Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Insulin fine-tunes self-renewal pathways governing naive pluripotency and extra-embryonic endoderm


Signalling downstream of Activin/Nodal (ActA) and Wnt can induce endoderm differentiation and also support self-renewal in pluripotent cells. Here we find that these apparently contradictory activities are fine-tuned by insulin. In the absence of insulin, the combination of these cytokines supports endoderm in a context-dependent manner. When applied to naive pluripotent cells that resemble peri-implantation embryos, ActA and Wnt induce extra-embryonic primitive endoderm (PrE), whereas when applied to primed pluripotent epiblast stem cells (EpiSC), these cytokines induce gastrulation-stage embryonic definitive endoderm. In naive embryonic stem cell culture, we find that insulin complements LIF signalling to support self-renewal; however, when it is removed, LIF, ActA and Wnt signalling not only induce PrE differentiation, but also support its expansion. Self-renewal of these PrE cultures is robust and, on the basis of gene expression, these cells resemble early blastocyst-stage PrE, a naive endoderm state able to make both visceral and parietal endoderm.

This is a preview of subscription content, access via your institution

Access options

Buy this article

Prices may be subject to local taxes which are calculated during checkout

Figure 1: Wnt3a and Activin mediate primitive endoderm differentiation of naive pluripotent cells.
Figure 2: Wnt3a and Activin mediate definitive endoderm differentiation of epiblast stem cells.
Figure 3: Self-renewal of naive extra-embryonic endoderm.
Figure 4: nEnd differentiates into both visceral and parietal endoderm.
Figure 5: Insulin supports naive pluripotency.
Figure 6: Insulin selectively supports the proliferation of naive pluripotent cells.
Figure 7: NACL supports naive pluripotency.
Figure 8: Nodal/ActA and Wnt can mediate primitive endoderm expansion both in vivo and in vitro.

Similar content being viewed by others

Accession codes

Primary accessions

Gene Expression Omnibus


  1. Rossant, J. & Tam, P. P. L. Blastocyst lineage formation, early embryonic asymmetries and axis patterning in the mouse. Development 136, 701–713 (2009).

    Article  CAS  Google Scholar 

  2. Kwon, G. S., Viotti, M. & Hadjantonakis, A.-K. The endoderm of the mouse embryo arises by dynamic widespread intercalation of embryonic and extraembryonic lineages. Dev. Cell 15, 509–520 (2008).

    Article  CAS  Google Scholar 

  3. Arnold, S. J. & Robertson, E. J. Making a commitment: cell lineage allocation and axis patterning in the early mouse embryo. Nat. Rev. Mol. Cell Biol. 10, 91–103 (2009).

    Article  CAS  Google Scholar 

  4. Morgani, S. M. & Brickman, J. M. The molecular underpinnings of totipotency. Phil. Trans. R. Soc. Lond. B 369, 20130549 (2014).

    Article  Google Scholar 

  5. Tanaka, S., Kunath, T., Hadjantonakis, A. K., Nagy, A. & Rossant, J. Promotion of trophoblast stem cell proliferation by FGF4. Science 282, 2072–2075 (1998).

    Article  CAS  Google Scholar 

  6. Kunath, T. et al. Imprinted X-inactivation in extra-embryonic endoderm cell lines from mouse blastocysts. Development 132, 1649–1661 (2005).

    Article  CAS  Google Scholar 

  7. Paca, A. et al. BMP signaling induces visceral endoderm differentiation of XEN cells and parietal endoderm. Dev. Biol. 361, 90–102 (2012).

    Article  CAS  Google Scholar 

  8. Artus, J. et al. BMP4 signaling directs primitive endoderm-derived XEN cells to an extraembryonic visceral endoderm identity. Dev. Biol. 361, 245–262 (2012).

    Article  CAS  Google Scholar 

  9. Kruithof-de Julio, M. et al. Regulation of extra-embryonic endoderm stem cell differentiation by Nodal and Cripto signaling. Development 138, 3885–3895 (2011).

    Article  CAS  Google Scholar 

  10. Cho, L. T. Y. et al. Conversion from mouse embryonic to extra-embryonic endoderm stem cells reveals distinct differentiation capacities of pluripotent stem cell states. Development 139, 2866–2877 (2012).

    Article  CAS  Google Scholar 

  11. Ying, Q.-L. et al. The ground state of embryonic stem cell self-renewal. Nature 453, 519–523 (2008).

    Article  CAS  Google Scholar 

  12. Marks, H. et al. The transcriptional and epigenomic foundations of ground state pluripotency. Cell 149, 590–604 (2012).

    Article  CAS  Google Scholar 

  13. Nichols, J. & Smith, A. Naive and primed pluripotent states. Cell Stem Cell 4, 487–492 (2009).

    Article  CAS  Google Scholar 

  14. Brons, I. G. M. Derivation of pluripotent epiblast stem cells from mammalian embryos. Nature 448, 191–195 (2007).

    Article  CAS  Google Scholar 

  15. Tesar, P. J. et al. New cell lines from mouse epiblast share defining features with human embryonic stem cells. Nature 448, 196–199 (2007).

    Article  CAS  Google Scholar 

  16. Davidson, K. C., Mason, E. A. & Pera, M. F. The pluripotent state in mouse and human. Development 142, 3090–3099 (2015).

    Article  CAS  Google Scholar 

  17. Gafni, O. et al. Derivation of novel human ground state naive pluripotent stem cells. Nature 504, 282–286 (2013).

    Article  CAS  Google Scholar 

  18. Ware, C. B. et al. Derivation of naive human embryonic stem cells. Proc. Natl Acad. Sci. USA 111, 4484–4489 (2014).

    Article  CAS  Google Scholar 

  19. Takashima, Y. et al. Resetting transcription factor control circuitry toward ground-state pluripotency in human. Cell 158, 1254–1269 (2014).

    Article  CAS  Google Scholar 

  20. Theunissen, T. W. et al. Systematic identification of culture conditions for induction and maintenance of naive human pluripotency. Cell Stem Cell 15, 471–487 (2014).

    Article  CAS  Google Scholar 

  21. Chan, Y.-S. et al. Induction of a human pluripotent state with distinct regulatory circuitry that resembles preimplantation epiblast. Cell Stem Cell 13, 663–675 (2013).

    Article  CAS  Google Scholar 

  22. Hayashi, K., Ohta, H., Kurimoto, K., Aramaki, S. & Saitou, M. Reconstitution of the mouse germ cell specification pathway in culture by pluripotent stem cells. Cell 146, 519–532 (2011).

    Article  CAS  Google Scholar 

  23. Irie, N. et al. SOX17 is a critical specifier of human primordial germ cell fate. Cell 160, 253–268 (2015).

    Article  CAS  Google Scholar 

  24. Grapin-Botton, A. & Constam, D. Evolution of the mechanisms and molecular control of endoderm formation. Mech. Dev. 124, 253–278 (2007).

    Article  CAS  Google Scholar 

  25. Murry, C. E. & Keller, G. Differentiation of embryonic stem cells to clinically relevant populations: lessons from embryonic development. Cell 132, 661–680 (2008).

    Article  CAS  Google Scholar 

  26. Wray, J. et al. Inhibition of glycogen synthase kinase-3 alleviates Tcf3 repression of the pluripotency network and increases embryonic stem cell resistance to differentiation. Nature 13, 838–845 (2011).

    CAS  Google Scholar 

  27. Ogawa, K. et al. Activin-Nodal signaling is involved in propagation of mouse embryonic stem cells. J. Cell Sci. 120, 55–65 (2007).

    Article  CAS  Google Scholar 

  28. Funa, N. S. et al. β-catenin regulates primitive streak induction through collaborative interactions with SMAD2/SMAD3 and OCT4. Cell Stem Cell 16, 639–652 (2015).

    Article  CAS  Google Scholar 

  29. Livigni, A. et al. Differentiation of embryonic stem cells into anterior definitive endoderm. Curr. Protoc. Stem Cell Biol. 10:G, 1G.3.1–1G.3.10 (2009).

    Google Scholar 

  30. Morrison, G. M. et al. Anterior definitive endoderm from ESCs reveals a role for FGF signaling. Cell Stem Cell 3, 402–415 (2008).

    Article  CAS  Google Scholar 

  31. Villegas, S. N. et al. PI3K/Akt1 signalling specifies foregut precursors by generating regionalized extra-cellular matrix. eLife 2, e00806 (2013).

    Article  Google Scholar 

  32. Duncan, S. A. et al. Expression of transcription factor HNF-4 in the extraembryonic endoderm, gut, and nephrogenic tissue of the developing mouse embryo: HNF-4 is a marker for primary endoderm in the implanting blastocyst. Proc. Natl Acad. Sci. USA 91, 7598–7602 (1994).

    Article  CAS  Google Scholar 

  33. Diez-Roux, G. et al. A high-resolution anatomical atlas of the transcriptome in the mouse embryo. PLoS Biol. 9, e1000582 (2011).

    Article  CAS  Google Scholar 

  34. Canham, M. A., Sharov, A. A., Ko, M. & Brickman, J. M. Functional heterogeneity of embryonic stem cells revealed through translational amplification of an early endodermal transcript. PLoS Biol. 8, e1000379 (2010).

    Article  Google Scholar 

  35. Morgani, S. M. et al. Totipotent embryonic stem cells arise in ground-state culture conditions. Cell Rep. 3, 1945–1957 (2013).

    Article  CAS  Google Scholar 

  36. Morgani, S. M. & Brickman, J. M. LIF supports primitive endoderm expansion during pre-implantation development. Development 142, 3488–3499 (2015).

    Article  CAS  Google Scholar 

  37. Hamazaki, T., Oka, M., Yamanaka, S. & Terada, N. Aggregation of embryonic stem cells induces Nanog repression and primitive endoderm differentiation. J. Cell Sci. 117, 5681–5686 (2004).

    Article  CAS  Google Scholar 

  38. Kaufman-Francis, K. et al. Differential response of epiblast stem cells to Nodal and Activin signalling: a paradigm of early endoderm development in the embryo. Phil. Trans. R. Soc. Lond. B 369, 20130550 (2014).

    Article  Google Scholar 

  39. Niwa, H., Burdon, T., Chambers, I. & Smith, A. Self-renewal of pluripotent embryonic stem cells is mediated via activation of STAT3. Genes Dev. 12, 2048–2060 (1998).

    Article  CAS  Google Scholar 

  40. Osorno, R. et al. The developmental dismantling of pluripotency is reversed by ectopic Oct4 expression. Development 139, 2288–2298 (2012).

    Article  CAS  Google Scholar 

  41. Factor, D. C. et al. Epigenomic comparison reveals activation of ‘seed’ enhancers during transition from naive to primed pluripotency. Cell Stem Cell 14, 854–863 (2014).

    Article  CAS  Google Scholar 

  42. Niakan, K. K., Schrode, N., Cho, L.T. Y. & Hadjantonakis, A.-K. Derivation of extraembryonic endoderm stem (XEN) cells from mouse embryos and embryonic stem cells. Nat. Protoc. 8, 1028–1041 (2013).

    Article  CAS  Google Scholar 

  43. Frum, T. et al. Oct4 cell-autonomously promotes primitive endoderm development in the mouse blastocyst. Dev. Cell 25, 610–622 (2013).

    Article  CAS  Google Scholar 

  44. Le Bin, G. C. et al. Oct4 is required for lineage priming in the developing inner cell mass of the mouse blastocyst. Development 141, 1001–1010 (2014).

    Article  CAS  Google Scholar 

  45. Rosas, M. et al. The transcription factor Gata6 links tissue macrophage phenotype and proliferative renewal. Science 344, 645–648 (2014).

    Article  CAS  Google Scholar 

  46. Brown, K. et al. A comparative analysis of extra-embryonic endoderm cell lines. PLoS ONE 5, e12016 (2010).

    Article  Google Scholar 

  47. Ohnishi, Y. et al. Cell-to-cell expression variability followed by signal reinforcement progressively segregates early mouse lineages. Nat. Cell Biol. 16, 27–37 (2014).

    Article  CAS  Google Scholar 

  48. Artus, J. et al. BMP4 signaling directs primitive endoderm-derived XEN cells to an extraembryonic visceral endoderm identity. Dev. Biol. 361, 245–262 (2012).

    Article  CAS  Google Scholar 

  49. Sokolik, C. et al. Transcription factor competition allows embryonic stem cells to distinguish authentic signals from noise. Cell Syst. 1, 117–129 (2015).

    Article  CAS  Google Scholar 

  50. Williams, M. R. et al. The role of 3-phosphoinositide-dependent protein kinase 1 in activating AGC kinases defined in embryonic stem cells. Curr. Biol. 10, 439–448 (2000).

    Article  CAS  Google Scholar 

  51. Sturm, O. E. et al. The mammalian MAPK/ERK pathway exhibits properties of a negative feedback amplifier. Sci. Signal. 3, ra90 (2010).

    Article  CAS  Google Scholar 

  52. McLean, A. B. et al. Activin a efficiently specifies definitive endoderm from human embryonic stem cells only when phosphatidylinositol 3-kinase signaling is suppressed. Stem Cells 25, 29–38 (2007).

    Article  CAS  Google Scholar 

  53. Chazaud, C., Yamanaka, Y., Pawson, T. & Rossant, J. Early lineage segregation between epiblast and primitive endoderm in mouse blastocysts through the Grb2-MAPK pathway. Dev. Cell 10, 615–624 (2006).

    Article  CAS  Google Scholar 

  54. Frankenberg, S. et al. Primitive endoderm differentiates via a three-step mechanism involving Nanog and RTK signaling. Dev. Cell 21, 1005–1013 (2011).

    Article  CAS  Google Scholar 

  55. Yasunaga, M. et al. Induction and monitoring of definitive and visceral endoderm differentiation of mouse ES cells. Nat. Biotechnol. 23, 1542–1550 (2005).

    Article  CAS  Google Scholar 

  56. Fujikura, J. et al. Differentiation of embryonic stem cells is induced by GATA factors. Genes Dev. 16, 784–789 (2002).

    Article  CAS  Google Scholar 

  57. McDonald, A. C. H., Biechele, S., Rossant, J. & Stanford, W. L. Sox17-mediated XEN cell conversion identifies dynamic networks controlling cell-fate decisions in embryo-derived stem cells. Cell Rep. 9, 780–793 (2014).

    Article  CAS  Google Scholar 

  58. Zhao, Y. et al. A XEN-like state bridges somatic cells to pluripotency during chemical reprogramming. Cell 163, 1678–1691 (2015).

    Article  CAS  Google Scholar 

  59. Parenti, A., Halbisen, M. A., Wang, K., Latham, K. & Ralston, A. OSKM induce extraembryonic endoderm stem cells in parallel to induced pluripotent stem cells. Stem Cell Rep. 6, 447–455 (2016).

    Article  CAS  Google Scholar 

  60. Ogawa, K. et al. Activin-Nodal signaling is involved in propagation of mouse embryonic stem cells. J. Cell Sci. 120, 55–65 (2007).

    Article  CAS  Google Scholar 

  61. Nishii, T., Oikawa, Y., Ishida, Y., Kawaichi, M. & Matsuda, E. CtBP-interacting BTB zinc finger protein (CIBZ) promotes proliferation and G1/S transition in embryonic stem cells via Nanog. J. Biol. Chem. 287, 12417–12424 (2012).

    Article  CAS  Google Scholar 

  62. Martin Gonzalez, J. et al. Embryonic stem cell culture conditions support distinct states associated with different developmental stages and potency. Stem Cell Rep. 7, 177–191 (2016).

    Article  CAS  Google Scholar 

  63. Kleinstiver, B. P. High-fidelity CRISPR–Cas9 nucleases with no detectable genome-wide off-target effects. Nature 529, 490–495 (2016).

    Article  CAS  Google Scholar 

  64. Gouon-Evans, V. et al. BMP-4 is required for hepatic specification of mouse embryonic stem cell-derived definitive endoderm. Nat. Biotechnol. 24, 1402–1411 (2006).

    Article  CAS  Google Scholar 

  65. Illingworth, R. S., Hölzenspies, J. J., Roske, F. V., Bickmore, W. A. & Brickman, J. M. Polycomb enables primitive endoderm lineage priming in embryonic stem cells. Elife 5, e14926 (2016).

    Article  Google Scholar 

Download references


We thank the entire Brickman laboratory and A. Grapin-Botton for critical discussion of this manuscript; D. Factor and P. Tesar for help importing and analysing enhancer data sets; J. M. Gonzalez, K. Bonderup and G. de la Cruz for technical assistance; A. Sharov for advice on bioinformatics analysis; and A. Watt (University of Edinburgh, UK) for reagents and discussion, and D. Alessi (Dundee University, UK) and H. Lickert (Helmholtz Zentrum Muenchen, Germany) for reagents. This work was supported by the Novo Nordisk Foundation Center for Stem Cell Biology, the Danish National Research Foundation (DNRF 116) and the MRC UK. K.G.V.A. was supported by a University of Copenhagen Studentship.

Author information

Authors and Affiliations



K.G.V.A. and J.M.B. conceived of the project and designed experiments. K.G.V.A., L.M.F., F.V.R., W.B.H. and M.A.C. designed reagents. K.G.V.A., W.B.H., F.V.R. and T.E.K. carried out experiments. A.A., K.G.V.A. and W.B.H. analysed bioinformatic data. A.A. designed bioinformatic pipelines. K.G.V.A. and J.M.B. wrote the manuscript. W.B.H., F.V.R. and T.E.K. contributed to revising the manuscript.

Corresponding author

Correspondence to Joshua M. Brickman.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Integrated supplementary information

Supplementary Figure 1 Characterisation of the HRHnG reporter cell line and PrE induction in vitro.

(A) Hnf4α and Hhex targeting constructs showing genomic locus, linearised and digested constructs with targeting arms, and incorporation into the locus by homologous recombination, both before and after treatment with Cre recombinase. Southern blot probes are shown in blue, and the respective size of the digested DNA is shown with two headed arrows. (B) Southern blot of HRHnG cell line showing the wild type gene size (22–23 kb) and the labelled targeted fragment (6 kb) before selection cassette removal. Clones HRHnG 44 and 45 are correctly targeted. HRHnG_44 ESCs were treated with Cre recombinase to give rise to the sub-clones 44_6, 44_7, 44_18, 44_19 and 44_22. The labelled band changes to 23 kb after successful cassette excision. 44_18 still contains the selection cassette and was not used. Targeted sub-clones were karyotyped: the chromosome count for each is 40 (average result of 10 cell counts, from 6 independent clones), scale bars 25 μm. (C) HRHnG ESCs were differentiated to ADE using our previously published protocol, shown in the cartoon. ADE (left) was differentiated for 5 further days to hepatic. Cells were not sorted before plating into hepatic differentiation, and hepatic colonies grew in small clusters surrounded by non-hepatic cells (right). Cells express Hnf4αGFP (green) or HexRS (red), scale bars 200 μm. Images are representative of 3 differentiations and more than 10 different photographed fields. (D) Dot plot illustrating qRT–PCR of HRHnG ESCs differentiated for 4 days in Activin and Wnt3a. Differentiated cells were sorted for RNA based on GFP expression and analysed for key markers: Oct4 (ESC), Pdgfra, Hnf4α Sox7, Gata6 and Hhex (PrE), each normalised to unsorted expression and log2 transformed. N = 3 biological replicates and the mean value is shown by the horizontal line and standard error shown. Source data provided in Supplementary Table 10. (E) Western blot showing Stat3−/− HexVenus ESCs generated using a high-fidelity CRISPR nuclease. Clones Nu1c1, Nu1c15, and Nu2c6 were used for the experiments in Fig. 1. Cells were cultured in 2i/LIF. (F) Dot plot showing qRT–PCR time course of HRHnG ESCs cultured in 2i/LIF and then differentiated to PrE over 7 days. XEN and definitive endoderm (ADE) were included for comparison. Key lineage markers are shown for pluripotency (Oct4, Nanog, Klf4), primitive streak/mesendoderm (Foxa2, Gsc, T), and endoderm (Sox17, Gata6, Gata4, Hnf4α, PDGFRA, Sox7). All expression data is normalised to ESC controls, with the mean and standard error shown for n = 3. The definitive endoderm samples are those shown in Supplementary Fig. 2C, but renormalized to ESCs for comparison. Source data provided in Supplementary Table 10.

Supplementary Figure 2 Activin and Wnt3a support EpiSC differentiation to definitive endoderm.

(A) HRHnG ESCS were differentiated to EpiSCs by passaging 12 times on fibronectin in standard EpiSC medium. Cells were compared by immunohistochemistry to an in vivo derived EpiSC line rtO, and stained for CDH1 (red) and OCT4 (green), scale bars 200 μm. In vitro derived HRHnG EpiSCs broadly resemble in vivo derived EpiSCs, both by morphology and by protein expression. Images are representative of 3 experiments and more than 7 different photographed fields. (B) ADE was analysed by FACS using CXCR4-APC antibody. The GFP positive population, isolated by the green box gate, is shown in the lower panels. On both days 3 and 4, the GFP positive cells are almost universally CXCR4 positive, precluding the possibility that they are PrE. Most transition to HexRS positive by day 4, indicating a foregut identity. Dot plots are representative of 8 differentiations. (C) Box plot showing qRT–PCR of HRHnG RNA during EpiSC differentiation to ADE over 4 days (n = 5). rtO controls are included both as EpiSCs (rtO) or differentiated to ADE (rtO d4) (n = 4). Distribution of all data points are shown and are normalized to the starting cells, EpiSCs (day 0). EpiSCs initially downregulated pluripotency markers (Oct4), and upregulated primitive streak/mesendoderm markers (Eomes, T, Gsc, Foxa2). Endodermal genes were upregulated towards the end of the protocol (Sox17, Lefty). Source data provided in Supplementary Table 10. (D) Flow cytometry time course showing HRHnG reporter behaviour through EpiSC differentiation to ADE and subsequent hepatic differentiation. EpiSCs first start to express CXCR4, then HexRS. On day 3, cells were plated into hepatic conditions. X axis, CXCR4 or Hnf4αGFP, as indicated. Y axis, HexRS. After plating into Hepatic conditions, Hnf4αGFP is significantly upregulated. Dot plots are representative of 3 differentiations. (E) Examples of changes in the regulatory regions of DE versus PrE genes in primed versus naïve pluripotency. IGV genome tracks showing Epi-specific and Esc-specific H3k27ac and DNase I hypersensitivity peaks enrichment at proximity of Lifr, Hnf4a, Cxcr4 and Gsc—genes that are differentially expressed in nEnd and ADE conditions. Data has sliding windows varying from 144 to 284 kb. Top two lines in each screen shot, EpiSC H3K27Ac, followed by DNAseI. Bottom two lines are the same, but for naïve ESCs. Data derived from ref. 41. (F) Contingency table of upregulated genes in nEnd (versus ESC) and ADE (versus Epi) conditions with number of overlapping peaks in H3K27ac and DNase I Hypersensitivity data for mEpiSC and mESC specific cell types. Numbers represented are overlapping peak counts specific for each cell type within +/−50 Kb of the significantly upregulated genes (FC > 2 and adj. P. val < 0.01). Gene lists are defined based on the set of genes changing when nEnd was compared to naïve ESCs, and ADE compared to EpiSCs. While the number of EpiSC specific enhancers is lower, they are enriched in the list of ADE specific induction events by >than 2-fold. The percentage of EpiSC specific enhancers in both datasets is listed to the right of the table. The H3K27ac and DNase I Hypersensitivity Data was obtained from ref. 41. Fisher’s Exact Test: P value < 2.2e−16, odds ratio at 95%, confidence Interval: 2.452–2.698.

Supplementary Figure 3 Additional characterisation of nEnd cultures.

(A) nEnd lines can be frozen and thawed. HRHnG nEnd showing the recovery after thawing, and expansion until confluency at 9 days, scale bars 200 μm. On day 9, the cells were analysed by flow cytometry, and expressed both Hnf4αGFP and PDGFRA (dot plot). Images are representative of 6 differentiations and more than 10 different photographed fields for each. (B) nEnd lines have a normal karyotype. Images show p13 HRHnG-BFP nEnd (left, n = 23 cells counted) and p6 HV5.1 (right, n = 9 cells counted) nEnd. Typical are shown. Both nEnd lines have an average chromosome counts of 40, scale bars 25 μm. The figure shows two lines, but the source data contains a third (p4 HRHnG, n = 10 cells counted). (C) Progression of PDFGRA and Hnf4αGFP expression during PrE differentiation. HRHnG ESCs differentiated from 2i/LIF to PrE over 8 days. PDGFRA expression emerges first on day 3, and this population begins to express both Hn4aGFP and PDGFRA on days 4–5. The proportion of PrE increases between days 5 and 8, and becomes more homogeneous. This increased homogeneity is also apparent by microscopy (lower panel, scale bar 200 μm). Dot plots are representative of 4 differentiations. (D) Expression panels for pluripotency, endoderm and immunity markers based on the set of significantly changing genes defined based on ANOVA. Values indicate intensity values from the microarray (from 0.2 to 5). (E) Pairwise comparisons of microarray expression data, comparing: Undifferentiated ESCs to differentiated cultures (Diff), Late passaged nEnd (Late) to XEN cells, and differentiated cultures (Diff) to Late passaged nEnd (Late). Comparisons were used to carried out using ExAtlas ( to generate lists of underexpressed and overexpressed genes (fold change cut off = 2, FDR threshold ≤ 0.05). The position of key markers is indicated in each comparison. Genes lists were uploaded to DAVID for GO analysis, and are shown in Supplementary Table 4. (F) Normalised data was used to generate the heatmap shown in Fig. 3i. This data was grouped into 20 significant expression clusters. Expression values of each cluster, and each triplicate, were averaged, to give the simplified heatmap overview shown here. Data has been sorted according to the expression in ESCs. (G) Key clusters from F and Fig. 3i were analysed further. Clusters were chosen based on high expression. GO analysis of all clusters is shown in Supplementary Table 4. A summary of the top 5 GO terms is shown for each cluster, sorted based on P value as calculated by DAVID (

Supplementary Figure 4 Testing nEnd potency by chimera injection.

(A) Schematic of injection experiment, indicating the process for preparing the injected cells with a brief accutase wash, and the potential locations where injected H2B-mCherry positive cells could contribute in late blastocysts (Epiblast, PrE and trophoblast) and E6.5 (epiblast, visceral endoderm, parietal endoderm). (B) Flow cytometry for PDGFRA and mCherry during HVMC nEnd adaptation to nEnd culture conditions. Cells were passaged in preparation for blastocyst injection, and maintained a high level of PDGFRA and H2B-mCherry expression. Numbers indicate the percentage of PDGFRA and mCherry double positive cells. Dot plots are representative of 3 HVMC passages. (C) Confocal sections showing contribution to both visceral and parietal endoderm. Data supports the images and experiment shown in Fig. 4f. Each row shows a pair of photos in different optical planes, from three different embryos with contribution in both Reichert’s membrane and the visceral endoderm. Images show confocal optical sections, with overlaid images from the brightfield, red (H2B-mCherry) and green (autofluorescence) fluorescent channels. (D) Confocal sections showing three embryos with chimeric contribution to only the visceral endoderm (top photo), or Reichert’s membrane (lower two photos, shown partially dissected). Images show confocal optical sections, with overlaid images from the brightfield, red (H2B-mCherry) and green (autofluorescence) fluorescent channels.

Supplementary Figure 5 The effect of media components on the choice between self-renewal or differentiation of naïve ESCs.

(A) Comparison between the base media N2B27 and RPMI + B27. While there are many differences, some of the largest differences were in the components highlighted in red including Insulin, Glucose, Galactose, L-Alanine, Sodium Pyruvate and Vitamin A. We then addressed these components during ESC self-renewal, and PrE differentiation. (B) Table of ingredients of B27. There is 2× the amount of B27 in RPMI + B27 than in N2B27, therefore one of these components could be enhancing PrE differentiation. Highlighted in red are two components that based on other studies could be responsible, Vitamin A and Galactose. (C) RPMI + B27 contains ready-mixed Glutamax, while N2B27 contains a homemade supplement of Glutamine, therefore we asked whether this could cause the difference in PrE differentiation of the two media. NanogGFP (transcriptional reporter) ESCs were differentiated to PrE for 6 days in either glutamine or Glutamax in RPMI + B27. We did not observe a significant difference. Dot plots are representative of 3 differentiations. (D) To address whether the supplements added to RPMI or N2B27 could affect pluripotency versus PrE differentiation, we asked whether making RPMI + B27 more comparable to N2B27 would reduce PrE differentiation. Thus, we added half the amount of B27, or N2 to RPMI PrE medium and differentiated the cells for 6 days. Both B27 and N2 have a small effect of reducing PrE differentiation, suggesting that there are multiple factors causing a difference in differentiation efficiency between the two media. Dot plots are representative of 3 differentiations. (E) NanogGFP ESCs were differentiated to PrE for 6 days in increasing doses of Sodium Pyruvate (NaPyr), or in a double dose of L-Alanine. Neither Sodium pyruvate nor L-alanine had a large effect on PrE differentiation (extremely high doses of NaPyr reduced PrE differentiation by 1/5, however this may be due to the secondary effects of such a high dose). Dot plots are representative of 3 differentiations. (F) NanogGFP ESCs were differentiated to PrE for 6 days in Galactose, which had little effect on differentiation efficiency. Dot plots are representative of 3 differentiations. (G) Retinoic acid can improve PrE differentiation. Retinoic acid is known to play a role in PrE specification. NanogGFP ESCs were differentiated to PrE for 6 days in DMSO control, or in varying concentrations of the RA inhibitor AGN 193109 (AGN). Cells were analysed for PrE differentiation by flow cytometry, with staining for PDGFRA. Graph shows quantification of percentages of gated cells from FACS data based on NanogGFP expression (ESC) or PDGFRA staining (PrE); n = 3, error bars show ± s.d., P values were calculated by t-test for each in comparison to 0 μM (DMSO) control for either PrE or ESC population. PrE calculations; P values 1 μM P = 0.0017; 10 μM P = 0.0032. ESC calculations; 1 μM P = 0.0060, 10 μM P = 0.0126. Dot plots are representative of 3 differentiations. (H) Antagonism of RA signalling can inhibit differentiation. AGN (1 μM) was added to self-renewing cells and differentiating cells. In both 2i/LIF and NACL there was less spontaneous differentiation to PrE in the presence of AGN. In PrE differentiation, AGN drastically reduced the percentage of differentiated PrE cells. Dot plots are representative of 4 differentiations.

Supplementary Figure 6 Signalling pathways in the decision between self-renewal and differentiation.

(A) The timing of insulin addition determines its impact on differentiation. NanogGFP ESCs were differentiated to PrE for 6 days. Insulin at 12.5 ng μl−1 was added for the time shown. Cells were assessed for PrE differentiation by staining for PDGFRA and analysing by flow cytometry. Flow cytometry dot plots show quadrants based on NanogGFP (x-axis) and PDGFRA (y-axis). Dot plots are representative of 3 differentiations. (B) Response of cell growth to insulin in PrE differentiation. Total cells from three different E14TG2A derived ESC cell clonal lines were counted at the end point of differentiation. The graph shows that insulin has a large impact on cell number. N = 3 biological replicates, data is the mean and errors bars are plus and minus s.d. Statistical significance was calculated by t-test and the P value was 0.013. (C) Response of ESCs to insulin signalling. Two pairwise comparisons were carried out based on the experiment described in Fig. 5g. The first is shown in Fig. 5i. The second compared PDK1 mutants in NACL versus 0-h (insulin withdrawal). For both comparisons, the list of genes significantly stimulated by insulin (downregulated upon insulin withdrawal) was compared via a Venn diagram (generated via (D) Quantification of western blot analysis for Gsk3 phosphorylation (S9/S21) in cells with or without insulin. N = 3 independent clones. Error bars represent the s.d. of, P = 0.022 (t-test). (E) ESCs were cultured in 2i/LIF and NACL for 5 passages, then assessed for gene expression by qRT–PCR. Bars show the mean (n = 3 biological replicates), error bars ± s.d., P = 0.0030 (t-test). (F) NANOG protein levels in different ESC culture conditions. Quantification of Western blot analysis for NANOG expression in E14TG2A ESCs at p2 and p4 in the indicated culture conditions. Error bars represent the s.d. for n = 3 experiments. (G) Karyotype of NACL cultured ESCs. The chromosome count for two high-passage (>10) NACL cultures is shown. The average for both is 40, a normal count in mouse ESCs. (H) Circles show Alkaline Phosphatase (AP) staining of ESCs plated in the same density into 2i/LIF or NACL. nEnd cultures are shown as AP negative controls. Left panel: Whole stained 6 well plate. Right panel: images of individual colonies in NACL and 2i/LIF. Both colonies have high levels of AP staining, and tight circular morphologies, scale bars 800 μm. Images are representative of 6 experiments, and images of individual colonies are representative of 6 different photographed fields. (I) Transcriptomes of three self-renewing cell types. RNA expression as determined by microarray for RNA. The average of three replicates is shown, fold change cut off = 2, FDR threshold ≤ 0.05. (J) PrE (H +) and Epi (H −) primed ESCs have different growth efficiencies in different ESC medias. HexVenus ESCs were sorted based on HexVenus expression (top and bottom 25%) and plated into 2i/LIF or NACL for 7 days. These cells were then counted. Bars show the mean (n = 3 independent sorting experiments), error bars show s.d. P values calculated by t test, from top to bottom: 0.031, 0.027, 0.008, 0.024, 0.046.

Supplementary Figure 7 Analysis of signalling in self-renewing 2i/LIF, NACL and nEnd conditions.

(A) Graphs displaying the total expression level of different transcription factors, signalling molecules/kinases, their activated forms, and the ratio between these in the three medium conditions 2i/LIF (2i/LIF), NACL and nEnd, and the expression levels of GP130, GATA6, KLF4 and NANOG. All were normalized to the expression of Histone 3 in the corresponding sample, and then averaged (n = 3 biological replicates). Error bars indicate the s.d. in each medium condition shown in B. (B) Western blots characterizing expression levels of markers for self-renewing ESCs (NANOG, KLF4, gp130), differentiated PrE (GATA6, PDGFRα) and activated kinases downstream of LIF, FGF, WNT, NODAL/ACTIVIN and insulin signalling. Protein lysates were collected from three cell lines grown in the self-renewing conditions 2i + LIF and NACL (N2B27 + ActA + CHI + LIF), and from the same cell lines after >5 passages as differentiated primitive endoderm (nEnd). The number of passages in these 3 media conditions is given in the top (for example, HVMC p5 PrE: passage 5 of cell line HVMC in PrE conditions). Full blots with indicated sizes are shown in Supplementary Fig. 9.

Supplementary Figure 8 Signalling in endoderm differentiation and expansion.

(A) To test the role of LIF signalling in nEnd expansion, LIF (1:1,000) or JAKi (either 1 μM or 10 μM) was added to passage 5 nEnd cultures, and to XEN cells. After 6 days, cells were counted and then stained for PDGFRA. Cells were analysed by flow cytometry, and the different numbers of live (DAPI-ve) nEnd cells (PDGFRA + ve) are shown, relative to the DMSO control in each replicate (white bars). N = 3 experimental replicates plotted seperately. (B) nEnd cells were cultured in the presence of LY (10 μM or PD03 (1 μM) for 6 days, then counted. Mean counts (n = 3 biological replicates) of independent E14TG2A-derived cell lines (NanogGFP, SOX2-GFP, and HVMC) are shown, error bars show ± s.d., P values: for LY to DMSO, P = 0.008, for PD03 to DMSO, P = 0.004. (C) The impact of insulin concentration on naïve ESC expansion. SOX2-GFP ESCs were differentiated for 5 days to PrE, then analysed by flow cytometry. On day 1, the concentrations of insulin shown were added to the media. Left graph shows the final percentage of PrE (PDGFRA + ve, SOX2-GFP-ve cells) at the end of differentiation. Three replicates are plotted (n = 3), with the mean shown in red. Both graphs show the same experiment, left graph with scale on the X-axis, right graph without to show effect at very low concentrations. (D) Flow cytometry dot plots of rtO EpiSCs differentiated to ADE for 4 days in Act and Wnt3a, and analysed for induction of definitive endoderm using the combination of CXCR4-PE and c-Kit-APC antibodies. Percentages of live cells shown for each quadrant. Dot plots are representative of 4 differentiations. (E) Insulin inhibits EpiSC to ADE differentiation. rtO EpiSCs were differentiated to ADE for 4 days in standard ADE medium, or medium supplemented with 12.5 ng μl−1 insulin. Flow cytometry dot plots show staining for CXCR4 and c-Kit. Quadrant percentages shown are double negative undifferentiated cells (red) and double positive ADE (black). Dot plots are representative of 7 differentiations (F) HRHnG EpiSCs were differentiated for 4 days to ADE in the presence of LIF (1:1,000), insulin (12.5 μg ml−1), insulin receptor inhibitor GSK 1838705, and analysed by flow cytometry. Populations were gated on HexRS and CXCR4 and categorized as follows: Epi (HexRS/CXCR4), Mesendoderm (HexRS/CXCR4+), and ADE (HexRS+/CXCR4+). Data represented as the mean (n = 4 independent experiments) +/−sd. P values by t test (left to right) = Epi 0.001617, 0.000098, 0.005683, Mes 0.003460, 0.010756, ADE 0.000833, 0.000249. (G) FACS gating strategy example, showing steps taken to set gates for live cell analysis or sorting based on DAPI and an example reporter (GFP) and conjugated antibody (CD31-APC).

Supplementary Figure 9 Unmodified Western blots accompanying Fig. 6, Supplementary Fig. 3E, Supplementary Fig. 6D.

Supplementary information

Supplementary Information

Supplementary Information (PDF 6025 kb)

Life Sciences Reporting Summary (PDF 85 kb)

Supplementary Table 1

Supplementary Information (XLSX 137 kb)

Supplementary Table 2

Supplementary Information (XLSX 12 kb)

Supplementary Table 3

Supplementary Information (XLSX 55 kb)

Supplementary Table 4

Supplementary Information (XLSX 2649 kb)

Supplementary Table 5

Supplementary Information (XLSX 6 kb)

Supplementary Table 6

Supplementary Information (XLSX 255 kb)

Supplementary Table 7

Supplementary Information (XLSX 42 kb)

Supplementary Table 8

Supplementary Information (XLSX 7 kb)

Supplementary Table 9

Supplementary Information (XLSX 32 kb)

Supplementary Table 10

Supplementary Information (XLSX 491 kb)

Three dimensional plot of tSNE analysis of all the transcriptomic data reported in this study. (MOV 8465 kb)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Anderson, K., Hamilton, W., Roske, F. et al. Insulin fine-tunes self-renewal pathways governing naive pluripotency and extra-embryonic endoderm. Nat Cell Biol 19, 1164–1177 (2017).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:

This article is cited by


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing