Polycomb Group Protein Bmi1 Promotes
Hematopoietic Cell Development from Embryonic Stem Cells
Xiaolei Ding,1,2Qiong Lin,1,2Roberto Ensenat-Waser,3Stefan Rose-John,4and Martin Zenke1,2
Bmi1 is a component of the Polycomb repressive complexes and essential for maintaining the pool of adult stem
cells. Polycomb repressive complexes are key regulators for embryonic development by modifying chromatin
architecture and maintaining gene repression. To assess the role of Bmi1 in pluripotent stem cells and on exit
from pluripotency during differentiation, we studied forced Bmi1 expression in mouse embryonic stem cells
(ESC). We found that ESC do not express detectable levels of Bmi1 RNA and protein and that forced Bmi1
expression had no obvious influence on ESC self-renewal. However, upon ESC differentiation, Bmi1 effectively
enhanced development of hematopoietic cells. Global transcriptional profiling identified a large array of genes
that were differentially regulated during ESC differentiation by Bmi1. Importantly, we found that Bmi1 induced
a prominent up-regulation of Gata2, a zinc finger transcription factor, which is essential for primitive hemato-
poietic cell generation from mesoderm. In addition, Bmi1 caused sustained growth and a >100-fold expansion of
ESC-derived hematopoietic stem/progenitor cells within 2–3 weeks of culture. The enhanced proliferative ca-
pacity was associated with reduced Ink4a/Arf expression in Bmi1-transduced cells. Taken together, our ex-
periments demonstrate distinct activities of Bmi1 in ESC and ESC-derived hematopoietic progenitor cells. In
addition, Bmi1 enhances the propensity of ESC in differentiating toward the hematopoietic lineage. Thus, Bmi1
could be a candidate gene for engineered adult stem cell derivation from ESC.
potential. Lossofself-renewaland inductionofdifferentiation
are accompanied by specific changes in gene expression.
Additionally, there is increasing evidence that cell identity
and function are determined and maintained by epigenetic
status, including DNA methylation and histone modification
[1–3]. We have previously shown that chromatin modifying
. The Polycomb group (PcG) proteins act as epigenetic
modifiers and have received significant attention due to their
role in stem cell self-renewal, commitment, and differentia-
tion, and in cancer stem cell formation [5,6].
PcG proteins were initially described in Drosphila, where
they control embryonic development by repressing homeotic
gene expression [5–7]. PcG proteins are conserved from
Drosphila to humans, and most of them are involved in
maintenance of cellular memory by modifying chromosome
structure and silencing gene expression. PcG proteins occur in
large protein complexes to exert transcriptional repressor
tem cells are characterized by an exceptionally high
self-renewal activity and their multilineage differentiation
activity, referred to as Polycomb repressive complexes (PRC)
[6,7]. In mammals, 2 PRC complexes, PRC1 and PRC2, have
been described. PRC2, which contains Eed, Suz12, and Ezh2
proteins, is recruited to chromatin and trimethylates lysine
residue 27 of histone 3 (H3K27me3). PRC1 contains Bmi1,
Ring1A/B, Cbx, Mel18, and Mph and is recruited to specific
sites formed by PRC2, referred to as maintaining complex .
PRC and H3K27me3 co-occupy cis-regulatory elements
of a large number of development-associated genes [8,9].
Genome-wide mapping of histone modification in embry-
onic stem cells (ESC) identified regions that are modified
with both H3K27me3 and histone 3 lysine 4 trimethylation
(H3K4me3) [9–12]. H3K4me3 is catalyzed through trithorax
group proteins and associated with active transcription units
H3K27me3, referred to as ‘‘bivalent domain,’’ effectively
controls gene transcription and poise genes in a ready-for-
transcription status [10,11].
Down-regulation of the PRC2 component Eed and Suz12
caused silenced differentiation-associated genes to become
re-expressed in ESC [9,12]. Further, ESC could not be es-
tablished from Ezh2 deficient mice . For PRC1, it has
of H3K4me3 and
1Department of Cell Biology, Institute for Biomedical Engineering, RWTH Aachen University Medical School, Aachen, Germany.
2Helmholtz Institute for Biomedical Engineering, RWTH Aachen University, Aachen, Germany.
3Joint Research Centre, Institute for Health and Consumer Protection, In Vitro Methods Unit, European Commission, Ispra, Italy.
4Institute of Biochemistry, Christian-Albrechts-University, Kiel, Germany.
STEM CELLS AND DEVELOPMENT
Volume 21, Number 1, 2012
? Mary Ann Liebert, Inc.
been shown that RingA/B is directly involved in the tran-
scriptional network associated with ESC pluripotency .
All these studies indicate that PcG proteins are highly critical
for embryonic development, the establishment and mainte-
nance of ESC pluripotency. Recent studies demonstrated that
PRC2 components, such as Ezh2 and Suz12, also orchestrate
gene expression in adult stem cells [15–17].
Unlike other PcG components that are ubiquitously ex-
pressed during development, Bmi1, a key component of PRC1,
is selectively expressed in diverse postnatal adult stem cells,
including neural stem cells (NSC) and hematopoietic stem cells
(HSC) [18–22]. Bmi1 activity in maintaining the pool of adult
stem cells is mainly due to repression of the Ink4a/Arf locus,
which encodes inhibitors of the cell cycle kinase p16Ink4a
and p19Arf [18,19,23,24]. Accordingly, in Bmi1 deficient
(Bmi1-/-) mice, the number of adult HSC is markedly
decreased. Additionally, in competitive transplantation ex-
periments, the repopulation capacity of Bmi1-/- HSC was
significantly decreased, thus mirroring their deficiency in
self-renewal [19,20]. In fact, with HSC differentiation, Bmi1
expression declines gradually . In contrast, Bmi1 over-
expression enhances HSC self-renewal in both the human
and mouse system [26,27]. Thus, there is multiple evidence of
Bmi1 activity in maintaining the pool of adult stem cells, yet
its role in pluripotent stem cells and on transition toward
adult stem/progenitor cells has so far not been studied.
Here, we examined the activity of Bmi1 on ESC and on
hematopoietic stem/progenitor cells derived from ESC. We
present evidence for a novel activity of Bmi1 in enhancing
hematopoietic cell development from ESC.
Materials and Methods
Cells and cell culture
ESC culture was performed as previously described .
Briefly, R1 ESC were cultured on irradiated mouse embryonic
fibroblasts (MEF) feeder. E14 and CCE ESC were cultured on
0.1% gelatin (Sigma-Aldrich) precoated dishes with ESC me-
dium. ESC medium was Dulbecco’s modified Eagle’s medium
inactivated fetal calf serum (FCS; Lonza), 25mM HEPES and
1,000U recombinant leukemia inhibitory factor (LIF, a kind
gift of Anna M. Wobus, IPK, Gatersleben, Germany), 2mM
L-glutamine, 100 units penicillin/100mg streptomycin, 0.1mM
nonessential amino acids, and 50mM b-mercaptoethanol. ESC
differentiation medium was ESC medium but with 10% FCS
and without LIF. ESC were cultured at 37?C with 5% CO2.
Medium was refreshed daily, and cells were passaged every
2–3 days before reaching confluency. OP9 stroma cells (a kind
gift of Ursula Just, University of Kiel, Kiel, Germany) were
grown in a-MEM supplemented with 20% FCS (both from
PAN Biotech), 2mM L-glutamine, 100 units penicillin/100mg
streptomycin, 25mM HEPES, and 100mM b-mercaptoethanol.
293T human embryonic kidney cells were maintained in
DMEM medium containing 10% FCS, 2mM L-glutamine, and
100 units penicillin/100mg streptomycin. Unless stated differ-
ently, all reagents were purchased from Invitrogen.
Bmi1 lentivirus vector and infection of ESC
Bmi1 cDNA was obtained from mouse NSC  by reverse
transcriptase (RT)-polymerase chain reaction (PCR) (forward
reverse primer 5¢-GGGATCCCTAACCAGATGAAGTTGCT
GATGACCC) and cloned into pJET vector (Fermentas Life
Sciences). Bmi1 sequence was released from the construct by
XbaI and BamH1 digestion and subcloned into XbaI/BamH1
sites of FUGIE vector (a kind gift from Filip Farnebo, Kar-
olinska Institute, Stockholm, Sweden). The Bmi1 cDNA was
verified by sequencing. FUGIE vector contains the human
ubiquitin C promoter driving Bmi1 expression and an in-
ternal ribosome entry site (IRES) for expression of enhanced
green fluorescent protein (GFP).
For infection of ESC, lentivirus was produced from 293T
cells. Briefly, 293T cells were transfected with 10mg FUGIE or
FUGIE-Bmi1 plasmid DNA, 7.5mg pCMVDR8.74 packaging
plasmid, and 2.5mg envelope vector pVSV-G. Virus super-
natant was collected and used to infect ESC in the presence
of 8mg polybrene (Sigma-Aldrich). After 2 passages, infected
GFP+ESC were sorted by flow cytometry, and GFP+ESC
were further expanded. Alkaline phosphatase staining was
performed with Alkaline Phosphatase Staining Kit II (Stem-
gent) according to the manufacturer’s protocol.
ESC were subjected to differentiation by embryoid body
(EB) assay. Briefly, ESC were trypsinized into single-cell
suspensions. EB were generated in hanging drops at 100 cells
per 10mL drops in an inverted bacterial Petri dish for 2 days
with ESC differentiation medium. EB were then collected
and cultured in bacterial Petri dishes for an additional 4
days. For some experiments, ESC were directly subjected to
EB formation in mass culture as previously described .
Hematopoietic cell differentiation from ESC
Hematopoietic cell differentiation from ESC was adapted
from the protocol by Carotta et al. . Briefly, ESC were
differentiated by EB formation as just described. On day 6,
EB were dissociated into single cells with 0.05% trypsin/
ethylenediaminetetraacetic acid (EDTA) solution (Invitro-
gen), and cells were passed through a 40mm cell strainer.
Cells were then plated on gelatin-coated dishes at 2·106
cells/mL in serum-free medium (StemPro34 plus nutrient
supplement; Invitrogen) supplemented with 25ng/mL Flt3
ligand (Flt3L; PeproTech), 30U/mL murine SCF, 5ng/mL
IL-6/soluble IL-6R fusion protein (hyper-IL-6) , 40ng/
mL long-range IGF-1 (Sigma-Aldrich), 2ng/mL murine IL-3
(PeproTech), and 1mM dexamethasone (Sigma-Aldrich).
After 2–3 days, nonadherent cells were harvested, passed
through a 40mm cell strainer, and further cultured in the
same culture medium plus growth factors as just described.
The cell concentration was maintained at 2·106cells/mL,
and the medium was refreshed every 1–2 days. Cumulative
cell numbers were determined with a CASY-1 cell counter
and analyzer system (Scha ¨rfe Systems).
Colony-forming assay and cytospins
Day 6 EB were dissociated into single cells with 0.05%
trypsin/EDTA as just described. Cells were collected and re-
suspended in DMEM, 10% FCS at 1·106cells/mL; and
100mL of cell suspension was added to 3mL MethoCult GF
M3434 methylcellulose, containing insulin, transferrin, SCF,
122DING ET AL.
IL-3, IL-6, and erythropoietin (StemCell Technologies). Me-
thylcellulose cultures were plated in two 35mm dishes and
incubated at 37?C with 5% CO2. The number of erythroid
colonies was scored on day 4–5; myeloid colonies were
evaluated at day 10. 1·104ESC-derived hematopoietic
progenitor cells (ESC-HPC) were subjected to colony-
forming units (CFU) assay in MethoCult GF M3434 meth-
ylcellulose as just described. Individual colonies were
picked, washed, and cyto-centrifuged on slides. Cells were
stained with neutral benzidine and histological dyes  and
Apoptosis and proliferation assays
To study the impact of Bmi1 expression on apoptosis
during ESC differentiation, day 6 EB were dissociated with
0.05% trypsin/EDTA and passed through 40mm cell strainer.
Single-cell suspensions were incubated with 7-amino-acti-
nomycin D (7-AAD) and allophycocyanin (APC)-conjugated
Annexin V (BD Bioscience) according to the manufacturer’s
instructions. Cells were then subjected to flow cytometry. To
determine the growth factor responses of Bmi1-ESC-derived
HPC, a total of 5·104cells were incubated in 200mL serum-
free medium supplemented with different combinations of
growth factors in a 96-well flat-bottom plate at 37?C for 48h.
Samples were then pulsed with 0.75mCi/well [3H] thymidine
(29Ci/mmol; Amersham Biosciences) for 4h and harvested
onto glass fiber filters. Radioactivity was measured by liquid
scintillation counting in a Microbeta counter (Wallac).
Western blot and immunofluorescence analysis
Cells were lysed in 2% sodium dodecyl sulfate with 5mM
EDTA. About 20mg protein per well were subjected to
polyacrylamide gel electrophoresis (12% sodium dodecyl
sulfate-polyacrylamide gels) and transferred onto nitrocel-
lulose membranes by Western blotting. Membranes were
blocked overnight with 2% nonfat milk in phosphate-
buffered saline (PBS) at 4?C and reacted with anti-Bmi1
antibody (Clone F6, 1:1,000 dilution; Millipore) at room
temperature for 2h. Anti-actin antibody (Clone AC-74,
Sigma) was used to detect actin loading control. Peroxidase-
conjugated anti-mouse secondary antibody (NA931, Amer-
sham Biosciences) was incubated for 1h at RT and detected
by chemiluminescence (ECL; Amersham Biosciences). Ima-
ges were acquired by Image Reader Las-1000 (Fujifilm).
For immunofluorescence staining, ESC were grown on
gelatin-coated chamber slides, fixed with 1% formaldehyde,
and permeabilized with PBS buffer containing 0.5% BSA,
0.1% Triton X-100. Cells on cover-slips were washed twice
with 0.5% BSA in PBS buffer and then incubated with anti-
Bmi1 antibody (Clone F6, Millipore). Secondary antibody
was anti-mouse Alexafluor 647 (Invitrogen), and 4¢,6-
diamidino-2-phenylindole dihydrochloride (DAPI) was used
to stain nuclei. Cover-slips were mounted on a glass slide
with mounting medium, and images were acquired under
bright and fluorescent fields with an Axiovert 200 micro-
scope (Carl Zeiss).
RNA isolation and RT-PCR
Total RNA was isolated from cells with RNeasy Mini Kit
(Qiagen). RNA quality and concentration were determined
by spectrophotometer. Total RNA (1 mg) was used as tem-
plate for reverse transcription with High Capacity cDNA
Reverse Transcription Kit (Applied Biosystems). cDNA was
then used for PCR amplification by Taq DNA polymerase
(Fermentas Life Sciences). PCR fragments were separated on
2% agarose gels, and images were recorded with Gel-Doc
The quantitative (q) PCR was carried out with 7300 Real-
Time PCR system (Applied Biosystems). Reactions were
performed with 50ng cDNA, SYBR Green PCR master mix
and primers (see Supplementary Table 2; Supplementary
Data are available online at www.liebertonline.com/scd).
The calculated threshold cycle (CT) value for each sample
was normalized against the corresponding b-actin CT value.
RNA was isolated using RNeasy Mini Kit with DNase I
digestion (Qiagen) and subjected to microarray analysis as
earlier . Briefly, sample preparation was performed accord-
ing to the Expression Analysis Technical Manual (Affymetrix).
GeneChip One-cycle Target Labeling Kit (Affymetrix) and 1mg
total RNA were used. Biotin-labeled cRNA was hybridized on
Affymetrix Mouse Genome 430 2.0 GeneChip arrays. Arrays
were stained, washed, and scanned according to the manu-
facturer’s protocols. Gene expression levels were determined
by GCRMA algorithm in R/Bioconductor. Hierarchical clus-
tering was performed using Pearson correlation coefficient
and the average linkage method and represented by dendro-
gram and heatmap. Differential expression between 2 condi-
tions was analyzed using Student’s t-test. The transcripts with
fold change >2 and P-values <0.05 were considered as being
differentially expressed. Raw P-values were adjusted by
Benjamini and Hochberg’s method. Data sets were submitted
to Gene Expression Omnibus database (www.ncbi.nlm.nih
.gov/geo) under accession number GSE20958.
Flow cytometry analysis
EB were dissociated into single cells by incubation with
0.05% trypsin/EDTA (5min) or 0.1% collagenase IV (Gibco;
20min) at 37?C and pipetting. Cells were passed through a
40mm cell strainer, washed with PBS, and incubated with the
following antibodies: phytoerythrin-conjugated anti-mouse
Flk1 antibody (clone Avas 12alpha1, BD Pharmingen), eFluor
450-conjugated anti-mouse CD41 (clone MWReg30), and
PE-Cy7-conjugated anti-mouse c-Kit (clone ACK2, both
eBioscience) in fluorescent activated cell sorting (FACS)
buffer. Isotype IgG was used as control. Samples were
measured by FACSCanto (BD Biosciences) and analyzed by
Flowjo software (Tree Star). CD41+c-Kit+cells from day 6,
8, and 10 EB-derived cells were sorted by FACSAria cell
sorting system (BD Biosciences).
For immunophenotyping of hematopoietic cells from
Bmi1-ESC, 1·106cells were washed with 1· PBS and incu-
bated with corresponding antibodies in FACS buffer for
30min. Samples were measured by FACSCanto, and results
were analyzed by Flowjo software. The following antibodies
were used: APC-Cy?7 rat anti-mouse CD45 (A95-1), PE-Cy7
anti-mouse c-Kit (ACK2), PE anti-mouse CD24 (M1/69), PE-
Cy5 anti-mouse Sca-1 (D7), PE anti-mouse CD34 (MEC 14.7),
APC anti-mouse CD133 (13A4), APC anti-mouse CD115
BMI1 PROMOTES HEMATOPOIESIS FROM ES CELLS123
(AFS98), pacific blue anti-mouse CD11b (M1/70), biotin anti-
mouse Ter119 (Ter119), biotin anti-mouse Gr-1 (RB6-8C5),
and unlabeled anti-mouse CD14 (mC5-3). For Ter119, Gr-1,
and CD14, the corresponding secondary antibodies were
added after incubation with the first antibody. CD34 anti-
body was from Invitrogen; all other antibodies were from BD
Biosciences or eBioscience.
Results are given as the mean–standard deviation. The
Student’s t-test was applied for paired samples of microarray
data. EB size was calculated with Image-Pro Plus 6.0 software
and quantified with SPSS software (version 10, SPSS, Inc.).
Forced Bmi1 expression in ESC leaves
Expression of Bmi1 and other PRC components was deter-
mined in ESC, NSC, MEF, and bone marrow (BM) mononu-
clear cells by q-PCR analysis. There was no or very low Bmi1
expression in ESC (R1, E14, and CCE lines), whereas Bmi1 was
abundantly expressed in NSC and BM cells (Fig. 1A). There
was also some Bmi1 expression in MEF. The PRC components
Ezh2 and Suz12 were highly expressed in ESC, corroborating
their transcriptional role in ESC [9,12]. As expected, the plur-
ipotency gene Oct4, used as a control, was expressed in ESC
but not in somatic cells. Importantly, Bmi1 expression was
rapidly induced during ESC differentiation in EB assays (Fig.
1B and below Supplementary Fig. S3).
The absence of Bmi1 in ESC and its up-regulation during
their differentiation was surprising, and we, thus, sought to
investigate the impact of enforced Bmi1 expression on ESC
growth and differentiation. CCE ESC were transduced with
lentivirus vector containing Bmi1 cDNA (Fig. 2A). By examin-
ing IRES-driven GFP expression by flow cytometry, infection
efficiencies were routinely *50% (data not shown). GFP+cells
were FACS sorted, thereby achieving a purity of >95% Bmi1-
GFP-transduced cells (Fig. 2B), and we refer to these cells as
Bmi1-ESC. Bmi1 expression in Bmi1-ESC was confirmed by
immunofluorescence and immunoblotting with anti-Bmi1 an-
tibody (Fig. 2C and D). Bmi1 protein was abundantly expressed
both under growth conditions and after ESC differentiation in
EB assays (day 14, Fig. 2D).
Bmi1-ESC had a colony morphology similar to parental
ESC and stained positive for alkaline phosphatase (Supple-
mentary Fig. S1A). ESC are characterized by a pluripotent
gene expression profile [32,33], and, thus, we subjected Bmi1-
ESC to genome-wide gene expression profiling with DNA
microarray. Scatter plot analysis and hierarchical clustering
show that Bmi1-ESC and parental ESC were very similar and
also clustered with a panel of other ESC lines (Supplemen-
tary Fig. S1B and C). Additionally, key genes of plur-
ipotency-transcriptional circuits, such as Oct4, Nanog, and
Sox2, had very similar expression levels in Bmi1-ESC and
control cells (Supplementary Fig. S1B). We also examined
expression of Oct4, Nanog, and Sox2 by RT-PCR and dem-
onstrate that their expression was the same in Bmi1-ESC,
empty vector ESC, and uninfected ESC control (Fig. 2E).
In adult stem cells, Bmi1 represses transcription of the
Ink4a/Arf locus, which encodes the CDK inhibitors p16Ink4a
and p19Arf [18,23], and thereby promotes cell proliferation.
In ESC expression of the Ink4a/Arf locus is repressed by
bivalent chromatin, and Ink4a/Arf mRNA levels are very
low (Fig. 2F) . Ink4a/Arf expression was further down-
regulated in Bmi1-ESC, showing that the repressive activity
on this locus is reinforced by exogenous Bmi1.
Repression of Ink4a/Arf by Bmi1 in adult cells directly
translates in an altered cell cycle profile [18,23], and we, thus,
proceeded to investigate the impact of Bmi1 on ESC prolif-
eration by cell-cycle analysis. Bmi1-ESC have a similar fre-
quency of cells in S phase as controls (Fig. 2G; 31%–34%),
which is in line with the same proliferation rate and growth
potential of Bmi1-ESC as controls (Fig. 2H). Taken together,
these results indicate that exogenous Bmi1 expression leaves
ESC self-renewal and pluripotency unaffected.
Gata2 is up-regulated during mesodermal
differentiation of Bmi1-ESC
To assess the impact of Bmi1 on differentiation, Bmi1-ESC
were subjected to differentiation by EB formation. Bmi1-ESC
gave rise to EB with similar morphology, size, and frequency
as controls (Fig. 3A and Supplementary Fig. S2A). The fre-
quency of apoptotic cells was the same for Bmi-ESC and
of Bmi1, Ezh2, and Suz12 in undifferentiated ESC (R1, E14,
and CCE lines), MEF, NSC, and BM cells was analyzed by
q-PCR. Expression levels were normalized to b-actin, and
relative RNA levels are shown (Bmi1, Ezh2, and Suz12 ex-
pression in MEF was set as a reference to 1.0; Oct4 expression
in CCE ESC was arbitrarily set to 10). Average values from 2
independent experiments are shown. (B) Kinetics of Bmi1
expression during ESC differentiation in EB assay was
measured by q-PCR and normalized to b-actin. BM, bone
marrow; EB, embryoid body; ESC, embryonic stem cells;
MEF, mouse embryonic fibroblasts; NSC, neural stem cells;
q-PCR; quantitative-polymerase chain reaction.
Bmi1 is not expressed in ESC. (A) Gene expression
124 DING ET AL.
controls, as determined by 7-AAD and annexin V staining
(Supplementary Fig. S2B).
Lineage tracing and reporter studies demonstrated that
during ESC differentiation, the hemangioblast is emerging
around day 3 of EB culture and is detected by expression of
the receptor kinase Flk1 (VEGF receptor-2) . He-
mangioblast development is followed by patterning of
specific subsets of mesodermal cells, including hematopoi-
etic, vascular, and cardiac somatic cells [36,37]. To deter-
mine the influence of Bmi1 expression on mesoderm
formation, Bmi1-ESC were subjected to differentiation in EB
assay, and Flk1 expression was examined by flow cytome-
try. Flk1+cells first emerged around day 3 of differentiation
in both control and the Bmi1-ESC (Fig. 3B). The frequency
of Flk1+cells peaked at day 5 (around 40% of Flk1+cells)
and started to decrease from day 6 onward. Thus, the
temporal pattern of Flk1+cell generation is similar between
Bmi1-ESC and ESC controls.
To study the impact of Bmi1 on early stages of develop-
ment, RNA was extracted at various periods of time of EB
culture, and expression of pluripotency genes, mesodermal
and hematopoietic markers was measured by q-PCR. Oct4
expression decreased gradually in both Bmi1-ESC and con-
trols. Importantly, we found a prominent up-regulation of
Gata2, which persisted over the entire period of analysis (day
10) and was not observed in controls (Fig. 3C and Supple-
mentary Fig. S3). Gata2, a zinc finger transcription factor, is
essential for primitive hematopoietic cell generation from
mesoderm [38,39]. Gata1, another member of the same
family, was also up-regulated. However, this up-regulation
was transient and the same for Bmi1-ESC and controls (Fig.
3C and Supplementary Fig. S3).
pression in ESC does not af-
fect self-renewal capacity. (A)
Schematic representation of
Bmi1 lentivirus vector. Bmi1
expression is driven by the
human Ubc. (B) CCE ESC
were transduced with Bmi1
vector or empty vector con-
trol (vector). GFP+cells were
obtained by FACS and are
transduced ESC, gray line.
(C) Immunostaining of Bmi1
in Bmi1-transduced ESC by
Bmi1-specific antibody (red).
Cell nuclei were stained with
DAPI (blue). Size bar, 200mm.
(D) Western blot analysis of
Bmi1 expression in Bmi1-ESC
(plus LIF) and after differen-
tiation in EB assays (day 14).
b-actin is shown as loading
control. (E) RT-PCR analysis
of pluripotency genes (Oct4,
Sox2, and Nanog) in Bmi1-
ESC under growth condi-
tions. GAPDH is shown as
loading control. (F) Expres-
sion of the Bmi1 target genes
p16Ink4a and p19Arf in ESC
as in (E) was determined
by q-PCR analysis. Relative
expression normalized to b-
actin is shown. (G) For cell-
cycle analysis, Bmi1-ESC were
stained with Hoechst 33342
and subjected to analysis by
flow cytometry. Gating for
cells in S phase is indicated
(bar). (H) Proliferation rates
were assessed by seeding 105
cell numbers 2 days later.
Control, untreated ESC; Vec-
tor, empty vector treated ESC. GFP, green fluorescent protein; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; IRES,
internal ribosome entry site; Ubc, ubiquitin C promoter; FACS, fluorescent activated cell sorting; DAPI, 4¢,6-diamidino-2-phe-
nylindole dihydrochloride; RT-PCR, reverse transcriptase-polymerase chain reaction.
BMI1 PROMOTES HEMATOPOIESIS FROM ES CELLS 125
The activation of Brachyury (T), an early mesoderm mar-
ker, showed a similar profile in Bmi1-ESC as controls (Fig.
3C). Scl, a member of the basic helix-loop-helix transcrip-
tional factor family, is involved in mesoderm patterning and
primitive hematopoietic cell generation. Scl and Flk1 are in-
volved in a combinatorial manner in regulating hemangio-
blasts fate . These 2 genes were activated sequentially in
both Bmi1-ESC and controls (Fig. 3C and Supplementary Fig.
S3). Runx1 is expressed in yolk sac mesodermal cells before
the establishment of the blood islands and also in the cor-
responding EB-derived hemangioblasts . The homeobox
gene Hoxb4 is implicated in growth of both embryonic and
adult HSC and is important for engraftment potential of
ESC-derived hematopoietic cells . Runx1 and Hoxb4
displayed a similar pattern of expression for Bmi1-ESC and
control ESC (Fig. 3C).
CD41 represents one of the early markers during hema-
topoiesis in EB assays and in the mouse embryo .
Therefore, to determine frequencies and kinetics of CD41+
c-Kit+hematopoietic precursor cells in differentiating Bmi1-
ESC, EB from day 2 to 10 were dissociated and analyzed by
flow cytometry. There was a prominent increase in CD41+
c-Kit+cells between day 4 and 6 (Fig. 3D). CD41+c-Kit+cells
were then obtained by cell sorting and analyzed for expres-
sion of hematopoietic genes. We observed strong up-regula-
tion of Gata2 in CD41+c-Kit+cells at day 8 and 10 (Fig. 3E).
Taken together, our results demonstrate that forced expres-
sion of Bmi1 allows ordered mesoderm and hemangioblast
Bmi1-ESC in EB assays. (A)
Bmi1-ESC were subjected to
differentiation by EB forma-
tion. Phase-contrast imagines
show EB morphology at days
3 and 6. (B) Kinetics of Flk1+
cell generation during EB for-
mation (days 3–6). Single-cell
stained with anti-Flk1 anti-
body and analyzed by flow
cytometry. The results shown
are the means of 3 indepen-
dent experiments. (C) Kinetics
of gene expression during EB
formation assessed by q-PCR
(days 0–10). Expression was
normalized to b-actin and rel-
ative RNA levels of mesoder-
mal genes (Brachyury, Flk1,
and Scl), hematopoietic genes
(Gata1, Gata2, Hoxb4, and
Runx1), and the pluripotency
gene Oct4 are shown. Nor-
malized expression levels at
day 0 were set to 1.0, and
normalized Oct4 expression at
day 10 was arbitrarily set to
1.0. One representative exper-
iment is shown. (D) Kinetics
of CD41+c-kit+cell genera-
Single-cell suspensions were
stained with CD41 and c-Kit
specific antibodies and ana-
lyzed by flow cytometry. The
results shown are the means
of 3 independent experiments.
(E) Expression of hematopoi-
etic genes in CD41+c-Kit+
cells at days 6, 8, and 10 of EB
were obtained by cell sorting
and analyzed by q-PCR. Ex-
pression values were normal-
ized to b-actin as in (C), and
expression in empty vector-
transduced cells was set to 1.0.
126DING ET AL.
development of ESC. Remarkably, Gata2, a gene expressed
in primitive hematopoietic cells, was efficiently increased by
Transcriptional regulation by Bmi1
in differentiating ESC
To obtain further insights into gene regulation by Bmi1
during differentiation in EB assays, we performed genome-
wide gene expression studies with DNA microarray. RNA
from Bmi1-ESC at day 10 of differentiation was extracted and
subjected to Affymetrix GeneChip arrays, and data were an-
alyzed by hierarchical clustering. Empty vector-transduced
ESC were used as control. We identified 304 genes that were
differentially expressed between both groups: 221 genes were
found to be up-regulated, and 83 genes were down-regulated
by Bmi1 (Supplementary Table 1). Many differentially regu-
lated genes were found to be involved in specific biological
functions, and Supplementary Fig. S4A shows a panel of
transcription factors and development-associated genes.
Several members of Hox gene family of transcription fac-
tors, which play a critical role in development, are known to be
dynamically regulated by Bmi1 or other PcG proteins [43,44].
We find 4 Hox genes, Hoxa3, Hoxa9, Hoxd10, and Hoxa11,
being repressed by Bmi1 (Supplementary Fig. S4A). The zinc-
finger domain transcription factor Zic1, which is required for
neuronal differentiation, was also repressed by Bmi1. This is
consistent with the observation that Zic1 represents a known
target of PcG proteins . Several members of hemoglobin
genes (Hbb-y, Hba-a1, and Hbax), markers for erythroid cell
differentiation, were also down-regulated in differentiating
Bmi1-ESC. However, repression of the Ink4a/Arf locus, which
is a direct target of Bmi1 in adult stem/somatic cells, could not
be detected by microarrays, probably due to very low ex-
pression levels (see also below Fig. 5E). Several genes were up-
regulated in Bmi1-ESC during differentiation (Supplementary
Fig. S4A), including transcription factors, cell structure pro-
teins, membrane receptors, and extracellular matrix proteins
(Supplementary Fig. S4; Supplementary Table 1). As expected,
Gata2 was also found to be up-regulated in Bmi-ESC, which is
in line with the PCR data just described (Fig. 3C, E and Sup-
plementary Fig. S3). Elevated expression of these up-regulated
genes is most likely due to secondary and indirect effects of the
repressive Bmi1 activity.
PCR analysis of the Hox genes Hoxa3, Hoxa9, Hoxd10,
and Hoxa11 and of the Zic1 and Gata2 transcription factors
confirmed the DNA microarray data (Supplementary Fig.
S4B). The myosin genes Myh6 and Myl2 showed elevated
expression in Bmi1-ESC, which is also consistent with mi-
croarray data (Supplementary Fig. S4A and B). The hema-
topoietic genes c-Kit, Hoxb4, and Gata1 and the endothelial
cell markers CD31 and CD144 (VE-Cadherin) remained un-
affected by Bmi1, by both microarray analysis and q-PCR.
Thus, forced Bmi1 expression during EB differentiation en-
tails repression of known Bmi1/PcG protein target genes but
also causes elevated expression of yet another group of
genes, most probably through indirect mechanisms.
Bmi1 promotes primitive hematopoiesis
Given the prominent up-regulation of Gata2 in Bmi1-ESC,
we chose to study the impact of Bmi1 on the generation he-
matopoietic cells from ESC. Primitive hematopoietic CFU,
particularly erythroid and erythroid-myeloid CFU cells, are
first detectable in differentiating ESC at day 5 to 6. Therefore,
single-cell suspensions of day 6 EB were plated in semi-solid
methylcellulose cultures supplemented with hematopoietic
cytokines, and the number and morphology of colonies was
evaluated, including CFU-E (erythroid CFU), CFU-GM (gran-
ulocyte and macrophage CFU), and CFU-GEM (erythrocyte,
granulocyte, and macrophage CFU). Strikingly, Bmi1 en-
hanced the number of hematopoietic CFUs up to 3-fold over
controls (Fig. 4A). Frequencies of CFU-GM and CFU-GEM for
Bmi1-ESC were also increased, whereas the morphology of
Bmi1-ESC and controls were quite similar (Fig. 4A and B).
Therefore, Bmi1 promoted development of hematopoietic cells
from ESC, but there appears to be no prominent impact of
Bmi1 on the development of specific hematopoietic lineages.
Bmi1 leads to robust proliferation of ESC-derived
hematopoietic progenitor cells
To further assess the impact of Bmi1 on hematopoietic cell
development from ESC, we studied hematopoietic progenitor
colony-forming assay of EB-derived cells at day 6 of differ-
entiation. Bmi1-ESC, empty vector ESC, and untreated ESC
control (CCE) were differentiated by EB formation for 6 days,
and 1·105cells were seeded into semi-solid methylcellulose
culture with cytokines (SCF, IL-3, IL-6, and erythropoietin).
After 5 and 10 days of culture, numbers of CFU-E, CFU-GM,
and CFU-GEM were scored by microscopy and are shown.
(B) Phase-contrast imagines were taken from representative
CFU-E, CFU-GM, and CFU-GEM colonies (upper panel);
neutral benzidine-stained isolated colonies (lower panel). A
representative result of 5 independent experiments is shown.
CFU, colony-forming unit.
CFU assay of EB-derived cells. (A) Hematopoietic
BMI1 PROMOTES HEMATOPOIESIS FROM ES CELLS127
cells (HPC) obtained from EB (Fig. 5A). Bmi1-ESC and control
ESC were subjected to differentiation in EB assays. At day 6,
single-cell suspensions were prepared and seeded in serum-
free medium with SCF, Flt3L, hyper-IL-6, IGF-1, and IL-3.
Cumulative cell numbers were calculated to assess the pro-
liferation potential of cultures. Strikingly, Bmi1 markedly
enhanced the proliferation capacity, and Bmi1-ESC-derived
HPC (Bmi1-HPC) were expanded up to nearly 100-fold
within 3 weeks of culture (Fig. 5B). Bmi1-HPC were main-
tained in culture for almost 2 months without loss of prolif-
eration potential and generated a vast number of cells. In
contrast, control ESC-HPC failed to expand and deteriorated
after a few days of culture (Fig. 5B). The low expansion rates
of control ESC-HPC made it difficult to obtain enough cells
for analysis. Bmi1-HPC had a primitive blast-like morphol-
ogy, as revealed by histological staining of cytospins (Fig.
5C), whereas cultures of control ESC-HPC showed dead cells
and cell debris (data not shown).
GEM (Fig. 5D). Bmi1 promotes self-renewal of adult HSC
mainly though repressing the expression of p16Ink4a and
p19Arf. Therefore, we determined expression of Ink4a/Arf in
ESC, day 6 EB, and day 12 ESC-HPC by q-PCR (Fig. 5E). We
found that Bmi1 efficiently repressed expression of p16Ink4a
and p19Arf in Bmi1-HPC. Thus, similar to the activity of Bmi1
on adult stem cells, Bmi1 might confer its growth promoting
The self-renewal of HSC is closely related to their adhesion
properties , and thus, we investigated the interaction of
Bmi1-HPC with stroma cells. HPC in suspension culture
were plated on OP9 stroma cells and cultured with growth
factor as just described. Bmi1-HPC formed typical cobble-
stone areas beneath the stroma cell layer (Supplementary Fig.
S5), whereas control HPC died during culture. Taken to-
gether, these data demonstrated that enforced Bmi1 expres-
sion enhanced hematopoietic cell differentiation, survival,
and caused continuous and extensive proliferation of ESC-
derived HPC in vitro.
matopoietic cell differentia-
tion from ESC. (A) Schematic
representation of the culture
system used for hematopoi-
etic cell differentiation from
ESC. ESC were subjected to
differentiation by EB forma-
tion. Day 6 EB were dis-
aggregated, and EB-derived
cells were plated on gelatin-
coated dishes with hemato-
poietic cell cytokines in se-
rum-free medium. At days
8–9 of culture, nonadherent
cells were harvested and ex-
panded. (B) Growth kinetics
of ESC-derived hematopoietic
cells. Cumulative cell numbers
of 2 independent experiments
is shown. (C) Morphology of
Bmi1-HPC on day 26 of dif-
ferentiation. Cytospin prepa-
from day 12 ESC-HPC. (E)
Expression of p16Ink4a and
ESC, day 6 EB and day 15
expression was examined by
q-PCR and normalized to b-
Bmi1 promotes he-
128DING ET AL.
The Bmi1-HPC are heterogeneous in phenotype
and dependent on hematopoietic cytokines
To characterize the phenotype of Bmi1-HPC, cells were
stained for a panel of hematopoietic cell surface antigens and
analyzed by flow cytometry (Fig. 6A and Supplementary Fig.
S6). More than 90% of the cells expressed c-Kit, probably
because cells were always cultured in the presence of c-Kit
ligand SCF. There was also expression of other stem cell
markers, such as Sca1, CD34, and CD133; however, these
cells represented only 10%–30% of the cell population. A ra-
ther high fraction of cells expressed myeloid lineage markers,
including CD11b, CD14, and Gr1 (30%–45%) and the ery-
throid lineage marker Ter119 (60%). We also examined the
surface phenotype of cells at days 15, 30, and 40 of culture,
and there were no obvious changes over extended culture
periods (data not shown). Thus, Bmi1-HPC are heteroge-
neous in phenotype, which is similar to Hoxb4 ESC-HPC .
To further characterize Bmi1-HPC, we compared the he-
matopoietic gene expression profile of Bmi1-HPC with fetal
liver cells and adult BM cells by RT-PCR (Fig. 6B). Bmi1-HPC
showed high expression of the hematopoietic transcription
factors Gata1, Gata2, and PU.1 and abundantly expressed
the cytokine receptors c-Kit and G-CSF receptor, but not Flt3.
Deregulation of Bmi1 is a common feature in several types
of leukemias, and high Bmi1 expression is frequently found
in primary leukemia cells . This raises the question
whether enforced Bmi1 expression is sufficient to promote
Bmi1-HPC self-renewal or to transform the cells. Therefore,
we tested the cytokine and growth factor response of Bmi1-
found that Bmi1-HPC were critically dependent on growth
factors for survival and proliferation. SCF and IL-3 con-
ferred some proliferative potential, yet multiple factor
combinations were required for maximal proliferation
rates (Fig. 6C). Since survival of the majority of leukemia
cells is aberrantly independent on cytokines signals ,
this finding suggests that forced Bmi1 expression causes
more potent proliferation rates but is not sufficient to
3H-thymidine incorporation assays (Fig. 6C). We
Bmi1 is crucial for maintaining the pool of adult stem cells,
and several types of malignancies are characterized by de-
regulated Bmi1 expression [18–21]. Here, we found that Bmi1
is not expressed in ESC and that forced Bmi1 expression
caused Gata2 up-regulation and enhanced frequencies of
hematopoietic cells derived from ESC. Such hematopoietic cell
derivatives expressing Bmi1 displayed robust proliferation
capacity in vitro. Our results uncover a so-far-not-recognized
activity of Bmi1 in ESC commitment and differentiation, and,
Bmi1-HPC. (A) Bmi1-HPC are
a heterogenous cell popula-
tion. Phenotype of day 20
Bmi1-HPC was analyzed by
flow cytometry. Data rep-
resent the mean of 2 inde-
Expression of hematopoietic
genes in Bmi1-HPC. RNA was
prepared from ESC, day 20
and 40 Bmi1-HPC, embryonic
day 14 fetal liver cells, and
adult BM cells. Expression of
mined by semi-quantitative
RT-PCR. GAPDH was used as
control. (C) Growth factor re-
vitro culture. Bmi1-HPC were
pulsed with3H-thymidine af-
ter 2 days of culture in the
presence of different combi-
nations of growth factors as
BMI1 PROMOTES HEMATOPOIESIS FROM ES CELLS 129
thus, Bmi1 might be a candidate factor to improve hematopoietic
stem/progenitor cell derivation from pluripotent cells.
The epigenetic modifier PcG proteins are important in
maintaining self-renewal in ESC and adult stem cells, in-
cluding HSC, NSC, and endothelial precursor cells. For ex-
ample, Ezh2 and Suz12 are required for both adult stem cells
and pluripotent cells [9,12,13,16,17]. RingA/B is directly in-
volved in the transcriptional network of ESC . Eed and
Bmi1 have antagonistic functions in HSC . Bmi1 is critical
for maintenance of adult stem cells, but as shown here, it is
not expressed or expressed at very low levels in ESC. Further,
we also show that forced Bmi1 expression leaves ESC self-
renewal unaffected. Thus, stem cell self-renewal appears to
require specialized PcG proteins. Understanding the re-
quirements and roles of PcG proteins for different stem cell
types is expected to help in elucidating the genetic programs
that drive stem cell self-renewal and cell fate decisions upon
differentiation. Thus, we set out to analyze the influence of
Bmi1 on differentiation and show that forced Bmi1 expres-
sion enhances hematopoietic cell development from ESC.
The activity of Bmi1 on promoting adult stem cell self-
renewal is, in part, through repressing p16Ink4a and p19Arf,
encoded by the Ink4a/Arf locus . In ESC, the promoter
region of Ink4a/Arf gene contains bivalent chromatin his-
tone H3 methylation marks, and Ink4a/Arf expression is
repressed . We show here that Bmi1 is not expressed in
ESC; therefore, repression of the Ink4a/Arf locus through
H3K27me3 might be performed by alternative PcG proteins.
We also found that exogenous Bmi1 expression in ESC re-
inforced the repressive activity on this locus. However, cell-
cycle analysis revealed that a further decrease in Ink4a/Arf
expression did not affect ESC proliferation.
Expression of the Ink4a/Arf locus remained low during
early stages of ESC differentiation in EB formation assays. At
later stages, Ink4a/Arf expression increased; however, it re-
mained effectively repressed in Bmi1-ESC-derived HPC. The
low expression of Ink4a/Arf gene in ESC and EB might
explain that ecotopic Bmi1 expression did not affect ESC self-
renewal and the early stages of ESC differentiation. Bmi-1-/-
mice displayed normal embryogenesis, as well as normal
numbers of fetal liver HSC . Ink4a/Arf expression was
not detected in fetal or young adult stem cells but increases
in aged stem cells, which are impaired in self-renewal [49,50].
Thus, the repressive activity of Bmi1 on Ink4a/Arf locus
seems to be confined to adult stem cells, whereas in ESC and
during embryo development, repression appears to be due to
other PcG proteins. How PcG dynamics change during de-
velopment is an interesting question. A large number of
genes, which are known to be regulated by PcG proteins, are
altered in expression during ESC differentiation, and this
coincides with dynamic changes of PcG protein regulatory
function [9,12,51]. ESC differentiation recapitulates aspects of
early stages of embryonic development, and, thus, studying
PcG dynamics in differentiating ESC will provide valuable
information on PcG activities during development.
Bmi1 acts as transcriptional repressor , and, thus, its
activity directly translates into changes in gene expression.
Gene expression profiling by microarray represents a pow-
erful tool to survey transcriptional pattern on a genome-wide
scale, and, thus, Bmi1 expressing cells were subjected to
microarray analysis. We identified a large number of genes
that had been repressed by Bmi1, which was expected. For
instance, several Hox genes were down-regulated, but Gata2
was up-regulated. This finding is interesting, as others and
we have shown earlier that Gata2 plays a key role in he-
matopoietic cell proliferation and in maintaining the pool of
adult HSC [31,39,52]. Additionally, doxycycline-induced
Gata2 expression in differentiating EB promoted hemato-
poietic mesoderm formation and suppressed endoderm and
ectodermal lineages .
Up-regulated Gata2 expression in Bmi1 ESC-derived EB is
unlikely to be due to an increased frequency of hematopoietic
cells, as other hematopoietic genes, such as Scl, c-Kit, Runx1,
and Hoxb4, showed no large difference in expression com-
pared with controls. Additionally, Bmi1 over-expression in
adult BM Flt3+HSC  also caused up-regulation of Gata2
hematopoietic precursor were unaffected by Bmi1 expression
and forced Bmi1 expression up-regulated Gata2 in CD41+
c-kit+precursor cells. Thus, elevated Gata2 expression in
CD41+c-kit+precursor cells might account for the enhanced
commitment and/or growth of hematopoieticcell derivatives.
Both Gata2 and Bmi1 are essential for adult HSC, and,
thus, it is tempting to speculate that both are important
components of a circuitry that is important for maintaining
the HSC pool and/or keeping progenitors immature. In
addition, HPC generated from Bmi1-ESC, despite showing
an enhanced proliferation capacity, were fully responsive
and dependent on cytokines for survival in culture. This
suggests that the extended lifespan observed for Bmi1-HPC
is not related to malignant transformation. In initial trans-
plantation experiments using standard protocols , we did
find neither short-term or long-term engraftment of Bmi1-
HPC nor tumor formation (data not shown).
The derivation of tissue-specific stem/progenitor cells and
terminally differentiated cells from ESC for stem cell-based
replacement therapies represents a major challenge in re-
generative medicine. In vitro generation of hematopoietic
cells from ESC has been demonstrated by several studies but
remained rather inefficient compared with other cell lineages
. Additionally, several previous studies demonstrated
that intrinsic regulators for adult HSC, such as Hoxb4, Stat5,
and Runx1, can promote hematopoietic cell development
from ESC [41,56–58]. In particular, Hoxb4 supports the gen-
eration of long-term repopulating HSC from mouse ESC
[41,56]. However, data on a similar activity of Hoxb4 on he-
matopoiesis of human ESC are still inconsistent [59,60]. Thus,
the identification of genes that enhance the hematopoietic cell
development from ESC, as reported here for Bmi1, represents
an important step toward the aim of obtaining high cell
numbers for treatment of hematopoietic diseases.
We would like to thank F. Farnebo (Karolinska Institute,
Stockholm, Sweden) for providing FUGIE vector, H. Klump
(University Hospital Essen, Essen, Germany) for the CCE ES
cell line, U. Just (University of Kiel, Kiel, Germany) for OP9
cells, and Anna M. Wobus (IPK, Gatersleben, Germany) for
LIF. We also thank A. M. Mu ¨ller (University of Wu ¨rzburg,
Wu ¨rzburg, Germany) for critical reading of the article and all
lab members for helpful comments and suggestions on this
work, C. Becker and B. Denecke for help with gene array
hybridization, I. Ortseifer for cell sorting, and A. Offergeld
130DING ET AL.
for expert secretary assistance. Part of this work was sup-
ported by funding within the SPP1356 program to M. Z.
Author Disclosure Statement
The authors have no financial conflict of interest.
1. Fisher AG. (2002). Cellular identity and lineage choice. Nat
Rev Immunol 2:977–982.
2. Jaenisch R and A Bird. (2003). Epigenetic regulation of gene
expression: how the genome integrates intrinsic and envi-
ronmental signals. Nat Genet 33 (Suppl.):245–254.
3. Reik W. (2007). Stability and flexibility of epigenetic gene
regulation in mammalian development. Nature 447:425–432.
4. Ruau D, R Ensenat-Waser, TC Dinger, DS Vallabhapurapu,
A Rolletschek, C Hacker, T Hieronymus, AM Wobus, AM
Muller and M Zenke. (2008). Pluripotency associated genes
are reactivated by chromatin-modifying agents in neuro-
sphere cells. Stem Cells 26:920–926.
5. Valk-Lingbeek ME, SW Bruggeman and M van Lohuizen.
(2004). Stem cells and cancer; the polycomb connection. Cell
6. Sauvageau M and G Sauvageau. (2010). Polycomb group
proteins: multi-faceted regulators of somatic stem cells and
cancer. Cell Stem Cell 7:299–313.
7. Ringrose L and R Paro. (2004). Epigenetic regulation of cel-
lular memory by the Polycomb and Trithorax group pro-
teins. Annu Rev Genet 38:413–443.
8. Bracken AP, N Dietrich, D Pasini, KH Hansen and K Helin.
(2006). Genome-wide mapping of Polycomb target genes un-
9. Boyer LA, K Plath, J Zeitlinger, T Brambrink, LA Medeiros,
TI Lee, SS Levine, M Wernig, A Tajonar, et al. (2006). Poly-
comb complexes repress developmental regulators in mu-
rine embryonic stem cells. Nature 441:349–353.
10. Bernstein BE, TS Mikkelsen, X Xie, M Kamal, DJ Huebert, J
Cuff, B Fry, A Meissner, M Wernig, et al. (2006). A bivalent
chromatin structure marks key developmental genes in
embryonic stem cells. Cell 125:315–326.
11. Azuara V, P Perry, S Sauer, M Spivakov, HF Jorgensen, RM
John, M Gouti, M Casanova, G Warnes, M Merkenschlager
and AG Fisher. (2006). Chromatin signatures of pluripotent
cell lines. Nat Cell Biol 8:532–538.
12. Lee TI, RG Jenner, LA Boyer, MG Guenther, SS Levine, RM
Kumar, B Chevalier, SE Johnstone, MF Cole, et al. (2006).
Control of developmental regulators by Polycomb in human
embryonic stem cells. Cell 125:301–313.
13. O’Carroll D, S Erhardt, M Pagani, SC Barton, MA Surani and T
Jenuwein. (2001). The polycomb-group gene Ezh2 is required
for early mouse development. Mol Cell Biol 21:4330–4336.
AP Otte, M Okano, N Brockdorff, M Vidal and H Koseki.
(2008). Polycomb group proteins Ring1A/B are functionally
linked to the core transcriptional regulatory circuitry to
maintain ES cell identity. Development 135:1513–1524.
15. Cui K, C Zang, TY Roh, DE Schones, RW Childs, W Peng
and K Zhao. (2009). Chromatin signatures in multipotent
human hematopoietic stem cells indicate the fate of bivalent
genes during differentiation. Cell Stem Cell 4:80–93.
16. Ezhkova E, HA Pasolli, JS Parker, N Stokes, IH Su, G Han-
non, A Tarakhovsky and E Fuchs. (2009). Ezh2 orchestrates
gene expression for the stepwise differentiation of tissue-
specific stem cells. Cell 136:1122–1135.
17. Majewski IJ, ME Blewitt, CA de Graaf, EJ McManus, M
Bahlo, AA Hilton, CD Hyland, GK Smyth, JE Corbin, et al.
(2008). Polycomb repressive complex 2 (PRC2) restricts he-
matopoietic stem cell activity. PLoS Biol 6:e93.
18. Molofsky AV, S He, M Bydon, SJ Morrison and R Pardal.
(2005). Bmi-1 promotes neural stem cell self-renewal and
neural development but not mouse growth and survival by
repressing the p16Ink4a and p19Arf senescence pathways.
Genes Dev 19:1432–1437.
19. Lessard J and G Sauvageau. (2003). Bmi-1 determines the
proliferative capacity of normal and leukaemic stem cells.
20. Park IK, D Qian, M Kiel, MW Becker, M Pihalja, IL Weiss-
man, SJ Morrison and MF Clarke. (2003). Bmi-1 is required
for maintenance of adult self-renewing haematopoietic stem
cells. Nature 423:302–305.
21. Liu S, G Dontu, ID Mantle, S Patel, NS Ahn, KW Jackson, P
Suri and MS Wicha. (2006). Hedgehog signaling and Bmi-1
regulate self-renewal of normal and malignant human
mammary stem cells. Cancer Res 66:6063–6071.
22. Oguro H, J Yuan, H Ichikawa, T Ikawa, S Yamazaki, H
Kawamoto, H Nakauchi and A Iwama. (2010). Poised line-
age specification in multipotential hematopoietic stem and
progenitor cells by the polycomb protein Bmi1. Cell Stem
23. Jacobs JJ, K Kieboom, S Marino, RA DePinho and M van
Lohuizen. (1999). The oncogene and Polycomb-group gene
bmi-1 regulates cell proliferation and senescence through the
ink4a locus. Nature 397:164–168.
24. Akala OO, IK Park, D Qian, M Pihalja, MW Becker and MF
Clarke. (2008). Long-term haematopoietic reconstitution
by Trp53-/-p16Ink4a-/-p19Arf-/- multipotent progenitors.
25. Hosen N, T Yamane, M Muijtjens, K Pham, MF Clarke and
IL Weissman. (2007). Bmi-1-green fluorescent protein-knock-
in mice reveal the dynamic regulation of bmi-1 expression in
normal and leukemic hematopoietic cells. Stem Cells
26. Iwama A, H Oguro, M Negishi, Y Kato, Y Morita, H Tsukui,
H Ema, T Kamijo, Y Katoh-Fukui, et al. (2004). Enhanced
self-renewal of hematopoietic stem cells mediated by the
polycomb gene product Bmi-1. Immunity 21:843–851.
27. Rizo A, B Dontje, E Vellenga, G de Haan and JJ Schuringa.
(2008). Long-term maintenance of human hematopoietic
stem/progenitor cells by expression of BMI1. Blood 111:
28. Ensenat-Waser R, ASN Vicente-Salar, JC Cigudosa, E Roche,
B Soria and JA Reig. (2006). Isolation and characterization of
residual undifferentiated mouse embryonic stem cells from
embryoid body cultures by fluorescence tracking. In Vitro
Cell Dev Biol Animal 42:115–123.
29. Carotta S, S Pilat, A Mairhofer, U Schmidt, H Dolznig, P
Steinlein and H Beug. (2004). Directed differentiation and
mass cultivation of pure erythroid progenitors from mouse
embryonic stem cells. Blood 104:1873–1880.
30. Fischer M, J Goldschmitt, C Peschel, JP Brakenhoff, KJ Kal-
len, A Wollmer, J Grotzinger and S Rose-John. (1997). I. A
bioactive designer cytokine for human hematopoietic pro-
genitor cell expansion. Nat Biotechnol 15:142–145.
31. Briegel K, KC Lim, C Plank, H Beug, JD Engel and M
Zenke. (1993). Ectopic expression of a conditional GATA-
2/estrogen receptor chimera arrests erythroid differen-
tiation in a hormone-dependent manner. Genes Dev 7:
BMI1 PROMOTES HEMATOPOIESIS FROM ES CELLS131
32. Young RA. (2011). Control of the embryonic stem cell state.
33. Kim JB, H Zaehres, G Wu, L Gentile, K Ko, V Sebastiano, MJ
Arauzo-Bravo, D Ruau, DW Han, M Zenke and HR Scholer.
(2008). Pluripotent stem cells induced from adult neural
stem cells by reprogramming with two factors. Nature
34. Ohm JE, KM McGarvey, X Yu, L Cheng, KE Schuebel, L
Cope, HP Mohammad, W Chen, VC Daniel, et al. (2007). A
stem cell-like chromatin pattern may predispose tumor
suppressor genes to DNA hypermethylation and heritable
silencing. Nat Genet 39:237–242.
35. Keller G. (2005). Embryonic stem cell differentiation: emer-
gence of a new era in biology and medicine. Genes Dev
36. Chung YS, WJ Zhang, E Arentson, PD Kingsley, J Palis and K
by FLK1 and SCL expression. Development 129:5511–5520.
37. Ema M, P Faloon, WJ Zhang, M Hirashima, T Reid, WL
Stanford, S Orkin, K Choi and J Rossant. (2003). Combina-
torial effects of Flk1 and Tal1 on vascular and hematopoietic
development in the mouse. Genes Dev 17:380–393.
38. Cantor AB and SH Orkin. (2001). Hematopoietic develop-
ment: a balancing act. Curr Opin Genet Dev 11:513–519.
39. Tsai FY, G Keller, FC Kuo, M Weiss, J Chen, M Rosenblatt,
FW Alt and SH Orkin. (1994). An early haematopoietic de-
fect in mice lacking the transcription factor GATA-2. Nature
40. Lacaud G, V Kouskoff, A Trumble, S Schwantz and G Keller.
(2004). Haploinsufficiency of Runx1 results in the accelera-
tion of mesodermal development and hemangioblast speci-
fication upon in vitro differentiation of ES cells. Blood
41. Kyba M, RC Perlingeiro and GQ Daley. (2002). HoxB4
confers definitive lymphoid-myeloid engraftment potential
on embryonic stem cell and yolk sac hematopoietic pro-
genitors. Cell 109:29–37.
42. Mikkola HK, Y Fujiwara, TM Schlaeger, D Traver and SH
Orkin. (2003). Expression of CD41 marks the initiation of
definitive hematopoiesis in the mouse embryo. Blood
43. Simon J, A Chiang and W Bender. (1992). Ten different Poly-
comb group genes are required for spatial control of the abdA
and AbdB homeotic products. Development 114:493–505.
44. van der Lugt NM, M Alkema, A Berns and J Deschamps.
(1996). The Polycomb-group homolog Bmi-1 is a regulator of
murine Hox gene expression. Mech Dev 58:153–164.
45. Wagner W, R Saffrich, U Wirkner, V Eckstein, J Blake, A
Ansorge, C Schwager, F Wein, K Miesala, W Ansorge and
AD Ho. (2005). Hematopoietic progenitor cells and cellular
microenvironment: behavioral and molecular changes upon
interaction. Stem Cells 23:1180–1191.
46. McKinney-Freeman SL, O Naveiras, F Yates, S Loewer, M
Philitas, M Curran, PJ Park and GQ Daley. (2009). Surface
antigen phenotypes of hematopoietic stem cells from em-
bryos and murine embryonic stem cells. Blood 114:268–278.
47. Van Etten RA. (2007). Aberrant cytokine signaling in leu-
kemia. Oncogene 26:6738–6749.
48. Lessard J, A Schumacher, U Thorsteinsdottir, M van Lo-
huizen, T Magnuson and G Sauvageau. (1999). Functional
antagonism of the Polycomb-Group genes eed and Bmi1 in
hemopoietic cell proliferation. Genes Dev 13:2691–2703.
49. Janzen V, R Forkert, HE Fleming, Y Saito, MT Waring, DM
Dombkowski, T Cheng, RA DePinho, NE Sharpless and DT
Scadden. (2006). Stem-cell ageing modified by the cyclin-
dependent kinase inhibitor p16INK4a. Nature 443:421–426.
50. Krishnamurthy J, MR Ramsey, KL Ligon, C Torrice, A Koh,
S Bonner-Weir and NE Sharpless. (2006). p16INK4a induces
an age-dependent decline in islet regenerative potential.
51. Ren X, C Vincenz and TK Kerppola. (2008). Changes in the
distributions and dynamics of polycomb repressive com-
plexes during embryonic stem cell differentiation. Mol Cell
52. Akala OO and MF Clarke. (2006). Hematopoietic stem cell
self-renewal. Curr Opin Genet Dev 16:496–501.
53. Lugus JJ, YS Chung, JC Mills, SI Kim, J Grass, M Kyba, JM
Doherty, EH Bresnick and K Choi. (2007). GATA2 functions
at multiple steps in hemangioblast development and dif-
ferentiation. Development 134:393–405.
54. Felker P, K Sere, Q Lin, C Becker, M Hristov, T Hieronymus
and M Zenke. (2010). TGF-b1 accelerates dendritic cell dif-
ferentiation from common dendritic cell progenitors and
directs subset specification toward conventional dendritic
Cells. J Immunol. 185: 5326–5335.
55. Olsen AL, DL Stachura and MJ Weiss. (2006). Designer
blood: creating hematopoietic lineages from embryonic stem
cells. Blood 107:1265–1275.
56. Pilat S, S Carotta, B Schiedlmeier, K Kamino, A Mairhofer, E
Will, U Modlich, P Steinlein, W Ostertag, et al. (2005).
HOXB4 enforces equivalent fates of ES-cell-derived and
adult hematopoietic cells. Proc Natl Acad Sci USA 102:
57. Schuringa JJ, K Wu, G Morrone and MA Moore. (2004).
Enforced activation of STAT5A facilitates the generation
of embryonic stem-derived hematopoietic stem cells that
contribute to hematopoiesis in vivo. Stem Cells 22:1191–
58. Sakai E, K Kitajima, A Sato and T Nakano. (2009). Increase of
hematopoietic progenitor and suppression of endothelial
gene expression by Runx1 expression during in vitro ES
differentiation. Exp Hematol 37:334–345.
59. Wang L, P Menendez, F Shojaei, L Li, F Mazurier, JE Dick, C
Cerdan, K Levac and M Bhatia. (2005). Generation of he-
matopoietic repopulating cells from human embryonic stem
cells independent of ectopic HOXB4 expression. J Exp Med
60. Unger C, E Karner, A Treschow, B Stellan, U Felldin, H
Concha, M Wendel, O Hovatta, A Aints, L Ahrlund-Richter
and MS Dilber. (2008). Lentiviral-mediated HoxB4 expres-
sion in human embryonic stem cells initiates early hemato-
poiesis in a dose-dependent manner but does not promote
myeloid differentiation. Stem Cells 26:2455–2466.
Address correspondence to:
Prof. Martin Zenke
Department of Cell Biology
Institute for Biomedical Engineering
RWTH Aachen University Medical School
Received for publication November 26, 2010
Accepted after revision April 28, 2011
Prepublished on Liebert Instant Online May 5, 2011
132 DING ET AL.
This article has been cited by:
1.P-Y Tung, P S Knoepfler. 2015. Epigenetic mechanisms of tumorigenicity manifesting in stem cells. Oncogene 34, 2288-2296.
2.D. Zdzieblo, X. Li, Q. Lin, M. Zenke, D. J. Illich, M. Becker, A. M. Müller. 2014. Pcgf6, a Polycomb Group Protein, Regulates
Mesodermal Lineage Differentiation in Murine ESCs and Functions in iPS Reprogramming. STEM CELLS 32:10.1002/
stem.v32.12, 3112-3125. [CrossRef]
3.Lorenzo Rinaldi, Salvador Aznar Benitah. 2014. Epigenetic regulation of adult stem cell function. FEBS Journal n/a-n/a.
4.Ding Xiaolei, Wang Xiaoying, Sontag Stephanie, Qin Jie, Wanek Paul, Lin Qiong, Zenke Martin. 2014. The Polycomb Protein
Ezh2 Impacts on Induced Pluripotent Stem Cell Generation. Stem Cells and Development 23:9, 931-940. [Abstract] [Full Text
HTML] [Full Text PDF] [Full Text PDF with Links] [Supplemental Material]
5.Sean M. Cullen, Allison Mayle, Lara Rossi, Margaret A. GoodellHematopoietic Stem Cell Development 39-75. [CrossRef]
6.Lorella Vecchio, Paul Faustin Seke Etet, Maulilio John Kipanyula, Mauro Krampera, Armel Hervé Nwabo Kamdje. 2013.
Importance of epigenetic changes in cancer etiology, pathogenesis, clinical profiling, and treatment: What can be learned from
hematologic malignancies?. Biochimica et Biophysica Acta (BBA) - Reviews on Cancer 1836, 90-104. [CrossRef]
7.Carlos Luzzani, Claudia Solari, Noelia Losino, Waisman Ariel, Leonardo Romorini, Carolina Bluguermann, Gustavo Sevlever,
Lino Barañao, Santiago Miriuka, Alejandra Guberman. 2011. Modulation of chromatin modifying factors’ gene expression in
embryonic and induced pluripotent stem cells. Biochemical and Biophysical Research Communications 410, 816-822. [CrossRef]