5-Aminoimidazole-4-carboxyamide Ribonucleoside Induces G1/S Arrest and Nanog Downregulation via p53 and Enhances Erythroid Differentiation§


  • Hee-Don Chae,

    1. Department of Microbiology and Immunology, Indiana University School of Medicine, Indianapolis, Indiana
    Search for more papers by this author
  • Man-Ryul Lee,

    1. Department of Microbiology and Immunology, Indiana University School of Medicine, Indianapolis, Indiana
    Search for more papers by this author
  • Hal E. Broxmeyer

    Corresponding author
    1. Department of Microbiology and Immunology, Indiana University School of Medicine, Indianapolis, Indiana
    • Department of Microbiology and Immunology, Indiana University School of Medicine, 950 West Walnut Street, R2-302 Indianapolis, Indiana 46202, USA
    Search for more papers by this author
    • Telephone: 317-274-7510; Fax: 317-274-7592

  • Author contributions: H.-D.C.: conception and design, collection and/or assembly of data, data analysis and interpretation, and manuscript writing; M.R.L.: collection of data.; H.E.B.: financial support, data analysis and interpretation, and manuscript writing.

  • Disclosure of potential conflicts of interest is found at the end of this article.

  • §

    First published online in STEM CELLSEXPRESS November 10, 2011.


Molecular mechanisms of how energy metabolism affects embryonic stem cell (ESC) pluripotency remain unclear. AMP-activated protein kinase (AMPK), a key regulator for controlling energy metabolism, is activated in response to ATP-exhausting stress. We investigated whether cellular energy homeostasis is associated with maintenance of self-renewal and pluripotency in mouse ESCs (mESCs) by using 5-aminoimidazole-4-carboxyamide ribonucleoside (AICAR) as an activator of AMPK. We demonstrate that AICAR treatment activates the p53/p21 pathway and markedly inhibits proliferation of R1 mESCs by inducing G1/S-phase cell cycle arrest, without influencing apoptosis. Treatment with AICAR also significantly reduces pluripotent stem cell markers, Nanog and stage-specific embryonic antigen-1, in the presence of leukemia inhibitory factor, without affecting expression of Oct4. H9 human ESCs also responded to AICAR with induction of p53 activation and repression of Nanog expression. AICAR reduced Nanog mRNA levels in mESCs transiently, an effect not due to expression of miR-134 which can suppress Nanog expression. AICAR induced Nanog degradation, an effect inhibited by MG132, a proteasome inhibitor. Although AICAR reduced embryoid body formation from mESCs, it increased expression levels of erythroid cell lineage markers (Ter119, GATA1, Klf1, Hbb-b, and Hbb-bh1). Although erythroid differentiation was enhanced by AICAR, endothelial lineage populations were remarkably reduced in AICAR-treated cells. Our results suggest that energy metabolism regulated by AMPK activity may control the balance of self-renewal and differentiation of ESCs. STEM CELLS 2012; 30:140–149.


Nutrient metabolism could play a critical role in cell fate decision as an adaptation mechanism to energy requirements [1–6]. Knockdown of metabolic enzymes induces myogenic or erythroid differentiation [1, 4, 7]. Overexpression of glycolytic enzymes is also involved in tumorigenesis or immortalization of fibroblasts [1, 2].

AMP-activated protein kinase (AMPK) is a master metabolic regulator that maintains cellular energy homeostasis to protect cells from an energy-shortage environment. AMPK, a heterotrimer composed of a catalytic subunit (α) and two regulatory subunits (β and γ), is activated by elevated intracellular AMP or AMP/ATP ratio caused by metabolic stresses such as glucose deprivation and hypoxia. AMPK is also activated by hormones or cytokines including leptin and interleukin (IL)-6 [3, 8]. Once activated, AMPK phosphorylates downstream target molecules to shut down ATP-consuming anabolic pathways such as synthesis of proteins, glycogens, and fatty acids and simultaneously to switch on ATP-generating catabolic pathways. The AMPK pathway couples metabolic stresses to cell growth, apoptosis, and differentiation by regulating p53, Rb, mammalian target of rapamycin (mTOR), and FOXO [2, 3, 8, 9].

Embryonic stem cells (ESCs) have the ability to undergo either self-renewal or differentiation into the three germ cell layers. Hence, ESCs are considered as a potential cell source for regenerative cell therapy [10, 11]. For ESC-based regenerative therapy, it is necessary to establish the techniques for efficient propagation of undifferentiated cells and differentiation to specific cell types. High glucose levels in culture medium are favored for proliferation of ESCs, and alteration of glucose concentration affects differentiation of ESCs [12–14]. AMPK activates various downstream signaling pathways to control cellular energy metabolism, including cell cycle checkpoints and apoptosis in response to energy stresses via the p53 tumor suppressor [2, 3, 8, 9], which can repress Nanog gene expression [15].

Nanog, Oct4, and Sox2 are intrinsic core factors for maintaining ESCs and preventing ESCs from spontaneous differentiation. Nanog is considered as a master transcriptional factor for self-renewal and pluripotency of ESCs and confers ESC pluripotency independent of leukemia inhibitory factor (LIF)-signal transducer and activator of transcription (STAT)3 signaling pathway [15–17]. Nanog expression is downmodulated at a transcriptional level in the cells under differentiation conditions. Binding of FoxD3 and Oct4/Sox2 to the Nanog promoter facilitates Nanog expression, while binding of transcription factor 3 (TCF3) and p53 to the promoter negatively regulates Nanog expression. LIF-STAT3 and bone morphogenetic protein (BMP)-brachyury (T) pathways were also reported to positively regulate Nanog expression [15]. Nanog gene expression in ESCs shows heterogeneous expression. Cells expressing lower levels of Nanog are more preferentially differentiated under differentiation conditions [18, 19]. Recently, Nanog protein stability was found to be regulated by its phosphorylation [20].

The mechanisms by which cellular energy metabolism affects self-renewal and pluripotency in ESCs remain unclear. Thus, we investigated the effects of 5-aminoimidazole-4-carboxyamide ribonucleoside (AICAR), an activator of AMPK, on self-renewal and differentiation of mESCs. We found that AMPK activated by AICAR induced p53/p21 activation, G1/S cell cycle arrest, and suppressed Nanog expression. Moreover, AICAR suppressed Nanog expression in mouse ESCs (mESCs) as well as human ESCs (hESCs) and promoted mESCs to differentiate into the erythroid lineage. These results suggest that metabolic energy control systems are closely coupled with cellular growth and differentiation fates of ESCs.


mESCs Culture and Differentiation

R1 mESCs [21] were maintained on mitomycin C-treated mouse embryonic fibroblasts (MEFs, Stem Cell Technology, Vancouver, Canada, http://www.stemcell.com) in knockout Dulbecco's modified Eagle's medium (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) supplemented with 15% fetal calf serum (FCS; Thermo Scientific, Waltham, MA, http://www.thermoscientific.com), 1% glutamine, 1% nonessential amino acids, antibiotics (Stem Cell Technology), 100 μM 2-mercaptoethanol (Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com), and LIF (1,000 U/ml; Millipore, Billerica, MA, http://www.millipore.com). For experiments, mESCs were cultured on gelatin-coated plates without MEF. mESCs were differentiated to embryoid bodies (EBs) in serum as reported [22]. Briefly, mESCs were trypsinized and replated on noncoated tissue culture plates for 30 minutes for MEF depletion. A total of 2,000 cells per milliliter were cultured in differentiation media (Iscove's Modified Dulbecco's Media (IMDM), 15% FCS, 1% glutamine, 450 μM monothioglycerol, 50 μg/ml ascorbic acid [Sigma-Aldrich], 0.2 mg/ml holo-transferrin [Roche, Indianapolis, IN, http://www.roche.com], and 5% protein-free hybridoma medium (PFHM-II) [Invitrogen]). AICAR was purchased from Sigma-Aldrich.

For proliferation assay, 5 × 104 mESCs were seeded in six-well plates. After 12 hours, cells were treated with AICAR (0.5 mM) for 24 hours. Viable cell number was determined by trypan blue exclusion using at least 300 cells in each group.

hESCs Culture and Immunocytochemistry

H9 hESCs were studied according to the research protocol of the WiCell Research Institute (WiCell, Madison, WI, http://www. wicell.org) and maintained as described previously [23]. hESCs were allowed to adhere to gelatin-coated cover glasses, cultured with or without AICAR (0.5 mM) for 1 day, and then fixed in 2% paraformaldehyde in phosphate buffered saline (PBS) for 10 minutes at room temperature. Cells were then refixed with cold 70% ethanol for 2 hours at −20°C. Cells were stained with anti-Ki-67-fluorescein isothiocyanate (FITC) Ab (clone B56; BD Biosciences, San Jose, CA, http://www.bdbiosciences.com) and anti-phospho-Histone H3 Ab (9701; Ser10; Cell Signaling Technology, Beverly, MA, http://www.cellsignal.com) followed by anti-rabbit Alexa555 (Molecular Probes, Eugene, OR, http://probes.invitrogen.com). Slides were mounted with ProLong Gold antifade reagent containing 4′,6-diamidino-2-phenylindole (DAPI) (Molecular Probes). Fluorescence images were captured with a Olympus FV1000-MPE confocal/multiphoton microscope (Olympus, Center Valley, PA, http://www.olympusamerica.com) at ×200 magnification.

RNA Extraction and Quantitative Reverse Transcription Polymerase Chain Reaction

Total RNA was extracted with the QIAGEN RNAeasy Mini Kit (Valencia, CA, http://www.qiagen.com) according to the manufacturer's instructions. Total RNA was reverse transcribed into cDNA using TAKARA BluePrint RT Reagent Kit (Takara Bio, Mountain View, CA, http://www.takara-bio.us/am) according to the manufacturer's instructions. Quantitative polymerase chain reactions (qPCR) were performed on an Agilent (Santa Clara, CA, http://www.genomics.agilent.com) MX3005P qPCR system with SYBR Green PCR Master Mix (SA Bioscience, Frederick, MD, http://www.sabiosciences.com). Levels of mRNA expression were normalized to β-tubulin (F: 5′-CTGGGAGGTGATAAG-CGATGA-3′, R: 5′-CGCTGTCACCGTGGTAGGT-3′) [24] or hypoxanthine phosphoribosyltransferase 1 (HPRT1) mRNA levels. qPCR primers except β-tubulin were purchased from SA Biosciences. miRNA expression was quantified using the RT2 miRNA qPCR Assays (SA Biosciences), normalizing with Rnu6 (RNA, U6 small nuclear 1) small RNA level. Quantitative reverse transcription PCR (qRT-PCR) was done in triplicate. Expression levels of target mRNAs or miRNAs were calculated by the ΔΔCT method [25].

Flow Cytometry Analysis

Cells were incubated with Fc Block antibody and then stained with anti-stage-specific embryonic antigen-1 (SSEA-1) (FAB2155P, R&D Systems, Minneapolis, MN, http:// www.RnDSystems.com), anti-CD45 (30-F11), anti-CD11b (M1/70) (BD Biosciences, San Jose, CA), anti-CD31 (390), anti-CD144 (eBioBV13), anti-Tie-2 (TEK4), or anti-Ter119 (TER119) (eBiosciences, San Diego, CA, http://www. ebiosciences.com). Apoptosis was measured by Annexin V/7-aminoactinomycin D (7AAD) staining (BD Biosciences) according to manufacturer's instructions. Cell cycle analysis was processed by bromodeoxyuridine (BrdU) incorporation and 7AAD staining using FITC BrdU Flow Kit (BD Biosciences). Briefly, cells were subjected to a pulse with 10 μM BrdU for 15 minutes at 37°C, trypsinized, washed with PBS, and fixed. Following permeabilization, refixation, and DNase treatment, cells were stained with FITC-labeled anti-BrdU antibody. After washing, DNA was stained with 7AAD. Flow cytometry data of stained cells were acquired by a LSRII using FacsDiva (BD Biosciences). Flow cytometry data were analyzed using CellQuest software (BD Biosciences) or FCS Express 3 (De Novo Software, Los Angeles, CA, http://www.denovosoftware.com).


Cells were harvested and lysed in ice-cold Mg2+ lysis/wash buffer (Millipore) containing protease and phosphatase inhibitors (1 mM Na3VO4, 10 mM NaF, complete protease inhibitors [Roche]). Total protein (20 μg) was separated on a 12% sodium dodecyl sulfate (SDS) polyacrylamide gel and transferred to polyvinylidene fluoride (PVDF) membranes. Proteins were detected with specific antibodies. The following antibodies were used in Western blot analyses at 1:1,000 dilution: anti-phospho-p53 (Ser15; mouse Ser18; 16G8), anti-phospho-p53 (Ser392; mouse Ser389; 9281), anti-acetyl-CoA carboxylase (ACC; C83B10), anti-phospho-ACC (3661), anti-phospho-STAT3 (Y705; 9131), anti-GATA-1 (D52H6), anti-Nanog (D73G4; Cell Signaling Technology); anti-β-actin (AC-15, Sigma-Aldrich), anti-p21 (F-5), anti-p53 (FL393), anti-hemoglobin β (M-19; Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com), anti-β-tubulin (TBN06, Thermo scientific), and anti-Nanog (AB5731, Millipore). Primary antibodies were detected with horseradish peroxidase-conjugated goat anti-mouse or goat anti-rabbit (1:3,000 dilution; Cell Signaling Technology) with enhanced chemiluminescence detection.

p53/p21 Knockdown

Lentiviral vectors expressing p53 (TRCN0000012359) and p21 short hairpin RNA (shRNA) (TRCN0000042583) were purchased from Sigma-Aldrich. Vesicular stomatitis virus G glycoprotein (VSV-G) pseudotyped lentiviral particles were produced by transient transfection of HEK293T cells by calcium phosphate transfection method. mESCs were infected twice with lentivirus in HEK293T cell supernatant on Retronectin (Takara Bio). Infected mESCs were selected by culturing with puromycin (Sigma-Aldrich) at 2 μg/ml for at least 2 days.

Statistical Analysis

Data were presented as mean ± SD. Statistical significances were determined by unpaired Student's t test comparisons for at least three experiments. Values of p < .05 were considered significant.


AICAR Treatment Induces the p53/p21 Pathway

AMPK is a key regulator of cellular metabolism in response to extracellular nutrients and cellular energy status. AICAR, an AMPK activator, regulates proliferation and cell death in several cell types [3, 8, 26]. As mESCs are sensitive to extracellular glucose levels in culture media [12–14], and decreases in glucose in culture media activate AMPK in mammalian cells types [3, 8, 26], we examined the role of AMPK in self-renewal and differentiation of mESCs. As AMPK phosphorylates ACC at Ser79 to inhibit biosynthesis of fatty acids [3], we first examined whether AICAR activated AMPK in mESCs by examining ACC phosphorylation. ACC phosphorylation at Ser 79 was observed after 6 hours of AICAR treatment in the presence of LIF and glucose (Fig. 1A).

Figure 1.

AICAR treatment leads to cell cycle arrest by p53/p21 induction, without effects on apoptosis. (A): AICAR activates p53/p21 pathway in R1 mouse embryonic stem cells (mESCs). Cells were treated with 0.5 mM AICAR for 1, 6, or 9 hours, and then cell lysates were prepared. Total lysates were analyzed by immunoblotting for phospho-ACC, ACC, phospho-STAT3, and p53 and p21 expression. Anti-β-tubulin antibodies were used as loading controls. A representative blot of at least three different experiments is shown. (B): AICAR inhibits proliferation of mESCs via p53/p21 pathway. Proliferation of mESCs expressing scrambled shRNA, p53 shRNA, or p21 shRNA 1 day after AICAR treatment (0.5 mM). A total of 5 × 104 cells were seeded in six-well plates. After 12 hours, cells were cultured with or without AICAR (0.5 mM) for 1 day, and viable cell number was measured by Trypan blue exclusion and results were displayed as % of number of control cells which were expressing corresponding shRNA but were not treated with AICAR. (C): AICAR does not affect mESC viability. Cells were treated with or without AICAR (0.5 mM, 24 hours), and cell viability was assessed by AnnexinV/7-AAD staining. (D): Knockdown of p53 and p21 expression. mESCs were infected by scrambled shRNA-, p53 shRNA-, or p21 shRNA-expressing lentiviruses. Expression of p53 and p21 was analyzed by immunoblotting using the cells treated with or without AICAR for 9 hours. (E): AICAR induces cell cycle arrest of mESCs. Cells were treated with AICAR (0.5 mM) for 1 day. Cells were subsequently incubated with BrdU for 15 minutes, fixed, and stained with anti-BrdU-FITC Ab and 7-AAD. Bivariate analysis of total DNA content and BrdU incorporation was performed by flow cytometry. Plots are representative of three or more experiments with similar results. Data represent % cell populations residing at each cell cycle stage determined by DNA content and BrdU incorporation and is expressed as mean ± SD (n = 3). *, p < .05, **, p < .01. Abbreviations: AAD, aminoactinomycin D; ACC, acetyl-CoA carboxylase; AICAR, 5-aminoimidazole-4-carboxyamide ribonucleoside; BrdU, bromodeoxyuridine; STAT, signal transducer and activator of transcription; shRNA, short hairpin RNA.

Activated AMPK enhances generation of ATP and inhibits consumption of ATP simultaneously. Lactate dehydrogenase (LDH) expression was examined after AICAR treatment. LDH has two isoforms, LDH-A and LDH-B. LDH-A is a heterotetramer of M-LDH, which favors conversion of pyruvate to lactate, while LDH-B is a homotetramer of H-LDH, which favors conversion of lactate to pyruvate [27]. As expected, LDH-B expression was enhanced in AICAR-treated mESCs (0.5 mM, 24 hours; 2.6 ± 0.4-fold, n = 3, p < .01). AMPK plays a critical role not only in cellular energy regulation but also in cell cycle checkpoints through modulating p53 activation [3, 8]. Levels of p53 and its downstream effector protein p21 were elevated in AICAR-treated cells (Fig. 1A). However, AICAR did not affect LIF-mediated STAT3 phosphorylation [28] (Fig. 1A). Sirtuin 1 (silent mating type information regulation 2, homolog) 1 (SIRT1) downregulates AMPK through deacetylating the serine-threonine kinase liver kinase B1 and antagonizes p53 activation by deacetylation [29, 30]. AMPK increases SIRT1 activity by enhancing cellular NAD+ levels [31]. However, SIRT1 deficiency did not influence AICAR-induced phosphorylation and acetylation of p53 (data not shown).

G1/S Arrest by AICAR Treatment

It is not clear whether cellular energy metabolism contributes to cell cycle control of mESCs. As the p53/p21 pathway is involved in cell cycle arrest and apoptosis [32, 33], we investigated whether AICAR inhibited proliferation and survival of mESCs. Cell proliferation was significantly inhibited in AICAR-treated cells in the presence of LIF (Fig. 1B) without influencing apoptosis (Fig. 1C). Cell cycle checkpoints tightly regulate cellular proliferation. Activation of p53/p21 reduces cell proliferation by inducing a halt to cell cycle progression [32]. To test whether p53/p21 pathway was required for AICAR-induced suppression of mESC proliferation, expression of p53 and p21was knocked-down with specific shRNAs (Fig. 1D). p53 knockdown significantly reduced expression of p21, its downstream target (Fig. 1D). As shown in Figure 1B, knockdown of p53 and p21 significantly blocked the suppressive effect of AICAR on proliferation. However, p53 knockdown alone was not sufficient to restore proliferation suppressed by AICAR as much as p21 knockdown (Fig. 1B, 1D), because p53-knocked-down cells still maintained expression of p21 at minimal levels. These results suggest that AICAR-induced suppression of proliferation in mESCs requires expression of both p53 and p21.

G1/S progression is regulated by cyclin D/cyclin-dependent kinase (CDK) 4, 6 or cyclin E/CDK2 complexes, and these CDK complexes are inhibited by p21 [32]. We hypothesized that AICAR-induced p53/p21 might inhibit these CDK complexes to induce cell cycle arrest in G1/S-phase. To assess the effect of AICAR on the cell cycle progression, cells were double-stained with BrdU and 7-AAD after AICAR treatment, and cycle distribution of cells was analyzed by flow cytometry. As mESCs grow very fast with a remarkably short doubling time (around 8 hours) [34], BrdU-positive cycling cells constituted 64% of mESCs (Fig. 1E). As expected, AICAR markedly decreased the percentage cycling S-phase of mESC population from 64%–37% as assessed by BrdU incorporation (Fig. 1E). This result is consistent with that obtained by cell counting (Fig. 1B). Instead, AICAR increased the cell population at G1 phase by 15%. AICAR also increased the cell population at noncycling S-phase (S BrdU) by fivefold (2.8% in control vs. 14.3% in AICAR-treated cells, Fig. 1E). This indicates that AICAR inhibits cell proliferation by inducing cell cycle arrest at G1 as well as S-phase.

Suppression of Nanog and SSEA-1 Expression by AICAR

Next, we examined whether AICAR affected the undifferentiated status of mESCs. First, we examined expression level of cell surface markers of undifferentiated mESCs, SSEA-1 by flow cytometry. SSEA-1 expression was substantially lower in AICAR-treated cells than nontreated control cells in the presence of LIF (Fig. 2A, 2B). We next addressed expression levels of Oct4-Nanog transcription factors, which play a critical role in maintaining self-renewal of ESCs [15, 35]. Downregulation of Nanog mRNA expression correlates with phosphorylation of p53 [15, 36]. AICAR induced phosphorylation of p53 at Ser 15 and 392 in mESCs (Fig. 2C). Consistent with this, when mESCs were treated with AICAR, Nanog protein expression level was downregulated (56 ± 7% of control, n = 3; Fig. 2C). However, the level of Oct4 protein was not affected by AICAR. We examined whether AICAR could suppress Nanog mRNA levels. Nanog mRNA levels were significantly reduced in cells after 9 hours of AICAR treatment but recovered at 24 hours (Fig. 2D). However, Nanog protein expression remained low even at 24 hours post-AICAR treatment (67 ± 5% of control, n = 3; Fig. 2D). Nanog expression has been reported to be regulated by miRNAs [37, 38]. However, AICAR did not induce the expression of miR-134, which suppresses Nanog expression (Fig. 2E). Another Nanog-suppressing miRNA, miR-296 was not detected even in AICAR-treated cells (data not shown).

Figure 2.

Expression of Nanog and SSEA-1 is downregulated by AICAR treatment. (A, B): Representative flow cytometric profile of SSEA-1 expression on mouse embryonic stem cells (mESCs). R1 mESCs were treated with AICAR (0.5 mM, 24 hours), and then cells were collected and analyzed for cell surface expression of SSEA-1 by flow cytometry. The graph shows relative expression levels of SSEA-1 in R1 mESCs treated with or without AICAR (0.5 mM, 1 day or 2 days). Values represent mean ± SD (n = 3). (C): AICAR treatment induces p53 phosphorylation and represses Nanog protein level. Cells were cultured with or without AICAR (0.5 mM) for 1 day. Total lysates were immunoblotted for phospho-p53, p53, Nanog, Oct4, and with β-actin as a loading control. A representative blot of at least three different experiments is shown. (D): AICAR treatment transiently represses nanog mRNA level (E) but does not affect the expression of miR-134. (D): R1 cells were cultured with AICAR (0.5 mM) for 9 or 24 hours and followed by quantitative polymerase chain reaction for examining nanog mRNA and immunoblotting for Nanog protein. Relative expression levels of nanog mRNA are shown as mean ± SD (n = 4). (E) R1 cells treated with AICAR for 9 hours were analyzed for the expression miR-134. Relative expression levels of mIR-134 are shown as mean ± SD (n = 3). **, p < .01. Abbreviations: AICAR, 5-aminoimidazole-4-carboxyamide ribonucleoside; HPRT, hypoxanthine phosphoribosyltransferase; MFI, median fluorescence intensity; SSEA, stage-specific embryonic antigen.

As p53 induces differentiation of mESCs by suppressing Nanog expression [15, 36], we investigated whether p53 also mediates the AICAR-induced suppression of Nanog and stage-specific embryonic antigen-1 (SSEA-1) expression. We found that suppression of SSEA-1 and Nanog induced by AICAR was abolished by p53 knockdown (Fig. 3A, 3B). These data suggest that AICAR-induced suppression of Nanog and SSEA-1 expression is mediated by p53. AICAR, up to 0.1 mM, had no significant effect on p53 expression (supporting information Fig. S1). Furthermore, up to 0.1 mM, AICAR did not suppress the expression of Nanog and SSEA-1, paralleling the activation of p53 (supporting information Fig. S1).

Figure 3.

p53 plays a critical role in AICAR-induced suppression of Nanog and SSEA-1 expression in embryonic stem cells (ESCs). (A): Mouse ESCs (mESCs) expressing scrambled shRNA or p53 shRNA were treated with AICAR (0.5 mM) for 24 hours and then analyzed for SSEA-1 expression by flow cytometry. Data are represented as mean values ± SD (n = 3). **, p < .01. (B): shRNA-expressing mESCs were treated with AICAR (0.5 mM) for 9 hours, and then expression levels of Nanog were compared by immunoblotting. The graph shows relative protein expression levels of Nanog in R1 mESCs treated with AICAR. Data are shown as mean values ± SD (n = 5). *, p < .05. (C): AICAR suppresses Nanog expression and proliferation of H9 hESCs. H9 hESCs were treated with or without AICAR (0.5 mM) for 1 day. Cell lysates were analyzed for p53, Nanog, Oct4, and with β-tubulin as a loading control, by immunoblotting. A representative blot of at least three independent experiments is shown. (D, E): H9 hESCs treated with AICAR for 1 day were analyzed for Ki-67 and p-H3 expression as markers for proliferating cells. % marker-positive cells are shown as mean ± SD (n = 3). **, p < .01. Abbreviations: AICAR, 5-aminoimidazole-4-carboxyamide ribonucleoside; DAPI, 4′,6-diamidino-2-phenylindole; MFI, median fluorescence intensity; SSEA, stage-specific embryonic antigen; shRNA, short hairpin RNA.

We also investigated whether hESCs would respond to AICAR similarly to mESCs by induction of p53 and inhibition of Nanog expression. Expression of p53 was induced and Nanog suppressed in H9 hESCs on AICAR treatment in a fashion similar to that seen in mESCs (Fig. 3C). To examine the effect of AICAR on proliferation of hESCs, we performed immunofluorescence staining assays with Ki-67 (a proliferation cell marker) and phospho-Histone H3 (p-H3, a mitotic cell marker). AICAR treatment significantly decreased p-H3 and Ki-67 labeling, which reflects an inhibitory effect of AICAR on proliferation of hESCs (Fig. 3D, 3E).

Recently, Moretto-Zita et al. [20] reported that Nanog proteins can be stabilized by phosphorylation and interaction with Pin1. We extended our study to investigate whether AICAR increased the degradation rate of Nanog protein by measuring the half-life of Nanog protein following treatment of mouse ESCs with cycloheximide, a protein synthesis inhibitor. Our data indicated that the half-life of Nanog protein was shortened by AICAR (from 2.5h to 1.5h; Fig. 4A). Furthermore, MG132, a proteasome inhibitor, recovered the Nanog protein level from the AICAR-induced degradation (Fig. 4B). These results indicate that AICAR suppresses Nanog expression via activating proteasome-dependent degradation of Nanog protein as well as p53-mediated inhibition of Nanog mRNA expression.

Figure 4.

AICAR treatment downregulates stability of Nanog protein. (A): R1 mouse embryonic stem cells (mESCs) were preincubated with or without AICAR (0.5 mM) for 6 hours, and then CHX (50 mg/ml) for indicated hours. Lysates were analyzed by immunoblotting for Nanog and β-tubulin (loading control). (B): Proteasome inhibitor (MG132) protects Nanog from AICAR treatment. After 6 hours of culture with or without AICAR (0.5 mM), mESCs were treated with MG132 (30 μM) or vehicle for 3 hours. Cells were harvested and lysed, and then Nanog and β-tubulin levels were analyzed. A representative blot of at three different experiments is shown. A diagram shows the experimental scheme. Abbreviations: AICAR, 5-aminoimidazole-4-carboxyamide ribonucleoside; CHX, cycloheximide.

Effects of AICAR on Erythroid Developmental Potential

We next investigated pluripotency of mESCs on treatment of cells with AICAR. To evaluate the effect of AICAR on the differentiation potential of mESCs, we compared EBs derived from AICAR-treated cells with those derived from control cells. mESCs can differentiate into a variety of specialized cell lineages in EBs [22, 39]. To induce differentiation of mESCs into EBs, mESCs were cultured in suspension without LIF. As shown in Figure 5A, AICAR treatment significantly inhibited EB formation, although both EBs derived from control and AICAR-treated mESCs were similar in size. However, EBs derived from AICAR-treated cells contained more red hemoglobinized cells compared with EBs from nontreated cells (supporting information Fig. S2). To determine whether AICAR enhanced erythroid cell formation in EBs, Ter119+ populations were measured by flow cytometry. The proportions of Ter119+ cells were increased for days 7 and 8 EBs derived from AICAR-treated cells (Fig. 5B, 5C). Given the enhanced erythroid differentiation potential of AICAR-treated mESCs, qRT-PCR was performed to analyze the expression of erythroid-specific genes in EBs. During EB development, primitive and definitive erythroid cells are generated. These two different stages of erythroid cells can be distinguishable based on the expression of different types of globin genes. Primitive erythroid cells express embryonic-specific globins (hbb-bh1), whereas definitive erythroid cells express adult-specific globins (hbb-b1) [40]. Expression levels of both embryonic (hbb-bh1) and adult globin (hbb-b1) genes were markedly enhanced in EBs derived from AICAR-treated cells (Fig. 6A, 6B). GATA1 and erythroid Kruppel-like factor (EKLF) are transcriptional factors that play a critical role in erythroid differentiation [40]. mRNA expression levels of gata1 and klf1 were also substantially higher in EBs derived from AICAR-treated cells (Fig. 6C, 6D). To confirm expression of Hemoglobinβ (Hbb) and GATA-1, we undertook immunoblotting assays. Significantly higher levels of Hbb and GATA-1 protein were detected in EBs derived from AICAR-treated mESCs compared to those from control cells (Fig. 6E). Taken together, these results suggest that AICAR enhances specifically erythroid lineage developmental potential in EBs although it reduces numbers of EBs formed.

Figure 5.

AICAR treatment affects EB formation and erythroid differentiation. (A): AICAR treatment inhibits EB formation. mESCs were cultured with or without AICAR (0.5 mM) for 1 day, and 20,000 cells were induced to form EB. Viable cells were counted at day 6 EB by Trypan blue exclusion. (B): Enhancement of EB differentiation into erythroid lineage. Cells were differentiated as described in (A). Cell surface expression of Ter119 was analyzed at EB days 6, 7, and 8. Percentage of Ter119+ cells was measured by flow cytometry. Representative flow cytometric plot of day 7 EB is shown in (C). Data are graphed as mean ± SD (n = 3). *, p < .05, **, p < .01. Abbreviations: AICAR, 5-aminoimidazole-4-carboxyamide ribonucleoside; EB, embryoid body.

Figure 6.

Expression of erythroid transcription factors is enhanced in embryoid body (EB) cells derived from AICAR-treated mouse embryonic stem cells (mESCs). After culture with or without AICAR (0.5 mM) for 1 day, mESCs were induced to form EBs. EBs were harvested at days 4, 6, and 8. Equal amounts of RNA from mESCs or from EB-derived cells. Relative mRNA expression of embryonic globin (hbb-bh1, A), adult globin (hbb-b1, B), gata1(C), and klf1(D) genes from days 4, 6, and 8 EBs by quantitative reverse transcription polymerase chain reaction. Expression of genes was normalized against β-tubulin expression level. Relative expression levels are presented as fold induction above expression levels in nontreated mESCs. Values are indicated as mean ± SD (n = 3). *, p < .05, **, p < .01. Expression levels of GATA-1, Hbb, and β-actin were analyzed by immunoblotting (E). A representative blot of at least three independent experiments is shown. Abbreviation: AICAR, 5-aminoimidazole-4-carboxyamide ribonucleoside.

AICAR Represses Endothelial Developmental Potential

To further analyze the hematopoietic and endothelial differentiating potential of mESC-derived EBs, we assessed cell surface markers by flow cytometry. CD45 is a marker for definitive multilineage hematopoietic cells and CD11b is for myeloid hematopoietic cells. The percentages of CD45+ or CD11b+ cells in day 8 EBs were not affected by AICAR treatment (Fig. 7A), while percentages of Ter119+ primitive and definitive erythroid cells were enhanced by AICAR (Figs. 5B, 6). Endothelial populations were analyzed using CD144, CD31, and Tie-2 as markers for endothelial cells. As shown in Figure 7B, 7C, endothelial cell populations were remarkably reduced in EBs derived from AICAR-treated cells. These data indicate that erythroid developmental potential was enhanced, whereas endothelial potential was decreased in EBs derived from AICAR-treated mESCs.

Figure 7.

Flow cytometry analysis of embryoid body (EB)-derived cells demonstrating decrease in endothelial lineage-specific markers from AICAR-treated mouse embryonic stem cells (mESCs). mESCs were induced to EB formation as described in Figure 5. Day 8 EBs were harvested and analyzed for the surface expression of CD11b/CD45 (A), CD45/CD144 (B), and Tie-2/CD31 (C). Flow cytometric profiles represent one out of three independent experiments. Values are shown as mean ± SD (n = 3). **, p < .01. Abbreviation: AICAR, 5-aminoimidazole-4-carboxyamide ribonucleoside.


There are increasing reports showing that metabolism is closely coupled with cell cycle progression and differentiation [1–7]. Mechanisms by which cellular energy metabolism controls self-renewal and pluripotency of ESCs have not been fully explained. Understanding effects of energy metabolism on ESC identity could provide us better tools for efficient control of propagation and directed differentiation of ESCs for regenerative therapy. We examined effects of AICAR, an AMPK activator, on proliferation, stemness, and subsequent differentiation potential of mESCs. Notably, AICAR treatment represses Nanog expression at both transcriptional and post-translational levels. These alternations in ESCs enhance erythroid differentiation, whereas general EB differentiation and endothelial lineage cell formation are suppressed. Thus, we report here that AMPK, a master regulator of energy metabolism [3], plays a critical role in the self-renewal and differentiation of ESCs.

Our data showed that AICAR activated p53/p21 signaling, concurrently induction of cell cycle arrest of mESCs at G1 and S-phases, with no noticeable effects on apoptosis. AMPK also has the ability to control proliferation and apoptosis in response to metabolic stresses directly or indirectly by regulating other key regulators such as p53 and mTOR [2, 3, 8, 9]. On glucose deprivation, AMPK phosphorylates p53 at Ser15 and activates the p53-dependent cell cycle checkpoint to induce cell cycle arrest that allows cells to survive under metabolic stress. mESCs lack a G1 checkpoint in response to DNA-damaging stress, which allows cells with damaged DNA to progress into S-phase to exacerbate DNA damage, resulting in inducing apoptosis. This G1 checkpoint missing is important for rapid proliferation and removing mutated genome to preserve genomic integrity [41]. However, mESCs have the ability to be arrested at the G0/G1 phase on serum starvation and these arrested cells have even higher capacity to differentiate into functional neuronal cells [42]. Serum starvation has been reported to activate the AMPK–p53–p21 pathway [43]. Our data suggests that the AMPK–p53–p21 pathway can be activated to induce cell cycle arrest at G1/S, likely to protect ESCs from metabolic stress.

Nanog and Oct4 are crucial factors for self-renewal and pluripotency of ESCs and they are cross-regulated [15]. Interestingly, AICAR inhibited expression of Nanog but not Oct4 proteins. It is reported that Nanog silencing by siRNA does not reduce Oct4 expression in mESCs [44]. Given that Oct4 is essential for antiapoptosis of ESCs in response to stress [45], sustained Oct4 expression might protect mESCs on AICAR treatment. Nanog mRNA was also downregulated transiently after AICAR treatment. Nanog mRNA expression is positively controlled by LIF-STAT3, Oct4/Sox2, and FoxD3, whereas p53 negatively regulates Nanog transcription [15]. AICAR enhanced phosphorylation and protein levels of p53, suggesting that AICAR-activated p53 might directly repress Nanog mRNA expression. Even though Nanog mRNA level was recovered at 24 hours after AICAR treatment, Nanog protein levels were still lower up to 24 hours. There are two possible mechanisms to explain this discrepancy, miRNAs and protein stability. miRNAs are post-transcriptional regulators that repress mRNA translation or modulate mRNA decay in a sequence-dependent manner. Nanog mRNA translation can also be inhibited by miRNAs [37, 38]. Our data show that expression of Nanog inhibitory miRNAs, miRNA-134 and 296, was not enhanced by AICAR. AICAR activated proteasome-dependent degradation of Nanog protein. Nanog protein is stabilized by phosphorylation (Serine 52 and 65)-dependent association with Pin1 prolyl isomerase, suppressing ubiquitination of Nanog [20]. As Nanog protein has the AMPK recognition motif for phosphorylation at Serine 104 and 267 [46], AMPK might directly control degradation of Nanog protein by phosphorylation. p53 knockdown significantly abrogated AICAR-induced suppression of Nanog expression. p53 regulates proteasomal degradation of Cdc6 and Topo I by transcriptional control of protein degradational machinery or modulation of CDK kinase activity [47, 48]. Taken together with previous reports, our data indicate that AICAR represses Nanog expression by increasing proteasome-dependent degradation of Nanog protein and suppression of Nanog mRNA expression in a p53-dependent manner. p53–p21 pathway suppresses induced pluripotent stem cells (iPSCs) generation, and the suppression of p53 enhances the efficiency of iPSC generation [49]. Moreover, Δ40p53, a dominant-negative transactivation-deficient isoform of p53, is highly expressed in ESCs and plays a key role in maintaining the ESCs by upregulating Nanog and SSEA-1 expression [50]. Suppression of SSEA-1 expression in AICAR-treated mESCs was restored by p53 knockdown. Thus, our data suggest that the AMPK–p53–p21 pathway plays a role in the maintenance of pluripotency of ESCs in response to energy stress.

Energy metabolism is tightly coupled with cellular fate decisions. Knockdown of three metabolic enzymes, phosphoglycerate kinase, hexose-6-phosphate dehydrogenase, and ATP citrate lyase (Acl) induces differentiation of C2C12 myoblasts to skeletal muscle. Moreover, Acl knockdown also induces erythroid differentiation of the K562 leukemia cell line [4, 7]. Overexpression of glycolytic enzymes, phosphoglycerate mutase and glucose phosphate isomerase, facilitates immortalization of MEFs [1]. Glucose metabolism is also crucial in embryogenesis. Metabolism is changed from pyruvate oxidation to glycolysis-based metabolism in blastocysts during the late cleavage stage [51]. Cardiac differentiation from ESCs requires metabolic switch from anaerobic glycolysis to mitochondrial oxidative metabolism [52]. Although high glucose levels are usually used for maintenance of ESC culture, the differentiation potential of mESCs can be modulated by changing the glucose level in culture medium [12–14]. AMPK activity is also closely related with myoblast [53], osteoblast [54], and adipocyte differentiation [55].

Nanog is a core transcriptional factor for controlling self-renewal and pluripotency of ESCs. Overexpression of Nanog can maintain the undifferentiated status of mESCs without LIF [16, 17]. Therefore, ESC differentiation requires downregulation of Nanog expression. Nanog-deleted or -downregulated ESCs are susceptible to differentiation [56]. Notably, Nanog functions as a regulator of fate decision of ESCs in a dose-dependent manner. While Nanog heterozygote (±) ESCs expressing half the amount of Nanog proteins can differentiate to endodermal, mesodermal, and ectodermal cells in the presence of LIF, complete knockdown of Nanog expression leads to exclusive differentiation to extraembryonic endoderm [17, 56, 57]. The ESC population is heterogeneous in terms of Nanog expression, with a distribution of Nanog-high and Nanog-low populations. Nanog-high ESCs show high proliferation rate and are more resistant to spontaneous differentiation. Nanog-low ESCs are unstable and can be easily differentiated [18, 19]. Inhibition of Nanog-stabilizing protein Pin1 reduces Nanog protein level, which also results in suppressing self-renewal and teratoma formation of ESCs [20]. Our data demonstrate that erythroid developmental potential was enhanced, whereas endothelial potential was impaired in EBs derived from AICAR-treated mESCs, implicating energy metabolism controlled by AMPK signaling in developmental fate decision during mESC differentiation.


We thank Dr. Andras Nagy (Samuel Lunenfeld Research Institute) for providing R1 mESCs. We thank Young-June Kim for helpful comments and discussion on the manuscript. These studies were supported by Public Health Service Grants R01-HL56416 and R01-HL67384 from the National Institutes of Health to H.E.B.


The authors indicate no potential conflicts of interest.