RISC-mediated control of selected chromatin regulators stabilizes ground state pluripotency of mouse embryonic stem cells

Abstract

Background: Embryonic stem cells are intrinsically unstable and differentiate spontaneously if they are not shielded from external stimuli. Although the nature of such instability is still controversial, growing evidence suggests that protein translation control may play a crucial role.

Results: We performed an integrated analysis of RNA and proteins at the transition between naïve embryonic stem cells and cells primed to differentiate. During this transition, mRNAs coding for chromatin regulators are specifically released from translational inhibition mediated by RNA-induced silencing complex (RISC). This suggests that, prior to differentiation, the propensity of embryonic stem cells to change their epigenetic status is hampered by RNA interference. The expression of these chromatin regulators is reinstated following acute inactivation of RISC and it correlates with loss of stemness markers and activation of early cell differentiation markers in treated embryonic stem cells.

Conclusions: We propose that RISC-mediated inhibition of specific sets of chromatin regulators is a primary mechanism for preserving embryonic stem cell pluripotency while inhibiting the onset of embryonic developmental programs.

Fig. 1

a ES cell in vitro neuralization. DIV days of in vitro differentiation. 0DIV corresponds to the time of leukemia inhibitory factor (LIF) withdrawal. N2 and B27 are the supplements used in the minimal medium of differentiation. Example of bright-field microphotographs of cells at different DIV are shown on the bottom. ELA epiblast-like aggregates, NPC neural progenitor cells, NPC/Neu neural precursors, Neu differentiated neurons. b RT-PCR gene expression analysis. Values are relative to β-actin mRNA expression. Highest and lowest expression levels were normalized to 1 in the left/middle histograms and in the right histogram, respectively. c, d Oct4 and Nanog immunodetection in ES cells (c) or ELA cells (d). e Violin plot shows the distribution of green fluorescent protein (GFP) intensity in a TNG-A Nanog::GFP line [16] in LIF/serum (ES cells, red) and 24 h (green) or 48 h (blue) after LIF/serum withdrawal (ELA) or Activin/fibroblast growth factor (FGF)2 induction (EpiSC), respectively. f, g Derivation of epiblast stem cells (EpiSC) and ELA-EpiSC from ES and ELA cells, respectively. h, i EpiSC and ELA-EpiSC bright-field images. j Expression correlation of markers of pluripotency and priming between EpiSC (y-axis) and ELA-EpiSC (x-axis). Values are expressed as log₂ΔCt of RT-PCR assay; R ² coefficient of determination. k Hierarchical clustering analysis on Spearman correlation between different microarray samples. l Flow cytofluorimetric analysis of Sox1::GFP cells (46C line), indicating the ratio of GFP-positive cells (y-axis) in different cell types or times of differentiation (x-axis). m, n Immunodetection of neural markers at 7 days of ELA-EpiSC neuralization. o RT-PCR gene expression analysis as in b in ELA-EpiSC after 4 (+4DIV) or 8 (+8DIV) days from FGF2/Activin A withdrawal. Error bars in b, l, and o show standard error. In b and o *p = 0.05, **p = 0.01 (REST randomization test). Scale bars are 30 microns in a, c, and d, 40 microns in h, i, m, and n

Fig. 2

a ES cell in vitro neuralization and RNA analysis. DIV days of in vitro differentiation. b The fraction of total mRNAs that are up- or down-regulated between consecutive steps of differentiation, with a threshold of |log₂FC| > 2.5 (where FC is fold change). c Number of genes whose mRNA is significantly released (or loaded) by Argonaute (Ago; see “Methods”); ***p = 0.001 (χ²-test). IP immunoprecipitation

Fig. 3

a GO terms significantly enriched in both gene sets of mRNAs released from Ago (Ago-RIP, gray) and of mRNAs significantly loaded on ribosomes (Polysome profiling, dark blue) at the ES–ELA transition, as obtained by DAVID (see “Methods”). Fold enrichment bars are grouped according to DNA replication, chromatin regulation, and embryonic development terms; a complete list of all enriched terms in each subset is provided in Additional file 3: Table S2. b Distribution of log₂ ribosome enrichment variation (RiboΔE) during the ES–ELA transition. c, d Box plots of Ago enrichment (c) and log₂ ribosome enrichment (Ribo Enrichment) (d) of mRNAs belonging to distinct classes of chromatin regulators, as listed in Histome [25] (for a complete list of genes taken into account see also Additional file 3: Table S2). In red, families showing statistically significant differences in enrichment. HAT histone acetyl-transferases, HDAC histone deacetylases, HMT histone methyl-transferases, PRC1/2 Polycomb repressor complex 1/2. *p = 0.05, **p = 0.01, ***p = 0.001 (Wilcoxon test)

Fig. 4

a Upper panel: the distribution of Ago enrichment variation (ΔE) during the ES–ELA transition. The most negative side of the x-axis (ΔE < −5), containing few values (density < −0.002), is not shown for clarity (for comparison, see Additional file 1: Figure S1e). Middle and lower panels: comparison between log₂ variation of ribosome enrichment (RiboΔE, middle panel) or log₁₀ protein fold changes (lower panel) of three subsets of genes displaying significant Ago release (ΔE ≤ −2, green), significant Ago loading (ΔE ≥ 2, pink) or non-significant change of Ago enrichment (−1 ≤ ΔE ≤ 1, gray). Asterisks indicate the p value (Student’s t-test, ***p = 0.001) against the null hypothesis of mean equal to 0. b–d Total mRNA fold change (Tot mRNA), linear Ago enrichment (Ago En), ribosome enrichment (Ribo En), and MS fold change (Protein) for the members of DNMT, KDM, and SWI/SNF classes of chromatin regulators detected in ES and ELA nuclear cell extracts. The plus sign marks values that are slightly out of the ranges indicated in “Methods” but are still relevant. Other genes unrelated to these families but displaying similar regulation are listed in Additional file 4: Table S3. **p = 0.01, ***p = 0.001 (Wilcoxon test). Error bars show standard error

Fig. 5

a Flow cytofluorimetry analysis of TNG-A Nanog::GFP cells at 4 DIV. b GFP-positive cell ratios at 2 DIV and 4 DIV after inhibition of SmarcA4 (shSA4), DNMT (AZA), or KDM (CHQ) compared with control (Ctrl). c, d Similar analysis as in a and b, with a mESC line carrying GFP under the human Nanog promoter (HNP). e, f RT-PCR gene expression analysis. Values are relative to β-actin mRNA expression. The lowest and highest expression levels were normalized to 1 in the left and right histograms, respectively. g Nanog proximal promoter methylation. The scheme shows the region of the mouse Nanog promoter amplified for bisulfite-treated DNA sequencing (bis-seq). Black circles in grid rows indicate CpG methylation sites. The histogram shows the percentage of CpG methylation in different culture conditions as above. *p = 0.05, **p = 0.01, ***p = 0.001 (b, d, g, Student’s t-test; e, f, REST randomization test). Error bars show standard error

Fig. 6

a CRISPR/Cas9 guide design: the single guide RNA (sgRNA; red square) was chosen to target a genomic region which is conserved between Ago1 and Ago2 but avoiding off-targets. b Western blot of Ago* proteins in control (Ctrl) or Ago1–2 CRISPR ES cells 4 days after transduction. c Western blot of Dnmt3b, SmarcA4, Kdm2b, Nanog, and GAPDH protein levels in ES cells cultured in 2i medium 4 days after transduction with Ago1–2 CRISPR lentiviral vector compared with control cells transduced with non-targeting CRISPR vector (Ctrl CRISPR). d mRNA levels of Dnmt3b, SmarcA4, and Kdm2b of cells as in c. e, f mRNA levels of pluripotency (e) or early neural commitment (f) in cells as in c and d 8 days after transduction. g, h GFP immunodetection in a 46C Sox1::GFP mESC line cultured in 2i medium 8 days after transduction with Ctrl CRISPR (g) or Ago1–2 CRISPR lentiviral vector (h). Scale bars, 50 microns. i, j Cell count (i) and cytofluorimetric analysis (j) of GFP-positive cells as in g and h. *p = 0.05, **p = 0.01, ***p = 0.001 (e–f, REST randomization test; i, Student’s t-test). Error bars show standard error

Fig. 7

a, b RT-PCR analysis of mRNA levels of pluripotency (a) or early neural commitment (b) markers in Dicer flox/- conditional ES cells [33] 8 days after transduction with a CRE-GFP carrying lentiviral vector. Values are relative to β-actin mRNA expression. Expression levels were normalized to GFP-transduced cells. c Log₂ fold change (FC; y-axis) and mean value of miRNA expression (x-axis) between ES and ELA cells. The blue dashed box indicates miRNAs significantly down-regulated (set “D”; mean log₂ RPM ≥ 5 and log₂ fold change ≤ −1) during priming. d Z-scores of the expression of miRNAs predicted to bind top Ago-released (ΔE < −2; left panel) or top Ago-loaded (ΔE > 2; right panel) mRNAs during the ES–ELA transition. Global miRNA/mRNA binding prediction was performed by the miRVestigator framework, which is designed to take as input a list of co-expressed genes and return the miRNA most likely regulating these genes [82]. miRNA expression was evaluated by small RNA-seq of ES cells cultured in 2i medium and LIF (ES#), ES cells cultured in LIF/serum (ES), ELA cells obtained from ES in 2iL (ELA#), and ELA obtained from ES. Ago enrichment variation provides a good estimation of miRNAs changing at the ES–ELA transition. e Z-scores of expression (RPM) of miRNAs selected in the blue box in c (set “D”), in ES cells cultured in 2i medium and LIF (ES#), ES cells cultured in LIF/serum (ES), ELA cells obtained from ES in 2iL (ELA#), and ELA obtained from ES (ELA). f The distribution of predicted binding affinity, calculated as cumulative Miranda scores of set “D” miRNAs in the 3′ untranslated region (UTR) of genes from the indicated families. The dashed red line marks the median score of a random gene set. g–i Normalized enhanced GFP (EGFP)/DsRed fluorescence ratios as obtained by co-transfection of plasmids and miRNA mimics/controls in ES and ELA cells (see “Methods”). All values, normalized on a DsRed plasmid as an internal control for transfection efficiency, are relative to the ratio displayed by ES cells transfected with an EGFP plasmid devoid of any 3′ UTR. AU arbitrary units.*p = 0.05, **p = 0.01, ***p = 0.001; a, b, REST randomization test; d–f, Wilcoxon test between pairs of conditions (d, e) or between each family and a randomized set of 3′ UTRs (f; see “Methods”); g–i, Student’s t-test

Fig. 8

Two-layer model of epigenetic control of pluripotency. A marked release from RISC characterizes the ES–ELA transition (priming), whereas transcriptional regulation occurs preferentially at the ELA–NPC transition (neuralization). Our data suggest a causal link between the two phenomena as the pool of de-repressed genes contains distinct chromatin regulators which could overcome the epigenetic barrier between the primed state and ground-state pluripotency. Thus, ground-state miRNAs, which are down-regulated during priming, could shield naïve pluripotent cells from the epigenetic transition required for the onset of embryonic differentiation programs