Naive Human Embryonic Stem Cells Can Give Rise to Cells with a Trophoblast-like Transcriptome and Methylome

Abstract

Human embryonic stem cells (hESCs) readily differentiate to somatic or germ lineages but have impaired ability to form extra-embryonic lineages such as placenta or yolk sac. Here, we demonstrate that naive hESCs can be converted into cells that exhibit the cellular and molecular phenotypes of human trophoblast stem cells (hTSCs) derived from human placenta or blastocyst. The resulting "transdifferentiated" hTSCs show reactivation of core placental genes, acquisition of a placenta-like methylome, and the ability to differentiate to extravillous trophoblasts and syncytiotrophoblasts. Modest differences are observed between transdifferentiated and placental hTSCs, most notably in the expression of certain imprinted loci. These results suggest that naive hESCs can differentiate to extra-embryonic lineage and demonstrate a new way of modeling human trophoblast specification and placental methylome establishment.

Keywords: DNA methylation; amnion; development; differentiation; embryonic stem cells; epigenetics; placenta; pluripotency; trophoblast.

Graphical abstract

Figure 1

Similarity of hTSCs to reported ITGA2⁺ EpCAM⁺ Progenitor Population (A) Expression of ITGA2 and EPCAM in hTSCs and primary placenta cells. Data are taken from Okae et al. (2018), with n = 3–4 independent experiments per cell type. (B) Seventy-four genes were identified as upregulated in ITGA2⁺ cells by Lee et al. (2018) and also present in Okae et al.'s RNA-seq dataset. The expression of these genes is plotted using RNA-seq data from Okae et al., with each gene represented as a single point in the boxplot. (C) Flow cytometry plot of ITGA2 and EpCAM in CT1 hTSCs and EVT. Representative of n = 5 independent experiments. (D) Downregulation of ITGA2 upon directed differentiation of CT1 (qRT-PCR, mean + SE of n = 2 independent experiments).

Figure 2

Transdifferentiation of hESCs to Putative hTSCs and Purification via FACS Sorting (A) Upper panel: light microscopy image of a colony of primed WIBR3 OCT4-ΔPE-GFP hESCs, with the colony circled with a dashed line. Lower panel: lack of GFP signal. (B) Light and fluorescent image of WIBR3 OCT4-ΔPE-GFP hESCs after 10 days of naive reversion. Many GFP⁺ colonies are present, with one representative colony circled. (C) Naive hESCs after 4 days of culture in hTSC medium. Note the presence of a colony of epithelial cells and the loss of GFP signal. (D) Immunofluorescent image of WIBR hESCs after 10 days in hTSC medium. Note a distinct population of TFAP2C⁺ TEAD4⁺ KRT7⁺ cells. (E) FACS of WIBR3 hESCs grown in hTSC medium for 11 days and comparison CT1 hTSCs. Note distinct population of ITGA2^hi EpCAM^hi cells which was sorted to produce WIBR3 tdhTSC line 1. (F) Light microscopy of CT1 and WIBR3 tdhTSC line 1. (G) Flow cytometry of WIBR3 OCT4-ΔPE-GFP hESCs in naive (green) and primed (black) conditions are overlaid. GFP^hi and GFP^lo cells from naive culture, populations indicated with boxes, were sorted into hTSC medium. (H) Flow cytometry of GFP^hi and GFP^lo cells after 16 days in hTSC medium. Note much higher EpCAM^hi ITGA2^hi population in the GFP^hi population. (I) Immunofluorescent staining for hTSC/CTB (TEAD4) and pan-placental (TFAP2C, KRT7) markers in lines indicated. Representative of n = 2 independent experiments.

Figure 3

Validation that Transdifferentiated hTSCs Express Placental Genes (A) Gene expression of indicated markers in each sample type are indicated by coloration. Expression from replicates of each sample type are averaged. n = 2 (all tdhTSC lines), n = 3 (UCLA1 hESC), n = 4 (WIBR3 hESCs, CT3, BT2), n = 7 (CT1) biological replicates. (B) Same as (A) except with a different color scale. (C) Principal-component analysis for gene expression of the lines indicated. Each dot is one biological replicate. (D) Expression of 89 trophoblast-specific genes, as identified by analysis of pre-implantation primate embryos (Nakamura et al., 2016), is indicated for each cell type. Expression of each gene, using an average of all replicates for a given cell type, is indicated as a single point on the violin plot. Box indicates 25th, 50th, and 75th percentiles. n = 1 (epithelial cell), n = 2 (all tdhTSC lines, naive hESCs), n = 3 (UCLA1 hESCs), n = 4 (WIBR3 hESCs, CT3, BT2), n = 7 (CT1) biological replicates.

Figure 4

Differentiation Capacity of tdhTSCs (A) qRT-PCR for markers of EVT (GCM1, HLA-G), and STB (GCM1, CGB7) differentiation of lines indicated, normalized to GAPDH. Error bars indicate mean + SE for n = 3–5 independent experiments. ^∗p < 0.05 in one-tailed t test. (B) ELISA assay for hCG secretion after directed differentiation to STB. Error bars indicate mean + SE for n = 2 biological replicates, except CT1 for which there is one replicate. (C) Flow cytometry of hTSCs and EVT, for the EVT markers HLA-G and ITGA1, for the lines indicated. Representative of n = 5 independent experiments.

Figure 5

Global Methylation Patterns of tdhTSC For all data, data in parentheses indicate data mined from published sources, data without parentheses indicate original data. (A) Global CpG methylation level in samples indicated. (B) Correlation of CpG island methylation in each of the two samples is indicated. Each CpG island represents as a single point, all CpG islands with adequate coverage are plotted. (C) Correlation of promoter methylation between two samples is indicated. All autosomal promoters with adequate coverage are plotted. (D) Violin plot indicating degree of CpG island methylation in samples indicated among 788 CpG islands that show higher DNA methylation in CT1 hTSCs relative to primed hESCs. (E) DNA methylation of a region of genome that includes the CpG island promoter of NODAL. Height of bars corresponds to percentage CpG methylation, from 0% to 100%. Data from WIBR3 tdhTSC L1 and L2 are merged to allow sufficient sequencing depth for visualization. (F) Violin plot showing methylation level of 2,107 promoters that show higher DNA methylation in primed hESCs relative to CT1 hTSCs. (G) Expression of 172 possible “gatekeeper” genes, genes that have upregulated expression in hTSCs and higher promoter methylation in hESCs. Expression of each gene, using an average of all replicates for a given cell type, is indicated as a single point on the violin plot. Box indicates 25th, 50th, and 75th percentiles. n = 2 (all tdhTSC lines), n = 3 (UCLA1 hESCs), n = 4 (WIBR3 hESCs, CT3, BT2), n = 7 (CT1) biological replicates.

Figure 6

Imprinting Abnormalities in tdhTSC (A) CpG methylation level of all universal (organism-wide) and placental imprints in cell types indicated. CTB, primary cytotrophoblasts; L1, WIBR3 tdhTSC L1; L2, WIBR3 tdhTSC L2. CTB data from Hamada et al. (2016), original data shown for all other samples. (B) Volcano plot of genes differentially expressed in hTSCs (CT1, CT3, and BT2) versus primed (WIBR3 and UCLA1 hESCs). Red dots correspond to differentially expressed genes (p_adj < 0.05, fold-change > 4). (C) Volcano plot of genes differentially expressed in hTSCs (CT1, CT3, and BT2) versus tdhTSCs (WIBR3 tdhTSC lines 1, 2, and 3, and UCLA1 line 1). Red dots correspond to differentially expressed genes (p_adj < 0.05, fold-change > 4).

Figure 7

Comparative Transdifferentiation Capacity from Different Media Conditions (A) Flow cytometry of UCLA1 hESCs cultured in various conditions and then grown in hTSC medium for 15 days. Percentage of ITGA2^hi EpCAM^hi cells is indicated. (B) Flow cytometry 20 days after sorting. Note higher ITGB6 signal in putative tdhTSCs derived from hESCs in non-naive conditions. (C) Flow cytometry histogram of ITGB6 signal for putative tdhTSCs derived from hESC lines, and growth condition indicated. (D) qRT-PCR quantification of placental markers and ITGB6 in indicated cell lines normalized to GAPDH. W3, WIBR3; U1, UCLA1. ^∗WIBR3 EPS-derived tdhTSCs could not be studied because conversion efficiency was too low to produce a pure ITGA2^hi EpCAM^hi line. n = 1 biological replicate.