Cdx2 efficiently induces trophoblast stem-like cells in naïve, but not primed, pluripotent stem cells

Abstract

Diverse pluripotent stem cell lines have been derived from the mouse, including embryonic stem cells (ESCs), induced pluripotent stem cells (iPSCs), embryonal carcinoma cells (ECCs), and epiblast stem cells (EpiSCs). While all are pluripotent, these cell lines differ in terms of developmental origins, morphology, gene expression, and signaling, indicating that multiple pluripotent states exist. Whether and how the pluripotent state influences the cell line's developmental potential or the competence to respond to differentiation cues could help optimize directed differentiation protocols. To determine whether pluripotent stem cell lines differ in developmental potential, we compared the capacity of mouse ESCs, iPSCs, ECCs, and EpiSCs to form trophoblast. ESCs do not readily differentiate into trophoblast, but overexpression of the trophoblast-expressed transcription factor, CDX2, leads to efficient differentiation to trophoblast and to formation of trophoblast stem cells (TSCs) in the presence of fibroblast growth factor-4 (FGF4) and Heparin. Interestingly, we found that iPSCs and ECCs could both give rise to TSC-like cells following Cdx2 overexpression, suggesting that these cell lines are equivalent in developmental potential. By contrast, EpiSCs did not give rise to TSCs following Cdx2 overexpression, indicating that EpiSCs are no longer competent to respond to CDX2 by differentiating to trophoblast. In addition, we noted that culturing ESCs in conditions that promote naïve pluripotency improved the efficiency with which TSC-like cells could be derived. This work demonstrates that CDX2 efficiently induces trophoblast in more naïve than in primed pluripotent stem cells and that the pluripotent state can influence the developmental potential of stem cell lines.

FIG. 1.

Epiblast stem cells (EpiSCs) do not give rise to trophoblast stem cells (TSCs) following Cdx2 overexpression. (A) Experimental outline of the Cdx2 overexpression assay. (B) Typical embryonic stem cell (ESC) morphology. (C) Typical morphology of TSCs in TSC medium [with fibroblast growth factor-4 (FGF4)/Heparin (Hep)]. (D) ESCs lacking Cdx2ER cultured in TSC medium do not exhibit TSC-like morphology, evidenced by ragged colony boundaries. (E) TSC-like cells derived from ESCs after Cdx2 overexpression. Note smooth colony boundary as in (C). (F) TSCs differentiated in the absence of FGF4/Hep for 7 days. (G) ESC-derived TSC-like cells differentiated in the absence of FGF4/Hep for 7 days resemble cells shown in (F). (H) Expression levels quantitative reverse transcription-polymerase chain reaction (qPCR) of TSC markers, relative to Hprt1, in ESCs following Cdx2 overexpression and in ESCs lacking Cdx2ER, relative to TSC levels of these genes, shows that TSC-like cell lines express levels of TSC genes that are at least as high as those expressed by TSCs. (I) Expression levels of TSC markers in TSCs and ESC-derived TSC-like cells after 7-day differentiation show that both cell lines downregulate TSC markers to a similar degree. (J) Expression levels of markers of differentiated trophoblast cell types in TSCs and ESC-derived TSC-like cells after 7-day differentiation show that both cell types increase junctional zone markers to a similar degree, but only TSCs efficiently upregulate labyrinth cell markers during in vitro differentiation. (K) Experimental outline of Cdx2 overexpression in EpiSCs. (L) Typical morphology of EpiSC colonies. (M) EpiSCs lacking Cdx2 cultured in TSC medium do exhibit TSC-like morphology. (N) EpiSCs do not exhibit TSC-like morphology after Cdx2 overexpression. (O) Expression levels of TSC markers in EpiSC overexpressing Cdx2 show that EpiSCs do not acquire TSC-like properties after Cdx2 overexpression. Scale bars=150 μm, error bars=standard error among three technical replicates; Endo, endogenous. Color images available online at www.liebertpub.com/scd

FIG. 2.

Embryonal carcinoma cells (ECCs) give rise to TSC-like cells efficiently upon Cdx2 overexpression (A) Experimental outline of the Cdx2 overexpression assay. (B) Typical ECC morphology. (C) ECCs lacking Cdx2ER do not acquire TSC-like morphology. (D) TSC-like cells derived from ECCs after Cdx2 overexpression in TSC medium. (E) ECC-derived TSCs after 7 days of differentiation in the absence of FGF4/Hep. (F) Fourteen-day differentiated ECC-derived TSC-like cells (compare with Fig. 1F). (G) Expression levels of TSC markers in indicated cell line after Cdx2 overexpression show that ECC-derived TSC-like cell lines express levels of TSC genes that are at least as high as those expressed by TSCs. (H) Expression levels of TSC markers are downregulated in ECC-derived TSC-like cells during differentiation. (I) Expression levels of differentiated trophoblast genes relative to Hprt1 in indicated cell lines show that ECC-derived TSC-like cells take twice as long as TSCs to upregulate differentiation markers. Scale bars=150 μm, error bars=standard error among three technical replicates. Color images available online at www.liebertpub.com/scd

FIG. 3.

Induced pluripotent stem cells (iPSCs) and ESCs give rise to TSC-like cells with variable efficiency. (A) Experimental outline of the Cdx2 overexpression assay. (B–F) The morphology and gene expression of one representative iPSC or ESC subclone after the Cdx2 overexpression assay, showing that some iPSC/ESC lines give rise to TSC-like cells with higher efficiency than others. (G) Comparison of combined total expression values of TSC genes for all tested subclones for ESCs, EpiSCs, ECCs, and iPSCs compared with TSC values (TSC=3.0), with genetic background indicated for each cell line. Scale bar=150 μm, error bars in (B–F)=standard error among technical replicates, error bars in (G)=standard error among subclones. Color images available online at www.liebertpub.com/scd

FIG. 4.

Cdx2 overexpression induces expression of non-TSC genes. (A) Mesendoderm gene expression levels in undifferentiated cell lines. (B) Germ layer marker expression levels, relative to Hprt1, in untreated EpiSCs and in five Cdx2-overexpressing subclones and control cell lines. (C) Germ layer marker expression levels, relative to Hprt1, in untreated ESC1 and iPSC3 cells and in five Cdx2-overexpressing subclones and control cells after Cdx2 overexpression for both ESC1 and iPSC3. Error bars=standard error among technical replicates. Color images available online at www.liebertpub.com/scd

FIG. 5.

Correlation between TSC gene expression levels and markers of pluripotency. Average TSC gene expression values relative to Hprt1 and normalized to TSCs for all pluripotent stem cell lines used in this study (Table 1), except ESC3, which is deficient for Hprt1 [71], plotted against the average expression levels of the indicated pluripotency genes. The degree of correlation (r value) was calculated using Pearson's correlation. Color images available online at www.liebertpub.com/scd

FIG. 6.

Pretreatment of ESC lines in 2i leads to increased levels of TSC gene expression following Cdx2 overexpression. (A) Overview of the experiment shown in (B). (B) Expression levels of the ground state pluripotency markers, Tert, Dazl, and Myc, relative to Hprt1 following treatments described in (A). (C) Overview of the experiment shown in (D, E). (D, E) TSC gene expression values for ESC2 and ESC3 subclones after treatment described in (C), showing that pretreatment with 2i leads to increased expression of TSC genes following Cdx2 overexpression. Error bars=standard error among technical replicates. Color images available online at www.liebertpub.com/scd

FIG. 7.

Relative efficiency of TSC-like cell formation reveals a continuum of pluripotent states. In our model, there exists a continuum of pluripotent states, ranging from naïve to primed. In this study, the degree of naiveté for a given cell line is predicted by the efficiency with which it gives rise to TSC-like cells following Cdx2 overexpression. Color images available online at www.liebertpub.com/scd