Fig 2: The PrE-/Epi-like niche spontaneously progresses into a post-implantation extra-embryonic endoderm/epiblast epithelialized pro-amniotic-like cavity (XEn/Epi EpiC) in minimal culture conditions. (A) Schematic depicting an E5.0 conceptus (left, middle) and corresponding tissues in an XEn/Epi EpiC (right). EC, ectoplacental cone; ExE, extra-embryonic ectoderm; PE, parietal endoderm; TGCs, trophoblast giant cells; VE, visceral endoderm; BM, basal lamina. (B) Schematic for XEn/Epi EpiC formation. (C) Bright-field image of XEn/Epi rosettes and EpiCs. Scale bar: 200 µm. (D) Immunofluorescence and bright-field images of individual XEn/Epi rosette images after 120 h of culture. Staining for nuclei (DNA), F-actin (pro-amniotic cavity), Pdgfrα-h2b-gfp (PrE), Oct4 (pluripotent Epi) (left) and Otx2 (right). (E) EB cultured under the same basic conditions but without PrE-induction molecules. (F) Immunofluorescence images depicting cell nuclei (DNA), Podxl (polarization) and laminin (basal lamina) in a XEn/Epi pro-amniotic-like cavity. (G,H) Effect of Lif on (G) the percentage of structures forming a pro-amniotic cavity or multiple cavities and (H) the resulting integrated surface area of the cavities. P-value calculated according to the Mann–Whitney U-test. Boxes and whiskers indicate the first, median and third quartile, and minimum and maximum data points excluding outliers, respectively. This result was repeatedly replicated (>10 times) in other experiments as inclusion of a negative control. (I) Immunofluorescence image of a non-cavitated and non-polarized structure resulting from continuous Lif supplementation, labeled for nuclei, Gata6 (PrE) and Podxl (polarization). (J) Immunofluorescence images of 120 h XEn/Epi EpiCs from double Nodal knockout (−/−) and control (+/+) ESCs. (K) Immunofluorescence images of 120 h XEn/Epi EpiCs treated with the Nodal/activin signaling inhibitor SB431542 and non-treated controls. (L) Percentage of structures (32 in total) that contained a laminated or delaminated XEn layer (outlined in black) that is either single or multilayered. Scale bars in D-F,I-K: 50 µm.

Fig 3: The NODAL exon2 promotes tumour progression and metastasis in A375 cells (A) Schematic of the human NODAL locus with its 3 exons (E1 to E3), showing the full-length NODAL mRNA above. The primers used to detect its expression are represented by arrows. (B) Representative images, acquired at t = 0 and t = 12 h, of scratch-wound healing assays performed with A375 and A375ΔE2a cells. (C) Top panel: Normalized gap closure rates of A375, A375ΔE2a, and A375ΔE2d cells during the scratch/wound healing assay shown in (B). Lower panel: Normalized gap closure rates of A375 cells treated with or without recombinant NODAL, recombinant ACTIVIN, or the ACTIVIN/NODAL signalling pathway inhibitor SB431542. The p-value was calculated by a two-way ANOVA test. (D) Top panels: Representative images of tumours (black circle) formed by A375 and A375ΔE2d cells on the upper CAM at the end of the assay (E18). The scale bar is 2000 μm. Lower panels: Representative sections of the tumours. Hematoxylin-eosin stain. Necrotic areas (N) appear in pink. Higher magnification views of selected areas (rectangles) are shown in Figure S1H. The scale bar is 500 μm. (E) Mean weight (mg) of tumours formed by A375 (n = 15) and A375ΔE2d (n = 14) cells on top of the upper CAM 9 days after the graft. The p-value was calculated via a one-way ANOVA test, **** p < 0.0001. (F) Relative presence of A375 and A375ΔE2d cells in the lower CAM (n = 7 for both cell types) and in the liver (n = 7 and n = 8, respectively), as measured by the qPCR detection of human Alu sequences. The p-value was calculated via a one-way ANOVA test * p < 0.05.

Fig 4: LADON, a natural antisense transcript of NODAL is a candidate oncogene in human melanoma. (A) Schematic of the human NODAL locus with its 3 exons (E1 to E3), showing the LADON transcript transcribed from the opposite strand below. The truncated LADON-ΔE2 transcript is expressed in A375 cells deleted for the NODAL exon2. The arrows represent the primers used to track these transcripts. The black boxes show the location of short interspersed nuclear elements (SINE), which are mostly Alu and MIR sequences. (B) The bar plot of the LADON expression profile across all tumour samples and paired normal tissue was obtained from TCGA and GTEx data. The height of the bars represents the median expression of tumour types and normal tissues. The tumour abbreviation chart is provided in supplementary Table S1. The arrow indicates the high differential of LADON expression found in skin cutaneous melanoma (SKCM). (C) Pathological stage plot of LADON expression in SKCM. (D) The overall survival plot of SKCM patients with tumours expressing high or low levels of LADON. The hazard ratio (HR) calculation is based on the Cox PH model. (E) The RT-PCR with primers (L1F-L1R) amplifying a 130 nt band from the 5′ regions of LADON detects the transcript in all cell lines of the panel: melanocytes, non-metastatic melanoma (MNT1), metastatic melanoma (A375, 888 mel, SLM8, FO1), breast cancer (MCF7), and an embryonic kidney. (HEK293). (F) The RT-qPCR measurements of the LADON expression were performed in keratinocytes (HaCat), melanocytes, non-metastatic (MNT1, DAJU, SK28) or metastatic melanoma cells (A375, FO1, 888 Mel, and SLM8), after 24 h and 96 h of culture. The LADON expression was normalized to that of endogenous RPL13. For each cell line, the value at 24 h was then set to 1. (G) The RT-qPCR measurements (dots) of LADON expression after 24 h culture were performed on the same panel as in (F) but normalized to the endogenous RPL13 expression. The histograms display mean values ± SD from a minimum of two independent replicates, p-values were calculated using the Student’s t-test, ** < 0.01.

Fig 5: Automated phenotypic quantification of structures exposed to signaling modulators of BMP and Tgfβ/Nodal pathway: A) Top: experimental design of exposing the pathway modulators from 0–72 h and 48h–120 h of development of XEn/EPiCs. Bottom: schematic showing the tissue compartment features measured on XEn/EPiCs, B) The effect of short‐listed signaling modulators of BMP and Tgfβ/Nodal pathways on the yield of different structures at 0–72 h and 48–120 h, C) Effect of the modulators on the overall size of XEn/EPiCs, Area of Epi, PAC, and XEn in XEn/EPiCs when exposed to BMP pathway modulators at D) 0–72 h, and F) 48–120 h, and Tgfβ/Nodal pathway modulators at E) 0–72 h, and G) 48–120 h; Control is base media + 0.1% DMSO (0–72, 48–120 h). Data are mean ± s.d. obtained from n = 3 wells, with each well containing ≈150 structures. After filtering for only XEn/EPiCs, sample size depended on the ratio of phenotypes observed in the overall population. In the graphs, each dot represents one XEn/EPiC. All statistical hypothesis testing was done using Dunnett's test; ∗ represents P< = 0.05, ∗∗ represents P < 0.01, ∗∗∗ represents P < 0.001, ∗∗∗∗ represents P < 0.0001 (One‐way ANOVA with Dunnet's post‐test).