Sirt1, p53, and p38^MAPK Are Crucial Regulators of Detrimental Phenotypes of Embryonic Stem Cells with Max Expression Ablation^†‡§

Abstract

c‐Myc participates in diverse cellular processes including cell cycle control, tumorigenic transformation, and reprogramming of somatic cells to induced pluripotent cells. c‐Myc is also an important regulator of self‐renewal and pluripotency of embryonic stem cells (ESCs). We recently demonstrated that loss of the Max gene, encoding the best characterized partner for all Myc family proteins, causes loss of the pluripotent state and extensive cell death in ESCs strictly in this order. However, the mechanisms and molecules that are responsible for these phenotypes remain largely obscure. Here, we show that Sirt1, p53, and p38^MAPK are crucially involved in the detrimental phenotype of Max‐null ESCs. Moreover, our analyses revealed that these proteins are involved at varying levels to one another in the hierarchy of the pathway leading to cell death in Max‐null ESCs. STEM CELLS2012;30:1634–1644

INTRODUCTION

Embryonic stem cells (ESCs) are derived from the inner cell mass of blastocysts and can self‐renew indefinitely while maintaining pluripotency, the ability to differentiate into all cell types of the body [1-3]. These remarkable characteristics are sustained by mechanisms encompassing epigenetic regulation, signaling pathways, and interplay of transcription factors. The fact that pluripotency can be acquired by forced expression of defined transcription factors in somatic cells implies a dominant role of specific transcription factors over others for ESC status [4-6]. Oct3/4, Sox2, and Nanog are collectively termed as core factors, which are crucially involved in preserving the pluripotent state of ESCs [7, 8]. Interestingly, genome‐wide chromatin immunoprecipitation analyses revealed that these factors share a large number of gene promoters at binding sites, thereby establishing specialized regulatory circuits in ESCs [7-12]. c‐Myc is also involved in sustaining the pluripotency and self‐renewal of ESCs [13-16]. However, c‐Myc target genes are not significantly overlapped with genes targeted by core factors. Indeed, c‐Myc predominantly binds to genes involved in cellular metabolism and cell cycle control, whereas targets of core factors are significantly enriched for genes related to developmental and transcription‐associated processes [10, 17, 18], suggesting that functions of c‐Myc in ESCs are reasonably independent from those of core factors.

The Myc family consists of three highly related transcription factors, c‐Myc, N‐Myc, and L‐Myc. These members are overexpressed in numerous types of human tumors, and forced expression of each factor has been shown to cause neoplastic transformation, although L‐Myc activity is rather low compared with that of other members [19]. Myc proteins dimerize with the apparently obligated basic‐helix‐loop‐helix (bHLH) Zip binding partner Max. Indeed, without Max association, Myc proteins fail to exert almost all biological activities including promotion of neoplastic transformation [20-22], although some Myc activities, such as those for Pol III gene transcription, are Max‐independent [23, 24]. We recently demonstrated that ablation of Max gene expression in ESCs leads to loss of pluripotency and extensive cell death. We also demonstrated that these phenotypes are due to the loss‐of‐function of Myc proteins using modified c‐Myc and Max proteins that efficiently dimerize with each other but do not bind to their endogenous partners [14]. Moreover, our analyses demonstrated that the two major phenotypes associated with Max expression ablation in ESCs, that is, loss of pluripotency and apoptotic cell death, do not occur randomly but follow a compulsory ordered rule with loss of pluripotency occurring first. However, the mechanisms underlying these phenotypes remain elusive.

Here, we demonstrate with chemical treatments and knockdown experiments that p38^MAPK, Sirt1, and p53 are crucially involved in the detrimental phenotypes of Max‐null ESCs. Our analyses also reveal that rescued cells show differences in their levels of pluripotency, dependent on which protein or gene was used as the target of inhibition, suggesting that these three factors contribute to the phenotypes of Max‐null ESCs at distinct points in the hierarchy.

MATERIALS AND METHODS

ESC Culture and Transfection

Max‐null ESCs [14] generated using EBRTcH3 ESCs [25] were maintained without feeder cells in standard ESC medium supplemented with fetal bovine serum (FBS) and leukemia inhibitory factor (LIF) unless indicated otherwise, as described elsewhere [26]. Expression vectors were transfected into ESCs with a Nucleofector using the A‐23 program (Lonza, Basel, Switzerland).

Reagents and Antibodies

The following regents were used at the indicated concentrations unless stated otherwise: 1 μg/ml doxycycline (Dox) (Clontech, Mountain View, CA, http://www.clontech.com); 3 μg/ml puromycin and 4 mM nicotinamide “Nam” (Sigma, St. Louis, MO); 20 μM “SB203580,” 5 μM pifithrin‐α “PFTα”, 20 μM 3,4‐Dihydro‐5[4‐(1‐piperindinyl)butoxy]‐1(2H)‐isoquinoline (DPQ), 20 μM Sirtinol, and anti‐N‐Myc (OP13) (Calbiochem, LA Jolla, CA, http://www.emdbiosciences.com); 1 μM PD0325901 (Axon Medchem, Groningen, Netherlands); 10 μM SB239063 (Tocris Bioscience, Bristol, U.K.); anti‐Max (sc‐197), anti‐β‐actin (sc‐47778), anti‐p38^MAPK (sc‐7972), anti‐Oct3/4 (sc‐5279), anti‐Tbx3 (sc‐31656), anti‐L‐Myc (sc‐28699), and anti‐Mad3 (sc‐770) (Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com); anti‐p‐p38^MAPK (#4631), anti‐p53 (#2524), anti‐c‐Myc (#9402), anti‐extracellular signal‐regulated kinase (ERK)^MAPK (#9102) and anti‐p‐ERK^MAPK (#9101) (Cell Signaling Technology, Danvers, MA, http://www.cellsignal.com); anti‐Sirt1 (07‐131), anti‐Sox2 (AB5603), and anti‐Caspase 3, active form (AB3623) (Millipore, Billerica, MA, http://www.millipore.com); anti‐KLF4 (AF3158) (R&D, Minneapolis, MN, http://www.rndsystems.com); anti‐Nanog (RCAB0001P) (Cosmo Bio, Tokyo, Japan); anti‐p53 (NCL‐p53‐CM5p) (Novocastra, Newcastle upon Tyne, U.K., http://www.novocastra.co.uk); anti‐Oct3/4 conjugated with Per‐CP‐Cy™5.5 (51‐9006267) and control IgG₁ κ isotype conjugated with Per‐CP‐Cy™5.5 (51‐9006272) (BD Pharmingen, San Diego, CA, http://www.bdbiosciences.com/index_us.shtml).

Western Blotting and Immunostaining

For Western blotting, proteins were resolved using SDS‐PAGE, transferred to a polyvinylidene fluoride (PVDF) membrane, and probed using the indicated primary antibodies and appropriate secondary antibodies conjugated with horseradish peroxidase. Specific protein bands were detected by an enhanced chemiluminescence system (GE Healthcare, Pittsburgh, PA, https://www3.gehealthcare.com). For immunostaining, cells were cultured on gelatin‐coated cell disks (Sumitomo Bakelite Co., Ltd., Tokyo, Japan) and then fixed with 4% paraformaldehyde (PFA) for 10 minutes at 37°C. After extensive washing with phosphate buffered saline (PBS), cells were permeabilized with 90% methanol for 5 minutes. Cells were then PBS washed and blocked with 3% FBS for 1 hour at room temperature followed by incubation with primary antibodies against Oct3/4 and activated caspase‐3 overnight at 4°C. After a PBS wash, cells were incubated with the appropriate Alexa Fluor dye‐conjugated secondary antibodies (Invitrogen, Carlsbad, CA, http://www.invitrogen.com). Cells were then washed and observed under a fluorescence microscope.

Terminal Deoxynucleotidyl Transferase‐Mediated dUTP‐Nick End Labeling (TUNEL) Assay

TUNEL assays were performed to detect apoptotic cells using an In situ Cell Death Detection Kit (Roche Diagnostics, Basel, Switzerland, http://www.roche‐applied‐science.com). TUNEL‐positive cells were detected microscopically or by flow cytometry with CELLQuest software (Becton Dickinson, Franklin Lakes, NJ, http://www.bd.com).

Sirt1 and Luciferase Activity Assays

To quantitate Sirt activity, nuclear extracts were prepared from both Dox‐treated (4 days) and Dox‐untreated Max‐null ESCs according to a method by Dignam et al. [27]. Nuclear extracts were used to measure deacetylase activity of an acetylated histone using an Epigenase Universal SIRT Activity/Inhibition Assay Kit (Epigentek, Farmingdale, NY). Luciferase assays were performed as described elsewhere [28].

Flow Cytometric Analysis of Oct3/4

Cells were fixed with 4% PFA for 10 minutes at 37°C, permeabilized with 90% methanol for 30 minutes on ice, and then blocked with 3% FBS for 1 hour at room temperature. Then, cells were incubated with anti‐Oct3/4 or control IgG conjugated with Per‐CP‐Cy™5.5 in PBS containing 3% FBS for 1 hour. After three PBS washes, cells were analyzed by flow cytometry.

Generation of Lentiviruses for Small Hairpin RNA‐Mediated Knockdown

Lentiviral vectors for small hairpin RNAs (shRNAs) against Sirt1 expression, and a scrambled control were generated by subcloning the following oligonucleotides together with their complementary sequences into a pLKO.1‐puro vector (Sigma). The sequence used for shRNA against p53 expression is described elsewhere [29]. Sirt1 KD1: 5′‐CCGGGATGAAGTTGACCTCCTCATTCAAGAGATGAGGAGGTCAACTTCATC TTTTTT‐3′; Sirt1 KD2: 5′‐CCGGGCCATGTTTGATATTGAGTTTCAAGAGAACTCAATA TCAAACATGGCT TTTTT‐3′; Sirt1 KD3: 5′‐CCGGCGCGGA TAGGTCCATATACTTCAAGAGAGTATATGGACCTATCCGC GT TTTTT‐3′; Scrambled: 5′‐CCGGTGCATACCAGTGGCTA TTTTTCAAGAGAAAATAGCCACTGGTATGCAT TTTTT‐3′.

To produce viruses carrying a specific shRNA expression unit, vectors bearing the above sequences or that for p53 knockdown were individually transfected into 293FT cells with Lipofectamine 2000 according to the manufacturer's instructions (Invitrogen). Max‐null ESCs were infected with the generated lentiviruses in the presence of 8 μg/ml polybrene and then selected with 3 μg/ml puromycin for 6 days.

Microarray Analysis

Biotin‐labeled cRNA was synthesized as recommended by the Affymetrix guidelines. Labeled samples were hybridized to Affymetrix GeneChip Mouse Genome 430 2.0 arrays according to the manufacturer's instructions. Microarray expression data were background‐subtracted and normalized using the robust multiarray analysis method as described elsewhere [14] and are deposited in NCBI's Gene Expression Omnibus under accession number GSE37917.

RESULTS

Identification of Chemicals that Mitigate the Detrimental Phenotype of Max‐Null ESCs

We recently generated inducible Max‐null ESCs in which both alleles of the Max gene were disrupted, while Max cDNA was introduced together with a tetracycline‐off system into the ROSA26 locus and demonstrated that ablation of Max expression is accompanied by loss of the undifferentiated state and cell death under an empirical ESC culture condition [14]. We also showed that these two phenomena do not occur randomly but under a compulsory ordered rule in which cell death only occurs after loss of the undifferentiated state. Therefore, Oct3/4‐ and TUNEL‐positive cells were mutually exclusive in Max‐null ESCs in which Max expression was eliminated by the addition of Dox as shown in Figure 1A. To elucidate molecular mechanisms underlying these events, we attempted to identify chemicals that alleviated the detrimental phenotype of ESCs after ablation of Max expression. We examined the effects of chemicals including kinase inhibitors (p38^MAPK, c‐Jun N‐terminal kinase, Ataxia telangiectasia mutated, Rho kinase, and DNA protein kinase), chromatin modifiers (Trichostatin A, 5‐azacytidine, and BIX‐01294), antioxidants (glutathione, N‐acetyl‐cysteine, and Trolox), and others (Nam, PFTα, nicotinamide adenine dinucleotide). This screening identified three chemicals (p38^MAPK inhibitor SB203580, Nam, and PFTα) that rescued the detrimental phenotype of Max‐null ESCs (Fig. 1B). Therefore, we focused on the rescued effects by these three chemicals for subsequent experiments.

Figure 1

Attenuation of the detrimental phenotypes of Max‐null embryonic stem cells (ESCs) by chemical inhibitors. (A): Oct3/4 and TUNEL staining of Max‐null ESCs treated with Dox for 6 days. Right panel shows a Western blot of Max protein in Dox‐treated Max‐null ESCs for the indicated days. (B): Microscopic inspection of cells rescued with SB203580, Nam, and PFTα. Max‐null ESCs were transferred to six‐well dishes at clonal density (500 cells per well) under the indicated conditions, and photos were taken after 12 days post‐transfer. Abbreviations: Dox, doxycycline; Nam, nicotinamide; PFTα, pifithrin‐α; TUNEL, terminal deoxynucleotidyl transferase‐mediated dUTP‐nick end labeling.

Involvement of p38^MAPK in the Cell Death Phenotype of Max‐Null ESCs

The rescue effect of SB203580, as shown in Figure 1B, prompted us to examine the phosphorylation level of p38^MAPK in Dox‐treated Max‐null ESCs. Western blot analysis showed that the phosphorylation level of p38^MAPK was significantly elevated in ESCs after Max expression ablation by Dox addition (Fig. 2A). We also examined Sirt1 and p53 protein levels and found that loss of Max expression at least did not noticeably alter the amounts of these proteins in ESCs. Then, we treated Max‐null ESCs with SB203580. We also used PD0325901, a mitogen‐activated protein kinase (MAPK) kinase inhibitor, which preserved the pluripotent state and cell viability of Max‐null ESCs [14] in parallel as a positive control for the rescue. We confirmed the rescue effect of SB203580 treatment, which was observed in the initial screening (Fig. 1B) by obtaining a number of viable ESC colonies (Fig. 2B). We also noted that the effect of SB203580 appeared to be more prominent than the previously demonstrated effect of PD0325901 [14] by evaluating the efficiency of generating viable colonies. We found that SB239063, a more selective inhibitor of p38^MAPK than SB203580 [30], also allowed efficient generation of viable colonies (Supporting Information Fig. S1), supporting the involvement of p38^MAPK in the detrimental phenotype of Max‐null ESCs. To approximate the apoptotic signaling level of SB203580‐treated Max‐null ESCs, we performed a TUNEL assay. We found that the emergence of TUNEL‐positive cells, which was strongly evident among Dox‐treated Max‐null ESCs, was almost completely suppressed by SB203580 treatment (Fig. 2C), indicating that SB203580 treatment suppresses the induction of apoptosis in ESCs after Max expression ablation. However, we noted declines in the amounts of pluripotency marker proteins in SB203580‐rescued cells, although the level of Klf4 was reasonably preserved (Fig. 2D). We also noted that rescued cells elevated differentiation marker gene expression showing a strong propensity toward differentiation into ectodermal and extra‐embryonic cells (Fig. 2E), as was the case with unrescued Dox‐treated Max‐null ESCs [14]. Flow cytometric analyses of Oct3/4 levels in SB239063‐rescued cells indicated that cells failed to mitigate the effect of Max expression ablation on lowing Oct3/4 expression levels in ESCs (Supporting information Fig. S2), similar to that of SB203580‐rescued cells (data not shown; Fig. 2D). We also found that, unlike PD032580, which unusually lowers the c‐Myc protein level irrespective of the presence or absence of Max protein, SB203580 alone did not affect the level of c‐Myc protein. Therefore, as was observed with Dox‐treated Max‐null ESCs, ablation of Max expression was accompanied by a substantial decline in the amount of c‐Myc protein in cells rescued with SB203580 (Fig. 2D; [14]). In summary, our data indicated that, although p38^MAPK did not significantly participate in the initial stage of the ESC response to Max expression ablation, this kinase was crucially involved in apoptotic cell death that subsequently occurs after loss of the undifferentiated state in Max‐null ESCs.

Figure 2

Participation of p38^MAPK in apoptosis but not disruption of pluripotency in Max‐null embryonic stem cells (ESCs). (A): p38^MAPK phosphorylation levels. (B): Effect on viability. After culturing for 12 days post‐transfer at clonal density under the indicated conditions, cells were inspected microscopically (upper panel) and then subjected to Leishman staining (lower panel). (C): Suppression of the emergence of TUNEL‐positive cells by SB20. TUNEL signals were examined among Max‐null ESCs with or without SB20 in the presence or absence of Dox for 6 days. (D): Pluripotency marker protein levels. (E): Quantitative reverse transcription polymerase chain reaction analyses of differentiation marker gene expression. Data are the mean with SD (n = 3). Each value from Dox‐untreated Max‐null ESCs (black bar) was arbitrarily set to one. Dox, doxycycline; SB20, SB203580; PD03, PD0325901; TUNEL, terminal deoxynucleotidyl transferase‐mediated dUTP‐nick end labeling.

Preservation of Pluripotency and Cell Viability of Max‐Null ESCs by Sirt1 Inhibition

Next, we scrutinized the rescue effect of Nam on Max‐null ESCs, because flow cytometric analyses suggested that this chemical appeared to exert a prominent activity to preserve Oct3/4 protein levels in the absence of Max expression in ESCs (Supporting Information Fig. S2). First, we confirmed a dose‐dependent effect of Nam on the viability of Max‐null ESCs (Fig. 3A). Then, we found that Nam treatment almost completely suppressed the emergence of TUNEL‐positive cells (Fig. 3B), similar to that by SB203580 treatment (Fig. 2C). Next, we examined the levels of Oct3/4 in Max‐null ESCs rescued with Nam in the absence of Max expression due to Dox treatment. As shown in Figure 4A, Nam appeared to exert a prominent effect on preventing the emergence of Oct3/4‐negative cells, which was comparable with that of the previously demonstrated effect obtained with PD0325901 [14]. Nam is known to inhibit the activity of poly (ADP‐ribose) polymerase (PARP) [31] as well as Sirt1 [32]. Therefore, to determine which inhibition caused the Nam‐dependent effect, we compared the abilities of Sirtinol and DPQ, specific inhibitors of Sirt1 and PARP, respectively, with respect to Nam activity for maintaining Oct3/4 expression in Max‐null ESCs. Flow cytometric analyses showed that PD0325901, Nam, and Sirtinol, but not DPQ, were able to preserve Oct3/4 expression in the absence of Max expression (Fig. 4A), suggesting that the effect of Nam was largely due to Sirt1 inhibition. This result was further supported by Western blotting that showed expression levels of pluripotency marker proteins in Max‐null ESCs treated with these chemicals (Fig. 4B). In parallel with examining the effects of these chemicals on Dox‐treated Max‐null ESCs, we also performed control experiments with Dox‐untreated cells and confirmed that none of these chemicals affected the levels of pluripotency marker gene expression (Supporting Information Fig. S3). We also performed DNA microarray analyses to assess the effect of Nam treatment on global gene expression in Max‐null ESCs. We found that Nam treatment significantly alleviated the magnitude of changes in gene expression caused by Max expression ablation in ESCs (Supporting Information Fig. S4A). Consistent with this finding, real‐time polymerase chain reaction analyses also showed that Nam treatment almost completely eliminated Max ablation‐coupled elevation of differentiation marker gene expression (Supporting Information Fig. S4B). Next, to determine whether Sirt1 enzymatic activity was indeed elevated in Dox‐treated Max‐null ESCs, we prepared nuclear extracts from both Dox‐treated (4 days) and Dox‐untreated Max‐null ESCs and measured Sirt1 deacetylase activity using an acetylated histone as a substrate. As shown in Figure 4C, we found that nuclear extracts from Dox‐treated Max‐null ESCs showed twofold or higher activity than those from untreated cells. Taken together, our results demonstrated that Sirt1, which showed elevated activity upon Max expression ablation in ESCs, is crucially involved in loss of the undifferentiated state.

Figure 3

Rescue of Max‐null embryonic stem cells (ESCs) by Nam. (A): Nam rescued Max‐null ESCs in a dose‐dependent manner. Viability testing was performed under the indicated conditions described in Figure 2B. (B): Effect of Nam on suppression of TUNEL signals. TUNEL assays were performed with Max‐null ESCs cultured for 6 days under the indicated conditions described in Figure 2C. Abbreviations: Dox, doxycycline; Nam, nicotinamide; TUNEL, terminal deoxynucleotidyl transferase‐mediated dUTP‐nick end labeling.

Figure 4

Rescue effect of Nam is due to Sirt1 inhibition. (A): Sirtinol, but not DPQ, exerted a rescue effect. Max‐null embryonic stem cells (ESCs) treated with Nam, Sirtinol, DPQ, and PD03 in the presence of Dox for 6 days were subjected to flow cytometry. (B): Western blot analyses showed that Sirt1 inhibition counteracted the decline of pluripotency marker expression. (C): Elevation of Sirt activity in Dox‐treated Max‐null ESCs. Sirt deacetylase activities were measured with extracts from Dox‐treated (4 days) and Dox‐untreated Max‐null ESCs using an acetylated histone as a substrate as described in Materials and Methods. Data are the mean with SD (n = 3). The Sirt activity of extracts from Dox‐untreated cells was arbitrarily set to one. Abbreviations: Dox, doxycycline; DPQ, 3,4‐Dihydro‐5[4‐(1‐piperindinyl)butoxy]‐1(2H)‐isoquinoline; Nam, nicotinamide; PD03, PD0325901.

Sirt1 and p53 Participate Differently in the Detrimental Pathway of Max‐Null ESCs

To obtain an independent line of evidence that showed impairment of the undifferentiated state occurred in a Sirt1‐dependent manner in Max‐null ESCs, we performed RNA interference (RNAi) experiments using lentiviral vectors. We prepared three sets of Sirt1 knockdown constructs, in which Sirt1 KD1 and Sirt1 KD2 performed reasonably well, whereas Sirt1 KD3 was almost nonfunctional. We found that transfection of Sirt1 KD1 and Sirt1 KD2 vectors led to a substantial decline in the number of TUNEL‐positive cells among Dox‐treated Max‐null ESCs, whereas the Sirt1 KD3 construct did not show an appreciable effect (Fig. 5A). Thus, there was a strong correlation between the efficiency of knockdown and the elimination of TUNEL‐positive cells among Max‐null ESCs, providing additional evidence that Sirt1 was involved in the detrimental pathway activated after Max expression ablation in ESCs. We also found that p53 knockdown appeared to be equally effective, compared with that of Sirt1 (Fig. 5B), at decreasing the number of TUNEL‐positive cells. However, results from luciferase assays with a reporter carrying the p53‐dependent cyclin G1 gene promoter indicated that, unlike Sirt1 activity, elevation of p53 activity in Dox‐treated Max‐null ESCs appeared to be very little, if any (Supporting Information Fig. 5). In accordance with these data, DNA microarray data of the expression of endogenous p53‐downstream genes indicated that the p53 pathway was not significantly activated in Max‐null ESCs, compared with that in nucleostemin‐null ESCs in which overall p53 activity was demonstrated to be prominently elevated [28] (Fig. 5C). Thus, our data imply that the rescue effects of p53 knockdown and PFTα treatment are not mainly due to suppression of p53‐dependent transcriptional activation (Discussion section for possible mechanisms).

Figure 5

Rescue of cell viability by Sirt1 and p53 knockdown. (A): Correlation between Sirt1 activity and TUNEL signals. Sirt1‐ and p53‐knockdown Max‐null ESCs were subjected to coimmunostaining of Oct3/4 and TUNEL after 6 days post‐doxycycline (Dox) addition. (B): Comparable activities of TUNEL signal suppression between Sirt1 and p53 knockdown. Each column represents the mean with SD (n = 3). *, p < .01 versus control. (C): Comparison of changes in expression levels of endogenous p53‐downstream genes between Max‐null and nucleostemin‐null ESCs. Induction folds due to Dox addition represent signal values of DNA microarray data from Dox‐treated cells for 4 days which were divided by those from corresponding Dox‐untreated cells with respect to 43 p53‐downstream genes. Actual signal values of Max‐null and nucleostemin‐null ESCs are shown in Supporting information Table S1. Abbreviations: ESCs, embryonic stem cells; TUNEL, terminal deoxynucleotidyl transferase‐mediated dUTP‐nick end labeling.

Our analyses showed differences in morphology of the rescued cell colonies and the intensity of Oct3/4 protein signals between Sirt1 and p53 knockdowns, although, similar to Sirt1 knockdown, suppression of caspase‐3 activation by p53 knockdown was very effective (Fig. 6A). Flow cytometric analyses clearly showed the difference in efficiency to suppress the emergence of Oct3/4‐negative cells between knockdowns of Sirt1 and p53 (Fig. 6B). Consistent results were obtained by flow cytometric analyses of cells rescued with chemicals, that is, Nam and PFTα (Supporting Information Fig. S2). Western blotting also revealed that Sirt1 KD1 strongly counteracted the decrease in expression levels of pluripotency factors, such as Oct3/4, Sox2, and Nanog, while the effect of p53 KD was less significant (Fig. 6C). In Sirt1 and p53 RNAi‐treated cells, levels of phosphorylated p38^MAPK were significantly lower than that of control cells infected with lentiviruses carrying scrambled shRNA, which was consistent with the assumption that Sirt1 and p53 were regulators functioning at an earlier stage than p38^MAPK in the compulsory ordered pathway leading to cell death in Max‐null ESCs. It is noteworthy that Sirt1 knockdown, but not that of p53, strongly reduced the phosphorylation levels of ERK, suggesting that Sirt1 was upstream of ERK in the hierarchy (Fig. 6C). We also noted that Sirt1 knockdown reduced c‐Myc protein levels regardless of Dox treatment, while p53 knockdown preserved the high level of c‐Myc protein even in the absence of Max protein in ESCs. We assume that the low level of c‐Myc protein in Sirt1 knockdown cells was due to reduced ERK activity in these cells, because ERK inhibition is known to significantly destabilize the c‐Myc protein [14, 33], while preservation of the high level of c‐Myc protein by p53 knockdown may indicate elimination of a negative feedback loop of the c‐Myc‐p53 control system [34]. However, we do not presently know the biological significance of the latter finding, given that this elevated level of c‐Myc protein failed to function as a transcription factor to provoke a prominent biological effect in the absence of Max association. Unlike c‐Myc, Max‐interacting N‐Myc, L‐Myc, and transcriptional repressor Mad3 did not show such a p53 knockdown‐mediated accumulation (Supporting Information Fig. S6).

Figure 6

Effect of Sirt1 and p53 knockdown on preservation of pluripotency. (A): Effect of Sirt1 and p53 knockdown on Oct3/4 expression and caspase‐3 activation. Coimmunostaining was performed as described in Figure 1A. (B): Comparison of the activities for preserving Oct3/4 expression between Sirt1 and p53 knockdown. (C): Western blotting to evaluate the ability to preserve the ESC status by Sirt1 and p53 knockdown. Abbreviation: Dox, doxycycline.

DISCUSSION

We previously demonstrated that ablation of Max expression in ESCs induces loss of pluripotency and caspase‐dependent apoptotic cell death in this order [14]. In this study, we studied molecules that are responsible for these phenotypes. Our analyses revealed that chemical inhibitors of p38^MAPK, Sirt1, and p53 allow generation of viable cell colonies even in the absence of Max expression in ESCs.

p38^MAPK is the mammalian ortholog of yeast HOG kinase, which participates in a signaling cascade after its activation by growth factors and various cellular stresses including osmotic shock and UV light [35]. p38^MAPK consists of four isoforms, α, β, γ, and δ, and only the p38α isoform is expressed in ESCs [36]. Whereas no role has been assigned to p38^MAPK in ESC pluripotency, this protein is known to control differentiation between ectodermal and mesodermal cell lineages when ESCs are induced to differentiate [37, 38]. We could obtain the rescue effect with SB203580 as well as SB239063 that is three times more selective toward p38^MAPK than SB203580 and displays more than 220‐fold selectivity than those of ERK and other kinases [30], indicating that the rescue effect obtained with these chemicals is indeed due to the inhibition of p38^MAPK activity. Consistent with the notion of no role of p38^MAPK in ESC pluripotency [37, 38], rescued Max‐null ESCs with either SB203580 or SB239063 in both cases failed to preserve the pluripotent state. On the other hand, two other chemicals, PFTα and in particular Nam, did preserve the pluripotent state of Max‐null ESCs. Nam can inhibit PARP and Sirt1 activities [31, 32]. However, our data obtained with the Sirt1‐specific inhibitor Sirtinol and from Sirt1‐knockdown studies strongly suggest that the rescue effect of Nam is due to its inhibition of Sirt1 activity. Involvement of p53 in the detrimental phenotypes of Max‐null ESCs was also confirmed not only by chemical inhibition with PFTα but also knockdown studies of p53 gene expression. p38^MAPK is a known activator of p53, and activated p53 functions as a negative regulator of p38^MAPK by elevating the expression level of the p53‐downstream gene Wip‐1 [39]. However, because Wip‐1 gene expression did not elevate, but rather decreased its expression level upon ablation of Max expression in ESCs (Supporting Information Table S1), we assume that such a regulatory relationship between p53 and p38^MAPK may not be prominent in Max‐null ESCs. Rather, our data indicated that p53 exerts its activity at a higher level than that of p38^MAPK in the hierarchy of the detrimental pathway that operates in Max‐null ESCs, because p53 inhibition prevented p38^MAPK activation in Max‐null ESCs. Although the p38^MAPK activation level was also kept low in Max‐null ESCs subjected to Sirt1 inhibition, it is noteworthy that rescued phenotypes obtained with Sirt1 and p53 inhibition are different with respect to activation of ERK signaling in which p53 inhibition did not suppress ERK activation, suggesting that the effect of p53 inhibition on the preservation of ESC pluripotency is independent of the Sirt1‐ERK axis. Figure 7 depicts our proposed model in which Sirt1, but not p53 and p38^MAPK, functions upstream of ERK signaling in the hierarchy. p38^MAPK does not contribute to disruption of the pluripotent state but is crucially involved in affecting cell viability. Although functional modes of p53 in Max‐null ESCs remain elusive, the p53 protein may participate in both facets of the detrimental pathway, loss of pluripotency, and cell death as detailed below.

Figure 7

Molecular scheme of the impairment of pluripotency and cell viability in a compulsory ordered manner occurring in embryonic stem cells (ESCs) after ablation of Max expression. Our data suggest that Sirt1 is upstream of ERK signaling in the detrimental pathway of ESCs with Max expression ablation. p53 is also involved in disruption of pluripotency but independently of ERK signaling. p53 may also affect cell viability directly via transcription‐dependent or ‐independent mechanisms, although no direct evidence of this pathway was obtained at present. p38^MAPK is restrictively involved in affecting cell viability but appears not to influence pluripotency. Abbreviations: Nam, nicotinamide; PD03, PD0325901.

Myc and p53 generally function antagonistically toward each other [34]. Therefore, p53 activity may be derepressed by the decline of Myc activity due to Max expression ablation. Moreover, the fact that p53 inhibition eliminated apoptotic induction almost completely in Max‐null ESCs, indicates that p53 is crucially involved in the detrimental pathway of these cells. However, our data revealed that elevation in overall p53 activity to upregulate its downstream genes was very little, if any in Dox‐treated Max‐null ESCs. A recent report [40] shows that p53, once activated in ESCs, exerts two different types of activities via drastically different molecular mechanisms. One is the well‐established pathway of transcriptional activation of p53‐downstream genes, while the other is the repression of genes involved in self‐renewal and pluripotency of ESCs. Repression of Nanog expression by p53 has also been demonstrated by a previous study [41]. Because Sirt1 is a well‐known negative regulator of p53 activity by the former mechanism [42, 43], we speculate that, although p53 was activated in response to Max expression ablation, the transcriptional activating activity of p53 was alleviated due to the high deacetylase activity of Sirt1 in Max‐null ESCs. However, transcription‐incompetent p53 still impairs pluripotency by exerting the remaining second activity. An alternative pathway, which is not mutually exclusive of the mechanism of p53‐mediated repression of pluripotency genes, may involve p53‐mediated upregulation of the p53‐responsive miR34 expression level in Max‐null ESCs which is known to regulate cell cycle progression, cellular senescence, and apoptosis [43]. It is also possible that p53 initiates apoptosis by translocating into mitochondria with the help of Sirt1, as was previously demonstrated [29]. Because it is generally believed that p53 needs to interact with some members of the apoptotic cascade such as Bax to participate in this transcription‐independent proapoptotic process [39], it will be important to examine whether such interactions exist in Max‐null ESCs.

A previous study [44] demonstrated that the Sirt1 protein preferentially binds to gene promoters involved in differentiation and developmental processes of ESCs, and the expression level of Sirt1 is downregulated during ESC differentiation to appropriately execute differentiation programs. Interestingly, this previous study also demonstrates that elimination of Sirt1 in ESCs, which were not induced to differentiate, has little impact on the expression of pluripotency and developmental factors, and this observation concurs with other Sirt1 studies [29, 45]. However, important roles of Sirt1 in ESCs appear to exist under suboptimal conditions. Indeed, it has been shown that ESCs subjected to DNA damage by oxidative stress localize Sirt1 to their DNA breakage sites to promote DNA repair [46]. On the other hand, Sirt1 has been also shown to contribute toward apoptotic cell death induced by oxidative stress in a different context [29]. Moreover, we demonstrated here that Sirt1 was crucially involved in disruption of the pluripotency state, which serves as an initiating cue for cell death in response to ablation of Max expression in ESCs. Based on these reports, we assume that Sirt1 does not participate in important functions when ESCs are maintained under healthy conditions. However, under conditions such as severe oxidative stress and Max ablation, cell death or survival of ESCs is determined by Sirt1, which promotes DNA repair in cells that are recoverable, but disrupts pluripotency as the prerequisite step for cell death if damage exceeds a repairable level. p53 has been shown to make a similar binary decision between cell death and survival in numerous cell types including ESCs that are subjected to sever stresses such as ionizing irradiation [47]. Therefore, it is tempting to speculate that Sirt1 and p53 collaborate or strongly influence each other to determine cell death or survival under stressful conditions.

We demonstrated both here and in previous work that loss of pluripotency and apoptotic cell death occur in this order in Max‐null ESCs. Notably, there is a previous report that describes a related observation. Guo et al. [48] demonstrated that forced elimination of Oct3/4 expression sensitizes ESCs to apoptotic signals upon UV exposure or other types of stresses, compared with that of ESCs with preserved Oct3/4 expression even within the time frame retaining other pluripotency marker gene expression. Moreover, our recent data demonstrated that ESCs follow the compulsory ordered rule even when cell death is induced by treatment with H₂O₂ or doxorubicin (T. Hishida and A. Okuda, unpublished observation). Thus, these results suggest that this compulsory ordered mechanism appears not to be restrictively enforced in Max‐null ES cells but generally operates in ESC death. Therefore, it is important to know whether p53 and in particular Sirt1, which we identified as initiators of disruption of pluripotency in Max‐null ESCs, also participate in a similar manner in ESCs subjected to other types of stresses.

ESCs and cancer cells share numerous biological characteristics such as rapid growth and prolonged self‐renewal [49, 50]. Similarities in transcriptional programs between ESCs and cancer cells have been recently characterized by higher‐ordered system‐level analyses [18, 51]. Moreover, Sirt1 has been shown to promote apoptotic cell death in both ESCs and cancer cells under certain conditions [29, 52]. Thus, it is possible that there is a commonality of cell death initiation between ESCs and cancer cells. Therefore, our future studies will address whether ablation of Max expression in cancer cells induces extensive cell death, as is the case with ESCs, under the compulsory ordered mechanism in which Sirt1 may disrupt the cancer state. In summary, we anticipate that our discovery of a role for Sirt1 as a disruptor of the pluripotent state in ESCs will provide new avenues for studying the mechanisms of pluripotency and cell death of ESCs and other cells such as cancer cells.

CONCLUSION

We demonstrate that Sirt1, p53, and p38^MAPK are crucially involved in the detrimental phenotypes observed with ESCs after ablation of Max expression. Our data also show that cells, rescued by chemical inhibitors or suppression of gene expression by RNA interference, possess variable levels of pluripotency dependent on which protein was targeted for inhibition. Indeed, rescue by inhibition of p38^MAPK activity does not lead to significant preservation of pluripotency, whereas inhibition of p53 and in particular Sirt1 preserves the pluripotent state of ESCs in which Max expression has been eliminated. To our knowledge, this is the first demonstration that Sirt1 is involved in disruption of ESC pluripotency, providing a new approach to understand the molecular basis of the pluripotent state of ESCs.

Acknowledgements

We thank Tomoko Okuda for technical assistance and Dr. Hitoshi Niwa for helpful discussions. This study was supported in part by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan and primarily by a Grant‐in‐Aid for the Support Project of Strategic Research Center in Private Universities to the Saitama Medical University Research Center for Genomic Medicine. This study was performed as a part of the Core Research for Evolutional Science and Technology (CREST) Agency. T.H. was a recipient of a Grant‐in‐Aid for Young Scientists (B) from the Japan Society for the Promotion of Science (JSPS). T.H. is currently affiliated with Gene Expression Laboratory (GEL‐B), Salk Institute for Biological Studies, La Jolla, California.

DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

The authors declare no potential conflicts of interest to this work.