Fig 2: Differential adhesion force in ETX embryos.a, Schematic showing cell–cell adhesion force measurement by AFM. b, The resulting force–distance curve, following the procedure depicted in a, enables quantification of the maximum adhesion force (Fmax). c, Fmax for the indicated homotypic and heterotypic adhesions between three different cell types. The experiments were performed three times independently. Total measured cell pairs: n = 60 (ES–ES), n = 177 (TS–TS), n = 101 (XEN–XEN), n = 124 (ES–TS), n = 148 (XEN–TS) and n = 134 (XEN–ES). Statistical significance was determined by one-way ANOVA with a multiple comparison test. d, Schematics of weakly and strongly adherent cell pairs at force equilibrium. ? is the contact angle of the two adhering cells. e, Distribution of the measured contact angles at all cell–cell contacts. Total measured cell pairs: n = 31 (ES–ES), n = 38 (TS–TS), n = 30 (XEN–XEN), n = 32 (TS–ES), n = 36 (XEN–TS) and n = 29 (XEN–ES). N = 3 for all conditions. Statistical significance was determined by one-way ANOVA with a multiple comparison test. f, Adhesion forces between cells and different cadherins. Left, schematic showing cell–cadherin adhesion force measurement by AFM. Right, quantification of the results. n = 42 (ES–E-cadherin), n = 35 (ES–P-cadherin), n = 41 (TS–E-cadherin) and n = 37 (TS–P-cadherin). N = 3 for all of the conditions. Statistical significance was determined by unpaired two-tailed Student’s t-test. g, Fmax for homotypic adhesion between the three different cell types after downregulation of Cdh1 or Cdh3. n = 60 (WT ES–ES), n = 18 (Cdh1 KD ES–ES), n = 19 (Cdh3 KD ES–ES), n = 177 (wild-type TS–TS), n = 20 (Cdh1 KD TS–TS), n = 20 (Cdh3 KD TS–TS), n = 101 (wild-type XEN–XEN), n = 19 (Cdh1 KD XEN–XEN) and n = 19 (Cdh3 KD XEN–XEN). N = 3 for all conditions. Statistical significance was determined by one-way ANOVA with a multiple comparison test. h, Heatmap of the adhesion parameter matrix, generated by sampling measured AFM adhesion forces, which parameterizes the CPM. i, Bootstrapping procedure to infer the distributions of conformations under the CPM (N = 498). The schematic represents all of the possible sorted conformations, demonstrating that the ETX-like configuration is the most represented. Conformations observed at a frequency of <5% are grouped. MCS, Markov Chain Steps. In the box and whisker plots in c and e–g, the line inside the box indicates the median value and the error bars show the minimum and maximum values. Box edges indicate lower and upper quartile value. Numerical data are available as source data.Source data

Fig 3: Differential cadherin code in ETX-embryo and natural embryo.(a) UMAP dimensional reduction shows Cdh1, Cdh3 and Cdh6 expression profile in different clusters as indicated by dashed lines. Each dot represents a single cell that is color-coded by sample type. (b) Heatmap showing average expression of cadherin and protocadherin related genes revealed by scRNA-seq in natural embryos (NE, n = 50) collected at 4.5, 5.5 and 6.5 days after fertilization and well-sorted ETX embryos (n = 50) at 4, 5, 6 days of culture. (c) Colonies of cultured ES and TS cells stained to reveal E-cadherin (green) and P-cadherin (red). Quantifications showing the mean intensity (A.U.) of E-cadherin or P-cadherin at cell-cell junctions. 20 colonies from 3 different experiments were selected for quantification. Scale bars represent 100 µm. Data are presented as mean ± SD. Statistics calculated by unpaired two-tailed Student’s t test. (d) Natural embryos (E5.5) and ETX embryos (Day 4) stained to reveal E-cadherin (green) and P-cadherin (red). Magnified insets show E- or P-cadherin staining in ExE and EPI compartments in natural embryos, TS and ES compartments in ETX embryos. Quantifications showing the mean intensity (A.U.) of E-cadherin or P-cadherin at cell-cell junctions. n = 20 ETX embryos and n = 19 natural embryos were used for quantification. Data are presented as mean ± SD. Statistics calculated by unpaired two-tailed Student’s t test. Scale bars represent 100 µm (main Figure) and 20µm (inset). (e) Representative images of E4.5 chimeras (8-cell stage embryos aggregated with Cdh6 or (f) Cdh3 OE ES stained for RFP (magenta), Sox17 (green), and DAPI (grey). Experiments were repeated 3 times. Scale bars represent 50 µm. Zoomed images are of regions indicated by dashed lines (scale bars represent 10 µm). Source numerical data are available in source data. Source data

Fig 4: Correct self-organization is necessary for proper morphogenesis.a, Time course of the assembly of ETX embryos stained to reveal E-cadherin (monochrome), Oct4 (red) and Gata4 (green). The bottom row of images are magnifications of the images above and show E-cadherin staining around a nascent cavity, as indicated by the dashed yellow lines. The dashed green line indicates the boundary between the ES and XEN compartment. Scale bar, 5 µm. b, Representative images showing Oct4 (red), Gata4 (green), E-cadherin (monochrome) and DAPI (grey) staining in day 4 cadherin OE ETX structures formed by combining E-cadherin OE ES cells with P-cadherin OE TS cells and wild-type XEN cells. ETX structures formed by combining wild-type cells were used as a control. Scale bars, 100 µm. c, Comparison and quantification of joined cavity formation in cadherin OE and control ETX structures. n = 361 (control group) and n = 253 (cadherin OE group). N = 5 for each condition. The data are presented as means ± s.d. Statistical significance was determined by unpaired two-tailed Student’s t-test. d, Representative image showing Oct4 (red), Gata4 (green), laminin (monochrome) and DAPI (blue) staining in day 4 cadherin OE ETX structures formed by combining E-cadherin OE ES cells with P-cadherin OE TS cells and wild-type XEN cells. ETX structures formed by combining wild-type cells were used as a control. Scale bars, 100 µm. e, Quantification of the structures that contained continuous or discontinuous laminin. n = 40 ETX structures per condition. N = 3. The data are presented as means ± s.d. Statistical significance was determined by unpaired two-tailed Student’s t-test. f, Self-organization principles in stem cell-derived ETX embryos. Differential expression of E-, K- and P-cadherins enables the sorting of ES (epiblast-like), XEN (VE-like) and TS (TE-like) stem cells. Wild-type ES cells with low E-cadherin expression and wild-type TS cells with low P-cadherin expression exhibited detrimental global sorting efficiency. This could be overcome by overexpressing E-cadherin in ES cells and P-cadherin in TS cells to increase the efficiency of ETX embryo formation. Proper morphogenesis, including cavity formation, basement membrane formation (purple) and symmetry breaking can only be observed in well-sorted structures. Numerical data are available as source data.Source data

Fig 5: Cadherin heterogeneity affects cell positioning in ETX embryos.(a) Schematics and representative images for assembled day 4 ETX-embryos from TS cells (upper) or (b) ES cells (lower) overexpressing (OE) or knocked-down (KD) for the indicated cadherins. Experiments were repeated 6 times. Scale bar represents 100 µm. P-cadherin-overexpressing TS cells and E-cadherin-overexpressing ES cells were pre-stained with Hoechst to distinguish them from cadherin knockdown cells. Scale bar represents 40 µm in zoomed panels. (c) Examples of structures made from E-cadherin-OE-ES, P-cadherin-OE-TS and XEN cells, stained at different time points to reveal ES cells (Oct4, red), TS cells (Tfap2c, blue) and XEN cells (Gata4, green). Scale bar represents 100 µm. (d) Quantification shows time course of formation of correctly-sorted ETX embryos following seeding. Control: 12 h: 24/332 structures; 24 h: 83/531 structures; 48 h: 71/448 structures; 72 h: 51/378 structures. Cadherin OE: 12 h: 80/276 structures; 24 h: 139/385 structures; 48 h: 136/374 structures; 72-h: 151/455 structures. N = 3 for all conditions. Data are presented as Mean ± SD. Statistics calculated by unpaired two-tailed Student’s t test. P values indicate significance between control and Cadherin OE ETX at the same time point. Source numerical data are available in source data. Source data