Identification of factors involved in the maintenance of embryonic stem cell self renewal and pluripotency

Chapter1

Introduction








1

1.1 Embryonic stem cells

1.1.1 Derivation and definition of embryonic stem cells

During the transition of an embryo from the morula to the blastocyst stage, a
differentiation event partitions a developing embryo into its extraembryonic and
embryonic components. The outer layer of cells of the blastocyst is the extraembryonic
components and forms an epithelium, the trophectoderm. Descendents of the
trophectoderm are restricted to the generation of the trophoblast components of the
placenta. The embryonic component, located on the interior is referred to as the inner cell
mass (ICM) and this embryonic component comprises a stem cell population. The ICM
represents a unique transitory cellular structure that differentiates into epiblast (also
known as embryonic ectoderm) derived cell types, namely the mesodermal, endodermal
and ectodermal lineages in a regulated manner during the course of embryo development.
Alternatively, the ICM may differentiate into the primitive endoderm which gives rise to

extraembryonic tissues instead of the pluripotent epiblast during this other differentiative
event. The three germ layers are the precursors to all tissues of the body. Ectoderm
develops into the epidermis, retina, brain and nervous system etc, while the mesoderm
gives rise to the bones, muscles and most of the cardiac and circulatory system.
Endoderm gives rise to the respiratory organs and gastrointestinal tract. Cells of the ICM
also have the potential to develop into germ cells (Chambers and Smith, 2004).

2

Of great significance is the fact that explant cultures of ICM can generate
pluripotent embryonic stem (ES) cell lines. The derivation of pluripotent cell lines from
blastocysts was first achieved in the murine system in 1981 by Martin Evans and
Matthew Kaufman and independently by Gail R. Martin (Evans and Kaufman, 1981;
Martin, 1981). Subsequently, successful derivation and propagation of rodent, rabbit,
primate and human ES cell lines have been reported (Fan and Collodi, 2006; Iannaccone
et al., 1994; Reubinoff et al., 2000; Thomson and Marshall, 1998; Vackova et al., 2007;
Thomson et al., 1998).

While ES cells are generally considered an in vitro phenomenon and are not true
equivalents of the ICM, ES cells are at least ICM-like in terms of their ability to
differentiate into cells of all three germ layers. This pluripotent nature of mouse ES cells
was first demonstrated by their ability to contribute to all tissues of adult mice, including
the germ line following injection into host blastocysts. Also, these cells can also form
teratomas - benign tumours which consist of a mixture of differentiated cell types from
the three germ layers following ectopic engraftment to immune-compromised mice
(Reubinoff et al., 2000). In the in vitro scenario, ES cells likewise display a remarkable
capacity to form a plethora of differentiated cell types of the three germ layers in culture.

Another defining feature of ES cells is their acquired capacity to proliferate
indefinitely in vitro without undergoing senescence while retaining a normal karyotype.

This characteristic, coupled with the potential to differentiate into more than 200 unique
cell types render ES cells a very attractive cell source for potential use in regenerative
3

medicine. The successful isolation and propagation of human ES cells in November 1998
by Thomson et al. further brought this promise of ES cell based therapy one step closer to
realization (Thomson et al., 1998).

1.1.2 Applications of ES cells


ES cells can contribute to the formation of chimeric organisms following
engraftment into a host ICM, and continue to give rise to all three germ lineages. This
potential for germline transmission in chimeras, coupled with the amenability of mouse
ES cells to genetic manipulation enable the generation of knock-out mice. As such,
mouse ES cells have since found widespread applications in applied pharmacogenetics
and basic functional genomics research (Pease and Williams, 1990; Tesar, 2005; Voss et
al., 1997; Wolf et al., 1994) .

As for the potential therapeutic application of ES cells, ES cell based therapies
and transplantation have been the most frequently discussed. This has yet to be realized
in clinic due to caveats such as less than optimal efficiency of in vitro differentiation, the
danger of tumorigenecity of transplanted tissues and graft versus host response triggered
in recipients upon recognition of non-autologous hES cells-derived cells as foreign.

Once
these limitations can be overcome, it is likely that monocellular deficiency states such as
Parkinson's disease and type I diabetes will be among the first examples of intractable
diseases that can be rectified through ES cell based replacement therapies (Burns et al.,
2006; Fukuda and Takahashi, 2005; Takagi et al., 2005).

4

Apart from cellular therapy, ES cell technology can also prove to be very useful
in many other aspects of medicine. For example, the availability of ES cell lines, coupled
with the recent development of various differentiative and purification regimes for the
generation of a broad spectrum of lineages from ES cells, have opened up exciting
opportunities to model mammalian embryonic development in vitro. The study of events
regulating the earliest stages of lineage induction and specification is tedious in mouse
embryos and prohibited in the human embryo due to ethical concerns. ES cell based
models provide a convenient means around these limitations. Understanding the events
that occur at the first stages of development has potential clinical significance for
preventing or treating birth defects, infertility and pregnancy loss. A thorough knowledge
of normal development could ultimately allow the prevention or treatment of abnormal
human development. For instance, testing drugs on cultured human embryonic stem cells
could help reduce the risk of drug-related birth defects (Seiler et al., 2006).

The understanding gained from the study of stem cell biology may also
profoundly improve the treatment of cancer as mechanistic links between ES cell self-
renewal and cancer stem cell proliferation might make it possible to improve the
treatment of cancer by targeting inappropriately activated self-renewal pathways.

Investigation of a number of human diseases is severely constrained by a lack of
animal and cell culture models. For instance, a number of pathogenic viruses including
human immunodeficiency virus and hepatitis C virus grow only in human or chimpanzee
5

cells. Primate and human ES cells might provide cell and tissue types that will greatly
accelerate investigation into some of these viral diseases (Yamamoto et al., 2003).

In short, the prospect for medical application of ES cell technology is vast and

promising. One great impediment to the unleashing of the immense potential of ES cells
is however the current incomplete understanding of the molecular mechanism underlying
self renewal and cell fate determination of ES cells.

1.2 Molecular basis underlying ES cell self renewal

Despite the importance of stem cell self renewal, we are only beginning to
understand how it is regulated. ES cell self renewal is a complex process that involves
both the proliferation and maintenance of pluripotency. Self renewal is an intricate
interplay instigated by instructive and permissive instructions provided by signaling
molecules in the microenvironment and also by intracellular regulators such as
transcriptional factors. Multiple factors are required to act in concert to maintain the
embryonic stem cell phenotype. Some factors regulate only proliferation, while others
regulate developmental potential and prevent differentiation. Some determinants regulate
proliferation and inhibit differentiation (Molofsky et al., 2004).

To date, several key signaling pathways, such as the LIF/gp130/STAT3, bone
morphogenetic protein (BMP) and wingless-type MMTV integration site (WNT) have
been established as important pathways maintaining ES cell self renewal and preventing
6

cell differentiation. Intracellularly, the transcriptional regulatory circuit governed by
OCT4, SOX2 and NANOG, have also been established as being crucial for maintaining
ES cells in an undifferentiated state.

1.2.1 Extrinsic regulators and signaling pathways governing ES cell self renewal

1.2.1.1 LIF/gp130/STAT3 pathway

When mouse ES cells were first derived in the 1980s, they were propagated in co-

culture with a layer of fibroblast in the presence of serum. An indication that fibroblasts
act by secreting a signal that inhibits ES cell differentiation was substantiated by the
ability of Buffalo rat liver cell line conditioned medium to replace the fibroblast
requirement (Smith and Hooper, 1987). Leukeamia inhibitory factor (LIF) was
subsequently identified to be the active component of the conditioned medium through
fractionation (Smith et al., 1988). Fibroblasts carrying deletions in Lif were also found to
have reduced capacity to support ES cells. This further supports the fact that LIF is a
major determinant of the ability of feeders to support ES cell self renewal. Today, most
investigators culture mouse ES cells feeder free in the presence of serum with
supplementation of LIF.

LIF is a member of the IL6 family of cytokines that signals through the
transmembrane receptor, gp130. LIF maintains ES cell self renewal through the
activation of STAT3, a member of the signal transducer and activator of transcription
7

(Stat) family (Raz et al., 1999). Evidence strongly indicates that STAT3 is the key
transcription factor downstream of the LIF/gp130 pathway. Forced expression of a
dominant negative Stat3 mutant caused mouse ES cell differentiation even in the
presence of LIF (Niwa et al., 1998). Also, point mutation of the tyrosine residue of gp130
responsible for STAT3 binding abrogated the ability of LIF to maintain self renewal. It
was also shown that mouse ES cells expressing a fusion molecule consisting of STAT3
and estrogen receptor could be maintained in the presence of estrogen derivative
tamoxifen (Matsuda et al., 1999) which translocates the fusion LIF to the nucleus.

The first event involved in the LIF signaling cascade involves binding of LIF to
the LIF receptor (LIFR) that contains a long cytoplasmic tail with a homology to the
gp130. The LIF/LIFR complex then recruits gp130 to form a trimeric complex (Zhang et
al., 1997). Heterodimer formation of the LIF receptor and gp130 receptor then results in
the activation of the tyrosine kinase JAK. The activated JAK phosphorylates tyrosine

residues of gp130 which then serves as a docking site for STAT3. STAT3 is then
activated and translocated into the nucleus to elicit transcriptional responses that prevent
differentiation (Niwa et al., 1998).


LIF and LIFR are expressed in blastocysts and expression of LIFR can be
detected in the ICM. However, mutant embryos deficient in LIF/gp130/STAT3 signaling
forms normal ICM. Lif deficient mice exhibit normal development (Stewart et al., 1992),
while Lifr deficient mice showed perinatal lethality (Li et al., 1995; Ware et al., 1995).
Embryos deficient in gp130 die progressively between dpc 12.5 and term. Stat3 deficient
8

embryos developed into egg cylinder stage until embryonic day 6.0 and rapidly
degenerate between E6.5-7.5. Lack of phenotype pertaining to pluripotency in
LIF/gp130/STAT3 suggests other mechanisms are normally responsible for maintenance
of pluripotency in vivo. In fact, a role for LIF/gp130 becomes only apparent for the
prolonged survival of blastocysts during diapause. Diapause is a physiological adaptation
to the presence of suckling litter that allows embryos to persist for several weeks without
implantation. Upon cessation of suckling, the embryos implant in the uterus and
development proceeds normally. gp130 mutants however lose the epiblast component
after 6 days in delay and can no longer generate a foetus upon implantation (Nichols et al.,
2001). This observation suggests LIF/gp130/STAT3 pathway is not fundamental for
pluripotency, but instead functions primarily to extend the period of pluripotency in vivo.

The LIF/gp130/Stat3 pathway also appears to be dispensable for the maintenance
of pluripotency and self- renewal of ES cells (Daheron et al., 2004; Ying et al., 2003).
Even though Lifr
-/-
mouse ES cells are less pluripotent than wildtype, it was found that
undifferentiated colony formation was not completely inhibited. This indicates that there

are LIFR-independent means by which fibroblast can support mouse ES cell self renewal.
Indeed, this may truly be the case as maintenance of several mouse ES cell lines does not
require LIF. Also, it has been reported that LIF is not necessary for human ES cell culture
and that maintenance of pluripotency in human ES cells is STAT3 independent (Daheron
et al., 2004; Thomson et al., 1998). Furthermore, STAT3 is expressed in a wide range of
cell types, and in some cases drives differentiation (Hirano et al., 2000). Therefore, other
9

core pathways and mechanisms that maintain pluripotency and ES cell self renewal may
exist.

1.2.1.2 Bone morphogenetic factor 4 and BMP signaling

In ES cell media without LIF, there is limited self renewal of mouse ES cells and
the induction of neural differentiation. It has recently been demonstrated by Ying et al
that the requirement for serum can be replaced by Bone Morphogenetic Factors (BMP) 4.
BMP4 treatment suppresses neural differentiation and in combination with LIF, is
sufficient to sustain ES cell self renewal without feeder or serum factors (Ying et al.,
2003). This concurs with the fact that BMP signaling inhibits premature neural
differentiation in the mouse embryo (Di-Gregorio A et al, 2007). The mechanistic action
of BMP signaling has been attributed to the activation of inhibitor of differentiation (Id)
genes by the downstream signal transducers, SMAD1/5/8. Forced Id1, Id2 and Id3
expression did not impair ES cell self-renewal nor block differentiation in the presence of
serum. ES cells transfected with Id1, Id2 and Id3 can however self-renew in serum-free
culture containing LIF strongly, thereby suggesting that BMP/SMAD1/5/8 acts through
Id1/2/3 proteins.

Id proteins exert a neuroectoderm lineage-specific block on ES cell
differentiation by preventing precocious expression of proneural basic helix loop helix
transcription factors as the Mash genes. It could also be that IDs exert their effect by

interaction with non bHLH proteins such as the PAXs.

10

The role played by BMP signaling in embryonic stem cells is however disparate
in mouse and human (Varga and Wrana, 2005). BMP signaling appears to play a
differentiation related role in the human context. Firstly, analysis of a hES cell line
cultured on Matrigel-coated plates in fibroblast-conditioned media showed that in the
continuous presence of FGF signaling, BMP induces hES cells to differentiate into the
trophoblast lineage (Xu et al., 2002). It has been found that blocking BMP activity in
serum with the BMP antagonist Noggin does not maintain human ES cell self-renewal,
but instead enhances neural differentiation by inhibiting non-neural differentiation (Pera
et al., 2004). Also, by culturing cells in the absence of conditioned media but in the
presence of high levels of bFGF, and with noggin to block BMP signaling, hES cells can
be maintained in a pluripotent state (Xu et al., 2005). Thus, one common thread arising
from analysis of mouse and human systems is that BMPs have maintained an
evolutionarily conserved role to block neural differentiation in early embryo and ES cells.


1.2.1.3 WNT signaling

To date, it appears that the WNT pathway is the only pathway that has been
demonstrated to be involved in maintenance of pluripotency in both mouse and human
ES cells, albeit its effect being short term (Sato et al., 2004). Expressed sequence tags
(EST) scan analysis of human ES cells and massively parallel signature sequencing
(MPSS) analysis of mouse and human cells suggest that the major components of the
WNT pathway are represented in detectable levels in undifferentiated cultures of mouse
and human ES cells. Sato et al have recently demonstrated the ability of the canonical
11

WNT signaling pathway to support the transient self renewal of both mouse and human
ES cells. Activation of the WNT pathway by 6-bromoindirubin-3’-oxime (BIO), a
specific pharmacological inhibitor of glycogen synthase kinase-3 (GSK3), maintains self
renewal and markers of pluripotency in both mES and hES cells. The actions of BIO are
functionally reversible and BIO withdrawal leads to normal differentiation. Future more,
the over-expression of Wnt1 or stabilized β-catenin or abrogation of the APC complex
leads to the inhibition of neural differentiation, mediated by the activation of downstream
target genes that includes Cyclins and c-Myc (Aubert et al., 2002).

However, an understanding of how exactly the various components of the WNT
signaling pathway affect the transcriptional activity of key pluripotency genes like Oct4
and Nanog, is still lacking. It is also not yet known which of these WNT-regulated genes
are fundamentally important for sustaining ES cell self renewal.

1.2.2 Intrinsic regulators and mechanism governing ES cell self renewal

1.2.2.1 Oct4

Oct4 (also known as Oct3 or Pou5f1) is a POU (Pit, Oct, Unc) family
transcriptional regulator of genes required in maintaining an undifferentiated pluripotent
state. Its expression is restricted to early embryos, primordial germ cells, undifferentiated
ES cells and embryonal carcinoma (EC) cells. In vivo, expression of Oct4 occurs in the
unfertilized egg and the early embryo prior to segregation of the ICM from the
12

trophectoderm (Pesce and Scholer, 2001). Subsequent to the allocation of ICM cells,
Oct4 mRNAs and proteins are readily detected in the ICM, but downregulated in the
trophectoderm. Oct4 expression is maintained in epiblast of pre- and post-implantation
embryos before becoming restricted to the migratory primordial germ cells where it
persists throughout the formation of genital ridges in both sexes (Nichols et al., 1998).

In vivo, Oct4 is necessary for pluripotency. Zygotic expression of Oct4 is required
for the development of the pluripotent ICM. Oct4 deficient mouse embryos are
embryonic lethal. Such mutant embryos can develop to a blastocyst-like stage, but lack
genuine ICM. The ICM is not pluripotent and is instead restricted to differentiation along
the trophoblast lineage. Oct4 is also necessary for pluripotency in vitro. ES cell lines
cannot be generated from Oct4
-/-
embryos (Nichols et al., 1998). Oct4 is highly expressed
in both human and mouse ES cells, and its expression is markedly reduced during
differentiation. Niwa et al found that the precise level of Oct3/4 governs three distinct
fates of mouse ES cells. A less than twofold increase in expression causes differentiation
into primitive endoderm and mesoderm. In contrast, repression of Oct-3/4 mRNA level
by more that 50% induces a loss of pluripotency and differentiation to trophectoderm
(Niwa et al., 2000). OCT4 RNAi in human ES cells also recapitulates a similar
differentiation pattern of human ES cells to trophectoderm (Matin et al., 2004).

As a transcription factor, OCT4 can act to activate or repress gene expression in
ES cells. Recent works have also confirmed that OCT4 is at the helm of the intricate
cascades of genetic events that orchestrates ES cell pluripotency. For example, Utf-1 and
13

Nanog are pluripotency-related genes that have been shown to be regulated by OCT4,
through a direct cooperative interaction with another ES cell transcription factor, SOX2
(Rodda et al., 2005). An auto-regulatory circuit of the SOX2/OCT-3/4 complex also
contributes to maintaining robustly, the precise expression level of Oct3/4 itself in ES
cells (Chew et al., 2005) .

The maintenance of Oct4 expression, although necessary, is however not
sufficient in itself to sustain the pluripotent phenotype. In mouse ES cells, Oct4 over-

expression is insufficient to block differentiation in absence of LIF. OCT4 must act in
combination with factors whose activity is influenced by the STAT3 pathway activated
by Leukemia Inhibitory Factor (LIF) to prevent ES cell differentiation.

In summary, Oct4 is necessary for pluripotency both in vitro and in vivo. The
level of Oct4 controls the first cell fate decision (trophoblast vs epiblast lineage)
undertaken by the mammalian embryo. As for ES cells, the function of Oct3/4 is to
specify cell fate into pluripotent cells or primitive endoderm cells by blocking
trophectodermal differentiation.

1.2.2.2 Sox2

Sox2 is a member of the HMG-domain DNA binding protein family. It has been
demonstrated that Sox2 has a role to play in maintaining epiblast pluripotency, as gene
targeting used to inactivate Sox2 in the mouse causes defective primitive ectoderm
14

(Avilion et al., 2003). Recent findings reported by Kuroda et al and Rodda et al also shed
light on the fact that Sox2 is critical for the ES cell state. SOX2 is known to play key
roles in transcription of several OCT4 targets such as Fgf4 (Ambrosetti et al., 1997). Also,
Nanog is regulated by OCT4, through a direct cooperative interaction with SOX2.

Recently, it has been found that unlike OCT4, SOX2 appears to be localized in
both the nuclei

and cytoplasm in pre-implantation embryos and ES cells. This subcellular
localization pattern suggests that SOX2 shuttles

between these two subcellular
compartments and that its nuclear

localization regulates its transcription activity. SOX2
contains two nuclear localization signals

(NLS). Li et al found that ablation of these two
NLS results in a dominant-negative

form of SOX2 that loses its ability to cooperate with
OCT4. When stably expressed, this dominant-negative form of mouse SOX2 induces
progressive polyploidy and trophectodermal differentiation in ES cells. Knockdown of
Sox2 by small interfering RNA (siRNA) similarly induces trophectoderm differentiation
and polyploid formation in mouse ES cells. The notion that SOX2 operates primarily in
pluripotent cells as a co-factor of OCT4 is reinforced by the fact that the mutant Sox2
phenotype mirrors that of Oct4 deficient mutants. In essence, SOX2 maintains ES cell
self renewal by

shuttling between the nucleus and cytoplasm in cooperation with

OCT4 to
prevent trophectoderm differentiation and polyploid

formation in ES cells (Li et al., 2007).





15

1.2.2.3 Nanog

Nanog is a homeodomain transcription factor that has been shown to confer upon
mouse ES cells, the ability to self renew without STAT3 activation (Chambers et al.,
2003; Mitsui et al., 2003). Identification of Nanog was first reported by both Mitsui et al
and Chamber et al independently in 2003. Mitsui et al employed digital differential
display to compare expressed sequence tag libraries from mouse ES cells and those from
various somatic tissues. Nanog was among the handful of genes that were found to be
highly enriched in ES cells. Interestingly, Mitsui et al found Nanog to specifically confer
upon mouse ES cells, the ability to self renew in the absence of LIF for a short period of
time when overexpressed. As for Chamber et al, they first isolated Nanog through
expression cloning. They similarly found that Nanog can bypass completely the
requirement for STAT3 activation and that maintained expression of Oct4 is integral to
its ability to sustain ES cell self-renewal. Chambers et al further showed that
overexpression of Nanog can even cause mouse ES cells to be more resistant to induction
of differentiation with compounds such as 3-methoxybenzamide or all trans-retinoic acid.
Since then, human NANOG has also been characterized and implicated to play a role in
human ES cells. Overexpression of NANOG in human ES cells has been shown to enable
feeder-free growth while inducing primitive ectoderm features (primitive ectoderm is the
pluripotent population in the embryo derived from the ICM) (Darr et al., 2006).

In vivo, Nanog expression is detected in pluripotent embryonic tissues but not in
adult tissues. During mouse embryonic development, Nanog is first detected in the
16

compacted morula. It then becomes localized to the ICM of the blastocyst, and is
subsequently restricted to the epiblast. Just prior to implantation, Nanog is down-
regulated. Chambers et al also reported that epiblast expression of Nanog persists in
implantation delayed blastocysts, a favoured source for ES cell derivation. Later during
embryo development, Nanog mRNA is present in primordial germ cells in E11.5 gonadal
ridges, the very cells from which pluripotent EG stem cells can be generated. The

knockout phenotype of Nanog
-/-
embryo is a lack of epiblast and primitive ectoderm
formations. Nanog deficient embryos only produce parietal endoderm like cells instead.
This is indicative of the absolute requirement for Nanog in directing both ICM and
primitive ectoderm development. Also, in contrast to Oct3/4 null ICM, no trophoblast
differentiation was observed. This data indicates that Nanog is essential for maintenance
of pluripotency at a stage after the first differentiative event of the ICM which requires
Oct3/4.

In vitro, Nanog expression is likewise confined to pluripotent cell lines such as ES
cells, EG cells and EC cells. It is now known that Nanog is regulated by OCT4, through a
direct cooperative interaction with SOX2 in ES cells. Nanog null

ES cells lose
pluripotency and differentiate into the extraembryonic endoderm lineage. A proposed
mechanism is that Nanog prevents differentiation of ES cells into primitive endoderm
through the transcriptional repression of differentiation promoting genes such as Gata6.
Gata6 overexpression was found to be sufficient in inducing extraembryonic endoderm
differentiation. NANOG also maintain ES cell self renewal by regulating ES cell specific
genes. For example, the Nanog consensus sequence was found in the enhancer region of
17

Rex-1, an ES cell related gene. The transcriptional regulation of Rex-1 by NANOG was
also subsequently demonstrated (Shi et al., 2006).

1.2.3 Current dogma of ES cell self renewal

Figure 1 summarises the current paradigm used to explain how NANOG, OCT3/4,
SOX2, and STAT3 cooperate to maintain pluripotency. Preimplantation embryos and ES

cells require both OCT3/4 and NANOG to prevent differentiation into trophectoderm and
primitive endoderm, respectively. However, these two transcription factors of normal
expression level are not sufficient for ES cell self-renewal. In the mouse system at least,
additional factor(s), such as STAT3 activated by LIF, are required to support prolonged
maintenance of pluripotency.

The current dogma is that transcription factors, NANOG, OCT4 and SOX2 are
key regulators of ES cell fate. These factors are at the top of the hierarchy of the ES cell
regulatory network and they keep ES cells undifferentiated by either the activation of
target genes that encodes pluripotency and self renewal mechanism, or the repression of
target genes that promote differentiation. The regulatory regions of hundreds of genes
that are expressed or repressed in the undifferentiated ES cell state are found to be co-
occupied by OCT4, NANOG and SOX2 (Boyer et al., 2005; Loh et al., 2006).
Maintenance of the ES cell state is hence dependent on the precise level and
transcriptional control of Oct4, Sox2 and Nanog, as perturbations will kick off drastic
changes in the downstream cascade, leading to differentiation.
18

Two landmark papers published recently have elucidated the binding landscape of
OCT4 and NANOG in mouse ES cells and OCT4, NANOG and SOX2 in human ES cells
(Boyer et al., 2005; Loh et al., 2006). These location maps are useful guides that can aid
in the identification of additional components in the ES cell self renewal regulatory
network. Also, if and what roles most putative OCT4, SOX2 and NANOG targets play in
ES cells are still very much a mystery at present. In this thesis, the functional dissection
of the role(s) of target genes such as Lefty1 and Lefty2 in ES cells is described.

Figure 1. Model proposed by Smith et al (2003) to describe the current dogma governing
ES cell self renewal. Nanog and Oct4 are both essential to sustain ES cell identity,
whereas Stat3 has an accessory function. Oct4 serves to block differentiation into
trophoblast but tends to promote differentiation into primitive endoderm and germ layers.

Nanog (and activated Stat3) may block this differentiation effect of Oct4.




Stat3
LIF
Nanog

Oct4
19

1.3 Lefty - a Nodal antagonist

1.3.1 Roles of Lefty in embryo development


Lefty proteins are members of the TGF-ß superfamily. The TGF-ß family has
more than 40 members, including Activin, Nodal and BMPs. TGF-ßs have been
implicated in regulating cell growth, differentiation, apoptosis and cellular patterning
amongst many other functions in embryonic development. The Lefty proteins are atypical
TGF-ßs. Typically, TGF-ß family members act as dimers and bring together

two
transmembrane serine/threonine kinases, the type I and II

receptors. The assembly and
oligomerization of these receptors

lead to phosphorylation of receptor-regulated SMADs

(R-SMADs)

and release them from their docking site on the receptor and

allow them to
heterodimerize with a common SMAD, SMAD4. These

complexes accumulate in the
nucleus, where they interact with

other transcription factors, bind to DNA, and activate
transcription

of TGF-ß-responsive genes. Also, typical TGF-ß family members have a
characteristic signature motif that is comprised of a series of seven cysteine residues at
their C-termini. All of these residues are used for the formation of intrapeptide bonds
with the exception of the third cysteine residue from the C terminus of mature proteins.
Active dimerized TGF-ß ligands are linked by the formation of disulfide bonds between
the fourth cysteine residues at the C terminus of the proteins. LEFTY and its Xenopus
homolog, Antivin however lack the cysteine residue necessary for the formation of
intermolecular disulfide bond Therefore, LEFTY appears to belong to a subgroup of the
TGF-ß superfamily with an unpaired cysteine residue that does not exist as a dimer. For
20

this reason, LEFTY2 was postulated to have an inhibitory role instead (Tabibzadeh and
Hemmati-Brivanlou, 2006).

Two homologues, Lefty1 and Lefty2 have been identified in mouse (Meno et al.,
1997). In mouse, the Lefty locus at chromosome 1(1q42.1) contains both genes with the
same transcriptional orientation. The mature regions of mouse Lefty1 and Lefty2 proteins

are 97% homologous to each other, excluding the four carboxyl terminal amino acid
residues which are variable. In comparison, the proprotein region is less conserved
between the two Lefty proteins, with 70% homolgy. In human, the Lefty locus contains a
third gene, Lefty3. This however appears to be a pseudogene because of its reverse
transcription orientation, presence of a single exon, interruption by Alu repeats, and
absence of human expressed sequence tag clones. In human, the Lefty genes (LEFTYA
and LEFTYB) have also been localized to chromosome one. The LEFTY B gene (human
Lefty 1) has 96% sequence identity and shares 46 amino acids with LEFTY A (human
Lefty 2). Based on structural features, it has been suggested that the Lefty proteins
evolved independently from the duplication of a single Lefty gene in both mouse and
human (Tabibzadeh and Hemmati-Brivanlou, 2006).

These non-prototypic TGF-ßs were named “Lefty” because mouse Lefty proteins
were first discovered to be expressed exclusively on the left side of developing embryos
and the loss of these Lefty protein led to disruption of left right axis formation in
developing embryos (Hamada et al., 2002; Meno et al., 1996). Conditional Lefty2
knockouts such as mutants with the Lefty2 asymmetric enhancer deleted could be carried
21

to term, but showed left isomerism and displayed severe defects such as abnormal heart
looping. Left isomerism was similarly observed in Lefty1 null mice (Hamada et al., 2002).
Other Lefty functions, aside from left-right asymmetry specification have since been
elucidated by knockout studies.

Dynamic spatial and temporal patterns of expression of Lefty2 proteins have been
observed during early embryonic development in mouse. A gene profiling study carried
out by Hamatani et al has revealed that Lefty 2 is co-expressed with transcription factors
such as Oct4 and Sall4 during mouse preimplantation development (Hamatani et al.,
2004). At gastrulation, Lefty2 is expressed in the nascent mesoderm arising in the mid-
distal region of the primitive streak. At early somite stage, Lefty2 is expressed

predominantly in left lateral plate mesoderm (LPM). Such an expression profile implies
that Lefty2 proteins may have a regulatory role during early embryo development (Saijoh
et al., 1999).

Indeed, Lefty2 has been found to play important roles during gastrulation and
mesoderm formation (Meno et al., 1999). Meno et al generated Lefty2 null mouse
embryos and that found that complete loss of Lefty2 activity in these embryos led to early
lethality even before left-right defects can be analyzed. In Meno et al’s paper, it was
stated that E8.25 homozygous mutant embryos were present at roughly the expected
frequency of 25%, but were found to be morphologically abnormal. Meno et al found that
their Lefty2
-/-
embryos failed to form organized structures such as the head folds, somites,
heart, node and notochord. In addition, they observed that the Lefty2
-/-
mouse mutants had
22

a dramatically expanded primitive streak and formed excess mesoderm. Put together,
their observations imply that Lefty2 is required for the correct specification of the
mesoderm generated along the proximal-distal length of the streak during gastrulation.

It is interesting to note that the phenotypes of Lefty2
-/-
embryos are the opposites
of Nodal null mutants (Meno et al., 1999). Nodal
-/-
mutants failed to form primitive
streaks and were observed to be largely devoid of mesoderm. Similarly, overexpression
of Antivin or mouse Lefty2 in zebrafish embryos blocked head and trunk mesoderm

formation, a phenotype identical to that of mutants caused by loss of Nodal signaling.
Also Nodal heterozygosity was found to partially rescue the Lefty2
-/-
mutant phenotype.
Overexpression of Nodal orthologs like cyclops or squint or the extracellular domain of
the ActRIIB receptor similarly salvaged Lefty2 depletion. These lines of evidence suggest
that LEFTY2 serves as an extracellular antagonist to modulate the range and duration of
Nodal signaling during embryonic patterning. Co-expression of Lefty2 and Nodal during
embryo development further supports this hypothesis.
With Lefty1, a similar dynamic expression pattern exists during early embryo
development (Hamatani et al., 2004). Gene profiling experiment performed by Hamatani
et al has revealed that Lefty1 is co-expressed with transcription factors such as Nanog and
Sox2 during mouse preimplantation development. Takaoka also detected asymmetric
expression of Lefty1 in primitive endoderm of implanting embryos. At gastrulation,
transient expression of Lefty1 has been detected in the left half of the gastrula. At the
early somite stage, Lefty1 is expressed predominantly in the prospective floor plate (PFP)
(Saijoh et al., 1999). Like Lefty2, Lefty1 plays an important role in modulating Nodal
23

signaling during early embryo development. Apart from left right asymmetry
determination, Perea-Gomez et al found that Lefty1 shares a redundant function with
Cerberus-like (Cerl) in regulating Nodal signaling during gastrulation. This was inferred
by the fact that a compound knockout of Lefty1 with Cerberus (Cerl
-/-
;Lefty1
-/-
mutants)
led to expanded anterior primitive streaks. A subset of these compound mutants also
exhibited an ectopic primitive streak. Interestingly, rescue of primitive streak defects in
Cerl

−/−
;Lefty1
−/−
mutant embryos lacking one copy of Nodal could be observed. Similar
to Lefty2
-/-
mutants, the defects found in Cerl
−/−
;Lefty1
−/−
mutant embryos are opposite to
those observed in Nodal mutants (Perea-Gomez et al., 2002). These lines of evidence
imply that Lefty1, like Lefty2 regulates Nodal signaling during early development.
Chen et al has previously shown two modes by which the Lefty proteins can
inhibit Nodal signaling using in vitro assays. They showed that Lefty proteins can interact
with Nodal proteins in solution and block Nodal from binding its receptors. Furthermore,
they also found that Lefty can also interact with EGF-CFC co-receptors and prevent their
ability to form part of a Nodal receptor complex (Chen and Shen, 2004).

1.3.2 Implications of a role for Lefty in ES cell self renewal

Recently, reports by several groups also implicate Lefty in the maintenance of
“stemness”. Mouse and human ES cells have large differences and share only a very
small number of similarities at the transcriptome level (Bhattacharya et al., 2004; Rao,
2004). In fact, human ES cells may represent a slightly later developmental stage as
compared to mouse ES cells. Recent work by Brons et al.(2007) and Tesar et al.(2007)
24

show that human ES cells share unique properties such as flattened colony morphology
and inefficient clonal propagation etc with stem cells derived from post implantation

mouse epiblasts. Human ES cells and mouse epiblast derived stem cells also share
patterns of gene expression and signaling requirements such as a dependence on
Nodal/Activin signaling for the maintenance of pluripotency that are characteristic of the
mouse epiblast. Global transcriptional profiling studies have however shown that Lefty1
and Lefty2 are among a core set of genes that is highly expressed in both human and
murine embryonic stem cells and blastocysts (Brandenberger et al., 2004; Wei et al.,
2005). Hence, it is possible that Lefty1 and Lefty2 may represent components of the
critical core pathway utilized for ES cell self renewal. Also, LEFTY1 and LEFTY2 are
the only secreted factors of the TGF- superfamily, apart from GDF3 expressed at high
level during pluripotency and is almost absent in terminally differentiated cell types
(Levine and Brivanlou, 2006). In fact, Lefty molecules are the most abundant inhibitors
both in mouse and human ES cells. Other inhibitors of the TGF family like TMEFF and
Follistatin-like, although present, are not enriched in ES cells. Study carried out by
Hamatani et al has also revealed that Lefty 2 is co-expressed with key ES cell
transcription factors such as Oct4 and Sall4 during mouse preimplantation development.
This suggests the likely regulation of Lefty2 by these transcription factors which have
notable effect in the regulation of ES cell fates. Genome scale location analysis
(chromatin immunoprecipitation coupled with DNA microarrays) carried out by Boyer et
al has also revealed that Lefty2 is a downstream target of the key ES cell transcriptional
factors, OCT4, SOX2 and NANOG in human ES cells. Ng et al has found that Lefty1 is a
gene bound by both NANOG and OCT4 in mouse ES cells. It has also found that