Volume 163, Issue 7, Pages 1678-1691 (December 2015) A XEN-like State Bridges Somatic Cells to Pluripotency during Chemical Reprogramming  Yang Zhao, Ting Zhao, Jingyang Guan, Xu Zhang, Yao Fu, Junqing Ye, Jialiang Zhu, Gaofan Meng, Jian Ge, Susu Yang, Lin Cheng, Yaqin Du, Chaoran Zhao, Ting Wang, Linlin Su, Weifeng Yang, Hongkui Deng  Cell  Volume 163, Issue 7, Pages 1678-1691 (December 2015) DOI: 10.1016/j.cell.2015.11.017 Copyright © 2015 Elsevier Inc. Terms and Conditions

Figure 1 The Induction of Cell Colonies Expressing XEN Cell Markers Is a Cornerstone Event during Chemical Reprogramming (A) Imaging tracing of the formation process of two CiPSC colonies (stage 3) in stage 1 and stage 2 during chemical reprogramming. Scale bars, 100 μm. (B) Numbers of CiPSC colonies generated from the inside (blue) and outside (red) of epithelial colonies in eight batches of experiments. For experiments 5 and 7, the cell confluence of epithelial colonies was <20%. (C) Immunofluorescence of typical epithelial colonies at the end of stage 1, with the expression of SALL4 (top, red) and GATA4 (top, green), SALL4 (bottom, red) and SOX17 (bottom, green) 4 days after replating at day 12. Scale bar, 100 μm. (D) qRT-PCR analysis of XEN cell markers (Gata4, Gata6, Sox17, Sox7, Sall4) and pluripotency marker Oct4 in MEFs, cells at the end of stage 1 (day 16) and stage 2 (day28) and eXEN (embryo-derived XEN cells). (E) CiPSC colony numbers generated at the end of stage 3 from EpCAM-negative (−), EpCAM-positive (+) and total cell populations sorted at day 20 (in stage 2 of chemical reprogramming). Data are represented as mean ± SD. See also Figure S1. Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions

Figure 2 Identification of Small Molecules that Promote the Transition from Fibroblasts to the XEN-like Cells (A) Numbers of SALL4 and GATA4 double-positive colonies after treatment with DMSO and VC6TF (CHIR, 10 μM and 20 μM) for 12 days. (B) qRT-PCR analysis of XEN cell marker expression after treatment with VC6TF (CHIR, 10 μM and 20 μM) for 12 days. eXEN was set as a positive control. (C) Numbers and phase images of XEN-like colonies after treatment with control cocktail (VC6TF with CHIR, 20 μM) and that with additional small molecule EPZ004777 (E) and AM580 (A) for 16 days. Cells were re-plated at day 12 by 1:2. Scale bar, 100 μm. (D) qRT-PCR analysis of Sall4 expression in cells treated with different small-molecule cocktails during stage 1. (E) qRT-PCR analysis of XEN cell markers expression in cells treated with different small-molecule cocktails at day 12. Data are represented as mean ± SD. See also Figure S2. Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions

Figure 3 Identification of Small Molecules that Promotes the Transition from a XEN-like State to a Pluripotency State (A) Numbers of CiPSC colonies at day 44 induced with a control cocktail (VC6TFAZ) and with 5-aza-dC, 5-aza-dC plus EPZ004777 (EPZ), or SGC0946 (SGC) in stage 2 of 12 days. (B) Phase and fluorescence images of primary CiPSC colonies in stage 3 (day 40) with treatment with VC6TFAZ plus 5-aza-dC and SGC0946 in stage 2. Scale bar, 100 μm. (C) High content screening fluorescent imaging of primary CiPSC colonies as indicated by pOct4-GFP expression at day 40 using different reprogramming conditions. The views were jointed together with all 4× magnification fluorescent views in a well by Meta Xpress software. The control in stage 2 represents VC6TFAZ, 2i in stage 3 represents 2i-medium used in initial protocol. Scale bar, 2,000 μm. (D) Numbers of CiPSC colonies at the end of reprogramming from different densities of cells in each well (6-well plate) re-plated at day 12. Numbers 1 and 2 were two independent experiments. (E) Numbers of CiPSC colonies at the end of reprogramming, with a different time course in stage 2. Data are represented as mean ± SD. See also Figure S3. Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions

Figure 4 A Robust CiPSC Induction Protocol Was Established by Modulating the Cell Transitions through a XEN-like State (A) Schematic comparison of the new protocol in this study and our initial protocol (Hou et al., 2013). In total, 1–20 CiPSC colonies (in 2 wells) were induced from initial 50,000 fibroblasts in a 60-day induction using the initial protocol. Whereas 1,000–9,000 CiPSC colonies (in 10–15 wells) were obtained from initial 50,000 fibroblasts after 40 days of small-molecule treatment by using the new protocol. 2iL represents 2i-medium with LIF. (B) Comparison of the original protocol and the new protocol in primary CiPSC colony numbers (day 40) induced from mouse neonatal fibroblasts (MNFs, left) and mouse adult fibroblasts (MAFs, right), respectively. (C) Numbers of CiPSC colonies generated under shortened durations for each stage. For example, “12+10+6” represents a sequential duration of 12, 10, and 6 days for stage 1, 2, and 3, respectively. As shown, the minimal time course required for CiPSC induction was 26 days (12+8+6). (D) Pluripotency marker expression in CiPSC colonies induced by the new protocol, analyzed by qRT-PCR. (E) Immunofluorescence of pluripotency markers in CiPSC colonies induced by the new protocol. Scale bar, 100 μm. (F) Teratoma generated from CiPSCs induced by the new protocol. Top left: gut-like epithelium (endoderm); top right: neural epithelium (ectoderm). Bottom left: cartilage (mesoderm); bottom right: muscle (mesoderm). Scale bar, 100 μm. (G) Chimeric mice generated from CiPSCs induced from MEFs (top left) and MNFs (top right). Germline contribution of CiPSCs in the gonad of a chimeric mouse embryo was indicated by pOct4-GFP expression (bottom left). Germ-line transmission offspring generated from CiPSCs were shown (bottom right). Scale bar, 100 μm. Data are represented as mean ± SD. See also Figure S4. Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions

Figure 5 Gene Expression Dynamics during CiPSC Generation (A) Gene expression heatmap of XEN, fibroblasts, pluripotency, and primitive streak-related genes. The gene expression patterns of MEFs, eXEN, CiPSCs, ESCs, and reprogrammed cells at day 12 (D12), day 16 (D16), day 20 (D20), and day 28 (D28) during chemical reprogramming were analyzed by RNA-seq. (B) Comparison of XEN-related gene expression pattern at indicated time points during chemical reprogramming (top) and OSKM-induced reprogramming (bottom) measured by qRT-PCR. CiPSCs, OSKM-iPSCs, ESCs, and eXEN were set as controls. (C) Dynamic expression change in pluripotency-associated genes and MET-related genes at the indicated time points during chemical reprogramming, examined by qRT-PCR. (D) Schematic representations of the two routes of somatic reprogramming by using chemical approach (left) and transgenic approach (right). In the transgene-induced reprogramming process, primitive streak-like mesendoderm (PSMN) has been reported as the transient state (Polo et al., 2012; Takahashi et al., 2014), whereas in chemically induced reprogramming, as reported in this study, a unique XEN-like state bridges the transition of fibroblasts to CiPSCs. Data are represented as mean ± SD. See also Figure S5. Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions

Figure 6 XEN Master Gene Expression Is Essential during Chemical Reprogramming (A) Numbers of SALL4 and GATA4 double-positive colonies with Sall4, Gata4, Gata6, or Sox17 knockdown at day 12 of chemical reprogramming. Non-targeting vector shRNA (shControl) was used as negative control. Sh1 and sh2 represent two shRNA vectors for each gene. (B) The expression of Oct4 on day 28 with Sall4, Gata4, Gata6, or Sox17 knockdown measured by qRT-PCR relative to that treated with a non-targeting vector (shControl). (C) Numbers of CiPSC colonies with Sall4, Gata4, Gata6, or Sox17 knockdown. (D) Numbers of XEN-like colonies by overexpression of SALL4 (S4), GATA4 (G4), GATA6 (G6), or their combinations in the presence of small-molecule cocktail VTAE (withdrawal of C6F from VC6TFAE) treatment. (−) represents cells treated with VTAE without the overexpression of XEN-related genes. DMSO-treated cells are shown as negative controls. (E) Fluorescence images of colonies with pOct4-GFP expression induced by the overexpression of SALL4 plus GATA4 or GATA6 in the presence of VTAE. Scale bar, 100 μm. (F) qRT-PCR analysis of Oct4 induced by overexpression of SALL4, GATA4, GATA6, and their combination in the presence of VTAE on day 20. (G) Phase and fluorescence images of primary iPS colonies induced by SALL4, GATA6 (or GATA4) plus Dox-induced SOX2. Scale bar, 100 μm. (H) The “seesaw” model in chemical reprogramming. Chemical reprogramming is a dynamic process with an unbalanced state inclining to extraembryonic endoderm in early reprogramming (top), which is further balanced in the end of reprogramming (bottom). Data are represented as mean ± SD. See also Figure S6. Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions

Figure 7 The XEN-like Intermediates Resemble Embryo-Derived XEN Cells in Gene Expression Patterns, In Vivo Development Potential, and Reprogramming Potential (A) Schematic representation of different types of XEN cells mentioned in this study. All three of these types of XEN cell lines can form parietal endoderm chimera conceptus after injection to blastocyst and can be further reprogrammed to CiPSCs with stage 2 and 3 medium of chemical reprogramming. Their expression patterns were indicated as shown. +, highly expressed; low, expressed in a low level; −, no expression detected. (B) Relative expression of XEN-related genes in different cell types as indicated measured by qRT-PCR. eXEN-1 and eXEN-2 were two sublines of eXENs, maintained in traditional XEN culture medium (Kunath et al., 2005) and stage 1 medium of chemical reprogramming, respectively. (C) Hierarchical clustering of global gene expression profiles in different cell types. XEN-like cell samples at different time points (day 16, 20, 26, and 28) during chemical reprogramming were indicated as D16, D20, D26, and D28. Controls were two batches of MEFs (MEFs-1, MEFs-2), eXEN-1, eXEN-2, two CeXEN cell lines (CeXEN-1, CeXEN-2), two CiPS cell lines (CiPSCs-1, CiPSCs-2), and two ES cell lines (ESCs-1, ESCs-2). (D) Principal component analyses (PCA) of global gene expression profiles of indicated cell types mentioned above. (E) Representative images of two E7.5 chimeras generated by blastocyst injection with GFP-labeled XEN-like cells at day 16 of chemical reprogramming (the end of stage 1) (top). Magnified images focused on the distal part of one conceptus show the integration of XEN-like cells to parietal yolk sac (bottom). The separation of parietal yolk sac from one conceptus indicates the chimeric integration of the XEN-like cells was exclusively in the extraembryonic region (top right). (F) Primary CiPSC colonies and immunostaining of NANOG in CiPSC lines induced from CeXEN cells (left) and eXEN cells (right). Scale bar, 100 μm. Data are represented as mean ± SD. See also Figure S7. Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions

Figure S1 The Induction of Cell Colonies Expressing XEN Cell Markers Is a Cornerstone Event during Chemical Reprogramming, Related to Figure 1 (A) Imaging tracing of the formation process of two CiPSC colonies (stage 3) in stage 1 and stage 2 during chemical reprogramming. Scale bars, 100 μm. (B) XEN-like colony numbers at day 16 generated from EpCAM positive (+), negative (-) populations after FACS sorting at day 12 (in stage 1 of chemical reprogramming). (C) CiPSC colony numbers induced from indicated populations from neural stem cells (NSC, left) and intestinal epithelium cells (IEC, right). EpCAM negative (-), positive (+) and total populations were sorted by FACS at day 13 (for IECs) and day 20 (for NSCs), respectively. (D) Immunofluorescence of epithelial colonies during reprogramming. Shown are a typical colony co-expressing SALL4 (red) and GATA4 (green) in stage 1 (top), a typical colony co-expressing GATA4 (red) and pOct4-GFP reporter (green) in stage 2 (middle), and a CiPSC colony with the expression of pOct4-GFP reporter in the middle (green) and GATA4 expression in the around (red) in stage 3 (bottom). Scale bars, 100 μm. Data are represented as mean ± SD. Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions

Figure S2 Identification of Small Molecules that Promote the Transition from Fibroblasts to the XEN-like Cells, Related to Figure 2 (A) qRT-PCR analysis of MET markers (EpCAM, Cdh1) expression after treatment of VC6TF (CHIR, 10 μM and 20 μM) for 12 days. (B) Numbers of SALL4 and GATA4 double-positive colonies after treatment with VC6TF (CHIR, 20 μM), and that with AM580, EPZ004777 (EPZ) and with their combination for 12 days. (C) qRT-PCR analysis of MET markers (Cdh1, EpCAM) expression after treatment with VC6TF (CHIR, 20 μM) and that with AM580 (A), EPZ004777 (E) and A plus E for 12 days. (D) qRT-PCR analysis of Gata4, Gata6, Sox17, Lin28a, EpCAM and Cdh1 expression in cells treated by different small molecule cocktails during stage 1. eXEN and CiPSCs were set as controls. Data are represented as mean ± SD. Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions

Figure S3 The Identification of Small Molecules that Promote the Transition from a XEN-like to a Pluripotency State, Related to Figure 3 (A) Structures of EPZ004777 and SGC0946. (B) Left: number of CiPSC colonies with treatment of control (VC6TFZ) and with EPZ004777 (EPZ) or SGC0946 (SGC) in stage 2. Right: numbers of CiPSC colonies in 2i-medium with different components in stage 3. (C) Optimization of the concentration (left) and duration (right) of 5-aza-dC. 5-aza-dC was added in stage 2. For duration optimization, 5-aza-dC was 0.5 μM. (D) Optimization of the concentration (left) and duration (right) of SGC0946 (SGC). SGC was added in stage 2. For duration optimization, SGC was 5 μM. (E) High content screening fluorescent imaging of primary CiPSC colonies generated on day 40 with different cell densities re-plated at day 12 in each well (6-well plate). Views were jointed together with all 4X magnification fluorescent views in a well by Meta Xpress software. Scale bars: 2000 μm. (F) Percentage of XEN-like colonies that reprogrammed to CiPSC colonies at day 40 by new protocol in five batches of experiments. (G) Numbers of CiPSC colonies generated inside (blue) and outside (red) XEN-like colonies by new protocol in three batches of experiments. Data are represented as mean ± SD. Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions

Figure S4 A Robust CiPSC Induction Protocol Was Established by Modulating the Cell Transitions through a XEN-like State, Related to Figure 4 (A) Scatter plots comparing the global gene expression profiles of CiPSCs (clone CiPSC-C2) to those of MEFs (left) and ESCs (right). The plots shown were values with Log2 (FPKM) > 0. (B) The heatmap of the spearman distance of global gene expression profiles between MEFs (MEFs-1 and MEFs-2), CiPSCs (CiPSCs-C1, CiPSCs-C2, CiPSCs-C3, CiPSCs-G1, CiPSCs-G3, CiPSCs-N4 and CiPSCs-N6) and ESCs (ESCs-1 and ESCs-2). (C) Typical karyotypes of CiPSC colonies (clone CiPSC-C2 at passage 7 and clone CiPSC-C5 at passage 8). Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions

Figure S5 Gene Expression Dynamics during CiPSC Generation, Related to Figure 5 (A) Scatter plot diagram of single cell qRT-PCR analysis in the expression of Sall4 and Gata4, Gata4 and Sox17, and Sall4 and Sox2, respectively, at the indicated time points. Gene expression levels were indicated by log2 (fold change to the minimal level in these samples). (B) qRT-PCR analysis of some pluripotency-associated genes in the cells treated with different cocktails as indicated in stage 2. (C) Relative expression of XEN cell markers Sall4, Gata6, Gata4, Sox17, Sox7 and Foxa2 in OSKM-induced reprogramming (Golipour et al., 2012; Mikkelsen et al., 2008; Polo et al., 2012; Sridharan et al., 2009). The expression of these genes in negative controls was set as 0 (Log10). As shown, the expression of Sall4, but not other XEN cell marker genes was significantly enhanced during OSKM-induced reprogramming. Data are represented as mean ± SD. Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions

Figure S6 XEN Master Gene Expression Is Essential during Chemical Reprogramming, Related to Figure 6 (A) The expression of Sall4, Gata4, Gata6 and Sox17 with Sall4, Gata6, Gata4 or Sox17 knockdown measured by qRT-PCR on day 16, relative to that in the non-targeting vector (shControl). (B) The expression of Sox2 with Sall4, Gata6, Gata4 or Sox17 knockdown measured by qRT-PCR on day 28, relative to that in the non-targeting vector (shControl). (C) Numbers of iPSC colonies induced from neural stem cells (NSCs) and intestinal epithelium cells (IECs) with Sall4, Gata4, Gata6 or Sox17 knockdown. (D) Numbers of iPSC colonies induced by OSKM with Sall4, Gata4, Gata6 or Sox17 knockdown. (E) qRT-PCR analysis of XEN cell markers Sall4, Gata4, Sox17, Gata6 and Sox7 by treatment of VC6TFAE or overexpression of Sall4 (S4) plus Gata4 (G4) or Gata6 (G6) in the presence of VTAE. eXEN was set as a positive control. (F) The expression of Sox2 by the overexpression of SALL4 (S4), GATA4 (G4), GATA6 (G6) and SALL4 plus GATA6 or GATA4 in the presence of VTAE. (G) Numbers of iPSC colonies generated by Dox-induced expression of Sox2 in the indicated time courses, with the expression of SALL4 plus GATA4 or GATA6 in the presence of VTAEZ. Data are represented as mean ± SD. Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions

Figure S7 The XEN-like Intermediates Resemble Embryo-Derived XEN Cells in Gene Expression Patterns, In Vivo Development Potential, and Reprogramming Potential, Related to Figure 7 (A) Relative expression of XEN-related genes after replating XEN-like cell cells (day 12) into different culture conditions: RPMI CM (traditional XEN culture medium; RPMI-medium with 70% MEF-condition medium) and stage 1 medium. MEF and eXEN were set as negative and positive controls, respectively. (B) Representative images of chimeras generated with EGFP-labeled CeXEN (E6.5, top) and two eXEN sublines, eXEN-1 (E7.5, middle) and eXEN-2 (E6.5, bottom). Magnified images on the right penal show the distal part of parietal yolk sac in the distal part of conceptus. (C) Statistical table of XEN integration chimeric ability of different cell types as indicated to parietal endoderm. EGFP-labeled fibroblasts were set as negative control. (D) qRT-PCR analysis of EpCAM, Cdh1 and Sox2 in different types of XEN cells as indicated. MEFs and ESCs were set as controls. (E) mRNA level of EPCAM, CDH1 and SOX2 in primitive endoderm (XEN cells in vivo) of human embryo and human ESCs at passage#0 and passage#10, analyzed by RNA-seq (Yan et al., 2013). (F) qRT-PCR analysis of pluripotency marker expression in two CiPSC colonies induced from eXEN and two CiPSC colonies induced from CeXEN. MEFs and ESCs were set as controls. (G) A typical chimeric mouse generated from CiPSCs derived from CeXENs. Data are represented as mean ± SD. Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions

Cell 2015 163, 1678-1691DOI: (10.1016/j.cell.2015.11.017) Copyright © 2015 Elsevier Inc. Terms and Conditions