REGENERATIVE MEDICINE
HIF-2a Suppresses p53 to Enhance the Stemness and Regenerative
Potential of Human Embryonic Stem Cells
BIKUL DAS,a,b,c,d
REZA BAYAT-MOKHTARI,c,d
MICKY TSUI,d
SHAMIM LOTFI,d
RIKA TSUCHIDA,e
DEAN W. FELSHER,a
HERMAN YEGER
c,d
a
Division of Oncology, Departments of Medicine and Pathology, Stanford University School of Medicine,
Stanford, California, USA; b
Division of Hematology and Oncology, Department of Paediatrics, c
Developmental
Biology and Stem Cell Program, and d
Department of Paediatric Laboratory Medicine, The Hospital for Sick
Children, University of Toronto, Toronto, Ontario, Canada; e
Department of Molecular Oncology, Tokyo Medical
and Dental University, Tokyo, Japan
Key Words. Embryonic stem cells • altruism • Hypoxia • Neoplastic stem cell biology • Reprogramming • Self-renewal • Oxidative stress
ABSTRACT
Human embryonic stem cells (hESCs) have been reported
to exert cytoprotective activity in the area of tissue
injury. However, hypoxia/oxidative stress prevailing in
the area of injury could activate p53, leading to death
and differentiation of hESCs. Here we report that when
exposed to hypoxia/oxidative stress, a small fraction of
hESCs, namely the SSEA31/ABCG21 fraction undergoes
a transient state of reprogramming to a low p53 and high
hypoxia inducible factor (HIF)-2a state of transcriptional
activity. This state can be sustained for a period of 2
weeks and is associated with enhanced transcriptional activity
of Oct-4 and Nanog, concomitant with high teratomagenic
potential. Conditioned medium obtained from
the post-hypoxia SSEA31/ABCG21 hESCs showed cytoprotection
both in vitro and in vivo. We termed this phenotype
as the ‘‘enhanced stemness’’ state. We then
demonstrated that the underlying molecular mechanism
of this transient phenotype of enhanced stemness involved
high Bcl-2, fibroblast growth factor (FGF)-2, and MDM2
expression and an altered state of the p53/MDM2 oscillation
system. Specific silencing of HIF-2a and p53 resisted
the reprogramming of SSEA31/ABCG21 to the enhanced
stemness phenotype. Thus, our studies have uncovered a
unique transient reprogramming activity in hESCs, the
enhanced stemness reprogramming where a highly
cytoprotective and undifferentiated state is achieved by
transiently suppressing p53 activity. We suggest that this
transient reprogramming is a form of stem cell altruism
that benefits the surrounding tissues during the process of
tissue regeneration. STEM CELLS 2012;30:1685–1695
Disclosure of potential conflicts of interest is found at the end of this article.
INTRODUCTION
Numerous studies have revealed that although the majority of
transplanted stem cells die, and do not engraft, stem cells
secrete growth factors and antioxidants that likely contribute
to the repair and regeneration of the inflamed and damaged
tissues [1]. This view is consistent with the findings that undifferentiated
stem cells are capable of secreting high levels
of antioxidants and growth factors [2]. However, the site of
injury/inflammation is particularly characterized by hypoxia
and production of reactive oxygen species (ROS). The ROS
could induce DNA damage, and subsequently trigger p53 activation
in stem cells, leading to differentiation and apoptosis
[3]. Hence, it is suggestive that stem cells may have evolved
specific mechanisms to maintain an undifferentiated state in
the area of injury to serve their cytoprotective activity. Therefore,
we posited that stem cells might have acquired molecular
mechanisms to suppress p53 and remain functionally undifferentiated
during hypoxia and oxidative stress in order to
be able to conduct their cytoprotective activity.
One of the potential ways that stem cells could suppress
p53 is by activating HIF-2a. Hypoxia upregulates the key
transcription factors HIF-1a and HIF-2a with numerous downstream
targets that play critical roles in survival, angiogenesis,
metabolism, and other important functions [4]. HIF-1a stabilizes
p53, and therefore enhances p53 activation [5], whereas
HIF-2a is shown to suppress p53 activity by reducing ROS
generation [6]. HIF-2a also upregulates several genes
involved in antioxidant mechanisms including SOD, catalase,
GSPx [7], and ABCG2 [8]. ABCG2 is an efflux pump [9–12],
which is found to secrete glutathione (GSH) [13], a primary
antioxidant and component of the GSH/glutathione-s-transferase
(GST) detoxification system [14]. HIF-2a also enhances
the secretion of growth factors having cytoprotective and
regenerative abilities, for example, vascular endothelial
Author contributions: B.D.: conceived the idea, initiated the study, designed the experimental strategies, and wrote the manuscript; B.D.,
R.B.M., M.T., and S.L.: performed the in vitro and in vivo experiments; B.D., R.T., D.W.F., and H.Y.: analyzed the data and
participated in manuscript revision.
Correspondence: Bikul Das, M.B.B.S., Ph.D., Division of Oncology, Department of Medicine, Stanford University School of Medicine,
Stanford, California, USA. Telephone: (650) 724-6467; Fax: (650) 725-1420; e-mail: bikuldas@stanford.edu Received November 29,
2011; accepted for publication May 4, 2012; first published online in STEM CELLS EXPRESS June 11, 2012; available online without
subscription through the open access option. VC AlphaMed Press 1066-5099/2012/$30.00/0 doi: 10.1002/stem.1142
STEM CELLS 2012;30:1685–1695 www.StemCells.com
growth factor (VEGF) and fibroblast growth factor (FGF)-2
[15]. Thus, one potential mechanism of p53 suppression
would be to enhance HIF-2a expression, which could not
only contribute to protect the stem cells from p53-induced
differentiation or death but also enhance their cytoprotective
activity. Therefore, we sought evidence that stem cells may
suppress p53 activity during hypoxia and oxidative stress by
upregulating HIF-2a in order to exert their cytoprotective and
regenerative activities. We also wanted to determine whether
the same mechanism might also be involved in the potential
transformation of stem cells to cancer stem cells as has been
recently documented [16, 17]
To investigate the above possibilities, we used human embryonic
stem cells (hESCs) as a model cell type. There are
several advantages in using ESCs to test our hypothesis. First,
p53 is highly active in hESCs and second, it directly inhibits
Nanog, one of the core regulators involved in maintaining the
stemness state [18–20]. Third, HIF-2a is active in ESCs [15].
We exposed hESCs to an extreme environment of hypoxia
and oxidative stress and applied a novel approach to isolate a
surviving, undifferentiated hESC fraction having low p53 and
high HIF-2a activity. We were successful in isolating a small
fraction of hESCs, the SSEA3þ/ABCG2þ fraction, which
exhibited high HIF-2a and low p53 activity as well as
enhanced cytoprotective activity. While investigating the undifferentiated
state, we serendipitously discovered that the
SSEA3þ/ABCG2þ cells undergo a transient state of reprogramming
to an even higher state of stemness, here termed the
‘‘enhanced stemness’’ state, having very high Nanog expression,
high antioxidant secretion, and high cytoprotective activity
following exposure to hypoxia and oxidative stress.
MATERIALS AND METHODS
Reagents and Chemicals
All the chemicals were obtained from Sigma-Aldrich (St. Louis,
MO, www.sigmaaldrich.com) unless otherwise noted. The HIF-
1a/2a inhibitor FM19G11 [21] was obtained from Millipore (Billerica,
MA, http://www.millipore.com [#400089-10MG]), Pifithrina
from ChemBridge Corporation (San Diego, CA, www.chem
bridge.com), caspase-3 substrate, N-acetyl-DEVD-7-amino-4methylcoumarin
peptide from Enzo Life Sciences (Farmingdale,
NY, www.enzolifesciences.com) [22] and recombinant FGF-2
and Matrigel from R&D Systems (Minneapolis, MN,
www.rndsystems.com). All the culture media reagents were
obtained from Invitrogen Technologies (CA, Invitrogen Life
Technologies, Grand Island, NY, www.invitrogen.com) until otherwise
noted.
hESC Culture and Induction of Hypoxia
Due permissions were obtained to culture hESCs both at Department
of Medicine, Stanford University (Stem Cell Research
Oversight protocol #285, director: BD) and at Department of
Laboratory Medicine and Pathobiology, Hospital for Sick Children.
BGO1 and H9 hESCs (WiCell Inc., Madison, WI, www.wi
cell.org) were expanded on Mitomycin-C (10 lg/ml; Sigma)-inactivated
mouse embryonic fibroblast (MEF) feeders, cultured in
knockout Dulbecco’s modified Eagle’s medium/F12 with
20% knockout serum replacement (KOSR, #12660012, and
#10828-028; Invitrogen Life Technologies, Grand Island, NY),
and 4 ng/ml FGF-2 (#PHG0024, Gibco, Life Technologies,
Grand Island, NY, www.invitrogen.com) as described elsewhere
[23, 24]. The expanded hESCs were then grown on feeder-free
Matrigel (##354230, BD Biosciences, San Jose, CA, www.bdbio
sciences.com)-coated dishes with MEF-derived conditioned media
(CM), supplemented with 8 ng/ml FGF-2 (complete hESC
media), and under standard conditions (37
C and 5% CO2 and
saturated humidity) as previously described [25] and used to perform
various experiments described in the text. To make singlecell
suspension for flow cytometry and to perform plating efficiency
experiments, the hESCs were dispersed with TrypLe select
(#12563-011, Invitrogen) for 3 minutes and then washed twice
with the complete hESC media supplemented with 10 lM Rhoassociated
protein kinase inhibitor Y-27632 (Calbiochem, Merck,
Rockland, MA, www.emd.us). The hESCs were then suspended
in the complete media plus Rho-associated protein kinase (ROCK)
inhibitor to maintain viability during flow cytometry sorting [26].
For the hypoxia experiments, day-2 hESCs culture having 30%–
40% confluence was used and hypoxia was achieved as described
previously [22, 23]. Briefly, hypoxic condition (<0.1% O2) was
established in a sealed chamber using the BBL GasPak plus
anaerobic system envelopes with a palladium catalyst (Becton
Dickinson, Cockeysville, MD, www.bd.com/ds). Immediately after
hypoxia, the cells were incubated in tissue culture incubator at
37
C and 5% CO2. Medium was changed only 24 hours after
hypoxia so that the cultured cells were exposed to ROS generated
immediately during the reoxygenation process [27].
Flow Cytometry
Flow cytometry was performed using FACScan and FACSAria II
(BD Bioscience) as described [23, 28]. Surface staining and flow
cytometry sorting was done as described elsewhere [23]. Briefly,
the single-cell suspensions of hESCs were incubated with primary
antibodies (Supporting Information Table 1) in staining media
(phosphate buffered saline (PBS) with 2% KOSR, Invitrogen), on
ice, for 15 minutes, washed thrice in staining medium, and then
incubated with Alexafluor secondary antibody for 30 minutes. After
washing, the stained cells were suspended in complete culture
media and kept at 4
C.
Immunostaining
Cells were fixed in 4% paraformaldehyde (Electron Microscopy
Sciences, Hatfield, PA, www.emsdiasum.com) for 10 minutes on
ice, permeabilized, washed thrice in PBS, and incubated overnight
in staining buffer (4% bovine serum albumin (BSA) þ
0.1% saponin) with primary antibodies against HIF-2a and p53
(Supporting Information Table 1). They were then washed and
incubated with secondary Alexa 488 or 594 antibodies for 1 hour
at RT. For immunostaining of teratoma tissues, frozen sections
were fixed with 4% paraformaldehyde, permeabilized with 0.1%
saponin, and subjected to immunostaining as described elsewhere
[23].
Cell Viability and Apoptosis Measurement
The hESCs were grown onto Matrigel-coated dishes in complete
hESCs media. At day 2 after plating, cells were exposed to hypoxia
for 24 hours and then reoxygenation for next 4 days. On
day-4 post-hypoxia, the viability was determined by Trypan blue
assay [22]. A portion of the cells was subjected to flow cytometry
sorting, and then the sorted population was subjected to caspase-3
activity assay as previously described [22].
ELISA Assay
The cell lysates were subjected to ELISA assay as previously
described [22]. Information about various ELISA kits used in the
study is given in Supporting Information Table 1.
InCell Western Assay
The assay was performed as per manufacturer instructions (Licor
Bioscience, NE, Lincoln, NE, www.licor.com). The hESCs (5 Â
103
per well) were plated in sterile 96-well plates (BG Falcon,
#353075) and kept at RT for 30 minutes before transfer to the incubator
to reduce excessive settling at the well edges. The plates
were incubated overnight at 37
C and then fixed with 3.7% formaldehyde
at RT for 10 minutes. After fixation, the plates were
washed with PBS, then primary antibodies were added at 1:300
1686 Enhanced Stemness Reprogramming
concentration in 2% BSA þ 0.1% saponin, and incubated overnight
at 4
C. After primary antibody incubations, plates were
washed thrice (5 minutes each) with 100 ll per well PBS-T (PBS
with Tween 20) at RT. Then, secondary antibodies conjugated to
IRDye 800CW (goat-anti-rabbit-IgG (Licor Biosciences, NE) or
goat-anti-mouse-IgG (Licor Biosciences, NE) were used at
1:1,000 dilution. The nuclear staining dye DRAQ5 (Biostatus
Ltd., Leicestershire, U.K., www.biostatus.com) combined with
Sapphire700 (Li-COR) at 1:10,000 and 1:1000, respectively, were
also added. Plates were then incubated with 50 ll per well secondary
antibody solutions for 90 minutes at R/T in the dark.
Background control wells were prepared by omitting primary
antibodies (i.e., secondary only). After secondary antibody incubations,
plates were washed thrice times (10 minutes each) with
PBS-T at R/T in the dark, and then plates were dried in air before
scanning. Plates were scanned and analyzed using an Odyssey IR
scanner using Odyssey imaging software 3.0. Scan settings used
were medium or high image quality, 169 lm resolution, intensity
5.0 for the both the 700-channel and 800-channel with an offset
of 4.0 mm. For signal quantification, antibody signals were analyzed
as the average 800-channel integrated intensities from
triplicate wells normalized to the 700-channel signal integrated
intensity to correct for well-to-well variations in cell number.
Results are expressed as fold change (means 6 standard errors of
the mean) compared to noninfected controls.
Measurement of Transcription Factor Activity
The transcriptional activity of HIFs and p53 were measured by a
transcription factor (TF) filter play assay system developed by
Signosis, Sunnyvale, CA, www.signosisinc.com (Supporting Information
Table 1). The assay was performed as per manufacturer
instructions. Briefly, the nuclear extracts of hESCs were mixed
with a specific biotin-labeled TF DNA binding sequence to obtain
TF-DNA complexes. Then, the TF-DNA complexes were retained
in a filter plate and hybridization was achieved. The captured
DNA probe was quantified by streptavidin-horseradish peroxidase
(HRP) method on a microplate luminometer. The Oct-4 and
Nanog transcriptional activities were measured by Cignal Lenti
Nanog reporter (Luciferase) kit (SABioscience, Valencia, CA,
www.sabio.sciences.com) as per manufacturer instructions.
Measurement of Oxidative Stress
The level of ROS was measured by DCFH-DA assay as described
[23]. The oxidative DNA damage was measured by a 8-hydroxy-
2-deoxy Guanosine (8-OH-dG) Enzyme Immuno Assay kit
(#589320; Cayman Chemical, MI, Ann Arbor, MI, www.cayman
chem.com) as per manufacturer’s instructions.
GSH Measurement
The GSH was measured using the glutathione assay kit (#GT10,
Oxford Biomedical Research, MI, www.oxfordbiomed.com) as
described previously [14].
siRNA Gene Silencing
We performed small interfering RNA (siRNA) gene silencing by
Accell siRNA (Thermoscientific Dharmacon, Lafayette, CO,
USA, www.dharmacon.com). The following Accell siRNAs were
used: the human HIF-2a (A-004814-17-0005), HIF-1a (A-
004018-24-0005), and p53 (A-003329-22-0005). The hESCs (5 Â
103
per well in 96-well plate) were cultured in the completed
hESC media, and 1 lM Accell siRNA was delivered to each well
as per manufacturer instructions. After incubation for 72 hours at
37
C, the media were replaced with fresh media, and the cells
were exposed to hypoxia. The gene knockdown was confirmed
by the real-time quantitative reverse transcription-polymerase
chain reaction (QPCR) and ELISA.
Real-Time QPCR
The QPCR assay was performed using Taqman Gene Expression
Assays (Applied Biosystems, Foster City, CA, www.appliedbio
systems.com) as previously described [22, 23]. RNA was quantified
with the Ct method, and glyceraldehyde 3-phosphate dehydrogenase
(GAPDH) was used as a housekeeping gene [22, 23].
Taqman primers used were HIF-1a: Hs00153153_m1, HIF-2a:
Hs01026149_m1, and p53: Hs00996818_m1. The Taqman primers
for Oct-4, Nanog, ABCG2, and GAPDH are as previously
described [23, 28].
Cytoprotective Assays
For the in vitro cytoprotective assay, primary rat cardiomycotes
were obtained from Cell Applications Inc. (San Diego, CA,
www.cellapplications.com) [29] and grown in the rat cardiomyocyte
growth medium as per manufacturer’s instructions. SHSY5Y
neuroblastoma cells were obtained from ATCC. The cells
were exposed to 24 hours of hypoxia followed by reoxygenation
for next 24 hours, and the viability was measured by Trypan blue
assay. The cell culture medium was replaced with hESCs CM 2
hours before exposing to hypoxia. The in vivo bone marrow
(BM) stem cell cytoprotection experiments were carried out in
BALB/c mice as described previously [14]. Briefly, 10–12-weekold
BALB/c mice (Charles River Laboratories International, MA)
were injected with hESCs CM i.p. 3 hours before the i.p. injection
of high-dose 120 mg/kg carboplatin. On the day 5 of injection,
mice were sacrificed, BMs were collected, and both the hematopoietic
stem cell (HSC) and mesenchymal stem cell (MSC)
fractions were isolated by using the easy-sep magnetic sorting
kits (#19756 for HSC and #19771 for MSC; Stem Cell Technologies,
BC, www.stemcell.com). The isolated cells were then subjected
to colony forming unit (CFU) assay for HSCs and MSCs
as described elsewhere [14].
Limiting Dilution Assay for Teratoma Formation
The in vivo teratoma study was performed by injecting hESCs
mixed with 1:2 Matrigel to non-obese diabetic severe combined
immunodeficient mice subcutaneously as described [23]. The number
of hESCs injected s.c. to mice were as follows: 2 Â 106
(5), 1
 105
(10), 1 Â 104
(10), 1 Â 103
(10), 1 Â 101
(10), and the number
within parentheses representing number of animal used for the
experiments. Mice were observed for 1–6 months for the teratoma
formation. To confirm teratoma formation, the xenografts were
fixed in formalin and subjected to H&E study [23]. In a separate
experiment, a portion of Matrigel plugs were removed 3 weeks
after implantation to mice as described [21] and frozen fixed to
perform immunostaining [23] for Nanog expressing cells by using
a anti-Nanog antibody (Supporting Information Table 1) [28].
Statistical Analysis
The statistical analysis was performed using GraphPad Prism 4.0
(GraphPad Software Inc., CA, www.graphpad.com) using either Student
t test or one-way analysis of variance with Newman-Keul post
hoc test. The flow cytometry data were analyzed by flowjo (Tree
Star, Inc., OR) [23]. The analysis of teratomagenic efficiency was
done by extreme limiting dilution software [30] (http://bioinf.wehi.e-
du.au/software/elda)
RESULTS
P53 and HIF-2a Are Upregulated in hESCs
Following Hypoxia/Oxidative Stress
To investigate whether exposure to hypoxia followed by reoxygenation
(i.e., oxidative stress) activates p53 as well as HIF-2a
in ESCs, BGO1, human ESCs were exposed to extreme hypoxia
(pO2 $ 0.1% for 24 hours) followed by reoxygenation
(Supporting Information Fig. 1a) [22], and changes in ROS,
p53, and HIFs were measured. We found that BGO1 cells
underwent marked morphological changes of differentiation
and cell death (Fig. 1A) with an accompanying fourfold
Das, Bayat-Mokhtari, Tsui et al. 1687
www.StemCells.com
increase in ROS level (Fig. 1B). While the p53 protein level
was increased fourfold by day 4 (Fig. 1B), HIF-2a protein was
already rapidly elevated threefold after 1 day post-hypoxia (not
shown) and remained sustained during the day-4 reoxygenation
period (Fig. 1B). In contrast, HIF-1a protein expression
increased immediately during hypoxia (day 0) and then showed
a rapid reduction in expression by day-4 post-hypoxia (Fig.
1B). These results suggest that activation of both p53 and HIF-
2a is sustained on day 4 during the reoxygenation phase following
hypoxia. In contrast, HIF-1a decreased during the reoxygenation
phase, which is consistent with the previous findings
that HIF-1a mRNA is unstable whereas HIF-2a mRNA could
be sustained for a prolonged period after hypoxia [6].
The High HIF-2a and Low p53 Cells Are Enriched
in the SSEA31/ABCG21 Fraction
We next considered that the high HIF-2a cell fraction might
be enriched within the undifferentiated ESC fraction and
would express ABCG2, since this is one of the target genes
of HIF-2a [8]. To investigate this possibility, we first evaluated
whether the high HIF-2a expressing hESCs on day-0 and day-
4 post-hypoxia treatment overlapped with ABCG2þ expression.
We found that most of the HIF-2a expressing cells were
positive for ABCG2 (Supporting Information Fig. 1b). Second,
we sorted by flow cytometry a surviving (propidium iodide
negative) undifferentiated SSEA3þ fraction expressing
ABCG2. The flow sorting showed that only 12% 6 3% (p <
.05; n ¼ 4) SSEA3þ cells survived by day-4 post-hypoxia/
reoxygenation (Fig. 1C), which was a eightfold reduction in
the SSEA3þ fraction (p < .05; n ¼ 4). We then found that the
ABCG2þ/SSEA3þ fraction (Fig. 1C) increased by fourfold
(Fig. 1C; 4.2 6 1.1, p < .05; n ¼ 4) by day-4 post-hypoxia/
reoxygenation.
Third, we compared the HIF-2a protein levels as well as
transcriptional activities between the SSEA3þ/ABCG2þ and
SSEA3þ/ABCG2À fractions using several assays. Immunofluorescence
labeling revealed a strong expression of HIF-2a
Figure 1. SSEA3þ/ABCG2þ fraction survives and exhibits high HIF-2a and low p53 activity during exposure to hypoxia/reoxygenation. (A):
Morphological evidence of BGO1 human embryonic stem cell differentiation and death (image to the right) following exposure to extreme hypoxia
followed by 4 days of reoxygenation. The hypoxia- and reoxygenation-treated cells showed elongated neurite processes including fibroblastic-like
features as well as evidence of cell fragmentation. (B): The fold change in reactive oxygen species (DCFH-DA), p53, and HIF-2a in the
BGO1 cells is shown in (A), following day-4 post-hypoxia/reoxygenation. (C): Flow cytometry showing expansion (a vs. b compartments) of the
SSEA3þ/ABCG2þ fraction on the day-4 post-hypoxia/reoxygenation. Inset shows the profile of isotype control. (D): Immunofluorescence labeling
of SSEA3þ/ABCG2À and SSEA3þ/ABCG2þ fractions on day-4 post-hypoxia/reoxygenation showing a high HIF-2a and low p53 signal in
the SSEA3þ/ABCG2þ fraction (DAPI, nuclear stain). (E): The corresponding fold change in DCFH-DA, p53, and HIFs in the SSEA3þ/
ABCG2þ versus SSEA3þ/ABCG2À fractions on the day-4 post-hypoxia/reoxygenation. (F): Protein levels of HIF-2a and p53 in the SSEA3þ/
ABCG2þ fraction were measured by ELISA. Note high HIF-2a versus low p53 until day 20. *, p < .05; n ¼ 4. Scale bar ¼ 40 lm (A); 10 lm
(D). Abbreviations: DAPI, 40
,6-diamidino-2-phenylindole; DCFH-DA, 20
,70
-dichlorfluorescein-diacetate; HIF, hypoxia inducible factor.
1688 Enhanced Stemness Reprogramming
in the ABCG2þ fraction versus the ABCG2À fraction (Fig.
1D), which was confirmed at the protein level as measured by
ELISA (Fig. 1E), and at gene expression and the transcriptional
activity level (Supporting Information Fig. 2a, 2b). The
HIF-1a expression was not sustained in the day-4 post-hypoxia/reoxygenation
ABCG2þ fraction (Fig. 1E; Supporting
Information Fig. 2a).
Taken together, our results suggest that HIF-2a expression
was enriched in the SSEA3þ/ABCG2þ fraction as compared
to the SSEA3þ/ABCG2À fraction.
Next we compared the protein level and transcriptional
activity of p53 in SSEA3þ/ABCG2þ versus SSEA3þ/
ABCG2À. Immunofluorescence labeling indicated a low
expression of p53 in the ABCG2þ fraction versus the
ABCG2À fraction (Fig. 1D), which was also confirmed quantitatively
by ELISA (Fig. 1E), and by transcriptional activity
as measured with the transcriptional filter assay (Supporting
Information Fig. 2b). These results suggest that the SSEA3þ/
ABCG2þ fraction expresses a low level of p53 as compared
to the SSEA3þ/ABCG2À fraction.
We reasoned that a high HIF-2a and low p53 activity could
be associated with higher survival and less DNA damage in the
SSEA3þ/ABCG2þ fraction versus the SSEA3þ/ABCG2À
fraction. Hence we measured levels of ROS, 8-OHdG, and caspase-3
and found these to be four- to fivefold higher in the
SSEA3þ/ABCG2À versus the SSEA3þ/ABCG2þ fraction
(Supporting Information Fig. 2c). The trypan blue viability
assay showed a 3.5-fold higher survival of SSEA3þ/ABCG2þ
versus SSEA3þ/ABCG2À (p < .05; n ¼ 4; data not shown).
Thus, we conclude that the SSEA3þ/ABCG2þ fraction was
characterized by higher survival and less DNA damage as
compared to the SSEA3þ/ABCG2À fraction.
We then investigated whether the high HIF-2a/low p53/
ABCG2þ phenotype generated through hypoxia/reoxygenation
hereon designated ABCG2þhox could persist in vitro for a prolonged
period of time. We found that this cell fraction actively
proliferated in the ESC medium for several weeks (three passages;
total of 25 days). During this period, p53 and HIF-2a levels
were measured by ELISA. The protein levels were compared
with day-0 post-hypoxia/reoxygenation ABCG2þ cell fraction.
We found that the post-hypoxia/reoxygenation ABCG2þhox
fraction showed a sustained phenotype of low p53 and high
HIF-2a expressions that lasted for 20 days (Fig. 1F).
Taken together, we conclude that hypoxia/reoxygenation
endows a subfraction of BGO1 ESCs with enhanced survival
and is associated with high HIF-2a and low p53 activities.
This high HIF-2a and low p53 phenotype could persist in
vitro for more than 2 weeks and was mainly confined to the
ABCG2þ fraction.
ABCG21Hox Cells Exhibit a Highly
Undifferentiated and Cytoprotective State
To investigate whether the ABCGþ2hox fraction maintained
an undifferentiated stemness state throughout the 3 week culture
period of low p53/high HIF-2a expression, we measured
the flow cytometry expression and transcriptional activity of
Oct-4 and Nanog in the ABCG2þhox fraction and compared
it with ABCG2þnox (designating the SSEA3þ/ABCG2þ
fraction maintained under normoxia and not exposed to hypoxia/reoxygenation).
First, we found that the levels of Oct-4
and Nanog protein expression measured by flow cytometry
were sustained in ABCG2þhox over the 2 week period (Fig.
2). Second, the transcriptional activity of Oct-4 and Nanog
were sustained three- to fourfold higher in ABCG2þhox versus
ABCG2þnox over the 2 week period (Fig. 3A). Thereafter
the transcriptional activity level returned to a baseline
level corresponding to the transcriptional activities of Oct-4
and Nanog in the ABCG2þnox fraction by day-20 of reoxygenation
(Fig. 3A), and similar to the period of high HIF-2a/
lowp53 state of ABCG2þhox as previously described in Figure
1F. This result suggested that the ABCG2þhox fraction,
having a sustained high HIF-2a activity, also sustained a
higher Nanog and Oct-4 transcriptional activity over the 2
week culture period.
Consequently, we speculated that the ABCG2 protein level
might also be sustained at a higher level in the ABCG2þhox
fraction for a correspondingly similar period. Previous studies
have shown that the ABCG2 efflux pump confers a cytoprotective
effect on ESCs by reducing intracellular ROS [9]. A recent
study reported that ABCG2 also secretes GSH, a potent cellular
antioxidant, into the extracellular space [13]. We, and
others, reported that GSH exerts a cytoprotective activity on
mammalian cells, including BM stem cells, against hypoxia/
reperfusion as well as cancer chemotherapy-induced toxicities
[14]. Hence, a corresponding increase in GSH secretion by the
ABCG2þhox fraction might confer cytoprotection on nearby
cells exposed to ROS-induced toxicity. We investigated these
possibilities as follows.
First, we measured the ABCG2 protein level in the
ABCG2þhox versus ABCG2þnox fractions by ELISA assay.
We found a sustained threefold increase of ABCG2 expression
in the ABCG2þhox versus the ABCGþ2nox fraction
(Fig. 3B). Second, we measured GSH secretion in the CM
obtained from the ABCG2þhox and ABCG2þnox fractions
with a GSH quantitative assay that is very sensitive for measurement
of GSH in stem cells [14]. We found a fivefold
increase in the GSH level in the CM obtained from the
ABCG2þhox fraction (Fig. 3B). We also noted that the level
of ABCG2 and GSH returned to the baseline level on day 20
corresponding to the return of the high-HIF-2a/low p53 state
to its baseline state as in the ABCG2þnox phenotype (Fig.
3B). Thus, it appears that the ABCG2þhox phenotype is associated
with high GSH secretion into the CM.
Third, we compared the in vitro cytoprotection potential
of the ABCG2þhox- versus ABCG2þnox-derived CM. CM
was collected from 1 Â 106
ABCG2þhox cells in 4 ml, on
day-8 post-hypoxia, and used to treat cardiomyocytes and
neuronal cells (SH-SY5Y cell line [31]) exposed to hypoxia/
reperfusion, and was compared to nonexposed medium as a
control. We then assessed the effect on survival and found a
10–12-fold increase in the survival of the cardiomyocytes and
neuronal cells when treated with the CM of ABCG2þhox
versus CM from the ABCG2þnox fraction (Fig. 3C). Thus
we demonstrated that the ABCG2þhox fraction CM exerted a
significant cytoprotective activity on differentiated cardiomyocytes
and neuronal cells, suggesting a mechanism whereby
the ABCG2þhox fraction is cytoprotective.
To obtain in vivo evidence of the cytoprotective activity
of the ABCG2þhox-derived CM, we used a murine model of
carboplatin-induced BM stem cell toxicity assay developed in
our laboratory [14]. In this assay, antioxidants were shown to
protect BM HSCs and MSCs from high-dose carboplatininduced
toxicity [14]. We pretreated carboplatin-exposed mice
with CM derived from ABCG2þhox compared to CM from
ABCG2þnox. The CM were collected from 1 Â 106
ABCG2þhox and ABCG2þnox cells in 2 ml, on day-8 posthypoxia,
and 1 ml was injected i.p. to mice. We then measured
both HSC and MSC survival using established assays
[14]. The CM was compared to N-acetyl-cysteine (NAC) a
potent antioxidant, the latter having been shown to cytoprotect
stem cells [32–34]. We found that the CM derived from the
ABCG2þhox cells showed a 17-fold cytoprotection of HSC
and 10-fold protection of MSCs (Fig. 3D). Interestingly, this
Das, Bayat-Mokhtari, Tsui et al. 1689
www.StemCells.com
cytoprotective activity in CM was roughly equivalent to
NAC. Thus, we conclude that the ABCG2þhox CM showed
potent cytoprotective activity in vivo against BM toxicity
induced by carboplatin, a commonly used anticancer agent in
clinics. Taken together, our observations above suggest that
the ABCG2þhox fraction attains a highly undifferentiated
state (henceforth known as enhanced stemness state) and
exhibits potent cytoprotective activity.
Considering the present interest in the potential use of
hESCs for tissue regeneration, the results obtained from the
BGO1 CM are provocative. To evaluate whether CM derived
from the ABCG2þhox fraction of other hESC lines could
also exhibit similar regenerative potential, we repeated the
above experiments with the H9 hESC line and obtained similar
results. These results are described in Supporting Information
Figure 3 and suggest that the H9 ABCG2þ hox cells
Figure 2. A representative flow cytometry profile depicting the sustained expression of high Nanog and Oct-4 in the BGO1 ABCG2þhox cells
following exposure to hypoxia/reoxygenation. The ABCG2þhox cells were flow cytometry sorted as shown in Figure 1D and maintained in the
human embryonic stem cells (hESCs) culture media for next 3 weeks. In specific interval of time, the ESCs were fixed and subjected to flow
cytometry expression of Nanog and Oct-4 as described [23]. The inset panel at the top of the prehypoxia panel indicates the isotype control. The
BGO1 hESCs were fixed and permeabilized before subjecting to incubation with primary and secondary (Alexa 488 for Oct-4 and phycoerythrin
for Nanog) antibodies. Results shown are representative of a typical experiment (n ¼ 4; data from 5,000 single-cell events).
Figure 3. ABCG2þhox fraction exhibits a highly undifferentiated and cytoprotective state. (A): The fold change in the transcriptional activity
of Oct-4 and Nanog in ABCG2þhox relative to ABCG2þnox cells. (B): The fold change of the GSH and ABCG2 levels in the ABCG2þhox relative
to ABCG2þnox cells. The GSH level was measured in the CM obtained from the cells. The ABCG2 level was measured by ELISA. (C):
Fold change in cell survival, measured by trypan blue exclusion, of human primary cardiomyocytes and neuronal-like cells (SHSY5Y N-type neuroblastoma
cell line) [31] following exposure to conditioned medium obtained from human embryonic stem cells versus nonexposed medium.
The antioxidant NAC was used as a positive control. (D): Fold change in the survival of the mouse HSCs and MSCs following treatment of carboplatin-treated
mice with CM obtained from ABCG2þhox versus ABCG2þnox cells. *, p < .05; **, p < .001; A, B, C n ¼ 4, and D, n ¼ 3.
Abbreviations: CM, conditioned media; GSH, glutathione; HSC, hematopoietic stem cell; MSC, mesenchymal stem cell; NAC, N-acetyl cysteine;
TF, transcription factor.
1690 Enhanced Stemness Reprogramming
exhibited a high HIF-2a, Oct-4, Nanog and low p53 phenotype,
and exhibited in vivo cytoprotective activity comparable
to that of NAC (Supporting Information Fig. 3). Thus, we
conclude that the ABCG2þ hox fraction in both the BGO1
and H9 hESCs exhibited cytoprotective activities. We then
decided to evaluate the characteristic of the ABCG2þhox
phenotype using the BGO1-derived ABCG2þhox cells.
The Enhanced Stemness State Portends an Extreme
Degree of Self-Sufficiency
We noted that the ABCG2þhox fraction showed a higher
level of Nanog activity (Figs. 2, 3A). Since very high Nanog
expression leads to an embryonal carcinomatous (EC) like
state of extreme self-sufficiency [16, 17, 35], we investigated
whether the ABCG2þhox fraction attained extreme self-
sufficiency.
First, we assayed for the expression of FGF-2 and Bcl-2,
involved in growth and regulation of apoptosis and associated
with extreme self-sufficiency of ESCs and EC cells [36]. We
found that FGF-2 and Bcl-2 expressions were upregulated
two- to threefold in the ABCG2þhox versus ABCG2þnox
fractions (Fig. 4A). Second, we performed a standard clonogenic
assay, minus supplementation with a ROCK inhibitor
[37], on flow-sorted ABCG2þhox and ABCG2þnox cells
grown on Matrigel as single cells, and the ESC colony formation
was measured after 4 days. We found that the
ABCG2þhox cells showed a 10-fold higher colony forming
capacity than the ABCG2þnox cells (Fig. 4B). The colony
forming efficiency was similar to that of Tera-2 cells (Fig.
4B). Our results are consistent with previous reports that high
Nanog and Bcl-2 levels enhance the ability of flow-sorted single
hESC to survive without the addition of the ROCK inhibitor
[36]. Thus, we suggest that the ABCG2þhox cells demonstrated
a significantly greater degree of self-sufficiency than
the ABCG2þnox fraction in the in vitro clonogenic assay.
Next, we performed the in vivo teratoma assay to further
characterize the enhanced state of self-sufficiency. The
ABCG2þhox and ABCG2þnox were injected subcutaneously
into immunocompromised mice [38] to obtain teratomas (Supporting
Information Fig. 4), and a serial dilution assay was
performed to compare the teratomagenic frequency between
the ABCG2þhox and ABCG2þnox. Tera-2 was used as a
control. We found a 15-fold increase in teratomagenic efficiency
of ABCG2þhox versus ABCG2þnox (Fig. 4C) and
comparable to the Tera-2 cells (Fig. 4C). We noted that the
ABCG2þhox fraction contained a eightfold higher level of
Nanog expression (Fig. 4D, and detail quantification given in
the Supporting Information Fig. 5) than the ABCG2þnox
fraction in the Matrigel plug, 10 days after implantation. This
is expected, since higher self-sufficiency would contribute to
higher cell survival and resistance to differentiation of ESCs
during subcutaneous implantation. We noted that a small
Figure 4. ABCG2þhox fraction exhibits a phenotype of self-sufficiency. (A): The fold change in FGF-2 and Bcl-2 in ABCG2þhox cells relative
to ABCG2þnox cells. (B): The single-cell clonogenic assay showing the alkaline phosphatase staining of abundant colony formation in the
ABCG2þhox cell fraction. The histogram shows the significant fold increase in colonies derived from the ABCG2þhox fraction. Tera-2-derived
colonies were used as a control. (C): The in vivo limiting dilution assay showing higher teratomagenic formation efficiency of ABCG2þhox versus
ABCG2þnox. Tera-2 teratoma formation was used as a control. (D): Immunofluorescence staining on a frozen section of ABCG2þ cells containing
Matrigel plug showing expression of Nanog positive cells. The quantification of the Nanog expressing cells is given in the Supporting Information
Figure 6. *, p < .05; **, p < .001; ***, p < .00001, n ¼ 4. Scale bar ¼ 40 lm (B) and (D). Abbreviations: DAPI, 40
,6-diamidino-2-phenylindole;
FGF-2, fibroblast growth factor-2.
Das, Bayat-Mokhtari, Tsui et al. 1691
www.StemCells.com
fraction of the ABCG2þ nox cells in the Matrigel plug continued
to express high Nanog (Fig. 4D; Supporting Information
Fig. 5). Considering that the Matrigel plug microenvironment
is hypoxic [22, 23], we speculate that a few of the
ABCG2þnox cells in the Matrigel plugs might have undergone
transient reprogramming to acquire ABCG2þhox phenotype.
We also noted that the ABCG2þhox xenograft-derived
cells did not form secondary teratomas suggesting that the
ABCG2þhox fraction exhibited EC-like characteristics of
self-sufficiency for a transient period of time in vivo but did
not exhibit the full-blown EC-like malignant phenotype of teratocarcinoma
as does Tera-2.
Taken together, we concluded that the ABCG2þhox fraction
underwent a temporary but measurable phenotypic switch
to an enhanced stemness state of extreme self-sufficiency but
did not exhibit malignant transformation to teratocarcinoma.
HIF-2a-Mediated p53 Suppression Maintained the
Enhanced Stemness State
Because the pluripotency state is stabilized in a metastable
state, wherein Nanog is maintained within a threshold level
[39], the high Nanog expression state is unstable and therefore
returns to a basal level within hours to days [40, 41]. Multiple
mechanisms, including p53 activity are likely involved in
maintenance of a balanced state of Nanog expression [39].
We investigated the molecular mechanism of how the
ABCG2þhox fraction maintained a state of high Nanog and
associated self-sufficiency for a prolonged period of time
without returning to the basal state of pluripotency. We
speculated that high HIF-2a might suppress p53 activity to
allow the high Nanog state to transiently retain stability and
tested this idea as follows.
First, we investigated the phenotypic switching of
ABCG2þnox to ABCG2þhox cells when they were made deficient
in HIF-2a and HIF-1a. Both HIF-1a and HIF-2a were
downregulated in BGO1 cells by siRNA gene silencing (Supporting
Information Fig. 6a) and also by treatment with FM19G11
(Calbiochem), an inhibitor of HIFs [21]. HIF-deficient BGO1
cells were then exposed to hypoxia/reoxygenation to obtain the
ABCG2þhox fraction. The inhibition of HIF-2a by siRNA as
well as FM19G11 led to an eight- to ninefold reduction of the
ABCG2þhox fraction (Supporting Information Fig. 6b). In contrast,
the inhibition of HIF-1a by siRNA gene silencing of HIF-
1a did not affect the ABCG2þhox fraction (Supporting Information
Fig. 6b). These results suggest that HIF-2a and not HIF-1a
was required for the survival and/or expansion of the
ABCG2þhox fraction during hypoxia/reoxygenation.
Next, we investigated the potential HIF-2a-mediated suppression
of p53 by treating the ABCG2þhox fraction with the
HIF-2a inhibitor FM19G11 or siRNA HIF-2a and then determining
cell survival and the expressions of p53. We found
that HIF-2a inhibition resulted in a fourfold increase in apoptosis
as measured by caspase-3 activity [22] (Fig. 5A) and
corresponding increase in p53 expression (Fig. 5B). The
enhanced stemness state Nanog and ABCG2 expressions were
downregulated (Fig. 5C). These results suggested that downregulating
HIF-2a led to increased expression of p53 and associated
differentiation of the enhanced stemness state.
We interrogated whether the increased p53 activity was
associated with increased apoptosis following HIF-2a inhibition.
First, the HIF-2a inhibitor-treated ABCG2þhox fraction was
treated with either siRNA p53 or pifithrin-a, a p53 inhibitor
[42]. We found that siRNA gene silencing of p53 enhanced survival
of the ABCG2þhox fraction (Fig. 5A). A similar result
was obtained with pifithrin-a treatment (data not shown). These
results suggested that HIF-2a was involved in the inhibition of
p53-mediated apoptosis/differentiation of the ABCG2þhox
fraction. We concluded that high HIF-2a suppressed the p53
activity to enhance the survival of ABCG2þhox cells.
HIF-2a Modulates the p53/MDM2 Oscillation
System
Because, p53 is associated with p53/MDM2 oscillation having a
period of high MDM2 and low p53 state [43–45], we evaluated
the potential role of HIF-2a in altering the oscillation to sustain
a high MDM2 and low p53 state in the ABCG2þhox cells.
Figure 5. HIF-2a is involved in the maintenance of the enhanced
stemness state. (A): The fold change in the apoptosis (caspase-3
activity) of the ABCG2þhox (day-8 post-hypoxia) cells versus
ABCG2þnox cells following treatment with various inhibitors of
HIF-2a and p53. (B): The fold change of HIF-2a and p53 protein
levels in the surviving cells of (A). (C): The fold change in Oct-4
and ABCG2 protein levels in the surviving cells of (A).*, p < .05;
**, p < .001; n ¼ 3. Abbreviations: HIF, hypoxia inducible factor;
siRNA, small interfering RNA.
1692 Enhanced Stemness Reprogramming
First, the ABCG2þhox cells were treated with Nutlin-3,
an MDM2 inhibitor, and the survival as well as differentiation
of the treated cells were evaluated by performing appropriate
assays. We noted that Nutlin-3 inhibition led to a ninefold
reduction in the ABCG2þhox fraction (Supporting Information
Fig. 6b). Next, we observed a four- to fivefold reduction
in Nanog, Oct-4, and ABCG2 levels (p < .05; data not
shown). These results suggest that a high MDM2 state was
associated with the enhanced stemness state.
Next, we characterized the effect of HIF-2a inhibition in
the p53/MDM2 oscillation state of ABCG2þhox cells by
measuring the protein levels of p53 and MDM2 by InCell
western and ELISA over a period of 3 weeks. The results
obtained by performing InCell western are shown in Figure 6.
We found a p53/MDM2 oscillation pattern in the ABCG2þhox
within the first 12 hours of hypoxia (Fig. 6A). The oscillation
then ceased, but a high MDM2 and low p53 state persisted for
2 weeks (Fig. 6B). We found similar evidence of the p53/
MDM2 oscillation by the ELISA measurement (data not
shown). Most importantly, HIF-2a inhibition by FM19G11 led
to a p53/MDM2 oscillation of high p53 amplitude, and low
MDM2 state (Fig. 6C), and then return of both p53 and
MDM2 levels to their basal state (Fig. 6D). We noted that the
p53/MDM2 oscillation in the SSEA3þ/ABCG2À fraction, that
did not show high HIF-2a, also did not exhibit a high MDM2
state (Supporting Information Fig. 7). These results suggest
that BGO1-derived ABCG2þhox exhibited an abnormal state
of p53/MDM2 oscillation characterized by a sustained high
MDM2 state that lasted for more than 2 weeks. Treatment with
HIF inhibitor FM19G11 prevented the occurrence of this sustained
high MDM2 state.
To further investigate the role of HIF-2a in the initiation
of the abnormal high MDM2 state of the p53/MDM2 oscillation,
we performed siRNA HIF-2a silencing of BGO1 before
subjecting cells to hypoxia/reoxygenation as described in Supporting
Information Figure 7. The post-hypoxia ABCG2þhox
was analyzed by ELISA assay to obtain p53 and MDM2 protein
levels. We also performed the experiment in H9 hESCs
that showed the presence of an ABCG2þhox phenotype (Supporting
Information Fig. 3). The result obtained with BGO1
was similar to data described in Figure 6D. The findings on
H9 are described in Supporting Information Figure 7c, 7d.
First, we noted that H9-derived ABCG2þhox showed a similar
pattern of an abnormal state of low p53 and high MDM2
that was sustained for 2 weeks. Second, siRNA HIF-2a treatment
lowered the MDM2 level to a basal state in H9 (Supporting
Information Fig. 7d). We concluded that the hypoxiainduced
HIF-2a was associated with the transient stabilization
of a high MDM2 and low p53 state in the ABCG2þ hox fraction
of both BGO1 and H9 hESCs.
DISCUSSION
hESCs are pluripotent stem cells derived from inner cell mass
of developing embryos and have adapted to specialized in
vitro culture conditions to maintain an undifferentiated state
with self-renewal capacity (stemness state). We now report
that a rare fraction of ESCs, the ABCG2þhox fraction, undergoes
transition under hypoxia/reoxygenation into a transient,
but reasonably persistent phenotype of high HIF-2a, Nanog,
Oct4, and ABCG2 activity. The phenotype was associated
with an abnormal p53 and MDM2 oscillation characterized by
low p53 and high MDM2 state. This phenotype was correlated
with a state of enhanced stemness that was permissive
for survival and self-sufficiency. Importantly, in the context
of tissue repair and regeneration, the enhanced stemness state
is characterized by high cytoprotective activity. Whether
Figure 6. HIF-2a is involved in the generation of an altered p53/MDM2 oscillation in the ABCG2þhox. (A): The kinetic changes of p53 and
MDM2 proteins in ABCG2þ hox cells immediately, (B) days after, and (C) and (D) following treatment with hypoxia inducible factor (HIF) inhibitor
FM19G11 [21]. The p53 and MDM2 levels were checked by InCell western and the induction fold of p53 and MDM2 after reoxygenation
was calculated at each time point. Data presented represent the mean of four experiments. Similar results of fold induction of p53 and MDM2
was obtained by ELISA measurement (described in text).
Das, Bayat-Mokhtari, Tsui et al. 1693
www.StemCells.com
similar activity could be attributed to adult stem cells are now
under active investigation.
Thus, we propose that hESCs use a novel mechanism of
enhanced stemness reprogramming to acquire a higher degree
of functionality for regeneration, where HIF-2a, turned on by
hypoxia and oxidative stress, sustains functional activity for at
least 2 weeks after the initial induction.
Although generalization of the enhanced stemness reprogramming
as a defense strategy common to stem cells would
be highly speculative, we suggest that if used, this potential
strategy of cytoprotection (altruism) might benefit hESCs as
well as adult stem cells to survive and maintain stemness and
regenerative potential at the site of injury.
Sites of tissue injury are subjected to cycles of hypoxia/
reoxygenation during the repair process. These sites appear to
attract stem cells whose roles are still poorly understood
although there is a growing understanding that stem cells may
assist in the repair process either directly or through production
of critical factors [1, 46]. Here we subjected hESCs to
hypoxia/reoxygenation mimicking the site of injury. The
hESC model has the advantage that distinct phenotypes can
be discriminated in vitro and in vivo and the factors
that maintain the stem cell state have been well elaborated
[16, 47].
The time frame of this hESC cell phenotype of enhanced
stemness state, characterized by higher degree of self-sufficiency
and cytoprotective ability, corresponds to prolonged and
abnormal state of p53/MDM2 oscillation. This altered state of
the oscillation system was associated with a high HIF-2a activity
since inhibiting HIF-2a led to phenotypic differentiation of
the enhanced stemness phenotype. Thus the HIF-2a and p53
systems, coordinately upregulated during hypoxia and oxidative
stress, could contribute to the emergence of a rare fraction of
hESCs harboring high HIF-2a and low p53, that is, a state of
enhanced stemness state of self-sufficiency.
This transient state now permitted hESCs to acquire a
highly cytoprotective, regenerative activity. This cytoprotective
activity appears to be mediated by a high level of functional
ABCG2, an efflux pump that secretes GSH, a potent
endogenous antioxidant. We showed that the enhanced stemness
state was associated with a high level of GSH secretion,
with GSH secreted into the CM, which was capable of cytoprotecting
cardiomyocytes and primitive neuronal cells represented
by SH-SY5Y cell line from the toxic effect of hypoxia/reperfusion
injury in vitro as well as chemotherapeutic
induced BM cell cytotoxicity in vivo. Whether other factors
in the CM act cooperatively with GSH remains to be determined,
however, these observations strongly support the
notion that stem cells in sites of injury and repair condition
the environment and may contribute effectively to the regenerative
process.
The suppression of p53 was required to initiate and maintain
the enhance stemness state, and therefore, the cytoprotective
activity. However, one potential trade-off of p53 suppression
would be the imbalance in the stemness state that could
lead to malignant transformation, a major concern of current
stem cell-based therapies [17]. In this regard, we noted the
observation that MDM2 is highly expressed in EC and positively
regulates FGF-2 [48, 49]. Although we found that the
ABCG2þhox cells were highly teratomagenic, ABCG2þhox
cells did not exhibit teratocarcinoma activity, suggesting that
the enhanced stemness state did not lead to malignant transformation.
One reason may be that the transient nature of the
enhanced stemness state functions as a possible safety mechanism
that is built into stem cells to mitigate against malignancy.
In this context, it would be of crucial importance to
determine how this safety mechanism is overridden, for example,
by an oncogenic event, in the switch to a malignant phenotype.
Our model might be ideally suited to probe this vital
question.
CONCLUSION
In summary, we now propose a novel mechanism whereby
hESCs might contribute to healing when exposed to an extreme
oxidative stress state of hypoxia/reoxygenation. In this potential
mechanism, a fraction of highly undifferentiated stem cells acquire
a transient state of enhanced stemness reprogramming
characterized by high antioxidants and their secretion, where
p53 is temporarily downregulated. Thus in this model, and perhaps
counterintuitive, a few stem cells can acquire a higher
degree of self-sufficiency to resist differentiation in the hypoxia/oxidative
stress microenvironment to protect other cells
from oxidative stress damage. Is this enhanced stemness
reprogramming then a form of cell altruism [50] that could benefit
the organism from the earliest developmental beginnings?
ACKNOWLEDGMENTS
We thank Dr. Peter W. Andrews, University of Sheffield, U.K.
for reviewing and making valuable comments on the manuscript.
B.D. thank Dr. Sylvain Baruchel, and Dr. David Malkin, Hospital
for Sick Children for encouragement and support during the
initiation of this study. This work was supported by grants from
Canadian Cancer Society Research Institute (B.D. and H.Y.),
Grand Challenges Exploration Initiative grant, Bill and Melinda
Gates Foundation (B.D.), Laurel Foundation, CA (B.D. and
D.W.F.), Restracomp, Sick Kids Foundation, Hospital for Sick
Children (B.D.), and James Birrell Neuroblastoma Fund (B.D.
and R.M.).
DISCLOSURE OF POTENTIAL
CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.
REFERENCES
1 Prockop DJ. Repair of tissues by adult stem/progenitor cells (MSCs):
Controversies, myths, and changing paradigms. Mol Ther 2009;17:
939–946.
2 Kim WS, Park BS, Sung JH. The wound-healing and antioxidant
effects of adipose-derived stem cells. Expert Opin Biol Ther 2009;9:
879–887.
3 Momcilovic O, Choi S, Varum S et al. Ionizing radiation induces
ataxia telangiectasia mutated-dependent checkpoint signaling and G(2)
but not G(1) cell cycle arrest in pluripotent human embryonic stem
cells. Stem Cells 2009;27:1822–1835.
4 Maxwell PH. Hypoxia-inducible factor as a physiological regulator.
Exp Physiol 2005;90:791–797.
5 An WG, Kanekal M, Simon MC et al. Stabilization of wild-type p53
by hypoxia-inducible factor 1a. Nature 1998;392:405–408.
6 Bertout JA, Majmundar AJ, Gordan JD et al. HIF2a inhibition promotes
p53 pathway activity, tumor cell death, and radiation responses.
Proc Natl Acad Sci USA 2009;106:14391–14396.
7 Scortegagna M, Ding K, Oktay Y et al. Multiple organ pathology,
metabolic abnormalities and impaired homeostasis of reactive oxygen
species in Epas1À/À mice. Nat Genet 2003;35:331–340.
1694 Enhanced Stemness Reprogramming
8 Martin CM, Ferdous A, Gallardo T et al. Hypoxia-inducible factor-2a
transactivates Abcg2 and promotes cytoprotection in cardiac side population
cells. Circ Res 2008;102:1075–1081.
9 Susanto J, Lin YH, Chen YN et al. Porphyrin homeostasis maintained
by ABCG2 regulates self-renewal of embryonic stem cells. PLoS One
2008;3:e4023.
10 Kim M, Turnquist H, Jackson J et al. The multidrug resistance transporter
ABCG2 (breast cancer resistance protein 1) effluxes Hoechst
33342 and is overexpressed in hematopoietic stem cells. Clin Cancer
Res 2002;8:22–28.
11 Suzuki M, Suzuki H, Sugimoto Y et al. ABCG2 transports sulfated
conjugates of steroids and xenobiotics. J Biol Chem 2003;278:
22644–22649.
12 Zhou S, Schuetz JD, Bunting KD et al. The ABC transporter Bcrp1/
ABCG2 is expressed in a wide variety of stem cells and is a molecular
determinant of the side-population phenotype. Nat Med 2001;7:
1028–1034.
13 Brechbuhl HM, Gould N, Kachadourian R et al. Glutathione transport
is a unique function of the ATP-binding cassette protein ABCG2. J
Biol Chem 2010;285:16582–16587.
14 Das B, Antoon R, Tsuchida R et al. Squalene selectively protects
mouse bone marrow progenitors against cisplatin and carboplatininduced
cytotoxicity in vivo without protecting tumor growth. Neoplasia
2008;10:1105–1119.
15 Covello KL, Kehler J, Yu H et al. HIF-2a regulates Oct-4: Effects of
hypoxia on stem cell function, embryonic development, and tumor
growth. Genes Dev 2006;20:557–570.
16 Andrews PW. From teratocarcinomas to embryonic stem cells. Philos
Trans R Soc Lond B Biol Sci 2002;357:405–417.
17 Baker DE, Harrison NJ, Maltby E et al. Adaptation to culture of
human embryonic stem cells and oncogenesis in vivo. Nat Biotechnol
2007;25:207–215.
18 Lin T, Chao C, Saito S et al. p53 induces differentiation of mouse embryonic
stem cells by suppressing Nanog expression. Nat Cell Biol
2005;7:165–171.
19 Maimets T, Neganova I, Armstrong L et al. Activation of p53 by nutlin
leads to rapid differentiation of human embryonic stem cells.
Oncogene 2008;27:5277–5287.
20 Xu Y. A new role for p53 in maintaining genetic stability in embryonic
stem cells. Cell Cycle 2005;4:363–364.
21 Moreno-Manzano V, Rodriguez-Jimenez FJ, Acena-Bonilla JL et al.
FM19G11, a new hypoxia-inducible factor (HIF) modulator, affects
stem cell differentiation status. J Biol Chem 2010;285:1333–1342.
22 Das B, Yeger H, Tsuchida R et al. A hypoxia-driven vascular endothelial
growth factor/Flt1 autocrine loop interacts with hypoxia-inducible
factor-1a through mitogen-activated protein kinase/extracellular
signal-regulated kinase 1/2 pathway in neuroblastoma. Cancer Res
2005;65:7267–7275.
23 Das B, Tsuchida R, Malkin D et al. Hypoxia enhances tumor stemness
by increasing the invasive and tumorigenic side population fraction.
Stem Cells 2008;26:1818–1830.
24 Draper JS, Pigott C, Thomson JA et al. Surface antigens of human
embryonic stem cells: Changes upon differentiation in culture. J Anat
2002;200:249–258.
25 Braam SR, Denning C, Matsa E et al. Feeder-free culture of human
embryonic stem cells in conditioned medium for efficient genetic
modification. Nat Protoc 2008;3:1435–1443.
26 Watanabe K, Ueno M, Kamiya D et al. A ROCK inhibitor permits
survival of dissociated human embryonic stem cells. Nat Biotechnol
2007;25:681–686.
27 Das B. The role of VEGF autocrine signaling in hypoxia and oxidative
stress driven ‘‘stemness switch’’: Implications in solid tumor progression
and metastasis [Vol PhD Thesis]. Toronto: Institute of
Medical Sciences, University of Toronto; 2007:215.
28 Tsuchida R, Das B, Yeger H et al. Cisplatin treatment increases survival
and expansion of a highly tumorigenic side-population fraction
by upregulating VEGF/Flt1 autocrine signaling. Oncogene 2008;27:
3923–3934.
29 Lionetti V, Cantoni S, Cavallini C et al. Hyaluronan mixed esters of
butyric and retinoic acid affording myocardial survival and repair
without stem cell transplantation. J Biol Chem 2010;285:9949–9961.
30 Hu Y, Smyth GK. ELDA: Extreme limiting dilution analysis for comparing
depleted and enriched populations in stem cell and other
assays. J Immunol Methods 2009;347:70–78.
31 Decressac M, Pain S, Chabeauti PY et al. Neuroprotection by neuropeptide
Y in cell and animal models of Parkinson’s disease. Neurobiol
Aging 2011 [Epub ahead of print].
32 Drowley L, Okada M, Beckman S et al. Cellular antioxidant levels
influence muscle stem cell therapy. Mol Ther 2010;18:1865–1873.
33 Ito K, Hirao A, Arai F et al. Regulation of oxidative stress by ATM is
required for self-renewal of haematopoietic stem cells. Nature 2004;
431:997–1002.
34 Lin TM, Tsai JL, Lin SD et al. Accelerated growth and prolonged
lifespan of adipose tissue-derived human mesenchymal stem cells in a
medium using reduced calcium and antioxidants. Stem Cells Dev
2005;14:92–102.
35 Draper JS, Smith K, Gokhale P et al. Recurrent gain of chromosomes
17q and 12 in cultured human embryonic stem cells. Nat Biotechnol
2004;22:53–54.
36 Ardehali R, Inlay MA, Ali SR et al. Overexpression of BCL2 enhances
survival of human embryonic stem cells during stress and obviates
the requirement for serum factors. Proc Natl Acad Sci USA 2011;108:
3282–3287.
37 Xu Y, Zhu X, Hahm HS et al. Revealing a core signaling regulatory
mechanism for pluripotent stem cell survival and self-renewal by
small molecules. Proc Natl Acad Sci USA 2010;107:8129–8134.
38 Cao F, Lin S, Xie X et al. In vivo visualization of embryonic stem
cell survival, proliferation, and migration after cardiac delivery. Circulation
2006;113:1005–1014.
39 Pan G, Thomson JA. Nanog and transcriptional networks in embryonic
stem cell pluripotency. Cell Res 2007;17:42–49.
40 Silva J, Smith A. Capturing pluripotency. Cell 2008;132:532–536.
41 Singh AM, Hamazaki T, Hankowski KE et al. A heterogeneous
expression pattern for Nanog in embryonic stem cells. Stem Cells
2007;25:2534–2542.
42 Solozobova V, Rolletschek A, Blattner C. Nuclear accumulation and
activation of p53 in embryonic stem cells after DNA damage. BMC
Cell Biol 2009;10:46.
43 Batchelor E, Mock CS, Bhan I et al. Recurrent initiation: A mechanism
for triggering p53 pulses in response to DNA damage. Mol Cell
2008;30:277–289.
44 Ciliberto A, Novak B, Tyson JJ. Steady states and oscillations in the
p53/Mdm2 network. Cell Cycle 2005;4:488–493.
45 Lev Bar-Or R, Maya R, Segel LA et al. Generation of oscillations by
the p53-Mdm2 feedback loop: A theoretical and experimental study.
Proc Natl Acad Sci USA 2000;97:11250–11255.
46 Prockop DJ, Kota DJ, Bazhanov N et al. Evolving paradigms for
repair of tissues by adult stem/progenitor cells (MSCs). J Cell Mol
Med14:2190–2199.
47 Ying QL, Wray J, Nichols J et al. The ground state of embryonic
stem cell self-renewal. Nature 2008;453:519–523.
48 Datta MW, Macri E, Signoretti S et al. Transition from in situ to invasive
testicular germ cell neoplasia is associated with the loss of p21
and gain of mdm-2 expression. Mod Pathol 2001;14:437–442.
49 Shaulian E, Resnitzky D, Shifman O et al. Induction of Mdm2 and
enhancement of cell survival by bFGF. Oncogene 1997;15:2717–2725.
50 Das B. Science Behind Squalene. Canada: Toronto Medical Publishing,
2000:187; ISBN 1–890412-95–3. Available at http://www.science-
forlife.eu/tekst%20sciencebehindsqualene.htm.
See www.StemCells.com for supporting information available online.
Das, Bayat-Mokhtari, Tsui et al. 1695