p53 regulates cell cycle and microRNAs to promote differentiation of human embryonic stem cells

Abstract

Multiple studies show that tumor suppressor p53 is a barrier to dedifferentiation; whether this is strictly due to repression of proliferation remains a subject of debate. Here, we show that p53 plays an active role in promoting differentiation of human embryonic stem cells (hESCs) and opposing self-renewal by regulation of specific target genes and microRNAs. In contrast to mouse embryonic stem cells, p53 in hESCs is maintained at low levels in the nucleus, albeit in a deacetylated, inactive state. In response to retinoic acid, CBP/p300 acetylates p53 at lysine 373, which leads to dissociation from E3-ubiquitin ligases HDM2 and TRIM24. Stabilized p53 binds CDKN1A to establish a G(1) phase of cell cycle without activation of cell death pathways. In parallel, p53 activates expression of miR-34a and miR-145, which in turn repress stem cell factors OCT4, KLF4, LIN28A, and SOX2 and prevent backsliding to pluripotency. Induction of p53 levels is a key step: RNA-interference-mediated knockdown of p53 delays differentiation, whereas depletion of negative regulators of p53 or ectopic expression of p53 yields spontaneous differentiation of hESCs, independently of retinoic acid. Ectopic expression of p53R175H, a mutated form of p53 that does not bind DNA or regulate transcription, failed to induce differentiation. These studies underscore the importance of a p53-regulated network in determining the human stem cell state.

Figure 1. p53 protein is induced during differentiation of hESCs.

(A) qRT-PCR. RNA from hESCs treated with RA in medium without FGF for 5 d (R0–R5), were subjected to qRT-PCR assay (data are presented as mean ± SEM). (B and C) Western blot. Lysates (total cell lysate [TCL]) prepared from hESCs cultured as in (A) were analyzed by Western blotting, the blots in (B) were quantitated (C): the average density of three different blots is plotted (*, p<0.05). (D) TP53 qRT-PCR. RNA samples prepared as in (A) were subjected to qRT-PCR assay (mean ± SEM). (E) Immunofluorescence. hESCs in complete CM or treated with RA for 3 d were stained with antibodies against p53 and OCT4, and nuclei were stained with DAPI. Scale bar is 50 µm. (F) p53 nuclear localization. Nuclear extracts prepared from hESCs cultured as in (A) were analyzed by Western blotting. (Also see Figure S1.).

Figure 2. Acetylation of Lys373 leads to stabilization of p53.

(A) p53 acetylation. Equal amounts of p53 were immunoprecipitated by adjusting the amounts of total cell lysates prepared from hESCs and probed with p53K373ac antibody. (B) Immunofluorescence. hESCs treated with RA for 3 d were stained with antibodies against p53K373ac and OCT4; nuclei were stained with DAPI. (C) Co-immunoprecipitation. Cell lysates from RA-treated hESCs were immunoprecipitated with p53 followed by Western blotting. (D) p53 acetylation. p53 immunoprecipitated from hESCs cultured under differentiating conditions and treated with either circumin on day 2 or nicotinamide on day 4 and probed with p53K373ac antibody. (E) Co-immunoprecipitation. Cell lysates from differentiating hESCs were immunoprecipitated with HDM2 or TRIM24 antibodies and analyzed by Western blotting. (F) Endogenous p53 ubiquitination. hESCs cultured under differentiating conditions were treated with MG132 + RA or MG132 + Adr; endogenous p53 was immunoprecipitated and probed for ubiquitin (top panel). Same blot was reprobed with p53 antibody to show the equal p53 pull down (bottom panel). (Also see Figure S2.) IP, mmunoprecipitation; Ub, ubiquitin; WB, Western blot.

Figure 3. The consequence of p53 accumulation in hESCs.

(A) Cell cycle analysis. hESCs transfected with non-target (siControl) or siRNA specific to p53 (siTP53) or p21 (siCDKN1A) and treated with RA were stained with PI and subjected to flow cytometry analysis to determine DNA content. Results quantitated as fold change in cell cycle are shown. (B) qRT-PCR. RNA from hESCs treated with RA for 4 d or Adr for 6 h were subjected to qRT-PCR assay using primers specific for human CDKN1A. (C) ChIP. p53-bound chromatin was immunoprecipitated from hESCs, and p53 enrichment on CDKN1A was analyzed by qRT-PCR using primers encompassing p53REs (*, p<0.05). Scheme representing location of p53RE and primers used for ChIP-qRT-PCR are shown on the top (asterisk indicates the 3′ end of the gene). (D) hESCs treated with RA or Adr were lysed, and cell lysates were blotted for γ-H2AX. (E) Apoptosis assay. hESCs treated as in (D) were stained with Annexin V and PI, and percent apoptotic cells was determined by flow cytometry (mean ± SEM). (Also see Figures S3 and S4.).

Figure 4. p53 drives differentiation of hESCs.

(A) AP staining. hESCs transfected with non-target (siControl) or siRNA specific to p53 (siTP53) or p21 (siCDKN1A) were treated with RA and stained for AP (blue colonies). (B and C) hESCs transfected and treated as in (A) were used in Western blotting (B) and qRT-PCR (C) analyses. The blots in (B) were quantitated, and average density of three different blots is plotted (bottom panel) (*, p<0.05) (mean ± SEM). (D) OCT4 + SSEA4 staining. hESCs transfected with siRNA followed by RA treatment were stained for SSEA4 and OCT4 and subjected to dual flow cytometry analysis. Triplicate samples were analyzed in each experiment, and data analyzed with FACSDiva software. Decreases in fraction of OCT4-positive cells, as compared to siControl untreated, are indicated in red. (E) AP staining. hESCs transfected with HDM2 or TRIM24 siRNA were treated with RA and stained for AP. (F) Quantified AP-stained colonies. Date shown are for 50 colonies per treatment in three separate experiments (in [A] and [E]), scored as undifferentiated, partially differentiated, or fully differentiated colonies, mean ± SEM. (G) OCT4 + SSEA4 staining and flow cytometry analysis as in (D) after transfection with siRNA targeting HDM2 or TRIM24. (H) Cell cycle analysis. hESCs transfected with HDM2 or TRIM24 siRNA were stained with PI and subjected to cell cycle analysis. (Also see Figures S3 and S5.).

Figure 5. DNA binding activity of p53 is required to induce differentiation of hESCs.

(A) hESCs stably expressing p53WT and mutant p53 (p53R175H and p53R175P) under control of tet-inducible promoter cultured in CM + FGF were treated with 100 ng/ml Dox for 2 d. p53 and OCT4 protein levels were analyzed by blotting. (B) AP staining. hESCs in (A) treated with Dox for 2 d (2D) or 4 d (4D) and AP stained. (Arrows indicate differentiated cells.) (C) qRT-PCR. hESCs treated with Dox for 1 d (1D) or 2 d (2D). RNA analyzed by qRT-PCR assay for expression of exogenous TP53, CDKN1A, OCT4, and AFP (*, p<0.05) (mean ± SEM). (D) Cell cycle analysis. hESCs treated with RA for 1 d or Dox for 1 d or 2 d, stained with PI, and subjected to cell cycle analysis (mean ± SEM). (E) hESCs stably expressing p53WT were transfected with siRNA and treated with Dox for 2 d. p53, p21, and OCT4 protein levels were analyzed. (F and G) hESCs expressing p53WT were treated with Dox for 2 d or 4 d, and lysed to analyze protein (F) or RNA (G) for various differentiation markers; AFP and GATA4 (endoderm), Brachyury (mesoderm), and PAX6 (ectoderm). (H) p53 acetylation. Lysates from hESCs treated with RA and Dox-inducible p53WT treated with Dox were blotted for p53K373ac and p53. (Also see Figure S6.).

Figure 6. p53 regulates miR-34a and miR-145 to drive differentiation of hESCs.

(A and B) miRNA-TaqMan assay. miRNAs were analyzed using total RNA from hESCs with probes specific for human miR-34a and miR-145 and were normalized to RNU6B as internal control (*, p<0.01). (B) Total RNA prepared from hESCs transfected with siRNA was analyzed for expression of miR-34a and miR-145. Data are presented as mean ± SEM. (C) ChIP. p53-bound chromatin was immunoprecipitated from hESCs, and p53 enrichment on miR-34a and miR-145 promoters was analyzed by qRT-PCR (*, p<0.05). Scheme representing location of p53RE and primers used for ChIP-qRT-PCR are shown on the top (asterisk indicates the 3′ end of the gene). Data are presented as mean ± SEM. (D and E) Anti-miRNA assay. hESCs transfected with anti-miRNA oligonucleotides for non-specific (NS), miR-34a, and miR-145 were subjected to qRT-PCR assay (D) and Western blotting (E) (mean ± SEM). (F) OCT4 + SSEA4 staining. hESCs transfected as in (D) followed by RA treatment for 3 d were stained for SSEA4 and OCT4 and subjected to flow cytometry analysis. Decreases in fraction of OCT4-positive cells, as compared to untreated control, are indicated in red. (G) Summary of miR-34a target sites in the 3′ UTRs of KLF4 and LIN28A. D, dog; H, human; M, mouse; R, rat. The underlined nucleotides in miRNA target sites were mutated in the mutant 3′ UTR constructs. (H) Luciferase assay. HEK293 cells were transfected with luciferase plasmids containing wild-type (WT) or mutated (Mut) 3′ UTRs along with miRNA precursors for scrambled or miR-34a. Relative luciferase levels were calculated, and Student's t test was used to compare the datasets (*, p<0.01). Error bars represent standard deviation for three independent experiments. (Also see Figure S7.).

Figure 7. Model depicting role of p53 in inducing differentiation of hESCs.

In pluripotent hESCs, p53 is negatively regulated by HDM2 and TRIM24. Differentiation induces acetylation at Lys373 of p53 via CBP/p300, p53K373ac then activates transcription by binding to p53REs on CDKN1A (p21), miR-34a, and miR-145. Induction of p21 leads to p53-dependent elongation of G₁ phase, whereas induction of miR-34a supports G1 elongation, blocks deactivation of p53 by targeting the deacetylase SIRT1, and counteracts pluripotency by targeting LIN28A. On the other hand, miR-145 targets OCT4, KLF4, and SOX2 and antagonize pluripotency. Thus, p53 exerts a cumulative pro-differentiation effect by elongating hESC G₁ phase via p21 and synergistically up-regulating miR-34a and miR-145 to counteract pluripotency. Ub, ubiquitin.