Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 Dec 6;17(10):2700-2714.
doi: 10.1016/j.celrep.2016.11.032.

Deletion of the Polycomb-Group Protein EZH2 Leads to Compromised Self-Renewal and Differentiation Defects in Human Embryonic Stem Cells

Affiliations
Free PMC article

Deletion of the Polycomb-Group Protein EZH2 Leads to Compromised Self-Renewal and Differentiation Defects in Human Embryonic Stem Cells

Adam Collinson et al. Cell Rep. .
Free PMC article

Abstract

Through the histone methyltransferase EZH2, the Polycomb complex PRC2 mediates H3K27me3 and is associated with transcriptional repression. PRC2 regulates cell-fate decisions in model organisms; however, its role in regulating cell differentiation during human embryogenesis is unknown. Here, we report the characterization of EZH2-deficient human embryonic stem cells (hESCs). H3K27me3 was lost upon EZH2 deletion, identifying an essential requirement for EZH2 in methylating H3K27 in hESCs, in contrast to its non-essential role in mouse ESCs. Developmental regulators were derepressed in EZH2-deficient hESCs, and single-cell analysis revealed an unexpected acquisition of lineage-restricted transcriptional programs. EZH2-deficient hESCs show strongly reduced self-renewal and proliferation, thereby identifying a more severe phenotype compared to mouse ESCs. EZH2-deficient hESCs can initiate differentiation toward developmental lineages; however, they cannot fully differentiate into mature specialized tissues. Thus, EZH2 is required for stable ESC self-renewal, regulation of transcriptional programs, and for late-stage differentiation in this model of early human development.

Keywords: differentiation; epigenetics; histone methylation; pluripotency.

Figures

None
Figure 1
Figure 1
Targeted Deletion of EZH2 in hESCs (A) Overview of EZH2 structure and targeting strategy. Exons encoding CXC and SET domains are indicated. The gRNA sequence is underlined and protospacer adjacent motif highlighted in red. DNA sequence of the deletions in one EZH2–/+ ESC line and one EZH2–/– ESC line is shown for both alleles. Mutation causes frameshift and premature stop codon. An additional line is shown in Figures S1 and S2. (B) mRNA expression levels from RNA-seq data revealing EZH2, EED, and SUZ12 transcript levels in EZH2+/+, EZH2–/+, EZH2–/–, and EZH2–/–+EZH2 ESCs. Data show mean ± SD; n = 3 biological replicates. (C) Immunofluorescent microscopy of colonies from EZH2–/– ESCs and control ESCs. This analysis reveals a strong reduction in EZH2 levels in EZH2–/– ESCs. The antibody was raised against a C-terminal epitope of EZH2; similar results were obtained using an alternative antibody raised against the N-terminal of EZH2 (Figure S2E). OCT4 expression indicates undifferentiated cells within a hESC colony. Arrows point to MEF. Scale bars, 100 μm. (D) EZH2, SUZ12, and the main isoform of EED are undetectable in EZH2–/– ESCs by western blot analysis and are restored upon expression of a DOX-induced EZH2 transgene. β-ACTIN is the loading control. Mass is in kilodaltons. (E) H3K27me3 and H3K27me2 levels are reduced to background levels, and H3K27me1 levels are partially reduced, in EZH2–/– ESCs. OCT4 expression in inset indicates undifferentiated ESCs within the field of view. Arrows point to MEF. Scale bars, 100 μm.
Figure 2
Figure 2
EZH2 Deficiency in hESCs Results in Loss of H3K27me3 (A) Quantitative trend plot of H3K27me3 normalized ChIP-seq reads over gene body ±5 kb. High CpG (HCP), intermediate CpG (ICP), and low CpG (LCP) promoters are shown separately. (B) Scatterplot of H3K27me3 (x axis) and H3K4me3 (y axis) normalized ChIP-seq reads in EZH2–/– relative to EZH2+/+ (left) and relative to EZH2–/– +EZH2 (center), and EZH2–/– +EZH2 versus EZH2+/+ (right). All transcriptional start sites (TSS) shown in gray; TSS that are positive for H3K27me3 in EZH2+/+ ESCs highlighted in blue. Disruption of EZH2 leads to a strong reduction in H3K27me3 levels at TSS, with little effect on H3K4me3 levels. Expression of a DOX-mediated EZH2 transgene in the EZH2-deficient cells causes restoration of H3K27me3 levels to levels equivalent to EZH2+/+. (C) ChIP-seq tracks of HOXB (left) and HOXD (right) loci illustrate the loss of H3K27me3 in EZH2–/– ESCs compared to control ESCs. H3K4me3 is relatively unaffected. All ChIP-seq data represent the average of three biological replicates for each cell line. These results were confirmed independently by qPCR analysis of ChIP DNA at several gene promoters (Figure S3C).
Figure 3
Figure 3
Genes Encoding Developmental Regulators Are Transcriptionally Derepressed in EZH2-Deficient hESCs (A) RNA-seq heatmap for EZH2–/– ESCs and control ESCs (three biological replicates per line). Shown are all differentially expressed genes between EZH2–/–and EZH2–/–+EZH2 ESCs. (B) Top GO terms of differentially expressed gene sets. Numbers of genes are shown; example genes within each GO category are listed (right). Corrected p values were calculated using a modified Fisher’s exact test followed by Bonferroni’s multiple comparison test. (C) Gene set enrichment analysis of PRC2 targets (n = 1,299; defined by high EZH2 and H3K27me3 promoter-localized ChIP-seq values in EZH2+/+ ESCs) in genes that have been ranked according to their fold change in transcription between EZH2–/– ESCs and EZH2–/–+EZH2 ESCs. The positive enrichment score (ES) reveals that genes selectively derepressed in the absence of EZH2 are enriched in PRC2 targets (p < 0.001; Kolmogorov-Smirnov statistic). (D) Genes within the upregulated category have higher levels of promoter-localized EZH2 (upper) and H3K27me3 (lower) in EZH2+/+ ESCs compared to an expression-matched set of genes and to downregulated genes. Data were compared using a Kruskal-Wallis test followed by Dunn’s multiple comparison test. (E) Genes encoding developmental regulators are transcriptionally derepressed in EZH2-deficient ESCs. A subset of direct EZH2 target genes is depicted as family groups. The color of each circle represents the log2 fold change in EZH2–/– ESCs relative to EZH2–/–+EZH2 ESCs. The size of each circle represents the expression value of the gene in EZH2–/– cells. A similar pattern of target gene derepression is observed when comparing EZH2–/– ESCs with EZH2+/+ ESCs (Figure S4C). (F) ChIP-seq and mRNA-seq tracks of four genes encoding key developmental regulators illustrate the association between loss of H3K27me3 and transcriptional upregulation in EZH2–/– ESCs.
Figure 4
Figure 4
Transcriptional Derepression Occurs Predominantly within Discrete Lineage-Specific Programs (A) Single-cell RNA-seq expression levels for six example PRC2-target genes in EZH2–/– ESCs and EZH2+/+ ESCs, where each dot represent the results from a single cell. A pluripotency gene (LIN28) and housekeeping gene (HMBS) are shown for comparison. Robust upregulation of PRC2-target genes occurs in a subset of EZH2–/– ESCs. (B) Heatmap of single-cell RNA-seq expression for EZH2–/– ESCs (right) and EZH2+/+ ESCs (left). Each column represents an individual cell. Each row represents an individual gene, grouped into three clusters corresponding to endoderm, mesoderm, and ectoderm cell lineages. Shown are PRC2-target genes from within the hESC scorecard assay, which is an assay that can classify differentiated cell lineages (Bock et al., 2011). Subsets of cells (boxed in purple) tend to mis-express many genes from within one lineage but rarely mis-express multiple genes derived from more than one lineage.
Figure 5
Figure 5
EZH2-Deficient hESCs Are Compromised in Self-Renewal and Proliferation (A) Phase contrast images show representative colonies of EZH2+/+, EZH2–/–, and EZH2–/–+EZH2 ESC lines. Note the variable morphology of EZH2-deficient colonies. Scale bars, 100 μm. (B) EZH2–/– ESCs show reduced ESC colony formation when plated as single SSEA4-positive cells at low density (6,000 cells seeded per well). Data show mean ± SD; n = 3 biological replicates. Data were compared using a one-way ANOVA followed by Bonferroni’s multiple comparison test (∗∗∗p < 0.0005). Representative AP staining is shown underneath. (C) EZH2–/– ESCs have reduced capacity to self-renew when plated at clonal density. ESC colonies were categorized as undifferentiated, mixed, or differentiated based on AP activity; examples shown underneath. Data show mean ± SD; n = 3 biological replicates. Over 150 colonies were scored for each cell line. Data were compared using a one-way ANOVA followed by Bonferroni’s multiple comparison test (∗∗∗p < 0.0005; ∗∗p < 0.005; p < 0.05). Scale bars, 100 μm. (D) Immunofluorescent microscopy for OCT4 (undifferentiated marker) and SOX17 (early differentiation marker) reveals an increased prevalence for mixed and fully differentiated colonies in EZH2–/– compared to control ESC lines. Representative images are shown underneath. Over 100 colonies were scored for each cell line. Scale bars, 100 μm. (E) Growth curve over 16 days reveals a significant proliferation defect in EZH2–/– ESCs compared to control ESCs. Data show mean ± SD; n = 3 biological replicates. Data were compared between EZH2–/– ESCs and each control ESC line using one-way ANOVA followed by Bonferroni’s multiple comparison test (p < 0.05 for each comparison). (F) Mitotic index was calculated for each ESC line by dividing the number of H3S10ph-positive cells by the total number of cells within a colony. The analysis was restricted to undifferentiated colonies (determined by OCT4 expression) of similar size in order to control for potential differences in cell state. Over 1,000 cells were scored for each cell line. Data show mean ± SD; n = 3 biological replicates. Data were compared using a one-way ANOVA followed by Bonferroni’s multiple comparison test (∗∗∗p < 0.0005). Scale bars, 100 μm.
Figure 6
Figure 6
EZH2-Deficient hESCs Can Initiate Differentiation but Are Severely Impaired in Generating Mature Cell Types (A) EZH2–/– ESCs fail to generate teratomas. Mass of teratoma and kidney samples for indicated ESC lines, with images shown underneath (scale bar, 5 mm). Additional images and histology analysis are provided in Figures S6A and S6B. (B) ESCs were induced to differentiate with retinoic acid for 5 days. Cell counts (upper) and crystal violet stain (lower) reveal that few EZH2–/– ESCs remain after 5 days compared to control ESCs. Short bars indicate mean values for the two biological replicates. (C) EZH2-deficient ESCs can generate early endoderm cells. Upper panel shows cell counts over endoderm differentiation time course. Lower panel shows flow cytometry analysis of endoderm markers KIT/CXCR4 in undifferentiated EZH2+/+ ESCs (black), day 5 endoderm differentiated EZH2+/+ (blue), undifferentiated EZH2–/– ESCs (red), and day 5 endoderm differentiated EZH2–/– (purple). Inset numbers show percentage positive cells for each cell population (mean of three biological replicates, with range). (D) EZH2-deficient ESCs can generate early mesoderm cells. Upper panel shows cell counts over 6 days of mesoderm differentiation. Lower panel shows flow cytometry analysis of mesoderm markers KDR/PDGFRα. (E) RT-qPCR analysis of endoderm, mesoderm, ectoderm and pluripotency genes in undifferentiated (black) and day 5 endoderm differentiated (blue) EZH2+/+ ESCs, and undifferentiated (red) and day 5 endoderm differentiated (purple) EZH2–/– ESCs. Note that POU5F1 and NANOG are also associated with ESC differentiation (Loh and Lim, 2011), which may underlie their elevated expression patterns in EZH2–/– ESCs. (F) qRT-PCR analysis of undifferentiated and day 2 mesoderm differentiated ESCs. (G) EZH2-deficient ESCs can generate early ectoderm cells, but with significantly reduced efficiency compared to EZH2+/+ ESCs. Upper panel shows cell counts over 10 days of ectoderm differentiation. Lower panel shows flow cytometry analysis of ectoderm marker CD56 and undifferentiated ESC marker CD326. Note the significantly decreased efficiency of ectoderm differentiation in EZH2–/– ESCs compared to EZH2+/+ ESCs (p = 0.01; unpaired two-sided t test). (H) RT-qPCR analysis of undifferentiated and day 10 ectoderm differentiated ESCs. For all panels, data show mean ± SEM of three biological replicates and were compared using an unpaired two-sided t test (p < 0.05).
Figure 7
Figure 7
Proposed Model Summarizing the Role of EZH2 in Regulating Transcriptional Programs and Cell Differentiation in hESCs EZH2+/+, above; and EZH2–/–, below. EZH2 deficiency leads to loss of PRC2, transcriptional derepression of developmental regulators and self-renewal defects in hESCs. Substantial cell loss (red crosses) and gene mis-regulation is observed upon differentiation of EZH2–/– ESCs (e.g., to endoderm in example shown).

Similar articles

See all similar articles

Cited by 25 articles

See all "Cited by" articles

References

    1. Azuara V., Perry P., Sauer S., Spivakov M., Jørgensen H.F., John R.M., Gouti M., Casanova M., Warnes G., Merkenschlager M., Fisher A.G. Chromatin signatures of pluripotent cell lines. Nat. Cell Biol. 2006;8:532–538. - PubMed
    1. Bernstein B.E., Mikkelsen T.S., Xie X., Kamal M., Huebert D.J., Cuff J., Fry B., Meissner A., Wernig M., Plath K. A bivalent chromatin structure marks key developmental genes in embryonic stem cells. Cell. 2006;125:315–326. - PubMed
    1. Bock C., Kiskinis E., Verstappen G., Gu H., Boulting G., Smith Z.D., Ziller M., Croft G.F., Amoroso M.W., Oakley D.H. Reference Maps of human ES and iPS cell variation enable high-throughput characterization of pluripotent cell lines. Cell. 2011;144:439–452. - PMC - PubMed
    1. Boyer L.A., Plath K., Zeitlinger J., Brambrink T., Medeiros L.A., Lee T.I., Levine S.S., Wernig M., Tajonar A., Ray M.K. Polycomb complexes repress developmental regulators in murine embryonic stem cells. Nature. 2006;441:349–353. - PubMed
    1. Bracken A.P., Pasini D., Capra M., Prosperini E., Colli E., Helin K. EZH2 is downstream of the pRB-E2F pathway, essential for proliferation and amplified in cancer. EMBO J. 2003;22:5323–5335. - PMC - PubMed

Publication types

MeSH terms

Substances

LinkOut - more resources

Feedback