Vol.2, No.4, 163-175 (2012) Stem Cell Discovery
http://dx.doi.org/10.4236/scd.2012.24021
Bone morphogenetic protein-4 affects both
trophoblast and non-trophoblast lineage-associated
gene expression in human embryonic stem cells
Margaret L. Shirley1,2*, Alison V enable1*, Raj R. Rao3, Nolan L. Boyd4, Steven L. Stice1,5,6,
David Puett1#, Prema Narayan7#
1Department of Biochemistry and Molecular Biology, University of Georgia, Athens, USA;
#Corresponding Author:
2Department of Psychiatry, University of California, San Francisco, USA
3Department of Chemical and Life Science Engineering, School of Engineering, Virginia Commonwealth University, Richmond, USA
4Cardiovascular Innovation Institute, University of Louisville, Louisville, USA
5Regenerative Bioscience Center, University of Georgia, Athens, USA
6Department of Animal and Dairy Sciences, University of Georgia, Athens, USA
7Department of Physiology, Southern Illinois University School of Medicine, Carbondale, USA;
#Corresponding Author:
Received 5 May 2012; revised 4 June 2012; accepted 1 July 2012
ABSTRACT
Human embryonic stem cells (hESC) can be in-
duced to differentiate to trophoblast by bone
morphogenetic proteins (BMPs) and by aggre-
gation to form embryoid bodies (EB), but there
are many differences and controversies regard-
ing the nature of the differentiated cells. Our
goals herein were to determine if BG02 cells
form trophoblast-like cells (a) in the presence of
BMP4-plus-basic fibroblast growth factor (FGF-2)
and (b) upon EB formation, and (c) whether the
BMP4 anta gonist noggin elicits direct effects on
gene expression and hormone production in the
cells. Transcriptome profiling of hESC incubated
with BMP4/FGF-2 showed a down-regulation of
pluripotency-associated genes, an up-regula tion
of trophoblast-associated genes, and either a
down-regulation or no change in gene expres-
sion for many markers of the three embryonic
germ layers. Yet, there was up-regulation of
several genes associated with mesoderm, ec-
toderm, and endoderm, strongly suggesting that
differentiation to trophoblast-like cells under the
conditions used does not yield a homogeneous
cell type. Several genes, heretofore unreported,
were identified that are altered in hESC in re-
sponse to BMP4-mediated differentiation. The
production of human chorionic gonadotropin
(hCG), progesterone, and estradiol in the dif-
ferentiated cells confirmed that trophoblast-like
cells were obtained. Gene expression by EB
was characterized by an up-regulation of a num-
ber of genes associated with trophoblast, ecto-
derm, endoderm, and mesoderm, and the pro-
duction of hCG and prog esterone confirmed that
trophoblast-like cells were formed. These re-
sults suggest that, in the presence of FGF-2,
BG02 cells respond to BMP4 to yield tropho-
blast-like cells, which are also obtained upon EB
formation. Thus, BMP4-mediated differentiation
of hESC represents a viable cell system for
studying early developmental events post-im-
plantation; however, up-regulation of non-tro-
phoblast genes suggests a somewhat diverse
response to BMP4/FGF-2. Noggin altered the
transcription of a limited number of genes but,
not surprisingly, did not lead to secretion of
hormones.
Keywords: Human Embryonic Stem Cells;
Trophoblasts; Bone Morphogenetic Prot ein-4;
Embryoid Bodies; Noggin
1. INTRODUCTION
In the human blastocyst, the first step of differentiation
from the morula, composed of totipotent cells, yields the
inner cell mass and trophoblast. The former differentiates
into the hypoblast, leading to the extraembryonic endo-
derm and the epiblast, that differentiates to give the am-
niotic ectoderm and the primitive ectoderm. The primi-
tive ectoderm, in turn, differentiates into the embryonic
*These authors contributed equally to this work.

M. L. Shirley et al. / Stem Cell Discovery 2 (2012) 163-175
164
ectoderm and the primitive streak, the latter giving rise to
extraembryonic mesoderm, primitive mesoderm, and em-
bryonic endoderm. The trophectoderm is composed of
epithelial cells and differentiates into several lineages.
The trophoblast precursor cells begin rapid proliferation
into cytotrophoblasts, with some fusing to form multinu-
cleate syncytiotrophoblasts, a terminally differentiated
trophoblast cell; in addition, villous and extravillous cy-
totrophoblasts are formed [1-3]. One of the hallmarks of
trophoblast differentiation is the production of the hete-
rodimeric glycoprotein hormone, human chorionic gona-
dotropin (hCG), that is secreted throughout gestation,
being essential for progesterone production by the corpus
luteum in the first trimester of human pregnancy [4,5].
There is a paucity of adequate cell models for studying
early trophoblast development, and a reliable system will
greatly facilitate progress in the area of human reproduc-
tion.
Recent studies have shown that human embryonic
stem cells (hESCs) can undergo differentiation into tro-
phoblast-like cells spontaneously from colonies [6],
growth of the cells beyond confluence [7], or embryoid
bodies (EBs) [3,8-12]. Similarly, trophoblast-like cells
can be obtained from hESCs via induced differentiation
from the bone morphogenetic proteins (BMPs) 2, 4, and
7 [13-24]. Fortunately, timely reviews are available [25-
28]. Others have reported that RNAi-mediated knock-
down of Oct4 in hESCs can also lead to trophoblast-like
cells [29,30]. The BMPs, members of the transforming
growth factor-b (TGFb) superfamily, function to regulate
many aspects of development which act by binding to
cell surface serine/threonine kinase receptors that, in turn,
phosphorylate particular Smads, thus enabling them to
enter the nucleus and act as transcriptional regulators
[31-34].
These cellular models for trophectoderm differentia-
tion begin to fill a long-awaited need for new systems to
study early development, particularly in view of the ma-
jor differences between human and mouse trophoblast
[35-37]. Yet, the characteristics of the human trophoblast
cells obtained by BMP-mediated differentiation vary,
sometimes quite significantly, depending upon the cell
type, culture conditions (including the presence or ab-
sence of growth factors, particularly FGF2), and mode of
differentiation [10,18,38]. Moreover, in a number of in-
stances there are considerable differences and controver-
sies regarding the product(s) of differentiation obtained
with BMP. For example, BMP4-mediated differentiation
was found to yield little evidence of trophoblast-like
cells under standard culture conditions [38]. Another stu-
dy found that 14 genes were highly up-regulated as de-
termined by microarrays, in addition to many others [21],
while another report did not identify six of these particu-
lar genes [17]. Also, differences in cell morphology and
gene markers have been noted in BMP-treated hESCs
[18,21]. Explanations for many of these controversial
reports were recently provided by two groups. Yu et al.
[22] reported that bFGF (FGF-2), which acts via the
MEK-ERK pathway, redirects BMP4-mediated differen-
tiation from trophectoderm to mesendoderm as judged by
the expression of brachyury. The activated MEK-ERK
pathway sustains NANOG expression leading to a
bFGF-independent induction of mesendoderm by BMP4.
Another recent paper concluded that in the cooperative
presence of FGF2, BMP4 (that acts through brachyury
and CDX2) leads to the induction of mesoderm, not tro-
phoblast [39].
The present study was undertaken with the goals of: 1)
extending the previous reports that were based on a va-
riety of cell lines, different culture conditions, and at
times conflicting results, and 2) to critically examine the
trophoblast and non-trophoblast BMP-regulated genes,
particularly since a number of the non-trophoblast BMP-
regulated genes have been noted in the earlier studies.
We chose to use the hESC line, BG02 (karyotype 46,
XY), that has not been thoroughly studied in hESC dif-
ferentiation to trophoblast. This cell line was established
from a 6-day embryo with an embryo grade of 3CC [7].
Herein, BMP4, in the presence of FGF2, was incubated
with adherent colonies of BG02 cells to initiate tro-
phoblast differentiation. Identical studies were also done
in the presence of the BMP4 antagonist noggin, since it
has been shown that noggin alters hESC gene expression
and morphology, perhaps leading to differentiation to-
ward early neuroectoderm [18]. In addition, trophoblast
differentiation was initiated by formation of EBs. Parti-
cular emphasis was placed on quantitative profiling of a
variety of genes via qRT-PCR, and measurements were
made to determine hCG and steroid hormone production
by the cells and EBs. Our results have many similarities
with the findings of others, including for example the
down-regulation of pluripotency genes, the up-regulation
of trophoblast genes, and placental hormone production,
but there are also notable differences, particularly with
the up-regulation of certain genes associated with the three
primary germ layers.
2. MATERIALS AND METHODS
2.1. Maintenance of Undifferentiated Cells
In order to ensure pluripotency, the BG02 cells, ob-
tained with proper authorization from Bresagen, Athens,
GA, were manually passaged every 2 - 3 days onto mito-
tically-inactivated mouse embryonic feeder (MEF) layers
derived from E13.5 mouse fetuses. The MEF layer was
removed from the hESC colony, which was then gently
dispersed with the colony pieces being transferred to
another 10 cm MEF-containing plate and treated with

M. L. Shirley et al. / Stem Cell Discovery 2 (2012) 163-175 165
hESC culture medium: 77% Dulbecco’s Modified Eagle
Medium (DMEM/F12; Gibco, Carlsbad, CA) supple-
mented with 15% fetal bovine serum (Hyclone, Logan,
UT); 5% knockout serum replacement, 1% non-essential
amino acids, 1% penicillin/streptomycin, and 1 mM
L-glutamine (all from Gibco); 0.1 mM
β -mercaptoeth-
anol and 4 ng/mL basic fibroblast growth factor (FGF2;
Sigma, St. Louis, MO); and 10 ng/mL leukemia inhibi-
tory factor (LIF; Chemicon, Temecula, CA). Cells were
passaged every 3 - 4 days. Two days after passage the
medium was aspirated and replaced daily.
2.2. BMP4-Mediated Differentiation
The cells were passaged by gentle enzymatic digestion
using cell dissociation buffer (Gibco) into 10 cm Ma-
trigel-coated dishes (BD Bioscience, Boca Raton, FL).
The BG02 cells were cultured in 50% DMEM/F12 me-
dium that was conditioned by MEF layers [21] and then
supplemented with 4 ng/mL FGF2. Experimental groups
included incubation with 100 ng/mL BMP-4 (Quest Di-
agnostics, Lyndberg, NJ) and with 250 ng/mL noggin
(Quest Diagnostics), with untreated hESC serving as
controls. The medium was collected each day for analy-
sis of secreted hormones. On day 7 the cells were har-
vested and quick-frozen for RNA extraction
2.3. Formation of Embryoid Bodies
Colonies were sliced into small pieces, then removed
gently from the MEF layer where they were allowed to
aggregate randomly in suspension on agarose plates. EBs
were grown on agarose dishes in 12 mL of the hESC
medium described above. Each culture began with ~50
EBs and ended with ~12 EBs of varying sizes. The loss
of EB was attributed to aggregation of the individual
units and/or to atresia/necrosis. On alternate days the
culture plates were swirled to aggregate the EB with 6
mL of medium being removed and frozen. The same vol-
ume of fresh medium was added with the EBs then dis-
persed to prevent clumping.
2.4. Hormone Assays
Media collected from cell cultures and EBs were ana-
lyzed for secretion of three placental hormones namely
hCG (the assay recognizes both heterodimer and hCG
β ),
progesterone, and estradiol using immunofluorescence-
based assays. The hormones were measured with an Im-
mulite 1000 with a tri-level internal control in human
serum, Con6, being used to standardize all kits (Diagnos-
tic Product Corporation, Los Angeles, CA).
2.5. qRT-PCR
Total RNA was isolated from BG02 (day 0, i.e. control,
and incubated for 7 days) and from EBs at days 5, 22,
and 50 of culture. hESCs were resuspended in 1 mL Tri-
zol (Invitrogen, Carlsbad, CA) and triturated until ho-
mogenized. The integrity of isolated RNA isolated from
the homogenates (Trizol, Molecular Research Corpora-
tion, Albany, NY) was verified and quantified using a
RNA 600 Nano Assay (Agilent Technologies, Foster City,
CA) and the Agilent 2100 Bioanalyzer. The cDNA Ar-
chive Kit (Applied Biosystems, Inc., Foster City, CA)
was used to reverse transcribe 5 μg total RNA with the
MultiScribe Reverse Transcriptase. Initially, reactions
were incubated at 25˚C for 10 min and subsequently at
37˚C for 120 min. Quantitative PCR (Taqman) assays
were selected for the transcripts to be evaluated from
Assays-On-Demand (Applied Biosystems, Inc.) and in-
corporated into 384-well Micro-Fluidic Cards. The cDNA
samples and 50 μl of GeneAmp Fast PCR master mix
(2×) (Applied Biosystems, Inc.) were loaded in duplicate
into respective channels on each microfluidic card and
briefly centrifuged.
qRT-PCR and relative quantification were performed
with the ABI PRISM 7900 Sequence Detection System
(Applied Biosystems, Inc.), with the expression levels of
all genes analyzed given relative to 18S rRNA expres-
sion. The results for differential expression between the
treated and control samples were expressed as means and
data with a Ct value greater than 35 were not analyzed.
The qRT-PCR data on gene expression of BMP4-treated
and noggin-treated hESCs on day 7 of culture were ana-
lyzed as ddCt relative to day 0, i.e. undifferentiated
hESCs. Data were collected for EBs at days 5, 22, and 50
of culture, in all cases using 18S rRNA as the internal
reference gene. The values of ddCt on days 22 and 50 are
expressed relative to day 5. In the equations below, std =
internal gene standard (18S rRNA), goi = gene of interest
under experimental conditions, and con = gene of interest
under control conditions. It is assumed that the amplify-
cation efficiency is identical under all conditions to give
the normalized fold-change of the mRNA of interest in
treated cells (experimental) relative to that of control cells
[40,41].
()() ()
tt
goigoi std
dCC C=−
t
)() ()
ttt
concon std
dCC C=− ,
(
() ()
tt t
goi con
ddC dCdC=−, Fold-change =
t
ddc
2−
As defined, ddCt is negative if the gene of interest un-
der experimental conditions, i.e. BG02 cells plus BMP4
or plus noggin on day 7, is expressed at a higher level
than the same gene under control conditions, i.e. un-
treated cells at day 0.
2.6. Data Analysis
Most of the PCR experiments were performed in trip-

M. L. Shirley et al. / Stem Cell Discovery 2 (2012) 163-175
166
licate, although in a few cases n was 2, 4, 5, or 6. Hor-
mone measurements were done in triplicate, and the re-
sults are given as mean ± SEM. Data exceeding a fold-
change of 2.0 (up-regulated) or –2.0 (down-regulated)
were analyzed using a one-way analysis of variance
(ANOVA) with the GraphPad Prism software. Results
are given for changes with P < 0.05 and P < 0.10. Al-
though the mean fold-change was large in some cases,
significance was at times not reached, often due to n = 2
or to one value in three being particularly different from
the others. Outliers were identified using the Grubbs’ test
(on-line GraphPad software). From over 400 hormone
measurements, only about 1% was deemed outliers, and
from over 700 qRT-PCR runs, only six and seven outliers
were identified in the cell and EB data, respectively.
3. RESULTS
3.1. Effects of BMP4 and Noggin on Gene
Expression and Hormone Production in
hESC
Gene expression: The selection of 177 genes to inves-
tigate via qRT-PCR was based on several criteria. For
example, it was important to include many that were
established markers of pluripotency, trophoblast, ecto-
derm, endoderm, and mesoderm. In addition to these
standard marker genes, others were also screened. In-
cluded in this list were genes encoding certain steroido-
genic enzymes even if not specific for trophoblast and
placenta, e.g. cytochrome P450 side-chain cleavage en-
zyme (CYP11A1) and aromatase (CYP19A1). Other genes
encoding proteins for extracellular matrix and various
aspects of cell function were monitored.
A heatmap, based on mean values of and de-
picting gene expression of 146 genes in hESCs incubated
with either BMP4 or with noggin for 7 days, both rela-
tive to control cells, is shown in Figure 1. Ta b l e I pro-
vides more information on many of the genes shown in
Figure 1 that are altered by 2-fold or more in response to
BMP4. It can be seen that a number of genes are up-
regulated by BMP4, a smaller number down-regulated,
and a much larger number exhibit no major change in
expression.
t
ddc
2−
BMP4 leads to down-regulation of the pluripotency-
associated genes: DNMT3B, EBAF, FGF2, FGF4, FGFR4,
FOXH1, GABRB3, GBX2, LDB2, POU5F1, and SOX2,
and an up-regulation of genes associated with differen-
tiation to trophoblast or involved in implantation and/or
placenta function: CGB, CYP11A1, CYP19A1, ENPEP,
EPAS1, GATA2, GATA3, HEY1, INHA, KRT7, MMP9,
MSX2, PGF, and WNT5A, genes that have also been
identified by others (cf. [17,21] and references therein).
Moreover, a comparison of BMP4-mediated hESC dif-
ferentiation with human trophectoderm [42] identified
Figure 1. Heatmap showing the alterations in gene
expression in hESC after seven days of incubation
with either BMP4 (B) or noggin (N) relative to con-
trol. The data reflect qRT-PCR results given as mean
values of . The gray bars denote sufficiently
low expression that could not be accurately deter-
mined, i.e. Ct greater than 35.
t
ddc
2−

M. L. Shirley et al. / Stem Cell Discovery 2 (2012) 163-175 167
Table 1. BMP4-mediated changes in gene expression of
hESCsa.
Gene Fold-change Description
Trophoblast
CGB 44.9b Chorionic gonadotropin-b
CYP11 A1 51.5c Cytochrome P450 family 11,
subfamily A
CYP19A1 8.9b Cytochrome P450 family 19,
subfamily A
ENPEP 244.6b Glutamyl aminopeptidase
EPAS1 185.6 d Endothelial PAS domain protein 1
GATA2 200.2 d GATA binding protein 2
GATA3 336.2 d GATA binding protein 3
HEY1 18.5b Hairy enhancement of split related with
YRPW motif
INHA 89.9c Inhibin a
KRT7 242.2c Keratin 7
MMP9 8.6c Matrix metalloproteinase 9
MSX2 174.1d Msh homeobox 2
PGF 26.5d Placental growth factor
WNT5A 101.8d Wingless-type MMTV integration site
family 5A
Pluripotency
DNMT3B −3.7b DNA (cytosine-5)-methlytransferase 3 b
EBAF −4.9b Endometrial bleeding associated factor
FGF2 −5.3b Fibroblast growth factor 2
FGF4 −4.5b Fibroblast growth factor 4
FGFR4 −6.2d Fibroblast growth factor receptor 4
FOXH1 −9.1d Forkhead box protein H1
GBX2 −3.5b Homeobox protein GBX2
POU5F1 −4.8b POU class 5 homoebox 1
SOX2 −26.8d SRY (sex determining region Y)-box 2
Mesoderm
BMP4 24.1d Bone morphogenetic protein 4
BMPR2 4.4d Bone morphogenetic protein receptor,
type II
CDH5 150.2b Cadherin 5, type 2
CDH11 7.1d Cadherin 1, type 2
CHRD 9.1d Chordin
GATA4 13.7b GATA binding protein 4
GJA1 –3.4d Connexin 43, gap junction a3
KDR –3.4b Kinase insert domain receptor
NKX2-5 15.5b Homeobox protein Nkx 2.5
PITX2 12.4d Paired-like homeodomain transcription
factor 2
RUNX1 60.9d Runt-related transcription factor 1
RUNX2 14.3b Runt-related transcription factor 2
SOX9 5.8c SRY (sex determining region Y)-box 9
Continued
T 38.1d Brachyury
TBX5 7.6c T box transcription factor
TIE −11. 8 b Receptor tyrosine kinase
VEGF 13.2b Vascular endothelial growth factor
Ectoderm
EN1 2.7b Engrailed homeobox 1
FN1 68.6c Fibronectin 1
HOXB4 24.5c Homeobox protein B4
Gene Fold-change Description
Ectoderm
MMP2 10.8d Matrix metalloproteinase 2
MSI2 2.6b Musashi homolog 2
MYL4 5.3b Myosin light chain 4
NOG 44.4c Noggin
SOX3 −5.2d SRY (sex determining region Y)-box 3
Endoderm
GATA6 7.3b GATA binding protein 6
HNF4A 3.5b Hepatocyte nuclear factor 4 a
NODAL −2.0b Nodal
Others
AKT1 4.3d Protein kinase B, PKB b
CEBPB 4.3c CCAAT/enhancer binding protein b
CEBPG 3.9c CCAAT/enhancer binding protein g
COL4A1 3.4c Collagen, type IV, a 1
CRABP1 −11 . 8 c Cellular retinoic acid binding protein
DAB2 10.5c Disabled homolog 2
FST 4.7b Follistatin
FZD1 6.3b Frizzled homolog 1
GREM1 −4.8b Gremlin 1
ID3 2.4b Inhibitor of DNA binding 3dominant
negative HLH protein
IL6 80.1b Interleukin 6
LAMA5 2.7b Laminin, a 5
NEFH −5.3b Neurofilament heavy polypeptide
NROB1 6.4c DAX-1 (nuclear receptor subfamily 0,
group B, member 1)
OTX2 −13.1b Orthodenticle homeobox 2
POMC −8.4b Proopiomelenocortin
PTCH 2.6b Protein patched homolog 1
PTGS2 5.2c Prostaglandin enteroperoxide synthase
PTEN −3.4b Phosphatase and tensin homolog
SMAD3 3.8b Smad 3 (Mad homolog 3)
VCAM1 30.5b Vascular cell adhesion protein 1
WNT4 3.2c Wingless type 4
aGenes altered in expression 2-fold or more (see Figure 1); bP > 0.10; cP ≤
0.10; dP ≤ 0.

M. L. Shirley et al. / Stem Cell Discovery 2 (2012) 163-175
168
many common genes, including ones we also found to be
up-regulated: FST, IL6, and VCAM1.
Overall, BMP4 resulted either in no increase, and of-
ten even a down-regulation, of many genes associated
with ectoderm (e.g. EN2, EOMES , EXO1, FGF13, MSI1,
NEFH, NES, NR4A2, PAX 6, PITX3, and SOX3), endo-
derm (e.g. AFP, CER1, and NODAL), and mesoderm (e.g.
CD34, GJA1, GSC, KD R, LMO2, SDNSF, TBX1, and
TIE). There are, however, important exceptions. Several
canonical mesodermal gene markers were up-regulated,
BMP4, BMPR2, CDH5, CDH11, CHRD , GATA4, NKX2-5,
PITX2, RUNX1, RUNX2, SOX9, T, TBX5, and VEGF, as
were the ectodermal markers, EN1, FN1, HOXB4, MMP2,
MSI2, MYL4, and NOG, and the endodermal markers,
GATA6 and HNF4A. Our transcriptome profiling also
revealed the up-regulation of additional genes, including
AKT1, CEBPB, CEBPG, COL4A1, DAB2, FST, FZD1,
IL6, LAMA5, NROB1, PTCH, PTGS2, SMAD3, VCAM1,
and WNT4, and the down-regulation of CRABP1, GREM1,
LHCGR, NEFH , OTX2, POMC, PTEN, and THBS2. The
regulation of CEBPB, RUNX2, SOX9, and VEGF by
BMP4 has not, to the best of our knowledge, been re-
ported by others.
The results in Figure 1 demonstrate that noggin alone
altered the transcription of several of the 177 genes sur-
veyed. Genes that were up-regulated include AFP, AKT1
CDH5, CEBPG, CER1, EBAF, EOMES, FGF4, GBX2,
GSC, HNF4A, IL6, LHCGR, OTX2, POU5F1, SOX2, T,
TBX5, THBS 2, and VCAM1. In addition, noggin down-
regulated a number of genes, CYP19A1, EN1, EN2,
EPAS1, FST, GREM1, MMP2, MMP9, NF KB 1, NOG,
NROB1, PGF, PITX2, PTGS2, SOX3, SOX9, and TFAP2A.
It is worthy of note that the transcription of three of these
down-regulated genes, CYP19A1, EN2, and TFAP2A, fall
below the level of detection following treatment with
noggin, i.e. Ct greater than 35.
Genes highlighted above that were altered similarly by
both BMP4 and noggin, albeit in some cases to a greater
or lesser degree, include, up-regulation: AKT1, CDH5,
CEBPG, IL6, TBX5, VCAM1; and down-regulation:
GREM1 and SOX3. These results may imply non-specific
action of BMP4 and/or direct effects of noggin.
Of the genes surveyed, 31 were not expressed suffi-
ciently to be accurately determined and hence could not
be included in Figure 1. These genes are listed in the
Supplement (Figure S1).
Phase-contrast microscopic images of the hESCs are
given in the Supplement (Figure S1) showing control
cells and cells incubated with BMP4 and with noggin.
Consistent with the transcriptome profiling results, he-
terogeneity is apparent, but the BMP4-treated cells give
more the appearance of syncytiotrophobla st while the
noggin-treated cells are quite distinct.
Hormone production: The media concentrations of
hCG, progesterone, and estradiol were measured daily, up
to seven days, for cells incubated with BMP4 and with
noggin (Figure 2). Following incubation with BMP4, the
three hormones increased on days 6 and 7 relative to
control cells. As expected, noggin itself had no effect on
hormone production. The concentrations reached on day
7 with BMP4 are about 15, 10, and 0.6 ng/mL for hCG,
progesterone, and estradiol, respectively.
3.2. Gene Expression and Hormone
Production in EB
Gene expression: Expression of 74 genes by EBs is
presented as a heatmap (derived from mean values of
) in Figure 3 at 22 and 50 days, each relative to
day 5. Some of the more prominently expressed genes
up-regulated at day 22 and/or day 50 include AFP, CGB ,
CHRD, DAB2, HSPG2, INHBA, PGF, PTGS2, RUNX1,
TIMP1, and VEGF, only a few of which are tropho-
blast-associated. Most of the genes shown in Figure 3
are, however, either down-regulated or exhibit no appre-
ciable change over time. Of the genes surveyed, an arbi-
trary selection of the 15 most highly expressed at day 5,
as assessed by dCt values, are (in decreasing order):
t
ddc
2−
(a) (b) (c)
Figure 2. Medium concentrations of hCG (a), progesterone (b), and estradiol (c) of hESC. The cells were incubated in medium
alone (CON), in medium containing BMP4 (BMP), or in medium with noggin (NOG).

M. L. Shirley et al. / Stem Cell Discovery 2 (2012) 163-175 169
Figure 3. Heatmap showing
changes in gene expression of
EB at 22 and 50 days, relative to
day 5 (mean values of ).
The most prominently altered
genes concomitant with time in
culture are discussed in the text.
t
ddc
2−
GJA1, CREBBP, CDH2, ID3, CDH1, DLX5, SOX2,
CREB1, HRAS, AR, AFP, CDH11, NES , GATA3, and
GREM1 (data not shown). Supplement Figure 2 (S. Fig-
ure 2) shows an H & E stained section of an EB at day
50, and the heterogeneous nature is quite apparent.
Hormone production: Figure 4(a) shows the media
concentrations of hCG by EBs up to 50 days incubation.
Considerable variability was noted in the hormone con-
centrations, attributed to the different sizes and numbers
of EBs in each agarose dish. Moreover, viability may be
decreasing after extended culture. hCG begins increasing
on about day 18 and reaches a maximum on days 32 - 36
days. Under the conditions of the assay, where 50% of
the EB medium is replaced every two days to ensure EB
viability, the concentrations measured reflect new syn-
thesis to a large extent, and, to a lesser degree, accumula-
tion. Assuming that hCG is stable in the medium, it is
possible to correct the concentrations for the total accu-
mulated values at each two days of measurement. Doing
so shows that hCG is continually synthesized between
days 20 - 40 and then reaches a plateau of about 660
ng/mL between days 40 - 50. Media concentrations of
progesterone follow a similar pattern to that of hCG
Figure 4(b) and, when concentrations were corrected as
described above, reach a plateau of about 1 ng/mL be-
tween days 40 - 50. The concentrations of estradiol were
also measured (data not shown) and, when corrected as
per hCG and progesterone, a maximal concentration of
only about 0.04 ± 0.01 ng/mL was achieved. In view of
the dynamic nature of EB size and number, quantifica-
tion of hormone data is precluded.
4. DISCUSSION
4.1. BG02 Cells
This study has shown that, in the presence of FGF2,
BMP4 leads to differentiation of the hESC line, BG02, to
trophoblast-like cells, as has been reported by several
groups using different cell lines, e.g. H1, H7, H9, H14,
HES-2, HES-3 and various culture conditions (cf. [17,19,
21,27]). We have also identified BMP4-mediated and
specific up-regulation of the genes AKT1, CEBPB,
RUNX2, SOX9, and VEGF. Of these, Xu et al. [21] re-
ported only minimal and non-significant changes in
AKT1 and CEBPB. These genes most likely reflect
BMP4 signaling and are not specific to trophoblast, al-
though VEGF is expected to become important in pla-
cental formation and development. AKT1 encodes an
isoform of serine/threonine kinase B (PKBα) involved in
cell survival, proliferation, metabolism, and angiogenesis.
Its regulation by BMP4 and noggin may imply that the
BMP4-regulated component may be via a non- canonical
pathway. The intronless gene CEBPB encodes the bZIP
transcriptional factor, CCAAT/enhancer binding pro-

M. L. Shirley et al. / Stem Cell Discovery 2 (2012) 163-175
170
(a)
(b)
Figure 4. Medium concentration of hCG (a) and
progesterone (b) production by EB cultured up to 50
days. The variability at each day is attributable to the
dynamic nature of EB, in particular the changing size,
number, and possibly atresia and necrosis. As dis-
cussed in the text, the conditions are such that new
hormone synthesis and secretion is the primary con-
tributor with accumulated prior secretion contribut-
ing some as well. A correction for the total accumu-
lated hCG and progesterone, assuming complete sta-
bility in the medium, shows that the majority of syn-
thesis occurs between days 20 - 40, after which pla-
teaus of about 660 ng/mL and 1 ng/mL, respectively,
are reached. The concentrations of estradiol were
also measured and found to be quite low, about 0.04
± 0.01 ng/mL (data not shown).
tein-b, that forms homodimers or heterodimers with
other members of the CEBP family, a, d, and g. RUNX2
is regulated by the BMP pathway and also encodes a
transcriptional factor involved in osteogenesis. Another
transcriptional factor, (sex determining region Y)-box 9,
is encoded by SOX9 and participates in chondrogenesis.
Vascular endothelial growth factor, a member of the cys-
tine-knot growth factor family and encoded by VEGF,
stimulates vasculogenesis and angiogenesis and is ex-
pected to function in placental development. The results
reported herein demonstrate an effect of noggin on the
expression of a number of genes. In some cases BMP4
and noggin alter gene expression similarly, suggesting
either that the regulation by BMP4 may not be via a ca-
nonical signaling pathway or that BMP4 and noggin may
alter transcription similarly. At this time it is not possible
to relate the observed changes accompanying treatment
with noggin with the early signs of neural differentiation
noted by Pera et al. [18].
Xu et al. [21] first reported that, in the presence of
FGF2, BMP2, BMP4, and BMP7 promoted differentia-
tion of H1, H7, H9, and H14 cells, cultured in mouse
embryonic fibroblast conditioned medium, into tro-
phoblast cells. Later, others found that the induction of
trophoblast by BMP4 required an inhibition of the Ac-
tivin/Nodal signaling [43]. The seminal study by Xu et al.
[21] was followed by other reports confirming BMP-me-
diated hESC differentiation to trophoblast-like cells. Das
et al. [44], for example, also showed that BMP4 directed
H1 and H9 cells to trophoblast, while FGF2 slowed this
differentiation, with oxygen accelerating it and promot-
ing formation of syncytiotrophoblast. Interestingly, it was
found that FGF2, acting to maintain NANOG levels via
the MEK-ERK pathway, is capable of switching BMP4-
mediated differentiation of hESC to mesendoderm as
documented by the expression of brachyury and other
primitive streak markers [22]. A similar conclusion was
reached by Bernardo et al. [39] who showed that differ-
entiation of hESCs by BMP4 in the presence of FGF2
led to formation of mesoderm and inhibition of endo-
derm. They also reported that this differentiation was via
the ERK pathway and mediated by brachyury and CDX2.
In a recent comprehensive study of BMP-mediated
differentiation of hESC to trophoblast, Marchand et al.
[17] performed microarray analysis using the Affymetrix
Human Gene version 1.0 ST array, along with qRT-PCR
on selected genes. H7 and H9 cells were incubated with
BMP4 for various times, 0, 2, 4, 6, 8, and 10 days, fol-
lowing the removal of FGF2. They found that after 2
days POU5F1 and NANOG were dramatically down-
regulated while trophoblast markers were up-regulated.
Many new genes were identified and suggested to be
involved in trophoblast formation, and pathway analysis
provided considerable insight into the myriad signaling
systems operative in the differentiation of hESCs to tro-
phoblast. This study was augmented by another report
[42] in which the transcriptome of trophectoderm cells
obtained from 13 human blastocysts were compared with
those of BMP4-mediated differentiation of hESCs [17].
Their results documented that BMP4-induced differentia-
tion of hESCs offers a good model for studying tro-
phoblasts and contributed significantly to a better de-
lineation of the associated transcriptome.
Our results with BMP4 are, by and large, in agreement
with the findings from the two major combined microar-
ray and PCR investigations on BMP-induced differentia-
tion of hESCs [17,21]. This is somewhat surprising since
we and Xu et al. [21] maintained FGF2 along with

M. L. Shirley et al. / Stem Cell Discovery 2 (2012) 163-175 171
BMP4, while Marchand et al. [17] removed FGF2 from
the medium when BMP4 was added. A summary of many
of the genes found to be altered in the present study
compared to other reports is given in the Supplement
(S.2). A major difference between our results and those
of Marchand et al. [17] is in the expression of the meso-
dermal markers, BMP4 and T. We observed increased
expression of these two genes, while they reported a de-
crease or no change, attributable to their removal of
FGF2 during BMP-mediated differentiation. Of interest,
they found increased expression of KDR, while we noted
a minimal decrease, albeit not significant. Increased ex-
pression of MMP9 was found herein, consistent with the
findings of Xu et al. [21] and Schultz et al. [19], but
Marchand et al. [17] reported down-regulation. With the
BG02 cells, we found reduced expression of the pluripo-
tent marker, FOXH1, whereas others did not [17,21].
These discrepancies may reflect cell-specific differences,
culture differences, e.g. ±FGF2, or other factors.
While it has been convincingly documented that
members of the BMP family lead to differentiation of
hESC to trophoblast, there is also considerable evidence
that experimental conditions have a profound effect on
the type of differentiation obtained. For example, an ear-
lier report on BMP4-mediated differentiation of BG02
cells identified the formation and outgrowth of an im-
mature vascular system when the cells are grown in a 3D
Matrigel substrate in an endothelial cell growth medium
[13]. Further, Pera et al. [18] found that, in response to
BMP2, BMP4, or BMP2/7, HES-2 and HES-3 hESC dif-
ferentiated to extra-embryonic endoderm, with only a
few percent of the cells having the appearance of the
trophoblast precursors described by Xu et al. [21]. They
also found that the BMP antagonist noggin blocks the
differentiation to extra-embryonic endoderm and directs
differentiation into neural precursors, differentiation that
may be mimicked by secretion of gremlin by mouse em-
bryo fibroblast feeder layers. Our results with noggin
reflect some type(s) of differentiation, but there is no
clear indication for preference of one major pathway. As
judged by increased expression of T, MIXL1, and WNT3,
others have found that short-term treatment of H1, H7,
and H9 hESC with BMP4 resulted in the induction of
mesoderm progenitor cells that can differentiate into he-
matopoietic and cardiac lineages [24]. Working with H7
cells, it was reported that BMP4 treatment failed to yield
trophectoderm using mouse embryonic fibroblasts as a
feeder layer; this, however, was overcome by using
feeder-free cells on Geltrex-coated plates in StemPro
[38]. Lastly, West et al. [45] have demonstrated that, in
the presence of BMP4, the KIT ligand enhances differen-
tiation to germ-like cells. Hence, additional work is needed
to clarify the many experimental parameters associated
with BMP-induced differentiation of hESC.
4.2. EB Derived from BG02 Cells
EB formation by these cells also yielded some degree
of differentiation to trophoblast as evidenced by the
up-regulation of CGB and the production of hCG and
progesterone. The most highly expressed genes in EB on
day 5 are: GJA1, CREBBP, CDH2, ID3, CDH1, DLX5,
SOX2, CREB1, HRAS, AR, and AFP. Most of these were
surveyed in the studies on cells receiving BMP4, and
none were up-regulated. The results suggest that tro-
phoblast-like cells are not forming to any significant ex-
tent by day 5, findings consistent with the absence of
hCG and progesterone production until days 18 - 20. These
data are consistent with those by Gerami-Naini et al. [8]
on H1 cell-derived EB growing in Matrigel. They could
not detect measureable hormone until about day 20 of
culture, and, depending upon the conditions used, maxi-
mal production of hCG was reached on days 35 - 40, fol-
lowed by a decline. In contrast to the results with the
Matrigel-embedded EB, they found that in suspension
culture EBs were producing hCG, progesterone, and es-
tradiol by 48 h.
In our studies, a comparison of gene expression by EB
on days 22 and 50, relative to day 5, with that of hESC
receiving BMP4 for 7 days, provides additional evidence
that differentiation to trophoblast is occurring by day 22,
as evidenced by the up-regulation of CGB, PGF, PTGS2,
RUNX1, and VEGF. These results are consistent with the
hormone secretion data for hCG and progesterone. In
addition to the above genes, there is also up-regulation of
AFP, BMP6, CHD11, CHRD , DAB2, HSPG2, INHBA,
PTGS2, and TIMP1. Not surprisingly, the EBs are ap-
parently more heterogeneous in terms of constituent cell
types than the BMP4-treated hESCs. For example, the
early appearance (day 5) of GJA1, CREBBP, CDH2, ID3,
CDH1, DLX5, SOX2, CREB1, HRAS, AR, AFP, CDH11,
NES, GATA3, and GREM1, followed by the later
up-regulation of AFP, BMP6, CHD11, CHR D, DAB2,
HSPG2, INHBA, PTGS2, and TIMP1, indicates forma-
tion of trophoblast, endoderm, ectoderm, and mesoderm.
Consistent with this finding was the observation some
years ago that markers for the three embryonic germ la-
yers were expressed in EBs prepared from H9 cells [46].
5. CONCLUSION
Overall, the results presented herein strongly support
other reports concluding that, under certain conditions,
BMP4 directs differentiation of hESCs to trophoblast-
like cells. This conclusion notwithstanding, it is clear that
experimental conditions, particularly the inclusion or
exclusion of FGF2 during BMB-mediated differentiation,
have a profound effect on the type of differentiation

M. L. Shirley et al. / Stem Cell Discovery 2 (2012) 163-175
172
achieved, and, moreover, based on our transcriptome
profiling, it is highly likely that other cell types may be
forming in response to BMP4. On the other hand, it may
emerge that some of the other genes we found to have
been up-regulated by BMP4 function in differentiation to
trophectoderm and then to cytotrophoblasts (villous and
extravillous) and syncytiotrophoblasts, or be involved in
placental formation and function. The heterogeneity of
the differentiated cells needs to be carefully established,
but the BMP-mediated differentiation of hESCs to tro-
phoblast, particularly in the absence of FGF2, certainly
provides an attractive in vitro system for studying early
differentiation events and gives a more homogeneous
system than that of embryoid bodies. Lastly, our obser-
vation that the BMP4 antagonist noggin alters gene tran-
scription of a subset of genes investigated may correlate
with the morphological changes reported by others;
however, more studies are required to map the transcrip-
tional changes to neural differentiation.
6. ACKNOWLEDGEMENTS
We thank Mr. Roger Nilsen of the University of Georgia Functional
Genomics Resource Facility for his expert assistance with the RNA
analysis. This study was supported by NIH: R01DK033973 (D.P.),
R01HD044119 (P.N.), and 5F32HL083741 (N.L.B.).
REFERENCES
[1] Rama, S. and Rao, A.J. (2003) Regulation of growth and
function of the human placenta. Molecular and Cellular
Biochemistry, 253, 263-268.
doi:10.1023/A:1026076219126
[2] Sullivan, M.H. (2004) Endocrine cell lines from the pla-
centa. Molecular and Cellular Endocrinology, 228, 103-
119. doi:10.1016/j.mce.2003.03.001
[3] Udayashankar, R., Baker, D., Tuckerman, E., Laird, S., Li,
T.C. and Moore, H.D. (2011) Characterization of invasive
trophoblasts generated from human embryonic stem cells.
Human Reproduction, 26, 398-406.
doi:10.1093/humrep/deq350
[4] Jameson, J.L. and Hollenberg, A.N. (1993) Regulation of
chorionic gonadotropin gene expression. Endocrine Re-
views, 14, 203-221.
[5] Mesiano, S. (2009) The endocrinology of human preg-
nancy and fetoplacental neuroendocrine development. In:
Strauss, J.F. and Barbieri, R.L., Eds., Yen and Jaffe’s Re-
productive Endocrinology. 6th Edition, Physiology, Patho-
physiology and Clinical Management, Philadelphia, 249-
281.
[6] Thomson, J.A., Itskovitz-Eldor. J., Shapiro, S.S., Waknitz,
M.A., Swiergiel, J.J., Marshall, V.S. and Jones, J.M.
(1998) Embryonic stem cell lines derived from human
blastocysts. Science, 282, 1145-1147.
doi:10.1126/science.282.5391.1145
[7] Mitalipova, M., Calhoun, J., Shin, S., Wininger, D., Sch-
ulz, T., Noggle, S., Venable, A., Lyons, I., Robins, A. and
Stice, S. (2003) Human embryonic stem cell lines derived
from discarded embryos. Stem Cells, 21, 521-526.
doi:10.1634/stemcells.21-5-521
[8] Gerami-Naini, B., Dovzhenko, O.V., Durning, M., Wegner,
F.H., Thomson, J.A. and Golos, T.G. (2004) Trophoblast
differentiation in embryoid bodies derived from human
embryonic stem cells. Endocrinology, 145, 1517-1524.
doi:10.1210/en.2003-1241
[9] Giakoumopoulos, M., Siegfried, L.M., Dambaeva, S.V.,
Garthwaite, M.A., Glennon, M.C. and Golos, T.G. (2010)
Placental-derived mesenchyme influences chorionic go-
nadotropin and progesterone secretion of human embry-
onic stem cell-derived trophoblasts. Reproductive Science,
17, 798-808. doi:10.1177/1933719110371853
[10] Golos, T.G., Pollastrini, L.M. and Gerami-Naini, B. (2006)
Human embryonic stem cells as a model for trophoblast
differentiation. Semin Reproductive Medicine, 24, 314-
321. doi:10.1055/s-2006-952154
[11] Harun, R., Ruban, L., Matin, M., Draper, J., Jenkins,
N.M., Liew, G.C., Andrews, P.W., Li, T.C., Laird, S.M.
and Moore, H.D. (2006) Cytotrophoblast stem cell lines
derived from human embryonic stem cells and their ca-
pacity to mimic invasive implantation events. Human
Reproduction, 21, 1349-1358.
doi:10.1093/humrep/del017
[12] Peiffer, I., Belhomme, D., Barbet, R., Haydont, V., Zhou,
Y.P., Fortunel, N.O., Li, M., Hatzfeld, A., Fabiani, J.N.
and Hatzfeld, J.A. (2007) Simultaneous differentiation of
endothelial and trophoblastic cells derived from human
embryonic stem cells. Stem Cells and Development, 16,
393-402. doi:10.1089/scd.2006.0013
[13] Boyd, N.L., Dhara, S.K., Rekaya, R., Godbey, E.A., Has-
neen, K., Rao, R.R., West, F.D., Gerwe, B.A. and Stice,
S.L. (2007) BMP4 promotes formation of primitive vas-
cular networks in human embryonic stem cell-derived
embryoid bodies. Experimental Biology and Medicine,
232, 833-843.
[14] Chen, G., Ye, Z., Yu, X., Zou, J., Mali, P., Brodsky, R.A.
and Cheng, L. (2008) Trophoblast differentiation defect
in human embryonic stem cells lacking PIG-A and
GPI-anchored cell-surface proteins. Cell Stem Cell, 2,
345-355. doi:10.1016/j.stem.2008.02.004
[15] Kee, K., Gonsalves, J.M., Clark, A.T. and Pera, R.A.
(2006) Bone morphogenetic proteins induce germ cell
differentiation from human embryonic stem cells. Stem
Cells and Development, 15, 831-837.
doi:10.1089/scd.2006.15.831
[16] Liu, Y.P., Dovzhenko, O.V., Garthwaite, M.A., Dambaeva,
S.V., Durning, M., Pollastrini, L.M. and Golos, T.G. (2004)
Maintenance of pluripotency in human embryonic stem
cells stably over-expressing enhanced green fluorescent
protein. Stem Cells and Development, 13, 636-645.
doi:10.1089/scd.2004.13.636
[17] Marchand, M., Horcajadas, J.A., Esteban, F.J., McElroy,
S.L., Fisher, S.J. and Giudice, L.C. (2011) Transcriptomic
signature of trophoblast differentiation in a human em-
bryonic stem cell model. Biology of Reproduction, 84,
1258-1271. doi:10.1095/biolreprod.110.086413
[18] Pera, M.F., Andrade, J., Houssami, S., Reubinoff, B.,

M. L. Shirley et al. / Stem Cell Discovery 2 (2012) 163-175 173
Trounson, A., Stanley, E.G., Ward-van Oostwaard, D. and
Mummery, C. (2004) Regulation of human embryonic
stem cell differentiation by BMP-2 and its antagonist
noggin. Journal of Cell Science, 117, 1269-1280.
doi:10.1242/jcs.00970
[19] Schulz, L.C., Ezashi, T., Das, P., Westfall, S.D., Livingston,
K.A. and Roberts, R.M. (2008) Human embryonic stem
cells as models for trophoblast differentiation. Placenta,
29, S10-S16.
[20] Xu R.H. (2006) In vitro induction of trophoblast from
human embryonic stem cells. Methods in Molecular Me-
dicine, 121, 189-202.
[21] Xu, R.H., Chen, X., Li, D.S., Li, R., Addicks, G.C., Glen-
non, C., Zwaka, T.P. and Thomson, J.A. (2002) BMP4
initiates human embryonic stem cell differentiation to tro-
phoblast. Nature Biotechnology, 20, 1261-1264.
doi:10.1038/nbt761
[22] Yu, P., Pan, G., Yu, J. and Thomson, J.A. (2011) FGF2
sustains NANOG and switches the outcome of BMP4-
induced human embryonic stem cell differentiation. Cell
Stem Cell, 8, 326-334. doi:10.1016/j.stem.2011.01.001
[23] Yu, X., Zou, J., Ye, Z., Hammond, H., Chen, G., Toku-
naga, A., Mali, P., Li, Y.M., Civin, C., Gaiano, N. and
Cheng, L. (2008) Notch signaling activation in human
embryonic stem cells is required for embryonic, but not
trophoblastic, lineage commitment. Cell Stem Cell, 2,
461-471. doi:10.1016/j.stem.2008.03.001
[24] Zhang, P., Li, J., Tan, Z., Wang, C., Liu, T., Chen, L.,
Yong, J., Jiang, W., Sun, X., Du, L., Ding, M. and Deng,
H. (2008) Short-term BMP-4 treatment initiates meso-
derm induction in human embryonic stem cells. Blood,
111, 1933-1941. doi:10.1182/blood-2007-02-074120
[25] Douglas, G.C., Vande-Voort, C.A., Kumar, P., Chang, T.C.
and Golos, T.G. (2009) Trophoblast stem cells: Models for
investigating trophectoderm differentiation and placental
development. Endocrine Reviews, 30, 228-240.
doi:10.1210/er.2009-0001
[26] Golos, T.G., Giakoumopoulos, M. and Garthwaite, M.A.
(2010) Embryonic stem cells as models of trophoblast
differentiation: Progress, opportunities, and limitations.
Reproduction, 140, 3-9. doi:10.1530/REP-09-0544
[27] Pera M.F. and Trounson A.O. (2004) Human embryonic
stem cells: Prospects for development. Development, 131,
5515-5525. doi:10.1242/dev.01451
[28] Roberts, R.M., Ezashi, T. and Das, P. (2004) Trophoblast
gene expression: Transcription factors in the specification
of early trophoblast. Reproductive Biology and Endocri-
nology, 2, 47. doi:10.1186/1477-7827-2-47
[29] Hay, D.C., Sutherland, L., Clark, J. and Burdon, T. (2004)
Oct-4 knockdown induces similar patterns of endoderm
and trophoblast differentiation markers in human and
mouse embryonic stem cells. Stem Cells, 22, 225-235.
doi:10.1634/stemcells.22-2-225
[30] Matin, M.M., Walsh, J.R., Gokhale, P.J., Draper, J.S.,
Bahrami, A.R., Morton, I., Moore, H.D. and Andrews,
P.W. (2004) Specific knockdown of Oct4 and beta 2-mi-
croglobulin expression by RNA interference in human
embryonic stem cells and embryonic carcinoma cells.
Stem Cells, 22, 659-668. doi:10.1634/stemcells.22-5-659
[31] Bragdon, B., Moseychuk, O., Saldanha, S., King, D.,
Julian, J. and Nohe, A. (2011) Bone morphogenetic pro-
teins: A critical review. Cell Signal, 23, 609-620.
doi:10.1016/j.cellsig.2010.10.003
[32] Rider, C.C. and Mulloy, B. (2010) Bone morphogenetic
protein and growth differentiation factor cytokine fami-
lies and their protein antagonists. The Biochemical Jour-
nal, 429, 1-12. doi:10.1042/BJ20100305
[33] Walsh, D.W., Godson, C., Brazil, D.P. and Martin, F.
(2010) Extracellular BMP-antagonist regulation in de-
velopment and disease: Tied up in knots. Trends in Cell
Biology, 20, 244-256. doi:10.1016/j.tcb.2010.01.008
[34] Zeng, S., Chen, J. and Shen, H. (2010) Controlling of
bone morphogenetic protein signaling. Cell Signal, 22,
888-893. doi:10.1016/j.cellsig.2009.12.007
[35] Malassine, A., Frendo, J.L. and Evain-Brion, D. (2003) A
comparison of placental development and endocrine
functions between the human and mouse model. Human
Reproduction, 9, 531-539. doi:10.1093/humupd/dmg043
[36] Rossant, J. (2001) Stem cells from the mammalian blas-
tocyst. Stem Cells, 19, 477-482.
doi:10.1634/stemcells.19-6-477
[37] Smith, A.G. (2001) Embryo-derived stem cells of mice
and men. Annual Reviews of Cell and Developmental Bi-
ology, 17, 435-462. doi:10.1146/annurev.cellbio.17.1.435
[38] Erb, T.M., Schneider, C., Mucko, S.E., Sanfilippo, J.S.,
Lowry, N.C., Desai, M.N., Mangoubi, R.S., Leuba, S.H.
and Sammak, P.J. (2011) Paracrine and epigenetic control
of trophectoderm differentiation from human embryonic
stem cells: the role of bone morphogenic protein 4 and
histone deacetylases. Stem Cells Development, 20, 1601-
1614. doi:10.1089/scd.2010.0281
[39] Bernardo, A.S., Faial, T., Gardner, L., Niakan, K.K., Ort-
mann, D., Senner, C.E., Callery, E.M., Trotter, M.W.,
Hemberger, M., Smith, J.C., Bardwell, L., Moffett, A. and
Pedersen, R.A. (2011) BRACHYURY and CDX 2 mediate
BMP-induced differentiation of human and mouse pluri-
potent stem cells into embryonic and extra embryonic
lineages. Cell Stem Cell, 9, 144-155.
doi:10.1016/j.stem.2011.06.015
[40] Livak, K.J. and Schmittgen, T.D. (2001) Analysis of rela-
tive gene expression data using real-time quantitative
PCR and the 2(-delta delta C(T)) method. Methods, 25,
402-408.
[41] Pfaffl, M.W. (2001) A new mathematical model for rela-
tive quantification in real-time RT-PCR. Nucleic Acids
Research, 29, e45. doi:10.1093/nar/29.9.e45
[42] Aghajanova, L., Shen, S., Rojas, A.M., Fisher, S.J., Irwin,
J.C. and Giudice, L.C. (2012) Comparative transcriptome
analysis of human trophectoderm and embryonic stem
cell-derived trophoblasts reveal key participants in early
implantation. Biology and Reproduction, 86, 1-21.
doi:10.1095/biolreprod.111.092775
[43] Wu, Z., Zhang, W., Chen, G., Cheng, L., Liao, J., Jia, N.,
Gao, Y., Dai, H., Yuan, J. and Xiao, L. (2008) Combina-
torial signals of activin/nodal and bone morphogenic pro-
tein regulate the early lineage segregation of human em-
bryonic stem cells. The Journal of Biological Chemistry,

M. L. Shirley et al. / Stem Cell Discovery 2 (2012) 163-175
174
283, 24991-25002. doi:10.1074/jbc.M803893200
[44] Das, P., Ezashi, T., Schulz, L.C., Westfall, S.D., Living-
ston, K.A. and Roberts, R.M. (2007) Effects of fgf 2 and
oxygen in the bmp4-driven differentiation of trophoblast
from human embryonic stem cells. Stem Cell Research, 1,
61-74. doi:10.1016/j.scr.2007.09.004
[45] West, F.D., Roche-Rios, M.I., Abraham, S., Rao, R.R.,
Natrajan, M.S., Bacanamwo, M. and Stice, S.L. (2010)
KIT ligand and bone morphogenetic protein signaling
enhances human embryonic stem cell to germ-like cell
differentiation. Human Reproduction, 25, 168-178.
doi:10.1093/humrep/dep338
[46] Itskovitz-Eldor, J., Schuldiner, M., Karsenti, D., Eden, A.,
Yanuka, O., Amit, M., Soreq, H. and Benvenisty, N.
(2000) Differentiation of human embryonic stem cells
into embryoid bodies compromising the three embryonic
germ layers. Molecular Medicine, 6, 88-95.

M. L. Shirley et al. / Stem Cell Discovery 2 (2012) 163-175 175
1. Supplement
1.1. S1. The Following Genes Were Not
Expressed Sufficiently to Be Measured
in Control and BMP4-Mediated
Differentiation of hESCs
AMH, AMHR2, BAPX1, BMP1, BMPR1B, DLX5,
ESR1, ESR2, FGF5, FIGF, FSHB, FSHR, GJA5, GJB3,
IGF1, INHBB, IPF1, LAMR1, LY6G6D, NKX2-2, PAX 8,
PECAM1, PGR, PP13, PROML1, PTGS1, PTPRC,
STAR, TAL1, TITF1, TSHB, and WT1.
Figure S1. Phase-contrast microscopy of hESCs before and
after incubation with BMP4 or noggin. (A) Cells at day 0 in
media; (B) Cells at day 7 in media; (C) Cells at day 7 in me-
dia-plus-100 ng/mL BMP4; (D) Cells at day 7 in media-
plus-250 ng/mL noggin. Cells incubated with BMP4 and
with noggin exhibit distinct morphological changes.
1.2. S2. A Comparison of Our Results with
BMP4-Induced Differentiation of
hESCs and the Results of Others Gives
the Following Similarities
We found increased expression of CDH11, CGB, EN-
PEP, EPAS1, FN1, GATA2, GATA3, HEY1, KRT7, MSX2,
PGF, PITX2, and WNT5A, as well as decreased expres-
sion of DNMT3B and POU5F1, in agreement with Xu et
al. [21] and Marchand et al. [17]. Consistent with the
data of Xu et al. [21], we found increased expression of
CDH5, DAB2, MMP9, and WNT4, along with decreased
expression of SOX3. Our results and those of Marchand
et al. [17] show increased expression of BMPR2,
COL4A1, CYP11A1, and CYP19A1, decreased expres-
sion of CRABP1, FGF2, GREM1, OTX2, and SOX2, and
either no changes or minimal changes in a number of
other genes, including AFP, GSC, NES, PA X6, and oth-
ers. Schultz et al. [19] also found increased expression in
CGB, GATA2, GATA3, KRT7, and MSX2, and decreased
expression of POU5F1 and SOX2.
Figure S2. Following formation of embryoid bod-
ies from hESCs and incubation for 50 days (d50),
an embryoid body was fixed in formalin, embedded
in paraffin, sectioned (5 microns), and stained with
H&E.