Short telomeres in ESCs lead to unstable differentiation

Abstract

Functional telomeres are critical for stem cell proliferation; however, whether they are equally important for the stability of stem cell differentiation is not known. We found that mouse embryonic stem cells (ESCs) with critically short telomeres (Tert(-/-) ESCs) initiated normal differentiation after leukemia inhibitory factor (LIF) withdrawal but, unlike control ESCs, failed to maintain stable differentiation when LIF was reintroduced to the growth medium. Tert(-/-) ESCs expressed higher levels of Nanog and, overall, had decreased genomic CpG methylation levels, which included the promoters of Oct4 and Nanog. This unstable differentiation phenotype could be rescued by telomere elongation via reintroduction of Tert, via suppression of Nanog by small hairpin RNA (shRNA) knockdown, or via enforced expression of the de novo DNA methyltransferase 3b. These results demonstrate an unexpected role of functional telomeres in the genome-wide epigenetic regulation of cell differentiation and suggest a potentially important role of telomere instability in cell fate during development or disease.

Graphical abstract

Figure 1

Analysis of Pluripotency Factors in WT and Tert^−/− ESCs (A) Quantification of Nanog levels normalized over DAPI (see Figure S1G for corresponding immunofluorescence images). Note a significant shift (p < 0.0001) from Nanog-low (DAPI to Nanog-488 ≥ 1.8) to Nanog-high (DAPI to Nanog-488 < 1.5) cells in Tert^−/−S in comparison to WT and Tert^−/−R ESCs (n ≥ 100 per cell population). (B) FACS analysis of the Nanog expression profile in the same genotypes as in (A). Note the rightward shift and increase in average Nanog signal intensity in Tert^−/−S ESCs. (C) Relative gene expression analyzed by qRT-PCR, normalized to GAPDH (n = 4). Data are represented as mean ± SD. (D) (Top) Nanog protein expression with LI-COR quantification below (n = 3). Data are represented as mean ± SD; L, long telomeres (passage 30); S, short telomeres (passage 70); R, Tert rescue (70 passages, followed by clonal selection and an additional 4 passages after Tert reintroduction). The superscripts 1 and 2 indicate two independently generated Tert^−/−R colonies. (E) ChIP analysis using an antibody to H3K27me3 and H3K4me3. Relative enrichment was quantified with the use of region-specific qPCR primers for Nanog, Oct4, and Gata6 promoters. Generic IgG was used as a control (n = 3). Data are represented as mean ± SD. ^∗, p < 0.05; ^∗∗, p < 0.01; ^∗∗∗, p < 0.0001. See also Figure S1 and Table S1.

Figure 2

Differentiation Analysis of Tert^−/− ESCs (A) Bright field images of ESC populations at day 0 and day 6 in media containing 5 μM all-trans retinoic acid (ATRA) and, after ATRA removal, an additional 6 days in LIF-containing media. The micrograph bar indicates 200 μm for bright field images and15 μm for immunofluorescence images. (B) Nanog immunofluorescence analysis (green, Nanog; red, Actin). (C) Nanog protein detection by western blot. Tub, β-tubulin (n = 3). (D) qRT-PCR analysis of pluripotency genes after ATRA-induced differentiation. Gene expression at day 0 was arbitrarily set as 100, and the expression through the time course was normalized to mRNA levels at day 0. Values were expressed as a ratio to GAPDH. (E) Single-colony formation assay after ATRA treatment (6 days) is shown, and, where indicated, the readdition of LIF-containing media (6 days) (n = 3) is shown. The difference in the incidence of colony formation between Tert^−/−S (after LIF readdition) and all the other genotypes (or Tert^−/−S without LIF) was statistically significant (p < 0.0001; ANOVA and related Dunnett’s test comparing every group with Tert^−/−S values). The y axis indicates colony number. Data are represented as mean ± SD. (F) Oct4-promoter-driven GFP expression analysis of WT and Tert^−/−S ESCs post-ATRA treatment and after cell sorting for GFP-negative cells. The y axis indicates the percentage of GFP-positive WT, or Tert^−/− cells after 12 days of ATRA treatment and FACS sorting (FACS; columns 1 and 3) and after the readdition of LIF-containing media to GFP-negative sorted cells (columns 2 and 4). The difference in the incidence of GFP-positive cells between Tert^−/−S and WT cells was statistically significant (p < 0.00001; Welch’s unpaired t test). Data are represented as mean ± SD. See also Figure S2.

Figure 3

Expression of DNA Methyltransferases in ESCs Lacking Tert (A) CpG methylation analysis of the Oct4 and Nanog promoters during ATRA treatment, followed by culture in LIF-containing media. Each column represents CpGs in a sequenced clone. Full dots symbolize methylated CpGs, and empty dots symbolize unmethylated CpGs. Percentage values indicate the proportion of methylated cytosine relative to total cytosine residues (n = 10). (B) Relative quantification of global DNA methylation (n = 3) is shown. Data are represented as mean ± SD. (C) Relative gene expression of Dnmt1, Dnmt3b, and Dnmt3a2 analyzed by qRT-PCR. Values were normalized to GAPDH (n = 4). Data are represented as mean ± SD. (D) (Top) Dnmt3b protein detection by western blot and (bottom) after LI-COR quantification (n = 3). Data are represented as mean ± SD. (E) Nanog protein detection by western blot. Tub, β-tubulin (n = 5); R, Tert rescue; 3b, Dnmt3b rescue. Passage numbers are as in Figure 1. ^∗, p < 0.05; ^∗∗, p < 0.01; ^∗∗∗, p < 0.0001. See also Figure S3.

Figure 4

Differentiation Ability of Tert^−/− ESCs after Enforced Expression of Dnmt3b (A) Nanog protein detection by western blot. Tub, β-tubulin (n = 3). The first two panels on the left are reproduced from Figure 2C. (B) qRT-PCR analysis of pluripotency genes upon ATRA-induced differentiation. Gene expression at day 0 was arbitrarily set as 100 and the expression through the time course was normalized to mRNA levels at day 0. Values were expressed as a ratio to GAPDH. The first two genotypes were reproduced from Figure 2D. (C) Single-colony formation assay after the removal of ATRA and the readdition of LIF-containing media (n = 3). The difference in the incidence of colony formation between Tert^−/−S and all the other genotypes, apart from short hairpin control-transduced Tert^−/−s cells, was statistically significant (p < 0.0001; ANOVA and related Dunnett’s test comparing every group with Tert^−/−S values). The y axis indicates colony number. Data are represented as mean ± SD. (D) Chromatin immunoprecipitation analysis of H3K27me3 enrichment at the Nanog promoter, as described in Supplemental Experimental Procedures. Data are represented as mean ± SD (n = 3). ^∗, p < 0.05; ^∗∗, p < 0.01; ^∗∗∗, p < 0.0001. (E) A schematic showing that telomere shortening impairs the expression of Dnmt3 isoforms, leading to genome-wide DNA hypomethylation, which, in turn, affects H3K27me3 enrichment on specific loci (e.g., Nanog), thus impairing the ability of ESCs to sustain pluripotency factor repression after differentiation and growth restimulation. See also Figure S4.