Fig 2: Trophoblastic differentiation of D-OSKM tumor cells.a Representative histological images of D-OSKM pancreatic tumors containing trophoblast giant cell (TGC) clusters. Dotted lines indicate TGC clusters (>3 cells/high-power field [×400]). Scale bars: 1000 µm (left), 200 µm (right). b Incidence of tumors containing trophoblast giant cells in S-/D-OSKM chimeric mice. c Quantification of TGC cluster area in S-/D-OSKM kidney tumors (normal kidney, n = 5; S-OSKM kidney tumor, n = 16; D-OSKM kidney tumor following Dox treatment at various doses [2.0 mg/mL, n = 17; 0.4 mg/mL, n = 4; 0.25 mg/mL, n = 7; 0.13 mg/mL, n = 4]). Data are presented as means ± SD of biologically independent samples (Kruskal–Wallis test and Dunn’s multiple-comparison test, two-sided). d Representative histological images and immunostaining for CDX2 and PL-1 of D-OSKM pancreatic tumors (upper panels). Dotted lines indicate TGC clusters. Scale bars: 200 µm. Immunofluorescence staining for OCT4 and CDX2 of D-OSKM kidney tumors (lower panels). Scale bars: 50 µm. e (Left) qRT-PCR analyses of trophoblast-related gene expression in S-/D-OSKM kidney tumors. Data are presented as means ± SD of biological triplicates. Relative expression levels to TSCs are shown (one-way ANOVA and Dunnett’s multiple-comparison test, two-sided) (Right) Heatmap showing relative expression of trophoblast-related genes in S-/D-OSKM kidney tumors. Color range is shown using a log2 scale. f Lineage-tracing system for Nanog-expressing cells. Mice were treated with tamoxifen (2.0 mg) six times every other day during Dox treatment. g Immunostaining for Venus in kidney tumors (n = 4). Arrows indicate TGCs. Dotted lines indicate the TGC cluster. Scale bars: 200 µm. h Schematic illustration of an experimental protocol for in vivo D-OSKM reprogramming of somatic cells deficient for Cdx2 or Tead4. i, Representative macroscopic and histological images of Cdx2 KO and Tead4 KO D-OSKM pancreatic tumors. Scale bars: 1000 µm (middle), 500 µm (right). j Quantification of TGC cluster area in control (n = 10), Cdx2 KO (n = 4), and Tead4 KO (n = 4) D-OSKM pancreatic tumors. Data are presented as means ± SD of biologically independent samples (Kruskal–Wallis test and Dunn’s multiple-comparison test, two-sided).

Fig 3: DMRT1 drives in vivo reprogramming and propagation of GCT-like tumor cells.a Schematic illustration of the experimental protocol for in vivo D-OSKM reprogramming of somatic cells deficient for Dmrt1. b Representative macroscopic and histological images of pancreatic tumors in Dmrt1-KO D-OSKM chimeric mice. Immunostaining for OCT4 and Ki67 in pancreatic tumors is shown in the right panels. Scale bars: 1000 µm (left), 200 µm (middle and right). c–e Quantification of tumor area (c), OCT4-positive cell area in the tumor (d), and Ki67-positive cell ratio in the OCT4-positive cell area (e) in the pancreas. D-OSKM mice that developed pancreatic tumors were evaluated in this study (control, n = 10; Dmrt1 KO18, n = 2; Dmrt1 KO21, n = 5). Data are presented as means ± SD of biologically independent samples (Mann–Whitney test, two-sided). f Immunostaining for DAZL and quantification of the DAZL-positive cell area in pancreatic tumors (control, n = 10; Dmrt1 KO18, n = 2; Dmrt1 KO21, n = 5). Data are presented as means ± SD of biologically independent samples (Mann–Whitney test, two-sided) scale bars: 200 µm. g Derivation efficiency of iPSC-like colonies from pancreatic tumors (control, n = 3; Dmrt1 KO21, n = 3). Colony counting was performed at Day 7. h Immunostaining for CDX2 and quantification of the CDX2-positive cell area in pancreatic tumors (control, n = 10; Dmrt1 KO18, n = 2; Dmrt1 KO21, n = 5). Data are presented as means ± SD of biologically independent samples (Mann–Whitney test, two-sided) scale bars: 500 µm. i qRT-PCR analyses of trophoblast-related gene expression in pancreatic tumors. Data are presented as means ± SD of biological triplicates. The mean expression level of control D-OSKM pancreatic tumors was defined as 1 (t-test, two-sided). j Schematic representation of in vivo D-OSKM reprogramming and derivation of PSCs with the expanded differentiation potential into trophoblasts.

Fig 4: Loss of Upf2 is incompatible with ICM survival and leads to a reduction in pluripotent and epiblast-specific gene expression in the blastocyst. (A) ICM grade after 72 and 96 h of outgrowth. The average grades for Upf2+/+, Upf2+/- and Upf2-/- ICMs were 2.1, 2.3 and 4.0 at 72 h, and 2.0, 2.3 and 4.0 at 96 h (**P<0.01; two-tailed Student's t-test). All Upf2-null embryos received an ICM grade of 4 (n=32). (B) TE outgrowth area (µm2) 72 and 96 h post-plating. Average expansion area is not significantly different for Upf2-null embryos and controls (72 h: Upf2+/+=27,556±2960; Upf2+/-=32,242±5919; Upf2-/-=34,596±2759; 96 h: Upf2+/+=30,373±2634; Upf2+/-=32,794±4551; Upf2-/-=33,952±2569) (n=32). (C) Representative brightfield images of blastocyst outgrowths at 72 and 96 h. TE expansion area is outlined in yellow. ICM is indicated by arrowheads. The ICM is clearly visible in Upf2+/+ and Upf2+/- controls but absent in Upf2-/- embryos. (D) The mean ratio of normalized reads in Upf2-null relative to control blastocysts is plotted for epiblast (EPI)- (***P<0.0001, n=50 genes), PrE- (P=0.606, n=77 genes) (Ohnishi et al., 2013) or TE- (P=0.078, n=128 genes) specific genes (Blakeley et al., 2015). A ratio of less than one indicates lower expression in Upf2-null embryos. (E) The core epiblast factors Fgf4 and Nanog, as well as the ‘Yamanaka factors’, Pou5f1, Klf4, c-Myc and Sox2, are downregulated in Upf2-null blastocysts (blue). Extra-embryonic markers in the TE (Cdx2, Dab2, Gata3; green) and PrE (Gata6, Gata4, Sox17; red) were unchanged. (*Q<0.1; Bonferroni-adjusted t-test). Data are mean±s.e.m. in A,B,D. In E, box plots show median values (middle bars) and interquartile ranges (boxes); whiskers indicate the maximum to minimum ranges.

Fig 5: Derivation of D-OSKM tumor-derived PSCs with the potential to differentiate toward the extraembryonic lineage.a Schematic illustration of iPSC derivation from S-/D-OSKM kidney tumor cells (S-/D-iPSCs). A total of nine S-/D-iPSC clones were established from three independent kidney tumors. b Representative macroscopic (upper) and histological (lower) images of ESC-, S-iPSC-, and D-iPSC-derived subcutaneous tumors in immunocompromised mice (upper panels). Immunostaining for PL-1 is shown in the lower panel. Scale bars: 500 µm (left), 200 µm (middle and right). c Incidence of tumors containing TGCs (upper) and quantification of TGC cluster area in ESC (n = 9), S-iPSC (n = 21), and D-iPSC tumors (n = 33). Nine independent S-/D-iPSC clones were used. Data are presented as means ± SD (Kruskal–Wallis test and Dunn’s multiple-comparison test, two-sided). d Representative images of ESC-, S-iPSC-, and D-iPSC-derived EBs (upper panels). Lower panels show immunofluorescence staining for TFAP2C and CDX2 of D-iPSC EB at differentiation Day 8. Scale bars: 200 µm (upper), 50 µm (lower). e qRT-PCR analyses for expression of trophoblast-related genes during EB formation. Data are presented as means ± SD of biological triplicates. The mean expression level at differentiation Day 0 ESCs was defined as 1. f Representative images of blastocysts injected with GFP-labeled ESCs, S-iPSCs, or D-iPSCs. Scale bars: 20 µm. g Representative images of E13.5 chimeric concepti. The chimeric blastocysts were transferred into the uterus of pseudopregnant ICR mice. Distinct GFP signals were observed in two of 16 placentae of D-iPSC chimeric concepti (arrow). Scale bars: 5 mm. h Adult D-iPSC chimeric mouse at 10 months of age. i DNA methylation analysis of individual CpG sites at the H19 and Nespas-Gnasxl ICRs in D-iPSCs.