Cell Stem Cell

Resource

Tracing the Derivation of Embryonic

Stem Cells from the Inner Cell Mass
by Single-Cell RNA-Seq Analysis
Fuchou Tang,1,3 Catalin Barbacioru,2 Siqin Bao,1 Caroline Lee,1 Ellen Nordman,2 Xiaohui Wang,2 Kaiqin Lao,2,*
and M. Azim Surani1,*
1Wellcome Trust/Cancer Research UK Gurdon Institute of Cancer and Developmental Biology, University of Cambridge,

Tennis Court Road, Cambridge CB2 1QN, UK

2Genetic Systems, Applied Biosystems, part of Life Technologies, 850 Lincoln Centre Drive, Foster City, CA 94404, USA
3Present Address: BIOPIC, School of Life Sciences, Peking University, Beijing 100871, China

*Correspondence: kai.lao@lifetech.com (K.L.), a.surani@gurdon.cam.ac.uk (M.A.S.)

DOI 10.1016/j.stem.2010.03.015

SUMMARY hampered by the limited number of cells available for analysis

(Niwa, 2007).
During the transition from the inner cell mass (ICM) Pluripotent E3.5-E4.5 primitive ectoderm/epiblast (PE) and
cells of blastocysts to pluripotent embryonic stem ESCs can both contribute to all three germ layers and the
cells (ESCs) in vitro, a normal developmental pro- germ line when injected into host blastocysts to form chimera
gram is replaced in cells that acquire a capacity for (Niwa, 2007). However, only the ESCs cultured in vitro have the
infinite self-renewal and pluripotency. We explored capacity for unlimited self-renewal while retaining their pluripo-
tency (Niwa, 2007; Smith, 2006). Some differences between
the underlying mechanism of this switch by using
the ICM and ESCs have been identified, such as the expression
RNA-Seq transcriptome analysis at the resolution of pramel5, pramel6, and pramel7 in the ICM, which are
of single cells. We detected significant molecular repressed in ESC (Kaji et al., 2007). Other genes, including Dicer
transitions and major changes in transcript variants, (Bernstein et al., 2003; Kanellopoulou et al., 2005; Murchison
which include genes for general metabolism. Fur- et al., 2005), Nanog (Chambers et al., 2007; Chambers et al.,
thermore, the expression of repressive epigenetic 2003; Mitsui et al., 2003), Mbd3 (Kaji et al., 2006; Kaji et al.,
regulators increased with a concomitant decrease 2007), and Ezh2 (O’Carroll et al., 2001; Shen et al., 2008), are
in gene activators that might be necessary to essential for the establishment of pluripotent PE cells in the
sustain the inherent plasticity of ESCs. Furthermore, ICM but dispensable for the maintenance of ESCs.
we detected changes in microRNAs (miRNAs), with There have been intensive studies on ESCs in recent years, but
one set that targets early differentiation genes while these have usually been on bulk cells by RNA-Seq, cDNA micro-
array, SAGE, and EST sequencing (Niwa, 2007; Ivanova et al.,
another set targets pluripotency genes to maintain
2006; Cloonan et al., 2008). However, the precise changes
the unique ESC epigenotype. Such genetic and
accompanying the process of conversion of ICM to ESCs remain
epigenetic events may contribute to a switch from to be fully elucidated. To gain insight into this process, we used
a normal developmental program in adult cells blastocysts from Oct4-DPE-GFP transgenic mice and cultured
during the formation of diseased tissues, including them in vitro under the classical conditions consisting of LIF
cancers. and FCS used for the derivation of ESCs (Niwa, 2007). The
Oct4-DPE-GFP reporter we used is under the control of only
the distal enhancer for Oct4 (also known as Pou5f1) and lacks
INTRODUCTION the proximal enhancer (Yeom et al., 1996). This GFP reporter
shows expression in the E3.5 ICM, E4.5 epiblast, primordial
The derivation of embryonic stem cells (ESCs) from the inner cell germ cells (PGCs), and ESCs, but not in the postimplantation
mass (ICM) of mouse blastocysts consisting of about 20 cells epiblast or in the epiblast stem cells (EpiSCs) (Yeom et al.,
occurs in vitro under a variety of culture conditions, such as in 1996; Bao et al., 2009). Notably, the distal enhancer of Oct4
the presence of leukemia inhibitory factor (LIF) and fetal calf represents the densest binding locus for the key pluripotency-
serum (FCS) (Evans and Kaufman, 1981; Ying et al., 2008). After specific transcription factors in ESCs (Chen et al., 2008), which
about 5 days in culture, the inner cell mass outgrowths of blasto- makes it an ideal reporter for tracing the course of changes
cysts are disrupted into small clusters of cells and passaged until during the establishment of ESCs from ICM. By analyzing single
the establishment of ESC lines. Thus, the ICM cells that, in vivo, Oct4-DPE-GFP-positive and Oct4-DPE-GFP-negative cells, we
are subject to a strict developmental program undergo a trans- set out to monitor changes in ICM cells during their progression
formation into cells with a capacity for infinite self-renewal while toward ESCs. We used our recently developed single-cell
retaining pluripotency. The precise molecular changes accom- RNA-Seq transcriptome analysis to investigate the critical early
panying this transition remain to be fully elucidated, which is changes during this process (Tang et al., 2009).

468 Cell Stem Cell 6, 468–478, May 7, 2010 ª2010 Elsevier Inc.
Cell Stem Cell
Single-Cell RNA-Seq Profiling of ICM Outgrowth

RESULTS AND DISCUSSION high expression, but this declined during the course of ICM
outgrowths so that ultimately only about 50% (7/14) of individual
Analysis of Individual ICM Outgrowth Cells ESCs retained Bmp4 expression (Ct = 25–40). This is compatible
First, we analyzed the three key pluripotency genes during the with the fact that maintenance of ESCs can be achieved by the
course of blastocyst culture and the formation of outgrowths addition of exogenous Bmp4 or serum, which contains Bmp4
(Figure 1). At each stage, we chose between 10 and 26 single (Ying et al., 2003).
cells for analysis. We generated cDNAs by whole transcriptome During the course of ICM outgrowth toward ESCs, we found
amplification (WTA) of these individual cells (see Experimental clear upregulation of several genes, including Tcf15, Prdm5,
Procedures for details). All ICM cells (22/22) tested showed Zic3, Ifitm1, Nodal, and Bex1, indicating that they may potentially
high expression of Oct4, Sox2, and Nanog. However, among be important during the transition to ESCs and/or for their subse-
cells from day 3 outgrowths that had high Oct4 expression, quent maintenance (Figure 2D). Indeed, Nodal is a known regu-
about 39% (7/18) had already lost expression of Nanog and/or lator of self-renewal but is not essential for the pluripotency of
Sox2, indicating that they might be losing pluripotency. By ESCs (see below). By contrast, there was clear downregulation
contrast, most of the cells from day 5 outgrowths (11/13) that of some genes during ICM outgrowth, such as Gata4, Gata6,
had high Oct4 expression also showed high expression of both Pramel7, Tbx3, Bmi1, Bcl2l14, Nr5a2, and Amhr2, which poten-
Sox2 and Nanog, suggesting that these may represent the tially have ICM specific development-related functions (Figure 2E).
earliest population that had acquired or were likely on course For example, ICM has the potential to develop into primitive endo-
to acquire the ESC-like fate with the potential for self-renewal. derm cells, for which Gata4 and Gata6 are crucial regulators
We were also able to establish an ESC line from a single cell iso- (Fujikura et al., 2002; Koutsourakis et al., 1999; Morrisey et al.,
lated from a day 5 outgrowth (data not shown). As expected, all 1998). Thus, repression of these genes may allow ICM cells to
the ESCs (23/23) had high expression of these three pluripotency exit from their inherent developmental program as they acquire
genes. the ability for self-renewal while retaining pluripotency as ESCs.

Expression Dynamics of 385 Genes in 74 Single Cells Molecular Changes during the Transition
from ICM to ESCs from ICM to ESCs
Next, we chose 385 pluripotency and early differentiation related To understand the dynamic nature of gene expression in indi-
genes to monitor their expression in cells from the ICM, as well as vidual cells at the whole-genome scale, we randomly selected
from day 3 and day 5 outgrowths, and from ESCs at single-cell 12 individual ESCs and generated their digital transcriptome
resolution (Table S1). All 14 ESCs analyzed had high expression profile (Figure 3A, Figure S2, and Tables S2 and S3) (Tang
(Ct = 19–28) of Oct4, Sox2, Nanog, Dppa4, Dppa5, Sall4, Utf1, et al., 2009). Indeed, all of the 12 ESCs analyzed had high
Rex2, and Rif1, indicating their pluripotent character (Figure 2A expression of Oct4, Sox2, Nanog, Rex1 (also known as Zfp42),
and Figure S1). By contrast, we detected little or no expression Dppa5, and Utf1, which indicates that all of them are in an undif-
(Ct = 40) of all 23 early differentiation marker genes (ectoderm ferentiated state and are pluripotent. To confirm the reliability of
markers: Pax6, Otx1, Neurod1, Nes, Lhx5, and Hoxb1; meso- our single-cell RNA-Seq approach, we compared our data with
derm markers: Tbx2, T, Nkx2-5, Myod1, Myf5, Mesdc1, Mesdc2, that obtained from bulk analysis of ESCs (Cloonan et al., 2008).
Kdr, Isl1, Hand1, and Eomes; endoderm markers: Onecut1, We found that on average, an individual ESC expresses 10,815
Gata4, Gata5, and Gata6; extraembryonic markers: Cdx2 and genes (RPM > 0.1), which means that we captured expression
Tpbpa) (see Table S1). Similarly, all 14 cells isolated and of at least 94.6% of the genes in a single cell of those detected
analyzed from ICM showed high expression of the nine pluripo- by deep sequencing in bulk assays of ESCs (Cloonan et al.,
tency-specific genes. However, expression of some genes, for 2008). Overall, 65.8% (13,326 out of 20,259) of known genes
example, c-Myc, which was shown to be an important reprog- were expressed in 12 single ESCs, which shows that our
ramming factor for pluripotency (Takahashi and Yamanaka, RNA-Seq data represent an accurate reflection of the entire
2006), was highly heterogeneous in cells from the ICM transcriptome in ESCs at single-cell resolution.
(Ct = 24–40); this variability was progressively reduced until, To understand the relationship between ESCs and the ICM/
finally, all ESCs consistently expressed c-Myc (Figure 2B). Inter- Epiblast cells from which they were derived, we compared the
estingly, we found that Tet1 and Tet2 (Table S1), which were single-cell RNA-Seq transcriptomes of these cells (Figure S2)
recently shown to mediate DNA demethylation in ESCs, were to determine the extent to which ESCs resemble E3.5 ICM or
highly expressed in both ICM and ESCs, but their expression E4.5 Epiblast cells (Nichols et al., 2009). We found that the
only decreased in Oct4-negative cells present in the ICM molecular signature of all undifferentiated ESCs maintained
outgrowths. Thus, our observations support their importance under our culture conditions are clearly different from both ICM
for pluripotency (Tahiliani et al., 2009). and epiblast cells based on the principal component analysis
Since ESCs can also be maintained in an undifferentiated state of their transcriptomes. This means that at the molecular level,
by LIF and BMP4 (Ying et al., 2003), we investigated the expres- ESCs are distinct from E3.5 ICM or E4.5 Epiblast (Figure 3A).
sion of a key receptor, Bmpr1a, and found it to be heterogeneous We detected a large set of genes, which show clear differential
in the ICM (Ct = 27–40). However during the ICM outgrowth, expression between ICM/Epiblast and ESCs. (Table S2 and Fig-
Bmpr1a expression was detected more consistently until, finally, ure S3. Note 2,475 genes with fold change, FC[ESC/ICM] > 4,
all ESCs (14/14) showed strong expression. This suggests that p < 0.01, and 2,362 genes with FC[ESC/ICM] < 0.25, p < 0.01;
all ESCs have the potential to respond to Bmp4 signaling 2,110 genes with FC[ESC/Epiblast] > 4, p < 0.01 and 1,170 genes
(Figure 2C). Conversely, for Bmp4, all ICM cells (14/14) showed with FC[ESC/Epiblast] < 0.25, p < 0.01).

Cell Stem Cell 6, 468–478, May 7, 2010 ª2010 Elsevier Inc. 469
 Cell Stem Cell
Single-Cell RNA-Seq Profiling of ICM Outgrowth

Figure 1. Morphology of ICM Outgrowth

Bright field and fluorescence image of mouse E3.5 blastocyst (day 0 ICM outgrowth) (A and B), E4.5 blastocyst (C and D), day 3 ICM outgrowth (E and F), day 5 ICM
outgrowth (G and H), and ESCs (I and J); real-time PCR measured gene expression in single cells of ICM outgrowth: Oct4, Sox2, and Nanog expression in 22 single
E3.5 ICM cells (K), in 10 single E4.5 epiblast cells (L), in 26 single day 3 ICM outgrowth cells (M), in 17 single day 5 ICM outgrowth cells (N), and in 23 single ESCs
(O). The y axis is normalized expression levels based on the mean expression values of all single cells for the given gene.

We found that genes expressed at the mid levels (read per related processes, such as DNA-dependent regulation of
million reads, 1 < RPM < 10), showed higher cell-to-cell varia- transcription (enrichment > 1.5, p < 2.5 3 1014), transcription
tions than the ones expressed at high levels (RPM > 10, Figure 4), factor activity (enrichment > 1.45, p < 2.6 3 1011), transcription
which indicates that the former set of genes have a higher regulator activity (enrichment > 1.51, p < 4.2 3 109), positive
propensity for a more dynamic regulation of expression among regulation of transcription from RNA polymerase II promoter
individual cells of the same type. These genes include Hoxd13, (enrichment > 1.51, p < 2.0 3 107), and transcription repressor
Hoxb3, Hoxb5, and Ddx3y that showed highly variable expres- activity (enrichment > 1.4, p < 1.8 3 105, Table S4). For
sion in ESCs, whereas Gm364, Tmem80, Hdx, Trpm3, Enox2, example, 29 out of the 34 genes of the nucleic acid metabolism
Ilvbl, Has3, Pygm, and Fbxw13 showed a great variation in pathway and 25 out of the 32 genes of lipid metabolism pathway
expression within ICM cells. Some genes, such as Tnk1, Myof, showed dramatic changes in their expression between ICM and
Adamts9, Tspan12, Rhox6, Epha7, Dhrs3, Fam189a1, and ESCs (Figures S5A and S5B). We also found clear enrichment of
Nudt18, showed highly variable expression in both ESCs and cell-fate-related pathways, such as cellular development (24 out
ICM (Table S2). These variations are probably not because of 31 genes), organismal development (27 out of 33 genes), and
technical reasons because genes expressed at low levels cellular assembly and organization (32 out of 35 genes), indi-
(RPM < 1) showed less cell-to-cell variation compared to genes cating a shift in the developmental potential from ICM to ESCs.
expressed at mid levels. For these genes with large coefficient of On the other hand, we found significant similarities in sig-
variations (CV > 1) between cells of ESCs and ICM, Gene naling-related processes (including signal transduction, signal
Ontology (GO) analysis showed that the genes involved in transducer activity, receptor activity, G protein coupled receptor
cellular growth, cellular assembly, amino acid metabolism, and activity, and G protein coupled receptor protein signaling
lipid metabolism were significantly enriched (p < 0.0001). pathway), indicating that their expression is relatively compa-
Next, we wanted to establish what general trends exist in gene rable in ICM and ESCs (Table S4). However, ESCs have the
expression changes during the transition from ICM to ESCs. GO ability to self-renew, whereas ICM does not. Thus, the 4,837
analysis showed that for genes that changed dramatically in genes differentially expressed between ICM and ESCs might
their expression from ICM to ESCs (FC[ESC/ICM] > 4, or < 0.25, provide insights into genes that regulate self-renewal ability
p < 0.01), we found strong enrichment of genes for transcription- and unique metabolism in ESCs. To confirm the reliability of

470 Cell Stem Cell 6, 468–478, May 7, 2010 ª2010 Elsevier Inc.
Cell Stem Cell
Single-Cell RNA-Seq Profiling of ICM Outgrowth

Figure 2. Gene Expression Measured by

Real-Time PCR
Gene expression measured by real-time PCR in
single cells of fourteen ICM (E3.5), six day 3 ICM
outgrowth cells, nine day 5 ICM outgrowth cells,
and 14 ESCs (Table S1).

Alternative Splicing during the ICM

Outgrowth at Whole-Genome Scale
Alternative splicing plays an important
role in defining tissue identity and speci-
ficity. It is estimated that nearly 95% of
the mammalian multiexon genes express
multiple transcript variants through alter-
native splicing (Chen and Manley, 2009).
We wished to know if alternative splicing
was a major feature during the outgrowth
process of ICM toward ESCs. We ad-
dressed the expression dynamics of all
the 6,331 transcript variants from the
2,567 RefSeq genes with multiple known
isoforms, which has not been addressed
previously. 1,852 transcript variants were
expressed (at least 5 counts) in either ICM
or ESCs. And from them, 417 transcript
variants were upregulated (fold change,
FC[ESC/ICM, splicing] > 2, p < 0.01),
and 586 transcript variants were clearly
downregulated (FC[ESC/ICM, splicing]
< 0.5, p < 0.01) (Figure 5; Table S2).
Thus, there was a dramatic change in
54.2% (1,003 out of 1,852) of the ex-
pressed transcript isoforms during ICM
outgrowth. Interestingly, we found that
epigenetic regulators, such as transcript
variants of heterochromatin binding pro-
tein Cbx5 (also known as Hp1a), Setdb1
(also known as Eset), Suv39h2, Ehmt2
(also known as G9a), Suv12, Kdm4a
our observations on differentially expressed genes by our single- (also known as Jmjd3), Sirt2, Smarcb1, and Kat2a (also known
cell RNA-Seq, we compared 108 genes that were detected by as Gcn5), showed between 2- and 132-fold change during ICM
both real-time PCR (Ct < 32) and RNA-Seq (>0.1 RPM) and found outgrowth. More importantly, there were 128 transcript variants
that their correlation coefficient is 0.92, confirming the accuracy expressed in ICM cells that were completely lost in ESCs.
of our single-cell RNA-Seq data (Figure S3). Conversely, 169 variants that were not expressed in ICM were

Figure 3. Transcriptome Analysis of ICM

Outgrowth Cells
(A) The principal component analysis of ICM
outgrowth cells. The nine E3.5 ICM, three E4.5
Epiblast, two day 3 Oct4+Sox2+Nanog+ outgrowth
cells, three day 5 Oct4+Sox2+Nanog+ out-
growth cells, two day 5 Oct4Sox2Nanog
outgrowth cells, and twelve ESCs are indepen-
dently clustered (Table S2).
(B) Expression dynamics of marker genes of early
primordial germ cells (PGCs). Averaged expres-
sion of different individual cells was shown. The
error bar represents the coefficient of variation
(CV) between individual cells.

Cell Stem Cell 6, 468–478, May 7, 2010 ª2010 Elsevier Inc. 471
 Cell Stem Cell
Single-Cell RNA-Seq Profiling of ICM Outgrowth

Figure 4. Plots of the Distribution of Coeffi-

cient of Variation between Individual Cells
158 (A) and 117 (B) genes are detected by TaqMan
real-time PCR (Ct < 32 in at least half of the cells)
in ESC and ICM cells, respectively, and used to
compare cell-to-cell measurement variance.
Density of coefficients of variation (CVs) for Ct
measurements (blue) and RPM measurements
representing RNA-Seq transcripts counts per
million reads (red) across both ESC (A) and ICM
(B) show a small difference between the two plat-
forms. A similar representation is generated using
the entire set of transcripts (24,435) separated
into three categories: highly expressed genes
(RPM > 10, aqua blue), mid-expressed genes
(1 < PRM < 10, blue), and low-expressed
genes (RPM < 1, green). CV density of each group
of transcripts is represented for ESC (C) and ICM
(D). The red curves are the sum of high-, mid-
and low-expressed genes density of CVs. The
mid-expressed genes tend to have higher cell-
to-cell variations (Table S2). For each transcript,
RPMs were used to calculate mean and standard
deviation across cells of the same type. CV was
defined as the ratio between the standard devia-
tion and the mean value.

of self-renewal while retaining pluripo-

tency, it is likely that epigenetic regula-
tors may have an important role during
this process. For this reason, we
clearly expressed in ESCs (Figure S4). Correspondingly, the analyzed the expression of 114 known key epigenetic regulators
expression of tissue-specific alternative splicing factors strongly (Table S4). We found that 37 of them showed strong upregula-
changed (5- to 60-fold) during ICM outgrowth, such as nPTB (also tion and 16 of them were clearly downregulated. Thus 46.5%
known as Ptbp2), Rbm35b (also known as Esrp2), Tia1, Slm2 (53 out of 114) of known epigenetic regulators showed changes
(also known as Khdrbs3), Celf4 (also known as Brunol4), and in expression, which may be necessary to underpin the pheno-
Celf5 (also known as Brunol5). Thus, there were global changes typic changes in these cells. More importantly, the majority of
in the regulation of alternative splicing during ICM outgrowth. the epigenetic regulators that show increased expression are
GO analysis of transcription variants showed that the general cell linked to a repressive epigenetic status. Thus, for DNA methyla-
metabolism-related processes, such as mRNA processing (enrich- tion, Dnmt3a, Dnmt3b, Dnmt3l, Mecp2, and Mbd2 increased
ment > 1.41, p < 0.02), RNA splicing (enrichment > 1.56, p < 0.015), between 2- and 12-fold from ICM to ESCs. Similarly, histone
protein modification processes (enrichment > 2.41, p < 0.0037), deacetylases Hdac5, Hdac6, Hdac7, Hdac11, H3K9 methyl-
cytoskeleton (enrichment > 1.27, p < 0.007), microtubule (enrich- transferase Ehmt1 (also known as Glp), H4K20 methyltransfer-
ment > 1.57, p < 0.008), carbohydrate metabolic processes (enrich- ase Suv420h2, and heterochromatin binding protein Cbx1
ment > 1.65, p < 0.03), modification-dependent protein catabolic (also known as Hp1b) increased from 2- to 40-fold. Conversely,
processes (enrichment > 1.3, p < 0.03), are clearly enriched for a large proportion of the epigenetic modifiers known to confer an
these transcript variants (Table S4). This suggests that one of the active epigenetic status were downregulated. These include
main sources of global regulation of alternative splicing from ICM histone acetyltransferases Ncoa3 (also known as ACTR),
to ESCs probably reflects the accommodation of ICM cells from Crebbp (also known as CBP/P300), Clock, H3K9 demethylases
in vivo embryonic environment to ESC in vitro culture condition, Kdm4a (also known as Jhdm3a), Kdm4d (also known as
which might be important for the regulation of basal cell metabo- Jmjd2d), H3K27 demethylase Kdm6b (also known as Jmjd3),
lism. Furthermore, cell-cycle-related processes (including cell and H3K4 methyltransferase Mll3 (also known as Kmt2c), which
division and mitosis) are also enriched in these significantly up- or showed between 2- and 10-fold downregulation from ICM
downregulated transcript variants, indicating the potential differ- outgrowth to ESCs.
ences in cell-cycle regulation between ICM and ESCs. These results indicate that ESCs probably have a globally
more repressive epigenetic status compared with the ICM. Since
Expression Dynamics of Epigenetic Regulators during ICM cells in vivo undergo rapid and transient phenotypic and
the ICM Outgrowth developmental changes in the course of early postimplantation
Since the process of ICM outgrowths and formation of ESCs development, these cells may require a greater epigenetic flexi-
involves arrest of a normal developmental program and initiation bility. By contrast, the more repressive epigenetic status of ESCs

472 Cell Stem Cell 6, 468–478, May 7, 2010 ª2010 Elsevier Inc.
Cell Stem Cell
Single-Cell RNA-Seq Profiling of ICM Outgrowth

Figure 5. Splice-Specific Differential Expression

The coverage plots of two junctions in (A) ESC, (B) E3.5 ICM, and (C) E4.5 epiblast of Dppa4 gene. The junction counts of Dppa4 transcript variant no. 1
(NM_001018002) are 12.47-fold more in ICM (88 reads, RPM = 150.16) than that in ESC (13 reads, RPM = 12.03), while the junction counts of Dppa4 transcript
variant no. 2 (NM_028610) are 2.2-fold less in the ICM (35 reads, RPM = 59.72) than that in ESC (142 reads, RPM = 131.43, Table S2). RPM, read per million
aligned reads.

is probably important for the maintenance and propagation of The first cluster of genes is upregulated during ICM outgrowth,
their undifferentiated pluripotent state while retaining the such as Nodal, Eras, Lin28, Smad1, Zic3, Id1, Id2, Tcf3, Kit (also
capacity for infinite self-renewal. known as c-Kit), Kitl (also known as Scf), Prdm5, Prdm16, Klf12,
Zfp41, Sox3, Ifitm3 (also known as Fragilis), Pim2, Cdh3 (also
Correlation between Gene Expression, Pluripotency, known as P-cadherin), and Nr0b1 (also known as Dax1); these
Cell Differentiation, and Cell Fate genes correlate positively with the capacity for self-renewal
Next, we examined in greater detail changes in gene expression exhibited by ESCs. Indeed, this trend is confirmed by the anal-
to further assess the differences between ICM and ESCs ysis of cells in day 5 ICM outgrowths that have ceased to express
(Figure S2). Whereas ESCs and ICM are both pluripotent, the pluripotency genes, where this first cluster of genes is downre-
phenotype of ESCs is relatively constant as they undergo self- gulated (Figure S6A). While some of the genes in this cluster
renewal, while ICM cells cannot do so in vivo because they are were shown to be important for pluripotency, we suggest that
poised to undergo further changes according to their develop- at least some of them are probably also important for self-
mental program (Smith, 2006; Niwa, 2007). From this analysis, renewal. In fact, Eras and Nodal that were recovered in this
we found four clusters of interesting genes related to pluripo- cluster are two known regulators that are crucial for the self-
tency and self-renewal (Figure S6). renewal of ESCs but are not essential for pluripotency (Takahashi

Cell Stem Cell 6, 468–478, May 7, 2010 ª2010 Elsevier Inc. 473
 Cell Stem Cell
Single-Cell RNA-Seq Profiling of ICM Outgrowth

et al., 2003; Ogawa et al., 2007). For example, interference with et al., 2006). We found a good correlation between our data and
the Nodal signaling has a strong effect on the self-renewal and that reported previously with respect to 200 genes that were
proliferation of ESCs, but there is little effect on their pluripotency downregulated in both instances (Figure S6E). It is likely that
(Ogawa et al., 2007). The loss of Eras has a similar effect on the our assay captured the earliest responding pluripotency genes,
properties of ESCs (Takahashi et al., 2003). because we compared pluripotent cells and the corresponding
The second cluster of genes is downregulated during ICM earliest differentiated cells cultured under the same conditions.
outgrowth, such as Gata3, Gata4, Gata6, Cdx1, Cdx2, Pramel5, Next, we asked if it is possible to determine the downstream
Pramel6, Pramel7, Sox17, Bmp15, Dppa1, Tbx3, Tbx20, Gdf9, targets of key pluripotency genes by analyzing genes that were
Hoxd8, Gsc, and Klf17. Some of these are developmental genes, downregulated in the day 5 outgrowth that had just ceased to
including Cdx2 and Gata6, which are required for development be pluripotent. We compared our data with the reported effects
of the extraembryonic cells, trophectoderm and primary endo- of knockdown of Nanog, Sox2, Oct4, or Esrrb—genes described
derm, respectively (Ralston and Rossant, 2008; Johnson et al., previously (Ivanova et al., 2006). We found clear overrepresenta-
2006; Koutsourakis et al., 1999), but they are apparently not tion of downregulated genes detected following loss of Nanog
essential for the self-renewal of ESCs (Figure S6B). (p = 2.2 3 1016), Oct4 (p = 2.2 3 1016), Sox2 (p = 2.2 3
The third cluster of genes is highly expressed during ICM 1016), or Esrrb (p = 1.6 3 107) (Figure S6F). Furthermore, there
outgrowth, except in those neighboring cells that cease to be was also strong overrepresentation of upregulated genes
pluripotent as judged by the loss of expression of Oct4, Sox2 following knockdown of Nanog (p = 2.0 3 108), Oct4 (p = 2.5 3
and Nanog, as well as Esrrb, Cdh1 (also known as E-cadherin), 1010), or Sox2 (p = 2.2 3 1016) (Figure S6G). This shows that
Pecam1, Pim1, Pim3, Notch1, Notch4, Fzd9, Frz10, Dazl, our analysis serves as a good indicator of targets of pluripotency
Prdm14, Bmp8b, and Dppa4, demonstrating a positive correla- genes. Moreover, detection of these probable target genes that
tion with the pluripotency of ICM and ESCs (Figure S6C). Indeed, are regulated by corresponding pluripotency genes is cell-
Esrrb, E-cadherin, Pim1, Pim3, and Prdm14 have been shown autonomous as seen by RNA-Seq within an individual cell, which
to be important for pluripotency of ESCs (Ivanova et al., 2006; excludes possible variable responses in individual cells analyzed
Chou et al., 2008; Aksoy et al., 2007; Tsuneyoshi et al., 2008). in bulk cultures.
The fourth cluster of genes is confined to cells that have To gain insight into the gene network underlying pluripotency,
ceased to be pluripotent as evident by the loss of expression we aligned our data to the known gene network for Oct4 associ-
of the Oct4-DPE-GFP reporter. This cluster of genes include ated with embryonic stem cell pluripotency (Loh et al., 2006,
Hoxd8, Bmp1, Bmp2, Tgfbr2, Tgfbr3, Jak2, Fgf3, Fgf10, Fgfr3, 2008; Sharov et al., 2008). We found that this network is enriched
Fgfr4, Sox7, Sox9, Sox17, Nanos1, Cdh5 (also known as VE-cad- in genes whose expression changed as judged by the compar-
herin), and Nkx6.2, showing a negative correlation with the pluri- ison of pluripotent and neighboring nonpluripotent cells present
potency of ESCs (Figure S6D). This cluster represents the within day 5 outgrowths. In fact, 22 out of the 45 Oct4 network
earliest set of differentiation genes that are activated when the genes changed their expression significantly by more than
cells have just ceased to be pluripotent. For example, Fgf 4-fold. Of these 22 genes, only two genes, namely Rb1 and
signaling has been shown to promote differentiation of ESCs Foxa2, were not directly regulated by Oct4 (Figure 6; Table S4).
(Kunath et al., 2007; Stavridis et al., 2007). Similarly, Sox7 and This suggests that through single-cell digital transcriptome
Sox17 also drive ESC differentiation (Séguin et al., 2008). approach, we captured a large proportion of genes that are
We also examined expression of primordial germ cell-specific directly regulated by Oct4 in the network (p = 4.02 3 106)
genes during the course of blastocyst outgrowths (Zwaka and (Loh et al., 2006; Sharov et al., 2008). To our knowledge, this
Thomson, 2005). We looked at the expression dynamics of early is the first time that the Oct4 gene network is validated in an
PGC markers during this process (Figure 3B; Table S2). We individual cell, proving the cell-autonomous regulation of the
detected upregulation of Ifitm3 (also known as Fragilis), network, which is an essential prerequisite for direct interactions
Prdm14, Ddx4 (also known as Vasa) and Nanos3, but Dppa3 within the gene network. Furthermore, we also aligned our data
(also known as stella) was downregulated. We also found upre- to the known gene network for human embryonic stem cell
gulation of Prdm1 (also known as Blimp1), but later in ESCs, pluripotency and found a similar enrichment pattern. Forty-five
and the levels were highly variable in individual cells (Table S2). out of the 137 human ESC network genes significantly changed
However, all these genes were repressed in the Oct4 negative their expression (p = 2.18 3 104, Figure S5C), suggesting
outgrowth cells that had lost pluripotency. Further studies are conservation of the core aspects of the gene network of pluripo-
required to assess the significance of these observations, if tency between mouse and human, despite some known differ-
any, for the derivation and properties of ESCs. ences between them.

Functional Downstream Genes of Pluripotency Segregation of MicroRNAs Repressing Early

Master Genes Differentiation and Those Repressing Pluripotency
The RNA-Seq analysis we have carried out provides an opportu- We next wanted to see if microRNAs (miRNAs) have a role in the
nity to identify pluripotency genes and their downstream targets, ICM outgrowths as they develop toward ESCs (Zamore and
by comparing neighboring cells that are pluripotent to those that Haley, 2005; Gangaraju and Lin, 2009). Expression analysis of
show a loss of pluripotency under the same culture conditions. 330 miRNAs showed similar expression of pluripotency related
These gene sets can also be compared with the expression of miR-290 to -295 cluster in ICM and ESCs (Table S5; Figure 7A),
the earliest genes that are downregulated when ESCs are while a set of 51 out of 330 miRNAs showed differential expres-
induced to differentiate in response to retinoic acid (RA) (Ivanova sion (Figure 7B). We found that let-7a, let-7e, let-7f, and let-7g

474 Cell Stem Cell 6, 468–478, May 7, 2010 ª2010 Elsevier Inc.
Cell Stem Cell
Single-Cell RNA-Seq Profiling of ICM Outgrowth

Figure 6. Gene Network Analysis of Oct4 in Embryonic Stem Cell Pluripotency Pathway
The 22 genes (including Oct4) up/downregulated for more than 4-fold when the pluripotent cells lose pluripotency (FC[day 5 Oct4+/day 5 Oct4] > 4 or < 0.25,
p < 0.01) were shown in red. The gray colored genes have FC[day 5 Oct4+/day 5 Oct4] < 4 and FC[day 5 Oct4+/day 5 Oct4] > 0.25 (Table S4). The p value was
estimated using Ingenuity systems software (http://www.ingenuity.com). FC, fold change.

were reduced between 4- and 12-fold in ESCs compared to ICM, preferably target the ESC-specific pluripotency genes
which is reflected in the reduced levels of let-7 expression (FC[ESC/ICM] > 4) (miR-669b, -298, -692, -204, -28, -149,
required for the self-renewal and maintenance of undifferenti- -34a, -182, [-129-5p, -133a, -320; the target enrichment is
ated cancer stem cells (Yu et al., 2007). Correspondingly, we 1.6-fold, p < 1.3 3 108). As a control, the targets of these eleven
found 5-fold upregulation of Lin28, which is the suppressor of miRNAs show no enrichment (p > 0.03) in ICM-specific genes
let-7 family miRNAs (Viswanathan et al., 2008). Recently, Card (FC[ESC/ICM] < 0.25). The loss of this class of miRNAs may
et al. (2008) reported that miR-302 cluster in human ESC is contribute to the phenotype of resistance to differentiation
actively regulated by Oct4 and Sox2, while miR-302a targets when Dicer or DGCR8, two key components of the miRNA
cell-cycle regulators and promotes an ESC-like cell cycle. Our processing pathway, are knocked out in established ESCs
miRNA profiling indicated that two members of miR-302 cluster, (Kanellopoulou et al., 2005; Wang et al., 2007). Taken together,
miR-302c and miR-367, increased 5- and 33-fold, respectively, miRNAs may contribute to ESC’s ability to maintain the balance
from ICM to ESCs, which may contribute to the capacity for between pluripotency and the potential for rapid differentiation,
self-renewal of ESCs. through one set of miRNAs targeting genes that drive differenti-
To explore the potential roles of miRNAs in regulating ESCs, ation, while a separate set of miRNAs target ESC-specific pluri-
we used the target prediction algorithms of PicTar, Miranda, or potency genes.
TargetScan (Krek et al., 2005; Lewis et al., 2005; Enright et al.,
2003). To reduce possible ‘‘prediction noise,’’ we considered Conclusion
only those targets that are predicted by at least two algorithms. Our study provides insight into the dynamic molecular changes
For the miRNAs highly expressed in both ICM and ESCs, there that accompany cell-fate changes. During the conversion of
are two classes: the first class of miRNAs preferably target early ICM cells to ESCs, there is an evident arrest of a normal develop-
differentiation genes (FC[Day5 Oct4+/Oct4] < 0.25, r < 0.6, mental program, which is subverted in vitro in favor of a potential
Table S6 and S7) (miR-19b, -19a, -106a, -20b, -106b, -9, -103, for unrestricted self-renewal while retaining the ability to undergo
-107, -124a, -145; the target enrichment is 2.0-fold, p < 3.8 3 differentiation into all the diverse cell types. We demonstrate
106). As a control, the targets of these ten miRNAs show no how both the retention of expression of key genes allows inher-
enrichment (p > 0.02) for pluripotency-related genes (FC[Day5 itance of a fundamental property of the ICM, namely pluripo-
Oct4+/Oct4] > 4, r > 0.6). The loss of this class of miRNAs tency, while other changes in the transcriptome permit exit
may contribute to the phenotype of loss of pluripotent Oct4- from a normal developmental program and confer a key property
positive epiblast cells when Dicer is knocked out in early of self-renewal. Changes in epigenetic regulators apparently
embryos (Bernstein et al., 2003). The second class of miRNAs allow for the stability of the newly acquired epigenotype, which

Cell Stem Cell 6, 468–478, May 7, 2010 ª2010 Elsevier Inc. 475
 Cell Stem Cell
Single-Cell RNA-Seq Profiling of ICM Outgrowth

Figure 7. MicroRNA Expression in ICM and

ESCs
(A) miR-290 295 cluster microRNA expression
in ICM and ESCs.
(B) microRNAs showing significant differential
expression between ICM and ESCs (p value <
0.01). The relative expression levels were shown
(Table S5). The error bar represents the standard
deviation calculated from three biological repli-
cates.

progenies. The inner core of cells in the outgrowth

was treated with EGTA-PBS for 10 min at room
temperature and trypsin for 5 min at 37 C. The
core of cells was dissociated by pipetting into
a single-cell suspension in GMEM medium with
15% FCS. Next, the GFP-positive and -negative
cells were separated manually under a fluores-
cence microscope. The single cells were washed
in BSA-PBS twice before they were picked individ-
ually for subsequent analysis.

Preparation of Single-Cell cDNAs

The single-cell RNA-seq method has been
described in detail previously (Tang et al., 2009,
2010). In brief, an individual cell was manually
picked and transferred into lysate buffer by
is crucial for the inherent plasticity of ESCs. The conversion from a mouth pipette, followed by reverse transcription directly on the whole-cell
ICM to ESC is also coupled with a role for distinct sets of miRNAs lysate. Following this procedure, terminal deoxynucleotidyl transferase was
that allow for both self-renewal while the cells retain the ability to used to add a poly(A) tail to the 30 end of first-strand cDNAs, which was fol-
lowed by 20 + 9 cycles of PCR to amplify the single-cell cDNAs.
respond rapidly to cues for differentiation. Our investigation may
serve as a paradigm for other studies, including regulation and
RNA-Seq Library Preparation, Sequencing, and Alignment
differentiation of small numbers of stem cells in adults. Our After generation of the target cDNA from a single cell, 100 ng cDNA (0.5–3 kb)
approach is applicable to studies on small groups of differenti- was sheared into 80–130 bp fragments. P1 and P2 adaptors were ligated to
ating cells and for gaining insight into how developmental each end, and the fragments were subjected to 8–10 cycles of PCR amplifica-
programs might be undermined, leading to the formation of tion. Emulsion PCR reactions were performed by combining 1.6 billion 1 mm
diseased tissues, including cancers. diameter beads that had P1 primers covalently attached to their surfaces
with 500 pg of single-cell libraries. Applied Biosystems SOLiD sequencer
generated 50-base sequences, and AB’s whole transcriptome software tools
EXPERIMENTAL PROCEDURES were used to analyze the sequencing reads (http://solidsoftwaretools.com/gf/
project/transcriptome/). The reads obtained from each cell were matched to
Isolation of Embryos and Single Cells the Mouse genome (mm 9, NCBI Build 37) and reads that aligned uniquely
All embryos were recovered from 129 females mated with Oct4-DPE-GFP were used in the downstream analysis. These reads were used to create
transgenic male mice. The transgenic GFP expression of the reporter is under base coverage files (in a wiggle format), which can be viewed directly in the
the control of Oct4 promoter and distal enhancer, but the proximal enhancer UCSC genome browser, or to detect known or novel exon-exon junctions.
region is deleted. This GFP transgene reporter shows expression in the E3.5 Unambiguously mapped reads were first used to generate exon counts and
ICM and E4.5 Epiblast of blastocysts and PGC in vivo and in ESC (Yeom then transcript or gene counts. Feature counts were normalized using the
et al., 1996). E3.5 and E4.5 blastocysts were flushed from the uterus of 129 RPM (read per million aligned reads) method, and no adjustment to gene/
pregnant females. For ESC outgrowth, E3.5 blastocysts were cultured in transcript size was made because our protocol has a limited coverage of
KSOM medium for the first day and then transferred to GMEM medium 0.5–3 kb from the 30 end of the transcripts. An alternative analysis was used
(GIBCO, cat. no. 21710-025) with 15% Fetal Calf Serum (FCS) (GIBCO, cat. for alignments that were not aligned to their full length, where reads were
no. 16000-044) and 1000 U/ml Lif on mitomycin C-treated MEF feeder cells aligned to a reference containing exon-exon junctions, using 42 bases on
for all later periods. The time when the E3.5 blastocysts were placed into each side for junctions, allowing up to four mismatches for the full length
culture was designated as day 0. of the read (50 bases) (Tang et al., 2009). The quality of the single-cell
For the isolation of single cells of E3.5 ICM or E4.5 epiblast, the blastocysts RNA-Seq data was analyzed (Figure S7). These analyses showed that our
were first placed in a mouse trophoblast antibody for 30 min. Then they were single-cell RNA-Seq data are highly reproducible, reliable, and accurate for
treated by complement for 30 min. After this, the lysed trophectoderm cells ICM outgrowth and ESCs.
were removed and the isolated ICM or epiblast was placed in EGTA-PBS for
10 min. After that, they were furthered treated by Trypsin at 37 C for 5 min. Real-Time PCR
Then they were transferred into GMEM medium with 15% FCS and dissoci- For TaqMan real-time PCR, 1.0 ml of diluted cDNAs was used for each 10 ml
ated into single-cell suspension. The resulting single cells were washed in real-time PCR (13 PCR Universal Master Mix, 250 nM TaqMan probe, 900
BSA-PBS twice and prepared to be picked as single cells. nM of each primer, that are commercially available as ready to use Assays,
For the isolation of blastocyst/ICM outgrowth, it was treated by trypsin for custom-plated in 384-plates or TaqMan low Density Array cards by Applied
5 min to dissociate the core part of outgrowth from surrounding trophectoderm Biosystems). All reactions were duplicated. The PCR was done as following

476 Cell Stem Cell 6, 468–478, May 7, 2010 ª2010 Elsevier Inc.
Cell Stem Cell
Single-Cell RNA-Seq Profiling of ICM Outgrowth

using an AB7900 with 384-well plates: first, 95 C for 10 min to activate the Taq Chou, Y.F., Chen, H.H., Eijpe, M., Yabuuchi, A., Chenoweth, J.G., Tesar, P.,
polymerase, then 40 cycles of 95 C for 15 sec and 60 C for 1 min. Lu, J., McKay, R.D., and Geijsen, N. (2008). The growth factor environment
defines distinct pluripotent ground states in novel blastocyst-derived stem
MicroRNA Profiling of ICM and ESCs cells. Cell 135, 449–461.
The detailed protocol is described previously (Tang et al., 2006). In brief, Cloonan, N., Forrest, A.R., Kolle, G., Gardiner, B.B., Faulkner, G.J., Brown,
10 cells were picked into a PCR tube by glass capillary and were lysed by M.K., Taylor, D.F., Steptoe, A.L., Wani, S., Bethel, G., et al. (2008). Stem cell
heat treatment at 95 C for 5 min. Then the microRNAs were reverse tran- transcriptome profiling via massive-scale mRNA sequencing. Nat. Methods
scribed into cDNAs by pool of 330 of stem-looped primers. After this, these 5, 613–619.
microRNA cDNAs were amplified by 18 cycles of PCR by 330 forward primers Enright, A.J., John, B., Gaul, U., Tuschl, T., Sander, C., and Marks, D.S. (2003).
and a universal reverse primer. Finally the cDNAs were split and each individual MicroRNA targets in Drosophila. Genome Biol. 5, R1.
microRNA was measured by TaqMan probe-directed real-time PCR. Three
Evans, M.J., and Kaufman, M.H. (1981). Establishment in culture of pluripoten-
biological replicates were done for each type of cell.
tial cells from mouse embryos. Nature 292, 154–156.
Fujikura, J., Yamato, E., Yonemura, S., Hosoda, K., Masui, S., Nakao, K.,
ACCESSION NUMBERS
Miyazaki Ji, J., and Niwa, H. (2002). Differentiation of embryonic stem cells
is induced by GATA factors. Genes Dev. 16, 784–789.
All of the raw data of single-cell RNA-Seq for ICM outgrowth have been depos-
ited in the high-throughput sequencing database of GEO (accession number Gangaraju, V.K., and Lin, H. (2009). MicroRNAs: key regulators of stem cells.
GSE20187). Nat. Rev. Mol. Cell Biol. 10, 116–125.
Ivanova, N., Dobrin, R., Lu, R., Kotenko, I., Levorse, J., DeCoste, C., Schafer,
X., Lun, Y., and Lemischka, I.R. (2006). Dissecting self-renewal in stem cells
SUPPLEMENTAL INFORMATION
with RNA interference. Nature 442, 533–538.
The Supplemental Information includes seven figures and seven tables and Johnson, B.V., Rathjen, J., and Rathjen, P.D. (2006). Transcriptional control
can be found with this article online at doi:10.1016/j.stem.2010.03.015. of pluripotency: decisions in early development. Curr. Opin. Genet. Dev. 16,
447–454.
ACKNOWLEDGMENTS Kaji, K., Caballero, I.M., MacLeod, R., Nichols, J., Wilson, V.A., and Hendrich,
B. (2006). The NuRD component Mbd3 is required for pluripotency of embry-
We would like to thank J. Bodeau, E. Dimalanta, S. Dadi, E. Langit, P. Brzoska, onic stem cells. Nat. Cell Biol. 8, 285–292.
J. Au-Young, and J. Sherlock for performing sequencing runs, design, and Kaji, K., Nichols, J., and Hendrich, B. (2007). Mbd3, a component of the NuRD
preparation of real-time PCR assays. We would also like to thank K. Hayashi, co-repressor complex, is required for development of pluripotent cells.
S. Jeffery, and W. Mifsud for insightful discussions. The work was supported Development 134, 1123–1132.
by a grant from the Wellcome Trust to M.A.S.
Kanellopoulou, C., Muljo, S.A., Kung, A.L., Ganesan, S., Drapkin, R.,
Jenuwein, T., Livingston, D.M., and Rajewsky, K. (2005). Dicer-deficient
Received: December 23, 2009
mouse embryonic stem cells are defective in differentiation and centromeric
Revised: February 5, 2010
silencing. Genes Dev. 19, 489–501.
Accepted: March 10, 2010
Published: May 6, 2010 Koutsourakis, M., Langeveld, A., Patient, R., Beddington, R., and Grosveld, F.
(1999). The transcription factor GATA6 is essential for early extraembryonic
development. Development 126, 723–732.
REFERENCES
Krek, A., Grün, D., Poy, M.N., Wolf, R., Rosenberg, L., Epstein, E.J.,
Aksoy, I., Sakabedoyan, C., Bourillot, P.Y., Malashicheva, A.B., Mancip, J., MacMenamin, P., da Piedade, I., Gunsalus, K.C., Stoffel, M., and Rajewsky,
Knoblauch, K., Afanassieff, M., and Savatier, P. (2007). Self-renewal of murine N. (2005). Combinatorial microRNA target predictions. Nat. Genet. 37, 495–
embryonic stem cells is supported by the serine/threonine kinases Pim-1 and 500.
Pim-3. Stem Cells 25, 2996–3004. Kunath, T., Saba-El-Leil, M.K., Almousailleakh, M., Wray, J., Meloche, S., and
Bao, S., Tang, F., Li, X., Hayashi, K., Gillich, A., Lao, K., and Surani, M.A. Smith, A. (2007). FGF stimulation of the Erk1/2 signalling cascade triggers
(2009). Epigenetic reversion of post-implantation epiblast to pluripotent transition of pluripotent embryonic stem cells from self-renewal to lineage
embryonic stem cells. Nature 461, 1292–1295. commitment. Development 134, 2895–2902.
Bernstein, E., Kim, S.Y., Carmell, M.A., Murchison, E.P., Alcorn, H., Li, M.Z., Lewis, B.P., Burge, C.B., and Bartel, D.P. (2005). Conserved seed pairing,
Mills, A.A., Elledge, S.J., Anderson, K.V., and Hannon, G.J. (2003). Dicer is often flanked by adenosines, indicates that thousands of human genes are
essential for mouse development. Nat. Genet. 35, 215–217. microRNA targets. Cell 120, 15–20.
Card, D.A., Hebbar, P.B., Li, L., Trotter, K.W., Komatsu, Y., Mishina, Y., and Loh, Y.H., Wu, Q., Chew, J.L., Vega, V.B., Zhang, W., Chen, X., Bourque, G.,
Archer, T.K. (2008). Oct4/Sox2-regulated miR-302 targets cyclin D1 in human George, J., Leong, B., Liu, J., et al. (2006). The Oct4 and Nanog transcription
embryonic stem cells. Mol. Cell. Biol. 28, 6426–6438. network regulates pluripotency in mouse embryonic stem cells. Nat. Genet.
Chambers, I., Colby, D., Robertson, M., Nichols, J., Lee, S., Tweedie, S., and 38, 431–440.
Smith, A. (2003). Functional expression cloning of Nanog, a pluripotency Loh, Y.H., Ng, J.H., and Ng, H.H. (2008). Molecular framework underlying
sustaining factor in embryonic stem cells. Cell 113, 643–655. pluripotency. Cell Cycle 7, 885–891.
Chambers, I., Silva, J., Colby, D., Nichols, J., Nijmeijer, B., Robertson, M., Mitsui, K., Tokuzawa, Y., Itoh, H., Segawa, K., Murakami, M., Takahashi, K.,
Vrana, J., Jones, K., Grotewold, L., and Smith, A. (2007). Nanog safeguards Maruyama, M., Maeda, M., and Yamanaka, S. (2003). The homeoprotein
pluripotency and mediates germline development. Nature 450, 1230–1234. Nanog is required for maintenance of pluripotency in mouse epiblast and ES
Chen, M., and Manley, J.L. (2009). Mechanisms of alternative splicing regula- cells. Cell 113, 631–642.
tion: insights from molecular and genomics approaches. Nat. Rev. Mol. Cell Morrisey, E.E., Tang, Z., Sigrist, K., Lu, M.M., Jiang, F., Ip, H.S., and Parmacek,
Biol. 10, 741–754. M.S. (1998). GATA6 regulates HNF4 and is required for differentiation of
Chen, X., Xu, H., Yuan, P., Fang, F., Huss, M., Vega, V.B., Wong, E., Orlov, Y.L., visceral endoderm in the mouse embryo. Genes Dev. 12, 3579–3590.
Zhang, W., Jiang, J., et al. (2008). Integration of external signaling pathways Murchison, E.P., Partridge, J.F., Tam, O.H., Cheloufi, S., and Hannon, G.J.
with the core transcriptional network in embryonic stem cells. Cell 133, (2005). Characterization of Dicer-deficient murine embryonic stem cells.
1106–1117. Proc. Natl. Acad. Sci. USA 102, 12135–12140.

Cell Stem Cell 6, 468–478, May 7, 2010 ª2010 Elsevier Inc. 477
 Cell Stem Cell
Single-Cell RNA-Seq Profiling of ICM Outgrowth

Nichols, J., Silva, J., Roode, M., and Smith, A. (2009). Suppression of Erk Takahashi, K., Mitsui, K., and Yamanaka, S. (2003). Role of ERas in promoting
signalling promotes ground state pluripotency in the mouse embryo. Develop- tumour-like properties in mouse embryonic stem cells. Nature 423, 541–545.
ment 136, 3215–3222. Tang, F., Hajkova, P., Barton, S.C., O’Carroll, D., Lee, C., Lao, K., and Surani,
Niwa, H. (2007). How is pluripotency determined and maintained? Develop- M.A. (2006). 220-plex microRNA expression profile of a single cell. Nat. Protoc.
ment 134, 635–646. 1, 1154–1159.
O’Carroll, D., Erhardt, S., Pagani, M., Barton, S.C., Surani, M.A., and Tang, F., Barbacioru, C., Wang, Y., Nordman, E., Lee, C., Xu, N., Wang, X.,
Jenuwein, T. (2001). The polycomb-group gene Ezh2 is required for early Bodeau, J., Tuch, B.B., Siddiqui, A., et al. (2009). mRNA-Seq whole-transcrip-
mouse development. Mol. Cell. Biol. 21, 4330–4336. tome analysis of a single cell. Nat. Methods 6, 377–382.
Ogawa, K., Saito, A., Matsui, H., Suzuki, H., Ohtsuka, S., Shimosato, D.,
Tang, F., Barbacioru, C., Nordman, E., Li, B., Xu, N., Bashkirov, V.I., Lao, K.,
Morishita, Y., Watabe, T., Niwa, H., and Miyazono, K. (2007). Activin-Nodal
and Surani, M.A. (2010). RNA-Seq analysis to capture the transcriptome
signaling is involved in propagation of mouse embryonic stem cells. J. Cell
landscape of a single cell. Nat. Protoc. 5, 516–535.
Sci. 120, 55–65.
Tsuneyoshi, N., Sumi, T., Onda, H., Nojima, H., Nakatsuji, N., and Suemori, H.
Ralston, A., and Rossant, J. (2008). Cdx2 acts downstream of cell polarization
(2008). PRDM14 suppresses expression of differentiation marker genes
to cell-autonomously promote trophectoderm fate in the early mouse embryo.
in human embryonic stem cells. Biochem. Biophys. Res. Commun. 367,
Dev. Biol. 313, 614–629.
899–905.
Séguin, C.A., Draper, J.S., Nagy, A., and Rossant, J. (2008). Establishment of
endoderm progenitors by SOX transcription factor expression in human Viswanathan, S.R., Daley, G.Q., and Gregory, R.I. (2008). Selective blockade
embryonic stem cells. Cell Stem Cell 3, 182–195. of microRNA processing by Lin28. Science 320, 97–100.
Sharov, A.A., Masui, S., Sharova, L.V., Piao, Y., Aiba, K., Matoba, R., Xin, L., Wang, Y., Medvid, R., Melton, C., Jaenisch, R., and Blelloch, R. (2007). DGCR8
Niwa, H., and Ko, M.S. (2008). Identification of Pou5f1, Sox2, and Nanog is essential for microRNA biogenesis and silencing of embryonic stem cell
downstream target genes with statistical confidence by applying a novel self-renewal. Nat. Genet. 39, 380–385.
algorithm to time course microarray and genome-wide chromatin immunopre- Yeom, Y.I., Fuhrmann, G., Ovitt, C.E., Brehm, A., Ohbo, K., Gross, M., Hübner,
cipitation data. BMC Genomics 9, 269. K., and Schöler, H.R. (1996). Germline regulatory element of Oct-4 specific for
Shen, X., Liu, Y., Hsu, Y.J., Fujiwara, Y., Kim, J., Mao, X., Yuan, G.C., and the totipotent cycle of embryonal cells. Development 122, 881–894.
Orkin, S.H. (2008). EZH1 mediates methylation on histone H3 lysine 27 and
Ying, Q.L., Nichols, J., Chambers, I., and Smith, A. (2003). BMP induction of
complements EZH2 in maintaining stem cell identity and executing pluripo-
Id proteins suppresses differentiation and sustains embryonic stem cell
tency. Mol. Cell 32, 491–502.
self-renewal in collaboration with STAT3. Cell 115, 281–292.
Smith, A. (2006). A glossary for stem-cell biology. Nature 441, 1060.
Ying, Q.L., Wray, J., Nichols, J., Batlle-Morera, L., Doble, B., Woodgett, J.,
Stavridis, M.P., Lunn, J.S., Collins, B.J., and Storey, K.G. (2007). A discrete Cohen, P., and Smith, A. (2008). The ground state of embryonic stem cell
period of FGF-induced Erk1/2 signalling is required for vertebrate neural self-renewal. Nature 453, 519–523.
specification. Development 134, 2889–2894.
Yu, F., Yao, H., Zhu, P., Zhang, X., Pan, Q., Gong, C., Huang, Y., Hu, X., Su, F.,
Tahiliani, M., Koh, K.P., Shen, Y., Pastor, W.A., Bandukwala, H., Brudno, Y.,
Lieberman, J., and Song, E. (2007). let-7 regulates self renewal and tumorige-
Agarwal, S., Iyer, L.M., Liu, D.R., Aravind, L., and Rao, A. (2009). Conversion
nicity of breast cancer cells. Cell 131, 1109–1123.
of 5-methylcytosine to 5-hydroxymethylcytosine in mammalian DNA by MLL
partner TET1. Science 324, 930–935. Zamore, P.D., and Haley, B. (2005). Ribo-gnome: the big world of small RNAs.
Takahashi, K., and Yamanaka, S. (2006). Induction of pluripotent stem cells Science 309, 1519–1524.
from mouse embryonic and adult fibroblast cultures by defined factors. Cell Zwaka, T.P., and Thomson, J.A. (2005). A germ cell origin of embryonic stem
126, 663–676. cells? Development 132, 227–233.

478 Cell Stem Cell 6, 468–478, May 7, 2010 ª2010 Elsevier Inc.