Author contributions: V.T.: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing, and read and approved the final version of the manuscript; L.O., Y.S., B.K., and P.C.: collection and assembly of data, data analysis and interpretation, and read and approved the final version of the manuscript; M.P.S., P.P., V.M., S.S., L.A., E.C.R., and K.M.: collection and assembly of data and read and approved the final version of the manuscript; A.C.: data analysis and interpretation and read and approved the final version of the manuscript; M.J.W. and C.G.: conception and design, financial support, manuscript writing, and read and approved the final version of the manuscript.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS May 24, 2012.
Phosphorylation of histone H2AX (γH2AX) is known to be the earliest indicator of DNA double-strand breaks. Recently, it has been shown that mouse embryonic stem cells (mESCs) have very high basal levels of γH2AX, even when they have not been exposed to genotoxic agents. As the specialized role of high basal γH2AX levels in pluripotent stem cells is still debated, we investigated whether H2AX phosphorylation is important in maintaining self-renewal of these cells. Here, we report that not only mESCs but also mouse-induced pluripotent stem cells (miPSCs), have high basal levels of γH2AX. We show that basal γH2AX levels decrease upon ESC and iPSC differentiation and increase when the cells are treated with self-renewal-enhancing small molecules. We observe that self-renewal activity is highly compromised in H2AX−/− cells and that it can be restored in these cells through reconstitution with a wild-type, but not a phospho-mutated, H2AX construct. Taken together, our findings suggest a novel function of H2AX that expands the knowledge of this histone variant beyond its role in DNA damage and into a new specialized biological function in mouse pluripotent stem cells. STEM CELLS2012;30:1414–1423
Histones are the fundamental components of chromatin's basic unit, the nucleosome. These highly evolutionarily conserved proteins possess tail motifs in both the C- and N-terminal regions, which are the target for different post-translational modifications [1, 2]. It has been shown that covalent modification in the tail of some histone variants results in specialized biological functions beyond their canonical function in the nucleosome [3, 4].
H2AX is a variant of the histone H2A family, characterized by an extended C-terminal tail containing an evolutionarily conserved serine-glutamine (SQ) motif. In response to DNA double-strand breaks (DSBs), H2AX is phosphorylated on a serine residue at position 139 of the SQ motif, yielding a phosphorylated form of the protein known as γH2AX . γH2AX interacts with several DNA repair proteins, indicating that it is involved in DNA repair processes .
Recent research has described potentially new and specialized roles for γH2AX in mouse embryonic stem cells (mESCs), in addition to the canonical DSB response. Banath et al. associated high basal γH2AX levels in ESCs with global chromatin decondensation rather than pre-existing DNA damage . High levels of γH2AX have also been associated with the single-strand breaks occurring in S-phase, and ESC populations, as rapidly dividing cells, have an increased proportion of cells in S-phase, compared to many somatic cells . Analogously, high γH2AX levels have also been linked to the replicative stress in the fast dividing cells of the early embryo in the absence of any induced DNA damage . Other authors have suggested a role of γH2AX in the negative regulation of ESC proliferation, via GABA-A receptor-dependent signaling [10, 11].
Until now, γH2AX has not been functionally linked to any of the key characteristics of ESCs, namely pluripotency and self-renewal. Pluripotency is the ability to differentiate into any tissue of the three embryonic layers, while self-renewal is the ability to maintain the undifferentiated state through cell division cycles. Self-renewal is finely regulated by different pathways: leukemia inhibitory factor (LIF)/STAT3 is considered the most important in mESCs [12–14], but other pathways such as the ones mediated by PI3K along with suppression of ERK 1/2 and GSK3 are also involved in optimal maintenance of self-renewal [15, 16], along with the transcription factors Oct4, Sox2, and Nanog . Histone modifiers also have an important role in ESC self-renewal. The chromatin of self-renewing ESCs exhibits a characteristic structure of increased accessibility due to fewer and more loosely bound histones . When ESCs undergo differentiation, their chromatin structure changes dynamically in response to global histone modifications. Examples are the differentiation-dependent increase in the silenced chromatin mark trimethylated residue K9 of histone H3 (H3-triMeK9) and a decrease in the global levels of acetylated histones H3 and H4, which are usually associated with active chromatin regions [19–21]. Besides core histones, modifications of histone variants might represent an additional means by which a pluripotent stem cell can regulate chromatin structure and function.
As the specialized role of high basal γH2AX levels in pluripotent stem cells is still debated, in this study, we investigated whether H2AX phosphorylation is important in maintaining self-renewal of pluripotent mouse ESCs and induced pluripotent stem cells (iPSCs), somatic cells reprogrammed to pluripotency by the over-expression of specific sets of genes [22–24]. Here, we report that not only mESCs but also miPSCs have high basal levels of γH2AX. We demonstrate that basal γH2AX levels decrease upon ESC and iPSC differentiation and increase when the cells are treated with self-renewal-enhancing small molecules. We observe that self-renewal activity is highly compromised in H2AX−/− cells and that it can be restored in these cells through reconstitution with a wild-type (WT), but not a phospho-mutated, H2AX construct. Taken together, our findings enrich the panel of the γH2AX-specific functions in pluripotent stem cells.
MATERIALS AND METHODS
ESC and iPSC Culture and Treatments
Four cultures of undifferentiated ESCs were used: a Nanog-green fluorescent protein (GFP) reporter ESC (Nanog-ESC) line , kindly provided by Prof. I. Chambers (Institute for Stem Cell Research, University of Edinburgh, Scotland) for use by Prof. Welham at University of Bath; the E14 ESC line ; the WT parental TC1 line and the H2AX−/− mESCs derived from TC1, kindly provided by Dr. F.W. Alt (Children's Hospital at Boston)  for use by Prof. Giachino at University of Turin. One single mouse Nanog-iPSC line was used  and was obtained from Riken Cell Bank for use by Prof. Welham at University of Bath. The ESCs and iPSCs were cultured on 0.1% (wt/vol) gelatin-coated tissue culture plates (Nunc, Roskilde, Denmark) in: knockout Dulbecco's modified Eagle's medium (KO-DMEM) supplemented with 15% (vol/vol) knockout serum replacement (KO-SR), 0.1 mM 2-mercaphtoethanol, 2 mM glutamine, 0.1 mM nonessential amino acids (all from Gibco-Invitrogen, Paisley, U.K.), and 1,000 units/ml murine LIF (Millipore Corporation, Bedford, MA); Glasgow Minimal Essential medium (GMEM) (Gibco-Invitrogen) containing 10% (vol/vol) fetal bovine serum (FBS) (ES tested Hyclone, Thermo Scientific, Rockford, IL), 0.1 mM 2-mercaphtoethanol, 1 mM sodium pyruvate, 2 mM L-glutamine, 0.1 mM nonessential amino acids (all from Gibco-Invitrogen), either with or without 1,000 units/ml murine LIF (Millipore Corporation). To generate embryoid bodies (EB), ESCs or iPSCs were placed into semisolid methylcellulose cultures to induce multilineage differentiation . NIH3T3 mouse embryonic fibroblast cell line was purchased from the American Type Culture Collection (Rockville, MD) and cultured in DMEM supplemented with 10% (vol/vol) FBS, 2 mM L-Glutamine, 50 μg/ml kanamycin, 1 mM sodium pyruvate, and 0.1 mM nonessential amino acids (all from Gibco Invitrogen). Ku-55933 and NU7026 were supplied by Sigma-Aldrich Co. (St. Louis, MO) and used at 10 μM. Caffeine was supplied by Sigma-Aldrich Co. and used at 10 mM. 1m, a GSK3 inhibitor , was used at 2 μM and PD0325901, a mitogen-activated protein kinase kinase (MEK) inhibitor , was supplied by Merck Chemicals (Darmstadt, Germany) and used at 1 μM. Cells were X-irradiated (2 Gy) using a RADGIL system (Gilardoni, Lecco, Italy) at a dose of 0.65 Gy/minute.
Three 14-week old female C57Bl/6 x CBA F1 mice were mated with C57Bl/6 males and had visible copulation plugs the following day. At 3.5 days postcoitum, blastocysts were isolated according to standard protocols , collected in M2 medium (Sigma-Aldrich Co.), and treated for 15 minutes with 3 mg/ml pronase (Sigma-Aldrich Co.).
Immunofluorescence and Confocal Microscopy Analyses
Cells were stained for γH2AX with the anti-γ-H2AX mAb (clone JBW301) (Millipore Corporation) as previously described . To stain with other antibodies, cells were either grown attached to gelatin-coated slides (IBIDI, Martinsried, Germany) or collected, fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100, and blocked with 6% (wt/vol) bovine serum albumin (BSA) (Sigma-Aldrich Co.) and 2.5% (vol/vol) normal goat serum (Sigma-Aldrich Co.). Cells were stained with either anti-phospho-ATM (Ser1981) (clone 10H11.E12) (Rockland Immunochemicals, Gilbertsville, PA) or anti-p53-binding protein 1 (anti-53BP1) (Novus Biological, Littleton, CO) as primary Abs and then with either Alexa 546-conjugated goat anti-mouse goat anti-rabbit as secondary Abs (Molecular Probes, Inc., Eugene, OR). Nuclei were stained with 0.1 μg/ml 4′-6-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich Co.). Fluorescence images were obtained with a Zeiss LSM 510 Meta confocal microscope using a ×63 objective. To quantify cell fluorescence, for each experiment, cell areas of at least 100 cells were selected and mean fluorescence intensity (MFI), expressed as arbitrary units per pixel, was quantified for each cell, using LSM Image Examiner software from Zeiss. For statistical analysis either Student's t tests or one-way analysis of variance (ANOVA) tests were performed and the significance was expressed with asterisks: *, p < .05; **, p < .01; ***, p < .005.
To determine the ability of ESCs to retain an undifferentiated phenotype, self-renewal assays were performed as previously described . In principle, this assay is based on alkaline phosphatase expression, which is high in undifferentiated ESCs and lost upon differentiation. Cells were plated at 1.8 × 103 to 3 × 103 cells per gelatin-coated 9.6 cm2 well in GMEM (Gibco-Invitrogen) containing 10% (vol/vol) FBS (Hyclone), 0.1 mM mercaptoethanol, 1 mM sodium pyruvate, 2 mM glutamine, 0.1 mM nonessential amino acids (all from Gibco-Invitrogen), and 1,000 units/ml murine LIF (Millipore Corporation). Where required, following a 6–15-hour period of cell adherence, cultures were supplemented with either 1m or PD0325901. To detect cells expressing alkaline phosphatase, a marker of pluripotency, after 5–6 days of culture, dishes for each treatment were washed, fixed, and then stained for 20 minutes with a solution containing 1 mg/ml Fast Red TR salt (Sigma-Aldrich Co.) dissolved in 0.1 m Tris, pH 9.2, containing 200 μg/ml naphthol AS-MX phosphate (Sigma-Aldrich Co.). Alkaline phosphatase-positive colonies, indicative of undifferentiated, self-renewing ESC colonies and unstained, differentiated colonies, were counted in duplicate or triplicate for each treatment. For statistical analyses, either paired Student's t tests or one-way ANOVA tests were performed and the significance indicated by asterisks: *, p < .05; **, p < .01; ***, p < .005.
Cells were plated in 12-well trays coated with gelatin at a starting concentration of either 3,500 cells per square centimeter or 7,000 cells per square centimeter in either complete KO-DMEM+15% (vol/vol) KO-SR plus 1,000 U/ml LIF or complete GMEM supplemented with 10% (vol/vol) FBS plus 1,000 U/ml LIF and then harvested after 24, 48, 72, and 96 hours. Where required, following a 6-hour period of cell adherence, cultures were supplemented with either 1m or PD0325901. Cells were counted in triplicate using a Neubauer chamber. Doubling times were calculated by the free-software available from the web site www.doubling-time.com.
Preparation of Cell Lysates and Immunoblotting
Cells were harvested and washed with PBS, pelleted, and lysed in Laemmli buffer (0.125 M Tris-HCl [pH 6.8], 5% SDS) containing a phosphatase and protease inhibitor cocktail (Roche, Mannheim, Germany). Total lysates were boiled for 2 minutes, sonicated, and quantitated by the microbicinchoninic acid method (Thermo Scientific, Rockford, IL). Twenty or 40 μg of total lysates were size fractionated by SDS-PAGE 12% gels and electroblotted onto polyvinylidene difluoride membranes (GE Healthcare, Buckinghamshire, U.K.). After blocking with 5% (wt/vol) nonfat dried milk in PBS plus 0.1% (vol/vol) Tween (Sigma-Aldrich Co.) or 5%(wt/vol) BSA in Tris-buffered saline (TBS), pH7.5, the membranes were incubated with anti-γH2AX (clone JBW301) (Millipore Corporation), anti-β-actin (clone AC-72) (Sigma-Aldrich Co.), anti-Nanog (ab80892, Abcam, Cambridge, U.K.), anti-Oct4 (sc 9081, Santa Criz Biotechnology), anti-Gapdh (AM4300, Ambion)-specific antibodies, and subsequently with peroxidase-conjugated secondary antibodies (GE Healthcare or Dako, Glostrup, Denmark). The immunoreactive bands were visualized by ECL Super Signal (Thermo Scientific) on autoradiographic films. Autoradiographic bands were scanned and quantified by Kodak 1D Image Analysis Software. Alternatively, signals were detected using EC Prime (GE Healthcare), images were captured using an ImageQuant RT-ECL system and digital signals quantified using ImageQuant TL software (GE Healthcare).
Cells were stained for γH2AX with the anti-γ-H2AX mAb (clone JBW301) (Millipore Corporation) as previously described . To assess γH2AX expression in the different phases of the cell cycle, after staining for γH2AX, cells were treated with 100 μg/ml RNase A (Sigma-Aldrich Co.) and stained with 50 μg/ml propidium iodide (Sigma-Aldrich Co.) for 45 minutes at 37°C. For human influenza hemaglutinin (HA) staining, approximately 400,000 cells were collected, fixed in cold 70% (vol/vol) ethanol, and stored at −20°C for up to 2 weeks before analysis. Cells were washed in TBS pH 7.4 and then rehydrated for 10 minutes at 4°C in TBS containing 4% (vol/vol) FBS and 0.1% (vol/vol) Triton X-100 (TST) (Sigma-Aldrich Co., St Louis, MO) prior to staining with a direct-conjugated anti-HA-PE Ab (Miltenyi Biotec, Bergisch Gladbach, Germany) diluted 1:100 in TST and incubated for 2 hours at 37°C. A minimum of 20,000 stained cells were acquired on a CyAn ADP flow cytometer (Beckman Coulter, Brea, CA) and analyzed with the Summit 4.0 software.
Production of Retroviral Supernatants and Transduction of ESCs
For retroviral production, 2 × 106 Plat-E packaging cells were cultured in 10-cm culture plates in DMEM supplemented with 10% (vol/vol) heat-inactivated FCS (Gibco-Invitrogen) and 100 μg/ml penicillin/streptomycin (Gibco-Invitrogen). On the following day, the cells were transfected using the calcium-phosphate technique with the following vectors: HA-tagged H2AX and HA-tagged S139A mutant of H2AX gateway compatible destination vectors kindly provided by Dr. M.S.Y. Huen (The University of Hong Kong)  and pIRES2-EGFP vector (Clontech laboratories, Palo Alto, CA). Culture medium was replaced 12–16 hours after transfection and supernatants were harvested and filtered 30–48 hours later. The retroviral supernatants were either used immediately or stored at −80°C until use. For stable retroviral transduction, 100,000 ESCs were seeded into 48-well Nunclon surface plates (Nunc), and after 24 hours, they were incubated with 1 ml of viral supernatants. The medium was replaced after 16–18 hours. Transduced cells were constantly selected with puromycin 0.75 μg/ml (Gibco-Invitrogen).
ESCs and iPSCs Have High Basal γH2AX Levels
A mouse ESC line (Nanog-ESCs)  and a mouse iPSC line (Nanog-iPSCs) , in which GFP expression is under the control of the endogenous Nanog promoter, were assessed for the presence and levels of γH2AX (Fig. 1). Immunofluorescence and confocal microscopy analyses highlighted the high numbers of endogenous γH2AX foci in both cell lines (Fig. 1A and supporting information Fig. S1A, S1B), with a distribution analogous to foci that are produced after exposure to ionizing radiation (IR) ( and Fig. 3A) and with the majority of the cells being positive for both γH2AX and the stemness marker Nanog (supporting information Fig. S1C). Western blot analyses confirmed high γH2AX levels in the two Nanog-GFP cell lines and showed analogous results in two other ESC lines (E14 and TC1) (Fig. 1B). No γH2AX antibody staining was observed in H2AX−/− ESCs, confirming a lack of nonspecific staining (Fig. 1B, lane 4). We then analyzed γH2AX levels in freshly isolated mouse day 3.5 blastocysts and found that cells located within the inner cell mass (ICM), which are the source of ESCs, were positive for both γH2AX and the pluripotency marker Oct4 (Fig. 1C).
Ataxia telangiectasia mutated (ATM) and DNA dependent protein kinase (DNA-PK) are the major kinases that mediate H2AX phosphorylation following IR [36, 37]. Considering that both kinases are expressed in ESCs (data not shown), we investigated whether they were both involved in maintaining the high γH2AX levels observed in mESCs by treating cells with Ku-55933, a specific ATM inhibitor , and NU7026, a specific DNA-PK inhibitor . We analyzed γH2AX levels by immunofluorescence, in terms of MFI, and by immunoblotting. After either ATM or DNA-PK inhibition, we observed a partial reduction of γH2AX levels, indicating that both kinases contribute to endogenous H2AX phosphorylation in ESCs (Fig. 2). However, dual inhibition of both ATM and DNA-PK did not abolish H2AX phosphorylation completely, indicating that other kinases might also be involved (Fig. 2). Another kinase implicated in H2AX phosphorylation is ataxia telangiectasia and rad-3 related kinase (ATR), which is associated with phosphorylation occurring in association with damaged replication forks . In accordance with previous results [41, 42], treatment with 10 mM caffeine, an ATM, and ATR inhibitor  was associated with γH2AX increment and massive apoptosis (our data, not shown) in these cells, thus it was not possible to directly evaluate the contribution of ATR to endogenous H2AX phosphorylation in ESCs; however, since ESCs are actively replicating cells, it appears conceivable that ATR will also take a role in this phosphorylation. These data suggest that a high γH2AX level is a particular feature of pluripotent cells, both those cultured in vitro as well as cells within the early embryo; moreover, our data suggest that iPSCs are comparable to ESCs in terms of γH2AX basal expression and that ATM, DNA-PK, and possibly ATR are all involved in maintaining the basal level of H2AX phosphorylation in ESCs.
High Basal γH2AX Levels Do Not Correlate with Either Proteins of the DNA Damage Response or the High Proliferation Rate of ESCs
Many proteins involved in the DNA damage response can aggregate as microscopically visible foci, colocalizing with γH2AX at sites of DSBs and mediating DSB repair via either homologous recombination or nonhomologous end joining . We examined ESCs for the damage sensors p53-binding protein 1 (53BP1) and phospho-ATM. Neither proteins form nuclear foci colocalizing with γH2AX foci in untreated ESCs (Fig. 3A). ESCs were then irradiated with 2Gy and the staining pattern of the same antibodies was analyzed. Both immunofluorescence and immunoblot analyses revealed that γH2AX itself increased remarkably (Fig. 3A, 3B) and 53BP1 and ATMpSer1981 foci promptly appeared (Fig. 3A). These data, in accordance with previous published results [7, 8], suggest that high basal γH2AX level does not correlate with other proteins of the DNA damage response and that ESCs can respond correctly to induced DSBs.
High levels of γH2AX have also been associated with the single-strand breaks occurring in S-phase, and ESC populations, as rapidly dividing cells, have an increased proportion of S-phase cells, compared to many somatic cells . To test the possibility that basal γH2AX may be ascribed to ESCs DNA replication, we evaluated γH2AX levels in relation to the cell cycle distribution and compared ESCs with differentiated cells (NIH3T3, a mouse embryonic fibroblast cell line). Flow cytometry analyses revealed that NIH3T3 cells are positive for γH2AX only in S and G2M phases, while the two ESC lines analyzed are positive also in G1 phase (Fig. 3B). These data demonstrate that although ESCs replicate very rapidly, the high basal level of γH2AX detected cannot be ascribed only to the increased proportion of cells in S-phase of the cell cycle.
γH2AX Levels Decrease During ESC and iPSC Differentiation
To examine whether γH2AX levels change during in vitro differentiation, we placed both ESCs and iPSCs into semisolid methylcellulose cultures to generate EB and so induce multilineage differentiation. We then analyzed γH2AX levels by immunofluorescence and confocal microscopy. Days 5 and 6 were chosen as the most indicative time points of EB formation and ESC differentiation as previous studies have shown that markers of all three germ layers are present at these time points . γH2AX decreased significantly in EBs derived from both ESCs and iPSCs at day 6, with most of the cells containing only a few foci and some becoming completely negative (Fig. 4A).
We also tested another model of in vitro differentiation, this time withdrawing LIF from monolayer cultures, which induces spontaneous, multilineage differentiation. We compared γH2AX levels in ESCs and iPSCs cultured in KO-DMEM medium supplemented with 15% KO-SR and 1,000 units/ml LIF with those in ESCs and iPSCs cultured in two different media (KO-DMEM and GMEM) supplemented with FBS but in the absence of LIF. γH2AX levels decreased significantly in both ESCs and iPSCs after 5 days of differentiation (Fig. 4B). We obtained analogous results when we similarly induced differentiation in two other ESC lines (E14 and TC1) (supporting information Fig. S2).
In the two Nanog-GFP cell lines, we observed that upon differentiation (through both EB formation and monolayer-based differentiation in the presence of FBS, but absence of LIF) Nanog-GFP expression decreased in the same direction, yet slightly earlier than γH2AX, in both ESCs and iPSCs (supporting information Fig. S3). As already observed by our group , iPSCs retained Nanog expression for longer period following LIF withdrawal than ESCs. Altogether, these results, in accordance with other published data , demonstrate that high γH2AX basal levels are a specific feature of pluripotent mouse ESCs and iPSCs and decrease during cell differentiation.
Small Molecules that Enhance ESC Self-Renewal Induce Increases in γH2AX Levels
The decrease in γH2AX levels observed during ESC and iPSC differentiation reinforced our hypothesis that there may be a correlation between high γH2AX expression and ESC self-renewal. This prompted us to investigate the effect of inhibiting key signaling pathways involved in regulation of self-renewal on γH2AX levels. Two selective chemical inhibitors were used, both have been reported to enhance ESC self-renewal: 1m, a GSK3 inhibitor , and PD0325901 (PD), a MEK inhibitor . After both stimuli, we observed a slight, but not significant, decrease in ESC proliferation (data not shown). We then performed clonal assays to assess self-renewal based on alkaline phosphatase expression, an enzyme that is expressed by pluripotent ESCs and rapidly downregulated upon differentiation. ESCs were plated in GMEM medium supplemented with FBS and 1,000 units/ml LIF (a condition that maintains both self-renewing and differentiating cells without favoring any specific population) either in the presence or in the absence of the two inhibitors. The results demonstrate that both 1m and PD treatments significantly enhance ESC self-renewal (Fig. 5A). We then analyzed γH2AX levels after either 1m or PD treatments and observed that both treatments were accompanied by a significant increase of γH2AX levels (Fig. 5B). These data demonstrate a correlation between self-renewal enhancement and increased levels of γH2AX.
H2AX Phosphorylation Is Necessary to Maintain Optimal ESC Self-Renewal
To further corroborate a potential role for H2AX in controlling self-renewal of pluripotent stem cells, we analyzed the influence of H2AX deficiency on ESC self-renewal capacity. We performed self-renewal assays using WT TC1 ESCs and TC1-derived H2AX−/− ESCs (kindly provided by F.W. Alt ). The morphology of the H2AX−/− ESCs already suggested a more differentiated status (Fig. 6A), and these cells grew more slowly than their WT parental TC1 cells, with doubling times of 30.1 ± 4.5 hours and 15.8 ± 1.7 hours, respectively (mean ± SEM) (Fig. 6B). Cell cycle profiles were similar in the two cell lines (data not shown). Analysis of self-renewal revealed a significant difference between TC1 and H2AX−/− cells, with only a small number of H2AX−/− colonies classified as self-renewing (Fig. 6C, left panel). Following treatment with both 1m and PD, H2AX−/− cells significantly enhance their self-renewal, showing that they can respond to signaling pathways normally active in ESCs (Fig. 6C, right panel). We then examined protein expression levels of Nanog and Oct4, two transcription factors important in self-renewal regulation, and found that they were both significantly lower in H2AX−/− cells compared to TC1 cells (supporting information Fig. S4). To examine whether the phosphorylation of Ser139 on H2AX is specifically required for the maintenance of ESC self-renewal, we reconstituted H2AX−/− ESCs by retroviral transduction with cDNAs encoding either HA-tagged H2AX (WT hereafter) or HA-tagged S139A H2AX, where Ser139 was replaced with an alanine residue (S139A) (supporting information Fig. S5) (plasmids kindly provided by M.S.Y. Huen ). H2AX−/− ESCs reconstituted with WT H2AX recovered their morphological and proliferation defects, showing round and compact colonies (Fig. 6D) and a doubling time of 15.1 ± 2.1 hours (mean ± SEM) (Fig. 6E). In contrast, the morphology and proliferation rate of H2AX−/− ESCs reconstituted with either S139A or a GFP control vector were as defective as H2AX−/− cells, with flattened and spiky colonies (Fig. 6D) and a doubling time of 29.4 ± 4.5 hours and 27.4 ± 0.9 hours (mean ± SEM), respectively (Fig. 6E). Finally, self-renewal assays showed that H2AX−/− ESCs reconstituted with WT H2AX recovered their self-renewal capacity, whereas those reconstituted with S139A did not (Fig. 6F). In these cells, self-renewal levels were similar to H2AX−/− cells transduced with the control vector (Fig. 6F). These results demonstrate that the phosphorylation of H2AX at Ser139 is required for its ability to promote the maintenance of ESC self-renewal.
As an epigenetic regulatory system, the histone code is thought to play important roles in regulating different cellular events, including pluripotency and differentiation . In this study, we demonstrate, for the first time, that γH2AX epigenetic modification contributes to sustaining the self-renewal ability of ESCs and iPSCs. First, we confirmed that γH2AX levels decrease during ESC differentiation  and demonstrate, for the first time, that iPSCs behave in the same way. Second, γH2AX levels increase when ESCs are treated with small molecule inhibitors that enhance self-renewal. Third, we show that an H2AX−/− ESC line has a reduced capacity for self-renewal, and fourth, we demonstrate that the self-renewal ability of H2AX−/− ESCs can be restored through reconstitution with WT H2AX but not with a mutant form of H2AX in which the S139 phosphorylation site is abolished. Our view is that H2AX may be involved in multiple pathways in stem cells, some ascribed to basal levels of γH2AX and others to induced levels of γH2AX (Fig. 7). Based on our data, mouse ESCs and iPSCs appear to require a basal level of γH2AX to sustain optimal levels of self-renewal (and proliferation), while it is clear that the basal H2AX phosphorylation is not involved in the DSB response.
Emerging data point to a key role for epigenetic mechanisms, including nuclear architecture, chromatin structure, chromatin dynamics, and histone modifications, in the molecular mechanisms regulating self-renewal, maintenance of pluripotency, and lineage specification [47, 48]. Chromatin in pluripotent stem cells is increasingly being recognized as “open” when compared with chromatin in somatic cells, implying that its overall structure is less condensed and that the ratio between euchromatin and heterochromatin is higher than in differentiating cells [19, 49]. This pattern of chromatin organization was recently found in vivo: cells in the ICM of the mouse blastocyst at day 3.5, which are the source of ESCs, share the same open chromatin conformation as cultured ESCs . As differentiation progresses, ESCs accumulate highly condensed, transcriptionally inactive heterochromatin regions [51, 52]. Consistent with changes in global genome activity, changes in histone-modification patterns accompany ESC differentiation . Our demonstration that high γH2AX levels sustain self-renewal of ESCs and iPSCs suggests this minor histone modification as another important epigenetic element involved in maintaining the chromatin architecture that contributes to the unique properties of pluripotent stem cells. Specific roles of H2AX in ESCs have already been suggested by the presence of high incorporation levels of this histone variant into the nucleosomes of ESCs in comparison to other cell types. H2AX histone is incorporated into chromatin at a significant level during the early preimplantation stage of development, after fertilization [9, 46, 53, 54]. This suggested a specific role for H2AX in the global changes of chromatin structure that occur after fertilization, with possible involvement in establishing and/or maintaining totipotency and the high plasticity necessary for genome reprogramming.
On the other hand, human ESC lines, as well as human iPSC lines, have been shown to have very low basal levels of H2AX phosphorylation (, and our data not shown). This could represent another of the several differences described between mouse and human pluripotent stem cells. These differences might be ascribed to the fact that human ESC lines correspond to a slightly different (later) stage of development, as, for some aspects, cells from mouse epiblast are more similar to the human ESCs; alternatively, they might reflect biological differences. For example, mouse but not human ESCs are deficient in rejoining of IR-induced DSBs .
In this report, we provide evidence for a redundant role of different kinases (ATM, DNA-PK, and possibly ATR) belonging to the phosphatidylinositol 3-kinase-related kinase (PIKK) family in the maintenance of the high basal levels of H2AX phosphorylation observed in mESCs. This redundant role had been previously suggested by Banath et al., although with a different experimental system .
No proteins typically associated with the DNA damage response colocalized with the γH2AX foci in ESCs, suggesting that the high basal numbers of γH2AX foci do not correlate with pre-existing DNA breaks. Chuykin et al. also concluded that endogenous γH2AX foci in ESCs did not constitute part of the DNA repair system, based on an apparent lack of phospho-ATM foci . In accordance, neutral comet assay did not reveal any evidence for the presence of DSBs in untreated ESCs . Moreover, we demonstrated that, although ESCs proliferate very rapidly, the γH2AX high basal level cannot be entirely ascribed to the increased proportion of cells in S-phase, as ESCs were positive for γH2AX also when in G1 phase. In the literature, a correlation between increased γH2AX levels and decreased ESC proliferation has been reported , further suggesting that in some conditions, there may not be a direct link between cell proliferation and γH2AX enhancement. However, we cannot formally rule out a link between high γH2AX and rapid proliferation, for example, the G2/M checkpoint in ESCs could be less sensitive, allowing cells with DNA damage to progress, giving rise to γH2AX-positive cells in G1. Detailed further studies are required to distinguish between these possibilities.
Banuelos et al. observed a reduced plating rate of H2AX−/− ESCs but similar cell cycle distribution and similar doubling time (12–14 hours) in WT and H2AX−/− cells . Our data confirm the similarity in cell cycle distribution and the reduced plating rate in H2AX−/− ESCs (data not shown), but they also clearly indicate a delayed doubling time (30 hours) in H2AX−/− cells. We believe that this discrepancy can be ascribed to the different cell culture media used: Banuelos et al. cultured ESCs in complete DMEM supplemented with 15% FBS, whereas we used the more specific ESC medium, complete KO-DMEM supplemented with 15% KO-SR. Our data are in accordance with other studies on H2AX−/− models, demonstrating that H2AX−/− mice were growth retarded and H2AX−/− mouse embryonic fibroblasts proliferated poorly in vitro  and a delayed doubling time in H2AX−/− ESCs had already been noticed (F.W. Alt, personal communication).
H2AX−/− cells show a greatly reduced ability for self-renewal and low Nanog and Oct4 protein levels. We show here that H2AX−/− cells can recover their self-renewal defect when reconstituted with WT H2AX, demonstrating the importance of H2AX in this ESC characteristic. Treatment with two selective chemical inhibitors acting on GSK3 and MEK [14, 30], both reported to enhance ESC self-renewal, was also able to recover self-renewal of H2AX−/− cells, showing that they can respond to signaling pathways normally active in ESCs. Therefore, these pathways could potentially compensate for the lack of H2AX and there is precedence in the literature for GSK3 and MEK inhibition to be able to support self-renewal of ESCs lacking, for example, Stat3 .
Recently, Andang et al. described a new pathway, active in ESCs and mediated by the GABA-A receptor, which they suggested regulates ESC proliferation . Activation of GABA-A receptors led to accumulation of stem cells in S-phase, causing a decrease in cell proliferation and an increase in γH2AX levels in nuclear foci of ESCs. Abdelalim and Tooyama convincingly showed that this pathway is under the control of brain natriuretic peptide (BNP) . Clearly, the focus of these two papers was the role of H2AX on control of stem cell proliferation via an S-phase effect. No impact on self-renewal/pluripotency was reported, and the correlation between increased γH2AX levels and decreased ESC proliferation they report contrasts with our observation of a delayed doubling time in H2AX−/− ESCs. However, both papers referred to a role of induced γH2AX (via GABA-A and/or BNP signaling) rather than of basal γH2AX, which was instead the focus of our work.
Here we demonstrate, for the first time, that high basal levels of phosphorylation of the H2AX histone contribute to the maintenance of self-renewal of mouse embryonic and iPSCs. Our findings suggest a novel function of H2AX that expands the knowledge of this histone variant beyond its role in DNA damage and into a new specialized biological function in pluripotent stem cells.
We thank F.W. Alt (Children's Hospital at Boston) for careful critical reading and suggestions and for the TC1 and H2AX−/− ESCs; M.S.Y. Huen (The University of Hong Kong) for the HA-WT H2AX and HA-S139A vectors; I. Chambers (Institute for Stem Cell Research, University of Edinburgh, Scotland) for the Nanog-GFP ESC line and RIKEN cell bank for provision of the iPSC line; H.K. Bone (The University of Bath, U.K.) for confocal image analyses. This work was supported by the British Council Italian Partnership program. LO and BK were supported by the Marie Curie Early-Stage Training program MEST-CT-2005-019822. YSR was the recipient of an MRC Capacity Building Studentship. Additional financial support for the work was provided by grants to MJW and CG.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.