munich, Germany). Author contributions NG, AMB, GK, JH, VK, and OG planned, performed, and analyzed experiments. MG, KU, and JS processed and analyzed RNA-sequencing datasets (bulk and scRNAseq). NG, MG, JS, PB, VK, OG designed data presentation. OG, VK initiated and guided the study. VK, MC, and OG interpreted data. OG wrote the manuscript with the help of VK, NG, and MG. Funding The project was funded by the German Research Council to OG (Gi 540/3-4) and a doctoral stipend from the Chinese Scholarship Council (CSC) to NG. Research was further funded by the German Research Council grant number INST 409/223-1 FUGG. Data availability All codes and R-packages used in the study are publicly available and have been disclosed in Methods or are available from the corresponding authors on reasonable request. RNA sequencing data have been uploaded at the Gene Expression Omnibus (GEO), GSE 293121, GSE318465. Additional materials are available upon request from the authors. Code availability All codes used in the present study are mentioned in the cognate section of data analysis and are public domain. Competing interests The authors have no competing interests. Footnotes Edited by Professor Mauro Piacentini Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Ningyue Gong, Mahesh Gouda. These authors jointly supervised this work: Martin Canis, Olivier Gires. Supplementary information The online version contains supplementary material available at 10.1038/s41419-026-08734-w. References 1. Evans M. Discovering pluripotency: 30 years of mouse embryonic stem cells. Nat Rev Mol Cell Biol. 2011;12:680–6. [ DOI ] [ PubMed ] [ Google Scholar ] 2. Evans MJ, Kaufman MH. Establishment in culture of pluripotential cells from mouse embryos. Nature. 1981;292:154–6. [ DOI ] [ PubMed ] [ Google Scholar ] 3. Martin GR. Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem cells. Proc Natl Acad Sci USA. 1981;78:7634–8. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 4. Desbaillets I, Ziegler U, Groscurth P, Gassmann M. Embryoid bodies: an in vitro model of mouse embryogenesis. Exp Physiol. 2000;85:645–51. [ PubMed ] [ Google Scholar ] 5. Gloss BS, Signal B, Cheetham SW, Gruhl F, Kaczorowski DC, Perkins AC, et al. High resolution temporal transcriptomics of mouse embryoid body development reveals complex expression dynamics of coding and noncoding loci. Sci Rep. 2017;7:6731. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 6. Boheler KR, Czyz J, Tweedie D, Yang HT, Anisimov SV, Wobus AM. Differentiation of pluripotent embryonic stem cells into cardiomyocytes. Circulation Res. 2002;91:189–201. [ DOI ] [ PubMed ] [ Google Scholar ] 7. Doevendans PA, Kubalak SW, An RH, Becker DK, Chien KR, Kass RS. Differentiation of cardiomyocytes in floating embryoid bodies is comparable to fetal cardiomyocytes. J Mol Cell Cardiol. 2000;32:839–51. [ DOI ] [ PubMed ] [ Google Scholar ] 8. Went PT, Lugli A, Meier S, Bundi M, Mirlacher M, Sauter G, et al. Frequent EpCam protein expression in human carcinomas. Hum Pathol. 2004;35:122–8. [ DOI ] [ PubMed ] [ Google Scholar ] 9. Ng VY, Ang SN, Chan JX, Choo AB. Characterization of Epithelial Cell Adhesion Molecule as a Surface Marker on Undifferentiated Human Embryonic Stem Cells. Stem Cells. 2010;29–5. [ DOI ] [ PubMed ] 10. Gonzalez B, Denzel S, Mack B, Conrad M, Gires O. EpCAM Is Involved in Maintenance of the Murine Embryonic Stem Cell Phenotype. Stem Cells. 2009;27:1782–91. [ DOI ] [ PubMed ] [ Google Scholar ] 11. Gires O, Klein CA, Baeuerle PA. On the abundance of EpCAM on cancer stem cells. Nat Rev Cancer. 2009;9:143. [ DOI ] [ PubMed ] [ Google Scholar ] 12. Dolle L, Theise ND, Schmelzer E, Boulter L, Gires O, van Grunsven LA. EpCAM and the biology of hepatic stem/progenitor cells. Am J Physiol Gastrointest Liver Physiol. 2015;308:G233–50. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Conigliaro A, Amicone L, Costa V, De Santis Puzzonia M, Mancone C, Sacchetti B, et al. Evidence for a common progenitor of epithelial and mesenchymal components of the liver. Cell Death Differ. 2013;20:1116–23. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 14. Galdos FX, Lee C, Lee S, Paige S, Goodyer W, Xu S, et al. Combined lineage tracing and scRNA-seq reveals unexpected first heart field predominance of human iPSC differentiation. Elife. 2023;7:12:e80075. [ DOI ] [ PMC free article ] [ PubMed ] 15. Sarrach S, Huang Y, Niedermeyer S, Hachmeister M, Fischer L, Gille S, et al. Spatiotemporal patterning of EpCAM is important for murine embryonic endo- and mesodermal differentiation. Sci Rep. 2018;8:1801. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 16. Chen HF, Chuang CY, Lee WC, Huang HP, Wu HC, Ho HN, et al. Surface Marker Epithelial Cell Adhesion Molecule and E-cadherin Facilitate the Identification and Selection of Induced Pluripotent Stem Cells. Stem Cell Rev. 2011;722−35. [ DOI ] [ PubMed ] 17. Kuan II, Lee CC, Chen CH, Lu J, Kuo YS, Wu HC. The extracellular domain of epithelial cell adhesion molecule (EpCAM) enhances multipotency of mesenchymal stem cells through EGFR-LIN28-LET7 signaling. J Biol Chem. 2019;294:7769–86. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 18. Kuan II, Liang KH, Wang YP, Kuo TW, Meir YJ, Wu SC, et al. EpEX/EpCAM and Oct4 or Klf4 alone are sufficient to generate induced pluripotent stem cells through STAT3 and HIF2alpha. Sci Rep. 2017;7:41852. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Lu TY, Lu RM, Liao MY, Yu J, Chung CH, Kao CF, et al. Epithelial cell adhesion molecule regulation is associated with the maintenance of the undifferentiated phenotype of human embryonic stem cells. J Biol Chem. 2010;285:8719–32. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 20. Maetzel D, Denzel S, Mack B, Canis M, Went P, Benk M, et al. Nuclear signalling by tumour-associated antigen EpCAM. Nat Cell Biol. 2009;11:162–71. [ DOI ] [ PubMed ] [ Google Scholar ] 21. Hachmeister M, Bobowski KD, Hogl S, Dislich B, Fukumori A, Eggert C, et al. Regulated intramembrane proteolysis and degradation of murine epithelial cell adhesion molecule mEpCAM. PLoS One. 2013;8:e71836. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 22. Liang KH, Tso HC, Hung SH, Kuan II, Lai JK, Ke FY, et al. Extracellular domain of EpCAM enhances tumor progression through EGFR signaling in colon cancer cells. Cancer Lett. 2018;433:165–75. [ DOI ] [ PubMed ] [ Google Scholar ] 23. Pan M, Schinke H, Luxenburger E, Kranz G, Shakhtour J, Libl D, et al. EpCAM ectodomain EpEX is a ligand of EGFR that counteracts EGF-mediated epithelial-mesenchymal transition through modulation of phospho-ERK1/2 in head and neck cancers. PLoS Biol. 2018;16:e2006624. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 24. Schinke H, Shi E, Lin Z, Quadt T, Kranz G, Zhou J, et al. A transcriptomic map of EGFR-induced epithelial-to-mesenchymal transition identifies prognostic and therapeutic targets for head and neck cancer. Mol Cancer. 2022;21:178. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 25. Pan M, Kohlbauer V, Soares AB, Schinke H, Huang YC, Kranz G, et al. Interactome analysis reveals endocytosis and membrane recycling of EpCAM during differentiation of embryonic stem cells and carcinoma cells. Iscience. 2021;24:103179. [ DOI ] [ PMC free article ] [ PubMed ] 26. Gires O, Pan M, Schinke H, Canis M, Baeuerle PA. Expression and function of epithelial cell adhesion molecule EpCAM: where are we after 40 years? Cancer Metastasis Rev. 2020;39:969–87. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 27. Ying QL, Wray J, Nichols J, Batlle-Morera L, Doble B, Woodgett J, et al. The ground state of embryonic stem cell self-renewal. Nature. 2008;453:519–23. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 28. Zhan T, Ambrosi G, Wandmacher AM, Rauscher B, Betge J, Rindtorff N, et al. MEK inhibitors activate Wnt signalling and induce stem cell plasticity in colorectal cancer. Nat Commun. 2019;10:2197. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 29. Mohammed H, Hernando-Herraez I, Savino A, Scialdone A, Macaulay I, Mulas C, et al. Single-Cell Landscape of Transcriptional Heterogeneity and Cell Fate Decisions during Mouse Early Gastrulation. Cell Rep. 2017;20:1215–28. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 30. Pijuan-Sala B, Griffiths JA, Guibentif C, Hiscock TW, Jawaid W, Calero-Nieto FJ, et al. A single-cell molecular map of mouse gastrulation and early organogenesis. Nature. 2019;566:490–5. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 31. Zeng B, Liu Z, Lu Y, Zhong S, Qin S, Huang L, et al. The single-cell and spatial transcriptional landscape of human gastrulation and early brain development. Cell Stem Cell. 2023;30:851–66.e7. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 32. Haghverdi L, Buttner M, Wolf FA, Buettner F, Theis FJ. Diffusion pseudotime robustly reconstructs lineage branching. Nat Methods. 2016;13:845–8. [ DOI ] [ PubMed ] [ Google Scholar ] 33. Li Q. scTour: a deep learning architecture for robust inference and accurate prediction of cellular dynamics. Genome Biol. 2023;24:149. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 34. Holtzinger A, Rosenfeld GE, Evans T. Gata4 directs development of cardiac-inducing endoderm from ES cells. Dev Biol. 2010;337:63–73. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 35. Schroder CM, Zissel L, Mersiowsky SL, Tekman M, Probst S, Schule KM, et al. EOMES establishes mesoderm and endoderm differentiation potential through SWI/SNF-mediated global enhancer remodeling. Dev Cell. 2025;60:735–48.e5. [ DOI ] [ PubMed ] [ Google Scholar ] 36. Evseenko D, Zhu Y, Schenke-Layland K, Kuo J, Latour B, Ge S, et al. Mapping the first stages of mesoderm commitment during differentiation of human embryonic stem cells. Proc Natl Acad Sci USA. 2010;107:13742–7. [ DOI ] [ PMC free article ] [ PubMed ] 37. Kim WT, Choi HS, Lee HM, Jang YJ, Ryu CJ B-cell receptor associated protein 31 regulates human embryonic stem cell adhesion, stemness, and survival via control of epithelial cell adhesion molecule. Stem Cells. 2014. [ DOI ] [ PubMed ] 38. Yu T, Ma Y, Wang H. EpCAM Intracellular Domain Promotes Porcine Cell Reprogramming by Upregulation of Pluripotent Gene Expression via Beta-catenin Signaling. Sci Rep. 2017;7:46315. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 39. Xiao D, Xiong M, Wang X, Lyu M, Sun H, Cui Y, et al. Regulation of the Function and Expression of EpCAM. Biomedicines. 2024;107:13742–7. [ DOI ] [ PMC free article ] [ PubMed ] 40. Schnell U, Cirulli V, Giepmans BN. EpCAM: structure and function in health and disease. Biochim Biophys Acta. 2013;1828:1989–2001. [ DOI ] [ PubMed ] [ Google Scholar ] 41. Schnell U, Kuipers J, Giepmans BN. EpCAM proteolysis: new fragments with distinct functions? Biosci Rep. 2013;33:e00030. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 42. Slanchev K, Carney TJ, Stemmler MP, Koschorz B, Amsterdam A, Schwarz H, et al. The epithelial cell adhesion molecule EpCAM is required for epithelial morphogenesis and integrity during zebrafish epiboly and skin development. PLoS Genet. 2009;5:e1000563. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 43. Maghzal N, Vogt E, Reintsch W, Fraser JS, Fagotto F. The tumor-associated EpCAM regulates morphogenetic movements through intracellular signaling. J Cell Biol. 2010;191:645–59. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 44. Mitsui K, Tokuzawa Y, Itoh H, Segawa K, Murakami M, Takahashi K, et al. The homeoprotein Nanog is required for maintenance of pluripotency in mouse epiblast and ES cells. Cell. 2003;113:631–42. [ DOI ] [ PubMed ] [ Google Scholar ] 45. Torres J, Watt FM. Nanog maintains pluripotency of mouse embryonic stem cells by inhibiting NFkappaB and cooperating with Stat3. Nat Cell Biol. 2008;10:194–201. [ DOI ] [ PubMed ] [ Google Scholar ] 46. Do DV, Ueda J, Messerschmidt DM, Lorthongpanich C, Zhou Y, Feng B, et al. A genetic and developmental pathway from STAT3 to the OCT4-NANOG circuit is essential for maintenance of ICM lineages in vivo. Genes Dev. 2013;27:1378–90. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 47. Lu H, Ma J, Yang Y, Shi W, Luo L. EpCAM is an endoderm-specific Wnt derepressor that licenses hepatic development. Dev Cell. 2013;24:543–53. [ DOI ] [ PubMed ] [ Google Scholar ] 48. Causeret F, Sumia I, Pierani A. Kremen1 and Dickkopf1 control cell survival in a Wnt-independent manner. Cell Death Differ. 2016;23:323–32. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 49. Zhang H, Fraser ST, Papazoglu C, Hoatlin ME, Baron MH. Transcriptional activation by the Mixl1 homeodomain protein in differentiating mouse embryonic stem cells. Stem Cells. 2009;27:2884–95. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 50. Kokity L, Czimmerer Z, Benyhe-Kis B, Poscher A, Belai E, Steinbach G, et al. Brachyury co-operates with polycomb protein RYBP to regulate gastrulation and axial elongation in vitro. Front Cell Dev Biol. 2024;12:1498346. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 51. Pfeiffer MJ, Quaranta R, Piccini I, Fell J, Rao J, Ropke A, et al. Cardiogenic programming of human pluripotent stem cells by dose-controlled activation of EOMES. Nat Commun. 2018;9:440. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 52. Russ AP, Wattler S, Colledge WH, Aparicio SA, Carlton MB, Pearce JJ, et al. Eomesodermin is required for mouse trophoblast development and mesoderm formation. Nature. 2000;404:95–9. [ DOI ] [ PubMed ] [ Google Scholar ] 53. van den Ameele J, Tiberi L, Bondue A, Paulissen C, Herpoel A, Iacovino M, et al. Eomesodermin induces Mesp1 expression and cardiac differentiation from embryonic stem cells in the absence of Activin. EMBO Rep. 2012;13:355–62. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 54. Bardot E, Calderon D, Santoriello F, Han S, Cheung K, Jadhav B, et al. Foxa2 identifies a cardiac progenitor population with ventricular differentiation potential. Nat Commun. 2017;8:14428. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 55. Scheibner K, Schirge S, Burtscher I, Buttner M, Sterr M, Yang D, et al. Epithelial cell plasticity drives endoderm formation during gastrulation. Nat Cell Biol. 2021;23:692–703. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 56. David R, Jarsch VB, Schwarz F, Nathan P, Gegg M, Lickert H, et al. Induction of MesP1 by Brachyury(T) generates the common multipotent cardiovascular stem cell. Cardiovasc Res. 2011;92:115–22. [ DOI ] [ PubMed ] [ Google Scholar ] 57. Liu Y, Schwartz RJ. Transient Mesp1 expression: a driver of cardiac cell fate determination. Transcription. 2013;4:92–6. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 58. Costello I, Pimeisl IM, Drager S, Bikoff EK, Robertson EJ, Arnold SJ. The T-box transcription factor Eomesodermin acts upstream of Mesp1 to specify cardiac mesoderm during mouse gastrulation. Nat Cell Biol. 2011;13:1084–91. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 59. Grigorash BB, van Essen D, Liang G, Grosse L, Emelyanov A, Kang Z, et al. p16(High) senescence restricts cellular plasticity during somatic cell reprogramming. Nat Cell Biol. 2023;25:1265–78. [ DOI ] [ PubMed ] [ Google Scholar ] 60. Li H, Collado M, Villasante A, Strati K, Ortega S, Canamero M, et al. The Ink4/Arf locus is a barrier for iPS cell reprogramming. Nature. 2009;460:1136–9. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 61. Rao J, Pfeiffer MJ, Frank S, Adachi K, Piccini I, Quaranta R, et al. Stepwise Clearance of Repressive Roadblocks Drives Cardiac Induction in Human ESCs. Cell Stem Cell. 2016;18:341–53. [ DOI ] [ PubMed ] [ Google Scholar ] 62. Lin CW, Liao MY, Lin WW, Wang YP, Lu TY, Wu HC. Epithelial cell adhesion molecule regulates tumor initiation and tumorigenesis via activating reprogramming factors and epithelial-mesenchymal transition gene expression in colon cancer. J Biol Chem. 2012;287:39449–59. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 63. Sankpal NV, Fleming TP, Sharma PK, Wiedner HJ, Gillanders WE. A double-negative feedback loop between EpCAM and ERK contributes to the regulation of epithelial-mesenchymal transition in cancer. Oncogene. 2017;36:3706–17. [ DOI ] [ PMC free article ] [ PubMed ] 64. Horns F, Martinez JA, Fan C, Haque M, Linton JM, Tobin V, et al. Engineering RNA export for measurement and manipulation of living cells. Cell. 2023;186:3642–58.e32. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 65. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15:550. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 66. Hao Y, Hao S, Andersen-Nissen E, Mauck WM 3rd, Zheng S, et al. Integrated analysis of multimodal single-cell data. Cell. 2021;184:3573–87.e29. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 67. Wolf FA, Angerer P, Theis FJ. SCANPY: large-scale single-cell gene expression data analysis. Genome Biol. 2018;19:15. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 68. Xu Y, Zhang T, Zhou Q, Hu M, Qi Y, Xue Y, et al. A single-cell transcriptome atlas profiles early organogenesis in human embryos. Nat Cell Biol. 2023;25:604–15. [ DOI ] [ PubMed ] [ Google Scholar ] 69. Schubert M, Klinger B, Klunemann M, Sieber A, Uhlitz F, Sauer S, et al. Perturbation-response genes reveal signaling footprints in cancer gene expression. Nat Commun. 2018;9:20. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Original Data (1.3MB, pdf) Supplementary Figures (9.9MB, pdf) Supplementary Table 1 (24.7KB, xlsx) 41419_2026_8734_MOESM4_ESM (17.8KB, docx) Data Availability Statement All codes and R-packages used in the study are publicly available and have been disclosed in Methods or are available from the corresponding authors on reasonable request. RNA sequencing data have been uploaded at the Gene Expression Omnibus (GEO), GSE 293121, GSE318465. Additional materials are available upon request from the authors. All codes used in the present study are mentioned in the cognate section of data analysis and are public domain. Articles from Cell Death & Disease are provided here courtesy of Nature Publishing Group ACTIONS View on publisher site PDF (4.9 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top