the stem cell hope : how stem cell medicine can change our lives
Post on 11-Sep-2021
1 Views
Preview:
TRANSCRIPT
Cell Stem Cell
Previews
Oxidative-Reductionist Approachesto Stem and Progenitor Cell Function
Mark Noble,1,* Chris Proschel,1 and Margot Mayer-Proschel11Department of Biomedical Genetics, University of Rochester Stem Cell and Regenerative Medicine Institute, University of RochesterMedical Center, 601 Elmwood Avenue, Rochester, NY 14642, USA*Correspondence: mark_noble@urmc.rochester.eduDOI 10.1016/j.stem.2010.12.005
Redox status is a critical modulator of stem and progenitor cell function. In this issue of Cell Stem Cell,Le Belle et al. (2011) demonstrate that oxidation promotes self-renewal of neuroepithelial stem cells,revealing fascinating differences—and surprising similarities—with how redox pathways regulate glialprogenitor cells.
The status of being oxidized or reduced is
one of the most fundamental regulators of
cell function. It has become increasingly
clear that small changes in redox status
are critical in regulating the function of
multiple signaling pathways and tran-
scription factors, that such regulation
is central to normal cell function and not
just in conditions of oxidative stress, and
that both signaling molecules and tran-
scriptional regulators exert many of their
effects through modulation of redox
status. Thus, despite the existing focus
on the regulation of stem/progenitor cell
function by specific signaling and tran-
scriptional events, it could be argued
that the regulation of these cells at the
level of redox modulation may be of
equal—if not greater—importance.
A welcome new addition to the litera-
ture on redox regulation of precursor cell
function is the current article by Kornblum
and colleagues (Le Belle et al., 2011) that
demonstrates the importance of reactive
oxygen species (ROS) in regulating self-
renewal and neurogenesis in central
nervous system (CNS) stem and progen-
itor cells. Their results provide highly
convincing evidence that increases in
oxidative status enhance neurosphere
generation by neuroepithelial stem cells
(NSCs) of the CNS. Specifically, exoge-
nous agents that elevate ROS levels
increased production of neurospheres,
one of the key in vitro assays for stem
cell activity of NSCs. Freshly isolated cells
from the subventricular zone (SVZ; the
predominant location of stem cells in the
CNS) that express stem cell antigens
exhibit high levels of ROS, while stem
cell antigen-negative cells harbor less
ROS. One key contributor to these
increased ROS levels is NADPH oxidase
(NOX), and pharmacological inhibition of
NOX inhibits neurosphere formation.
Moreover, cells isolated from the SVZ of
NOX2�/� mice showed lower ROS levels
and diminished capacity for NSC self-
renewal and retention of multipotency
during passaging in vitro. Brain-derived
neurotrophic factor (BDNF), which can
further enhance neurosphere generation
in cultures exposed to adequate levels
of EGF and FGF, increased ROS levels
in these cells. Furthermore, NOX inhibition
or treatment with the antioxidant and
glutathione pro-drug N-acetyl-L-cysteine
(NAC) inhibited the effects of BDNF on
NSCs. BDNF was also not able to stimu-
late self-renewal in cells isolated from
NOX2�/� mice.
One of the most striking aspects of the
findings of Le Belle et al. (2010) is that they
represent, in many respects, a reverse
image of previous studies that examined
redox regulation of oligodendrocyte/
type-2 astrocyte progenitor cells (also
known as oligodendrocyte precursor
cells, and here abbreviated as O-2A/
OPCs). In O-2A/OPCs, it is the more
reduced cells that exhibit enhanced self-
renewal properties, while cells that are
relatively oxidized have a higher proba-
bility of differentiating into nondividing
oligodendrocytes (Power et al., 2002;
Smith et al., 2000). Moreover, increasing
glutathione with NAC in O-2A/OPCs
promotes self-renewal, whereas expo-
sure to chemical pro-oxidants inhibits
cell division.
Remarkably, despite the opposite
effects of redox changes on NSC and
O-2A/OPC proliferation and differentia-
tion, there are multiple similarities that
Cell Stem Ce
reveal certain common principles at
work. For example, in both cases, the
correlation between redox status in vitro
and in vivo is strongly conserved, such
that NSCs freshly isolated from regions
where they normally undergo more self-
renewal are more oxidized (Le Belle
et al., 2011) and O-2A/OPCs isolated
from developing regions of CNS in which
self-renewal occurs for extended periods
are more reduced (Power et al., 2002;
Smith et al., 2000). In addition, cells puri-
fied from the animal on the basis of their
redox status exhibit the predicted differ-
ences in self-renewal for both NSCs and
O-2A/OPCs. Moreover, in both cases,
cells more prone to self-renewal exhibit
some ability to maintain their redox set
point when grown in conditions that
would otherwise alter their redox state.
In other words, NSCs remained relatively
oxidized when grown in 4% (physio-
logical) O2 levels, and the more reduced
O-2A/OPCs remained reduced when
grown in 21% (atmospheric) O2. The
presence of homeostatic regulation of
redox set points suggests strongly that
regulation of a particular redox balance
is of critical importance in the function of
stem/progenitor cells in the CNS.
Common principles also are apparent
when considering the essential nature of
redox regulation as a mediator of the
effects of signaling molecules relevant to
NSC and O-2A/OPC function. In both cell
types, cell-signaling ligands that alter the
balance between self-renewal and differ-
entiation alter redox state in precisely the
direction predicted by the effects on self-
renewal probability of chemical redox
modulators. In NSCs, BDNF promotes
self-renewal and exposure to this cytokine
ll 8, January 7, 2011 ª2011 Elsevier Inc. 1
Cell Stem Cell
Previews
makes these cells more oxidized. In O-2A/
OPCs, fibroblast growth factor-2 and neu-
rotrophin-3 enhance self-renewal and
make cells more reduced, while thyroid
hormone and bone morphogenetic
protein-4 promote differentiation and
make cells more oxidized. Critically, in
every case, inhibiting the redox changes
caused by the signaling molecules abro-
gates their effects on self-renewal and
differentiation. Such findings make it clear
that analysis of cell signaling function
purely in terms of phosphorylation
cascades, transcriptional regulation, etc.,
provides only a partial understanding of
the means by which signaling regulates
precursor cell function. In addition, it is
clear for both O-2A/OPCs and NSCs the
effects of redox modulation are quite
specific (Li et al., 2007), lending support
to the idea that rather than acting as
a mere cofactor in general cell-biological
processes, redox state can act as a
specific regulator of stem/progenitor cell
function.
The current findings on NSCs are not
the only example in which being more
oxidized enhances self-renewal and/or
division. In the CNS, hippocampal cells
that give rise to neurons are stimulated
to divide by oxidation (Limoli et al.,
2004), as are a variety of other non-CNS
cells (Sauer et al., 2001). But when
considering stem cells, it is important to
consider the biological function of rapidly
dividing cells. Outside of the earliest
stages of development, stem cells are
thought to exist mainly in a slowly
dividing, ‘‘quiescent’’ state, and studies
of hematopoietic stem cells (HSCs)
suggest that oxidation is associated with
the transition from quiescence to a rapidly
dividing stage. This proliferative pool
retains the capacity for multilineage
reconstitution but loses the ability for
long-term, serial repopulation of the
bone marrow (Kim et al., 1998), which is
considered a gold standard functional
2 Cell Stem Cell 8, January 7, 2011 ª2011 El
assay for self-renewal. It is intriguing to
speculate whether the generation of
rapidly dividing cells is a universal stem
cell response to injury and whether the
increased ROS production seen in most
or all injuries might be a universal signal
to stem cells to exit quiescence. But it
is clear that even cells that find oxidation
beneficial generate cells that have a re-
dox response more like O-2A/OPCS, as
evidenced by the death of neurons in the
same oxidative conditions that promoted
their generation from NSCs (Le Belle
et al., 2011).
How are alterations in redox status
translated into changes in self-renewal
and differentiation? In O-2A/OPCs, small
increases in oxidative status cause acti-
vation of Fyn kinase, leading to activation
of the ubiquitin ligase c-Cbl and acceler-
ated degradation of its target proteins,
including several critical receptor tyrosine
kinases (RTKs) (Li et al., 2007). Loss of
RTKs leads to suppression of down-
stream signaling through ERKs and Akt.
In contrast, in NSCs, oxidative suppres-
sion of PTEN activity leads to elevated
Akt activity, and the Akt pathway appears
to be essential for NSC self-renewal
(Le Belle et al., 2011). But connections to
other components of the cell-cycle
machinery still need to be made. It is
also particularly intriguing that many of
the signaling players identified thus far
(e.g., PTEN, Fyn, c-Cbl) are present in
virtually all cell types, which raises the
question of what regulatory network
enables distinct outcomes in different
cell types.
Redox regulation of stem/progenitor
cell function should also be considered
carefully by the developing field of tissue
repair by stem/progenitor cells. It is
already clear that differences in redox
status can be used to isolate cells of
differing self-renewal potential (Le Belle
et al., 2011; Smith et al., 2000) and there
are growing numbers of examples in
sevier Inc.
which oxygen concentrations modulates
stem/progenitor cell function (Mazumdar
et al., 2009; Mohyedin et al., 2010).
But will the redox status of the host
also determine the ability of endogenous
or transplanted stem/progenitor cells to
carry out repair? Given that, in some
populations, even a 15% increase in
glutathione content causes a >1000%
increase in cell survival (Mayer and Noble,
1994), relatively small metabolic fluctua-
tions may greatly change the outcome of
experiments and clinical trials. Consid-
ering that the redox state is altered in
almost every type of tissue injury, efforts
to understand how the repair response
of specific cell types may be altered by
particular redox states may prove essen-
tial to achieving an optimal clinical benefit.
REFERENCES
Kim, M., Cooper, D., Hayes, S., and Spangrude, G.(1998). Blood 91, 4106–4117.
Le Belle, J.E., Orozco, N.M., Paucar, A.A., Saxe,J.P., Mottahedeh, J., Pyle, A.D., Wu, H., and Korn-blum, H.I. (2011). Cell Stem Cell 8, this issue, 59–71.
Li, Z., Dong, T., Proschel, C., and Noble, M. (2007).PLoS Biol. 5, e35. 10.1371/journal.pbio.0050035.
Limoli, C.L., Rola, R., Giedzinksi, E., Mantha, S.,Huang, T.-T., and Fike, J.R. (2004). Proc. Natl.Acad. Sci. USA 101, 16052–16057.
Mayer, M., and Noble, M. (1994). Proc. Natl. Acad.Sci. USA 91, 7496–7500.
Mazumdar, J., Dondeti, V., and Simon, M.C.(2009). J. Cell. Mol. Med. 13, 4319–4328.
Mohyedin, A., Garzon-Muvdi, T., and Quinones-Hinojosa, A. (2010). Cell Stem Cell 6, 150–161.
Power, J., Mayer-Proschel, M., Smith, J., andNoble, M. (2002). Dev. Biol. 245, 362–375.
Sauer, H., Wartenberg, M., and Hescheler, J.(2001). Cell. Physiol. Biochem. 11, 173–186.
Smith, J., Ladi, E., Mayer-Proschel, M., and Noble,M. (2000). Proc. Natl. Acad. Sci. USA 97, 10032–10037.
Cell Stem Cell
Previews
Aging by Telomere Loss Can Be Reversed
Bruno Bernardes de Jesus1 and Maria A. Blasco1,*1Telomeres and Telomerase Group, Molecular Oncology Program, Spanish National Cancer Centre (CNIO), Melchor Fernandez Almagro 3,Madrid E-28029, Spain*Correspondence: mblasco@cnio.esDOI 10.1016/j.stem.2010.12.013
Recently in Nature, Jaskelioff et al. (2010) demonstrated that multiple aging phenotypes in a mouse model ofaccelerated telomere loss can be reversed within 4 weeks of reactivating telomerase. This raises the majorquestion of whether physiological aging, likely caused by a combination of molecular defects, may also bereversible.
Accumulation of short/damaged telo-
meres with increasing age is considered
one of the main sources of aging-associ-
ated DNA damage responsible for the
loss of regenerative potential in tissues
and during systemic organismal aging
(Harley et al., 1990; Flores et al., 2005).
Mounting evidence suggests that telome-
rase is a longevity gene that functions
by counteracting telomere attrition. Thus,
telomerase-deficient mice age prema-
turely, and telomerase overexpression
results in extended longevity in mice
(Tomas-Loba et al., 2008). Moreover,
human mutations in telomerase compo-
nents produce premature adult stem cell
dysfunction and decreased longevity
(Mitchell et al., 1999).
Previous work had shown that restora-
tion of telomerase activity in mouse
zygotes with critically short telomeres,
owing to a deficiency in the
telomerase RNA component
(Terc), rescues critically short
telomeres and chromosomal
instability in the resulting
mice (Samper et al., 2001).
Restoration of telomerase
activity in zygotes also pre-
vented the wide range of
degenerative pathologies
that would otherwise appear
in telomerase-deficient mice
with critically short telomeres,
including bone marrow apla-
sia, intestinal atrophy, male
germ line depletion, and
adult stem cell dysfunction
(Samper et al., 2001; Siegl-
Cachedenier et al., 2007),
and resulted in a normal
organismal life-span (Siegl-
Cachedenier et al., 2007).
Together, all the above find-
ings indicate that aging provoked by crit-
ical telomere shortening can be prevented
or delayed by telomerase reactivation.
From these grounds, reversion of aging
caused by telomere loss was the next
frontier. A recent study in Nature takes
an important step forward from these
previous findings by using a new mouse
model for telomerase deficiency, de-
signed to permit telomerase reactivation
in adultmice after telomere-induced aging
phenotypes have been established (Jas-
kelioff et al., 2010). Specifically, DePinho
and colleagues generated a knockin allele
encoding a 4-OH tamoxifen (4-OHT)-
inducible mouse telomerase (TERT-ER)
under the control of the TERT endogenous
promoter. In the absence of tamoxifen,
these mice exhibit premature appearance
of aging pathologies and reduction in
survival (Figure 1). Thesemice phenocopy
previously described Terc-deficient mice,
which highlights that elongation of short
telomeres by telomerase is the main
mechanism by which telomerase protects
from aging pathologies. Importantly,
4 weeks of tamoxifen treatment to induce
TERT re-expression in adult TERT-ER
mice with clear signs of premature
aging was sufficient to extend their
telomeres and rescue telomeric DNA
damage signaling and associated check-
point responses. Dramatically, tamox-
ifen-induced TERT re-expression also
led to resumption of proliferation in quies-
cent cultured cells and eliminated the
degenerative phenotypes across multiple
organs, including testis, spleen, and intes-
tines (Figure 1). Reactivation of telome-
rase also ameliorated the decreased
survival of TERT-ER mice. These findings
represent an important advance in the
aging field, as they show that
aging induced by telomere
loss can be reversed in
a broad range of tissues and
cell types, including neuronal
function.
Looking to the future, the
next key question is to what
extent natural, physiological
aging is caused by the pres-
ence of critically short telo-
meres and, consequently, to
what extent telomere restora-
tion will be able to reverse
physiological aging. In this re-
gard, other recent findings
support the idea that telomere
shortening does impact
natural mouse aging. On one
hand, despite the long-
standing belief that mouse
aging was not linked to telo-
mere shortening given that
Figure 1. Antiaging Effects of TelomeraseSchematic showing the major findings of Jaskelioff et al. (2010). Telomerasereactivation in late generation telomerase-deficient mice (G4TERT-ER) couldrevert some of the aging phenotypes observed, demonstrating the regenera-tive potential capacity of different tissues.
Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc. 3
Cell Stem Cell
Previews
mice are born with very long telomeres—
much longer than human telomeres—
mouse telomeres do suffer extensive
shortening associated with aging (Flores
et al., 2008). Inparticular,whilemousecells
maintain relatively long telomeres during
their first year of life, there is a dramatic
loss of telomeric sequences at 2 years of
age, even in various stem cell populations,
and this change is concomitant with the
loss of regenerative capacity associated
with mouse aging. In addition, telome-
rase-deficient mice from the first genera-
tion (G1Terc�/�) exhibit a significant
decrease in median and maximum
longevity and a higher incidence of age-
related pathologies and stem cell dysfunc-
tion compared with wild-type mice (Flores
et al., 2005; Garcia-Cao et al., 2006), indi-
cating that, as in humans, telomerase
activity is rate limiting for natural mouse
longevity and aging. These results suggest
that strategies aimed to increase telome-
rase activity may delay natural mouse
aging. Further supporting this notion, it
was recently shown that overexpression
of TERT in the context of mice engineered
to be cancer resistant owe to increase
HGPS-Derived iPS
Tom Misteli1,*1National Cancer Institute, NIH, Bethesda, MD*Correspondence: mistelit@mail.nih.govDOI 10.1016/j.stem.2010.12.014
In this issue of Cell Stem Cell, Zhangture aging diseases, Hutchinson-Gilto study HGPS, and their use may l
Some problems in biology are more
difficult to study than others. Human
aging is certainly one of them. Most
conclusions regarding molecular mecha-
nism of human aging rely onmere correla-
tion, and direct experimental testing
is generally not feasible. One approach
to dissect the molecular basis of human
aging is to study naturally occurring
premature aging disorders. One of the
most dramatic and prominent of such
4 Cell Stem Cell 8, January 7, 2011 ª2011 E
expression of tumor suppressor genes
(Sp53/Sp16/SARF/TgTERT mice) was
sufficient to decrease telomere damage
with age, delay aging, and increasemedian
longevity by 40% (Tomas-Loba et al.,
2008). However, it remains to be seen
whether telomerase reactivation late in life
would be sufficient to delay natural mouse
aging and extend mouse longevity without
increasing cancer incidence.
In summary, these proof-of-principle
studies using genetically modified mice
are likely to encourage the development
of targeted therapeutic strategies based
on reactivation of telomerase function.
Indeed, small molecule telomerase acti-
vators have been reported recently and
have demonstrated some preliminary
health-span beneficial effects in humans
(Harley et al., 2010). Identifying drugable
targets and candidate activators clearly
opens a new window for the treatment of
age-associated degenerative diseases.
REFERENCES
Flores, I., Cayuela, M.L., and Blasco, M.A. (2005).Science 309, 1253–1256.
Cs For The Ages
20892, USA
et al. (2011) generate patient-derivedford Progeria Syndrome (HGPS). Thesead to novel insights into mechanism
diseases is Hutchinson-Gilford Progeria
Syndrome (HGPS). Zhang et al. (2011)
now report the generation of induced
pluripotent stem cells (iPSCs) from
HGPS cells, providing a powerful new
tool to unravel the molecular and physio-
logical mechanisms of premature and
normal aging.
HGPS is a truly remarkable disease in
many ways. To start with, it affects an
unusually wide spectrum of tissues and
lsevier Inc.
Flores, I., Canela, A., Vera, E., Tejera, A., Cotsare-lis, G., and Blasco, M.A. (2008). Genes Dev. 22,654–667.
Garcia-Cao, I., Garcia-Cao, M., Tomas-Loba, A.,Martin-Caballero, J., Flores, J.M., Klatt, P., Blasco,M.A., and Serrano, M. (2006). EMBO Rep. 7,546–552.
Harley, C.B., Futcher, A.B., and Greider, C.W.(1990). Nature 345, 458–460.
Harley, C.B., Liu, W., Blasco, M., Vera, E.,Andrews, W.H., Briggs, L.A., and Raffaele, J.M.(2010). Rejuvenation Res. 14, in press. Publishedonline September 7, 2010. 10.1089/rej.2010.1085.
Jaskelioff, M., Muller, F.L., Paik, J.H., Thomas, E.,Jiang, S., Adams, A.C., Sahin, E., Kost-Alimova,M., Protopopov, A., Cadinanos, J., et al. (2010).Nature. 10.1038/nature09603.
Mitchell, J.R., Wood, E., and Collins, K. (1999).Nature 402, 551–555.
Samper, E., Flores, J.M., and Blasco, M.A. (2001).EMBO Rep. 2, 800–807.
Siegl-Cachedenier, I., Flores, I., Klatt, P., andBlasco, M.A. (2007). J. Cell Biol. 179, 277–290.
Tomas-Loba, A., Flores, I., Fernandez-Marcos,P.J., Cayuela, M.L., Maraver, A., Tejera, A., Borras,C., Matheu, A., Klatt, P., Flores, J.M., et al. (2008).Cell 135, 609–622.
iPSCs for one of the major prema-e cells are amuch-needed new tools of aging.
leads to the development of highly diverse
symptoms ranging from depletion of
subcutaneous fat to loss of hair and
tendon contractures. The diversity of
affected tissues pointed early on to stem
cell defects as a likely disease mecha-
nism. Most relevant in patients are
vascular defects and recurring strokes,
which invariably are fatal in patients in
their mid- to late teens (Hennekam,
2006). The disease is exceedingly rare
mice are born with very long telomeres—
much longer than human telomeres—
mouse telomeres do suffer extensive
shortening associated with aging (Flores
et al., 2008). Inparticular,whilemousecells
maintain relatively long telomeres during
their first year of life, there is a dramatic
loss of telomeric sequences at 2 years of
age, even in various stem cell populations,
and this change is concomitant with the
loss of regenerative capacity associated
with mouse aging. In addition, telome-
rase-deficient mice from the first genera-
tion (G1Terc�/�) exhibit a significant
decrease in median and maximum
longevity and a higher incidence of age-
related pathologies and stem cell dysfunc-
tion compared with wild-type mice (Flores
et al., 2005; Garcia-Cao et al., 2006), indi-
cating that, as in humans, telomerase
activity is rate limiting for natural mouse
longevity and aging. These results suggest
that strategies aimed to increase telome-
rase activity may delay natural mouse
aging. Further supporting this notion, it
was recently shown that overexpression
of TERT in the context of mice engineered
to be cancer resistant owe to increase
expression of tumor suppressor genes
(Sp53/Sp16/SARF/TgTERT mice) was
sufficient to decrease telomere damage
with age, delay aging, and increasemedian
longevity by 40% (Tomas-Loba et al.,
2008). However, it remains to be seen
whether telomerase reactivation late in life
would be sufficient to delay natural mouse
aging and extend mouse longevity without
increasing cancer incidence.
In summary, these proof-of-principle
studies using genetically modified mice
are likely to encourage the development
of targeted therapeutic strategies based
on reactivation of telomerase function.
Indeed, small molecule telomerase acti-
vators have been reported recently and
have demonstrated some preliminary
health-span beneficial effects in humans
(Harley et al., 2010). Identifying drugable
targets and candidate activators clearly
opens a new window for the treatment of
age-associated degenerative diseases.
REFERENCES
Flores, I., Cayuela, M.L., and Blasco, M.A. (2005).Science 309, 1253–1256.
Flores, I., Canela, A., Vera, E., Tejera, A., Cotsare-lis, G., and Blasco, M.A. (2008). Genes Dev. 22,654–667.
Garcia-Cao, I., Garcia-Cao, M., Tomas-Loba, A.,Martin-Caballero, J., Flores, J.M., Klatt, P., Blasco,M.A., and Serrano, M. (2006). EMBO Rep. 7,546–552.
Harley, C.B., Futcher, A.B., and Greider, C.W.(1990). Nature 345, 458–460.
Harley, C.B., Liu, W., Blasco, M., Vera, E.,Andrews, W.H., Briggs, L.A., and Raffaele, J.M.(2010). Rejuvenation Res. 14, in press. Publishedonline September 7, 2010. 10.1089/rej.2010.1085.
Jaskelioff, M., Muller, F.L., Paik, J.H., Thomas, E.,Jiang, S., Adams, A.C., Sahin, E., Kost-Alimova,M., Protopopov, A., Cadinanos, J., et al. (2010).Nature. 10.1038/nature09603.
Mitchell, J.R., Wood, E., and Collins, K. (1999).Nature 402, 551–555.
Samper, E., Flores, J.M., and Blasco, M.A. (2001).EMBO Rep. 2, 800–807.
Siegl-Cachedenier, I., Flores, I., Klatt, P., andBlasco, M.A. (2007). J. Cell Biol. 179, 277–290.
Tomas-Loba, A., Flores, I., Fernandez-Marcos,P.J., Cayuela, M.L., Maraver, A., Tejera, A., Borras,C., Matheu, A., Klatt, P., Flores, J.M., et al. (2008).Cell 135, 609–622.
Cell Stem Cell
Previews
HGPS-Derived iPSCs For The Ages
Tom Misteli1,*1National Cancer Institute, NIH, Bethesda, MD 20892, USA*Correspondence: mistelit@mail.nih.govDOI 10.1016/j.stem.2010.12.014
In this issue of Cell Stem Cell, Zhang et al. (2011) generate patient-derived iPSCs for one of the major prema-ture aging diseases, Hutchinson-Gilford Progeria Syndrome (HGPS). These cells are amuch-needed new toolto study HGPS, and their use may lead to novel insights into mechanisms of aging.
Some problems in biology are more
difficult to study than others. Human
aging is certainly one of them. Most
conclusions regarding molecular mecha-
nism of human aging rely onmere correla-
tion, and direct experimental testing
is generally not feasible. One approach
to dissect the molecular basis of human
aging is to study naturally occurring
premature aging disorders. One of the
most dramatic and prominent of such
4 Cell Stem Cell 8, January 7, 2011 ª2011 El
diseases is Hutchinson-Gilford Progeria
Syndrome (HGPS). Zhang et al. (2011)
now report the generation of induced
pluripotent stem cells (iPSCs) from
HGPS cells, providing a powerful new
tool to unravel the molecular and physio-
logical mechanisms of premature and
normal aging.
HGPS is a truly remarkable disease in
many ways. To start with, it affects an
unusually wide spectrum of tissues and
sevier Inc.
leads to the development of highly diverse
symptoms ranging from depletion of
subcutaneous fat to loss of hair and
tendon contractures. The diversity of
affected tissues pointed early on to stem
cell defects as a likely disease mecha-
nism. Most relevant in patients are
vascular defects and recurring strokes,
which invariably are fatal in patients in
their mid- to late teens (Hennekam,
2006). The disease is exceedingly rare
Cell Stem Cell
Previews
with only about 200 patients in the world
at any time, making access to relevant
tissues very difficult. HGPS is also re-
markable in how much we know about
its molecular and cellular basis. HGPS is
caused by a mutation in the LMNA gene
encoding the intermediate filament
proteins lamin A and C, key architectural
components of the cell nucleus and both
involved in higher-order genome organi-
zation (Worman et al., 2010). The disease
mutation leads to activation of a cryptic
splice site in LMNA and the production
of a dominant gain-of-function isoform of
lamin A, referred to as progerin. This
protein is permanently farnesylated at its
C terminus and accumulates in the
nuclear lamina, where it disrupts normal
lamina function.
Progerin is not only relevant to HGPS,
but also to normal aging, because the
cryptic splice site which creates progerin
is also used at low frequency in healthy
individuals and progerin can be found in
normal tissues (Scaffidi and Misteli,
2006). Further parallels between HGPS
and normal aging are suggested, given
that several cellular defects such as loss
of epigenetic marks and increased DNA
damage are observed in both settings. In
addition, HGPS patients and normally
aged individuals exhibit similar vascular
defects. Due to the rarity of the disease
and the fragility of the patients it is diffi-
cult, however, to obtain relevant biolog-
ical materials for molecular analysis,
and much of what we know about the
disease’s mechanisms comes from cul-
tured skin cells and animal models. The
generation of HGPS-derived iPSCs now
reported by Zhang et al. (2011) now
provides a much needed source for
tissue-specific cell lines with which to
probe the effect of progerin on tissue
function and differentiation.
The HGPS-derived iPSCs were gener-
ated from patient skin fibroblasts using
the standard Yamanaka method (Zhang
et al., 2011). The derived cells appeared
pluripotent since they form teratomas
and exhibit gene expression profiles akin
to established human embryonic stem
cell (hESC) lines. Interestingly, though,
the efficiency of iPSC generation from
HGPS patient cells was lower than from
wild-type control cells. This might be
due, as the authors suggest, to early
onset of senescence in HGPS cells, but
it might also have something to do with
an inhibitory role of progerin on the
large-scale chromatin reorganization
required during reprogramming. We
know that lamins tether chromatin to the
periphery and clamp it down into hetero-
chromatin and that progerin solidifies the
normally dynamic nuclear lamina (Dahl
et al., 2006). ESCs are one of few human
cell types that do not express lamins
A and C, and at the same time, they lack
heterochromatin, possibly as a means to
maintain broad genome plasticity. It is
conceivable that the presence of progerin
in HGPS cells prevents the dynamic
reorganization of chromatin required for
efficient reprogramming.
The derivation of HGPS-iPSCs is of
significant practical importance. The
described cells are able to differentiate
into five lineages, including vascular
smooth muscle cells (VSMCs) and
mesenchymal stem cells (MSCs) (Zhang
et al., 2011), confirming their multipo-
tency. These cells now offer a useful
experimental system to probe the effect
of progerin on the differentiation of
various cell lineages, something that
could not be done before because of the
inability to obtain tissue samples from
patients. These cells also open the door
to performing critical experiments, such
as transplantation of HGPS-derived
MSCs into the vasculature of animal
models to probe the physiological mech-
anisms that participate in the vascular
defects experienced by HGPS patients.
The HGPS-iPSCs, and their derivatives,
will also be useful for drug discovery. At
present, the only clinical strategy for
HGPS is farnesyltransferase inhibitors
(FTIs), which prevent the addition of the
C-terminal farnesyl group on progerin
(Capell and Collins, 2006). While FTIs
have been shown to reverse cellular
phenotypes and have a positive effect
on vasculature and on extension of life-
span in animal models, the nonspecific
nature of the drug might become limiting
in clinical applications. Lineage-differenti-
ated cell lines derived from HGPS-iPSCs
will provide ample and well-controlled
biological materials for the search of novel
drugs in high-throughput screens.
Although the HGPS-derived iPSCs
appear to differentiate normally in vitro,
they are functionally compromised, pro-
viding some insights into disease mecha-
nism (Zhang et al., 2011). HGPS-iPSC-
derived cells are hypersensitive to various
Cell Stem C
forms of stress. Survival of HGPS-iPSC-
derived VSMCs was significantly reduced
under hypoxic conditions or when sub-
jected to extended electrical stimulation.
The latter is potentially relevant to their
pathological function because VSMCs
undergo extensive mechanical stress
in vivo due to the pulsing of the vascula-
ture, and the reduced survival and prolif-
eration observed in vitro may suggest
increased cell death in the vasculature
of HGPS patients. HGPS-iPSC-derived
MSCs were also functionally compro-
mised in vivo. When transplanted into an
ischemic hind-limb muscle, they were
unable to prevent necrosis, whereas
MSCs derived in parallel from control
iPSCs did. This failure may be due to the
inability of HGPS-derived MSCs to
replace vascular cells that are removed
due to their normal turnover and/or the
poor survival of these cells in the hypoxic
environment of the muscle. Although it
remains unclear why exactly the HGPS-
iPSC-derived MSCs failed to rescue
these defects, it is tempting to consider
that MSC transplantation may offer a
novel therapeutic option for HGPS. An
intriguing, albeit distant, goal may be the
generation of patient-derived MSCs in
which the LMNA mutation has been
corrected using recombination-based
approaches.
These observations onmuscle regener-
ation are also directly relevant to our
thinking about normal aging. Loss of
regeneration capacity has become a pre-
vailing, albeit quite obvious, model for
aging (Sharpless and DePinho, 2007). If
tissue cells, and particularly stem cells,
which are lost from a tissue due to normal
turnover, are not replaced efficiently,
tissues will, of course, deteriorate. It ap-
pears that in the case of HGPS, and likely
in normal aging, tissue stem cells become
increasingly unable to keep upwith regen-
eration of lost tissue cells. This pattern
may arise for several reasons. Tissue
stem cell numbers may be reduced due
to increased apoptosis, in the case of
HGPS possibly due to their inability to
cope with stress, for example, under
hypoxic conditions in tissues. In addition,
tissue stem cells might fail to self-renew,
or they may produce fewer and function-
ally impaired offspring. TheHGPS-derived
iPSCs should be useful in further resolving
the relevanceof thesevariouspathways to
organismal aging.
ell 8, January 7, 2011 ª2011 Elsevier Inc. 5
Cell Stem Cell
Previews
HGPS is an extraordinary disease,
and the generation of patient-derived
iPSCs is a significant milestone. This
step continues the remarkable progress
made in the last few years. After discovery
of the disease-causing gene in 2003, it
only took four years to initiate several
clinical trials. Much has been learnt along
the way about the biology of HGPS and its
relevance to normal aging. The generation
of iPSCs fromHGPS patients now heralds
another wave of rapid progress with
A Roundabout Wa
Kateri Moore1,2,*1Departments of Gene and Cell Medicine2Department of Developmental and RegeneraMount Sinai School of Medicine, New York, N*Correspondence: kateri.moore@mssm.eduDOI 10.1016/j.stem.2010.12.011
A new player in hematopoietic stemSmith-Berdan et al. (2010) demonstrso in cooperation with Cxcr4 to gui
Bone marrow (BM) transplantation has
been used for treatment of hematopoietic
disorders for some fifty years and repre-
sents a paradigm for all future stem cell
therapies. A number of cytokines, espe-
ciallygranulocytecolony-stimulating factor
(G-CSF), are known to mobilize hemato-
poietic stem and progenitor cells (HSPCs)
from their BM niches into the peripheral
blood (PB) (Papayannopoulou and Scad-
den, 2008). Indeed, mobilization is the
preferred method for obtaining transplant-
able HSC. Despite the number of currently
available HSPC mobilizing agents, a
significant number of donors mobilize
poorly. Therefore, identifying novel and
more efficient mobilization approaches is
of paramount clinical importance.
Understanding the molecular frame-
work of how the niche regulates retention
and release of stem cells provides the
ground onwhich to base alternativemobi-
lization strategies. The basic processes of
transplantation are homing to, engraft-
ment in, and retention of HSCs in the
niche. Mobilization may thus be under-
6 Cell Stem Cell 8, January 7, 2011 ª2011 E
implications for HGPS disease mecha-
nisms, for aging in general, and potentially
as a tool to develop novel strategies to
combat vascular disease.
REFERENCES
Capell, B.C., and Collins, F.S. (2006). Nat. Rev.Genet. 7, 940–952.
Dahl, K.N., Scaffidi, P., Islam, M.F., Yodh, A.G.,Wilson, K.L., and Misteli, T. (2006). Proc. Natl.Acad. Sci. USA 103, 10271–10276.
y to the Niche
tive BiologyY 10029, USA
cell (HSC)-niche interactions is introdate that Robo4 is involved in HSC engde stem cells to and secure them in
stood as the process of breaking the
bonds of stem cell retention in the BM
niche or enhancement of the existing
means that allow HSCs to enter the PB.
The cellular milieu and molecular mecha-
nisms that mediate these processes are
starting to be revealed but, at best, remain
poorly understood (Garrett and Emerson,
2009). The Cxcr4/Cxcl12 axis has been
identified as critically important in homing,
engraftment, and retention in theBM (Lap-
idot et al., 2005). Previouswork has shown
that the Cxcr4 antagonist AMD3100 can
mobilize both mouse and human HSPCs
and has found use clinically as an adjunct
therapy for poor G-CSF mobilizers (Brox-
meyer et al., 2005). In this issue of Cell
Stem Cell, Smith-Berdan et al. show that
Roundabout 4 (Robo4), a neuronal guid-
ance molecule, regulates engraftment
and mobilization and, in cooperation with
Cxcr4, localizes HSCs to the niche.
Previous profiling studies by the senior
author had revealed that Robo4 was ex-
pressed at high levels in long-term HSCs
(Forsberg et al., 2005). In the present
lsevier Inc.
Hennekam, R.C. (2006). Am. J.Med. Genet. A. 140,2603–2624.
Scaffidi, P., and Misteli, T. (2006). Science 312,1059–1063.
Sharpless, N.E., and DePinho, R.A. (2007). Nat.Rev. Mol. Cell Biol. 8, 703–713.
Worman, H.J., Ostlund, C., and Wang, Y. (2010).Cold Spring Harb. Perspect. Biol. 2, a000760.
Zhang, J.L., Zhu, Q., Zhou, G., Sui, F., Tan, L.,Mutalif, A., Navasankari, R., Zhang, Y., Tse, H.-F.,Stewart, C., et al. (2011). Cell Stem Cell 8, thisissue, 31–45.
uced in this issue of Cell Stem Cell.raftment andmobilization and doesthe niche.
work, the authors show that Robo4
becomes downregulated upon differenti-
ation, consistent with the observations of
Shibata et al., who also demonstrated
that repopulating cells segregated to the
Robo4+ fraction of HSPCs (Shibata et al.,
2009). Notably, Smith-Berdan et al. also
found that Robo4 expressionwas dramat-
ically downregulated in mobilized HSCs.
To determine a functional role for Robo4
in HSCs, the authors investigated Robo4
knockout mice. Robo4�/� mice appear
normal but have defects in vascular integ-
rity and angiogenesis (Jones et al., 2008).
An analysis of the stem cell compartments
revealed that Robo4�/� mice had a spe-
cific decrease of HSCs in the BM with a
reciprocal increase in PB, suggesting
poor BM retention. Upon transplantation,
Robo4�/� HSCs engrafted poorly, but
those that did engraft contributed to a
normal spectrum of blood cell lineages.
In addition, the ability of Robo4�/� HSC
tomake spleen colonies was normal, sug-
gesting that the engraftment defect was
likely because of a specific impairment of
HGPS is an extraordinary disease,
and the generation of patient-derived
iPSCs is a significant milestone. This
step continues the remarkable progress
made in the last few years. After discovery
of the disease-causing gene in 2003, it
only took four years to initiate several
clinical trials. Much has been learnt along
the way about the biology of HGPS and its
relevance to normal aging. The generation
of iPSCs fromHGPS patients now heralds
another wave of rapid progress with
implications for HGPS disease mecha-
nisms, for aging in general, and potentially
as a tool to develop novel strategies to
combat vascular disease.
REFERENCES
Capell, B.C., and Collins, F.S. (2006). Nat. Rev.Genet. 7, 940–952.
Dahl, K.N., Scaffidi, P., Islam, M.F., Yodh, A.G.,Wilson, K.L., and Misteli, T. (2006). Proc. Natl.Acad. Sci. USA 103, 10271–10276.
Hennekam, R.C. (2006). Am. J.Med. Genet. A. 140,2603–2624.
Scaffidi, P., and Misteli, T. (2006). Science 312,1059–1063.
Sharpless, N.E., and DePinho, R.A. (2007). Nat.Rev. Mol. Cell Biol. 8, 703–713.
Worman, H.J., Ostlund, C., and Wang, Y. (2010).Cold Spring Harb. Perspect. Biol. 2, a000760.
Zhang, J.L., Zhu, Q., Zhou, G., Sui, F., Tan, L.,Mutalif, A., Navasankari, R., Zhang, Y., Tse, H.-F.,Stewart, C., et al. (2011). Cell Stem Cell 8, thisissue, 31–45.
Cell Stem Cell
Previews
A Roundabout Way to the Niche
Kateri Moore1,2,*1Departments of Gene and Cell Medicine2Department of Developmental and Regenerative BiologyMount Sinai School of Medicine, New York, NY 10029, USA*Correspondence: kateri.moore@mssm.eduDOI 10.1016/j.stem.2010.12.011
A new player in hematopoietic stem cell (HSC)-niche interactions is introduced in this issue of Cell Stem Cell.Smith-Berdan et al. (2010) demonstrate that Robo4 is involved in HSC engraftment andmobilization and doesso in cooperation with Cxcr4 to guide stem cells to and secure them in the niche.
Bone marrow (BM) transplantation has
been used for treatment of hematopoietic
disorders for some fifty years and repre-
sents a paradigm for all future stem cell
therapies. A number of cytokines, espe-
ciallygranulocytecolony-stimulating factor
(G-CSF), are known to mobilize hemato-
poietic stem and progenitor cells (HSPCs)
from their BM niches into the peripheral
blood (PB) (Papayannopoulou and Scad-
den, 2008). Indeed, mobilization is the
preferred method for obtaining transplant-
able HSC. Despite the number of currently
available HSPC mobilizing agents, a
significant number of donors mobilize
poorly. Therefore, identifying novel and
more efficient mobilization approaches is
of paramount clinical importance.
Understanding the molecular frame-
work of how the niche regulates retention
and release of stem cells provides the
ground onwhich to base alternativemobi-
lization strategies. The basic processes of
transplantation are homing to, engraft-
ment in, and retention of HSCs in the
niche. Mobilization may thus be under-
6 Cell Stem Cell 8, January 7, 2011 ª2011 E
stood as the process of breaking the
bonds of stem cell retention in the BM
niche or enhancement of the existing
means that allow HSCs to enter the PB.
The cellular milieu and molecular mecha-
nisms that mediate these processes are
starting to be revealed but, at best, remain
poorly understood (Garrett and Emerson,
2009). The Cxcr4/Cxcl12 axis has been
identified as critically important in homing,
engraftment, and retention in theBM (Lap-
idot et al., 2005). Previouswork has shown
that the Cxcr4 antagonist AMD3100 can
mobilize both mouse and human HSPCs
and has found use clinically as an adjunct
therapy for poor G-CSF mobilizers (Brox-
meyer et al., 2005). In this issue of Cell
Stem Cell, Smith-Berdan et al. show that
Roundabout 4 (Robo4), a neuronal guid-
ance molecule, regulates engraftment
and mobilization and, in cooperation with
Cxcr4, localizes HSCs to the niche.
Previous profiling studies by the senior
author had revealed that Robo4 was ex-
pressed at high levels in long-term HSCs
(Forsberg et al., 2005). In the present
lsevier Inc.
work, the authors show that Robo4
becomes downregulated upon differenti-
ation, consistent with the observations of
Shibata et al., who also demonstrated
that repopulating cells segregated to the
Robo4+ fraction of HSPCs (Shibata et al.,
2009). Notably, Smith-Berdan et al. also
found that Robo4 expressionwas dramat-
ically downregulated in mobilized HSCs.
To determine a functional role for Robo4
in HSCs, the authors investigated Robo4
knockout mice. Robo4�/� mice appear
normal but have defects in vascular integ-
rity and angiogenesis (Jones et al., 2008).
An analysis of the stem cell compartments
revealed that Robo4�/� mice had a spe-
cific decrease of HSCs in the BM with a
reciprocal increase in PB, suggesting
poor BM retention. Upon transplantation,
Robo4�/� HSCs engrafted poorly, but
those that did engraft contributed to a
normal spectrum of blood cell lineages.
In addition, the ability of Robo4�/� HSC
tomake spleen colonies was normal, sug-
gesting that the engraftment defect was
likely because of a specific impairment of
Cell Stem Cell
Previews
Robo4�/� HSCs to home, engraft, and
remain in the BM.
On the basis of these results, the Fors-
berg group hypothesized that Robo4
mediates HSC adhesion to the niche and
that downregulation of Robo4 was a crit-
ical step enabling exit from the niche to
the bloodstream. Consistent with this
idea, the authors predicted that mobiliza-
tion induced by G-CSF treatment would
be elevated in Robo4 null mice. Instead,
they found that Robo4�/� HSCs were
delayed in their ability to mobilize in
response to G-CSF. Smith-Berdan et al.
next examined the well-known Cxcr4/
Cxcl12 axis and found that Cxcr4 expres-
sion in HSCs and Cxcl12/Sdf1 expression
in stromal cells was elevated in Robo4�/�
mice. Thus, a compensatory upregulation
of the Cxcr4/Cxcl12 axis likely explains
why Robo4�/� HSCs were slower to
mobilize. Mobilization experiments using
AMD3100, a Cxcr4 antagonist, in con-
junction with G-CSF or as the sole mobili-
zation agent, revealed that HSCs were
specifically mobilized at higher levels in
Robo4�/� mice. In order to test whether
inhibition of the Cxcr4/Cxcl12 axis specif-
ically affects stem cell homing, HSCs
were pretreated with AMD3100 before
transplantation. HSCs from both strains
homed less efficiently to BM after
AMD3100 pretreatment but even less so
when lacking Robo4, suggesting that
Robo4 cooperates with Cxcr4 in stem
cell homing. Taken together, these results
suggest that a Robo4 antagonist would
aid in specific mobilization of HSCs into
the bloodstream andmay have a potential
clinical use in combination with other
agents. As such, these experiments pro-
vide enticing evidence for a novel path-
way in stem cell homing, engraftment,
and mobilization from the niche.
The findings of Smith-Brennan et al.
point to an exciting new line of investiga-
tion in stem/niche cell interactions with
many questions to be probed in future
work. At the forefront of these questions
is whether the pattern of Robo4 expres-
sion in human HSCsmimics that in mouse
and whether nongenetic approaches
targeting Robo4 would be useful for
mobilization and purification of HSCs.
Mechanistically, the reciprocal loss of
Robo4 and the upregulation of the
Cxcr4/Cxcl12 axis remain to be defined.
Is there a point where the two pathways
intersect in their downstream signaling?
Of interest, Robo4 is expressed in endo-
thelium and functions in vascular sprout-
ing upon activation by its ligand Slit2. It
will be interesting to determine if Robo4
in this context acts via Slit2 and if there
is an additional coreceptor. Activated
Robo4 also stabilizes the vascular
network through inhibition of endothelial
permeability (Jones et al., 2008). Thus,
how loss of Robo4 affects the endothelial
function will be an important topic to
address in future studies. Finally, where
are the Robo4+ HSC in the BM normally
localized and to where do they home?
Osteoblasts upregulate the expression
of Slit2 after 5-FU treatment (Shibata
et al., 2009), and Slit2 expression has
very recently been found in the extramural
cells surrounding endothelium in devel-
oping mammary tissue (Marlow et al.,
2010). It would be very interesting if Slit2
Cell Stem C
expression were found in the Cxcl12
abundant reticular (CAR) cells that sur-
round endothelium, localize near the
endosteum, and are thought to play a
role in the stem cell niche (Sugiyama
et al., 2006). Indeed, it should be very
revealing to pursue this roundabout way
into and out of the niche.
REFERENCES
Broxmeyer, H.E., Orschell, C.M., Clapp, D.W.,Hangoc, G., Cooper, S., Plett, P.A., Liles, W.C.,Li, X., Graham-Evans, B., Campbell, T.B., et al.(2005). J. Exp. Med. 201, 1307–1318.
Forsberg, E.C., Prohaska, S.S., Katzman, S.,Heffner, G.C., Stuart, J.M., and Weissman, I.L.(2005). PLoS Genet. 1, e28.
Garrett, R.W., and Emerson, S.G. (2009). Cell StemCell 4, 503–506.
Jones, C.A., London, N.R., Chen, H., Park, K.W.,Sauvaget, D., Stockton, R.A., Wythe, J.D., Suh,W., Larrieu-Lahargue, F., Mukouyama, Y.S., et al.(2008). Nat. Med. 14, 448–453.
Lapidot, T., Dar, A., and Kollet, O. (2005). Blood106, 1901–1910.
Marlow, R., Binnewies, M., Sorensen, L.K., Mon-ica, S.D., Strickland, P., Forsberg, E.C., Li, D.Y.,and Hinck, L. (2010). Proc. Natl. Acad. Sci. USA107, 10520–10525.
Papayannopoulou, T., and Scadden, D.T. (2008).Blood 111, 3923–3930.
Shibata, F., Goto-Koshino, Y., Morikawa, Y., Ko-mori, T., Ito,M., Fukuchi, Y., Houchins, J.P., Tsang,M., Li, D.Y., Kitamura, T., et al. (2009). Stem Cells27, 183–190.
Smith-Berdan, S., Nguyen, A., Hassanein, D., Zim-mer, M., Ugarte, F., Ciriza, J., Li, D., Garcıa-Ojeda,M., Hinck, L., and Forsberg, C. (2010). Cell StemCell 8, this issue, 72–83.
Sugiyama, T., Kohara, H., Noda, M., and Naga-sawa, T. (2006). Immunity 25, 977–988.
ell 8, January 7, 2011 ª2011 Elsevier Inc. 7
Cell Stem Cell
Previews
There and Back Again:Hair Follicle Stem Cell Dynamics
Katherine A. Fantauzzo1 and Angela M. Christiano1,2,*1Department of Dermatology2Department of Genetics and DevelopmentColumbia University, New York, NY 10032, USA*Correspondence: amc65@columbia.eduDOI 10.1016/j.stem.2010.12.018
Recently in Cell, Hsu et al. (2011) defined the relationship between stem cells and differentiated progenywithin a hair follicle lineage. Their work reveals that stem cell descendants that havemigrated out of the bulgecan return to this niche and actively contribute to its function.
Stem cells are defined by self-renewal
and multipotency and participate in
homeostasis and injury repair in numerous
tissues within the adult organism. They
are often characterized by their relative
quiescence, as well as residence in
specialized niches throughout the body.
While differentiated stem cell progeny
have beendescribed formultiple lineages,
thecircumstancesunderwhichadaughter
cell, or descendant, adopts a permanently
committed state remain unclear. Recently
in Cell, Hsu et al. (2011) used the murine
hair follicle (HF) as a model system to
address questions of fate commitment
and function for multiple cell types in
a stem cell lineage, both within and
outside of the niche. Their findings
demonstrate that recent HF stem cell
derivatives return to the bulge niche to
serve as future stem cells, while more
committed progeny home back to a
distinct layer of the niche to maintain
stem cell quiescence.
Throughout the postnatal hair cycle, the
follicle undergoes phases of regression
(catagen), rest (telogen), and regeneration
(anagen), producing a new hair fiber
during each cycle. Over 20 years ago,
a reservoir of slow-cycling, label-retaining
cells was identified by nucleotide pulse-
chase experiments in the permanent,
upper portion of the murine follicle,
continuous with the outer root sheath
(ORS), in a compartment known as the
‘‘bulge’’ (Cotsarelis et al., 1990). While
this local expansion of the ORS is not
visible in murine pelage (coat) follicles
until approximately 3 weeks after birth,
recent findings have established that
slow-cycling bulge progenitors exist
much earlier and are specified during
8 Cell Stem Cell 8, January 7, 2011 ª2011 El
embryonic development (Nowak et al.,
2008). Clonal and in vivo lineage analyses
of bulge cells, coupled with reconstitution
assays, revealed that these undifferenti-
ated cells are able to self-renew and
contribute to all epithelial lineages in the
skin, including the HF, sebaceous gland,
and interfollicular epidermis (Blanpain
et al., 2004; Morris et al., 2004).
During periods of HF growth, previous
transplantation and genetic marking
studies havedemonstrated that stemcells
from the bulge migrate downward along
the ORS to the base of the HF, giving rise
to transit-amplifying matrix cells, which in
turn proliferate and differentiate to
generate the various layers of the inner
root sheath and hair shaft (Oshima et al.,
2001; Nowak et al., 2008). The character-
istics of these migratory cells upon exiting
the bulge have not previously been
defined, though several lines of evidence
point to retained stem cell properties. For
example, portions of the vibrissa (whisker)
follicle ORS located below the bulge are
able to generate clonogenic keratinocytes
and form skin epithelial lineages upon
embryo transplantation in a hair-cycle-
dependent manner (Oshima et al., 2001).
Moreover, ORS cells express numerous
bulge stem cell markers that are not found
in the more differentiated epithelial cells
at the base of the follicle (Fuchs, 2009),
lending further support to the notion that
early bulge descendants may retain
some properties of their stem cell precur-
sors. However, the in vivo dynamics of
thesecells beyond follicle growth and their
particular relationship to the bulge stem
cell niche have remained elusive.
Hsu and colleagues (2011) have used
a sophisticated combination of lineage
sevier Inc.
tracing and nucleotide pulse-chase
experiments at various time points to
monitor the activity of ORS cells
throughout the HF cycle and precisely
determine the timing and nature of their
lineage commitment. The authors first
employed a Tet-Off system whereby
administration of doxycycline repressed
expression of a histone H2B-GFP trans-
gene throughout the skin epithelium. A
long doxycycline chase that began before
the first postnatal growth phase revealed
that ORS cells along the length of the
follicle display a range of proliferative
activity during anagen, with the cells
closest to their bulge predecessors
cycling the slowest and, further, that
these upper ORS cells survive the
destructive phase of the cycle. By prefer-
entially labeling upper ORS cells during
midanagen utilizing a tamoxifen-inducible
LacZ transgene driven by the Lgr5
promoter or a short BrdU pulse in combi-
nation with the Tet-Off H2B-GFP model,
the authors demonstrated that upper
ORS cells are the main contributors to
the new bulge and hair germ during
telogen.
Postponing the BrdU pulses until late
anagen using the Tet-Off H2B-GFP
system revealed that cells in the mid-
zone of the ORS supply additional cells
to the telogen hair germ. The authors
then employed a Tet-On H2B-GFP
lineage tracing model under the control
of the keratin 14 (K14) promoter to induce
GFP expression in the ORS upon applica-
tion of doxycycline during midanagen.
Coupling this system with a BrdU pulse
in late anagen, the authors demonstrated
that lower ORS cells are also able to home
back to the stem cell niche, giving rise to
Cell Stem Cell
Previews
cells in the CD34�K6+ inner layer of the
new bulge.
The cells in this unique inner bulge
population expressed numerous HF
stem cell transcription factors and were
shown to remain quiescent and stationary
during the following hair cycle through
further nucleotide pulse-chase experi-
ments. Additional lineage tracing analysis
in the Tet-Off H2B-GFP system with a
chase throughout multiple hair cycles
revealed that, importantly, CD34+ new
bulge and hair germ cells are the sole
contributors to newly developing hair folli-
cles, effectively ruling out a role for the
inner bulge layer in HF homeostasis.
The authors next explored functional
differences between the bulge layers
using wounding and cell ablation experi-
ments, together with BrdU pulses applied
at the time of injury. Upon introduction of
punch wounds to the skin or ablation of
CD34+ bulge cells by means of an induc-
ible K15-DTR (diphtheria toxin receptor)
model, CD34+ new and old bulge cells
briefly proliferated during wound repair,
whereas K6+ inner bulge cells remained
quiescent. Alternatively, targeted ablation
of K6+ bulge cells through an inducible
Sox9-DTRmodel led to hair loss and rapid
re-entry into anagen, marked by a pro-
longed increase in CD34+ bulge cell prolif-
eration. In examining the mechanism by
which K6+ bulge cells might contribute
to HF quiescence, the authors revealed
high expression of Fgf18 and Bmp6 in
these cells and demonstrated that injec-
tion of each factor was capable of inhibit-
ing activation of CD34+ bulge cells at the
time of K6+ cell ablation.
Several novel findings of broad impor-
tance to both HF and stem cell biology
are introduced in this study. First, slow-
cycling stem cell descendants persist
outside of the niche during hair growth.
These cells survive the widespread
apoptosis of the lower follicle during cata-
gen and, furthermore, serve as functional
stem cells during the next cycle of follicle
regeneration. Hsu and colleagues (2011)
thus provide direct evidence to support
the hypothesis foreshadowed by previous
studies (Oshima et al., 2001; Jaks et al.,
2008) that HF stemness is not wholly
maintained by the bulge niche but is an
intrinsic characteristic of the cell itself,
consistent with evidence from the hema-
topoietic stem cell field.
Second, rapidly cycling ORS cells are
also able survive catagen and return to
the bulge, albeit in a distinct layer. This
observation puts into context the prior
finding that actively cycling Lgr5+ bulge
and hair germ descendants in the mature
follicle return to these structures by the
following telogen (Jaks et al., 2008). While
these lower ORS cells are permanently
committed and no longer possess prolif-
erative potential, they serve two vital roles
in the stem cell niche, namely, anchoring
the club hair and maintaining stem cell
quiescence during telogen. The cellular
dynamics demonstrated here lend sup-
port to key aspects of the HF predetermi-
nation hypothesis proposed by Pante-
leyev et al. (2001), in that lower ORS
cells are spared from apoptosis during
catagen and retain a memory of the
previous hair cycle that shapes their
future function in the follicle.
Finally, the authors contribute signifi-
cant functional data to substantiate the
heterogeneity of cell types in the bulge
described by Blanpain et al. (2004). They
clearly demonstrate that cells in the
CD34+ outer bulge layer function as
bona fide stem cells capable of follicle
regeneration and wound repair, consis-
tent with previous genetic lineage tracing
results (Morris et al., 2004; Ito et al.,
2005), while CD34�K6+ inner bulge cells,
though quiescent, actively contribute to
the niche environment. Future studies in
the field must now take into account that
Cell Stem Ce
HF stem cells beyond the first postnatal
cycle are not naive and immobile resi-
dents of their niche, but that their move-
ments during previous cycles may have
exposed them to various signaling
climates along the length of the follicle
that may have imparted these cells with
as yet unrecognized attributes.
Having established a range of proper-
ties and fates for HF stem cell descen-
dants, it will now be interesting to address
how these characteristics are acquired
and maintained outside of the bulge
niche. In particular, the question of
whether HF stemness is directly corre-
lated with the number of cell divisions or
influenced by additional signaling and
architectural cues in the local environ-
ment. The unique combination of lineage
tracing and labeling techniques employed
in this study provide a robust model with
which to explore these questions.
REFERENCES
Blanpain, C., Lowry, W.E., Geoghegan, A., Polak,L., and Fuchs, E. (2004). Cell 118, 635–648.
Cotsarelis, G., Sun, T.-T., and Lavker, R.M. (1990).Cell 61, 1329–1337.
Fuchs, E. (2009). Cell 137, 811–819.
Hsu, Y.-C., Pasolli, H.A., and Fuchs, E. (2011). Cell144, 92–105.
Ito, M., Liu, Y., Yang, Z., Nguyen, J., Liang, F.,Morris, R.J., and Cotsarelis, G. (2005). Nat. Med.11, 1351–1354.
Jaks, V., Barker, N., Kasper, M., van Es, J.H.,Snippert, H.J., Clevers, H., and Toftgard, R.(2008). Nat. Genet. 40, 1291–1299.
Morris, R.J., Liu, Y., Marles, L., Yang, Z., Trempus,C., Li, S., Lin, J.S., Sawicki, J.A., and Cotsarelis, G.(2004). Nat. Biotechnol. 22, 411–417.
Nowak, J.A., Polak, L., Pasolli, H.A., and Fuchs, E.(2008). Cell Stem Cell 3, 33–43.
Oshima, H., Rochat, A., Kedzia, C., Kobayashi, K.,and Barrandon, Y. (2001). Cell 104, 233–245.
Panteleyev, A.A., Jahoda, C.A., and Christiano,A.M. (2001). J. Cell Sci. 114, 3419–3431.
ll 8, January 7, 2011 ª2011 Elsevier Inc. 9
Cell Stem Cell
Previews
Transition of Endothelium to Cartilage and Bone
Ofer Shoshani1 and Dov Zipori1,*1Department of Molecular Cell Biology, Weizmann Institute of Science, Rehvot 76100, Israel*Correspondence: dov.zipori@weizmann.ac.ilDOI 10.1016/j.stem.2010.12.004
Mesenchymal stromal cells (MSCs) are capable of differentiating into bone-forming osteoblasts. A recentNature Medicine study (Medici et al., 2010) shows that the mislocalized bone in the human disease fibrodis-plasia ossificans progressiva (FOP) originates from vascular endothelium that gives rise to MSCs.
Ectopic bone formation in soft tissues is
a common occurrence following trauma,
internal muscular bleeding, osteoarthritis
(OA), inflammation, and also in specific
genetic disorders. One such condition
is fibrodisplasia ossificans progressiva
(FOP), in which cartilage and bone form
pathologically within soft tissues rather
than only within the skeleton. Olsen
and colleagues studied the source of
ectopic bone in individuals inflicted
with FOP (Medici et al., 2010). Mesen-
chymal stromal cells (MSCs) are multipo-
tent cells with bone-, fat-, and cartilage-
forming potential that are widespread
in calcified and soft tissues and have
been presumed to be the source of
mislocalized bone. In FOP, heterotopic
ossification is thought to occur through
mesenchymal condensation, followed
by chondrogeneis, and finally endochon-
dral ossification. Olsen and colleagues
show that vascular endothelial cells that
undergo endothelial-to-mesenchymal
transition (EndMT) are the source of cells
that generate cartilage and bone lesions
(Medici et al., 2010). This phenomenon
of transdifferentiation of endothelium into
bone, as demonstrated in the FOPmodel,
shows that the human disease recapitu-
lates hallmarks of embryonic plasticity.
The ability of FOP-derived endothelial
cells to undergo EndMT is related to
a mutation in the receptor ALK2, which
causes its constitutive activation. This
observation leaves open the possibility
that the unmutated form of ALK2 might
not mediate EndMT. However, the
authors also demonstrate that activation
of endothelial cells with ALK2 ligands,
such as transforming growth factor
(TGF)-b superfamily cytokines (Figure 1),
results in the transition of endothelium
into mesenchyme. Therefore, EndMT
may be a physiological occurrence, and
10 Cell Stem Cell 8, January 7, 2011 ª2011 E
not necessarily restricted to a diseased
state.
The Olson et al. study makes a strong
case that EndMT provides a mechanism
for heterotopic bone formation, based,
in part, on their analysis of diseased tis-
sues. Both humans with FOP and mice
with mutated ALK2 develop heterotopic
bone, the phenotype of which includes
expression of relevant cartilage and
bone markers, as well as the endothelial
markers TIE2 and vWF. These observa-
tions are substantiated through the
use of reporter mice that express an
enhanced green fluorescence protein
(EGFP) transgene under the control of
the endothelial-specific Tie2 promoter.
Analysis of EGFP expression in sections
of ligand-induced heterotopic cartilage
and bone revealed that many green endo-
thelial-derived cells are also Sox9 (carti-
lage) and osteocalcin (bone) positive
(Medici et al., 2010). The hybrid endothe-
lial/mesenchymal phenotype observed
in vivo suggests that mutant ALK2 medi-
ates the transition from endothelium to
cartilage and bone, and results from
subsequent culture experiments support
this hypothesis. Specifically, expression
of the mutant ALK2 in human cultured
endothelial cells (HUCEC) and in human
cutaneous microvascular endothelial
cells (HCMEC) resulted in the acquisition
of fibroblast morphology, associated
with the expression of classical markers
of epithelial-to-mesenchymal transition
(EMT), including Snail and Slug. The tran-
sition of endothelium into mesenchyme is
also supported by the appearance of the
fibroblast marker FSP-1 in early lesions
of the mutant mice induced with the
ALK2 ligand, bone morphogenic protein
(BMP)-4. In both in vitro experiments
and an in vivo immunocompromized
mouse model, the mutant ALK2 express-
lsevier Inc.
ing endothelial cells gave rise to osteo-
genic, adipogenic, and chondrogenic
mesodermal lineages, consistent with
the proposal that the endothelial cells de-
differentiated into MSCs. This pathway,
involving the acquisition of MSC pheno-
type and function by endothelium, is not
dependent on the presence of the consti-
tutively active, mutant ALK2. Indeed,
endothelial cells exposed to the ALK2
ligands TGF-b2 and BMP4 also differenti-
ated, both in vitro and in vivo, into the
aforementioned three mesodermal line-
ages. Finally, because the knockdown of
this receptor prevented the transition,
the study provides evidence that EndMT
in this system is dependent on signals
downstream of ALK2.
The combination of in vivo observa-
tions, in vitro findings, and the analysis
of the molecular mechanism of EndMT
(Medici et al., 2010) constitute a solid
study that demonstrates an alternate
pathway of chondrogenesis and osteo-
genesis. One caveat to the findings pre-
sented by Olsen and colleagues that will
require further investigation relates to the
current dependence on the expression
of specific cell markers. Surface pheno-
type determination may not always iden-
tify cell lineages faithfully. Further analysis
that establishes specific endothelial func-
tion is required in order to complement
the existing assessment of functional
mesenchymal traits, namely, multilineage
differentiation potential. Future studies
should also explore the possibility that
other cases of ectopic ossification might
be due to EndMT. In osteoarthritis (OA),
as one example, ectopoic ossification
causes severe pain and disability. The
mechanism of OA is not well understood,
and elucidation of the possible contribu-
tion of themicrovasculature is now neces-
sary. Futhermore, EndMT may not be
Figure 1. A Putative Cycle of Cell-Fate TransitionsVascular endothelium activated by appropriate ALK2 ligands, such as TGF-b2, undergoes an endothelial-mesenchymal transition (EndMT), leading to acquisition of fibroblast morphology and markers, and multi-potency that defines mesenchymal stromal cells (MSCs). Multipotency is demonstrated by the ability ofthe cells produced by EndMT to differentiate, upon specific induction, into osteoblasts, adipocytes,and chondrocytes. The reported potential of MSCs to differentiate into endothelial cells completes theputative cycle. The question mark indicates that this portion of the cycle has not been demonstrated inthe present study.
Cell Stem Cell
Previews
restricted to pathological conditions, and
bone remodeling and fracture repair may
entail similar processes in which the
vasculature serves as the source of oste-
ogenic cells. In addition, it is tempting to
speculate that EndMT may represent
a physiological mechanism for the gener-
ation of MSCs. Perivascular cells, specif-
ically pericytes (Crisan et al., 2008), have
been suggested to be the in vivo counter-
parts of cultured MSCs. The present
study provides evidence that the endo-
thelium itself serves as an alternative
source.
The observation of EndMT in adult
tissues, albeit diseased, reawakens the
debate as to the plasticity of cell behavior
in the adult. Studies published almost
10 years ago proposed that adult hemato-
poietic stem cells, adult MSCs, and
a variety of tissue-specific progenitors
can undergo transdifferentiation. For
example, Sharkis and colleagues pub-
lished that bone-marrow-derived cells
could produce mature cells of epithelial
organs, such as the liver and lung (Krause
et al., 2001). Other examples of transitions
from one fully differentiated cell type into
mature cells of a different lineage/tissue
have been reported and were suggested
to entail dedifferentiation. The present
report by Olsen et al. can be added to
the list of studies supporting the notion
of cellular plasticity in adult mammalian
tissues. Notably, this report is not iso-
lated. Several other recent studies also
support the possibility that cellular plas-
ticity is neither restricted to the embryo
nor to diseased adult tissues. Studies of
mouse and human spermatogonia high-
light the fact that these cells are easily
reprogrammable under mild conditions
(Conrad et al., 2008), which do not require
the use of harsh genetic manipulations.
Even more striking is the finding that the
dedifferentiation of maturing germ cells
back into spermatogonial stem cells
occurs under stress (Nakagawa et al.,
2007), and even spontaneously and
frequently (Klein et al., 2010), supporting
themodel that dedifferentiation is a physi-
ological phenomenon. An example of
mammalian dedifferentiation and trans-
differentiation has also been recently
observed in the pancreas (Thorel et al.,
2010).
A fraction of the MSC population
constitutes multipotent cells that give
Cell Stem Cel
rise to a variety of cell types, including
endothelium (Conrad et al., 2009). Thus,
a complete cycle may exist in which
EndMT leads to the formation of MSCs,
which, in turn, differentiate back into
endothelium through a mesenchymal-to-
endothelial transition (MEndT) (Figure 1).
This reversibility in cell-fate determination
has been used to propose the model of
a ‘‘stem state’’ (Zipori, 2004), in which
stemness is considered a transient state
in a cell’s life cycle. In other words, cells
may differentiate, but this change does
not determine their status permanently.
Upon demand for tissue repair, cells
downstream in the differentiation cas-
cade may ‘‘turn back’’ and re-exhibit
stemness by regaining additional lineage
potentials that had previously been lost.
The stem state notion predicts that dedif-
ferentiation is possible in mammalian
tissues (Zipori, 2009), and this proposal
is supported by the current findings
that supposedly unipotent adult endothe-
lium can, when prompted, re-exhibit
multipotency.
REFERENCES
Conrad, S., Renninger, M., Hennenlotter, J., Wies-ner, T., Just, L., Bonin, M., Aicher, W., Buhring,H.J., Mattheus, U., Mack, A., et al. (2008). Nature456, 344–349.
Conrad, C., Niess, H., Huss, R., Huber, S., vonLuettichau, I., Nelson, P.J., Ott, H.C., Jauch,K.W., and Bruns, C.J. (2009). Circulation 119,281–289.
Crisan, M., Yap, S., Casteilla, L., Chen, C.W.,Corselli, M., Park, T.S., Andriolo, G., Sun, B.,Zheng, B., Zhang, L., et al. (2008). Cell Stem Cell3, 301–313.
Klein, A.M., Nakagawa, T., Ichikawa, R., Yoshida,S., and Simons, B.D. (2010). Cell Stem Cell 7,214–224.
Krause, D.S., Theise, N.D., Collector, M.I., Hene-gariu, O., Hwang, S., Gardner, R., Neutzel, S.,and Sharkis, S.J. (2001). Cell 105, 369–377.
Medici, D., Shore, E.M., Lounev, V.Y., Kaplan, F.S.,Kalluri, R., and Olsen, B.R. (2010). Nat. Med. 16,1400–1406.
Nakagawa, T., Nabeshima, Y., and Yoshida, S.(2007). Dev. Cell 12, 195–206.
Thorel, F., Nepote, V., Avril, I., Kohno, K., Desgraz,R., Chera, S., and Herrera, P.L. (2010). Nature 464,1149–1154.
Zipori, D. (2004). Nat. Rev. Genet. 5, 873–878.
Zipori, D. (2009). Biology of Stem Cells and theMolecular Basis of the Stem State (New York:Humanna Press Inc.).
l 8, January 7, 2011 ª2011 Elsevier Inc. 11
Cell Stem Cell
Forum
In Vitro Fertilization, the Nobel Prize,and Human Embryonic Stem Cells
John Gearhart1,* and Christos Coutifaris2,*1Institute for Regenerative Medicine, University of Pennsylvania, 421 Curie Boulevard, Philadelphia, PA 19104, USA2Division of Reproductive Endocrinology and Infertility, University of Pennsylvania School of Medicine, 3701 Market Street, Philadelphia, PA19104, USA*Correspondence: gearhart@upenn.edu (J.G.), ccoutifaris@obgyn.upenn.edu (C.C.)DOI 10.1016/j.stem.2010.12.015
Robert Edwards was awarded the 2010 Nobel Prize in Physiology or Medicine for the development of humanin vitro fertilization. His work not only provided the means to overcome many forms of infertility, but it alsoenabled research on early stages of human embryos and the derivation of human embryonic stem cells.
It was with great excitement that investi-
gators and clinicians in the field of re-
production received the news that the
2010 Nobel Prize in Physiology or Medi-
cine was awarded to Professor Robert
G. Edwards for his contributions to the
development of human in vitro fertilization
(IVF). With the exception of transfusion
medicine, human IVF and embryo transfer
represents the only other medical inter-
vention that involves the removal of cells
from the body, processing of these cells
in the laboratory, and the eventual re-
introduction of the ‘‘processed’’ cells re-
sulting in a successful therapy of
a medical condition. Infertility, which is
defined as the inability to conceive after
1 year of unprotected intercourse, affects
approximately one in seven couples of
reproductive age in the United States. It
is a major medical and social problem,
and it was not until the development of
clinical human IVF that many diverse
causes of infertility could be successfully
overcome. With the exception of infertility
secondary to anovulation, which was
easily ‘‘cured’’ once ovulation induction
hormonal regimens were developed, no
other fertility treatment has met with the
success of IVF. It is estimated that 2%–
3% of all births in developed countries
are the result of IVF procedures. In addi-
tion, there are strong prospects for
applying this treatment in a cost-effective
way to wider infertility populations. The
births made possible by IVF, now and in
the future, are clear tangible results of
this important basic research. However,
the development of IVF has another
significant impact as well. Edwards’ No-
bel Prize-winning work has also enabled
research that could improve the quality
of life for millions more by providing the
12 Cell Stem Cell 8, January 7, 2011 ª2011 E
basis for deriving human embryonic
stem cells (hESCs), which may be used
to restore tissues lost or damaged
because of disease or injury.
The history of both the research and the
clinical application leading to human IVF
is very instructive, and clear parallels can
be drawn with the modern, growing field
of hESC research. Here, we offer an
abbreviated historical perspective of the
development of human IVF and discuss
how some of the lessons learned might
help inform the current debate over poli-
cies regulating hESC research.
The Path to the Birth of the First IVFBabiesThe very first in vitromanipulation of eggs/
embryos was performed byWalter Heape
(1890), when he transferred in vivo fertil-
ized eggs from one female rabbit to
another and achieved pregnancy and
subsequent delivery of Angora rabbits
similar to the biological parents’ breed. It
is interesting that successful embryo
transfers in other species did not happen
until much later, with rat, sheep, goat,
and mouse pregnancies reported in the
1930s, and eventually cow and pig
embryo transfers in the 1950s (for histor-
ical reviews, see Biggers, 1981; Wolf
and Quigley, 1984). These experiments
all involved in vivo conceptions and
subsequent transfer of the resulting
embryos to a pseudopregnant recipient,
usually of a different breed.
Attempts at IVF also date back to the
late 1800s. Specifically, Schenk attemp-
ted to fertilize rabbit and guinea pig
oocytes in vitro; however, there was no
unequivocal proof that sperm had entered
the eggs. It was not until 1959 when M.C.
Chang, using rabbits, provided unequiv-
lsevier Inc.
ocal proof of successful IVF (Chang,
1959).
Parallel laboratory work refined the
culture techniques for mammalian em-
bryo development in vitro. While the
development of these methods aided
the eventual establishment of clinically
relevant IVF, it could be argued that the
more significant contribution of these
efforts was to uncover molecular mecha-
nisms behind the physiology and cell
biology of oocyte maturation and early
embryo development. Even though
many individuals may have considered
proceeding with human IVF during this
period, it was Robert Edwards who first
put these thoughts into action and
achieved IVF of human eggs that were
obtained from excised ovaries and
matured in vitro prior to fertilization (Ed-
wards et al., 1969). The fertilization effi-
ciency using this approach was extremely
low, largely due to the complexities of
in vitro maturation of the developmentally
arrested oocytes. The subsequent break-
through of retrieving human eggs that
were first matured in vivo and shown to
achieve efficient fertilization and early
development in vitro (Edwards et al.,
1970) was quickly translated into clinical
practice. Despite this promising finding,
when additional attempts were made by
Edwards and his clinical collaborator,
Patrick Steptoe, to obtain multiple mature
human eggs following treatment of
women with ovulation-inducing agents,
pregnancies were not achieved and so
they abandoned this approach.
Finally, in 1977, a mature egg obtained
during a natural cycle was fertilized
in vitro and transferred back to the egg
donor, resulting in the first pregnancy
and the birth of Louise Brown in July of
Cell Stem Cell
Forum
1978 (Steptoe and Edwards, 1978).
Almost concurrently, the Australian team
of Lopata and colleagues also succeeded
using the natural ovulatory cycle, and then
Trounson and theMonash group reported
the use of fertility drugs, ovulatory control-
ling strategies, and delayed insemination
that substantially increased embryo
production and pregnancy success rates
for IVF (see Cohen et al., 2005 for specific
references and amore complete historical
accounting). These major breakthroughs
were then quickly transferred to the UK,
France, Belgium, and the United States.
IVF and the associated technologies
developed with and around it are now
collectively referred to as ‘‘assisted repro-
ductive technologies,’’ or ART.
Technical Developments ContinueDuring the decade following the first IVF
births, progress continued with three
major technical advances that contrib-
uted to innovative treatments and to our
understanding of basic molecular and
cellular processes involved in fertilization
and early development in the human.
The first such advance, establishing
safe cryopreservation techniques, came
in response to the collection of multiple
eggs and embryos (see Cohen et al.,
2005). This method enabled the storage
of excess embryos for the patient’s future
use, thus avoiding further ovarian stimula-
tion and allowing clinicians to restrict the
number of embryos transferred to the
patient on any one occasion in order to
limit high-order multiple births. Cryopres-
ervation techniques made it possible for
couples who did not desire additional
children to donate stored embryos to
other infertile couples or to research. Indi-
rectly, therefore, the combination of IVF
and embryo cryopreservation made the
generation of human embryonic stem
cells possible.
The second advance was the develop-
ment of intracytoplasmic sperm injection
(ICSI), which showed that the injection of
a single sperm into a human oocyte was
sufficient to achieve fertilization, preg-
nancy, and live birth (see Cohen et al.,
2005). This technique not only offered an
alternative to male factor infertility, which
affects approximately one-third of infertile
couples, but also provided clues to under-
standing functional aspects of sperm
physiology and elements of egg activation
and early development.
The third technical advance in the field
was the introduction of blastomere
biopsy, which allowed for the diagnosis
of genetic diseases at the level of the pre-
implantation embryo and also provided
the opportunity to uncover molecular
mechanisms regulating early embryonic
cell differentiation (Handyside et al.,
1990). Clearly, this technology provided
at least the technical means that subse-
quently allowed the development of ap-
proaches to generate human embryonic
stem cells from single blastomeres
without destroying the embryo.
Clinical IVF, hESCs, Science,and SocietyThis brief historical overview clearly de-
monstrates the importance of the devel-
opment of IVF to the birth of the field of
hESC biology. The availability of spare
human embryos generated via IVF,
made available by choice and consent of
the parents, opened the door for their
use in research. As such, the develop-
ment of human IVF and its associated
laboratory methodologies, culture tech-
niques, and other technical aspects
played a critical role in enabling hESC
research and its potential future clinical
applications.
As is observed for many great innova-
tions that impact society, IVF raised its
share of ethical, moral, religious, and
political issues. Among these concerns
were that any children born would not be
normal, that society was poised on a slip-
pery slope that carried the risk of playing
God or would lead to eugenics, baby
farms, human cloning, an explosion in
the world’s population, and so on.
Edwards was also faced with criticism
from some prominent scientists and the
continued need for research funding. For
example, the MRC rejected his applica-
tion to fund his IVF studies (see Johnson
et al., 2010 for a more detailed account).
Yet, with the successful clinical demon-
stration that IVF could overcome infertility
in many patients, the technique became
accepted, widely practiced, and the loud
criticisms diminished. Edwards engaged
the public with his advocacy of IVF and
strongly promoted oversight and regula-
tion of this field (Edwards, 1974), which,
in the UK, eventually resulted in the
passage of the Human Fertilisation and
Embryology Act in 1990. This act
provided oversight and regulation not
Cell Stem Cel
only of IVF but also for human embryo
research. Edwards’s experiences have
provided lessons for those pursuing other
promising yet controversial medical
advances, none more so than the work
IVF has directly enabled: the derivation
of hESCs.
In the 1990s, several laboratories were
pursuing the derivation of hESCs using
procedures that resulted in embryo
destruction. These efforts were under-
taken because investigators recognized
the potential importance of hESCs in
basic research and ultimately as a source
of cells for therapies, as Edwards had
foreseen and promoted (Edwards, 1982).
Indeed, it could be argued that Edwards
himself was the intellectual founder of
hESC research. With the first publication
of hESC derivation in 1998 came a
pronounced vocal opposition that echoed
the objections Edwards experienced in
response to human IVF. In contrast,
however, the hESC debate, which con-
tinues to this day, has been largely
focused on the destruction of embryos
(e.g., that an embryo is a human being
or a nascent human being). Given that
typical IVF practices give rise to embryos
that are not used for reproduction, this
technique has always been faced with
the contentious issue of the frequent dis-
carding of human embryos. However,
this point had not been widely debated
until after the derivation of hESCs brought
the practice more visibly into the public
domain. Lewis Wolpert (Wolpert, 2001),
and others, have pointed out repeatedly
that there is no ethical difference between
IVF and deriving hESCs in that both prac-
tices require the creation and destruction
of embryos. With IVF, a significant
number of embryos are discarded either
because they do not meet the criteria for
uterine transfer or because patients have
completed their treatments and no longer
have need for their cryopreserved em-
bryos. Although the embryos were pro-
duced with the intent of reproduction,
patients have been given the opportunity
to provide the embryos for research,
including for hESC derivation. IVF is per-
formed regularly in countries where
hESC research (or the derivation of
hESC lines) is banned. One could ques-
tion whether it is rational to support clin-
ical IVF and yet oppose ESC derivation.
(For the legal status of hESC research in
countries and in U.S. states, check the
l 8, January 7, 2011 ª2011 Elsevier Inc. 13
Cell Stem Cell
Forum
ISSCR database at www.isscr.org/public/
regions).
It is important to acknowledge the
invaluable contributions that infertile cou-
ples, particularly the women subjected to
medical treatments, failures, and surgical
procedures in the hope of achieving a
pregnancy, have made to the develop-
ment of clinical human IVF. In a sense,
they should share, in spirit, the Nobel
prize with Robert Edwards. Furthermore,
couples that have provided embryos for
research purposes are largely unsung
heroes who have enabled the develop-
ment of the hESC field which, it could be
argued, holds an even greater promise
than clinical IVF in terms of potential
impact on basic research and therapeutic
development.
hESC Research, Funding,Regulation, and OversightThe pace of stem cell research and inno-
vation and the utilization of the knowledge
gained from the study of ESCs will
continue to change the strategies em-
ployed for developing clinical therapies.
At present, there is still a need for
research with hESCs and some of the
newer developments, such as induced
pluripotency, which is assumed widely
to replace hESCs, are not without their
own ethical and moral issues. In the
U.S., remarkable progress has been
made despite numerous political obsta-
cles, thanks mainly to dedicated investi-
gators and funding from philanthropic
donors, supportive states, and disease-
and patient-based organizations. The
FDA has now approved two clinical trials
that will transplant hESC-derived cells:
oligodendrocyte progenitors for spinal
cord injury and retinal pigmented ep-
ithelial cells for Stargardt’s macular
dystrophy. Should these and other
upcoming trials prove successful, it
seems likely that support for the clinical
utility of hESCs will follow. Indeed, if the
parallels with IVF’s journey into main-
stream clinical practice continue, a thera-
peutic success for hESCs may well
overshadow any lingering objections to
ongoing basic research efforts and tech-
nological development that remain essen-
tial to the growth of this field.
For the past 40 years, the U.S. govern-
ment has not followed through on recom-
mendations of committees that have been
empanelled to propose scientifically
14 Cell Stem Cell 8, January 7, 2011 ª2011 E
sound, ethical, and regulated policies,
including funding, on human embryo
research. This lack of progress has led
to the development of hESC research
guidelines by the National Academy of
Sciences/National Research Council and
the International Society for Stem Cell
Research for voluntary adherence. The
NIH policy, more limited than the non-
public organizations on what research is
eligible to receive funding, has evolved
over time as well. For ART in the U.S.,
there are voluntary guidelines that have
been developed by the American Society
for Reproductive Medicine, the clinical
field’s professional society.
In our opinion, the lack of a rational,
widely acceptable policy on the use of
human embryos in research has compro-
mised hESC research in the U.S. Unlike
IVF, whose wide acceptance came on
the heels of a relatively rapid and highly
visible demonstration of clinical success,
hESC research will take years to ‘‘trans-
late’’ into routine clinical use. Opponents
of hESC research are quick to point out
that no one has been cured using hESCs
even 12 years after their derivation.
Recent years have seen an escalation of
much needed federal funding for hESC
research, but this support is now jeopar-
dized by a legal challenge on the use of
federal funds for human embryo
research. The lawsuit before the U.S.
District Court in Washington D.C. illus-
trates the vulnerability of current policy
reflecting differences of opinions on the
interpretation of a law. If patients are to
benefit from the impressive progress
made in ESC research over the past
decade, it is clear that federal funding
and legislative action are both required.
Congress must define what is eligible
for federal funding, address the current
law and provide for authorization of
expenditures of funds for human embryo
research. Either the NIH (or a public body
established for this purpose) could
resolve the complex issues surrounding
the use of embryos in research. Only
with transparent public deliberations
among scientific experts, social scien-
tists, and legislators can much needed
guidance emerge. This will not be easy
to achieve, given the existing perceptions
of ‘‘medical naivete’’ and political consid-
erations and pressures. Nevertheless, it
is imperative to achieve an outcome
that would permit and support the fund-
lsevier Inc.
ing of sound science involving a legitimate
use of human embryos in research,
particularly the use of existing embryos
that couples have no further plans to
use and do not wish to donate to other
infertile couples. The legislative process
may prove difficult politically but each of
us must realize our responsibility to
pursue every opportunity to alleviate the
suffering and to improve the quality of
life for those citizens in desperate need
of therapies. It is not surprising that in
our pluralistic society there are wide
differences of opinions on the moral and
ethical values of the earliest stages of
human development. We must accept
that there are compelling views and
sound science for a legitimate use of
embryos in research that could improve
the quality of life for many, as well as
save lives. It is accurate to say that
a majority of Americans have come to
a consensus that we should pursue
hESC research with proper guidelines,
oversight, and government funding.
Legislation should reflect this consensus.
Human IVF, despite initial resistance by
society and, indeed, from within the
medical community, has proven to be
a key treatment of infertility. This practice
is now firmly established in clinical medi-
cine, although additional improvements
continue to be needed and sought.
Although Robert Edwards was awarded
the Nobel Prize for his scientific con-
tributions in the development of IVF tech-
nologies, his vision extended beyond
treatments for infertility, and included
embryonic stem cells. As a society, we
must now manage this newer offspring
of IVF with policies that will enable the
pursuit of human embryo research that
will serve to benefit all people.
REFERENCES
Biggers, J.D. (1981). N. Engl. J. Med. 304, 336–342.
Chang, M.C. (1959). Nature 184, 466–467.
Cohen, J., Trounson, A., Dawson, K., Jones, H.,Hazekamp, J., Nygren, K.G., and Hamberger, L.(2005). Hum. Reprod. Update 11, 439–459.
Edwards, R.G. (1974). Q. Rev. Biol. 49, 3–26.
Edwards, R.G. (1982). The case for studyinghuman embryos and their constituent tissuesin vitro. In Human Conception In Vitro, R.G.Edwards and J.M. Purdy, eds. (London: AcademicPress), pp. 371–387.
Cell Stem Cell
Forum
Edwards, R.G., Bavister, B.D., and Steptoe, P.C.(1969). Nature 221, 632–635.
Edwards, R.G., Steptoe, P.C., and Purdy, J.M.(1970). Nature 227, 1307–1308.
Handyside, A.H., Kontogianni, E.H., Hardy, K., andWinston, R.M. (1990). Nature 344, 768–770.
Heape, W. (1890). Proc. R. Soc. Lond. 48, 457–458.
Johnson, M.H., Franklin, S.B., Cottingham,M., andHopwood, N. (2010). Hum. Reprod. 25, 2157–2174.
Steptoe, P.C., and Edwards, R.G. (1978). Lancet 2,366.
Cell Stem Cel
Wolf, D.P., and Quigley, M.M. (1984). Historicalbackground and essentials for a program inin vitro fertilization and embryo transfer. Chapter 1,In Human In Vitro Fertilization and Embryo Trans-fer, D.P. Wolf and M.M. Quigley, eds. (New York:Plenum Press), pp. 1–9.
Wolpert, L. (2001). Nature 413, 107–108.
l 8, January 7, 2011 ª2011 Elsevier Inc. 15
Cell Stem Cell
Review
DNA-Damage Response in Tissue-Specificand Cancer Stem Cells
Cedric Blanpain,1,* Mary Mohrin,2 Panagiota A. Sotiropoulou,1 and Emmanuelle Passegue2,*1Universite Libre de Bruxelles, IRIBHM, B1070 Bruxelles, Belgium2The Eli and Edythe Broad Center for Regenerative Medicine and Stem Cell Research, Department of Medicine, Division ofHematology/Oncology, University of California San Francisco, San Francisco, CA 94143, USA*Correspondence: cedric.blanpain@ulb.ac.be (C.B.), passeguee@stemcell.ucsf.edu (E.P.)DOI 10.1016/j.stem.2010.12.012
Recent studies have shown that tissue-specific stem cells (SCs) found throughout the body respond differ-entially to DNA damage. In this review, we will discuss how different SC populations sense and functionallyrespond to DNA damage, identify various common and distinct mechanisms utilized by tissue-specific SCsto address DNA damage, and describe how these mechanisms can impact SC genomic integrity by poten-tially promoting aging, tissue atrophy, and/or cancer development. Finally, we will discuss how similar mech-anisms operate in cancer stem cells (CSCs) and can mediate resistance to chemo- and radiotherapy.
Stem cells (SCs) are often referred to as the mother of all cells,
meaning they sit at the apex of a cellular hierarchy and, upon
differentiation, give rise to all the mature cells of a tissue (Rossi
et al., 2008). More specifically, SCs are described as having
the unique capacity to self-renew, in order to establish and
replenish the SC pool, and also to differentiate, thereby gener-
ating progeny that carry out specific tissue functions. SCs are
essential for specification and morphogenesis of tissues during
embryonic development (organogenesis) and for the mainte-
nance and repair of adult tissues throughout life by replacing
cells lost during normal tissue turnover (homeostasis) or after
injury. Although tissue-specific SCs are found in many highly
regenerative organs, such as blood, skin, and the digestive tract,
they are also found in nonrenewing organs such as muscle,
where they allow repair after tissue damage.
Like every other cell in the body, SCsmust constantly contend
with genotoxic insults arising from both endogenous chemical
reactions, such as reactive oxygen species (ROS) generated
by cellular metabolism, and exogenous insults coming from their
surrounding environment (Sancar et al., 2004). It has been esti-
mated that every cell undergoes about 100,000 spontaneous
DNA lesions per day (Lindahl, 1993). As SCs ensure the lifetime
maintenance of a given tissue, anymisrepair of DNA damage can
be transmitted to their differentiated daughter cells, thereby
compromising tissue integrity and function. Consequently,
mutations that diminish the renewal and/or differentiation poten-
tial of SCs can result in tissue atrophy and aging phenotypes,
whereas mutations providing a selective advantage to the
mutated cells can lead to cancer development (Rossi et al.,
2008).
As such, a delicate balance must be struck to prevent exhaus-
tion and transformation of the SC pool while maintaining the
ability of SCs to preserve homeostasis and to respond to injury
when necessary. To fulfill these demands, the numbers of SCs
and their functional quality must be strictly controlled through
a balance of cell-fate decisions (self-renewal, differentiation,
migration, or death), which are mediated by a complex network
of cell-intrinsic regulation and environmental cues (He et al.,
2009; Weissman, 2000). Specific protective mechanisms also
16 Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc.
ensure that SC genomic integrity is well preserved and include
localization to a specific microenvironment, resistance to
apoptosis, limitation of ROS production, and maintenance in
a quiescent state (Orford and Scadden, 2008; Rossi et al.,
2008). Altogether, these attributes of SCs ensure tissue mainte-
nance and function throughout the lifetime of an organism, while
limiting atrophy and cancer development.
DNA-Damage ResponseAll living cells, including tissue-specific SCs, must constantly
contend with DNA damage (Sancar et al., 2004) (Figure 1). Due
to its chemical structure, DNA is particularly sensitive to sponta-
neous hydrolysis reactions which create abasic sites and base
deamination. Furthermore, ongoing cellular metabolism gener-
ates ROS and their highly reactive intermediate metabolites,
which can create 8-oxoguanine lesions in DNA as well as
a variety of base oxidations and DNA strand breaks that are all
highly mutagenic and can lead to genomic instability. DNA is
also constantly assaulted by mutagens present in the external
environment. UV light from the sun, as well as various chemical
reagents, can react with DNA and induce nucleotide chemical
modifications. Ionizing radiations (IR) generated by the cosmos,
X-rays, and exposure to radioactive substances, as well as treat-
ment with certain chemotherapeutic drugs, can induce base
modifications, interstrand crosslinks, single- and double-strand
breaks (DSBs), which can all lead to genomic instability.
Consistent with the wide diversity of potential DNA lesions,
eukaryotic cells exhibit many highly conserved DNA repair
mechanisms that can recognize and repair different types of
DNA damage with varying fidelity and mutagenic consequences
(Lombard et al., 2005) (Figure 1). For instance, base modifica-
tions induced by spontaneous chemical reactions and ROS-
mediated DNA lesions are repaired by base excision repair
(BER), whereas nucleotide modifications induced by chemicals
and UV light are repaired by the nucleotide excision repair
(NER) pathway. The pathways that mediate the repair of DSBs
vary depending on the cell-cycle status of the damaged cells.
During the G0/G1 phase, DSBs are repaired by the nonhomolo-
gous end-joining (NHEJ) pathway, while, during the S-G2/M
Abasic sitesSingle strand breaks8-oxoguanine lesions
Bulky adductsPyrimidine dimers
Double strand breaksSingle strand breaks
Intrastrand crosslinksInterstrand crosslinks
Bases mismatchInsertionsDeletions
Base ExcisionRepair (BER)
NucleotideExcision
Repair (NER)
HomologousRecombination (HR)
MismatchRepair (MMR)
Oxygen radicalsHydrolysis
Alkylating agents
UV-lightchemicals
Ionizing radiationX-rays
Anti-tumor drugs
Replication errors
C
8-oxo8-oxoG
T
CCC
TT
AG
T C
A
Non Homologous End Joining (NHEJ)
FIDELITY
+
++++
++++++
++++++
++++++
DNA REPAIRPATHWAYS
DNA DAMAGINGAGENTS
DNALESIONS
Figure 1. DNA-Repair Pathways in Mammalian CellsEach type of DNA assault results in a different type of lesion, which can be repaired with different fidelity by distinct and highly specialized repair pathways.
Cell Stem Cell
Review
phase, these lesions are repaired by the homologous recombi-
nation (HR) pathway. These two modes of DNA repair are not
equally faithful. HR is an error-free DNA repair mechanism due
to the use of the other intact strand as a template, while NHEJ
is an error-prone repair mechanism, which may result in small
deletions, insertions, nucleotide changes, or chromosomal
translocations due to the absence of an intact template for
repair. Lastly, replication errors leading to insertion, deletion,
and base misincorporation resulting in base mispairing are cor-
rected by the mismatch repair (MMR) pathway.
Irrespective of the type of lesion and the repair mechanism,
DNA damage is rapidly sensed and activates evolutionarily
conserved signaling pathways, known collectively as the DNA-
damage response (DDR), whose components can be separated
into four functional groups: damage sensors, signal transducers,
repair effectors, and arrest or death effectors (Sancar et al.,
2004) (Figure 2). Ultimately, activation of DDR leads to the
phosphorylation and stabilization of p53, inducing its nuclear
accumulation and upregulation of its target genes (d’Adda di
Fagagna, 2008). Depending upon the extent of DNA damage,
the type of cell undergoing DNA damage, the rapidity of DNA
repair, the stage of the cell cycle, the strength and the duration
of p53 activation, and the genes transactivated by p53, cells
can either undergo transient cell-cycle arrest (through induction
of the cyclin-dependant kinase inhibitor p21), programmed cell
death (through induction of the pro-apototic bcl2 gene family
members bax, puma and noxa), or senescence (through induc-
tion of the cyclin-dependant kinase inhibitor p16/Ink4a and the
tumor suppressor gene p19/ARF).
Diversity of DNA Repair Mechanisms in Tissue-SpecificStem CellsThe critical role of the different DNA repair mechanisms for over-
all tissue integrity and function is well illustrated by the severe
clinical consequences observed in both humans and mice for
mutations in genes regulating these pathways (Hakem, 2008).
The involvement of tissue-specific SCs in mediating such symp-
toms and the role of the diverse DNA-damage recognition and
DNA-repair mechanisms in maintaining tissue-specific SC func-
tion is now starting to emerge (Kenyon and Gerson, 2007).
Defects in DSB recognition machinery lead to premature
aging, neurodegeneration, and increased cancer susceptibility.
ATM (ataxia-telengiectasia mutated), ATR (ATM and Rad3
related), and DNA-PKs are DNA-damage-sensing protein
kinases that, through a series of phosphorylation events, signal
the presence of DNA lesions and initiate DNA repair or cell-cycle
Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc. 17
Sensors
Effectors
Transducers
Mediators
Cellular
Outcome
Brca1
p53
p21 BAX
NOXA
PUMA
p16 p19
Cell cycle
arrest SenescenceApoptosis
DNA-PK ATM ATR
H2AX
MRN ATRIP
H2AX
MRN
ATM
MRN
H2AX
53BP1
KU70/80
DNA-PKATR
CHK2CHK1
DNA repair
PARP
DNA damage
Figure 2. DNA-Damage ResponsePathwaysUpon DNA damage, distinct factors detect, trans-mit, and amplify the DNA-damage signal. DNAdouble-strand breaks can be repaired by homolo-gous recombination (mediated among otherfactors by the MRN complex, ATM, and Brca1)or by nonhomologous end-joining (in which theKu70/Ku80/DNA-PKcs complex plays a majorrole). This DNA-damage response convergesupon p53 which, depending on the target genesactivated, regulates different cellular outcomes.
Cell Stem Cell
Review
arrest (Figure 2). Patients with mutations in ATM present blood
vessel abnormalities, cerebelar degeneration, immunodefi-
ciency, and increased risk of cancers (Hoeijmakers, 2009).
Mice lacking Atm, like ATM patients, are extremely sensitive to
IR exposure and have decreased somatic growth, neurological
abnormalities, decreased T cell numbers, and exhibit premature
hair graying and infertility (Barlow et al., 1996). Many of these
phenotypes can be linked to defects in SC function, which high-
lights the critical role of this DDR component for the survival and
preservation of various SC compartments. Atm-deficient hema-
topoietic SCs (HSCs) harbor increased ROS levels and display
an overall decrease in number and function over time, leading
to eventual hematopoietic failure (Ito et al., 2004, 2006).Atm defi-
ciency also sensitizes mice to IR-induced prematuremelanocyte
SC differentiation, resulting in hair graying (Inomata et al., 2009).
Germ cell development is also altered in Atm-deficient mice, and
mutant animals experience a progressive loss in germ SCs
(spermatogonia) and become infertile (Takubo et al., 2008).
Mutations in ATR also cause developmental defects in mice
(pregastrulation lethality) and humans (Seckel syndrome)
(Hakem, 2008; Hoeijmakers, 2009; Seita et al., 2010). Condi-
tional deletion of Atr in adult mice leads to the rapid appearance
18 Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc.
of age-related phenotypes, such as hair
graying, alopecia, kyphosis, osteopo-
rosis, thymic involution, and fibrosis,
which are associated with SC defects
and exhaustion of tissue renewal and
homeostatic capacity (Brown and Balti-
more, 2000; Ruzankina et al., 2007).
The MRE11, RAD50, and NBS1 (MRN)
complex senses DSBs, unwinds the
damaged region of DNA, serves as part
of the repair scaffolding, and induces
downstream signaling including ATM
activation (Figure 2). Deletion of any
component of the MRN complex results
in embryonic lethality in mice (Hakem,
2008). However,micebearing ahypomor-
phic Rad50k22m mutation are viable but
die around 2.5 months from of B cell
lymphoma or bone marrow failure due,
in part, to p53-dependent DDR-mediated
apoptosis and loss of HSC function
(Bender et al., 2002). Moreover, muta-
tions in BRCA1 and BRCA2, two DSB
mediators that trigger DNA repair through
the HR pathway (Figure 2), lead to a major increase in the risk of
developing breast and ovarian cancers in women, which, at least
in the breast, has recently been linked to the accumulation of
genetically unstable mammary SCs (Liu et al., 2008).
While no spontaneous mutations in NHEJ pathway compo-
nents have been reported so far in human syndromes associated
with premature aging or increased risk of cancers, the inactiva-
tion of various NHEJ genes in mice has demonstrated their
essential function in lymphocyte development and prevention
of lymphoma. The core components of the NHEJ repair pathway
include the end-binding and end-processing proteins Ku70,
Ku80, DNA-PKcs, and Artemis, as well as the ligation complexes
XRCC4, LigIV, and Cerrunos (Lombard et al., 2005). As NHEJ is
critical for V(D)J recombination during lymphocyte maturation,
many of the mutant mouse models deficient in particular NHEJ
components exhibit arrested lymphoid development. Mice
carrying a Lig4y288c hypomorphic mutation also display growth
retardation, immunodeficiency, and pancytopenia associated
with severe HSC defects (Kenyon and Gerson, 2007; Nijnik
et al., 2007). Mice lacking the end-binding and end-processing
components of NHEJ, Ku70, and Ku80 have stress-induced
HSC self-renewal defects associated with poor transplantability,
Cell Stem Cell
Review
increased apoptosis, decreased proliferation, and impaired
lineage differentiation (Kenyon and Gerson, 2007; Rossi et al.,
2007).
Mutations in NER pathway components induce human
syndromes known as Xeroderma Pigmentosum (XP), Cockayne
syndrome (CS), and Trichothiodistrophy (TTD), which are char-
acterized by premature aging, neurodegeneration, and extreme
photosensitivity, especially in XP syndromes (Hoeijmakers,
2009). XP patients often completely lack NER repair activity
and have increased incidence of skin cancer, while CS and
TTD patients have defects in transcription-coupled repair, which
has little mutagenic effect because it only deals with lesions in
the transcribed strand. Mice expressing XPDTTD, a mutated
formof an essential NER component, have decreasedHSC func-
tionwith reduced self-renewal potential and increased apoptosis
levels (Rossi et al., 2007). Mice deficient in Ercc1, a component
of both NER and intrastrand crosslink (ICL) repair, die within
4 weeks of birth, have multilineage hematopoietic cytopenia
due to progenitor depletion, HSC senescence, and a defective
response to DNA crosslinking by mitomycin C (Hasty et al.,
2003; Prasher et al., 2005).
Mutations in MMR pathway components induce hereditary
nonpolyposis human colorectal cancer known as Lynch
syndrome, which presents with about an 80% lifetime risk of
developing colorectal cancers as well as other malignancies
(Hoeijmakers, 2009). Mice mutant for genes important for the
MMR pathway, including Msh2 and Mlh1, also display higher
frequencies of hematological, skin, and gastrointestinal tumors,
consistent with a critical role of the MMR in preventing accumu-
lations of oncogenic mutations (Hakem, 2008). In addition, mice
lacking Msh2 exhibit defective HSC activity, with enhanced
microsatellite instability observed in their progeny (Reese et al.,
2003).
Other human conditions associated with defects in DNA-
damage recognition and repair pathways include Fanconi’s
Anemia (genetic defects in the FANC family of proteins), Bloom’s
or Werner’s syndromes (both caused by mutations in DNA heli-
cases), and a range of diseases associated with telomerase
dysfunction and telomere instability (Kenyon and Gerson,
2007). These diseases are not specifically reviewed here, but
their complex pathologies involve defects in various tissue-
specific SCs.
DNA-Damage Response in Tissue-Specific SCsWhile tissue-specific SCs share the same purpose of maintain-
ing organ functionality, recent studies have shown that the
mechanisms of their responses to DNA damage, the outcome
of their DDR, and the consequences of DNA repair for their
genomic stability vary greatly between tissues.
Hematopoietic SCs
The hematopoietic (blood) system is one of the best-studied
adult tissues in terms of its hierarchical development, in that all
blood cell lineages derive from a small number of quiescent
HSCs via a highly proliferative amplifying progenitor compart-
ment (Orkin and Zon, 2008). Being a highly regenerative
compartment, it is also one of the most radiosensitive tissues
in the body (<4 Gy), and one of the first organ systems to fail after
total body irradiation. IR exposure differentially affects hemato-
poietic cells depending on their state of maturity, with HSCs
being more radioresistant than their downstream progeny
(Meijne et al., 1991). By comparing thewayHSCs and their differ-
entiated progeny respond to low doses of IR (2 to 3 Gy), recent
work has begun to clarify the ways in which HSCs at different
stages of ontogeny deal with DNA damage and the mutagenic
consequences of different DNA repair mechanisms in this
tissue-specific SC population (Figure 3A).
HSCs are specified in the aorta-gonad-mesonephros (AGM)
region of the developing fetus, are actively expanded in several
anatomic locations, including the liver and placenta, during fetal
development, and are finally seeded in the bone marrow cavity
during late embryogenesis. In the bone marrow, HSCs progres-
sively mature after birth to become the quiescent adult HSCs
that are maintained during the lifetime of the organism. Fetal
and adult HSCs differ in many aspects of their biological regula-
tion, including cell-cycle status and transcriptional control (Orkin
and Zon, 2008). Using human umbilical cord blood (CB)-derived
HSCs, which are highly proliferative, circulating cells that are still
considered to be of fetal origin, Milyavsky and colleagues found
that irradiated (3 Gy) CB-derived HSCs had a slower rate of DSB
repair than more mature progenitors and increased levels of
apoptosis mediated in part through the ASPP1 protein, which
could be reversed if p53 expression was silenced or bcl2 expres-
sion was enhanced (Milyavsky et al., 2010). Upon primary trans-
plantation, irradiated CB-derived HSCs could not successfully
engraft into immunodeficient mice. In contrast, irradiated cells
with disabled p53 or bcl2 overexpression could be serially trans-
planted, albeit with decreased efficiency compared to nonirradi-
ated normal cells. In this context, transplanted CB-derived HSCs
with disabled p53 reconstituted even less well than cells with
bcl2 overexpression, and their progeny harbored high levels of
DSBs that were not observed in the progeny of bcl2 overex-
pressing cells. This study emphasizes the role of p53-mediated
DDR and the Bcl2 family of prosurvival genes in HSC function
(Asai et al., 2010; Seita et al., 2010; Weissman, 2000), and indi-
cates that the main outcome of the DDR in fetal HSCs is induc-
tion of apoptosis and overt cell elimination (Figure 3A). On the
other hand, using adult mouse HSCs that are kept mostly quies-
cent within the bone marrow cavity, Mohrin and colleagues
showed a very different response to irradiation, with overt cell
survival and DNA repair being the main outcomes of the DDR
(Mohrin et al., 2010). Adult HSCs, either quiescent or induced
to proliferate by cytokine pretreatment, engage specialized
response mechanisms that protect them from low doses of IR
(2 Gy). In quiescent HSCs, these mechanisms include enhanced
prosurvival gene expression (bcl2, bcl-xl, mcl1, a1), which
inhibits cell death induced by p53 proapototic genes (bax,
noxa, puma), likely allowing p53-mediated induction of p21 to
engage a transient growth-arrest response and to permit DNA
repair. While the exact mechanism of the survival response in
proliferating HSCs is less clear, they were found to be as radio-
resistant as quiescent HSCs (Mohrin et al., 2010). Dictated by
their cell-cycle status, proliferating HSCs use the high-fidelity
HR pathway to repair DSBs, while quiescent HSCs employ the
error-prone NHEJ pathway. Irradiated quiescent HSCs display
high levels of chromosomal abnormalities when compared to
proliferating HSCs, and their progeny show persistent genomic
instability associated with misrepaired DNA and engraftment
defects in secondary recipient mice. Since NHEJ appears to
Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc. 19
A
Quiescent adult mousebone marrow
hematopoietic SCs
Proliferating humanumbilical cord blood hematopoietic SCs
DNA repair
NHEJ
Mcl1Bcl-xl
Bcl2, a1
Celldeath
SC Survival
p21Cell cycle
arrest
DNArepair
Bcl2
Celldeath
SC depletion
DNA damage
ASSP1
Genetic instability
p53 p53
Quiescent and proliferating
adult mouse hair follicle bulge SCs
DNA repair
NHEJDNA-PK
Bcl2
Celldeath
SC Survival
p21
Cell cyclearrest
Geneticinstability
DNA damage
?
p53
B
Figure 3. DNA-Damage Response in Hematopoietic and Hair Follicle Bulge Stem Cells(A) Human umbilical cord blood-derived HSCs and mouse bone marrow-derived HSCs exhibit opposite outcomes following irradiation-induced DNA damage,with different consequences for their overall maintenance and genomic integrity.(B) Upon irradiation, mouse hair follicle bulge stem cells exhibit transient p53 activation due, in part, to high levels of DNA-PK-mediated NHEJ repair and higherBcl2 expression that block apoptosis, resulting in enhanced survival.
Cell Stem Cell
Review
be the initial andmost commonly used DNA repair mechanism in
quiescent HSCs, these results help explain why most mouse
models lacking functional components of DSB recognition and
repair pathways undergo hematopoietic failure upon genotoxic
stress (Hakem, 2008). Moreover, this study indicates that while
adult HSCs, in contrast to fetal HSCs, may survive DNA-
damaging insults, they do not emerge unscathed (Figure 3A),
which might have direct implications for aging and cancer devel-
opment. It may also explain why cancer patients treated with
radiotherapy or chemotherapy may develop leukemias and
lymphomas (blood cancer) or myelodysplasias (bone marrow
failure) because the use of error-prone DNA repair in quiescent
HSCs may be at the heart of these dangerous side effects of
cancer treatment.
Taken together, these two studies (Milyavsky et al., 2010;
Mohrin et al., 2010) unveil some striking differences in the
outcome of irradiation-induced DDR in HSCs from different
species and at different developmental stages. While it is
possible that different organisms with vastly different lifespans
have evolved distinct strategies to cope with DNA damage, it
is tempting to speculate that these differences reflect an adapta-
tion in the stress responsemechanisms used by HSCs at distinct
stages of ontogeny to ensure optimal function of the blood
system. During embryogenesis and until birth, the goal is to
expand the SC population while protecting its genomic integrity
in order to establish a pool of pristine HSCs that will ensure blood
homeostasis for the lifetime of the organism. In this context, the
efficient elimination of irradiated human CB-derived HSCs
described by Milyavsky and colleagues fulfill this demand by
eliminating damaged fetal HSCs that could be detrimental to
the organism and its reproductive purpose. Conversely, in
20 Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc.
adults, the main function of the HSC compartment is to preserve
blood homeostasis and to quickly respond to hematopoietic
needs (blood loss, infection, etc.). The fact that adult HSCs
reside in hypoxic niches in the BM cavity and are mostly kept
in a quiescent phase of the cell cycle contribute to their overall
maintenance (self-renewal) and protect their genomic integrity
(fitness) by minimizing DNA damage associated with ROS
production, cellular respiration, and cell division (Orford and
Scadden, 2008; Rossi et al., 2007). In this context, the survival
and efficient DNA repair of irradiated mouse adult HSCs
described by Mohrin and colleagues fulfills the same purpose
by protecting the most important cells of the tissue. Since both
quiescent and proliferating mouse adult HSCs show similar
radioresistance, it is likely that the radiosensitivity displayed by
human CB-derived HSCs reflect cell-intrinsic differences in
transcriptional programs or chromatin states between HSCs at
various stages of development. Additional investigations are
clearly needed to fully understand the mechanisms underlying
these differences in DDR outcomes between fetal and adult
HSCs.
However, the short-term survival strategy used by adult HSCs
likely comes at a cost for their long-term genomic integrity. While
quiescence is one of the very mechanisms that protects adult
HSC function, it also renders damaged HSCs intrinsically vulner-
able to mutagenesis because it forces them to use the error-
prone NHEJ pathway to repair DSBs, thereby increasing the
risk of creating mutations in this self-renewing population. In
fact, the accrual of chromosomal translocations resulting from
unfaithful DNA repair following DSBs is a hallmark of human
blood malignancies (Look, 1997). Such accumulation over time
of NHEJ-mediated mutations may hinder cellular performance
Cell Stem Cell
Review
and could be a major contributor to the loss-of-function occur-
ring with age in the HSC compartment and to the development
of age-related hematological disorders (Rossi et al., 2007).
Epidermal SCs
The skin epidermis is composed by the juxtaposition of themany
pilosebaceous units consisting of a hair follicle, its associated
sebaceous gland, and its surrounding interfollicular epidermis.
Different classes of SCs ensure homeostasis of the skin
epidermis (Blanpain and Fuchs, 2009). Multipotent hair follicle
bulge SCs (BSCs) contribute to the cyclic regeneration of the
hair follicle and to the repair of the interfollicular epidermis
following wounding. In the absence of injury, the interfollicular
epidermis can self-renew independently of BSCs through the
presence of unipotent progenitors scattered throughout the
basal region of the epidermis. Specialized SCs and progenitor
cells are also found in the infundibulum and sebaceous glands
(Blanpain and Fuchs, 2009).
Since the epidermis serves as a barrier between the body and
the external environment, it is constantly assaulted by genotoxic
stress such as UV irradiation. As discussed earlier, UV radiation
causes the formation of thymidine dimers, (6-4) pyrimidine
photoproducts, and ROS-induced DNA lesions that are repaired
by the NER, NHEJ, or HR pathways, depending on the type of
damage and the state of the cell cycle. Upon UV irradiation,
basal epidermal cells exhibit sustained p53 activation compared
to the more differentiated suprabasal cells (Finlan et al., 2006).
Following chronic administration of UV radiation, slow-cycling
SCs and progenitor cells of the infundibulum and sebaceous
glands also retain UV-induced photoproducts longer than
more differentiated cells of the epidermis, suggesting a decrease
in the repair activity of these cells (Nijhof et al., 2007). Recently,
Nrf2 has been shown to regulate the expression of critical
regulators of oxidative stress (such as several enzymes of the
glutathione metabolism) and to protect the epidermis from UV-
induced apoptosis. The gradient of apoptosis levels observed
between basal (high) and suprabasal (low) cells following UV irra-
diation is inversely correlated with Nrf2 expression. Surprisingly,
while Nrf2 overexpression protects basal cells from UV induced
apoptosis, it does not decrease the proportion of cells that
harbor thymidine dimers. In addition, suprabasal expression of
Nrf2 offers some protection from UV-induced apoptosis to basal
cells through a paracrine mechanism (Schafer et al., 2010).
These data indicate that proliferative cells of the interfollicular
epidermis are more sensitive to UV-mediated apoptosis relative
to their more committed progeny.
While the skin epidermis is more radioresistant than the blood
system, acute administration of more than 5 Gy results in severe
skin reactions consisting of inflammation (erythema) and loss of
differentiated skin layers (desquamation) that rapidly appear
following IR, whereas hair loss and chronic ulcerations appear
with a delay of 2 to 3 weeks after IR administration. The sensi-
tivity of the epidermis to IR is also illustrated by the common
side effects of radiotherapy, which include acute and chronic
dermatitis and an increased incidence of skin cancer (Gold-
schmidt and Sherwin, 1980). While the field is still in search of
specific cell-surface markers that will allow high purity isolation
of interfollicular epidermal progenitors, a combination of
markers, including a6 integrin and CD71, have been used to
enrich SCs from the mouse and human interfollicular epidermis
(Li et al., 1998; Tani et al., 2000). Following exposure to low
doses of IR, rapidly cycling human epidermal progenitor cells
(a6Hhi/CD71+) undergo apoptosis and display decreased
in vitro colony forming efficiency, whereas slow-cycling human
epidermal SCs (a6H/CD71�) were resistant to IR-induced cell
death (Rachidi et al., 2007). The enhanced survival of human
epidermal SCs upon IR exposure has been linked to a higher
secretion of FGF2 following DNA damage, which increases
DNA repair activity in epidermal SC by autocrine/paracrine
mechanisms (Harfouche et al., 2010). While these studies have
been performed ex vivo, Sotiropoulou and colleagues have
recently investigated how epidermal cells respond to DNA
damage within their native niche and showed that multipotent
hair follicle BSCs, like HSCs, are more resistant to DNA-
damage-induced cell death compared to the other cells of the
epidermis (Sotiropoulou et al., 2010). At least two important
mechanisms contribute to the higher resistance of BSCs to
IR-mediated DNA damage (Figure 3B), both which are indepen-
dent of the relative quiescence of these cells and of the induction
of premature senescence. First, BSCs express higher levels of
the antiapoptotic protein Bcl2, and the proportion of BSCs
undergoing apoptosis is increased in bcl2 null mice, demon-
strating that similar to HSCs, a higher expression of prosurvival
factors contributes to the resistance of BSCs to apoptosis. The
other contributing mechanism is the transient nature of DDR
activation in BSCs. Soon after IR exposure, p53 is expressed
in the nuclei of almost all epidermal cells, including BSCs, and
is required for DNA-damage-induced cell death in the epidermis
(Botchkarev et al., 2000; Song and Lambert, 1999; Sotiropoulou
et al., 2010). However, unlike other cells of the epidermis, the
number of BSCs expressing p53 is greatly decreased by 24 hr
following irradiation, and mutant mice exhibiting sustained
expression of p53 show increased IR-induced apoptosis in
BSCs. This indicates that the short duration of IR-mediated
p53 activation promotes BSC survival following DNA damage.
Interestingly, BSCs also display accelerated DNA repair and
enhancedNHEJ repair activity. In SCIDmice, which have amuta-
tion in DNA-PK and thus exhibit decreased NHEJ activity, BSCs
are radiosensitive, suggesting that accelerated NHEJ-mediated
DSB repair contributes to their protection against IR exposure.
The importance of DDR in BSCs is also illustrated by the SC
exhaustion and progressive alopecia that occurs in mice where
Atr has been deleted in hair follicle BSCs and their progeny
(Ruzankina et al., 2007).
Because NHEJ is an error-prone DNA repair mechanism, the
higher resistance of BSCs to DNA-damage-induced apoptosis
and the accelerated NHEJ-mediated DNA repair activity could
be, like in HSCs, a double-edged sword that promotes short-
term survival of BSCs at the expense of their long-term genomic
integrity and could potentially allow for the accumulation of
cancerous mutations (Figure 4). Consistent with this notion,
SCID mice and mice deficient for Bcl-XL, a prosurvival gene,
show decreased susceptibility to chemical carcinogenesis
(Kemp et al., 1999; Kim et al., 2009), which has been attributed
to the elimination of mutated BSCs by apoptosis.
Melanocyte SCs
Melanocytes are neural crest-derived cells responsible for the
pigmentation of skin and hair. The mature melanocytes respon-
sible for hair color are derived from melanocyte SCs (MSCs),
Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc. 21
DNA
damage
DNA
damage response
pro-apoptotic
genes
Cell death
Pro-survival
genes
DNA repair
Terminal
differentiation
Hematopoietic SCs
Hair Follicle SCs
Stem cell
maintenance
DNA
damage
DNA
damage response
pro-apoptotic
genes
Cell death
Pro-survival
genes
DNA repair
Terminal
differentiation
Melanocyte SCs
Stem cell
maintenance
DNA
damage
DNA
damage response
pro-apoptotic
genes
Cell death
Pro-survival
genes
DNA repair
Terminal
differentiation
Intestinal SCs
Stem cell
maintenance
p53 p53 p53
ATM
Figure 4. DNA-Damage Response in Tissue-Specific Stem CellsCommon and distinct pathways of DNA-damage response in different types of tissue-specific SCs.
Cell Stem Cell
Review
which reside in the same niche as hair follicle BSCs. At each
cycle of hair regeneration, MSCs are stimulated to proliferate
and give rise to transit amplifying cells, which will expand in
the lower hair follicle before undergoing terminal differentiation,
which results in the integration of their pigment into the new
hair. At the end of each hair cycle, mature melanocytes undergo
apoptosis and are eliminated with the rest of the follicle, to be
subsequently replenished by the renewal and differentiation of
MSCs during the next cycle (Robinson and Fisher, 2009). Hair
graying, which is one of the most common signs of aging, results
from the depletion of MSCs from the hair follicle. The onset of
hair graying in mice and humans is accompanied by the pres-
ence of ectopically pigmented melanocytes, suggesting prema-
ture differentiation of MSCs within their niche (Nishimura et al.,
2005). Premature hair graying can also result fromahypomorphic
mutation in Mitf, the main regulator of MSC differentiation, that
results in a downregulation of bcl2 and in premature differentia-
tion of MSCs in the hair follicle (McGill et al., 2002). Bcl2 is critical
for MSCmaintenance as bcl2 null mice lose their coat pigmenta-
tion after the first hair cycle due to massive MSC apoptosis
(Nishimura et al., 2005). Premature hair graying and progressive
MSC loss also occur following administration of DNA damaging
agents such as IR, mitomycin C, or hydrogen peroxide (Inomata
et al., 2009). While the mechanisms underlying the DDR in MSCs
are not yet fully understood, p53, p16, and p19ARF, although
transiently activated by DNA damage, are not responsible for
the premature differentiation and loss of MSCs. Indeed, mice
deficient for p53 or the Ink4a locus (p16 and p19ARF) are not pro-
tected fromDNA-damage-induced hair graying, contrasting with
the requirement of p53 in mediating DNA-damage-induced cell
death in other tissue-specific SCs. In contrast, DNA damage
induces prolonged activation of the canonical differentiation
program of MSCs, including sustained upregulation of Mitf,
a key regulator of melanocyte differentiation and melanogenic
enzymes, which in turn stimulates the premature and ectopic
22 Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc.
differentiation of MSCs within their niche. The ATM checkpoint
regulator also exerts a protective function in MSCs because
Atm null mice and ATM-deficient patients exhibit premature
hair graying (Hakem, 2008) and loss of Atm sensitizes mice to
IR-induced premature MSC differentiation (Inomata et al., 2009).
Despite being located in the same hair follicle niche, BSCs and
MSCs respond very differently to DNA damage. Both types of
SCs do not senesce or commit apoptosis upon DNA damage,
but while BSCs repair their DNA rapidly and express high
levels of antiapoptotic molecules in order to avoid programmed
cell death, MSCs are eliminated by premature differentiation
(Figure 4). These different outcomes imply that cell intrinsic prop-
erties are more important than the local microenvironment in
controlling DDR in skin SCs. It is interesting to note that mela-
noma, a malignant tumor of melanocytes, does not arise from
hair follicle MSCs but rather from skin melanocytes. These cells
are located along the interfollicular epidermis, suggesting that
the premature differentiation of MSCs following DNA damage
may serve to eliminate precancerous MSCs residing in the hair
follicle.
Intestinal SCs
The intestinal tissue is very sensitive to DNA damage. Acute
whole-body irradiation (<6 Gy) induces considerable damage
to the intestine, resulting in severe diarrhea and electrolyte
imbalances, which can be lethal in extreme cases. The intestinal
lining is a simple epithelium composed of a single layer of cells
that can be divided into two compartments: the proliferative
base of the intestine, called the crypt, and the differentiated
intestinal cells forming the villi that face the intestinal lumen.
The intestinal SCs (ISCs) are localized at the bottom of the crypt,
where they proliferate to give rise to transit amplifying cells,
which are found along the crypt, and divide faster and migrate
to the upper part of the crypt where they undergo cell-cycle
arrest and terminal differentiation (Barker et al., 2010; Casali
and Batlle, 2009; Marshman et al., 2002). Although the exact
Cell Stem Cell
Review
position of the ISCs within the crypt is still under intense debate,
it has long been suggested that ISCs reside at the +4 position
from the base of the crypts and that these SCs are more quies-
cent compared to the other crypt cells. Consistent with that
notion, Bmi1, which is preferentially expressed in +4 crypt cells,
induced long-term labeling of the crypto-vilus unit inBmi1CREER
reporter mice, consistent with the labeling of long-lived multipo-
tent ISCs (Sangiorgi and Capecchi, 2008). A second population
of ISCs expressing Lgr5, a leucine-rich orphan G protein-
coupled receptor and Wnt pathway activated gene, has recently
been identified (Barker et al., 2007). Lgr5+ cells cycle more
frequently than the +4 cells and are located at the bottom
of the crypt intercalated between the paneth cells. Lineage
tracing experiments using Lgr5-GFP-IRES-Cre-ERT;;RosaLacZ
reporter mice demonstrated that Lgr5+ cells give rise to all intes-
tinal cell lineages and result in the long-term labeling of the
cryptovilus unit, also consistent with the labeling of long-lived
multipotent ISCs.
ISCs are extremely sensitive to DNA damage and undergo
massive apoptosis upon low doses of irradiation (1 Gy). Interest-
ingly, while it is generally assumed that radiosensitivity is corre-
lated with cell-cycle status (Gudkov and Komarova, 2003), the
apoptosis sensitivity of intestinal crypt cells is inversely corre-
lated with their relative quiescence. The most quiescent ISCs
located at +4 position are the most sensitive to IR-induced cell
death, followed by the more active Lgr5+ ISCs, whereas the
rapidly cycling transit-amplifying cells appear to be the most
radioresistant (Barker et al., 2007; Potten et al., 2002; Wilson
et al., 1998). Different mechanisms are responsible for the
extreme sensitivity of ISCs to DNA damage, including an
enhanced activation of the p53 pathway, lower expression of
the antiapoptotic protein Bcl2 (Merritt et al., 1995), and general
lack of DNA repair activity (Potten, 2004). Upon irradiation,
expression of p53 and its downstream target genes p21 and
puma increases throughout the crypts, but the frequency of
p53-positive cells and the levels of expression of its target genes
are higher at the base of the crypt and progressively decrease
along the crypts toward the vilus (Merritt et al., 1994; Qiu et al.,
2008; Wilson et al., 1998). Furthermore, IR does not induce
apoptosis in the intestine of p53 null mice (Merritt et al., 1994;
Qiu et al., 2008; Wilson et al., 1998). IR-induced ISC apoptosis
is also blocked in puma-deficient mice, and ISC survival is pro-
longed after administration of puma antisense nucleotides,
thereby demonstrating that Puma is the main proapoptotic
target of the p53-mediated DDR in ISCs (Qiu et al., 2008). In
contrast to other SC populations described above, bcl2 expres-
sion is not detected in ISCs and irradiated bcl2 null mice only
show a modest increase in ISC apoptosis, suggesting that
Bcl2 does not play a critical role in protecting ISCs from DNA-
damage-induced cell death (Merritt et al., 1995). Finally, the
absence of an irradiation dose response of crypt degeneration
suggests that quiescent ISCs lack DNA repair capacity, thereby
increasing their propensity to undergo apoptosis following DNA
damage (Hendry et al., 1982; Potten, 2004).
The architecture of the colon resembles that of the small intes-
tine. Similar to ISCs, colonic SCs (CoSCs) are also localized at
the bottom of the crypt and express Lgr5, although CoSCs
exhibit a longer cell-cycle time than ISCs. Interestingly, the
DDR of CoSCs differs significantly from that of ISCs, with CoSCs
being considerably more radioresistant than ISCs (Figure 4). It is
estimated that CoSCs require eight times the dose of irradiation
needed by ISCs to reach similar levels of apoptosis (Barker et al.,
2007; Potten and Grant, 1998; Pritchard et al., 2000). The greater
radioresistance of CoSCs has been attributed to a lower expres-
sion of p53 (Hendry et al., 1997; Merritt et al., 1994) and higher
expression of bcl2 (Merritt et al., 1995; Qiu et al., 2008). Further-
more, in contrast to ISCs, CoSCs from bcl2 null mice show
a much greater increase in DNA-damage-induced apoptosis,
demonstrating that bcl2 expression in CoSCs does contribute
to their higher relative radioresistance. The altruistic suicide
of ISCs in response to DNA damage could decrease the acquisi-
tion of precancerous mutations in these cells and potentially
explain the rarity of intestinal neoplasia compared to the higher
frequency of colonic cancers, despite the higher cellular turnover
of the intestine.
Germline SCs
Primordial germ cells (PGCs) are transient precursors of germ
SCs (GSCs), which uponmeiosis give rise to the gametes (sperm
and egg), which are the only cells capable of transferring genetic
information from one generation to the next (Chuva de Sousa
Lopes and Roelen, 2010; Laird et al., 2008; Richardson and
Lehmann, 2010). PGCs are specified in the embryo, migrate to
the gonadal ridges were they undergo sex determination, and
give rise to the female (oogonia) or the male (spermatogonia)
GSCs. The spermatogonia exhibit an almost unlimited life
span, remaining quiescent until puberty, at which point they re-
acquire the ability to self-renew, undergo meiosis, and produce
mature male gametes for the lifetime of the organism. In sharp
contrast, the pool of oogonia is established during embryogen-
esis, and consequently, females are born with a finite number
of oogonia.
The generation of haploid chromosomes during meiosis
requires many of the proteins involved in DNA repair (Sasaki
et al., 2010). During PGC maturation, genome-wide DNA
demethylation occurs in order to erase genomic imprinting.
DNAdemethylation inmousePGCs is initiatedby theappearance
of single-strand breaks and activation of theBERpathway,which
may be linked to deamination of methylcytosine or to other yet-
to-be-discovered mechanisms (Hajkova et al., 2010). Mutations
in the germ line can be extremely dangerous and can either
directly lead to sterility (Loft et al., 2003) or transmission of heri-
table genetic diseases by the gametes. Genetic aberrations in
GSCs may occur upon radiation exposure, such as radiotherapy
and radiological examination, or after exposure to teratogenic or
mutagenic chemicals, but the main source of DNA damage is
their normal metabolic activity and ROS production (Kujjo et al.,
2010).Microarray analysis uncovered that DNA-damage sensors
and multiple components of the NHEJ, BER, NER, and MMR
pathways are expressed in human oocytes (Menezo et al.,
2007), with a similar high expression of DNA repair proteins found
in human sperm (Galetzka et al., 2007), which suggest that GSCs
and gametes are well equipped to respond to DNA damage.
Accordingly, spermatogonia in Atm-deficient mice are progres-
sively lost, undergo meiotic arrest, accumulate DNA damage,
and lose their self-renewal potential in a p21-dependent manner
(Takubo et al., 2008).Mice expressing the hypomorphicmutation
of Rad50k22m also show severe attrition of spermatogonia, which
could be minimized by loss of p53 (Bender et al., 2002).
Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc. 23
Cell Stem Cell
Review
The cell-cycle duration of human spermatogonia is estimated
to be around 16 days, with male GSCs being mostly kept in the
G0/G1 phase of the cell cycle. Consequently, NHEJ is the first line
of DNA repair in these cells. Interestingly, in vitro studies in mice
showed that spermatogonia are more sensitive to IR when they
are quiescent than when they are proliferating (Forand et al.,
2009; Moreno et al., 2001). In oogonia, the homologous chromo-
somes are close to each other and female GSCs preferentially
repair their DNA using HR (Baker, 1971). Mutations in the HR
repair pathway render female GSCs more susceptible to DNA-
damage-mediated cell death as shown by the increase sensi-
tivity to doxorubicin-induced apoptosis in oocytes from mice
deficient in Rad51 (Kujjo et al., 2010). Contrary to most SC pop-
ulations and somatic cells, the DDR in female GSCs does not
depend on p53. Instead, TAp63, an isoform of the p63 gene
and a p53 homolog, is constitutively expressed in oocytes and
is rapidly phosphorylated following DNA damage. Deletion of
TAp63 in mice results in a major increase in oocyte radioresist-
ance, consistent with the notion that TAp63 is the primary medi-
ator of DDR pathway in oocytes (Suh et al., 2006).
Mammary SCs
The mammary gland alternates between cycles of growth and
degeneration in relation to the estrus cycle. Mammary stem cells
(MaSCs) are responsible for homeostasis of the breast tissue
and for the massive tissue expansion and remodeling that
occurs during pregnancy and lactation (Visvader, 2009). MaSCs
have been isolated from mice and humans and represent multi-
potent SCs that have the ability to self renew as well as to differ-
entiate into ductal, alveolar, and myoepithelial cell lineages
(Ginestier et al., 2007; Shackleton et al., 2006; Stingl et al.,
2006). Breast cancer is the most common form of malignancies
in women. Mutations in genes involved in DNA repair such
as BRCA1 and BRCA2 are found in the majority of patients
with hereditary breast cancers, demonstrating the importance
of the HR-repair pathway in preventing the occurrence of
mammary tumors (Bradley and Medina, 1998). Mice deficient
for Brca1 are embryonic lethal, but mice with a conditional dele-
tion of Brca1 in the mammary epithelium are viable, display
severe abnormalities in mammary morphogenesis, and develop
undifferentiated breast cancers (Hakem, 2008). Knockdown of
BRCA1 in human MaSCs leads to a decrease of differentiated
luminal cells and an increase in cells with SC characteristics,
which suggests that BRCA1 is required for normal MaSC differ-
entiation and that BRCA1 loss may result in the accumulation of
genetically unstable MaSCs that are susceptible to cancer
development (Liu et al., 2008).
While the role of DNA repair in mammary development, main-
tenance, and prevention of breast tumors is well established, the
mechanisms underlying the DDR in MaSCs have only just begun
to emerge. Mouse MaSCs are more radioresistant than their
differentiated progeny, and their numbers increase following
IR (Woodward et al., 2007). Interestingly, MaSCs present less
DNA damage and rapidly activate the Wnt/b-catenin signal-
ing pathway following IR. Furthermore, increasing b-catenin
signaling by overexpression of Wnt1 or stabilized b-catenin
increases the survival of MaSCs following DNA damage, indi-
cating that Wnt/b-catenin signaling is an important component
of the DDR in MaSCs that may promote MaSC survival through
upregulation of survivin, a direct Wnt/b-catenin target gene
24 Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc.
(Chen et al., 2007; Woodward et al., 2007). It would certainly
be interesting to determine whether the selective activation of
Wnt/b-catenin pathway observed in MaSCs also occurs in other
tissue-specific SCs and promotes their survival following DNA
damage. Another mechanism that might promote MaSCs resis-
tance to DNA damage is their low level of ROS compared their
differentiated progeny (Diehn et al., 2009).
DNA-Damage Response in Cancer Stem CellsA number of human cancers, including leukemia, glioblastoma,
breast, and skin cancers, contain cells with higher clonogenic
potential that are capable of reforming the parental tumors
upon transplantation. These cells functionally resemble tissue-
specific SCs, albeit with aberrant self-renewal and differentiation
abilities, and have been collectively referred to as cancer SCs
(CSCs), despite their variable developmental origin (Clarke and
Fuller, 2006; Jordan et al., 2006). It has been suggested that
CSCs are responsible for disease progression and tumor relapse
after therapy. Recent studies indicate that CSCs may take
advantage of the mechanisms of DNA repair used by tissue-
specific SCs to mediate resistance to chemo- and radiotherapy.
CSCs in Leukemia
Leukemias are cancers of the blood system, which often arise
due to deregulated HSC functions or acquisition of extended
self-renewal capabilities by more mature progenitor cells
(Passegue, 2005). Leukemia CSCs exist in acute myeloid
leukemia (AML) and chronic myelogenous leukemia (CML) and
have been shown to be more resistant to cancer therapies
than the bulk of the leukemia cells, indicating that their survival
may be responsible for disease persistence and cancer relapse
(Elrick et al., 2005; Jordan et al., 2006). Leukemia CSCs also use
to their advantage some protective mechanisms of HSCs,
including quiescent cell-cycle status, localization to a hypoxic
niche, and DDR mechanisms, to specifically escape chemo-
and radiotherapy that kill the bulk of the tumor cells (Guzman
and Jordan, 2009).
CML is a two-stage blood disease caused by the acquisition of
the chromosomal translocation fusion product BCR/ABL in
HSCs, which can be separated into chronic and acute phases.
The transition from chronic to acute disease is still poorly under-
stood, but the presence of DNA damage and the acquisition of
additional chromosomal aberrations resulting in overall genomic
instability in both HSCs and their downstream progeny is
believed to play a critical role in this transition (Burke and Carroll,
2010). BCR/ABL expression increases intracellular ROS levels,
which in turn enhances oxidative stress and DNA damage
and deregulates DNA repair mechanisms, thereby promoting
unfaithful and/or inefficient DNA repair leading to mutations
and chromosomal aberrations (Perrotti et al., 2010). Malfunction-
ing MMR, mutagenic NER, and compromised DSB repair (both
HR and NHEJ) are all hallmarks of cells expressing BCR/ABL
(Burke and Carroll, 2010; Deutsch et al., 2001; Slupianek et al.,
2002, 2006). Once DNA damage occurs, BCR/ABL-mediated
signaling can also inhibit apoptosis, thereby allowing cells to
survive DNA damage with which they normally would not be
able to cope (Burke and Carroll, 2010; Deutsch et al., 2001;
Slupianek et al., 2002, 2006). The genomic instability induced
by BCR/ABL has major implications for the pathogenesis and
treatment of CML since it can facilitate disease progression
Cell Stem Cell
Review
from chronic to acute phase and promote the acquisition of
resistance against the current drugs used to treat CML (tyrosine
kinase inhibitors such as imatinib). Indeed, evolution from HSC-
derived CSCs to myeloid progenitor-derived CSCs has been
observed during the transition to myeloid blast crisis in human
CML and has been linked to activated mutations in the Wnt/
b-catenin pathway and acquisition of aberrant self-renewal
activity in HSC progeny (Rice and Jamieson, 2010). Preventing
oxidative stress and correcting defects in DNA repair pathways
in BCR/ABL-expressing CSCs at all stages of the disease may
therefore be beneficial to limit the acquisition of drug resistance
and slow down CML progression (Koptyra et al., 2006; Perrotti
et al., 2010).
Leukemia CSCs maintain some of the same protective mech-
anisms as normal HSCs. CSCs in both CML and AML have been
found to be quiescent (Elrick et al., 2005; Guan et al., 2003; Ishi-
kawa et al., 2007), suggesting that cell-cycle restriction is one of
the protective mechanisms that leukemia CSCs utilize to their
advantage (Guzman and Jordan, 2009). Indeed, human AML
CSCs transplanted into immunodeficient mice use quiescence
as a protective mechanism against chemotherapy (Saito et al.,
2010). When these cells are induced to exit quiescence and to
enter the cell cycle by treating the mice with the cytokine
G-CSF, AML CSCs become more sensitive to chemotherapy
and are effectively eliminated in vivo. Leukemia CSCs are also
able to co-opt other mechanisms used by normal HSCs for their
protection, such as p53-mediated induction of p21 and resulting
growth arrest that has recently been found to be critical in pro-
tecting adult HSCs from IR (Mohrin et al., 2010). Expression of
the PML/RAR or AML1/ETO fusion oncoproteins in murine
HSCs induces high levels of DNA damage and activates a p21-
dependent cell-cycle arrest in AML CSCs, which allows them
to repair excessive DNA damage and to escape apoptosis,
thereby maintaining their leukemic self-renewal capacity (Viale
et al., 2009). While it may seem paradoxical that a leukemia-initi-
ating oncogene promotes cell-cycle arrest instead of prolifera-
tion, the hijacking of such a protective mechanism provides
a strong selective advantage to the CSCs. In the absence of
p21, AML CSCs were more sensitive to replicative and thera-
peutic stress, and p21 null HSCs expressing PML/RAR or
AML1/ETO were unable to transplant the disease into recipient
mice, indicating a failure to maintain CSC activity (Viale et al.,
2009).
CSCs in Breast Cancer
The first evidence that solid tumors also contained cells with
CSC properties came with the demonstration that in human
breast cancer, CD44+CD24�/lo cells are more clonogenic and,
when transplanted in immunocompromized mice, are able to
generate tumors that recapitulate the parental disease (Al-Hajj
et al., 2003). Transcriptional profiling of murine mammary gland
CSCs revealed increased expression of many DDR and DNA
repair associated genes (Zhang et al., 2008), suggesting that
mammary gland CSCs might be more resistant to chemo- and/
or radiotherapy. Comparison of tumor biopsies before and
after neoadjuvant chemotherapy showed an increase in the
proportion of mammary gland CSCs with mammosphere-form-
ing capacity following chemotherapy, hence confirming that
mammary gland CSCs are more resistant to chemotherapy (Li
et al., 2008; Shafee et al., 2008). Like normal MaSCs, mammary
gland CSCs harbor lower levels of ROS compared to the rest of
the tumor cells, due to increased levels of genes regulating free
radical scavenging systems, such as those of the glutathione
metabolism. Mammary gland CSCs from human xenografts
(Phillips et al., 2006) or MMTV-Wnt1 tumor-bearing mice (Diehn
et al., 2009) exhibited higher survival upon IR treatment. Consis-
tent with the fact that ROS levels control IR-induced DNA
damage and apoptosis in CSCs, inhibition of glutathione metab-
olism decreased the clonogenic potential and sensitized
mammary gland CSCs to IR (Diehn et al., 2009). Furthermore,
p53-deficient mammary gland CSCs show accelerated DNA
repair activity as well as high Akt and Wnt signaling activity,
which promotes CSC survival following IR treatment (Zhang
et al., 2010). Interestingly, administration of an Akt inhibitor
inhibits b-catenin signaling and sensitizes mammary gland
CSCs to radiotherapy.
Understanding the role of DNA repair genes in the pathogen-
esis of breast cancer has been exploited for the development
of novel anticancer strategies. Tumors derived from Brca1-defi-
cient cells are extremely sensitive to the inhibition of PARP,
which plays an important role in the repair of single-strand
breaks by the BER pathway. In the absence of Brca1 and HR-
mediated DNA repair, persistent single-strand breaks need to
be repaired by the BER pathways, and as a consequence, inhi-
bition of PARP blocks this alternative pathway of DNA repair,
inducing cell death preferentially in cancer cells. A PARP inhibitor
prolonged disease-free survival when administered alone or in
combination with chemotherapeutic drugs in a mouse model
of brca1-deficient mammary gland tumors (Rottenberg et al.,
2008) and also exhibits clinical efficacy in human breast cancers
(Fong et al., 2009).
CSCs in Glioblastoma
Glioblastoma multiform (GBM) represents the most aggressive
type of brain tumor. The standard treatment combines surgery
and radiotherapy, but still, most patients relapse after therapy,
with a median survival of less than 12 months (Prados and Levin,
2000). CSCs from human glioblastoma have been isolated
based on the expression of prominin (CD133) (Singh et al.,
2004). Irradiation of human GBM xenografts led to increased
proportions of CD133+ cells, indicating that CSCs may be
responsible for tumor relapse after radiotherapy (Bao et al.,
2006). CSCs from GBM are more resistant to IR-induced cell
death compared to non-CSCs and show more robust activation
of DNA-damage checkpoint proteins, including ATM, Chk1, and
Chk2, as well as more efficient DNA repair activity. Importantly,
treatment with inhibitors of Chk1 and Chk2 kinases sensitizes
CSCs to IR-induced cell death, suggesting that inhibition of
DNA-damage checkpoint in CSCs may improve the efficiency
of radiotherapy in GBM (Bao et al., 2006). However, this increase
in DNA repair activity was not observed in all glioma-derived cell
lines (Ropolo et al., 2009), and loss of Chk2 instead potentiates
GBM radioresistance in mice (Squatrito et al., 2010), indicating
that this characteristic may be related to certain glioblastoma
subtypes. Moreover, glioma stem cell-like cells have been
shown to exhibit elevated levels of the antiapoptotic protein
Mcl1 that contributes to their radioresistance (Tagscherer
et al., 2008). Temozolomide, the most commonly used chemo-
therapy in the treatment of GBM that induces cell death by trig-
gering the methylation of guanine at position 6, which can be
Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc. 25
Cell Stem Cell
Review
removed by the methylguanine DNAmethyltransferase (MGMT),
induced CSC depletion in MGMT-negative, but not in MGMT-
positive, GBM (Beier et al., 2008).
Future DirectionsThe study of DDR in different types of tissue-specific SCs has
clearly highlighted the existence of common mechanisms acting
in certain adult SC populations to limit the amount of DNA
damage, to restrain them from undergoing massive apoptosis
and being exhausted following DNA damage, and to preserve
overall tissue function. These protective mechanisms may
have a cost for these tissue-specific SC populations, such as
blood HSCs and hair follicle BSCs, as they preserve immediate
survival at the expense of long-term maintenance of genomic
integrity, which may lead to aging, tissue atrophy, and/or cancer
development. Further studies are required to fully understand
and ultimately prevent the long-term deleterious consequences
of these protective mechanisms. In contrast, some tissue-
specific SCs, such as intestinal SCs, are not well protected
and undergo massive death after DNA damage. More studies
are needed to better understandwhy someSCs prefer to commit
suicide after DNA damage while others decide to survive, as well
as to understand how altruistic suicide might provide a selective
advantage to overall tissue function and what molecular mecha-
nisms dictate these very different outcomes.
Most of the studies on DDR in tissue-specific SCs have been
performed in adult animals during normal, or homeostasic,
conditions. Since the activity and relative quiescence of SCs
varies considerably during organogenesis, adult homeostasis,
and tissue repair following injuries, the consequence of DNA
damage might be very different in SCs at different ontogenic
stages or levels of activity, as it has now been shown for fetal
and adult HSCs. During organogenesis and tissue regeneration,
SCs divide more frequently, whereas during homeostasis,
SCs are more quiescent. Since different mechanisms of DNA
repair are used depending on the cell-cycle stage of the
damaged cells, are HR and NHEJ repair pathways differentially
important to preserve SC fitness depending on their activation
state? Are DNA repair-associated genes differentially activated
during morphogenesis, homeostasis, and regeneration? Do
mice with defective NHEJ or HR repair genes present different
phenotypes when these genes are ablated during embryonic
development compared to adult life? Future investigations are
needed to fully comprehend the role of these different DNA repair
mechanisms in SC biology.
In addition to the conserved set of genes that act in DDR and
DNA repair pathways, some miRNAs have recently been shown
to be induced by p53 in response to DNA damage and play an
important role in DDR outcomes of survival versus apoptosis
by interacting with key tumor-suppression networks (He et al.,
2007). Irradiation of cultured cells uncovered the involvement
of miR-34a in promoting apoptosis (Chang et al., 2007) and of
miR-192 and miR-215 in cell-cycle arrest induction (Georges
et al., 2008). Moreover, miR-34a is lost in several cancer cell
lines (Chang et al., 2007). Future studies will determine whether
DNA damage and repair-associated miRNAs are differentially
expressed in tissue-specific SCs compared to their differenti-
ated progeny and whether these miRNAs modulate the DDR in
different types of tissue-specific and cancer SCs. Another
26 Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc.
important question is whether CSCs from different types of
cancer also exhibit a survival advantage following chemo- and
radiotherapy. If so, is this resistance related to enhanced DNA
repair mechanisms or higher expression of antiapoptotic
factors? Do CSCs retain the DNA repair properties of the SCs
of their tissue of origin, or do they acquire functionally similar
characteristics during cancer progression through a selective
pressure? Do DDR abnormalities in CSCs versus bulk cancer
cells account for the vast genomic instability present within the
bulk of the tumors? Progresses in next generationwhole genome
sequencing and further studies of defined CSC populations will
be needed to assess how defects in their DDR contribute to
cancer evolution and associated genomic or base-pair level
changes.
Addressing these open questions will have profound implica-
tions for our understanding of how tissue-specific SCs respond
to DNA damage and maintain the integrity of their genome, how
deregulation of these mechanisms leads to cancer and aging,
how CSCs respond to chemo- and radiotherapy, and how these
characteristics may be exploited to increase the efficacy of
current anticancer treatments.
ACKNOWLEDGMENTS
We thank Drs. E. Pietras and M. Warr for their insightful comments. C.B. andP.A.S are chercheur qualifie of the Fonds de la Recherche Scientifique(F.R.S.)/Fonds National de la Recherche Scientifique (FNRS). M.M. is sup-ported by a CIRM predoctoral training grant. This work was supported bythe program CIBLES of the Wallonia Region, a research grant from the Fonda-tion Contre le Cancer and the fond Gaston Ithier, a starting grant of the Euro-pean Research Council (ERC) and the EMBO Young Investigator Program toC.B., and a CIRM New Faculty Award and Rita Allen Scholar Award to E.P.
REFERENCES
Al-Hajj, M., Wicha, M.S., Benito-Hernandez, A., Morrison, S.J., and Clarke,M.F. (2003). Prospective identification of tumorigenic breast cancer cells.Proc. Natl. Acad. Sci. USA 100, 3983–3988.
Asai, T., Liu, Y., Bae, N., and Nimer, S.D. (2010). The p53 tumor suppressorprotein regulates hematopoietic stem cell fate. J. Cell Physiol. Published onlineDecember 6, 2010. 10.1002/jcp.22561.
Baker, T.G. (1971). Comparative aspects of the effects of radiation duringoogenesis. Mutat. Res. 11, 9–22.
Bao, S., Wu, Q., McLendon, R.E., Hao, Y., Shi, Q., Hjelmeland, A.B., Dewhirst,M.W., Bigner, D.D., and Rich, J.N. (2006). Glioma stem cells promote radiore-sistance by preferential activation of the DNA damage response. Nature 444,756–760.
Barker, N., van Es, J.H., Kuipers, J., Kujala, P., van den Born, M., Cozijnsen,M., Haegebarth, A., Korving, J., Begthel, H., Peters, P.J., et al. (2007). Identi-fication of stem cells in small intestine and colon by marker gene Lgr5. Nature449, 1003–1007.
Barker, N., Bartfeld, S., and Clevers, H. (2010). Tissue-resident stem cellpopulations of rapidly self-renewing organs. Cell Stem Cell 7, 656–670.
Barlow, C., Hirotsune, S., Paylor, R., Liyanage, M., Eckhaus, M., Collins, F.,Shiloh, Y., Crawley, J.N., Ried, T., Tagle, D., et al. (1996). Atm-deficientmice: a paradigm of ataxia telangiectasia. Cell 86, 159–171.
Beier, D., Rohrl, S., Pillai, D.R., Schwarz, S., Kunz-Schughart, L.A., Leukel, P.,Proescholdt, M., Brawanski, A., Bogdahn, U., Trampe-Kieslich, A., et al.(2008). Temozolomide preferentially depletes cancer stem cells in glioblas-toma. Cancer Res. 68, 5706–5715.
Bender, C.F., Sikes, M.L., Sullivan, R., Huye, L.E., Le Beau, M.M., Roth, D.B.,Mirzoeva, O.K., Oltz, E.M., and Petrini, J.H. (2002). Cancer predisposition andhematopoietic failure in Rad50(S/S) mice. Genes Dev. 16, 2237–2251.
Cell Stem Cell
Review
Blanpain, C., and Fuchs, E. (2009). Epidermal homeostasis: a balancing act ofstem cells in the skin. Nat. Rev. Mol. Cell Biol. 10, 207–217.
Botchkarev, V.A., Komarova, E.A., Siebenhaar, F., Botchkareva, N.V.,Komarov, P.G., Maurer, M., Gilchrest, B.A., and Gudkov, A.V. (2000). p53 isessential for chemotherapy-induced hair loss. Cancer Res. 60, 5002–5006.
Bradley, A., and Medina, D. (1998). Introduction: BRCA1 and BRCA2 inmammary gland development and tumorigenesis. J. Mammary Gland Biol.Neoplasia 3, 363–364.
Brown, E.J., and Baltimore, D. (2000). ATR disruption leads to chromosomalfragmentation and early embryonic lethality. Genes Dev. 14, 397–402.
Burke, B.A., and Carroll, M. (2010). BCR-ABL: a multi-faceted promoter ofDNA mutation in chronic myelogeneous leukemia. Leukemia 24, 1105–1112.
Casali, A., and Batlle, E. (2009). Intestinal stem cells in mammals andDrosophila. Cell Stem Cell 4, 124–127.
Chang, T.C., Wentzel, E.A., Kent, O.A., Ramachandran, K., Mullendore, M.,Lee, K.H., Feldmann, G., Yamakuchi, M., Ferlito, M., Lowenstein, C.J., et al.(2007). Transactivation of miR-34a by p53 broadly influences gene expressionand promotes apoptosis. Mol. Cell 26, 745–752.
Chen,M.S.,Woodward,W.A., Behbod, F., Peddibhotla, S., Alfaro,M.P., Buch-holz, T.A., and Rosen, J.M. (2007). Wnt/beta-catenin mediates radiation resis-tance of Sca1+ progenitors in an immortalized mammary gland cell line. J. CellSci. 120, 468–477.
Chuva de Sousa Lopes, S.M., and Roelen, B.A. (2010). On the formation ofgerm cells: The good, the bad and the ugly. Differentiation 79, 131–140.
Clarke, M.F., and Fuller, M. (2006). Stem cells and cancer: two faces of eve.Cell 124, 1111–1115.
d’Adda di Fagagna, F. (2008). Living on a break: cellular senescence as aDNA-damage response. Nat. Rev. Cancer 8, 512–522.
Deutsch, E., Dugray, A., AbdulKarim, B., Marangoni, E., Maggiorella, L.,Vaganay, S., M’Kacher, R., Rasy, S.D., Eschwege, F., Vainchenker, W., et al.(2001). BCR-ABL down-regulates the DNA repair protein DNA-PKcs. Blood97, 2084–2090.
Diehn, M., Cho, R.W., Lobo, N.A., Kalisky, T., Dorie, M.J., Kulp, A.N., Qian, D.,Lam, J.S., Ailles, L.E., Wong, M., et al. (2009). Association of reactive oxygenspecies levels and radioresistance in cancer stem cells. Nature 458, 780–783.
Elrick, L.J., Jorgensen, H.G., Mountford, J.C., and Holyoake, T.L. (2005).Punish the parent not the progeny. Blood 105, 1862–1866.
Finlan, L.E., Nenutil, R., Ibbotson, S.H., Vojtesek, B., and Hupp, T.R. (2006).CK2-site phosphorylation of p53 is induced in DeltaNp63 expressing basalstem cells in UVB irradiated human skin. Cell Cycle 5, 2489–2494.
Fong, P.C., Boss, D.S., Yap, T.A., Tutt, A., Wu, P., Mergui-Roelvink, M.,Mortimer, P., Swaisland, H., Lau, A., O’Connor, M.J., et al. (2009). Inhibitionof poly(ADP-ribose) polymerase in tumors from BRCA mutation carriers.N. Engl. J. Med. 361, 123–134.
Forand, A., Fouchet, P., Lahaye, J.B., Chicheportiche, A., Habert, R., andBernardino-Sgherri, J. (2009). Similarities and differences in the in vivoresponse of mouse neonatal gonocytes and spermatogonia to genotoxicstress. Biol. Reprod. 80, 860–873.
Galetzka, D., Weis, E., Kohlschmidt, N., Bitz, O., Stein, R., and Haaf, T. (2007).Expression of somatic DNA repair genes in human testes. J. Cell. Biochem.100, 1232–1239.
Georges, S.A., Biery, M.C., Kim, S.Y., Schelter, J.M., Guo, J., Chang, A.N.,Jackson, A.L., Carleton, M.O., Linsley, P.S., Cleary, M.A., et al. (2008). Coor-dinated regulation of cell cycle transcripts by p53-Inducible microRNAs,miR-192 and miR-215. Cancer Res. 68, 10105–10112.
Ginestier, C., Hur, M.H., Charafe-Jauffret, E., Monville, F., Dutcher, J., Brown,M., Jacquemier, J., Viens, P., Kleer, C.G., Liu, S., et al. (2007). ALDH1 isa marker of normal and malignant human mammary stem cells and a predictorof poor clinical outcome. Cell Stem Cell 1, 555–567.
Goldschmidt, H., and Sherwin, W.K. (1980). Reactions to ionizing radiation.J. Am. Acad. Dermatol. 3, 551–579.
Guan, Y., Gerhard, B., and Hogge, D.E. (2003). Detection, isolation, and stim-ulation of quiescent primitive leukemic progenitor cells from patients withacute myeloid leukemia (AML). Blood 101, 3142–3149.
Gudkov, A.V., and Komarova, E.A. (2003). The role of p53 in determining sensi-tivity to radiotherapy. Nat. Rev. Cancer 3, 117–129.
Guzman, M.L., and Jordan, C.T. (2009). Lessons learned from the study ofJunB: new insights for normal and leukemia stem cell biology. Cancer Cell15, 252–254.
Hajkova, P., Jeffries, S.J., Lee, C., Miller, N., Jackson, S.P., and Surani, M.A.(2010). Genome-wide reprogramming in the mouse germ line entails thebase excision repair pathway. Science 329, 78–82.
Hakem, R. (2008). DNA-damage repair; the good, the bad, and the ugly. EMBOJ. 27, 589–605.
Harfouche, G., Vaigot, P., Rachidi, W., Rigaud, O., Moratille, S., Marie, M.,Lemaitre, G., Fortunel, N.O., and Martin, M.T. (2010). Fibroblast growth factortype 2 signaling is critical for DNA repair in human keratinocyte stem cells.Stem Cells 28, 1639–1648.
Hasty, P., Campisi, J., Hoeijmakers, J., van Steeg, H., and Vijg, J. (2003). Agingand genomemaintenance: lessons from themouse? Science 299, 1355–1359.
He, L., He, X., Lowe, S.W., and Hannon, G.J. (2007). microRNAs join the p53network–another piece in the tumour-suppression puzzle. Nat. Rev. Cancer7, 819–822.
He, S., Nakada, D., and Morrison, S.J. (2009). Mechanisms of stem cell self-renewal. Annu. Rev. Cell Dev. Biol. 25, 377–406.
Hendry, J.H., Potten, C.S., Chadwick, C., and Bianchi, M. (1982). Cell death(apoptosis) in the mouse small intestine after low doses: effects of dose-rate, 14.7 MeV neutrons, and 600 MeV (maximum energy) neutrons. Int.J. Radiat. Biol. Relat. Stud. Phys. Chem. Med. 42, 611–620.
Hendry, J.H., Cai, W.B., Roberts, S.A., and Potten, C.S. (1997). p53 deficiencysensitizes clonogenic cells to irradiation in the large but not the small intestine.Radiat. Res. 148, 254–259.
Hoeijmakers, J.H. (2009). DNA damage, aging, and cancer. N. Engl. J. Med.361, 1475–1485.
Inomata, K., Aoto, T., Binh, N.T., Okamoto, N., Tanimura, S., Wakayama, T.,Iseki, S., Hara, E., Masunaga, T., Shimizu, H., et al. (2009). Genotoxic stressabrogates renewal of melanocyte stem cells by triggering their differentiation.Cell 137, 1088–1099.
Ishikawa, F., Yoshida, S., Saito, Y., Hijikata, A., Kitamura, H., Tanaka, S.,Nakamura, R., Tanaka, T., Tomiyama, H., Saito, N., et al. (2007). Chemo-therapy-resistant human AML stem cells home to and engraft within thebone-marrow endosteal region. Nat. Biotechnol. 25, 1315–1321.
Ito, K., Hirao, A., Arai, F., Matsuoka, S., Takubo, K., Hamaguchi, I., Nomiyama,K., Hosokawa, K., Sakurada, K., Nakagata, N., et al. (2004). Regulation ofoxidative stress by ATM is required for self-renewal of haematopoietic stemcells. Nature 431, 997–1002.
Ito, K., Hirao, A., Arai, F., Takubo, K., Matsuoka, S., Miyamoto, K., Ohmura, M.,Naka, K., Hosokawa, K., Ikeda, Y., et al. (2006). Reactive oxygen species actthrough p38 MAPK to limit the lifespan of hematopoietic stem cells. Nat. Med.12, 446–451.
Jordan, C.T., Guzman, M.L., and Noble, M. (2006). Cancer stem cells. N. Engl.J. Med. 355, 1253–1261.
Kemp, C.J., Vo, K., and Gurley, K.E. (1999). Resistance to skin tumorigenesisin DNAPK-deficient SCIDmice is not due to immunodeficiency but results fromhypersensitivity to TPA-induced apoptosis. Carcinogenesis 20, 2051–2056.
Kenyon, J., and Gerson, S.L. (2007). The role of DNA damage repair in aging ofadult stem cells. Nucleic Acids Res. 35, 7557–7565.
Kim, D.J., Kataoka, K., Sano, S., Connolly, K., Kiguchi, K., and DiGiovanni, J.(2009). Targeted disruption of Bcl-xL in mouse keratinocytes inhibitsboth UVB- and chemically induced skin carcinogenesis. Mol. Carcinog. 48,873–885.
Koptyra, M., Falinski, R., Nowicki, M.O., Stoklosa, T., Majsterek, I., Nieborow-ska-Skorska, M., Blasiak, J., and Skorski, T. (2006). BCR/ABL kinase induces
Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc. 27
Cell Stem Cell
Review
self-mutagenesis via reactive oxygen species to encode imatinib resistance.Blood 108, 319–327.
Kujjo, L.L., Laine, T., Pereira, R.J., Kagawa,W., Kurumizaka, H., Yokoyama, S.,and Perez, G.I. (2010). Enhancing survival of mouse oocytes following chemo-therapy or aging by targeting Bax and Rad51. PLoS ONE 5, e9204.
Laird, D.J., von Andrian, U.H., and Wagers, A.J. (2008). Stem cell trafficking intissue development, growth, and disease. Cell 132, 612–630.
Li, A., Simmons, P.J., and Kaur, P. (1998). Identification and isolation of candi-date human keratinocyte stem cells based on cell surface phenotype. Proc.Natl. Acad. Sci. USA 95, 3902–3907.
Li, X., Lewis, M.T., Huang, J., Gutierrez, C., Osborne, C.K., Wu, M.F., Hilsen-beck, S.G., Pavlick, A., Zhang, X., Chamness, G.C., et al. (2008). Intrinsic resis-tance of tumorigenic breast cancer cells to chemotherapy. J. Natl. Cancer Inst.100, 672–679.
Lindahl, T. (1993). Instability and decay of the primary structure of DNA. Nature362, 709–715.
Liu, S., Ginestier, C., Charafe-Jauffret, E., Foco, H., Kleer, C.G., Merajver, S.D.,Dontu, G., and Wicha, M.S. (2008). BRCA1 regulates human mammary stem/progenitor cell fate. Proc. Natl. Acad. Sci. USA 105, 1680–1685.
Loft, S., Kold-Jensen, T., Hjollund, N.H., Giwercman, A., Gyllemborg, J., Ernst,E., Olsen, J., Scheike, T., Poulsen, H.E., andBonde, J.P. (2003). Oxidative DNAdamage in human sperm influences time to pregnancy. Hum. Reprod. 18,1265–1272.
Lombard, D.B., Chua, K.F., Mostoslavsky, R., Franco, S., Gostissa, M., andAlt, F.W. (2005). DNA repair, genome stability, and aging. Cell 120, 497–512.
Look, A.T. (1997). Oncogenic transcription factors in the human acute leuke-mias. Science 278, 1059–1064.
Marshman, E., Booth, C., and Potten, C.S. (2002). The intestinal epithelial stemcell. Bioessays 24, 91–98.
McGill, G.G., Horstmann, M., Widlund, H.R., Du, J., Motyckova, G., Nishimura,E.K., Lin, Y.L., Ramaswamy, S., Avery, W., Ding, H.F., et al. (2002). Bcl2 regu-lation by the melanocyte master regulator Mitf modulates lineage survival andmelanoma cell viability. Cell 109, 707–718.
Meijne, E.I., van der Winden-van Groenewegen, R.J., Ploemacher, R.E., Vos,O., David, J.A., and Huiskamp, R. (1991). The effects of x-irradiation on hema-topoietic stem cell compartments in the mouse. Exp. Hematol. 19, 617–623.
Menezo, Y., Jr., Russo, G., Tosti, E., El Mouatassim, S., and Benkhalifa, M.(2007). Expression profile of genes coding for DNA repair in human oocytesusing pangenomic microarrays, with a special focus on ROS linked decays.J. Assist. Reprod. Genet. 24, 513–520.
Merritt, A.J., Potten, C.S., Kemp, C.J., Hickman, J.A., Balmain, A., Lane, D.P.,and Hall, P.A. (1994). The role of p53 in spontaneous and radiation-inducedapoptosis in the gastrointestinal tract of normal and p53-deficient mice.Cancer Res. 54, 614–617.
Merritt, A.J., Potten, C.S., Watson, A.J., Loh, D.Y., Nakayama, K., andHickman, J.A. (1995). Differential expression of bcl-2 in intestinal epithelia.Correlation with attenuation of apoptosis in colonic crypts and the incidenceof colonic neoplasia. J. Cell Sci. 108, 2261–2271.
Milyavsky, M., Gan, O.I., Trottier, M., Komosa, M., Tabach, O., Notta, F., Lech-man, E., Hermans, K.G., Eppert, K., Konovalova, Z., et al. (2010). A distinctiveDNA damage response in human hematopoietic stem cells reveals anapoptosis-independent role for p53 in self-renewal. Cell Stem Cell 7, 186–197.
Mohrin, M., Bourke, E., Alexander, D., Warr, M.R., Barry-Holson, K., Le Beau,M.M., Morrison, C.G., and Passegue, E. (2010). Hematopoietic stem cellquiescence promotes error-prone DNA repair and mutagenesis. Cell StemCell 7, 174–185.
Moreno, S.G., Dutrillaux, B., and Coffigny, H. (2001). Status of p53, p21,mdm2, pRb proteins, and DNA methylation in gonocytes of control andgamma-irradiated rats during testicular development. Biol. Reprod. 64,1422–1431.
Nijhof, J.G., van Pelt, C., Mulder, A.A., Mitchell, D.L., Mullenders, L.H., andde Gruijl, F.R. (2007). Epidermal stem and progenitor cells in murine epidermisaccumulate UV damage despite NER proficiency. Carcinogenesis 28,792–800.
28 Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc.
Nijnik, A., Woodbine, L., Marchetti, C., Dawson, S., Lambe, T., Liu, C.,Rodrigues, N.P., Crockford, T.L., Cabuy, E., Vindigni, A., et al. (2007). DNArepair is limiting for haematopoietic stem cells during ageing. Nature 447,686–690.
Nishimura, E.K., Granter, S.R., and Fisher, D.E. (2005). Mechanisms of hairgraying: incomplete melanocyte stem cell maintenance in the niche. Science307, 720–724.
Orford, K.W., and Scadden, D.T. (2008). Deconstructing stem cell self-renewal: genetic insights into cell-cycle regulation. Nat. Rev. Genet. 9,115–128.
Orkin, S.H., and Zon, L.I. (2008). Hematopoiesis: an evolving paradigm forstem cell biology. Cell 132, 631–644.
Passegue, E. (2005). Hematopoietic stem cells, leukemic stem cells andchronic myelogenous leukemia. Cell Cycle 4, 266–268.
Perrotti, D., Jamieson, C., Goldman, J., and Skorski, T. (2010). Chronicmyeloid leukemia: mechanisms of blastic transformation. J. Clin. Invest. 120,2254–2264.
Phillips, T.M., McBride, W.H., and Pajonk, F. (2006). The response of CD24(-/low)/CD44+ breast cancer-initiating cells to radiation. J. Natl. Cancer Inst.98, 1777–1785.
Potten, C.S. (2004). Radiation, the ideal cytotoxic agent for studying the cellbiology of tissues such as the small intestine. Radiat. Res. 161, 123–136.
Potten, C.S., and Grant, H.K. (1998). The relationship between ionizing radia-tion-induced apoptosis and stem cells in the small and large intestine. Br.J. Cancer 78, 993–1003.
Potten, C.S., Owen, G., and Booth, D. (2002). Intestinal stem cells protect theirgenome by selective segregation of template DNA strands. J. Cell Sci. 115,2381–2388.
Prados, M.D., and Levin, V. (2000). Biology and treatment of malignant glioma.Semin. Oncol. 27, 1–10.
Prasher, J.M., Lalai, A.S., Heijmans-Antonissen, C., Ploemacher, R.E.,Hoeijmakers, J.H., Touw, I.P., and Niedernhofer, L.J. (2005). Reduced hema-topoietic reserves in DNA interstrand crosslink repair-deficient Ercc1�/�mice. EMBO J. 24, 861–871.
Pritchard, D.M., Print, C., O’Reilly, L., Adams, J.M., Potten, C.S., andHickman,J.A. (2000). Bcl-w is an important determinant of damage-induced apoptosisin epithelia of small and large intestine. Oncogene 19, 3955–3959.
Qiu,W., Carson-Walter, E.B., Liu, H., Epperly, M., Greenberger, J.S., Zambetti,G.P., Zhang, L., and Yu, J. (2008). PUMA regulates intestinal progenitor cellradiosensitivity and gastrointestinal syndrome. Cell Stem Cell 2, 576–583.
Rachidi, W., Harfourche, G., Lemaitre, G., Amiot, F., Vaigot, P., and Martin,M.T. (2007). Sensing radiosensitivity of human epidermal stem cells. Radio-ther. Oncol. 83, 267–276.
Reese, J.S., Liu, L., and Gerson, S.L. (2003). Repopulating defect of mismatchrepair-deficient hematopoietic stem cells. Blood 102, 1626–1633.
Rice, K.N., and Jamieson, C.H. (2010). Molecular pathways to CML stem cells.Int. J. Hematol. 91, 748–752.
Richardson, B.E., and Lehmann, R. (2010). Mechanisms guiding primordialgerm cell migration: strategies from different organisms. Nat. Rev. Mol. CellBiol. 11, 37–49.
Robinson, K.C., and Fisher, D.E. (2009). Specification and loss of melanocytestem cells. Semin. Cell Dev. Biol. 20, 111–116.
Ropolo, M., Daga, A., Griffero, F., Foresta, M., Casartelli, G., Zunino, A., Poggi,A., Cappelli, E., Zona, G., Spaziante, R., et al. (2009). Comparative analysisof DNA repair in stem and nonstem glioma cell cultures. Mol. Cancer Res. 7,383–392.
Rossi, D.J., Bryder, D., Seita, J., Nussenzweig, A., Hoeijmakers, J., andWeissman, I.L. (2007). Deficiencies in DNA damage repair limit the functionof haematopoietic stem cells with age. Nature 447, 725–729.
Rossi, D.J., Jamieson, C.H., and Weissman, I.L. (2008). Stems cells and thepathways to aging and cancer. Cell 132, 681–696.
Cell Stem Cell
Review
Rottenberg, S., Jaspers, J.E., Kersbergen, A., van der Burg, E., Nygren, A.O.,Zander, S.A., Derksen, P.W., de Bruin, M., Zevenhoven, J., Lau, A., et al.(2008). High sensitivity of BRCA1-deficient mammary tumors to the PARPinhibitor AZD2281 alone and in combination with platinum drugs. Proc. Natl.Acad. Sci. USA 105, 17079–17084.
Ruzankina, Y., Pinzon-Guzman, C., Asare, A., Ong, T., Pontano, L., Cotsarelis,G., Zediak, V.P., Velez, M., Bhandoola, A., and Brown, E.J. (2007). Deletion ofthe developmentally essential gene ATR in adult mice leads to age-relatedphenotypes and stem cell loss. Cell Stem Cell 1, 113–126.
Saito, Y., Uchida, N., Tanaka, S., Suzuki, N., Tomizawa-Murasawa, M., Sone,A., Najima, Y., Takagi, S., Aoki, Y.,Wake, A., et al. (2010). Induction of cell cycleentry eliminates human leukemia stem cells in a mouse model of AML. Nat.Biotechnol. 28, 275–280.
Sancar, A., Lindsey-Boltz, L.A., Unsal-Kacmaz, K., and Linn, S. (2004). Molec-ular mechanisms of mammalian DNA repair and the DNA damage check-points. Annu. Rev. Biochem. 73, 39–85.
Sangiorgi, E., and Capecchi, M.R. (2008). Bmi1 is expressed in vivo in intestinalstem cells. Nat. Genet. 40, 915–920.
Sasaki, M., Lange, J., and Keeney, S. (2010). Genome destabilization byhomologous recombination in the germ line. Nat. Rev. Mol. Cell Biol. 11,182–195.
Schafer, M., Dutsch, S., auf dem Keller, U., Navid, F., Schwarz, A., Johnson,D.A., Johnson, J.A., and Werner, S. (2010). Nrf2 establishes a glutathione-mediated gradient of UVB cytoprotection in the epidermis. Genes Dev. 24,1045–1058.
Seita, J., Rossi, D.J., and Weissman, I.L. (2010). Differential DNA damageresponse in stem and progenitor cells. Cell Stem Cell 7, 145–147.
Shackleton, M., Vaillant, F., Simpson, K.J., Stingl, J., Smyth, G.K., Asselin-Labat, M.L., Wu, L., Lindeman, G.J., and Visvader, J.E. (2006). Generation ofa functional mammary gland from a single stem cell. Nature 439, 84–88.
Shafee, N., Smith, C.R., Wei, S., Kim, Y., Mills, G.B., Hortobagyi, G.N.,Stanbridge, E.J., and Lee, E.Y. (2008). Cancer stem cells contribute to cisplatinresistance in Brca1/p53-mediated mouse mammary tumors. Cancer Res. 68,3243–3250.
Singh, S.K., Hawkins, C., Clarke, I.D., Squire, J.A., Bayani, J., Hide, T., Henkel-man, R.M., Cusimano, M.D., and Dirks, P.B. (2004). Identification of humanbrain tumour initiating cells. Nature 432, 396–401.
Slupianek, A., Hoser, G., Majsterek, I., Bronisz, A., Malecki, M., Blasiak, J.,Fishel, R., and Skorski, T. (2002). Fusion tyrosine kinases induce drug resis-tance by stimulation of homology-dependent recombination repair, prolonga-tion of G(2)/M phase, and protection from apoptosis. Mol. Cell. Biol. 22,4189–4201.
Slupianek, A., Nowicki, M.O., Koptyra, M., and Skorski, T. (2006). BCR/ABLmodifies the kinetics and fidelity of DNA double-strand breaks repair inhematopoietic cells. DNA Repair (Amst.) 5, 243–250.
Song, S., and Lambert, P.F. (1999). Different responses of epidermal and hairfollicular cells to radiation correlate with distinct patterns of p53 and p21induction. Am. J. Pathol. 155, 1121–1127.
Sotiropoulou, P.A., Candi, A., Mascre, G., De Clercq, S., Youssef, K.K.,Lapouge, G., Dahl, E., Semeraro, C., Denecker, G., Marine, J.C., et al.(2010). Bcl-2 and accelerated DNA repair mediates resistance of hair folliclebulge stem cells to DNA-damage-induced cell death. Nat. Cell Biol. 12,572–582.
Squatrito, M., Brennan, C.W., Helmy, K., Huse, J.T., Petrini, J.H., and Holland,E.C. (2010). Loss of ATM/Chk2/p53 pathway components accelerates tumordevelopment and contributes to radiation resistance in gliomas. Cancer Cell18, 619–629.
Stingl, J., Eirew, P., Ricketson, I., Shackleton, M., Vaillant, F., Choi, D., Li, H.I.,and Eaves, C.J. (2006). Purification and unique properties of mammary epithe-lial stem cells. Nature 439, 993–997.
Suh, E.K., Yang, A., Kettenbach, A., Bamberger, C., Michaelis, A.H., Zhu, Z.,Elvin, J.A., Bronson, R.T., Crum, C.P., and McKeon, F. (2006). p63 protectsthe female germ line during meiotic arrest. Nature 444, 624–628.
Tagscherer, K.E., Fassl, A., Campos, B., Farhadi, M., Kraemer, A., Bock, B.C.,Macher-Goeppinger, S., Radlwimmer, B., Wiestler, O.D., Herold-Mende, C.,et al. (2008). Apoptosis-based treatment of glioblastomas with ABT-737,a novel small molecule inhibitor of Bcl-2 family proteins. Oncogene 27,6646–6656.
Takubo, K., Ohmura, M., Azuma, M., Nagamatsu, G., Yamada, W., Arai, F.,Hirao, A., and Suda, T. (2008). Stem cell defects in ATM-deficient undifferen-tiated spermatogonia through DNA damage-induced cell-cycle arrest. CellStem Cell 2, 170–182.
Tani, H., Morris, R.J., and Kaur, P. (2000). Enrichment for murine keratinocytestem cells based on cell surface phenotype. Proc. Natl. Acad. Sci. USA 97,10960–10965.
Viale, A., De Franco, F., Orleth, A., Cambiaghi, V., Giuliani, V., Bossi, D.,Ronchini, C., Ronzoni, S., Muradore, I., Monestiroli, S., et al. (2009). Cell-cyclerestriction limits DNA damage and maintains self-renewal of leukaemia stemcells. Nature 457, 51–56.
Visvader, J.E. (2009). Keeping abreast of themammary epithelial hierarchy andbreast tumorigenesis. Genes Dev. 23, 2563–2577.
Weissman, I.L. (2000). Stem cells: units of development, units of regeneration,and units in evolution. Cell 100, 157–168.
Wilson, J.W., Pritchard, D.M., Hickman, J.A., and Potten, C.S. (1998). Radia-tion-induced p53 and p21WAF-1/CIP1 expression in the murine intestinalepithelium: apoptosis and cell cycle arrest. Am. J. Pathol. 153, 899–909.
Woodward, W.A., Chen, M.S., Behbod, F., Alfaro, M.P., Buchholz, T.A., andRosen, J.M. (2007). WNT/beta-catenin mediates radiation resistance ofmousemammary progenitor cells. Proc. Natl. Acad. Sci. USA 104, 618–623.
Zhang, M., Behbod, F., Atkinson, R.L., Landis, M.D., Kittrell, F., Edwards, D.,Medina, D., Tsimelzon, A., Hilsenbeck, S., Green, J.E., et al. (2008). Identifica-tion of tumor-initiating cells in a p53-null mouse model of breast cancer.Cancer Res. 68, 4674–4682.
Zhang, M., Atkinson, R.L., and Rosen, J.M. (2010). Selective targeting ofradiation-resistant tumor-initiating cells. Proc. Natl. Acad. Sci. USA 107,3522–3527.
Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc. 29
Cell Stem Cell
Article
A Human iPSC Model of Hutchinson Gilford ProgeriaReveals Vascular Smooth Muscleand Mesenchymal Stem Cell DefectsJinqiu Zhang,1 Qizhou Lian,3,4 Guili Zhu,1 Fan Zhou,1 Lin Sui,1 Cindy Tan,1 Rafidah Abdul Mutalif,2 Raju Navasankari,2
Yuelin Zhang,3 Hung-Fat Tse,3 Colin L. Stewart,2,* and Alan Colman1,*1Stem Cell Disease Models2Developmental and Regenerative BiologyA*STAR Institute of Medical Biology, Singapore 138648, Singapore3Cardiology Division, Department of Medicine4Eye Institute, Li Ka Shing Faculty of Medicine
University of Hong Kong, Pokfulam, Hong Kong, China*Correspondence: colin.stewart@imb.a-star.edu.sg (C.L.S.), alan.colman@imb.a-star.edu.sg (A.C.)
DOI 10.1016/j.stem.2010.12.002
SUMMARY
The segmental premature aging disease Hutchinson-Gilford Progeria syndrome (HGPS) is caused bya truncated and farnesylated form of Lamin A calledprogerin. HGPS affects mesenchymal lineages,including the skeletal system, dermis, and vascularsmooth muscle (VSMC). To understand the under-lying molecular pathology of HGPS, we derivedinduced pluripotent stem cells (iPSCs) from HGPSdermal fibroblasts. The iPSCs were differentiatedinto neural progenitors, endothelial cells, fibroblasts,VSMCs, and mesenchymal stem cells (MSCs). Pro-gerin levels were highest in MSCs, VSMCs, and fibro-blasts, in that order, with these lineages displayingincreased DNA damage, nuclear abnormalities, andHGPS-VSMC accumulating numerous calponin-staining inclusion bodies. Both HGPS-MSC and-VSMC viability was compromised by stress andhypoxia in vitro and in vivo (MSC). Because MSCsreside in low oxygen niches in vivo, we proposethat, in HGPS, this causes additional depletion oftheMSC pool responsible for replacing differentiatedcells lost to progerin toxicity.
INTRODUCTION
Hutchinson-Gilford Progeria syndrome (HGPS) is a rare congen-
ital disease that may cause some aspects of premature aging in
children (Hennekam, 2006). Afflicted individuals generally die in
their early teens due to myocardial infarction or stroke, but it is
the wizened facial features and wasted bodies that have made
this harrowing condition familiar to the population at large. The
disease progression displays many symptoms of normal aging,
such as severe growth retardation, alopecia, loss of subcuta-
neous fat, and progressive atherosclerosis, although other
symptoms associated with aging such as neural degeneration,
diabetes, malignancies, and cataracts are absent (Ackerman
and Gilbert-Barness, 2002; Gordon et al., 2007; Merideth et al.,
2008). This disease, which seems to affect mainly mesenchymal
lineages, is caused by an autosomal dominant mutation in the
LMNA gene (De Sandre-Giovannoli et al., 2003; Burke and Stew-
art, 2006; Capell and Collins, 2006). The most common (in 80%–
90% of cases) mutation is a C-T transition at position 1824 in
exon 11 that creates an efficient alternative splice donor site.
This leads to the production of a truncated lamin A protein
(progerin) with an internal deletion of 50 amino acids in the
C-terminal globular domain. As a result of this mutation, pro-
gerin, but not mature lamin A, retains a C-terminal farnesyl tail
that is normally only transiently present in the Lamin A precursor.
Farnesyl retention is widely thought to underlie the intracellular
disruption associated with progerin, although the exact roles of
the farnesyl group and the deletion in the etiology of the disease
are controversial (Yang et al., 2008).
The cell type-specific pathologies in HGPS have been attrib-
uted to a variety of causes, including progerin-mediated stem
cell pool exhaustion (Halaschek-Wiener and Brooks-Wilson,
2007), mesenchymal lineage differentiation defects (Scaffidi
and Misteli, 2008), a diminished DNA-damage-repair response
(Musich and Zou, 2009), and nuclear fragility in mechanically
stressed cells such as cardiomyocytes (Verstraeten et al.,
2008). Interestingly, the same aberrant splicing event may also
occur at much lower levels in normal cells (Scaffidi and Misteli,
2006). Although low progerin RNA levels may not increase with
age, several reports have suggested that progerin protein levels
do increase, (McClintock et al., 2007; Scaffidi and Misteli, 2006),
possibly reflecting a low turnover of the protein or an age-related
inability to remove cells with high progerin loads. This has led to
speculation that studies on progeria may provide insight into the
normal human aging process.
Due to the rarity and juvenile mortality of this disease, biopsy
and autopsy analysis has been limited, although pronounced
vascular smooth muscle loss and artherosclerosis appear to
be critical factors contributing to the death of patients (Stehbens
et al., 2001; Olive et al., 2010). Much of the information relating to
the pathophysiology of HGPS has come from studies on patient-
derived skin fibroblasts, wild-type and mutant lamin A overex-
pression in established cell lines (Cao et al., 2007; Goldman
et al., 2004), and the development of various mouse models in
Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc. 31
N1
N2Phase contrast
APG1
PG2
DA LMNA LMNAPI PROGERIN
66%
17%
14%
p26
52%
p15
21%
29%
C
0
50
100
150
iP
S c
olo
ny
n
um
be
r (/1
05 c
ells
)
N1 N2 N2N1PG2PG1 PG1 PG2
p15-20 p25-30
p15-20 p25-30
* *
B
GAPDH
N
AP
OCT4
NANOG
SOX2
SSEA4
TRA-1-80
1 N2 PG1 PG2
LMNAprogerinLMNC
Fib-control (N1, N2)
Fib-HGPS (PG1, PG2)
iPS-control (N1-iPS-1, N2-iPS-1)
iPS-HGPS (PG1-iPS-1, PG2-iPS-1)
HES (HES3, H9)0.20 0.10 0.05 00.15
E
11%
3%
D
Figure 1. Generation of Patient-Specific iPSCs
(A) HGPS patient fibroblasts AG11498 (PG1) and AG06297 (PG2) were obtained from the Coriell Institute, and two unaffected HGPS parental fibrolast lines,
AG03512 (N1) and AG06299 (N2), were used as controls. Immunofluorescence microscopy of fibroblast cells using an antibody specifically recognizing mutant
LMNA (progerin) shows specific expression of progerin in HGPS patient fibroblasts, but not controls. Immunostaining with JOL2 antibody recognizing human
LMNA/C (right panels) shows increasing nuclear deformation (arrows) in HGPS fibroblasts undergoing extended passaging from p15 to p26. The percentage
of cells showing aberrant nuclei is indicated for respective passages. Scale bar, 20 mm.
(B) Progerin expression in donor fibroblasts. Western blot analysis of fibroblast lysates using the JOL2 antibody recognizes both human LMNA and -C.
Cell Stem Cell
Progeria iPSC Model Reveals MSC and VSMC Defects
32 Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc.
Cell Stem Cell
Progeria iPSC Model Reveals MSC and VSMC Defects
which the lamin A gene is deleted, mutated, and/or overex-
pressed (Mounkes et al., 2003; Yang et al., 2005; Varga et al.,
2006; Sagelius et al., 2008; Hernandez et al., 2010). The use of
mouse models has been particularly informative; however, no
one mouse model recapitulates all the symptoms seen in the
human disease.
Recently, a number of human disease models have been
established by the transcription-factor-mediated reprogram-
ming of somatic cells taken from patients with Lou Gehrig’s
disease (Dimos et al., 2008), spinal muscular atrophy (Ebert
et al., 2009), familial dysautonomia (Lee et al., 2009), and dysker-
atosis congenita (Agarwal et al., 2010). In all cases, the reprog-
rammed cells (induced pluripotent stem cells [iPSCs]) were
used to derive cell types that, in vivo, display a distinctive
disease phenotype. The underlying hope behind these studies
is that a disease pathology will emerge in a relatively short time
(compared to disease progression in vivo) and generate insight
into early disease pathophysiology, as well as providing cell
types for drug screening and discovery.
Here, we describe an iPSC model of HGPS. iPSC lines were
made from patient-derived fibroblasts and differentiated into
mesenchymal and nonmesenchymal lineages to analyze the
impact of progerin on the functional properties of the different
cell types. We find that progerin levels are highest in mesen-
chymal stem cells, VSMCs, and fibroblasts, and lowest in the
neural progenitors. Progerin expressing VSMCs and MSCs,
but not controls, are sensitive to hypoxia, and HGPS-MSCs fail
to mediate circulatory restoration in a murine hind limb recovery
model. We speculate that one significant cause of progeria
pathology is a shortage of MSCs needed for tissue replacement,
and this shortage is exacerbated by a loss of specific differenti-
ated types due to progerin. Our experiments support the hypoth-
esis that the MSC pool becomes exhausted due to replicative
overload in HGPS patients (Halaschek-Wiener and Brooks-Wil-
son, 2007), which is compounded by a parallel depletion due
to progerin-induced sensitivity of the stem cells to their niche
conditions.
RESULTS
Generation of HGPS-iPSCsSkin-derived fibroblast cultures from two HGPS patients
(AG11498 [PG1], AG06297 [PG2]) and two HGPS parents
(AG03512 [N1], AG06299 [N2]) were obtained from the Coriell
Institute. Both patient genomes contain the typical C-T mutation
of LMNA gene at position 1824 of exon 11. The mutant protein,
progerin, was detected exclusively in HGPS fibroblasts by
western analysis and immunofluorescence using a progerin-
specific antibody (Figure 1). Nuclear membrane deformation
(C) Reduced reprogramming efficiency of HGPS fibroblasts. Efficiencies were c
transduction of 105 HGPS (PG1, PG2) or control (N1, N2) fibroblasts by retroviru
passage (p25–p30) HGPS fibroblasts.
(D) Immunohistology of iPSC clones. Phase contrast (row 1) with alkaline phospha
rows) for the following pluripotent markers: OCT4, SOX2, NANOG, SSEA-4, and
(E) Cluster analysis ofmicroarray data fromHGPS fibroblast (PG1, PG2), control fib
N2-iPS-1) and human ESC (HES3, H9) RNA. The Pearson correlation coefficient (
expression level of all transcripts. Hierarchical cluster analysis was carried out w
See also Figures S1 and S2 for characterization of iPSCs.
‘‘blebbing’’ was seen in approximately 20%–30% HGPS fibro-
blasts at p15-20, in contrast to 3%–11% in normal fibroblast cells
at similar passage. This increased to 60%–70% in HGPS fibro-
blasts by p25-30, withmoremodest increases seen in the control
fibroblasts. PG1, PG2, and N2 fibroblasts displayed normal
karyotypes. However, in N1 fibroblasts, 55% of cells had an
abnormal karyotype with trisomy 7. We reprogrammed the four
fibroblast lines by retrovirus infection of cells at p15-20 using
the ‘‘Yamanaka’’ factor cocktail: OCT4, SOX2, KLF4, and C-
MYC. Infected fibroblasts were cultured on amurine feeder layer
with 0.5 mM valproic acid in human ESCmedium. iPSC colonies
were observed at 2 to 3weeks and picked between 3 to 4 weeks.
Though the efficiency of reprogramming ofHGPSfibroblastswas
4-fold lower than parental control fibroblasts (Figure 1C), all the
colonies picked were expanded and displayed morphologies
indistinguishable from human ESCs (Figure 1D): no colonies
were obtained from the two HGPS fibroblast cultures at late
passage (p26). We attribute this failure to the onset of senes-
cence observed in these cultures beginning at p22 since efficient
iPSC generation is associated with a cell’s proliferative potential
(Hanna et al., 2009). Five to ten colonieswere picked to represent
each patient or control, and for the N1-iPSC, about 50% of the
colonies displayed a normal karyotype.
Characterization of iPSCsExogenous expression of the four reprogramming factors was
screened for by RT-PCR in all the iPSC clones. No transgenic
transcripts were found in any clone (Figure S1A available online).
Two clones from each patient or control (given the suffix iPSC-1
or iPSC-2) each with a normal karyotype (Figure S1B) were
selected for further characterization. DNA sequencing revealed
HGPS-iPSC but no control clones had the C-T mutation of the
LMNA gene (Figure S1C). All clones express the pluripotent
markers OCT4, SOX2, NANOG, SSEA4, and TRA1-80, as deter-
mined by immunocytochemistry (Figure 1D). In addition, all iPSC
lines showed reactivation of three endogenous pluripotency-
related genes with similar level of expression as seen in hESCs
(Figure S2A). As expected for hiPSCs, theOCT4 promoter region
in all iPSCs (and hESC3) was hypomethylated in contrast to its
hypermethylated state in the parental fibroblasts (Figure S2B).
To test the pluripotency of iPSCs, teratoma assays were per-
formed in SCID mice. All the clones developed teratomas
comprised of tissues from all three germ layers (Figure S2C).
The transcriptomes of two independently derived sister clones
from each hiPSC genotype were compared to that in two
different hESC lines (hESC3 and H9) by microarray analysis on
24,000 gene Illumina chips. Clustering analysis revealed a high
degree of similarity (r = 0.99) between the reprogrammed
HGPS-iPSCs (PG1-iPSCs, PG2-iPSCs) and parental control
alculated as number of alkaline phosphatase positive colonies at day 21 after
ses carrying OSKM. *p < 0.01, n = 3. No iPS colonies were obtained for late
tase (AP) staining (inset) and immunofluorescence staining of iPSCs (remaining
Tra-1-81. Two iPS clones were analyzed for each line.
roblast (N1, N2), HGPS-iPSC (PG1-iPS-1, PG2-iPS-1), control iPSC (N1-iPS-1,
PCC) was calculated for each pair of samples (see Table S1) using the relative
ith PCC as the distance measurement using Illumina GenomeStudio software.
Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc. 33
A
B
LMNA
iPSC
SOX2LMNA
HES3
LMNA
iPSCiPSC
whole colony center
whole colony center0.0
0.5
1.0
1.5
LMNA
progerin
fo
ld d
ifferen
ce
DAPI MERGE iPSCiPSC
Figure 2. Expression of LMNA and Progerin Is Suppressed in iPSCs
(A) Immunofluorescence staining of LMNA/C on human ESC line HES3 and
HGPS iPSCs (PG1) with JOL2 antibody shows LMNA/C expression in differen-
tiated cells lining the edge of colonies. Costaining of LMNA/C with SOX2
shows suppressed expression of LMNA/C in pluripotent iPSCs. Scale bar,
50 mm.
(B) qPCR of LMNA and progerin expression in HGPS-iPSCs after colony
dissection. Most expression was seen at the colony edges where differentia-
tion was occurring.
Cell Stem Cell
Progeria iPSC Model Reveals MSC and VSMC Defects
iPSC (N1-iPSCs, N2-iPSCs) that clustered with the hESC group
(r = 0.99). The three groups were distant from the HGPS and
control fibroblasts (r = 0.82) (Figure 1E, Table S1). The above
data indicate that, despite the presence of progerin in the
nucleus, somatic cells from HGPS patients can be reprog-
rammed into iPSCs with characteristics that are highly similar
to embryo-derived human ESCs.
LaminA/C and Progerin Expression Are Silencedby ReprogramingLamins A and C are transcribed from the same promoter and
share the first 566 amino acids. Undifferentiated mESCs and
hESCs do not express LMNA transcripts (Constantinescu
et al., 2006). Immunofluorescent analysis of the iPSC colonies re-
vealed that LMNA proteins were expressed in differentiated cells
at the edge of the colony, whereas the bulk of the colony stained
for the pluripotent marker SOX2 (Figure 2A). No specific staining
for progerin was detected in any region. Quantitative RT-PCR
analysis confirmed the presence of LMNA transcripts but
showed progerin transcripts present at lower levels. Colony
dissection indicated that the majority of these transcripts were
at the rim of the colonies (Figure 2B). In addition, RNA levels of
LMNA and progerin expression in intact HGPS-iPSC colonies
were less than 10% of those present in fibroblasts (Figure 3A).
These data demonstrate that LMNA expression in patient and
control fibroblasts was suppressed by reprogramming.
Differentiation of iPSCsClinical and autopsy data indicate that in HGPS, the main line-
ages affected are mesenchymal in origin with neural lineages
seemingly unaffected (McClintock et al., 2006; Merideth et al.,
2008). Since tissue samples from HGPS patients are extremely
rare, we used existing or modified protocols to derive MSC,
neural, and other candidate lineages for further examination.
LMNA and Progerin Are Re-expressed during iPSCDifferentiationDifferentiation of the HGPS-iPSC clones always resulted in
expression of lamins A, C, and progerin. We quantified lamin A
and C expression by both western and real-time RT-PCR anal-
ysis (Figures 3A and 3B). Progerin levels were highest in iPSC-
derived mesenchymal stem cells (MSCs), followed by vascular
smooth muscle cells (VSMCs), fibroblast, and endothelial cells,
in that order. Neural progenitors consistently showed the lowest
levels. Interestingly, we observed a 3- to 5-fold increase in
progerin levels over prolonged culture for iPSC-MSCs (Fig-
ure 3C), whereas in only one of the two HGPS-iPSC fibroblast
lines was an increase noted (data not shown). A similar inconsis-
tency in progerin accumulation, with passage number in patient
fibroblasts, was previously reported (Goldman et al., 2004;
Verstraeten et al., 2008). Overall, our observations are consistent
with skin biopsy data in finding significant presence of progerin in
endothelial and fibroblast lineages.
Impact of the HGPS Mutation on Fibroblast, NeuralProgenitor, Endothelial, and MSC Functions underNormal Culture ConditionsExtended culture of HGPS and control fibroblasts revealed an
accumulation of abnormal nuclear morphologies (Bridger and
34 Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc.
Kill, 2004). We obtained fibroblasts by FACS sorting of Thy1+
cells formed during differentiation of the iPSC lines. The purified
cells expressed prolyl 4-hydroxylase and displayed a similar
surface marker profile to parental fibroblasts—high in CD29,
CD44, and CD90 (Thy1) and low in CD106 (Figure 4A and Fig-
ure S3A). As with the HGPS-iPSC fibroblasts, we found
0
10
20
30
1 2 3 4 5 6 7 8 9 10 11 12
LMNAprogerin
B
LMNAprogerinLMNC
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
LMNAprogerin
A fibroblast HES fibroblast VSMC neural endothelial MSCiPSC
GAPDH
0
10
20
30
LMNAprogerinLMNC
LMNAprogerinLMNC
GAPDH
PG2-iPS-MSC PG1-iPS-MSC
p3 p5p13 p20C
Figure 3. Lamin A/C Expression in iPSC-Derived Fibroblasts, Endothelial Cells, Neural Progenitor Cells, VSMCs, and MSCs
(A) Relative Lamin A and progerin expression in iPSC differentiated samples by RT-PCR using Taqman probes. LMNA and progerin expression were each set as
100% for PG1 samples. Commercially available cell lines were used as references for each lineage: (1) N1 fibroblast p10, (2) PG1 fibroblast p10, (3) HESC3 p83, (4)
N1-iPSC-1p11, (5) PG1-iPSC-1 p13, (6)N1-iPSC-fibroblast p10, (7) PG1-iPSC-fibroblast p10, (8) N1-iPSC-VSMCp3, (9) PG1-iPSC-VSMCp3, (10) ReNcell human
fetal neural stem cell line, (11) N1-iPSC-neural p5, (12) PG1-iPSC-neural p5, (13) human umbilical vein endothelial cell line HUVEC p5, (14) N1-iPSC-endo p5, (15)
PG1-iPSC-endo p6, (16) human adult bone-marrow MSC p5, (17) N1-iPSC-MSC p6, and (18) PG1-iPSC-MSC p6. Data were normalized to GAPDH expression.
(B) Quantitative analysis of LMNA and progerin by western blot. Twenty micrograms of total protein extracts were loaded, and protein expression was quan-
titatively analyzed with the Odyssey infrared imaging system. Values represent relative densitometry normalized to GAPDH. The passage number of each
sample is indicated. The picture is combination of three separate blots which contained N1 and PG1 fibroblast lysates in each blot as the internal control.
(C) Accumulation of progerin in HGPS-iPSC-MSCs during extended passaging. Quantitative western blot shows progerin protein accumulation in two different
patient-derived HGPS-iPSC-MSCs.
Cell Stem Cell
Progeria iPSC Model Reveals MSC and VSMC Defects
Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc. 35
F
N1-
iPS-
MSC
p17
PG1-
iPS-
MSC
p17
PG1-
iPS-
MSC
p10
LAP2 PROGERIN DAPI MERGE LMNA H2AX/DAPI
28%
37%
86%
11%
4%
34%66%
3%
15%
P4H/DAPI H2AX/DAPI
7.5%
6.4%
12.2%
10.7%
3.5%
8.8%
32%
45%
9%
N1-iPS-fib
PG1-iPS-fib
PG2-iPS-fib
N2-iPS-fib
D7-FIB/DAPI
N1 (03512)
LMNA
8.2%
10.6%
30.5%
28.5%
5.2%
PROGERIN LAP2 MERGE/DAPI
15.2%
BTuJ1/nestinDAPI
-
-
Progerin
DAPI
A
Vimentin H2AX/DAPI
N1-iPS-neural
PG1-iPS-neural
H9-neural
Nestin
C N1-iPS-endo PG1-iPS-endo
CD31/DAPI
VE-CADHERIN/DAPI
LMNA
2.1% 3.5%
PROGERIN/DAPI
H2AX/DAPI
1.2% 2.8%
LDL/DAPI
matrigel
D BM-MSC N1-iPS-MSC PG1-iPS-MSCN2-iPS-MSC PG2-iPS-MSC
Oil
red
Aliz
arin
red
Safra
nin-
O
0.0
0.5
1.0
1.5
osteogenesisadipogenesis
OD
500
/405
Ki67
7.9%
6.5%
LMNA
5.4%
E
Cell Stem Cell
Progeria iPSC Model Reveals MSC and VSMC Defects
36 Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc.
Cell Stem Cell
Progeria iPSC Model Reveals MSC and VSMC Defects
increased nuclear dysmorphology, DNA damage (% nuclei with
3 or more foci staining for g-H2AX), and mislocalization of the
nuclear protein, LAP2, in progerin-expressing lines (Figure 4A).
Progerin accumulation was heterogeneous in the HGPS popula-
tions, and loss of LAP2 from the nuclei was more apparent in
nuclei displaying higher progerin levels and more pronounced
dysmorphology (Figure 4A, arrows). These results are similar to
those for HGPS patient fibroblasts after extended culture (Scaf-
fidi and Misteli, 2006).
Neither increased DNA-damage foci nor nuclear dysmorphol-
ogies were observed in iPSC-derived neural progenitors and
endothelial cells (Figures 4B and 4C). Neural progenitors were
propagated by neurosphere culture and characterized by nestin,
vimentin, and Ki67 staining (Figure 4B). Endothelial cultures were
CD31+ve/CD43�ve, stained for VE-cadherin (Figure 4C) and were
also positive for CD31, VE-cadherin, C-KIT, and KDR transcripts
(Figure S3B). Neural progenitors made from different hESC or
iPSC lines at passages 7–10 were differentiated mainly into
Tuj-1+ve neurons (Figure 4B), although some GFAP-staining glial
cells were detected (data not shown). No differences were
observed between HGPS and normal neurospheres in growth
or neuronal differentiation. Both HGPS and control iPSC-derived
endothelial cells formed lattice-like vessel structures on Matrigel
(Figure 4C), a characteristic feature of endothelial cells. HGPS-
iPSC endothelial cells also displayed a normal lipid uptake func-
tion (Figure 4C).
MSCs were prepared from iPSC-derived EBs and were char-
acterized by FACS analysis as negative or low for the surface
markers of CD24, CD31, and CD34 and positive for CD29,
CD44, CD73, CD105, and CD166 (Figure S3C). Cluster analysis
of microarray data indicated that the iPSC-MSCs were more
closely related to hESC-derived and fetal bone marrow-derived
MSCs than adult bone marrow MSCs (Figure S3D).
MSCs differentiate into osteogenic, chondrogenic, and adipo-
genic lineages in vitro when provided with appropriate growth
conditions. All three lineages were formed from all the HGPS-
and control-iPSC-MSC lines (Figures 4D and 4E). Under
extended culture, the HGPS-MSCs showed considerable
nuclear lobulation that correlated with increasing levels of pro-
Figure 4. Characterization of iPSC-Derived Fibroblast-like Cells, Endo
(A) Differentiation of iPSCs into fibroblast-like cells. Left panel: phase contrast pic
bodies against P4H, fibroblasts (D7-FIB), LMNA and H2AX. Right panel, Progeri
zation of LAP2 and progerin by immunostaining showed increased number of bleb
The percentage of cells showing aberrant phenotypes is indicated. The parental
(B) Characterization of iPS-derived neural progenitor cells. Immunostaining show
progenitor markers Ki67, nestin and vimentin and were able to differentiate into ne
Progerin and H2AX staining (green fluorescence) is rarely detected. The percenta
color. Scale bar, 50 mm.
(C) Differentiation of iPSCs into endothelial cells. Phase contrast pictures showing
specific markers, CD31 and VE-Cadherin. The percentage of cells showing aberra
of Dil-AC-LDL uptake of iPSC-derived endothelial cells. Scale bar, 20 mm.
(D) Differentiation of MSCs into bone, cartilage, and adipocyte. Bone marrow MS
teogenesis, adipogenesis, and chondrogenesis. Oil red, Safranin-O, and Alizarin
glycosaminoglycans (cartilage), and mineralization (bone), respectively.
(E) Semiquantitative analysis of adipogenesis and osteogenesis of MSCs. Oil red O
ing was semiquantitatively analyzed at 405 nm using a plate reader. Values repr
(F) Progerin accumulation leads to aberrant expression of LAP2 in late passageMS
number of blebbing nuclei associated with mislocalization of LAP2 in HGPS-MSC
is associated with mislocalization of LAP2. The percentage of cells showing abn
See also Figure S3 and Tables S1–S3 for further characterization of iPSC-derive
gerin. The high percentage of nuclear malformation was accom-
panied by mislocalization of LAP2 (Figure 4F). A significant
increase in nuclei-containing DNA damage foci was also
observed in late passage HGPS-MSCs (Figure 4F).
VSMC were obtained from iPSC derived MSCs by treatment
with a combination of SPC and TGFb1 (Jeon et al., 2006). After
3 weeks of induction, 50%–60% cells showed specific VSMC
marker expression of a-smooth muscle actin, calponin 1, and
smooth muscle myosin heavy chain (Figure 5A) with the VSMC
lineage marker transcripts being confirmed by RT-PCR (Fig-
ure 5B). The VSMCs displayed a characteristic spindle-like
morphology and were induced to contract by carbachol admin-
istration (Figure 5C), supporting their VSMC identity. Like fibro-
blasts and MSCs, the HGPS-VSMCs displayed nuclear defor-
mations, LAP2 mislocalization, and increased DNA damage on
culture (Figure 5D). We also noticed that many, though not all,
the calponin 1-staining HGPS-VSMC cells had vesicular-like cal-
ponin 1 inclusions (Figure 5A), which were absent in control and
N-VSMCs. Although calponin decorates the filamentous actin
cytoskeleton, we observed no costaining of actin with these
bodies.
Patient iPSC-Derived VSMCs and MSCs ShowFunctional Defects under StressOf the five lineages we derived, two (HGPS-iPSC-derived
VSMCs and MSCs) seem most adversely affected by progres-
sive culture under normoxic conditions. This might reflect the
higher levels of progerin these cell types appear to accumulate,
although fibroblasts levels are only slightly lower (Figure 3B). We
investigated whether other functional properties of these cells
were impaired.
VSMCs were subjected to three different conditions of stress:
hypoxia with substratum deprivation, hypoxia alone, and recur-
rent electrical stimulation. When HGPS-VSMCs were immersed
under mineral oil for 4 to 5 hr (substratum deprivation/hypoxia),
their survival was more than halved (Figure 6A). Hypoxia (2%
O2) for 3 days also increased senescence in HGPS-VSMCs,
shown by b-galactosidase staining (5.1% to 38.5%, Figure 6B).
To mimic the mechanical stresses endured by VSMCs in vivo
thelial Cells, Neural Progenitor Cells, and MSCs
tures and immunofluorescence staining of iPSC derived fibroblasts using anti-
n accumulation leads to aberrant expression of LAP2 in fibroblasts. Co-locali-
bed nuclei associated with mis-localization of LAP2 (arrows) in iPSC-fibroblast.
fibroblast N1 (03512) was used as a control. Scale bar, 15 mm.
ed that both HGPS and control iPSC-derived neurospheres expressed neural
urons with expression of bIII-tubulin (TuJ1) upon withdrawal of growth factors.
ge of cells showing aberrant phenotypes is indicated. DAPI stains nuclei a blue
the endothelial morphology and immunofluorescence staining with endothelial
nt phenotypes and DNA damage is indicated. Lower panels show live imaging
Cs, HGPS-iPSC-MSCs, and control-iPSC-MSCs were induced to undergo os-
red were used for staining of lipid oil droplet (adipocyte), proteogylcans and
was eluted with isopropanol and measured OD at l500 nm. Alizarin red stain-
esent mean ± SEM from three replicates.
Cs. Colocalization of LAP2 and progerin by immunostaining showed increased
s from p10 to p17. Nuclear blebbing with strong expression of progerin (arrows)
ormal phenotypes is indicated. Scale bar, 15 mm.
d lineages.
Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc. 37
B
SM -22
caldesmon
C
A
SM-MHC/DAPI
N1-VSMC N2-VSMC PG1-VSMC PG2-VSMC BM-MSC-VSMC
α-SMA/DAPI
Calponin/DAPI
phalloidin
D
-SMA
GAPDH
calponin
Smoothelin-B
EGREM2PAL NIREGORP IPAD/ANMLIPAD H2AX/DAPI
1.8%
2.6%
9.8%
11.2%
5.4%
7.6%
23.8%
19.6%
18.3%
22.5%
44.6%
39.7%
α
α
Figure 5. Characterization of iPSC-Derived Vascular Smooth Muscle Cells
(A) Immunostaining of vascular smooth muscle cells (VSMCs) showed expression of a-smooth muscle actin (a-SMA), calponin, and VSMC exclusive marker
smooth muscle myosin heavy chain (SM-MHC). Lower panel shows colocalization of F-actin (phalloidin) with calponin. Scale bar, 15 mm.
(B) RT-PCR analysis of VSMC-specific contractile protein transcripts, a-SMA, calponin, smoothelin-B, h-caldesmon, and SMa-22. Neural cell line ReNcell was
used as control.
(C) Induction of contraction by carbachol treatment. Phase contrast image shows contraction of VSMCs under carbachol (1 3 10�5M) treatment for 1 hr (right
panel).
Cell Stem Cell
Progeria iPSC Model Reveals MSC and VSMC Defects
38 Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc.
Cell Stem Cell
Progeria iPSC Model Reveals MSC and VSMC Defects
due to pulsatile circulation, repeated contraction of VSMCs was
chronically induced with electrical pulses (40V/cm, 1Hz) for
3 days. We observed an increase in nuclear dysmorphology
and accelerated senescence in both HGPS-VSMC and
N-VSMCs, although the effect was significantly greater with
the HGPS-VSMCs (Figure 6C).
Although the exact roles of MSCs in vivo are unclear, human
MSCs from bone marrow (Li et al., 2009) or from human embry-
onic stem cells (Lian et al., 2010) significantly improve vascular
circulation after their transplantation into the ischemic hind limbs
of immunocompromised mice. As a measure of their capacity to
effect such improvement, we tested parental control (N-MSC)-
and HGPS-iPSC-derived MSCs (HGPS-MSC) to protect against
ischemia in this mouse model. Mouse limb ischemia was
induced by ligation of the femoral artery and its branches in the
left hind-limb of SCID mice. The MSCs were transplanted by
intramuscular injection into the left hind-limb immediately after
ligation. After 28 days of transplantation, the culture medium-in-
jected group (vehicle group) displayed severe necrosis of the
ischemic limbs leading to limb loss (87.5%; Figure 7A). In the
15 mice given an N-MSC injection, limb loss was only 20%.
Most of the ischemic limbswere fully rescued (60%) or displayed
moderate necrosis from knee to toe. However, in the 15 mice
given HGPS-MSCs, rescue of ischemic limb occurred in only
one mouse (6.7%). Most mice suffered limb loss (60%), which
is significantly different from that seen with N-MSCs (Figure 7A).
Histological analysis of the adductor muscle of the ‘‘rescued’’
limbs revealed extensive muscle degeneration and pronounced
interstitial fibrosis in the HGPS–MSC group. The N-MSCmice, in
contrast, exhibited significantly less fibrosis and more muscle
regeneration (Figure S4).
The simplest explanation for limb salvage was a restoration of
blood flow following transplantation. To monitor blood flow after
MSC transplantation, laser Doppler imaging was performed at
days 0, 14, and 28 after surgery. The N-MSC-treated animals
had significantly improved blood flow in contrast to the HGPS
and culture medium groups (p < 0.001) (Figure 7B). Histological
examination for the presence of human cells in the affected limbs
by human nuclear antigen staining indicated that HGPS-MSC
disappear much faster than N-MSC in ischemia limbs. At day
35 after surgery, while some HNA positive cells could be seen
in N-MSCs transplanted samples, no positive cells were found
in HGPS-MSC treated limbs (Figure 7C).
MSC-mediated, postischemia recovery is attributed to neo-
vasculogenesis due to the secretion of paracrine factors by the
transplanted MSCs (Horwitz and Prather, 2009). However,
analysis of media conditioned by the various control and
HGPS-MSC preparations failed to reveal any differences in the
levels of those factors often implicated in neovasculogenesis
including VEGF, bFGF, and IL-6 (Figure S5A). We conclude
that the poor survival of HGPS-MSCs in ischemic limbs may
underlie limb-rescue failure.
MSCs in vivo occupy low oxygen niches and normally exhibit
a faster and longer proliferation potential under hypoxic condi-
tions (Dos Santos et al., 2010; Rosova et al., 2008). To determine
(D) Immunostaining of progerin, LMNA, LAP2, and H2AX showed increased numb
VSMCs at p3. The percentage of cells showing abnormal phenotypes is indicate
whether HGPS-MSC loss is due to an acquired sensitivity to
ischemia-induced hypoxic conditions, we subjected the cells
to hypoxia and substratum deprivation as described earlier.
Under normal growth conditions, HGPS-MSC and N-MSCs
proliferated at same rate; however, after the hypoxia, only 40%
of the HGPS-MSCs survived compared to 80% of the N-MSCs
(Figure 6A). TUNEL assay showed that double the number of
HGPS-MSCs were apoptotic (Figure 7D). In parallel experi-
ments, HGPS fibroblasts and endothelial cells and their normal
controls derived from iPSCs survived, as well as human bone
marrow MSCs after hypoxia and substrate deprivation (Fig-
ure 6A). These results indicate that HGPS-MSCs are more sensi-
tive to the combination of hypoxia and substrate deprivation.
Interestingly, with hypoxia alone (3 days in 2% O2), very little
sensescence was noted in HGPS- and N-MSC populations in
contrast to the higher levels in HGPS-VSMCs (Figure 6B).
Antisense morpholinos or shRNAs specifically target and
suppress progerin expression (Huang et al., 2005). To demon-
strate that the increased sensitivity of the HGPS-MSCs were
a consequence of progerin expression, we infected N- or
HGPS-MSCs with lentiviral vectors expressing control shRNA
and shRNA against progerin. Resistance to hypoxia and
substratum deprivation is restored in HGPS-MSCs, when pro-
gerin levels were reduced by 65% (Figure 7E and Figure S6),
indicating progerin accumulation is responsible for this defect.
Finally, we determined if HGPS-MSCs were susceptible to
other forms of stress. Both HGPS-MSC and N-MSCs were
cultured in serum-free medium for 10 days. HGPS-MSC
numbers declined rapidly; in contrast, N-MSCs and adult bone
marrow MSCs survived serum starvation (Figure 7F).
In summary, HGPS fibroblasts, -MSCs, and -VSMCs all
display enhanced DNA damage, LAP2 mislocalization, and
pronounced nuclear dysmorphology. When exposed to addi-
tional stress in vitro and (MSC) in vivo, the viability of VSMCs
and MSCs were significantly reduced, with VSMCs showing
a particular sensitivity to low oxygen.
DISCUSSION
Mutations in LMNA are responsible for more than ten distinct
diseases (Mounkes and Stewart, 2004). HGPS is the best known
of these laminopathies, with the most common form of HGPS
being characterized by the production of the mutant lamin A,
progerin. It is widely believed that the pathological effects of
progerin aremediated by its disruption of the structural and func-
tional integrity of the nuclear lamina. Autopsies indicated that
death is associated with premature atherosclerosis (Olive
et al., 2010), which may be accompanied by vascular smooth
muscle loss (Stehbens et al., 2001) The limited biopsy data avail-
able shows progerin is mainly detected in some keratinocytes,
vascular smooth muscle, dermal fibroblast, and endothelial cells
(McClintock et al., 2006; Olive et al., 2010). Apart from the patient
data, most information on the disease pathophysiology has been
inferred from mouse models expressing mutated endogenous
LMNA alleles or mutated human or mouse LMNA transgenes,
er of cells associated with mislocalization of LAP2 and DNA damage in HGPS-
d. Scale bar, 20 mm.
Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc. 39
% o
f cel
ls w
ith
abno
rmal
nu
clea
r env
elop
B
A
C
Before s mula onA er s mula on
N-VSMC
Hypoxia/VSMC Normoxia/VSMC Hypoxia/MSC
Before s mula onA er s mula on
% o
f β-g
al p
osi
ve c
ells
2.2%10.5%
5.5%
35.6%
0%
10%
20%
30%
40%
50%
N-VSMC PG-VSMC
1.2% 2.3%5.7%
18.4%
0%
5%
10%
15%
20%
25%
N-VSMC PG-VSMC
* *
N-MSC/β-gal
PG-MSC/β-gal
4.6±1.2%
7.4±0.5%
β-gal
β-gal
5.1±1.3%
38.5±3.3%
PG-VSMC
N1(A
G03512)
N2(A
G06299)
BM
-MSC
N1-M
SC
N2-M
SC
HU
VEC
N1-e
ndo
N2-e
ndo
N1-V
SM
C
N2-V
SM
C
PG
1(A
G11498)
PG
2(A
G06297)
PG
1-M
SC
PG
2-M
SC
PG
1-e
ndo
PG
2-e
ndo
PG
1-V
SM
C
PG
2-V
SM
C
0.0
0.2
0.4
0.6
0.8
1.0
1.2
surv
iving
ratio
* ** *
N-VSMC
PG-VSMC
LMNA
8±2.3%
LMNA
52±4.8%
LMNA
2.2±1.1%
LMNA
10.5±1.8%
Figure 6. Stress Testing of iPSC-Derived Lineages in Culture
(A) Cell survival rate after treatment with oil immersion and substrate deprivation. Values represent ratio of surviving cells relative to input. Results were obtained
from three biological replicates, and experiments were repeated three times. *p < 0.01 compared with controls in left panel, e.g., BM-MSC, N1-iPSC-MSC,
N2-iPSC-MSC, HUVEC, N1-iPSC-endo, N2-iPSC-endo, N1-VSMC, and N2-VSMC.
(B) Hypoxia treatment of VSMCs. Cells were incubated in 2% O2 with normal culture medium for 72 hr. Representative phase contrast and immunostaining
images show morphology of VSMC and LMNA staining. Cell senescence was detected by b-gal staining. The percentage of cells showing b-gal positive and
abnormal nuclei is indicated. Cell senescence is not observed in iPS-MSCs under the same treatment.
(C) Electrical stimulation of VSMCs. Electrical pulses (40V/cm, 1Hz) were applied to VSMCs in culture for a period of 3 days. Increased nuclear dysmorphology by
LMNA staining was observed in HGPS-VSMCs (left panel). Accelerated sensescence of HGPS-VSMCs was detected by b-gal staining (right panel). Experiment
was repeated, and *p < 0.01 compared with N-iPSC-VSMCs.
Cell Stem Cell
Progeria iPSC Model Reveals MSC and VSMC Defects
studies on cultured patient fibroblasts, or through overexpres-
sion of progerin in primary and immortalized human cells.
Although some mouse models with LMNA changes show
a severe HGPS-like growth retardation and bone disease
(Fong et al., 2004; Mounkes et al., 2003; Varga et al., 2006;
40 Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc.
Yang et al., 2005), no mouse model captures all the human
symptoms and, until recently, only one (Varga et al., 2006)
displays a cardiovascular phenotype. Hernandez et al. (2010)
showed in the murine progeria model (LmnaD9/D9) that,
although the mutant lamin A in this strain has a different internal
Cell Stem Cell
Progeria iPSC Model Reveals MSC and VSMC Defects
deletion to the one found in progerin, it retains a farnesylated tail,
resulting in number of characteristic progeric pathologies,
including thinning of the VSMC intima. These pathologies were
attributed inter alia to defective extracellular matrix synthesis.
Microarray comparisons of the HGPS and control iPSC-MSCs
indicated significant misregulation of transcripts encoding extra-
cellular matrix proteins, as previously reported by Csoka et al.
(2004) and Scaffidi and Misteli (2008), although we could not
detect the alteration in the named components of the Notch
signaling pathway reported by these latter authors (Table S3).
HGPS fibroblasts demonstrate increased nuclear blebbing,
mislocalization of the nuclear protein (LAP2), DNA damage,
and aberrant chromatin modifications in HGPS cells at late
passage as reported for. HGPS fibroblasts or MSCs expressing
exogenously added progerin genes (Scaffidi and Misteli, 2008).
Apart from fibroblasts, very few other patient-derived cell
lineages are available for study. Here, we describe an approach
to HGPS modeling that should, in time, allow the production and
investigation of multiple HGPS cell lineages with endogenous
levels of progerin and other lamins.
HGPS fibroblasts were reprogrammed to iPSC by retroviral-
mediated integration of the Yamanaka factors OCT4, SOX2,
KLF4, and C-MYC (Takahashi et al., 2007). Though the reprog-
ramming efficiency is low compared with normal fibroblasts,
the HGPS-iPSCs, once established, were all karyotypically
normal and indistinguishable from normal iPSCs and human
ESCs in many aspects, including specific marker expression,
germ layer formation, epigenetic status, and global transcrip-
tional expression. As expected for pluripotent cell types (Con-
stantinescu et al., 2006), very little progerin, lamin A, or lamin C
were detected in the undifferentiated iPSCs. Upon differentiation
into fibroblasts, neural progenitors, vascular endothelial cells,
VSMCs, andMSCs, the LMNA gene is transcriptionally activated
and progerin levels increase. We observed the highest levels of
progerin (and the highest levels of DNA damage, LAP2 mislocal-
ization, and nuclear dysmorphology) in HGPS-iPSC derived
MSCs, VSMCs, and fibroblasts, slightly less in endothelial cells
and very little in neural progenitors. These differences are likely
to reflect inherent variation in the production and/or turnover of
progerin in the different cell types and the relationship between
progerin level and damage may, if extended over many more
cell types, track the more affected tissues in HGPS patients.
Progerin did not affect either the efficiency of differentiation of
iPSCs into MSCs, VSMCs, endothelial cells, and neural progen-
itors, or the proliferation rate of these cell types. It did not func-
tionally interfere with lattice formation or LDL uptake in the endo-
thelial cells, or with neuron formation from the neural progenitors.
We also noticed a new and completely unexpected phenotype in
many of the HGPS-VSMCs: the appearance of heterogeneously
sized, calponin 1-staining inclusion bodies in the cytoplasm
(Figure 5). Calponin1 is an actin-binding protein involved in the
regulation of smooth muscle contraction, possibly by inhibiting
actin-activated myosin ATP-ase activity (Takahashi and Yama-
mura, 2003). We speculate that its sequestration into aggregates
could affect the contractile properties of the VSMC in situ.
HGPS-VSMCs were also very sensitive to 2% hypoxia and to
the combination of hypoxia and substratum deprivation. Further-
more, whenHGPS-VSMCswere subjected to repeated pulses of
electrical stimulation, they rapidly senesced; in the context of the
vascular system, electrical stimulation has been used to
enhance angiogenesis (Zhao et al., 2004), to improve engineered
myocardium (Radisic et al., 2004) and, in this study, to act as
a surrogate means of mimicking the hemodynamic shear stress
normally endured by VSMCs in vivo. This pronounced sensitivity
of the HGPS-VSMCs to various imposed insults, as well as the
possible perturbation of contractile properties due to calponin
sequestration, may explain why this lineage features prominently
in the pathology of progeria.
HGPS and control iPSC-MSCs were differentiated into bone,
cartilage, and fat. We could not determine whether the efficiency
of differentiation into the various lineages was affected by the
presence of progerin, unlike Scaffidi and Misteli (2008), who
reported that adipogenesis was impaired while osteogenesis
was stimulated by progerin. We did not note any impact of pro-
gerin on adipogenesis, although osteogenenic differentiation
varied, but not in a progerin-specific manner (Figure 4E).
However, we believe the two sets of experiments are not compa-
rable since Scaffidi and Misteli (2008) obtained their results by
manipulation of a single MSC genotype, while each iPSC-MSC
population used in our study represented a different genotype,
and it is known that human MSC differentiation is influenced by
donor genotype (Leskela et al., 2006). We, therefore, elected to
functionally test iPSC-MSCs in vivo. Using a hind limb ligation
mouse model that measures limb survival and the restoration
of blood flow after cell transplantation, we found that HGPS-
MSCs were only slightly better than the vehicle in saving the
ischemic limb, in contrast to the successful rescue mediated
by control iPSC-MSCs or adult bone marrow MSCs. Microarray
comparisons between HGPS-MSCs and control MSCs showed
no obvious change in angiogenic factors such as VEGF, bFGF,
and Il-6, results confirmed by direct analysis of conditioned
media (Figure S5A). The histology of salvaged limbs showed
no long-term integration of any of the transplanted populations
and that HGPS-MSCs were cleared more rapidly than control
cells. We speculate that the HGPS-MSCs are cleared before
they can exert any trophic effect on the neighboring tissues.
This rapid clearance could reflect a greater sensitivity of HGPS-
MSCs to the ischemic conditions or to stress in general. Accord-
ingly, we stressed the cells in two ways: first, we studied their
reaction to serum starvation and found the HGPS-MSC numbers
rapidly declined (Figure 7F) and second, we deprived cells of
oxygen and substratum (Weil et al., 2009). Compared to the
control cells, survival of the HGPS-MSCs was significantly
reduced, an effect that was mostly reversed by a 65% decrease
in cellular progerin as a result of specific shRNA knockdown.We,
therefore, conclude that HGPS-MSC are particularly sensitive to
this type of hypoxic condition and believe that this is a significant
finding because MSCs normally reside in low O2 niches within
the body (Rosova et al., 2008). We do not yet have a molecular
explanation for this difference; transcript levels of one obvious
candidate-Hypoxia-Inducing Factor 1 were similar in control
and HGPS-MSCs, at least under normoxic conditions (Fig-
ure S5B). In vivo, the exact roles of MSCs are unclear, although
they may be a source of VSMCs and pericytes, but it is generally
agreed that they are important for tissuemaintenance and repair.
One hypothesis, for the underlying pathology of HGPS, is that
progerin inhibits cell replacement in the cardiovascular system,
hair follicles, fat, and cartilage due to premature exhaustion of
Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc. 41
A
1 week 2 week 4 week
PG1-1, n=5PG1-2, n=3PG2-1, n=7
N1-1, n=5N1-2, n=3N2-1, n=7
left right left right left right
0
100
200
300
400
0
100
200
300
400
perf
usio
n un
it
B
D
0
10
20
30
% c
ells
0
0.2
0.4
0.6
0.8
fetal-MSC
N1-iPSC-MSC
N2-iPSC-MSC
PG1 -iPSC-MSC
PG2-iPSC-MSC
OD
450
* *
Perc
enta
ge (%
)
0
10
20
30
40
50
60
70
80
90
100
Vehicle BM-MSC PG-iPS-MSC N-iPS-MSC
Limb LossFoot NecrosisLimb Salvage
C
* *
0.0
0.2
0.4
0.6
0.8
1.0
surv
ivin
g ra
tio
EF
N1-
MS
C
1 week 5 week
PG
1-M
SC
Figure 7. Rescue of Ischemic Murine Hind Limb Using Transplanted iPSC-Derived MSCs
(A) Transplantation of HGPS-iPSC-MSCs failed to attenuate hind-limb ischemia. At day 28 after transplantation of bonemarrow or iPSC-derivedMSCs, the phys-
iological status of ischemic limbs were rated for limb salvage, foot necrosis, and limb loss. The BM-MSC (p = 0.0058) and N-iPSC-MSC (p = 0.0092), but not
HGPS-iPSC-MSC (p = 0.2779), showed significant rescue of ischemia. Student’s t test (two tail) comparison with vehicle is used for analysis (n = 8 for vehicle,
n = 15 per group).
Cell Stem Cell
Progeria iPSC Model Reveals MSC and VSMC Defects
42 Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc.
Cell Stem Cell
Progeria iPSC Model Reveals MSC and VSMC Defects
stem cell pools (Halaschek-Wiener and Brooks-Wilson, 2007),
perhaps by inhibiting Wnt signaling (Hernandez et al., 2010;
Meshorer and Gruenbaum, 2008). Our findings lead us to
propose a refinement to their model and suggest that in addition
to the ‘‘exhaustion’’ caused by the need to replace lost mesen-
chymal tissue, the MSC pool is also depopulated due to
increased hypoxia sensitivity caused by progerin. Given the
historical (Horwitz and Prather, 2009) and current clinical trials
(http://clinicaltrials.gov/ct2/show/NCT01061099) using alloge-
neic MSCs for another congenital disease, osteogenesis imper-
fecta, our data suggest MSCs may be of therapeutic use for
HGPS patients.
To our knowledge, this is the first report of an iPSC-based
disease model of HGPS. It complements existing approaches
using animal and cell models, and the achievement of tissue-
specific expression of a disease-associated gene at endoge-
nous levels must be considered an attractive feature. This
general approach is still young and faces numerous challenges,
particularly regarding the appropriateness of using embryonic
starting material to model neonatal and adult onset diseases
(Colman and Dreesen, 2009; Saha and Jaenisch, 2009). We
are encouraged by the demonstration that the HGPS-iPSCs
can yield distinctive phenotypes, e.g., VSMCs and MSCs, that
lead to specific predictions that could be tested with suitable
animal and cellular models and/or patient material.
EXPERIMENTAL PROCEDURES
Cell Culture
HGPS patient fibroblast cells, adult human bone marrowMSCs, human umbil-
ical cord endothelial cells (HUVECs), and the human neural stem cell line (ReN-
cells) were purchased from Coriell cell repositories (http://ccr.coriell.org/),
Lonza, and Millipore, respectively. Fetal bone marrow MSC was obtained
from S.K. Lim (Institute of Medical Biology, Singapore). All cell cultures were
maintained at 37�C with 5% CO2. Human embryonic stem cell lines (hESCs)
HES3, H9, and induced pluripotent stem cells (iPSCs) were cultured on irradi-
ated mouse embryonic fibroblasts (MEFs) with Knockout DMEMmedium sup-
plemented with 20% knockout serum replacement, nonessential amino acid,
2-mercaptoethanol, penicillin/streptomycin, GlutaMax, and bFGF. Cells were
passaged with collagenase IV (all GIBCO). The hypoxia and substrate depriva-
tion assay was performed according to Weil et al. (2009). Cells were harvested
using 0.25% Trypsin-EDTA and pelleted by centrifugation. 3 3 105 cells were
(B) The dynamic change of blood flow after MSC transplantation. Hind limb ischem
was used as control for blood flow measurement using laser Doppler flow imagin
and restoration of blood flow 28 days after MSC transplantation (right panel). Rest
lower than N-iPSC-MSC group at 4 weeks. (p < 0.001, n = 15 per group). Three
patient fibroblasts were used to derive the MSCs used for transplantation.
(C) Immunostaining of human nuclear antigen (HNA) in ischemic limb tissues. Mic
immunostaining using antibody against HNA. Human cells stained brown (red ar
5 weeks. Scale bar, 200 mm.
(D) Percentage of cell death after hypoxia and substrate deprivation treatment. Ap
analyzed by FACS. Results were each obtained from three biological replicates
iPSC-MSC/PG2-iPSC-MSC (p < 0.01).
(E) Knocking down progerin in HGPS-iPSC-MSC improves survival in hypoxia and
shRNAs before being challenged. Values represent ratio of surviving cells relativ
ments were repeated two times. *p < 0.01 compared with H9-derived MSC, N
expressing control shRNA), and PG1-shD50 (HGPS-iPSC-MSC expressing shRN
(F) Cell proliferation in serum-free medium. HGPS or control iPSC-derived MSCs
for each line and cell survival wasmeasured using aWST-1 kit (Roche) according t
with three normal control groups.
See also Figure S4 for representative images of fibrosis in hind limbs of ischemic m
progerin shRNA and lamin expression.
exposed to 2–4 hr of hypoxia via mineral oil immersion followed by reoxygena-
tion in normal conditions for 24 hr. Surviving cells were counted using Trypan
blue stain.
Retroviral Production and iPS Generation
The pMX-based retroviral vector encoding the human cDNAs of KLF4, SOX2,
OCT4, and C-MYCwere obtained from Addgene. Retrovirus was produced as
described (Dimos et al., 2008). Briefly, pMXs plasmids were cotransfectedwith
packaging plasmid gag-pol and VSV-G into 293T packaging cells (ATCC)
using SuperFect (QIAGEN). Viral supernatant fractions were harvested after
60 hr, filtered through a 0.45 mm low protein binding cellulose acetate filter,
and concentrated by centrifugation. To produce patient-specific iPSCs, two
rounds of viral transduction of 100,000 fibroblast cells were performed. After
4 days, cells were transferred onto MEFs in human ESC medium containing
0.5 mM valproic acid (VPA, Sigma). iPSC colonies were manually picked after
2 to 3 weeks.
Differentiation of iPSCs into Fibroblast-like Cells
The human ESCs and iPSCs were harvested using collagenase IV and
embryoid bodies (EBs) were formed and transferred to gelatin-coated plates
in differentiation medium as described (Xu et al., 2004). The cells were subse-
quently passagedwithmedium containing 90%DMEM (Invitrogen), 10%heat-
inactivated FBS (Hyclone), 2mM L-glutamine, and 1% nonessential amino
acids. Cells were further purified with Thy1 antibody by FACS sorter. The iden-
tity of the established fibroblast-like cells was confirmed with immunostaining
by antibody against human fibroblast/epithelial cells (Novus Biologicals,
D7-FIB) and prolyl4-hydroxylase, an enzyme required for collagen synthesis.
Differentiation of VSMCs from MSCs
VSMCs were differentiated from iPSC-MSCs by culturing in EGM-2 medium
(Lonza) with sphingosylphosphorylcholine (SPC, 5 mM) and TGFb1 (2 ng/ml)
for 3 weeks. The identity of VSMCs was verified by specific marker expression
of smooth muscle actin, smooth muscle myosin heavy chain and calponin by
RT-PCR and immune-staining (all DAKO 1:100). Contraction of VSMCs was
induced by carbachol at 13 10�5 M for 1 hr. Electrical stimulation was applied
to VSMCs using C-pace/C-dish cell culture stimulation system (IonOptix, MA)
at 40V, 1 Hz with pulse duration of 2 ms.
Knockdown of Progerin in HGPS-iPS-MSCs by Lentiviral Infection
ShRNA-specific knockdown of lamin AD50 (progerin), but not lamin A or C,
was designed according to Huang et al. (2005). Oligonucleotides encoding
the hairpin shRNA (shD50) targeting the sequence (50-GGC TCA GGA GCC
CAG AGC CCC-30) were cloned into the lentiviral vector plko.3G (Addgene).
A shRNA that does not target any mammalian gene (50-TTC TCC GAA CGT
GTC ACGT-30) was used as control (shcon). For lentivirus production, lentiviral
vectors were cotransfected with packaging vectors into 293FT cells, and the
ia was created by ligation of the left hind limb femoral artery. The right hind limb
g. Representative Doppler photo shows no blood flow upon ligation (left panel)
oration of blood flow in the HGPS-iPSC-MSC transplanted group is significantly
iPSC clones (PG1-iPSC-1, PG1-iPSC-2, and PG2-iPSC-1) from two different
e tissues were paraformaldehyde fixed, paraffin embedded, and sectioned for
row) were not found anywhere in HGPS-iPSC-MSC transplanted tissues after
optotic cells were label by TUNEL (TdT-mediated dUTP nick end labeling) and
. N1-iPSC-MSC/N2-iPSC-MSC shows significantly more resistant than PG1-
substrate deprivation assay. Cells were transfected with lentivirus-containing
e to input. Results were obtained from three biological replicates, and experi-
1-NTC (N1-iPSC-MSC Non-Transduction Control), N1-shcon (N1-iPSC-MSC
A against progerin).
or bone marrow MSCs were seeded at a density of 43 104 per well in triplicate
o themanufacturer’s protocol. Value representsmean ± SD *p < 0.01 compared
ice, Figure S5 for characterization of hypoxia related factors, and Figure S6 for
Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc. 43
Cell Stem Cell
Progeria iPSC Model Reveals MSC and VSMC Defects
supernatant was harvested and concentrated by ultracentrifugation for 1.5 hr
at 25,000 r.p.m. in a Beckman SW28 rotor. Titers were determined by infecting
NIH/3T3 cells with a serial dilution of the concentrated virus. For a typical prep-
aration, the titer was approximately 1–53 107/ml. 23 105 cells were incubated
in suspension with 1 3 106 particles and 8 mg/ml polybrene for 3 hr in a 37�Cincubator. The cells were then replated and cultured as described. GFP-posi-
tive cells were sorted after 4 days of culture.
Limb Ischemia and Transplantation Studies
SCID mice were anesthetized with xylazine (20 mg/kg) and ketamine
(100 mg/kg), and critical limb ischemia was induced as described previously
(Lian et al., 2010). The femoral artery and its branches were ligated through
a skin incision with 5-0 silk (Ethicon, Somerville, NJ). The external iliac artery
and all of the above arteries were then ligated. The femoral artery was excised
from its proximal origin as a branch of the external iliac artery to the distal point
where it bifurcates into the saphenous and poplite arteries. After arterial liga-
tion, SCID mice were immediately assigned to the following experimental
groups: (1) N-iPS-MSC group: the mice were injected with MSCs derived
from normal control iPSCs (3.0 3 106 cells per mouse in 200 mL) intramuscu-
larly at four sites of the gracilis muscle in the medial thigh with 29-gauge tuber-
culin syringes; (2) PG- iPS-MSC: the mice were injected as above with MSCs
derived from HGPS patient iPSCs. Fetal bone marrow (BM)-MSC and culture
medium (vehicle) were used as controls. To exclude the possibility of bacterial
infection following surgery being responsible for poor recovery, specimens of
the ischemic muscle were cultured on day 7 after surgery and no bacterial
growth was detected in each group. All animal experiments were approved
by Committee on the Use of Live Animals in Teaching and Research (CULTAR)
at the University of Hong Kong.
Laser Doppler Imaging Analysis
Laser Doppler imaging analysis was performed as described previously (Lian
et al., 2010). A laser Doppler perfusion imager (Moor Instruments, Devon,
United Kingdom) was used for serial scanning of surface blood flow of hind-
limbs on days 0, 7, 14, and 28 after treatment. The digital color-coded images
were analyzed to quantify the blood flow in the region from the knee joint to the
toe, and mean values of perfusion were calculated.
ACCESSION NUMBERS
Microarray data has been deposited in the GEO database (GSE26093).
SUPPLEMENTAL INFORMATION
Supplemental Information includes six figures, three tables, and Supplemental
Experimental Procedures and can be found with this article online at doi:10.
1016/j.stem.2010.12.002.
ACKNOWLEDGMENTS
We thank the Singapore Biomedical Research Council and the Singapore
Agency for Science, Technology and Research (A*STAR) for funding this
work. The animal work was supported by Hong Kong Research Grant Council
(HKU 8/CRF/09). We also thank M. Costa for helpful advice.
Received: June 3, 2010
Revised: October 18, 2010
Accepted: December 6, 2010
Published online: December 23, 2010
REFERENCES
Ackerman, J., and Gilbert-Barness, E. (2002). Hutchinson-Gilford progeria
syndrome: a pathologic study. Pediatr. Pathol. Mol. Med. 21, 1–13.
Agarwal, S., Loh, Y.H., McLoughlin, E.M., Huang, J., Park, I.H., Miller, J.D.,
Huo, H., Okuka, M., Dos Reis, R.M., Loewer, S., et al. (2010). Telomere elon-
gation in induced pluripotent stem cells from dyskeratosis congenita patients.
Nature 464, 292–296.
44 Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc.
Bridger, J.M., and Kill, I.R. (2004). Aging of Hutchinson-Gilford progeria
syndrome fibroblasts is characterised by hyperproliferation and increased
apoptosis. Exp. Gerontol. 39, 717–724.
Burke, B., and Stewart, C.L. (2006). The laminopathies: the functional architec-
ture of the nucleus and its contribution to disease. Annu. Rev. Genomics Hum.
Genet. 7, 369–405.
Cao, K., Capell, B.C., Erdos, M.R., Djabali, K., and Collins, F.S. (2007). A lamin
A protein isoform overexpressed in Hutchinson-Gilford progeria syndrome
interferes with mitosis in progeria and normal cells. Proc. Natl. Acad. Sci.
USA 104, 4949–4954.
Capell, B.C., and Collins, F.S. (2006). Human laminopathies: nuclei gone
genetically awry. Nat. Rev. Genet. 7, 940–952.
Colman, A., and Dreesen, O. (2009). Pluripotent stem cells and diseasemodel-
ling. Cell Stem Cell 5, 244–247.
Constantinescu, D., Gray, H.L., Sammak, P.J., Schatten, G.P., and Csoka,
A.B. (2006). Lamin A/C expression is amarker of mouse and human embryonic
stem cell differentiation. Stem Cells 24, 177–185.
Csoka, A.B., English, S.B., Simkevich, C.P., Ginzinger, D.G., Butte, A.J.,
Schatten, G.P., Rothman, F.G., and Sedivy, J.M. (2004). Genome-scale
expression profiling of Hutchinson-Gilford progeria syndrome reveals wide-
spread transcriptional misregulation leading to mesodermal/mesenchymal
defects and accelerated atherosclerosis. Aging Cell 3, 235–243.
De Sandre-Giovannoli, A., Bernard, R., Cau, P., Navarro, C., Amiel, J.,
Boccaccio, I., Lyonnet, S., Stewart, C.L., Munnich, A., Le Merrer, M., et al.
(2003). Lamin a truncation in Hutchinson-Gilford progeria. Science 300, 2055.
Dimos, J.T., Rodolfa, K.T., Niakan, K.K., Weisenthal, L.M., Mitsumoto, H.,
Chung, W., Croft, G.F., Saphier, G., Leibel, R., Goland, R., et al. (2008).
Induced pluripotent stem cells generated from patients with ALS can be differ-
entiated into motor neurons. Science 321, 1218–1221.
Dos Santos, F., Andrade, P.Z., Boura, J.S., Abecasis, M.M., da Silva, C.L., and
Cabral, J.M. (2010). Ex vivo expansion of human mesenchymal stem cells:
a more effective cell proliferation kinetics and metabolism under hypoxia.
J. Cell. Physiol. 223, 27–35.
Ebert, A.D., Yu, J., Rose, F.F., Jr., Mattis, V.B., Lorson, C.L., Thomson, J.A.,
and Svendsen, C.N. (2009). Induced pluripotent stem cells from a spinal
muscular atrophy patient. Nature 457, 277–280.
Fong, L.G., Ng, J.K., Meta, M., Cote, N., Yang, S.H., Stewart, C.L., Sullivan, T.,
Burghardt, A., Majumdar, S., Reue, K., et al. (2004). Heterozygosity for Lmna
deficiency eliminates the progeria-like phenotypes in Zmpste24-deficient
mice. Proc. Natl. Acad. Sci. USA. 101, 18111–18116.
Goldman, R.D., Shumaker, D.K., Erdos, M.R., Eriksson, M., Goldman, A.E.,
Gordon, L.B., Gruenbaum, Y., Khuon, S., Mendez, M., Varga, R., et al.
(2004). Accumulation of mutant lamin A causes progressive changes in nuclear
architecture in Hutchinson-Gilford progeria syndrome. Proc. Natl. Acad. Sci.
USA 101, 8963–8968.
Gordon, L.B., McCarten, K.M., Giobbie-Hurder, A., Machan, J.T., Campbell,
S.E., Berns, S.D., and Kieran, M.W. (2007). Disease progression in
Hutchinson-Gilford progeria syndrome: impact on growth and development.
Pediatrics 120, 824–833.
Halaschek-Wiener, J., and Brooks-Wilson, A. (2007). Progeria of stem cells:
stem cell exhaustion in Hutchinson-Gilford progeria syndrome. J. Gerontol.
A Biol. Sci. Med. Sci. 62, 3–8.
Hanna, J., Saha, K., Pando, B., van Zon, J., Lengner, C.J., Creyghton, M.P.,
van Oudenaarden, A., and Jaenisch, R. (2009). Direct cell reprogramming is
a stochastic process amenable to acceleration. Nature 462, 595–601.
Hennekam, R.C. (2006). Hutchinson-Gilford progeria syndrome: review of the
phenotype. Am. J. Med. Genet. 140, 2603–2624.
Hernandez, L., Roux, K.J., Wong, E.S., Mounkes, L.C., Mutalif, R.,
Navasankari, R., Rai, B., Cool, S., Jeong, J.W., Wang, H., et al. (2010).
Functional coupling between the extracellular matrix and nuclear lamina by
Wnt signaling in progeria. Dev. Cell 19, 413–425.
Horwitz, E.M., and Prather, W.R. (2009). Cytokines as the major mechanism of
mesenchymal stem cell clinical activity: expanding the spectrum of cell
therapy. Isr. Med. Assoc. J. 11, 209–211.
Cell Stem Cell
Progeria iPSC Model Reveals MSC and VSMC Defects
Huang, S., Chen, L., Libina, N., Janes, J., Martin, G., Campisi, J., and Oshima,
J. (2005). Correction of cellular phenotypes of Hutchinson-Gilford Progeria
cells by RNA interference. Hum. Genet. 118, 444–450.
Jeon, E.S., Moon, H.J., Lee, M.J., Song, H.Y., Kim, Y.M., Bae, Y.C., Jung, J.S.,
and Kim, J.H. (2006). Sphingosylphosphorylcholine induces differentiation of
human mesenchymal stem cells into smooth-muscle-like cells through
a TGF-beta-dependent mechanism. J. Cell Sci. 119, 4994–5005.
Lee, G., Papapetrou, E.P., Kim, H., Chambers, S.M., Tomishima,M.J., Fasano,
C.A., Ganat, Y.M., Menon, J., Shimizu, F., Viale, A., et al. (2009). Modelling
pathogenesis and treatment of familial dysautonomia using patient-specific
iPSCs. Nature 461, 402–406.
Leskela, H.V., Olkku, A., Lehtonen, S., Mahonen, A., Koivunen, J., Turpeinen,
M., Uusitalo, J., Pelkonen, O., Kangas, L., Selander, K., et al. (2006). Estrogen
receptor alpha genotype confers interindividual variability of response to
estrogen and testosterone in mesenchymal-stem-cell-derived osteoblasts.
Bone 39, 1026–1034.
Li, Q., Yao, D., Ma, J., Zhu, J., Xu, X., Ren, Y., Ding, X., and Mao, X. (2009).
Transplantation of MSCs in combination with Netrin-1 improves neoangiogen-
esis in a rat model of hind limb ischemia. J. Surg. Res. Published online
September 23, 2009. 10.1016/j.jss.2009.08.031.
Lian, Q., Zhang, Y., Zhang, J., Zhang, H.K., Wu, X., Zhang, Y., Lam, F.F., Kang,
S., Xia, J.C., Lai, W.H., et al. (2010). Functional mesenchymal stem cells
derived from human induced pluripotent stem cells attenuate limb ischemia
in mice. Circulation 121, 1113–1123.
McClintock, D., Gordon, L.B., and Djabali, K. (2006). Hutchinson-Gilford pro-
geria mutant lamin A primarily targets human vascular cells as detected by
an anti-Lamin A G608G antibody. Proc. Natl. Acad. Sci. USA 103, 2154–2159.
McClintock, D., Ratner, D., Lokuge, M., Owens, D.M., Gordon, L.B., Collins,
F.S., and Djabali, K. (2007). The mutant form of lamin A that causes
Hutchinson-Gilford progeria is a biomarker of cellular aging in human skin.
PLoS ONE 2, e1269.
Merideth, M.A., Gordon, L.B., Clauss, S., Sachdev, V., Smith, A.C., Perry,
M.B., Brewer, C.C., Zalewski, C., Kim, H.J., Solomon, B., et al. (2008).
Phenotype and course of Hutchinson-Gilford progeria syndrome. N. Engl. J.
Med. 358, 592–604.
Meshorer, E., and Gruenbaum, Y. (2008). Gone with theWnt/Notch: stem cells
in laminopathies, progeria, and aging. J. Cell Biol. 181, 9–13.
Mounkes, L.C., and Stewart, C.L. (2004). Aging and nuclear organization: lam-
ins and progeria. Curr. Opin. Cell Biol. 16, 322–327.
Mounkes, L.C., Kozlov, S., Hernandez, L., Sullivan, T., and Stewart, C.L.
(2003). A progeroid syndrome in mice is caused by defects in A-type lamins.
Nature 423, 298–301.
Musich, P.R., and Zou, Z. (2009). Genomic instability and DNA damage
responses in progeria arising from defective maturation of prelamin A. Aging
1, 28–37.
Olive, M., Harten, I., Mitchell, R., Beers, J., Djabali, K., Cao, K., Erdos, M.R.,
Blair, C., Funke, B., Smoot, L., et al. (2010). Cardiovascular Pathology in
Hutchinson-Gilford Progeria: Correlation With the Vascular Pathology of
Aging. Arterioscler. Thromb. Vasc. Biol. 30, 2301–2309.
Radisic, M., Park, H., Shing, H., Consi, T., Schoen, F.J., Langer, R., Freed, L.E.,
and Vunjak-Novakovic, G. (2004). Functional assembly of engineered myocar-
dium by electrical stimulation of cardiac myocytes cultured on scaffolds. Proc.
Natl. Acad. Sci. USA 101, 18129–18134.
Rosova, I., Dao,M., Capoccia, B., Link, D., and Nolta, J.A. (2008). Hypoxic pre-
conditioning results in increasedmotility and improved therapeutic potential of
human mesenchymal stem cells. Stem Cells 26, 2173–2182.
Sagelius, H., Rosengardten, Y., Hanif, M., Erdos, M.R., Rozell, B., Collins, F.S.,
and Eriksson, M. (2008). Targeted transgenic expression of the mutation
causing Hutchinson-Gilford progeria syndrome leads to proliferative and
degenerative epidermal disease. J. Cell Sci. 121, 969–978.
Saha, K., and Jaenisch, R. (2009). Technical challenges in using human
induced pluripotent stem cells to model disease. Cell Stem Cell 5, 584–595.
Scaffidi, P., and Misteli, T. (2006). Lamin A-dependent nuclear defects in
human aging. Science 312, 1059–1063.
Scaffidi, P., and Misteli, T. (2008). Lamin A-dependent misregulation of adult
stem cells associated with accelerated ageing. Nat. Cell Biol. 10, 452–459.
Stehbens, W.E., Delahunt, B., Shozawa, T., and Gilbert-Barness, E. (2001).
Smooth muscle cell depletion and collagen types in progeric arteries.
Cardiovasc. Pathol. 10, 133–136.
Takahashi, K., and Yamamura, H. (2003). Studies and perspectives of calponin
in smooth muscleregulation and cancer gene therapy. Adv. Biophys. 3,
91–111.
Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Ichisaka, T., Tomoda, K.,
and Yamanaka, S. (2007). Induction of pluripotent stem cells from adult human
fibroblasts by defined factors. Cell 131, 861–872.
Varga, R., Eriksson, M., Erdos, M.R., Olive, M., Harten, I., Kolodgie, F., Capell,
B.C., Cheng, J., Faddah, D., Perkins, S., et al. (2006). Progressive vascular
smooth muscle cell defects in a mouse model of Hutchinson-Gilford progeria
syndrome. Proc. Natl. Acad. Sci. USA 103, 3250–3255.
Verstraeten, V.L., Ji, J.Y., Cummings, K.S., Lee, R.T., and Lammerding, J.
(2008). Increased mechanosensitivity and nuclear stiffness in Hutchinson-
Gilford progeria cells: effects of farnesyltransferase inhibitors. Aging Cell 7,
383–393.
Weil, B.R., Markel, T.A., Herrmann, J.L., Abarbanell, A.M., and Meldrum, D.R.
(2009). Mesenchymal stem cells enhance the viability and proliferation of
human fetal intestinal epithelial cells following hypoxic injury via paracrine
mechanisms. Surgery 146, 190–197.
Xu, C., Jiang, J., Sottile, V., McWhir, J., Lebowski, J., and Carpenter, M.K.
(2004). Immortalized fibroblast-like cells derived from human embryonic
stem cells support undifferentiated cell growth. Stem Cells 22, 972–980.
Yang, S.H., Bergo, M.O., Toth, J.I., Qiao, X., Hu, Y., Sandoval, S., Meta, M.,
Bendale, P., Gelb, M.H., Young, S.G., et al. (2005). Blocking protein farnesyl-
transferase improves nuclear blebbing in mouse fibroblasts with a targeted
Hutchinson-Gilford progeria syndrome mutation. Proc. Natl. Acad. Sci. USA
102, 10291–10296.
Yang, S.H., Andres, D.A., Spielmann, H.P., Young, S.G., and Fong, L.G. (2008).
Progerin elicits disease phenotypes of progeria in mice whether or not it is far-
nesylated. J. Clin. Invest. 118, 3291–3300.
Zhao, M., Bai, H., Wang, E., Forrester, J.V., andMcCaig, C.D. (2004). Electrical
stimulation directly induces pre-angiogenic responses in vascular endothelial
cells by signaling through VEGF receptors. J. Cell Sci. 117, 397–405.
Cell Stem Cell 8, 31–45, January 7, 2011 ª2011 Elsevier Inc. 45
Cell Stem Cell
Article
Calcineurin-NFAT Signaling Critically RegulatesEarly Lineage Specificationin Mouse Embryonic Stem Cells and EmbryosXiang Li,1,2,5 Lili Zhu,1,2,5 Acong Yang,1,3 Jiangwei Lin,4 Fan Tang,1,3 Shibo Jin,1 Zhe Wei,1,2 Jinsong Li,4 and Ying Jin1,3,*1Key Laboratory of Stem Cell Biology, Institute of Health Sciences, Shanghai Institutes for Biological Sciences,
Chinese Academy of Sciences/Shanghai JiaoTong University School of Medicine, Shanghai, 200025, China2Graduate School of Chinese Academy of Sciences, Beijing, 100000, China3Shanghai Stem Cell Institute, Shanghai JiaoTong University School of Medicine, Shanghai, 200025, China4Laboratory of Molecular Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences,
Chinese Academy of Sciences, Shanghai, 200031, China5These authors contributed equally to this work*Correspondence: yjin@sibs.ac.cn
DOI 10.1016/j.stem.2010.11.027
SUMMARY
Self-renewal and pluripotency are hallmarks ofembryonic stem cells (ESCs). However, the signalingpathways that trigger their transition from self-renewal to differentiation remain elusive. Here, wereport that calcineurin-NFAT signaling is both neces-sary and sufficient to switch ESCs from an undiffer-entiated state to lineage-specific cells and that theinhibition of this pathway can maintain long-termESC self-renewal independent of leukemia inhibitoryfactor. Mechanistically, this pathway converges withthe Erk1/2 pathway to regulate Src expression andpromote the epithelial-mesenchymal transition(EMT), a process required for lineage specificationin response to differentiation stimuli. Furthermore,calcineurin-NFAT signaling is activated when theearliest differentiation event occurs in mouse em-bryos, and its inhibition disrupts extraembryoniclineage development. Collectively, our results de-monstrate that the NFAT and Erk1/2 cascades forma signaling switch for early lineage segregation inmouse ESCs and provide significant insights intothe regulation of the balance between ESC self-renewal and early lineage specification.
INTRODUCTION
During early mouse development, lineage specification begins at
embryonic day 2.5 (E2.5) in 8-cell embryos. The first lineage
decision leads to the establishment of the inner cell mass (ICM)
and the trophectoderm at the blastocyst stage. The second
lineage decision gives rise to the epiblast and the primitive
endoderm when the latter delaminates from the ICM (Rossant,
2007; Rossant et al., 2003). The epiblast contains pluripotent
cells that generate the three germ layers as well as germ cells.
Embryonic stem cells (ESCs), derived from the preimplantation
blastocyst, have the potential to differentiate into all cell types
46 Cell Stem Cell 8, 46–58, January 7, 2011 ª2011 Elsevier Inc.
of an organism and to grow indefinitely in culture (Martin, 1981;
Smith, 2001). Mouse ESCs can be maintained in an undifferenti-
ated self-renewal state in the presence of leukemia inhibitory
factor (LIF) and either bone morphogenetic protein 4 (BMP4) or
serum without feeder cells (Ying et al., 2003). Withdrawal of
LIF results in extensive ESC differentiation with downregulation
of the pluripotency-associated core transcription factors
Oct4, Sox2, and Nanog (Kim et al., 2008). Recently, maintaining
ESCs at a ground state of self-renewal in the absence of LIF and
serum was reported via two inhibitors (2i) of fibroblast growth
factor/extracellular signal-related kinase 1/2 (Fgf/Erk1/2) and
glycogen synthase kinase 3 (GSK3) (Ying et al., 2008). Despite
these major advances, the molecular basis for ESCs to transit
from the state of self-renewal to early differentiation has not
been fully elucidated. The Fgf/Mek/Erk1/2 pathway is
considered important for the formation of the first two extraem-
bryonic lineages (trophectoderm and primitive endoderm) in
early murine development and for ESC differentiation in vitro
(Chazaud et al., 2006; Lu et al., 2008; Nichols et al., 2009; Yama-
naka et al., 2010). We were interested in the question of whether
there are additional signaling pathways critical for the early
lineage specification and, if so, how they may be integrated to
orchestrate early development. To address this question,
we employed the piggyBac (PB) transposon, which randomly
inserts into a host genome and disrupts gene function (Ding
et al., 2005), and then selected transfected ESC colonies with
undifferentiated ESC morphology after LIF withdrawal. One
gene that was identified through this approach was Cnb1, also
known as Ppp3r1, which encodes calcium binding B (CnB),
a subunit of calcineurin.
Calcineurin is a Ca2+ influx-activated serine/threonine-
specific phosphatase composed of the CnA and CnB subunits
(Crabtree, 1999; Crabtree and Schreiber, 2009). Three genes
(Ppp3ca, Ppp3cb, and Ppp3cc) encode three members of the
catalytic CnA subunit, whereas two members of the regulatory
CnB unit of calcineurin are products of two genes (Ppp3r1 and
Ppp3r2). Calcineurin dephosphorylates cytoplasmic NFAT
(nuclear factor of activated T cell, the products of four NFATc
genes, NFATc1-c4) to promote translocation of NFAT to the
nucleus, where NFAT and its usual partner AP1 (Fos/Jun,
substrates of Erk1/2) bind target promoters to control the
Cell Stem Cell
NFAT Pathway Regulates Early Lineage Specification
expression of genes with diverse functions. Although NFAT
proteins were first recognized for their central role in T lympho-
cyte activation, they have since been demonstrated to orches-
trate diverse developmental programs, including nervous,
cardiovascular, hematopoietic, and muscular-skeletal develop-
ment, as well as to maintain the quiescent state of stem cells
in skin (Chin et al., 1998; Clipstone and Crabtree, 1992; Graef
et al., 2001, 2003; Molkentin et al., 1998; Muller et al., 2009;
Stankunas et al., 1999). Recently, an intrinsic role of calcineurin
signaling in keratinocyte tumor suppression was reported (Wu
et al., 2010). Mice with disrupted Cnb1 gene, or NFATc1 gene,
or both NFATc3 and NFATc4 genes, die around E11 to E14
(de la Pompa et al., 1998; Graef et al., 2001). However, the
expression and function of calcineurin-NFAT signaling in
ESCs, as well as in early embryonic development at the peri-
implantation stage, have remained unnoticed, possibly because
of genetic redundancy among family members of the calci-
neurin-NFAT signaling pathway. The advent of pharmacologic
inhibitors of NFAT translocation has greatly facilitated our under-
standing of the functions of calcineurin-NFAT signaling. The
specificity of these inhibitors for calcineurin, and potentially
NFAT, was established based upon the nearly identical pheno-
types observed for Cnb1 null mice, NFATc3/c4 double null
mice, and embryos of mothers given the inhibitor at E7.5 to
E8.5 (Crabtree and Olson, 2002; Graef et al., 2001). Here,
utilizing the inhibitors and genetic approaches, we demonstrate
that calcineurin-NFAT signaling is both necessary and sufficient
to trigger lineage commitment through the upregulation of
Src expression to promote epithelial-mesenchymal transition
(EMT) in mouse ESCs.
RESULTS
Calcineurin-NFAT Signaling Is Required forMultilineageDifferentiation of ESCsA PB plasmid (PGK-Neo) and a PBase expression plasmid
(Act-PBase) were coelectroporated into ESCs and transfectants
were selected with G418 in the presence of LIF and serum, as
described (Ding et al., 2005). Approximately 1000G418-resistant
colonies were obtained from each 10 cm dish. Colonies that
maintained an undifferentiated morphology in the absence of
LIF after several passages in culture were expanded. Integration
sites were analyzed with inverse PCR. Among these sites, genes
were identified that potentially associate with the Ras-MAPK,
Akt, and calcineurin signaling pathways, such as Grb10,
Ptpn21, Inpp4b, and Ppp3r1 (Table S1 available online). We
were particularly interested in the Ppp3r1 gene (Figure S1A),
because it encodes the regulatory subunit of calcineurin, and
the role of calcineurin-NFAT signaling in ESCs has not been
reported.
The disruption of Ppp3r1 expression by PB insertion and its
role in the prevention of ESCs from LIF withdrawal-induced
differentiation were validated by quantitative real-time RT-PCR
(qRT-PCR) analysis of the levels of expression of Ppp3r1 and
other marker genes in both control and Ppp3r1-disrupted
ESCs (Figure S1B). Subsequently, to determine the function of
the calcineurin-NFAT pathway, we treated ESCs with cyclo-
sporine A (CsA), a specific inhibitor of calcineurin (Clipstone
and Crabtree, 1992), and examined cell morphology and the
expression of the marker genes in the presence and absence
of LIF. As expected, the removal of LIF resulted in extensively
differentiated cell morphology, increased transcript levels for
various lineage markers (including Fgf5, Cdx2, Dab2, Nestin, T,
and Mixl1), and downregulation of pluripotency markers such
as Oct4, Nanog, and Rex1 (Figures 1A and 1B). Moreover, at
the protein level, nuclear staining of Nanog was dramatically
reduced after LIF withdrawal (Figure 1C). All these LIF with-
drawal-induced changes were efficiently abolished by CsA treat-
ment (Figures 1A–1C), although CsA-treated ESCs displayed
a relatively reduced growth rate (Figure S1C). In contrast, CsA
did not affect the LIF withdrawal-induced reduction in the level
of phosphorylated Stat3, although it did block the LIF with-
drawal-induced decrease in total Stat3 protein levels
(Figure S1D). The inhibitory effect of CsA on calcineurin-NFAT
signaling under such conditions was verified by a reporter assay
(Figure S1E). Furthermore, FK506 (the unrelated calcineurin
inhibitor) and the NFAT-selective inhibitory peptide VIVIT
(Yu et al., 2007) significantly attenuated ESC differentiation
induced by LIF removal (Figures S1F–S1I). A similar effect was
also observed when RNA interference (RNAi) targeting Ppp3r1
was introduced (Figure 1D; Figure S1J), suggesting that the
attenuation was calcineurin-NFAT signaling dependent. In addi-
tion, CsA could block neural differentiation of ESC-derived
epiblast stem cells as well as ESC differentiation in the presence
of retinoic acid (RA) and during embryoid body (EB) formation
(Figures S1K–S1N). Collectively, our findings demonstrate that
calcineurin-NFAT signaling is crucial for ESC differentiation in
response to various differentiation stimuli.
We next sought to determine whether, in the absence of LIF,
CsA couldmaintain ESCs in an undifferentiated state indefinitely.
After 40 days, ESCs cultured under such conditions displayed an
undifferentiated morphology. After withdrawal of CsA, they
retained a normal differentiation potential and formed EBs with
differentiated cells expressing markers of all three germ layers
(Figures 1E and 1F). Teratomas exhibiting cells of the three
germ layers were also detected when these cells were injected
into nude mice (Figure 1G). Finally, the ESCs re-entered embry-
onic development in the chimeric mice (Figure 1H) when
CsA-expanded ESCs carrying a histone 2B-GFP gene were in-
jected into mouse blastocysts. Therefore, the inhibition of the
calcineurin-NFAT pathway can sustain ESC properties for
a long period in the absence of LIF.
Finally, we cultured CsA-treated ESCs in the serum-free
condition in the absence of LIF (Ying et al., 2008). Similar to 2i,
CsA or CsA plus the GSK3 inhibitor (CHIR) could maintain
the expression of Oct4 and Nanog as well as a compact undiffer-
entiated morphology (Figures S1O and S1P). Moreover, the
combination of CsA and CHIR could sustain a cell growth rate
comparable to that observed with 2i treatment (Figure S1Q).
Calcineurin-NFAT Signaling Is Activated upon ESCDifferentiationWe then examined the expression patterns of calcineurin and
NFAT in mouse ESCs and their progeny. The transcriptional
levels of the calcineurin subunits increased when ESCs were
induced into differentiation by LIF withdrawal (Figure 2A). Pub-
lished microarray data (Ivanova et al., 2006) showed low
NFATc1/c2 expression and abundant NFATc3/c4 expression
Cell Stem Cell 8, 46–58, January 7, 2011 ª2011 Elsevier Inc. 47
Figure 1. Calcineurin-NFAT Signaling Is Required for Multilineage Differentiation of ESCs
(A) Phase contrast images of CGR8 ESCs grown for 3.5 days under the indicated conditions. Scale bars represent 100 mm.
(B) qRT-PCR analysis of gene expression levels in cells described in (A). The mRNA level in ESCs cultured with LIF was set at 1.0. Data are shown as the
mean ± SD (n = 3).
(C) Confocal images of ESCs grown under the indicated conditions and incubated with Nanog antibody. Scale bars represent 50 mm.
(D) qRT-PCR analysis of marker gene expression levels in CGR8 ESCs cultured in the presence or absence of LIF, the control oligonucleotide, and the Ppp3r1
oligonucleotide. The mRNA level in ESCs cultured with LIF and the control oligonucleotide was set at 1.0. Data are shown as the mean ± SD (n = 3).
(E) RT-PCR analysis of marker genes in EBs derived from CsA-expanded ESCs. Gapdh was used as a loading control.
(F) Immunofluorescence staining of EBs after they adhered to culture dishes. Cells were stained with antibodies against Sox17 (endoderm), Nestin (ectoderm),
and Flk1 (mesoderm). Scale bars represent 50 mm.
(G) H&E staining of teratomas from ESCs grown in medium containing CsA (15 mM) and subcutaneously injected into nude mice. Scale bars represent 50 mm.
(H) A chimeric embryo at day E18.5 produced from CsA-expanded ESCs, expressing a histone 2B-GFP fusion protein.
See also Figure S1 and Table S1.
Cell Stem Cell
NFAT Pathway Regulates Early Lineage Specification
in mouse ESCs (Figure 2B). Results from our western blot anal-
ysis indicated that the steady-state level of NFATc4 proteins
increased markedly after ESC differentiation, whereas the
NFATc3 level remained consistently high (Figure 2C). The spec-
ificity of NFATc3 and NFATc4 antibodies was verified by
RNAi-specific expression knockdown (Figures S2A and S2B).
Notably, NFATc3 was primarily found in the cytoplasm of undif-
ferentiated ESCs but moved to the nucleus after LIF withdrawal
48 Cell Stem Cell 8, 46–58, January 7, 2011 ª2011 Elsevier Inc.
(Figure 2D), indicative of NFATc3 activation upon ESC differenti-
ation. In line with this observation, the activity of a luciferase
reporter containing three tandem copies of the murine Il2
promoter element, a prototypical NFAT:AP-1 composite (Macian
et al., 2000), was significantly upregulated after the removal of
LIF (Figure 2E). In later experiments of this study, we primarily
focused on NFATc3 because of its high expression level in
ESCs and rapid activation upon differentiation.
Figure 2. Calcineurin-NFAT Signaling Is Activated
during ESC Differentiation
(A) qRT-PCR analysis of calcineurin subunit mRNA levels
in ESCs after LIF withdrawal. The mRNA level in ESCs
cultured with LIF was set at 1.0. Data are shown as the
mean ± SD (n = 3).
(B) Expression of NFATc1-4 at different days after RA
induction. The microarray expression data were obtained
from the literature (Ivanova et al., 2006).
(C) Western blot analysis of NFATc3 and NFATc4 steady-
state levels in ESCs after LIF withdrawal. Tubulin was used
as a control.
(D) Subcellular localization of NFATc3 in ESCs 3.5 days
after LIF withdrawal was detected by immunostaining.
DAPI was used to visualize the nucleus. Scale bars repre-
sent 25 mm.
(E) Activation of the NFAT:AP-1 reporter in ESCs upon LIF
removal. The activity in the cells cultured with LIF was set
at 1.0. Data are shown as the mean ± SD (n = 3).
See also Figure S2.
Cell Stem Cell
NFAT Pathway Regulates Early Lineage Specification
Activation of Calcineurin-NFAT Signaling Triggers anESC Transition from Self-Renewal to DifferentiationTo addresswhether calcineurin-NFAT signaling acts as apermis-
sive signal or if its activation is sufficient to initiate ESC differenti-
ation,weoverexpressed constitutively active formsof calcineurin
(DCnA) or NFATc1-4 (CA-NFATc1-4) in ESCs. Strikingly, in the
presence of LIF, forced expression of CA-NFATc1-4 rapidly
induced the differentiated cell morphology (Figure 3A), substan-
tially reduced the mRNA levels of pluripotency genes (Oct4,
Nanog, and Rex1), and enhanced the expression of differentia-
tion markers, especially those of the trophectoderm (Cdx2,
Hand1) and the primitive endoderm (Gata6, Ihh) (Figure 3B);
DCnA induced relatively weaker differentiation than did CA-
NFATc1-4 (Figures 3C). After injecting ESCs overexpressing
CA-NFATc3 into immunodeficient mice, teratomas containing
regional hemorrhages were subsequently detected (Figure 3D),
Cell Stem Cell
suggesting the existence of trophoblastic cell
types in the tumor. RT-PCR analysis showed
that levels of trophoblast and primitive endo-
derm markers were markedly higher in tera-
tomas derived from CA-NFATc3-expressing
cells than from control cells (Figure 3E). There-
fore, the activation of calcineurin-NFAT
signaling is sufficient to induce ESCs into early
differentiated lineages.
NFAT Directly Activates Src ExpressionTo identify the downstreammolecules regulated
by calcineurin-NFAT signaling, we examined
the expression of a repertoire of genes while
calcineurin-NFAT signaling was either activated
or inhibited. We uncovered 306 candidate
targets, including molecules involved in cell
migration and focal adhesions (Figures S3A
and S3B; Table S2). To find direct NFAT targets,
the promoter sequences of these candidate
genes were examined for binding sites of
NFAT and AP-1, the common cofactor of
NFAT (Macian et al., 2001; Sanna et al., 2005).
The consensus NFAT and AP-1 binding sites were foundupstream of the nonreceptor tyrosine kinase Src (also known
as c-Src) gene, and these elements were well conserved (Fig-
ure S3C). AP1 is required for Src expression (Jiao et al., 2008;
Kumagai et al., 2004), whereas the involvement of NFAT in Src
expression has not been reported. The results of qRT-PCR
analysis showed that the tetracycline (Tc)-inducible expression
of a constitutively active NFATc3 significantly elevated Src
transcript levels in an induction-time-dependent manner (Fig-
ure 4A), and the transient overexpression of CA-NFATc3 also
increased Src protein levels (Figure 4B). Moreover, the transient
expression of DCnA or CA-NFATc1-4 substantially upregulated
Src transcript levels (Figure S3D), comparable to the effect of
activated Ras (Figure S3E). Strikingly, the activation of Src
expression and kinase activity by LIF withdrawal was completely
abolished by CsA treatment (Figure 4C; Figure S3F), indicating
8, 46–58, January 7, 2011 ª2011 Elsevier Inc. 49
Figure 3. Activation of Calcineurin-NFAT Signaling Triggers ESC Transition from Self-Renewal to Lineage Commitment
(A) Morphological changes of ESCs induced by the transient overexpression of the constitutively active form of calcineurin (DCnA) or the constitutively active form
of NFATc1-4 (CA-NFATc1-4). CGR8 ESCs were grown under selection for 3 days after transfection. Scale bars represent 50 mm.
(B and C) qRT-PCR analysis of marker expression levels in cells described in (A). The mRNA level in cells transfected with the control vector was set at 1.0. Data
are shown as the mean ± SD (n = 3).
(D) The images of teratomas derived from control cells or CA-NFATc3 transiently expressing cells.
(E) RT-PCR analysis of gene expression of three teratomas from control cells and hemorrhagic teratomas from CA-NFATc3 transiently expressing cells.
Cell Stem Cell
NFAT Pathway Regulates Early Lineage Specification
a regulatory role for calcineurin-NFAT signaling in endogenous
Src expression. To test whether NFAT directly regulates Src
expression, electrophoretic mobility shift assays (EMSAs)
were conducted with a Src oligo probe (containing the NFAT
consensus sequence from the Src promoter), an Il2 control
probe, and nuclear extracts from ESCs overexpressing
CA-NFATc3 (Figure 4D). NFATc3-containing DNA-protein com-
plexes were identified by the appearance of a super-shifted
band when NFATc3 antibody was included in the incubation
mixture, which was identical to the super-shifted band observed
with the control Il2 probe. To test the functional significance
of the NFAT binding site for Src expression, we performed
luciferase reporter assays and verified critical roles for NFAT
and AP-1 binding sites in Src transcription in response to
NFATc3 (Figure 4E). Finally, chromatin immunoprecipitation
(ChIP) assays showed an enrichment of NFATc3 at the Src
gene, but not at the Cdx2 regulatory region during ESC differen-
50 Cell Stem Cell 8, 46–58, January 7, 2011 ª2011 Elsevier Inc.
tiation induced by either LIF withdrawal or CA-NFATc3 overex-
pression (Figure 4F). These observations revealed that endoge-
nous NFATc3 or ectopically expressed NFATc3 binds to the
Src regulatory sequence in differentiating ESCs but not in undif-
ferentiated ESCs.
Src Is Required for Calcineurin-NFAT-Induced ESCDifferentiationWe next examined whether Src could initiate ESC differentiation.
Transient expression of a sustained active form of Src (Src F)
(Brabek et al., 2004) resulted in the rapid appearance of the
differentiated cell morphology (Figure 5A) with the simultaneous
downregulation of pluripotency markers and the activation of
differentiation genes. Markers robustly activated by Src F
included trophectoderm markers (Cdx2, Hand1, Prl2c2, and
Psx1) followed by primitive endoderm lineage markers, such as
Gata6 (Figure 5B), which is similar to the expression pattern
Figure 4. NFAT Directly Activates Src Expression
(A) qRT-PCR analysis of dynamic expression patterns of Src after the induction of CA-NFATc3 expression by removing Tc (0.5 mM) in iNFATc3ES cells. The rela-
tive mRNA values in control cells were set at 1.0. Data are shown as the mean ± SD (n = 3).
(B) Western blot analysis of Src protein levels after transient overexpression of CA-NFATc3. Tubulin was used as a control. The number indicates the relative
density of specific bands from the western blot measured by densitometer and normalized by the density of Tubulin (n = 3).
(C) qRT-PCR analysis of Src mRNA levels in ESCs cultured under the indicated conditions for 4 days. Data are shown as the mean ± SD (n = 3).
(D) EMSAs via the Src or Il2 probe and the nuclear extract of E14T ESCs transiently overexpressing CA-NFATc3 for 2 days.
(E) Luciferase assays with the wild-type (wt) Src promoter luciferase reporter with or without mutations in the NFAT (NFATm) or AP-1 (AP-1 m) element in ESCs.
The activity of wt Src promoter reporter cotransfected with empty vector was set at 1.0. Data are shown as the mean ± SD (n = 3).
(F) ChIP assays demonstrating the capacity of NFATc3 to bind to the Src upstream fragment during ESC differentiation induced by LIF withdrawal or Tc-induced
CA-NFATc3 expression.
See also Figure S3 and Table S2.
Cell Stem Cell
NFAT Pathway Regulates Early Lineage Specification
induced by activated NFAT. To determine whether Src is
essential for NFATc3- or LIF withdrawal-initiated ESC differenti-
ation, PP2 (Hamadi et al., 2009), a specific Src inhibitor, was
used. Strikingly, PP2 efficiently abolished ESC differentiation
induced by CA-NFATc3 overexpression or LIF withdrawal,
as determined by cell morphology and marker expression
(Figures 5C–5F). Interestingly, CsA and PP2 combined treatment
sustained a faster cell growth rate than CsA treatment alone
(Figure S1C). In addition to PP2, SKI-1, another Src inhibitor,
also abrogated LIF withdrawal-induced differentiation (Figures
S4A and S4B). Specifically, the knockdown of Src by specific
RNAi also markedly attenuated NFATc3- or LIF withdrawal-
induced differentiation (Figure 5G; Figure S4D), whereas the
depletion of another Src family kinase, Lck, did not block
differentiation (Figures S4C and S4D). Our results indicate that
Src, like NFAT, is both necessary and sufficient for ESC
differentiation.
NFAT-Src-Mediated EMT Is an Essential Stepfor Lineage SpecificationSrc is closely associated with EMT, and many EMT inducers
(Tgfb, Fgf family members, Ras-Erk1/2) also induce ESC differ-
entiation (Guarino, 2010). We therefore asked whether EMT
was an essential step for ESC differentiation. In addition to the
induction of ESC differentiation, Src F overexpression led to
the rapid appearance of typical EMT characteristics (Mandal
et al., 2008; Thiery et al., 2009), including the activation of
EMT markers such as Igf2, SIP1, Ncad, and Snai1, as well as
upregulated matrix metalloproteinases (MMPs), which are crit-
ical components of EMT (Figure 6A; Cavallaro and Christofori,
2004). The redistribution of E-cadherin, an epithelial cell
marker, was also observed from themembrane to the cytoplasm
in Src F-expressing cells (Figure 6B). As upstream activators of
Src, both NFAT and Ras also increased expression of the EMT
markers (Figures S5A and S5B). Furthermore, E-cadherin redis-
tribution was observed upon either overexpressing active
NFATc3 or withdrawing LIF (Figures S5C and S5D).
To define the relationship between EMT and ESC lineage
commitment, the kinetics of gene expression were examined.
After active NFATc3 induction, the elevation of EMT genes
(on day 1) occurred prior to the activation of lineage markers
(on day 1.5) and the downregulation of pluripotency genes (on
day 3) (Figure 6C), suggesting that the EMT process precedes
ESC differentiation. Furthermore, the broad-spectrum MMP
inhibitor GM6001, known to prevent EMT (Tan et al., 2010),
blocked NFATc3- induced ESC differentiation (Figures 6D and
6E) and partially blocked LIF withdrawal-induced ESC differenti-
ation (Figures S5E and S5F). The calcineurin inhibitor CsA and
Cell Stem Cell 8, 46–58, January 7, 2011 ª2011 Elsevier Inc. 51
Figure 5. Src Is Required for Calcineurin-NFAT-Induced ESC Differentiation
(A) Morphological changes in ESCs induced by the constitutively active form of Src (Src F). CGR8 ESCs were grown under selection for 3 days after transient
transfection. Scale bars represent 50 mm.
(B) qRT-PCR analysis of marker expression levels in cells described in (A). The expression level in cells transfected with the control vector was set at 1.0. Data are
shown as the mean ± SD (n = 3).
(C) Blockade ofNFATc3-mediated ESC differentiation by Src inhibitor PP2. Phase contrast images of iNFATc3ES cells cultured under the indicated conditions are
shown. CA-NFATc3 was induced by removing Tc (0.5 mM) for 3.5 days in iNFATc3 ESCs. Scale bars represent 100 mm.
(D) qRT-PCR analysis of marker expression levels in cells described in (C). The expression level in control cells without PP2 was set at 1.0. Data are shown as the
mean ± SD (n = 3).
(E) Abolishment of LIF withdrawal-induced ESC differentiation by PP2. Phase contrast images of CGR8 ESCs cultured under the indicated conditions for 3 days
are shown. Scale bars represent 100 mm.
(F) qRT-PCR analysis of marker expression levels in cells described in (E). The relative mRNA level in cells cultured with LIF was set at 1.0. Data are shown as the
mean ± SD (n = 3).
(G) Blockade of NFATc3-mediated ESC differentiation by Src knockdown via RNAi. qRT-PCR analysis of marker expression levels in cells in the presence or
absence of Src siRNAs. Expression of CA-NFATc3 was induced by removing Tc (0.5 mM) for 2 days in iNFATc3ES cells. The level in control cells with
control-siRNA was set at 1.0. Data are shown as the mean ± SD (n = 3).
See also Figure S4.
Cell Stem Cell
NFAT Pathway Regulates Early Lineage Specification
the Src inhibitor PP2 also efficiently abrogated the activation of
EMT markers caused by LIF withdrawal in a dose-dependent
manner (Figures S5G and S5H), indicating that these inhibitors
52 Cell Stem Cell 8, 46–58, January 7, 2011 ª2011 Elsevier Inc.
may maintain ESC self-renewal through the inhibition of EMT.
Taken together, our data indicate that EMT is an essential step
required for differentiation events.
Figure 6. NFAT-Src-Mediated EMT Is an Essential Step for Lineage Specification
(A) qRT-PCR analysis of EMT-relatedmarkers in CGR8 ESCs transiently overexpressingSrc F. The value in cells transfectedwith the control vector was set at 1.0.
Data are shown as the mean ± SD (n = 3).
(B) Redistribution of E-cadherin from the membrane to the cytoplasm stimulated by the inducible expression of Src F. Immunostaining of E-cadherin was con-
ducted in iSrcES cells after removing Tc for 3 days. Scale bars represent 25 mm.
(C) The time course of marker gene expression after inducible expression of NFATc3 in ESCs for different days. The qRT-PCR analysis was conducted to deter-
mine the gene expression level and the value in control cells was set at 1.0. Data are shown as the mean ± SD (n = 3).
(D) Blockade ofNFATc3-mediated ESC differentiation by the broad-spectrumMMP inhibitor GM6001. Phase contrast images of iNFATc3ES cells cultured under
the indicated conditions for 3 days are shown. Scale bars represent 100 mm.
(E) qRT-PCR analysis of marker expression levels in cells described in (D). The value in control cells was set at 1.0. Data are shown as the mean ± SD (n = 3).
See also Figure S5.
Cell Stem Cell
NFAT Pathway Regulates Early Lineage Specification
The Calcineurin-NFAT Signaling Cascade IsIndispensable for Early Embryo DevelopmentFinally, we investigated the expression and possible role for
NFAT in mouse embryos prior to implantation. Intriguingly,
immunofluorescence staining of mouse embryos revealed that
NFATc3 was primarily localized to the cytoplasm in ICM cells
at E3.5 and the epiblast cells at E4.5, which is the same as in
undifferentiated ESCs. However, NFATc3 was detected in the
nucleus of the differentiated trophectoderm of the blastocyst.
As a control, Oct4 was exclusively found in the nucleus of
epiblast cells (Figure 7A). This expression pattern indicates that
NFATc3 was inactive in undifferentiated cells and became acti-
vated upon the earliest lineage commitment. Thus, the differen-
tial activation of NFAT signaling is clearly identifiable among the
earliest lineages during mouse embryonic development.
To test whether this calcineurin-NFAT pathway is functionally
relevant to early lineage segregation, CsA was used to treat
8-cell mouse embryos, which markedly increased the percent-
age of embryos that stopped development at the morula stage
and attenuated the percentage of embryos developing into the
blastocoele formation (Figure 7B). Confocal immunofluores-
cence analysis showed that embryos treatedwith CsA contained
few Cdx2-positive trophectoderm cells, and the majority of cells
expressed Oct4 (Figure 7C). Moreover, NFATc3 was detected in
the nucleus of trophectoderm cells in control embryos and in the
cytoplasm of CsA-treated embryos (Figure S6A). These observa-
tions indicate that the inhibition of calcineurin-NFAT signaling
blocks embryo development at the morula stage. To test
whether the effect of CsA on trophectoderm formation is revers-
ible, 8-cell embryos were treated with CsA for 1.5 days and then
cultured for an additional 1.5 days in the control medium. Inter-
estingly, embryo development resumed, and numerous Cdx2-
positive cells appeared in the developing blastocyst after CsA
withdrawal. Moreover, CsA treatment after culture of the 8-cell
embryo in the control medium for 1.5 days, when the trophecto-
derm is thought to have already formed, did not disturb
Cell Stem Cell 8, 46–58, January 7, 2011 ª2011 Elsevier Inc. 53
Figure 7. Calcineurin-NFAT Signaling Is
Indispensable for Early Mouse Embryonic
Development
(A) Optical sections of themouse embryo immuno-
fluorescently stained with NFATc3 (red) and Oct4
(green) antibodies. Scale bars represent 25 mm.
(B) The effect of CsA and PP2 on the kinetics of
blastocoele formation. Mouse embryos were
treated from 8-cell stage (E2.5) and observed for
the presence of blastocoeles after 1.5 days.
(C) Confocal images of embryos grown from the
8-cell stage (E2.5) for 1.5 days in control medium
or in the medium containing CsA (2.5 mM).
Embryos were immunostained with antibodies
against Oct4 (green) and Cdx2 (red), respectively.
Scale bars represent 25 mm.
(D) The effect of PP2 on the trophectoderm devel-
opment. Embryos from the 8-cell stage were
treated with or without PP2 for 1.5 days and
were immunostained as in (C). Scale bars repre-
sent 25 mm.
(E) A bar chart showing the percentage of cell
numbers of the ICM and the trophectoderm of
embryos cultured under the conditions described
in (D). Data are shown as the mean ± SD.
(F) The effect of CsA and PP2 on the primitive
endoderm development. Confocal images of
embryos grown from E3.25 stage for 1.5 days in
control medium or in medium containing CsA
(1.75 mM) or PP2 (10 mM) are shown. Embryos
were immunostained with antibodies against
Nanog (green) and Gata4 (red), and nuclei were
counterstained with DAPI. Scale bars represent
25 mm.
(G) Bar chart showing the percentage of cell
numbers of the epiblast (green) and primitive
endoderm (red) of embryos cultured in the condi-
tions described in (F). Data are shown as the
mean ± SD.
(H) A proposed model for the signaling circuit
involved in ESC differentiation.
See also Figure S6.
Cell Stem Cell
NFAT Pathway Regulates Early Lineage Specification
trophectoderm development evidently (Figure S6B). Therefore,
the blockage of embryos at the morula stage might be due to
the inability of morula cells to differentiate into trophectoderm
cells when calcineurin-NFAT signaling is inhibited. We also
54 Cell Stem Cell 8, 46–58, January 7, 2011 ª2011 Elsevier Inc.
examined the effect of the Src inhibitor,
PP2, on trophectoderm formation of early
mouse embryos. PP2 treatment partially
blocked embryos at the morula stage
(Figure 7B), and in embryos developing
into blastocysts, PP2 treatment signifi-
cantly reduced Cdx2-positive trophecto-
derm cells and enhanced the Oct4-posi-
tive ICM cells compared to control
embryos (Figures 7D and 7E).
Finally, to determine whether calci-
neurin-NFAT signaling was also required
for formation of the primitive endoderm,
embryos of E3.25 were treated with CsA
or PP2 and costained for Nanog and
Gata4. Immunofluorescence staining revealed that CsA or PP2
treatment eliminated Gata4-positive primitive endoderm cells
and increased the Nanog-positive cell number (Figures 7F and
7G), implicating an indispensable role of the signaling pathway
Cell Stem Cell
NFAT Pathway Regulates Early Lineage Specification
for the segregation of the primitive endoderm from the ICM.
Taken together, these data suggest that the calcineurin-NFAT-
Src signaling cascade is activated and indispensable for extra-
embryonic development in early mouse embryos.
DISCUSSION
Our data establish a model whereby calcineurin-NFAT signaling
collaborates with Ras-Erk-AP-1 to activate Src, promote EMT,
and initiate lineage specification in mouse ESCs (Figure 7H).
We demonstrate that the calcineurin-NFAT-Src cascade criti-
cally regulates the transition of ESCs from self-renewal to lineage
commitment, having a similar role to Ras-Erk1/2 signaling, which
was previously shown to be crucial for ESC differentiation
(Kunath et al., 2007; Nichols et al., 2009). Interestingly, the acti-
vation of Ras-Erk1/2 and calcineurin-NFAT signaling resulted in
nearly identical gene expression profiles in mouse ESCs (Fig-
ure 3B; Figure S4E). Moreover, we found that the two pathways
were mutually dependent as shown by the fact that the blockade
of either one significantly attenuated the ESC differentiation
phenotypes induced by the other one (Figures S4F–S4I), which
is in line with a previous study conducted in cardiomyocytes
(Sanna et al., 2005). The cooperation that was discovered
between Ras-Erk1/2 and calcineurin-NFAT signaling provides
new insights into early ESC differentiation events; these two
distinct signaling pathways could be integrated to precisely
regulate ESC fates, depending upon whether both pathways
are concomitantly activated and which distinct sets of target
genes are activated.
Mechanistically, Ras-Erk1/2 augments NFAT transcriptional
activity through the activation of AP-1, which complexes with
NFAT at their coregulated genes (Macian et al., 2001). Impor-
tantly, we uncovered Src as one of these genes, which function-
ally phenocopied both NFATc3 and Ras in ESCs. Moreover, Src
inhibition abrogated NFAT- or Ras-triggered ESC differentiation
(Figures 5C and 5D; Figures S4J and S4K). The involvement of
Src family kinases in the regulation of ESC self-renewal and
differentiation has been previously studied; individual members
of the Src family play distinct and possibly opposite roles in
the control of ESC fates (Meyn et al., 2005; Meyn and Smithgall,
2009). One Src family member, c-Yes, was reported to be impor-
tant for ESC self-renewal (Anneren et al., 2004), whereas Src was
found to activate primitive ectoderm formation in mouse ESCs
(Meyn and Smithgall, 2009). The latter finding is consistent with
our observation that the expression of constitutively active Src
induces ESC differentiation. Therefore, we propose that calci-
neurin-NFAT and Ras-Erk1/2 signaling pathways converge to
regulate Src expression and that Src might be a critical mediator
for both signaling pathways. Nevertheless, we do not rule out
other possible mechanisms for ESC differentiation induced by
these two important signaling pathways.
The molecular mechanism by which Src regulates ESC differ-
entiation remains unknown. In this study, we propose that the
promotion of EMTmight account for the role of Src in ESC differ-
entiation. EMT regulates multiple critical processes during early
development, and development cannot proceed past the blas-
tula stage without EMT (Thiery and Sleeman, 2006). Recently,
the involvement of EMT in ESC differentiation has been
described (Eastham et al., 2007; Spencer et al., 2007), although
its role in ESC differentiation is not well defined. It is not clear
whether EMT is essential or merely a concomitant event during
differentiation. Our study indicates that differentiation stimuli
induce EMT prior to the differentiation processes and that inhibi-
tion of EMT blocks ESC differentiation, thus placing EMT as an
early and essential step in lineage specification. In addition, we
found that the GSK3b inhibitor CHIR99021 suppressed the
expression of EMT-related genes after LIF withdrawal (Figures
S5I and S5J) in a manner similar to the calcineurin-NFAT, Ras-
Erk1/2, Src, and MMP inhibitors. GSK3b regulates focal adhe-
sion kinase (FAK) (Bianchi et al., 2005; Kobayashi et al., 2006),
an important player in EMT, which might explain the effect of
GSK3b inhibitors in suppressing ESC differentiation (Ullmann
et al., 2008; Ying et al., 2008). Furthermore, some ESC-specific
transcription factors were found to bind promoters of EMT-
related genes (Figure S5K; Chen et al., 2008). Interestingly,
EMT inhibition was recently found to promote somatic cell
reprogramming (Lin et al., 2009). Therefore, EMT appears to be
essential for ESC differentiation, and its inhibition may promote
differentiated cells to revert to a pluripotent state.
Another important advance of this study is the discovery that
calcineurin-NFAT signaling is activated during the first differenti-
ation event in the preimplantation embryo and is essential for
early embryo development. We show that the subcellular locali-
zation of NFATc3 proteins is different between ICM and trophec-
toderm cells of the blastocyst, although they are detected in
both cell types. The NFAT activity is tightly regulated by phos-
phorylation and dephosphorylation; NFAT is activated by calci-
neurin-dependent dephosphorylation, which stimulates NFAT
translocation from the cytoplasm into the nucleus. Therefore,
our data showing that NFAT localizes to the nucleus of the
trophectoderm suggest that NFAT is activated in the trophecto-
derm. Moreover, the reversible and selective effect of CsA on
the trophectoderm formation at the restricted stage, as well as
the inhibitory effect of CsA and PP2 on the primitive formation,
argues for the specific and essential role of the calcineurin-
NFAT-Src cascade in extraembryonic lineage specification
during early embryonic development.
In summary, this study reveals a role for a well-characterized
signaling pathway in ESC differentiation and early embryonic
development. The identification of calcineurin-NFAT signaling
as one of the initial triggers of the ESC exit from self-renewal,
and of Src as a key player downstream of NFAT and Erk1/2 in
ESC early fate determination, provides new insights into how
the self-renewal of pluripotent stem cells is orchestrated.
Improved understanding of the role that EMT plays in ESC differ-
entiation also opens up additional avenues for developing more
efficient platforms for ESC programming and somatic cell
reprogramming.
EXPERIMENTAL PROCEDURES
Cell Culture and Differentiation
E14T and CGR8 mouse ESCs (gift of Austin Smith) were grown as previously
described (Li et al., 2010). To induce differentiation, CGR8 cells were cultured
with 0.1 mMRA (Sigma). Tetracycline (Tc)-inducible iDCnAES, iNFATc3ES, and
iSrcES cell lines were established with the Rosa-Tet system (Masui et al.,
2005), and the cells were maintained in medium containing Tc (0.5 mM). Exog-
enous gene expression was induced after removal of Tc.
Cell Stem Cell 8, 46–58, January 7, 2011 ª2011 Elsevier Inc. 55
Cell Stem Cell
NFAT Pathway Regulates Early Lineage Specification
Luciferase Assays
CGR8 ESCs were seeded at a density of 1.13 105 per well into 24-well tissue-
culture plates 24 hr before transfection. For NFAT:AP-1 reporter assays, LIF
was removed on different days, as indicated in the text. To determine the
endogenous activity, cells were transfected with reporter plasmids (500 ng)
and vector pRL-TK (20 ng, Promega) as a control for the transfection efficiency
via Lipofectamine 2000 (Invitrogen). To examine the effect of exogenously
expressed factors on the reporters, control vector and CA-NFATc3 expression
plasmids were cotransfected with reporters and pRL-TK. Cell extracts were
prepared 48 hr after transfection. Luciferase activity was evaluated with the
Dual-Luciferase Assay System (Promega) according to the manufacturer’s
recommendations.
ChIP Assays
CGR8 ESCs were cross-linked by incubating cells on plates with 1% formal-
dehyde, and ChIP assays were carried out as described (Li et al., 2010). Ten
percent of the total genomic DNA from the nuclear extract was used as the
input. The primers used are provided in Table S3.
EMSA
Nuclear extracts isolated from E14T ESCs overexpressing CA-NFATc3 or the
vector alone for 2 days were prepared, and EMSAs were performed as
described previously (Yang et al., 2008). In brief, the oligonucleotide probes
were synthesized and labeled with biotin at the 50 end of the forward oligonu-
cleotide. For the competitive assay, an additional 200-fold molar excess of the
unlabeled probe was added. For super-shift analysis, NFATc3 antibody (2 ml
per reaction) was added. Probe sequences are provided in Table S3.
RNAi and Oligonucleotides
siRNAs were introduced into cells according to the manufacturer’s instruc-
tions. The oligonucleotide sequences for mouse Ppp3r1, Src, and Lck are
provided in Table S3. The negative controls were obtained from Invitrogen
(Stealth RNAi Negative Control Med GC).
Western Blot Analysis
Protein (30 mg) from whole ESC extracts was used for western blot analysis as
described previously (Li et al., 2010).
RT-PCR and qRT-PCR Analysis
RT-PCR and qRT-PCR were conducted as previously described (Li et al.,
2010), and the primers used are provided in Table S3.
Generation of Teratomas
For teratoma generation, 53 106 ESCs were harvested and injected intramus-
cularly into nude mice. 6 to 8 weeks later, teratomas were harvested and pro-
cessed with hematoxylin and eosin staining.
Embryo Chimeras
CGR8 cells constitutively expressing the H2B-GFP fusion gene were cultured
in medium containing CsA (15 mM) without LIF for 40 days. These ESCs were
injected into blastocysts and then transferred to the uteri of pseudopregnant
mice. Embryos were dissected at E18.5.
Embryo Collection and Culture
Mouse embryos were collected from F1 hybrids between C57/B6 and DBA2
mice. For experiments studying trophectoderm development, zygotes in
cumulus masses were dissected from oviduct ampullae at E0.5 and incubated
in KSOM+AA (Millipore) for 2 days. Then embryos were cultured in KSOM+AA
containing CsA (2.5 mM) or PP2 (15 mM) for 1.5 days before fixation and stain-
ing. For experiments studying primitive endoderm development, zygotes were
dissected at E0.5 and incubated in KSOM+AA for 2.75 days. Then embryos
were cultured in N2B27 medium (Ying and Smith, 2003) with CsA (1.75 mM)
or PP2 (10 mM) for 1.5 days. The embryos were cultured in wells of 4-well
dishes (Nunc) without mineral oil covering.
Immunostaining
Embryos were fixed, permeabilized, and stained as described (Li et al., 2010).
Reconstructions of three-dimensional images from confocal sections and cell
counts were performed with Leica software and Adobe Photoshop.
56 Cell Stem Cell 8, 46–58, January 7, 2011 ª2011 Elsevier Inc.
Statistical Analysis
All values are shown as means ± SD. To determine the significance between
groups, comparison was made with Student’s t test. For all statistical tests,
the 0.05 confidence level was considered statistically significant. In all figures,
* denotes p < 0.05 and ** denotes p < 0.01 in an unpaired Student’s t test.
ACCESSION NUMBERS
Microarray data are accessible at the GEO database under accession number
GSE21378.
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental Procedures,
six figures, and three tables and can be found with this article online at
doi:10.1016/j.stem.2010.11.027.
ACKNOWLEDGMENTS
The authors wish to thank Drs. H. Niwa, A. Rao, S. Miyatake, X. Cao, S. Hanks,
T. Xu, and T. Takeya for generously providing plasmids; Drs. A. Smith and
G. Daley for their mouse ESC lines, and Drs. A. Smith, R.D. McKinnon, D. Li,
and A. Chong for critical reading of the manuscript. This study was supported
by grants from the National Natural Science Foundation (91019929,
30911130361, and 31000625) and National High Technology Research
and Development Program of China (2009CB941100, 2007CB947904,
2007CB948004, 2101CB945200, and 2007CB947101) and from Shanghai
Science & Technology Developmental Foundations (08dj1400502 and
07DZ22919). The study was also supported by the Shanghai Leading
Academic Discipline Project (S30201).
Received: June 2, 2010
Revised: October 2, 2010
Accepted: October 25, 2010
Published: January 6, 2011
REFERENCES
Anneren, C., Cowan, C.A., and Melton, D.A. (2004). The Src family of tyrosine
kinases is important for embryonic stem cell self-renewal. J. Biol. Chem. 279,
31590–31598.
Bianchi, M., De Lucchini, S., Marin, O., Turner, D.L., Hanks, S.K., and Villa-
Moruzzi, E. (2005). Regulation of FAK Ser-722 phosphorylation and kinase
activity by GSK3 and PP1 during cell spreading and migration. Biochem. J.
391, 359–370.
Brabek, J., Constancio, S.S., Shin, N.Y., Pozzi, A., Weaver, A.M., and Hanks,
S.K. (2004). CAS promotes invasiveness of Src-transformed cells. Oncogene
23, 7406–7415.
Cavallaro, U., and Christofori, G. (2004). Cell adhesion and signalling by
cadherins and Ig-CAMs in cancer. Nat. Rev. Cancer 4, 118–132.
Chazaud, C., Yamanaka, Y., Pawson, T., and Rossant, J. (2006). Early lineage
segregation between epiblast and primitive endoderm in mouse blastocysts
through the Grb2-MAPK pathway. Dev. Cell 10, 615–624.
Chen, X., Xu, H., Yuan, P., Fang, F., Huss,M., Vega, V.B.,Wong, E., Orlov, Y.L.,
Zhang, W., Jiang, J., et al. (2008). Integration of external signaling pathways
with the core transcriptional network in embryonic stem cells. Cell 133,
1106–1117.
Chin, E.R., Olson, E.N., Richardson, J.A., Yang, Q., Humphries, C., Shelton,
J.M., Wu, H., Zhu, W., Bassel-Duby, R., and Williams, R.S. (1998). A calci-
neurin-dependent transcriptional pathway controls skeletal muscle fiber
type. Genes Dev. 12, 2499–2509.
Clipstone, N.A., and Crabtree, G.R. (1992). Identification of calcineurin as a key
signalling enzyme in T-lymphocyte activation. Nature 357, 695–697.
Crabtree, G.R. (1999). Generic signals and specific outcomes: signaling
through Ca2+, calcineurin, and NF-AT. Cell 96, 611–614.
Cell Stem Cell
NFAT Pathway Regulates Early Lineage Specification
Crabtree, G.R., and Olson, E.N. (2002). NFAT signaling: choreographing the
social lives of cells. Cell Suppl. 109, S67–S79.
Crabtree, G.R., and Schreiber, S.L. (2009). SnapShot: Ca2+-calcineurin-NFAT
signaling. Cell 138, 210.
de la Pompa, J.L., Timmerman, L.A., Takimoto, H., Yoshida, H., Elia, A.J.,
Samper, E., Potter, J., Wakeham, A., Marengere, L., Langille, B.L., et al.
(1998). Role of the NF-ATc transcription factor in morphogenesis of cardiac
valves and septum. Nature 392, 182–186.
Ding, S., Wu, X., Li, G., Han, M., Zhuang, Y., and Xu, T. (2005). Efficient trans-
position of the piggyBac (PB) transposon in mammalian cells and mice. Cell
122, 473–483.
Eastham, A.M., Spencer, H., Soncin, F., Ritson, S., Merry, C.L., Stern, P.L.,
and Ward, C.M. (2007). Epithelial-mesenchymal transition events during
human embryonic stem cell differentiation. Cancer Res. 67, 11254–11262.
Graef, I.A., Chen, F., Chen, L., Kuo, A., and Crabtree, G.R. (2001). Signals
transduced by Ca(2+)/calcineurin and NFATc3/c4 pattern the developing
vasculature. Cell 105, 863–875.
Graef, I.A., Wang, F., Charron, F., Chen, L., Neilson, J., Tessier-Lavigne, M.,
and Crabtree, G.R. (2003). Neurotrophins and netrins require calcineurin/
NFAT signaling to stimulate outgrowth of embryonic axons. Cell 113, 657–670.
Guarino, M. (2010). Src signaling in cancer invasion. J. Cell. Physiol. 223,
14–26.
Hamadi, A., Deramaudt, T.B., Takeda, K., and Ronde, P. (2009). Src activation
and translocation from focal adhesions to membrane ruffles contribute to
formation of new adhesion sites. Cell. Mol. Life Sci. 66, 324–338.
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
with RNA interference. Nature 442, 533–538.
Jiao, X., Katiyar, S., Liu, M., Mueller, S.C., Lisanti, M.P., Li, A., Pestell, T.G.,
Wu, K., Ju, X., Li, Z., et al. (2008). Disruption of c-Jun reduces cellular migration
and invasion through inhibition of c-Src and hyperactivation of ROCK II kinase.
Mol. Biol. Cell 19, 1378–1390.
Kim, J., Chu, J., Shen, X., Wang, J., and Orkin, S.H. (2008). An extended
transcriptional network for pluripotency of embryonic stem cells. Cell 132,
1049–1061.
Kobayashi, T., Hino, S., Oue, N., Asahara, T., Zollo, M., Yasui, W., and Kikuchi,
A. (2006). Glycogen synthase kinase 3 and h-prune regulate cell migration by
modulating focal adhesions. Mol. Cell. Biol. 26, 898–911.
Kumagai, N., Ohno, K., Tameshige, R., Hoshijima, M., Yogo, K., Ishida, N., and
Takeya, T. (2004). Induction of mouse c-src in RAW264 cells is dependent on
AP-1 and NF-kappaB and important for progression to multinucleated cell
formation. Biochem. Biophys. Res. Commun. 325, 758–768.
Kunath, T., Saba-El-Leil, M.K., Almousailleakh, M., Wray, J., Meloche, S., and
Smith, A. (2007). FGF stimulation of the Erk1/2 signalling cascade triggers tran-
sition of pluripotent embryonic stem cells from self-renewal to lineage commit-
ment. Development 134, 2895–2902.
Li, L., Sun, L., Gao, F., Jiang, J., Yang, Y., Li, C., Gu, J.,Wei, Z., Yang, A., Lu, R.,
et al. (2010). Stk40 links the pluripotency factor Oct4 to the Erk/MAPK pathway
and controls extraembryonic endoderm differentiation. Proc. Natl. Acad. Sci.
USA 107, 1402–1407.
Lin, T., Ambasudhan, R., Yuan, X., Li, W., Hilcove, S., Abujarour, R., Lin, X.,
Hahm, H.S., Hao, E., Hayek, A., and Ding, S. (2009). A chemical platform for
improved induction of human iPSCs. Nat. Methods 6, 805–808.
Lu, C.W., Yabuuchi, A., Chen, L., Viswanathan, S., Kim, K., and Daley, G.Q.
(2008). Ras-MAPK signaling promotes trophectoderm formation from embry-
onic stem cells and mouse embryos. Nat. Genet. 40, 921–926.
Macian, F., Garcıa-Rodrıguez, C., and Rao, A. (2000). Gene expression elicited
by NFAT in the presence or absence of cooperative recruitment of Fos and
Jun. EMBO J. 19, 4783–4795.
Macian, F., Lopez-Rodrıguez, C., and Rao, A. (2001). Partners in transcription:
NFAT and AP-1. Oncogene 20, 2476–2489.
Mandal, M., Myers, J.N., Lippman, S.M., Johnson, F.M., Williams, M.D.,
Rayala, S., Ohshiro, K., Rosenthal, D.I., Weber, R.S., Gallick, G.E., and
El-Naggar, A.K. (2008). Epithelial to mesenchymal transition in head and
neck squamous carcinoma: association of Src activation with E-cadherin
down-regulation, vimentin expression, and aggressive tumor features.
Cancer 112, 2088–2100.
Martin, G.R. (1981). Isolation of a pluripotent cell line from early mouse
embryos cultured in medium conditioned by teratocarcinoma stem cells.
Proc. Natl. Acad. Sci. USA 78, 7634–7638.
Masui, S., Shimosato, D., Toyooka, Y., Yagi, R., Takahashi, K., and Niwa, H.
(2005). An efficient system to establish multiple embryonic stem cell lines
carrying an inducible expression unit. Nucleic Acids Res. 33, e43.
Meyn, M.A., 3rd, and Smithgall, T.E. (2009). Chemical genetics identifies c-Src
as an activator of primitive ectoderm formation inmurine embryonic stem cells.
Sci. Signal. 2, ra64.
Meyn, M.A., 3rd, Schreiner, S.J., Dumitrescu, T.P., Nau, G.J., and Smithgall,
T.E. (2005). SRC family kinase activity is required for murine embryonic stem
cell growth and differentiation. Mol. Pharmacol. 68, 1320–1330.
Molkentin, J.D., Lu, J.R., Antos, C.L., Markham, B., Richardson, J., Robbins,
J., Grant, S.R., andOlson, E.N. (1998). A calcineurin-dependent transcriptional
pathway for cardiac hypertrophy. Cell 93, 215–228.
Muller, M.R., Sasaki, Y., Stevanovic, I., Lamperti, E.D., Ghosh, S., Sharma, S.,
Gelinas, C., Rossi, D.J., Pipkin, M.E., Rajewsky, K., et al. (2009). Requirement
for balanced Ca/NFAT signaling in hematopoietic and embryonic develop-
ment. Proc. Natl. Acad. Sci. USA 106, 7034–7039.
Nichols, J., Silva, J., Roode, M., and Smith, A. (2009). Suppression of Erk
signalling promotes ground state pluripotency in the mouse embryo.
Development 136, 3215–3222.
Rossant, J. (2007). Stem cells and lineage development in the mammalian
blastocyst. Reprod. Fertil. Dev. 19, 111–118.
Rossant, J., Chazaud, C., and Yamanaka, Y. (2003). Lineage allocation and
asymmetries in the early mouse embryo. Philos. Trans. R. Soc. Lond. B Biol.
Sci. 358, 1341–1348, discussion 1349.
Sanna, B., Bueno, O.F., Dai, Y.S., Wilkins, B.J., and Molkentin, J.D. (2005).
Direct and indirect interactions between calcineurin-NFAT and MEK1-extra-
cellular signal-regulated kinase 1/2 signaling pathways regulate cardiac
gene expression and cellular growth. Mol. Cell. Biol. 25, 865–878.
Smith, A.G. (2001). Embryo-derived stem cells: of mice and men. Annu. Rev.
Cell Dev. Biol. 17, 435–462.
Spencer, H.L., Eastham, A.M., Merry, C.L., Southgate, T.D., Perez-Campo, F.,
Soncin, F., Ritson, S., Kemler, R., Stern, P.L., and Ward, C.M. (2007). E-cad-
herin inhibits cell surface localization of the pro-migratory 5T4 oncofetal
antigen in mouse embryonic stem cells. Mol. Biol. Cell 18, 2838–2851.
Stankunas, K., Graef, I.A., Neilson, J.R., Park, S.H., and Crabtree, G.R. (1999).
Signaling through calcium, calcineurin, and NF-AT in lymphocyte activation
and development. Cold Spring Harb. Symp. Quant. Biol. 64, 505–516.
Tan, T.K., Zheng, G., Hsu, T.T., Wang, Y., Lee, V.W., Tian, X., Wang, Y., Cao,
Q., Wang, Y., and Harris, D.C. (2010). Macrophage matrix metalloproteinase-9
mediates epithelial-mesenchymal transition in vitro in murine renal tubular
cells. Am. J. Pathol. 176, 1256–1270.
Thiery, J.P., and Sleeman, J.P. (2006). Complex networks orchestrate epithe-
lial-mesenchymal transitions. Nat. Rev. Mol. Cell Biol. 7, 131–142.
Thiery, J.P., Acloque, H., Huang, R.Y., and Nieto, M.A. (2009). Epithelial-
mesenchymal transitions in development and disease. Cell 139, 871–890.
Ullmann, U., Gilles, C., De Rycke, M., Van de Velde, H., Sermon, K., and
Liebaers, I. (2008). GSK-3-specific inhibitor-supplemented hESC medium
prevents the epithelial-mesenchymal transition process and the up-regulation
of matrix metalloproteinases in hESCs cultured in feeder-free conditions. Mol.
Hum. Reprod. 14, 169–179.
Wu, X., Nguyen, B.C., Dziunycz, P., Chang, S., Brooks, Y., Lefort, K., Hofbauer,
G.F., and Dotto, G.P. (2010). Opposing roles for calcineurin and ATF3 in squa-
mous skin cancer. Nature 465, 368–372.
Yamanaka, Y., Lanner, F., and Rossant, J. (2010). FGF signal-dependent
segregation of primitive endoderm and epiblast in the mouse blastocyst.
Development 137, 715–724.
Yang, L., Jiang, Y., Wu, S.F., Zhou, M.Y., Wu, Y.L., and Chen, G.Q. (2008).
CCAAT/enhancer-binding protein alpha antagonizes transcriptional activity
Cell Stem Cell 8, 46–58, January 7, 2011 ª2011 Elsevier Inc. 57
Cell Stem Cell
NFAT Pathway Regulates Early Lineage Specification
of hypoxia-inducible factor 1 alpha with direct protein-protein interaction.
Carcinogenesis 29, 291–298.
Ying, Q.L., and Smith, A.G. (2003). Defined conditions for neural commitment
and differentiation. Methods Enzymol. 365, 327–341.
Ying, Q.L., Nichols, J., Chambers, I., and Smith, A. (2003). BMP induction of Id
proteins suppresses differentiation and sustains embryonic stem cell self-
renewal in collaboration with STAT3. Cell 115, 281–292.
58 Cell Stem Cell 8, 46–58, January 7, 2011 ª2011 Elsevier Inc.
Ying, Q.L., Wray, J., Nichols, J., Batlle-Morera, L., Doble, B., Woodgett, J.,
Cohen, P., and Smith, A. (2008). The ground state of embryonic stem cell
self-renewal. Nature 453, 519–523.
Yu, H., van Berkel, T.J., and Biessen, E.A. (2007). Therapeutic potential of
VIVIT, a selective peptide inhibitor of nuclear factor of activated T cells, in
cardiovascular disorders. Cardiovasc. Drug Rev. 25, 175–187.
Cell Stem Cell
Article
Proliferative Neural Stem Cells Have HighEndogenous ROS Levels that Regulate Self-Renewaland Neurogenesis in a PI3K/Akt-Dependant MannerJanel E. Le Belle,1 Nicolas M. Orozco,1 Andres A. Paucar,1 Jonathan P. Saxe,2 Jack Mottahedeh,1 April D. Pyle,3,4,5
Hong Wu,2,4,5,6 and Harley I. Kornblum1,2,4,5,*1NPI-Semel Institute for Neuroscience & Human Behavior and Department of Psychiatry and Biobehavioral Sciences2Department of Molecular and Medical Pharmacology3Department of Microbiology, Immunology, and Molecular Genetics4Eli and Edythe Broad Center of Regenerative Medicine and Stem Cell Research5Jonsson Comprehensive Cancer Center6Institute for Molecular MedicineDavid Geffen School of Medicine, University of California Los Angeles, Los Angeles, CA 90095, USA
*Correspondence: harley@ucla.edu
DOI 10.1016/j.stem.2010.11.028
SUMMARY
The majority of research on reactive oxygen species(ROS) has focused on their cellular toxicities. Stemcells generally have been thought to maintain lowlevels of ROS as a protection against these pro-cesses. However, recent studies suggest that ROScan also play roles as secondmessengers, activatingnormal cellular processes. Here, we investigatedROS function in primary brain-derived neural progen-itors. Somewhat surprisingly, we found that prolifer-ative, self-renewing multipotent neural progenitorswith the phenotypic characteristics of neural stemcells (NSC) maintained a high ROS status and werehighly responsive to ROS stimulation. ROS-mediatedenhancements in self-renewal and neurogenesiswere dependent on PI3K/Akt signaling. Pharmaco-logical or genetic manipulations that diminishedcellular ROS levels also interfered with normal NSCand/or multipotent progenitor function both in vitroand in vivo. This study has identified a redox-medi-ated regulatory mechanism of NSC function thatmay have significant implications for brain injury,disease, and repair.
INTRODUCTION
Oxidative stress caused by the cellular accumulation of reactive
oxygen species (ROS) is a major contributor to disease and to
cell death. In contrast to the damaging effects of ROS, there is
evidence that in some systems ROS at lower, nontoxic levels
can actually promote cell proliferation and survival (Blanchetot
and Boonstra, 2008; Chiarugi and Fiaschi, 2007; Leslie, 2006).
These findings suggest a much more complex role for redox
balance in cellular biology than was first understood by models
of oxidative stress. For example, in the hematopoietic system
a low endogenous cellular ROS status has been associated
with maintaining the quiescence of hematopoietic stem cells
(HSCs), whereas a higher ROS state is associated with a greater
proliferation leading to a premature exhaustion of self-renewal in
these cells (Jang and Sharkis, 2007). This has led to the hypoth-
esis that keeping ROS levels low within the stem cell niche is an
important feature of ‘‘stemness’’ that is directly related to the
relatively quiescent state of stem cells in vivo. Although it is
thought that the resident neural stem cells (NSCs) within the
neurogenic niches of the brain are also relatively quiescent, it
is not yet known how ROS status or ROS stimuli may affect
this population of cells. One might hypothesize that NSCs would
utilize and defend against ROS in the same manner as HSCs,
maintaining low endogenous levels of ROS. However, despite
similarities in the core functions of self-renewal and multipo-
tency, HSCs and NSCs also have many biological differences.
For example, the premature replicative senescence observed
in HSCs as a result of the hyperproliferation caused by deletion
of the tumor suppressor gene PTEN is not observed in PTEN-
deleted NSCs (Zhang et al., 2006; Yilmaz et al., 2006; Groszer
et al., 2006).
Emerging evidence now suggests that in addition to the
passive production of ROS by the mitochondria, we have
evolved a redox mechanism to utilize cellular ROS in a directed
manner by NADPH oxidase (NOX) enzymes, which are the
predominant source of ROS in many cells (Lambeth et al.,
2008). NOX was originally characterized in phagocytes, which
utilize a NOX-generated burst of superoxide to defend against
pathogens. It has now become clear that other cell types utilize
NOX-generated ROS as second messengers in tightly con-
trolled signal transduction networks. The realization that ROS
production is an essential component of cellular signaling has
led to the discovery that many ligands essential to normal cell
function including peptide and angiogenic growth factors,
hormones, and interleukins require the generation of ROS via
NOX activation in some nonphagocytotic cells (Kwon et al.,
2004; Wang and Lou, 2009; Garrido and Griendling, 2009;
Cell Stem Cell 8, 59–71, January 7, 2011 ª2011 Elsevier Inc. 59
A B
D
C
E F
Figure 1. Stimulation of Neurosphere
Cultures by Reactive Oxygen Species
Promotes Proliferation and Self-Renewal
(A) A diagram of NADPH oxidase (NOX) signaling in
the PI3K/Akt/mTOR pathway.
(B) Serial clonal density neurosphere formation,
sphere diameters, and multipotency in response
to exogenous H2O2 stimulation.
(C) Clonal density neurosphere formation inmouse
(ms) embryonic day 14 (E14) cortical cultures,
adult subventricular zone (aSVZ) cultures, and
human fetal (HF) cortical cultures as a percentage
of control (untreated) conditions.
(D) Cells sorted according to their relative endoge-
nous ROS levels or unselected (US) cells from
primary tissue microdissections.
(E) Secondary sorts of cells according to their
relative endogenous ROS levels.
(F) Human ES-derived monolayer cell prolifera-
tion after sorting according to relative endogenous
ROS levels.
Data expressed as mean ± SEM. See also Figures
S1A–S1C.
Cell Stem Cell
Neural Stem Cell Redox Regulation
Goldstein et al., 2005; Behrens et al., 2008). The NOX-stimu-
lated production of ROS, in turn, can activate pathways that
have been previously associated with enhanced cell prolifera-
tion and survival, including the MAPK and PI3K/Akt pathways
(Figure 1A; Kwon et al., 2004; Sundaresan et al., 1995). NOX
isoforms have been identified in a number of different tissues,
including the brain, although apart from the deleterious produc-
tion of high levels of ROS in brain injuries, their function in the
CNS is not known (Infanger et al., 2006; Lambeth et al., 2007;
Park et al., 2008).
In this study we sought to determine the role of ROS-
mediated signaling in NSCs. We have found a surprising sensi-
tivity to redox regulation in the neural stem cell-enriched pool
of cells compared to the more generalized proliferative pool
of limited progenitors, as shown by the fact that the manipula-
tion of cellular ROS levels predominantly affects self-renewal
and neurogenesis. In contrast to what has been previously
60 Cell Stem Cell 8, 59–71, January 7, 2011 ª2011 Elsevier Inc.
fa
s
R
EafT
p
m
o
m
a
a
o
a
observed with O-2A progenitors (Smith
et al., 2000; Power et al., 2002; Li
et al., 2007), we have found that a
decrease in normal cellular ROS levels
can have an unexpectedly negative
impact on self-renewal and neurogene-
sis both in vitro and in vivo. We observed
a higher level of endogenous ROS in
NSCs and within the neural stem cell
niche, the SVZ, in vivo. We found that
the regulation of endogenous ROS levels
in NSCs was highly dependent on
NADPH oxidase and PI3k/Akt signaling.
The prominent effects of cellular ROS
levels that we have observed on neural
stem and progenitor cell function may
be of particular relevance in injury and
disease because of the large number of
ctors that may influence and deregulate ROS-mediated
ignaling.
ESULTS
levated ROS Enriches for Self-Renewing Neural Stemnd Progenitor Cells in Clonal Neurosphere Culturesrom Different Species and Developmental Ageshe addition of low, nontoxic concentrations of hydrogen
eroxide (H2O2) to culture media produced a large increase in
ultipotent—capable of producing neurons, astrocytes, and
ligodendrocytes—clonal density neurosphere formation over
ultiple serial passages (Figure 1B; p = 0.001). There was
more modest increase in overall cell proliferation (Figure S1A
vailable online). Exogenous ROS had a similar stimulatory effect
n clonal neurosphere formation in embryonic and adult mouse
nd fetal human neurosphere cultures (Figure 1C; p < 0.001).
A
C
B
F
D
EGFR+GFAP+
CD24-ID1+GFAP+ Lex+
% E
GF
R-/
ID1-
/Lex
-N
egat
ive
po
pu
lati
on
Endogenous ROS levels in NSC-enriched populations
Nestin GFAP DCX Sox2 Mash1 Dlx20
100
200
300
400
500
% R
OS
lo
ROShi cell phenotypes
020406080
100120140160180200
P1 P2 P3 P4
% U
nse
lect
ed C
ells
Sphere Formation (self-renewal)Sphere Diameter (proliferation)Multipotency
Serial Clonal ROShi Neurosphere Formation
050
100150200250300350400
EGFR+GFAP+
CD24-ID1+GFAP+
% G
FA
P -
Neg
ativ
e p
op
ula
tio
n
Endogenous ROS levels in NSC-enriched populations
020406080
100120140160180200
0
3
6
9
12
15
18
Norm Hyp
Rat
io E
xpre
ssio
n
NOX2 (gp91phox) expression
EClonal neurosphere formation
0
5
10
15
20
25
30
C H A
Normoxia Hypoxia
% C
ells
see
ded
C H A
Figure 2. Neural Stem Cells Are Associated
with a High-ROS Status and NOX Is a Signif-
icant Endogenous Source of Cellular ROS
Regulating NSC Function in Low-Oxygen
Conditions
(A) Cells derived from adult SVZ were sorted for
the highest (top 10%) endogenous ROS levels
via DCFDA dye fluorescence (propidium iodide-
negative population) or for unselected (US) propi-
dium iodide-negative cells and then serially
cultured at clonal density to determine relative
stem cell numbers.
(B) The phenotypes of the high and low ROS cells
immediately after sortingwere evaluated by immu-
nocytochemistry and flow cytometry.
(C) The relative endogenous ROS levels were
measured in the EGFR+GFAP+CD24� and ID1+
GFAP+ cell populations (the stem cell containing
fractions) compared to the GFAP-negative popu-
lations.
(D) The relative endogenous ROS levels were
measured in the EGFR+GFAP+CD24�, ID1+
GFAP+, and Lex+ cells compared to the cells
negative for those markers.
(E) Relative expression of the NOX2 homolog in
neurosphere cultures grown in normoxic, room-
air-oxygen levels (Norm) or in low-oxygen (4%)
conditions (Hyp) normalized to 18S housekeeping
expression.
(F) Clonal neurosphere formation in room-air-
oxygen levels (Normoxia) or low oxygen (Hypoxia)
in control media (C) or treated with hydrogen
peroxide (H) or the NOX inhibitor Apocynin (A).
Data expressed as mean ± SEM. See also Figures
S2A and S2B.
Cell Stem Cell
Neural Stem Cell Redox Regulation
Hematopoietic stem cells have relatively low levels of endog-
enous ROS (Jang and Sharkis, 2007). To determine whether
neural stem cells were also low-ROS cells, we used FACS and
the ROS-sensitive dye DCFDA to separate cells into ROShi (top
10%) and ROSlo (bottom 10%) populations and assessed their
serial clonal density neurosphere-forming capacity. The ROShi
population contained almost all of the multipotent sphere-form-
ing cells in primary and secondary clonal cultures (p < 0.001; Fig-
ure 1D). We replicated this finding with multiple different ROS-
sensitive dyes (see Figure S1B). In addition, a high-ROS status
provided an enrichment in clonal neurosphere formation com-
pared to unselected (US), sorted cells from the same sample.
ROSlo cells formed only primary clonal neurospheres and there-
fore displayed a limited capacity for self-renewal. Culture and
resorting of sorted cells demonstrated that ROShi cells gave
rise to both ROShi and ROSlo cells in secondary neurospheres
but ROSlo cells were not capable of giving rise to ROShi cells (Fig-
ure 1E). Consistent with these results in murine cells, we also
observed that the ROShi population in human ESC-derived
neural progenitors had a greater proliferative capacity compared
to ROSlo or unselected cells from the same sample (p < 0.001;
Figure 1F).
Elevated ROS Levels and NOX Signaling Are Associatedwith Increased NSC EnrichmentSerial clonal density neurosphere formation (self-renewal),
sphere diameter (proliferation), and multipotency were assessed
in the ROShi cells compared to unselected cells from the
same samples over multiple passages. ROShi cells were highly
enriched for clonal neurosphere-forming cells at all passages
although a gradual decrease in this enrichment was observed
(p < 0.001; Figure 2A). There was also an initial significant
increase in neurosphere diameter (p < 0.05), but this returned
to control levels with successive passages. ROShi spheresmain-
tained a high level of multipotency over serial clonal passages.
In agreement with our data utilizing exogenous ROS stimulation,
a high endogenous ROS status was also associated with a
greater positive effect on self-renewing divisions than on overall
proliferation.
We next sought to identify differences in cellular phenotypes
between the ROShi and ROSlo populations because they dis-
played different capacities for long-term clonal self-renewal.
Therefore, we sorted primary adult SVZ cells for three different
neural stem cell-enriching marker sets: (1) EGFR+GFAP+CD24�
cells (Pastrana et al., 2009), (2) ID1+GFAP+ cells (Nam and Ben-
ezra, 2009), and (3) Lex (SSEA1)+ cells (Capela and Temple,
2002). Then, we evaluated their relative endogenous ROS levels.
We found that the enriched populations maintained significantly
elevated endogenous ROS levels compared to the negative,
non-NSC enriched populations from the same samples in each
case, indicating that the ROShi fraction contains the neural
stem cell fraction (p < 0.001; Figures 2C and 2D). The ‘‘stem
cell astrocytes’’ (EGFR+GFAP+CD24� cells) had 48% higher
ROS levels than the (EGFR+GFAP�CD24�) transit-amplifying
Cell Stem Cell 8, 59–71, January 7, 2011 ª2011 Elsevier Inc. 61
A B
D
C
0
2
4
6
8
10
12
% C
ells
see
ded
+H2O2 +H2O2
Low GF GF
Clonal neurosphereformation
0
1
2
3
4
Low GF GF
DC
FD
A [
RF
U x
103 ]
Endogenous ROS levels Clonal neurosphere formation with NOX inhibition and rescue
120
0
20
40
60
80
100
DPI D+HE14
% C
on
tro
l (u
ntr
eate
d)
DPI D+HaSVZ
8101214
See
ded
Serial Clonal Neurosphere Formation
6080
100
Sp
her
es
Multipotency
E
0246
WT MUT WT MUT WT MUT MUT
P1 P2 P3+H2O2
% C
ells
S
02040
WT MUT WT MUT MUT
P1 P2+H2O2
% T
ota
l S
Figure 3. Reactive Oxygen Species Are
Required for Stimulation of Normal Neural
Stem Cell Self-Renewal
(A) Clonal neurosphere formation by adult SVZ cells
in low growth factor media (Low GF; 1/20th normal
growth factor concentrations) is compared to
normal growth factor concentrations (GF) or sup-
plemented with hydrogen peroxide (H2O2).
(B) The corresponding endogenous ROS levels
detected by DCFDA dye and expressed in relative
fluorescent units (RFU) in the same culture condi-
tions described in (A).
(C) Clonal neurosphere formation in response to
NOX inhibition (DPI) and rescue with hydrogen
peroxide (H) in cells from embryonic and adult
brain.
(D) Serial clonal neurosphere formation by NOX2
mutant (MUT) and wild-type (WT) cells with H2O2
rescue.
(E) Multipotency of NOX2 MUT and WT neuro-
spheres over serial clonal passages with H2O2
rescue.
Data expressed as mean ± SEM. See also Figures
S3A and S3B.
Cell Stem Cell
Neural Stem Cell Redox Regulation
cells and approximately 200% more than the EGFR-negative
niche astrocyte-containing fraction of cells. The ID1+GFAP+
cells also had 57% higher ROS levels than the GFAP-negative
cells.
Clonal neurosphere formation was greatly enhanced by the
addition of exogenous ROS to the stem cell-enriched fractions
derived from mouse SVZ, whereas the stem cell-negative frac-
tions had limited or no response, an effect that was inhibited
by the NOX inhibitor apocynin (Figure S2A).
When cells were sorted directly from the SVZ according to
their ROS status and then analyzed for other markers, we found
no differences in the expression of Dlx2 in ROShi compared to
ROSlo cells, whereas Mash1-positive cells were enriched in the
ROSlo fraction (Figure 2B). These data suggest that transit-
amplifying cells are not responsible for differences observed in
neurosphere formation between the two populations. The ROShi
population was significantly enriched for cells expressing nestin
and doublecortin (DCX). No significant differences in Sox2- or
GFAP-expressing populations were observed.
The previous experiments were performed under room-
oxygen conditions. However, low-oxygen conditions are known
to stimulate NSC self-renewal (Studer et al., 2000). We found
that low, physiological oxygen conditions (4% O2) resulted in
elevated endogenous ROS levels (Figure S2B), consistent with
findings of others in different cell models (Guo et al., 2008),
increased clonal neurosphere formation (Figure 2H; p < 0.001),
and upregulation of the NOX2 homolog (Figure 2G; p < 0.01).
Conversely, lowering endogenous ROS levels in the low-oxygen
cultures through NOX inhibition eliminated the positive effects
of hypoxia and resulted in decreased clonal density neurosphere
formation (Figure 2H). These data suggest that the enhancement
62 Cell Stem Cell 8, 59–71, January 7, 2011 ª2011 Elsevier Inc.
of self-renewal by lower-oxygen concentrations is at least
partially mediated through enhanced NOX activity, which in
turn leads to elevated ROS levels.
ROS Augments Growth and Trophic Factor Stimulationand Is Required for Normal NSC Self-RenewalWe next wanted to determine whether NOX-generated ROS
played an important role in facilitating growth factor signal trans-
duction. To do this we placed neurosphere-derived cells in low
concentrations of EGF and bFGF, which led to a marked reduc-
tion in neurosphere formation (Figure 3A). However, clonal neu-
rosphere formation could be restored to levels observed with
high growth factor concentrations by supplementing low growth
factor conditions with exogenous ROS (p < 0.001; Figure 3A).
The addition of exogenous ROS to the low-growth factor
cultures elevated intracellular ROS levels to those observed in
the high growth factor conditions (Figure 3B). No clonal neuro-
spheres were formed in cultures without any growth factors
even with the addition of exogenous ROS (data not shown), indi-
cating that ROS on its own is not sufficient to replace growth
factor-initiated signaling.
Because the addition of small amounts of ROS resulted in
a gain of function, we next investigated the effects of a loss of
function in NOX signaling. In growth factor-supplemented SVZ
neurosphere cultures, we found that NOX inhibition (DPI) signif-
icantly decreases clonal neurosphere formation but this inhibi-
tion can be rescued by adding exogenous ROS (H2O2) back to
the culture medium. We also observed that cells derived from
the SVZ of NOX2 mutant mice had significantly lower endoge-
nous ROS levels (Figure S3A) and subsequently displayed
a significantly diminished NSC self-renewal and multipotency
Cell Stem Cell
Neural Stem Cell Redox Regulation
over serial clonal passages (Figures 3D and 3E; p < 0.01). Mutant
neurospheres produce approximately 29% more glial-only
spheres (astrocytes and oligodendrocytes) compared to wild-
type cultures (Figure S3B). Clonal neurosphere formation and
multipotency in the NOX2 mutants could also be significantly
rescued by the addition of exogenous ROS (H2O2) in the mutant
cultures (Figures 3A and 3B).
Because NOX has been implicated in both growth factor and
neurotrophin signaling, we next examined whether it may play
a role in the proliferative effects of brain-derived neurotrophic
factor (BDNF) on neural stem and progenitor cells (Islam et al.,
2009). In the presence of standard concentrations of NSC
growth factors (EGF and bFGF), we observed that BDNF could
significantly increase clonal neurosphere formation. Therefore,
we used inhibition of NADPH oxidase or treatment with the
antioxidant N-acetyl-cysteine (NAC) in order to determine that
NOX signaling played a significant role in the positive effects of
BDNF on clonal neurosphere formation (p < 0.001; Figure 4A).
In addition, we demonstrated that endogenous superoxide
(the ROS species directly produced by NOX) was increased
upon BDNF treatment, which could be blocked by NOX inhibition
(p < 0.001; Figure 4B). However, BDNF was not able to stimulate
NSC self-renewal in cells derived from NOX2 mutant mice but
was stimulatory only to wild-type cells (p < 0.01; Figure S4A),
suggesting that NOX-dependent signaling plays a significant
role in the stimulatory effects of BDNF on neural stem and
progenitor proliferation.
NOX Regulation of Neural Stem and Progenitor CellsIs Dependent on PI3K/Akt/mTOR SignalingPrevious studies have suggested that ROS can activate the
PI3K/Akt/mTOR pathway through the reversible inactivation of
the PTEN protein (Kwon et al., 2004; Leslie, 2006). Consistent
with this, we found in neurospheres that the addition of stimula-
tory concentrations of H2O2 induced direct oxidation of the PTEN
protein (Figure 4C). To more directly assess the requirement for
PTEN expression in the mechanisms underlying the stimulatory
effect of ROS, we used cells derived from PTEN-deficient,
PTEN heterozygous, and wild-type mice (Groszer et al., 2001),
demonstrating that the addition of exogenous ROS is not
capable of stimulating the PTEN-deficient cells (p < 0.001; Fig-
ure 4D). As would be predicted from this model, ROS stimulated
clonal neurosphere formation in heterozygous cells to a greater
extent than it did wild-type cells (Figure 4D; p < 0.01). Likewise,
inhibition of NOX resulted in dramatically reduced clonal neuro-
sphere formation in WT but not in PTEN-deficient cells (Fig-
ure S4B). Finally, BDNF stimulation of clonal neurosphere
formation was also observed only in wild-type but not in PTEN-
deficient cells (see Figure S4A).
We examined activation status of key downstream nodes
of the pathway. Exogenous ROS (H2O2 and Gox) enhanced,
whereas inhibition of endogenous NOX-generated ROS with
DPI, inhibited the phosphorylation of Akt (Figure 4E). Further-
more, we observed increased phospho Akt (pAkt) in the ROShi
compared to the ROSlo population of cells, increased pAkt in
BDNF-treated neurosphere cultures, and increased pAkt after
the addition of H2O2 into low-growth-factor conditions media
(Figure 4E). We observed similar results from flow cytometry
analysis of S6 phosphorylation (Figure 4F). In addition, the ROShi
population from human ES-derived neural cells also had
elevated pAkt and pS6 activation (Figure S1C).
Pharmacological experiments also support a role for the PI3K
pathway. The effects of exogenous ROS on neurosphere forma-
tion were abolished by the PI3K inhibitor LY294002 (LY),
suggesting that exogenous ROS do not exert their effects by
either bypassing the pathway or by providing enough stimulation
downstream of PI3K to overcome this inhibition. Because ROS
can also mediate effects via activation of the MAPK pathway,
we compared the relative effects of LY and the ERK inhibitor
U0126 in ROShi and unselected cells (Figure 4G). In both cases,
pathway inhibition had a greater effect on the ROShi compared to
unselected cells. However, LY had a much greater inhibitory
effect on the ROShi cells than the U0126, indicating a greater
dependence of these cells on the PI3K pathway than on the
MAPK pathway. Acute LY treatment inhibition significantly
decreased endogenous cellular ROS levels (Figure S4C), in
agreement with our hypothesized pathway for NOX signaling in
neural stem cells (Figure 1A).
Cellular ROS Levels Influence Neurogenic PotentialConditional deletion of PTEN results in both enhanced NSC
self-renewal and a sustained increase in neurogenesis (Groszer
et al., 2001, 2006; Gregorian et al., 2009). Therefore, we deter-
mined whether ROS stimulation of PI3K/Akt signaling had similar
effects on neurogenesis. Treatment of clonal density cultures
with low, nontoxic levels of exogenous ROS during sphere
formation produced significantly higher numbers of neurons as
a percentage of total cells when differentiated in the presence
of standard conditions (p < 0.001; Figures 5A and 5B). However,
treatment of cells with the same exogenous ROS during differen-
tiation resulted in increased cell death and few, if any, neurons
were produced (data not shown). Conversely, inhibition of NOX
or inhibition of PI3K (LY294002) prior to differentiation signifi-
cantly reduced neuron numbers (p < 0.01; Figures 5A and 5B).
In combination with exogenous ROS stimulation, inhibition of
the PI3K pathway (LY294002) eliminated the positive effects of
ROS on neurogenesis (p < 0.001; Figures 5A and 5B). In agree-
ment with our data demonstrating that NOX inhibition decreased
neurogenesis, we found that neurosphere cultures derived from
NOX2mutant mice produced significantly fewer neurons as well
(p < 0.01; Figures 5C and 5D).
NOX-Generated ROS Regulates SVZ Proliferationand Neurogenesis In VivoWe next tested whether our ex vivo findings extend to an in vivo
stem cell system. To this end, we tested the effects of the NOX
inhibitor apocynin (Apo) on SVZ proliferation. We first assessed
the effects of Apo treatment on endogenous ROS levels by using
the in vivo ROS-sensitive dye hydroethidine (HEt). Even in control
(vehicle-treated) animals, the SVZ had significantly higher ROS
levels than surrounding brain tissues such as the striatum and
cortex (p < 0.01; Figures 6A–6C). The SVZ also had approxi-
mately 8-fold enriched expression for the NOX2 homolog
compared to neighboring cortical tissue (p < 0.001; Figure 6B).
The 3 week Apo treatment resulted in a significant reduction in
SVZ ROS levels (p < 0.01; Figures 6A and 6D) and in the number
of Ki67 (proliferative) cells within the SVZ (p < 0.02; Figure 6E).
Cells acutely dissociated from the SVZ of mice similarly treated
Cell Stem Cell 8, 59–71, January 7, 2011 ª2011 Elsevier Inc. 63
A B
C D
E F
G
Figure 4. ROS Augment Trophic Factor Signaling and Are Dependent on the PI3K/Akt Signaling Pathway for Their Effects
(A) Clonal neurosphere formation after stimulation of adult SVZ cells by BDNF (B), BDNF plus the NOX inhibitor DPI (B+D), and BDNF plus the antioxidant
N-acetyl-cysteine (B+N) all expressed as a percentage of control (untreated) cells.
(B) Endogenous superoxide production in adult SVZ cultures treated with BDNF (B) and BDNF plus DPI (B+D).
(C) Oxidized and reduced PTEN are visualized on a redox-sensitive western blot.
(D) Clonal density neurosphere formation in response to stimulation by hydrogen peroxide (H2O2) and glucose oxidase (Gox) was determined in PTEN-deficient
(KO), PTEN heterozygous (HET), and wild-type (WT) cells.
(E) Phospho-Akt activation in exogenous ROS-stimulated cells (H2O2 and Gox), NOX-inhibited cells (DPI), ROShi and ROSlo cells, BDNF-stimulated, low growth
factor, and low growth factor supplemented with exogenous H2O2.
(F) Phospho-S6 activation detected by immunocytochemistry and flow cytometry in ROShi, ROSlo, NOX2 mutant, and wild-type and LY294002-treated cells.
(G) IC50 calculations for LY294002 (Pi3K inhibitor) and U0126 (ERK inhibitor).
Data expressed as mean ± SEM. See also Figures S4A–S4C.
Cell Stem Cell
Neural Stem Cell Redox Regulation
64 Cell Stem Cell 8, 59–71, January 7, 2011 ª2011 Elsevier Inc.
A B
0
50
100
150
200
250
H G A D LY HL GL
% C
on
tro
l
Neurons generated by pretreated clonal neurospheres
Control LY294002
H2O2 H2O2 + LY
Gox + LYGox
10micron
C D
20
0
4
8
12
16
WT MUT
% T
ota
l Cel
ls
WT
MU
T
10micron
Neurons generated NOX2 mutant and wild-type neurospheres
Figure 5. ROS Stimulation during Mitogenic
Expansion Enhances Neurogenesis in
a PI3K-Dependent Manner
(A) TuJ1-positive neurons produced in hydrogen
peroxide (H), glucose oxidase (G), Apocynin (A),
DPI (D), or LY294002 (LY; H+LY = HL; G+LY =
GL) -supplemented conditions during mitogenic
expansion.
(B) Picomicrographs of TuJ1 staining (green) and
Hoechst (blue) counterstain in differentiated neu-
rospheres under the conditions described above
taken at 103 magnification.
(C) Neuron numbers (as a percentage of total
Hoechst cells) produced by differentiated neuro-
spheres from NOX2 mutant (MUT) and wild-type
(WT) cultures.
(D) Picomicrographs of TUJ1 (red) and Hoechst
(blue) expression in differentiated cultures from
NOX2 mutants and wild-type cultures.
Data expressed as mean ± SEM.
Cell Stem Cell
Neural Stem Cell Redox Regulation
with Apo in vivo produced significantly fewer clonal neuro-
spheres in primary cultures compared to vehicle-treated mice
(p < 0.01; Figure 6F), indicating decreased neural stem or pro-
genitor cell numbers. However, this deficit recovered in subse-
quent serial clonal passages, demonstrating that although APO
administration acutely inhibited proliferation in vivo, the compe-
tency for self-renewal in the SVZ-derived cells was not affected.
Consistent with our observations on apocynin-treated ani-
mals, we found that NOX2 mutant mice also had diminished
numbers of Ki67 (proliferating) cells within the SVZ compared
to wild-type mice (p < 0.03; Figure 7A). NOX2 mutant and
wild-type mice were pulsed with BrdU followed by a 4 week
wash-out period during which time labeled SVZ cells would be
expected to leave the SVZ andmigrate through the rostral migra-
tory stream to the olfactory bulb where they normally differen-
tiate into postmitotic neurons. We found that a larger number
of BrdU-positive cells remained within the SVZ of mutant mice,
Cell Stem Cell 8, 59–7
and there were also fewer BrdU+ cells
in the olfactory bulb of mutant mice
and fewer new neurons (BrdU+/NeuN+)
produced there (p < 0.01; Figures 7A
and 7B). As a result of this defect in cell
proliferation and possibly also in migra-
tion, we observed that the granule cell
layer of the olfactory bulb in mutant
mice was smaller than those of wild-
type mice (p < 0.05; Figures 7C and 7D).
By using flow cytometry analysis of
acutely dissociated SVZ, we found that
the NOX2 mutants have more immature
progenitor cells (nestin+ and Sox2+) and
fewer cells expressing markers for neuro-
blasts (DCX) or transit-amplifying cells
(Mash1 and Dlx2; Figure 7G). Although
these data suggest an increase in some
progenitor cells, our in vitro findings indi-
cate a diminished capacity for the gener-
ation of clonal, serially passagable neuro-
spheres, suggesting a diminished number of neural stem cells in
NOX2 mutants. Therefore, the ex vivo cell phenotypes we have
observed indicate that there may also be defects in cell matura-
tion and differentiation.
In addition to the negative effects on NSCs caused by
decreased NOX activity, we have also conversely demon-
strated that increased NOX activity in vivo can have stimulatory
effects. Systemic administration of a low, nontoxic dose of the
neuroinflammatory stimulus lipopolysaccharide (LPS) resulted
in a significant enhancement in SVZ proliferation (p < 0.001;
Figures 7E and 7F) whereas inhibition of NOX activity by Apo
cotreatment eliminated the stimulatory effects of LPS on SVZ
proliferation (p < 0.03; Figures 7E and 7F). Although neuro-
inflammatory cells are likely to play a role in this effect in vivo,
low-dose LPS stimulates NSC self-renewal in vitro, which is
also blocked by NOX inhibition and antioxidant treatment
(Figure S5).
1, January 7, 2011 ª2011 Elsevier Inc. 65
Inte
nsi
ty (R
FU
)
150
200
250
300
A
B C D
LVIn
ten
sity
(R
FU
)
150
200
250
300Endogenous ROS levels in vivo
Endogenous ROS levels following NOX inhibition in vivo
Control APO
HEt
Hoescht
LVLV
6
8
10
12
Rat
io E
xpre
ssio
n
NOX2 Expression
Flu
ore
scen
t I
0
50
100
Veh APO
0
1000
2000
3000
4000
Veh APO
Ki6
7+ c
ell n
um
ber
E
Flu
ore
scen
t
0
50
100
SVZ STR CTX
SVZ proliferation after NOX inhibition in vivo
0
2
4
6
8
10
12
14
16
18
P1 P2 P3
% C
ells
See
ded
Serial Clonal Neurosphere Formation after In Vivo NOX inhibition
VehicleApocynin
0
2
4
CTX SVZ
Rat
io E
xpre
ssio
n
F
Figure 6. In Vivo Inhibition of NADPH
Oxidase by Apocynin Decreases SVZ Prolif-
eration, Endogenous ROS Levels, and NSC
Self-Renewal
(A) Picomicrographs of hydroethidine (HEt) fluores-
cence in the subventricular zone of Apocynin- and
vehicle-treated mice. The lateral ventricle (LV) is
indicated.
(B) Relative expression of the NOX2 (gp91phox)
homolog of NADPH oxidase in the adult SVZ com-
pared to neighboring cortex.
(C) Hydroethidine fluorescence (ROS levels) in the
SVZ and the surrounding cortical (CTX) or striatal
(STR) tissue.
(D) Hydroethidine fluorescence intensity (ROS)
levels within the SVZ after a 3-week daily apocynin
(Apo) or vehicle (control) treatment.
(E) Cell proliferation (Ki67) in the SVZ of apocynin-
and vehicle-treated animals.
(F) Serial clonal density neurosphere formation by
cells derived from the SVZ of mice that received
apocynin or vehicle treatment in vivo.
Data expressed as mean ± SEM.
Cell Stem Cell
Neural Stem Cell Redox Regulation
DISCUSSION
Reactive Oxygen Species Regulate Neural StemCell FunctionIn the current manuscript we have demonstrated that both
exogenous and endogenous ROS can have a significant impact
on neural stem and progenitor cell proliferation, self-renewal,
and neurogenesis. Our observations of the effects of ROS on
these cells are surprising for the fact that the neural stem cell
compartment appears to be disproportionately dependent on
ROS-mediated signaling in the brain. This is not inconsistent
with observations by others that embryonic and neural stem
cells have enhanced antioxidant capacity compared to more
differentiated progeny (Madhavan et al., 2006) because this
activity may be a protective mechanism in stem cell populations
with active oxidant-mediated signaling to prevent excessive or
toxic levels of ROS from being generated. Stem cell populations
have been observed to possess an enhanced resistance
66 Cell Stem Cell 8, 59–71, January 7, 2011 ª2011 Elsevier Inc.
R
re
A
th
in
a
g
(C
re
c
a
R
o
to
m
a
d
to oxidative stress-mediated cell death
(Madhavan et al., 2006, 2008; Romanko
et al., 2004). One such mechanism impor-
tant for cellular redox regulation could be
FOXO proteins. When FOXO genes are
deleted from neural stem and progenitor
cells, antioxidant defenses are signifi-
cantly depleted and endogenous ROS
levels undergo large increases (Renault
et al., 2009; Paik et al., 2009). As a result
of this elevated cellular ROS, there is an
initial hyperproliferation of NSCs leading
to brain overgrowth on par with what
has been observed with PTEN deletion
in the developing brain. However, toxic
levels of ROS build up over time, leading
to a premature senescence in the cells,
suggesting that control of endogenous
OS levels may play a significant role in the regulation of self-
newal and proliferation in neural stem and progenitor cells.
ccordingly, Yoneyama et al. (2010) have recently observed
at NOX inhibition and antioxidant treatments significantly
hibit hippocampal progenitor proliferation. On the other hand,
nother recent study has identified a novel ROS-regulating
ene, Prdm16, which results in brain undergrowth when deleted
huikov et al., 2010). Prdm16was identified by the authors as a
sult of BMI-1 inhibition, which has also been shown to regulate
ellular ROS levels in hematopoetic stem cells by specifically
ltering mitochondrial ROS and not NADPH oxidase-generated
OS (Liu et al., 2009). Thus, the contradictory inhibitory effects
f Prdm16-mediated ROS regulation on NSCs may be related
the endogenous source of the ROS and the cellular compart-
ent in which they act.
Previous studies have disagreed on whether stem cells gener-
lly have lower or higher endogenous ROS levels than their
ifferentiated progeny (Madhavan et al., 2006; Tsatmali et al.,
Cell counts within the SVZ and olfactory bulb of NOX2 mutant and wild-type mice
A
C
WTMUT
Ki-67
BrdU
1.5
2
2.5
ea (
mm
2 )
B
D E
0
40
80
120
160
200
Ki67 BrdU BrdU BrdU/NeuN
SVZ OB
% W
T
OB size
WT
OB
100
150
200
% W
T
Cell Phenotypes in MUT SVZ
WT
MUT
0
0.5
1
GC
L A
re
F
0
20
40
60
80
100
120
140
160
180
APO LPS LPS+APO
Co
ntr
ol (
veh
icle
)
SVZ proliferation after NOX inhibition and neuro-inflammatory stimuli
LV
LVLV
Veh
LPS LPS+APO
MU
T
0
50
Nestin GFAP DCX Sox2 Mash1Dlx2
%
G
Figure 7. In Vivo SVZ Proliferation and Neurogenesis Are Significantly Impacted by Changes in Cellular ROS(A) SVZ proliferation (Ki67+) and olfactory bulb (OB) neurogenesis (BrdU+/NeuN+) was stereologically quantitated in mutant and wild-type mice.
(B) Picomicrographs of Ki67 and BrdU labeling in the adult SVZ at 203 magnification.
(C) Area measurements of the granule cell layer (GCL) of the olfactory bulb in mutant (MUT) and wild-type (WT) mice.
(D) Pictomicrograph of olfactory bulb (NeuN, red; BrdU, green; Hoechst, blue).
(E) Cell phenotypes in NOX2 mutant SVZ compared to wild-type cells.
(F) SVZ proliferation (Ki67+ cells) was quantitated in wild-type mice treated with the NOX inhibitor apocynin (APO), the neuroinflammatory mediator lipopolysac-
charide (LPS), or both. Results are expressed as a percentage of control (vehicle) treated.
(G) Picomicrographs of Ki67 immunostaining the SVZ of the mice described in (F).
Data expressed as mean ± SEM. See also Figure S5A.
Cell Stem Cell
Neural Stem Cell Redox Regulation
2005; Limoli et al., 2004; Jang and Sharkis, 2007; Diehn et al.,
2009). Definitive NSCs might be expected to have a lower
endogenous ROS status than that of the highly proliferative,
transit-amplifying progenitors because the adult neural stem
cell in vivo is thought to be a relatively quiescent cell under
normal circumstances (Doetsch et al., 1997). Thus, the higher
Cell Stem Cell 8, 59–71, January 7, 2011 ª2011 Elsevier Inc. 67
Cell Stem Cell
Neural Stem Cell Redox Regulation
ROS status of the SVZ that we observed may play a role in main-
taining the proliferation of progenitor cells within this neurogenic
niche. However, in order for the stem cell population to maintain
a more quiescent state in this environment, it would necessitate
that they are able to maintain a lower endogenous ROS level
when not dividing, suggesting a robust antioxidant regulation
in a subset of specialized cells in vivo. Our ex vivo and in vitro
data are consistent with high endogenous ROS levels in neural
stem cells but could be reflective of an ‘‘activated’’ state in the
cells as a result of removal from their normal in vivo environment.
In vivo we observed significantly reduced SVZ proliferation and
neurogenesis when endogenous ROS levels are reduced in the
NOX2 mutants and APO-treated mice. This suggests that in
order to maintain normal levels of neurogenesis, the neural
stem cells must need to be able to increase ROS levels when
required for cell division but does not rule out the possibility
that NSCs maintain a low ROS state in vivo when they are in
a quiescent state.
The Effects of ROS on NSC Function Are Dependenton PI3k/Akt SignalingThe most often cited mechanism by which ROS contribute to
cellular signaling is by modifying the actions of proteins through
the reversible oxidation of essential cysteine residues (Ross
et al., 2007; Leslie et al., 2003; Kwon et al., 2004), although other
mechanisms have been proposed such as cell cycle targets
(cyclin D1 and forkhead proteins) (Abid et al., 2004; Burch and
Heintz, 2005; Blanchetot and Boonstra, 2008). Our data are
consistent with a model of posttranslational oxidative inactiva-
tion of the tumor suppressor PTEN, a negative regulator of
PI3K signaling. Although the involvement of other pathways
such asMAPK signaling has not been ruled out, our data suggest
a critical role for the PI3K/Akt pathway and are similar to the
phenotype observed after genetic deletion of PTEN (Groszer
et al., 2001, 2006; Gregorian et al., 2009).
Perhaps more surprising than the stimulatory effects of exog-
enous ROS, we have found that the inhibition of normal endog-
enous ROS production by NOX inhibition or mutation negatively
regulated the PI3K/Akt pathway and NSC function. Thus, the
high ROS status of NSCs appears to be required to maintain
their self-renewal and neurogenesis by maintaining adequate
levels of PI3K signaling.
The Effects of ROS-Mediated PI3K Pathway SignalingAre Context DependentDespite the broad influence of ROS-mediated signaling indi-
cated by the stimulatory effects of exogenous ROS and the
negative effects of NOX inhibition in neural stem cell-enriched
populations, there are many cases in which cellular response
to ROS is highly dependent on other factors such as cell pheno-
type, cell differentiation state, or other signaling cofactors.
For example, conditional deletion of PTEN in nestin-expressing
neural stem and progenitors in the developing brain and in
GFAP-expressing stem cells in the SVZ of the adult brain leads
to an enhanced and sustained neural stem cell self-renewal and
neurogenesis, contributing to brain overgrowth (Groszer et al.,
2001, 2006; Gregorian et al., 2009). However, studies in the
hematopoetic system indicate that although PTEN deletion
68 Cell Stem Cell 8, 59–71, January 7, 2011 ª2011 Elsevier Inc.
results in a similar enhancement in self-renewal in hematopoi-
etic stem cells (HSCs), it also results in a premature senescence
in these cells (Zhang et al., 2006; Yilmaz et al., 2006; Chen
et al., 2008). The effects of cellular ROS levels may also be simi-
larly cell type dependent. For example, HSCs have been shown
to have lower endogenous ROS levels than their more dif-
ferentiated hematopoietic cell counterparts (Jang and Sharkis,
2007).
Previous work has established that O2-A progenitor cells are
modulated by changes in cellular redox status, namely that
they maintain a low ROS status that promotes cell division and
maintains an undifferentiated state (Li et al., 2007; Power et al.,
2002; Smith et al., 2000). Li et al. (2007) determined that one
mechanism by which higher ROS inhibit O-2A progenitors is
through c-Cbl-mediated receptor tyrosine kinase (RTK) ubiquiti-
nation and breakdown. On the other hand it has recently been
shown in other cell types that PTEN deletion prevents c-Cbl-
mediated RTK breakdown (Vivanco et al., 2010). Therefore,
NOX-mediated oxidative inactivation of PTEN should have
similar RTK-stabilizing effects. Additionally, EGF signaling acti-
vates NOX and is required for aSVZ neurosphere cultures,
whereas EGF is not utilized by O2-A progenitors (Kondo and
Raff, 2000). Thus phenotypic differences in cell NOX activity
and EGF signaling could be important factors in the functional
differences we have observed between NSCs and O-2A
progenitors.
The effects of ROS are also dependent on the differentiation
state of the cells. For example, the neurotrophic factor BDNF
promotes differentiation of postmitotic neurons, but we found
that in undifferentiated cells in the presence of growth factors,
it will promote NSC self-renewal in a NOX- and ROS-dependent
manner. Similarly, we found that the effects of exogenous ROS
stimulation are dependent on the differentiation state of cells.
ROS stimulation of undifferentiated cells in the presence of
growth factors promotes both NSC self-renewal and neurogenic
potential but, on the other hand, the same levels of ROS that
were stimulatory to proliferative cells were found to be toxic to
the same cells when present during differentiation after growth
factor withdrawal. Consistently, the effect of PTEN deletion is
also dependent on the differentiation state of the cells. For
example, whereas PTEN deletion in undifferentiated, mitotic
cells produced enhanced NSC proliferation and neurogenesis
(Groszer et al., 2001), PTEN deletion in postmitotic neurons
does not influence cell phenotypes or cause cells to re-enter
the cell cycle and divide (Kwon et al., 2006). Rather, enhanced
PI3K pathway signaling in differentiated neurons results in
cellular hypertrophy, which can also contribute to a macroce-
phalic phenotype in vivo (Zhou et al., 2009).
In conclusion, we have identified a redox-mediated regulatory
mechanism of self-renewal and differentiation potential that
is required for normal neural stem cell function and to support
normal ontogeny. However, a large number of environmental
factors and genetic mutations can potentially influence and
deregulate ROS-mediated signaling, which may contribute to
abnormal brain development or transformation and tumorigen-
esis. Thus, understanding how normal and transformed cells
utilize ROS may play an important role in identifying new tar-
gets for anticancer treatments or points of vulnerability in brain
development.
Cell Stem Cell
Neural Stem Cell Redox Regulation
EXPERIMENTAL PROCEDURES
Animals
Unless otherwise specified all experiments were carried out on adult CD1mice
from Charles River, USA. PTEN mutants were generated as described in
Groszer et al. (2001) and Gregorian et al. (2009). NADPH oxidase (gp91phox)
mutant mice and wild-type controls were obtained from Jackson Labs (USA)
and backcrossed onto a CD-1 background. In vivo administration of apocynin
(5 mg/kg/day; Sigma) was performed by daily intraperitoneal injections for
3 weeks. Lipopolysaccharide (LPS E. coli serotype 0111:B4; 0.1 mg/kg,
Sigma) was administered in vivo via i.p. injection 48 hr prior to perfusion-fixa-
tion. In vivo administration of BrdU (50 mg/kg/injection) was performed every
2 hr for 8 hr. BrdU-injected mice were perfused 4 weeks later. All procedures
were approved by the UCLA Chancellor’s Committee for Animal Research.
Human and Mouse Cell Culture
Standard high-density neurosphere cultures and clonal density neurosphere
assays were established for mouse and human cells according to themethods
of Groszer et al. (2006) and Gregorian et al. (2009). See Supplemental Exper-
imental Procedures for more detailed information. Exogenous ROS used were
hydrogen peroxide (2–4 mM H2O2) and glucose oxidase (2 mU; GOx). NOX
inhibitors used were apocynin (100 mM; APO) and diphenylene iodonium
(1 nM; DPI). The antioxidant N-acetyl-cysteine (1 mM; NAC) was also used.
Flow Cytometry
The isolation of Lex- and EGFR-positive and -negative cells for clonal analysis
was performed with fluorescent activated cell sorting (FACS) to place one cell
per well in 96-well plates. FACS was performed with a FACSDiVa cell sorter
(BD Biosciences) with a purification-mode algorithm. Sort gates were set by
side and forward scatter to eliminate dead and aggregated cells and by Alexa
secondary fluorophores to define positive cells. Purity of the sorted cells was
confirmed by flow cytometric reanalysis of positive and negative cell samples.
Stem Cell Sorting
A combination of live cell sorting for extracellular EGFR and CD24 was per-
formed with a FACSDiVa cell sorter followed by DCFDA dye-labeling, fixation,
permeablization, intracellular staining for GFAP, ID1, and flow analysis.
Western Blotting
All primary antibodies (total Akt and phospho-specific Akt) and positive and
negative controls were purchased from Cell Signaling Technologies. Neuro-
spheres from each condition were lysed in buffer containing 0.1% Triton
X-100 in 50 mM Tris-HCl and 150 mM NaCl and Protease Inhibitor Cocktail
(Sigma). Samples were prepared according to standard western blot protocol.
See Supplemental Experimental Procedures for details. Oxidized PTEN was
visualized according to the methods of Delgado-Esteban et al. (2007).
Quantitative Real-Time PCR
RNA was isolated with Trizol reagent (Invitrogen) according to the manufac-
turer’s protocol. 1 mg of total RNA was treated with 1 unit of Amplification
Grade DNase I (Sigma-Aldrich) at room temperature for 15 min followed by
inactivation at 70�C for 10 min as described by the manufacturer. See Supple-
mental Experimental Procedures.
Measuring Endogenous ROS Levels
In cell culture the ROS-sensitive dye DCFDA (5 mM; Molecular Probes), Hydro-
ethidine (2 mM; Sigma), and HPF-APF (5 mM; Invitrogen) was used to measure
endogenous cellular ROS levels in control and treated cultures as well as in
cells from mutant and wild-type animals. In vivo ROS levels were determined
with the ROS-sensitive dye hydroethidine (10 mg/kg; Invitrogen, Kunz et al.,
2007). See Supplemental Experimental Procedures.
Immunohistochemistry
Perfused-fixed mouse brains were stabilized by incubation in 10% sucrose for
48 hr. Brains were cryo-sectioned at 20 mM. Brain sections were immuno-
stained for Ki67 and BrdU according to the methods of Tang et al. (2007).
Double-labeling with the neuronal marker NeuN (Abcam 1:200) were carried
out on sections. Ten serial sections, spaced 120 mm apart, through the SVZ
and olfactory bulb (OB), were quantified with the unbiased optical fractionator
approach (Tsai et al., 2006) (StereoInvestigator; MicroBrightField, Colchester,
VT). Hoescht counterstain was used to measure olfactory bulb granule cell
layer area with image analysis software (MCID, Imaging Research, St. Cather-
ines, ON, Canada).
Statistical Analysis
All data are expressed as mean ± SEM, unless otherwise indicated. t tests
were performed with Microsoft Excel to determine statistical significance of
treatment sets. For multiple comparisons, one- or two-way ANOVA was per-
formed, as appropriate, and Bonferroni post-hoc t tests were done to deter-
mine significance. Alpha values were 0.05 except when adjusted by the
post-hoc tests.
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental Procedures
and five figures and can be found with this article online at doi:10.1016/
j.stem.2010.11.028.
ACKNOWLEDGMENTS
This work is supported by the following grants and awards: Cure Autism Now
Fellowship (to J.E.L.), Autism Speaks Basic and Clinical grant (to H.I.K.),
Autism Speaks Environmental Sciences grant (to J.E.L.), Center for Autism
Research and Treatment (CART) Pilot Grant Award #06LEB2008, which is sup-
ported by NIH/NICHD grant # P50-HD-055784 (to J.E.L.), NIH MH65756 (to
H.I.K. and H.W.), Henry Singleton Brain Cancer Research Program and James
S. McDonnell Foundation Award (to H.W.), Miriam and Sheldon Adelson
Program in Neural Repair and Rehabilitation (to H.W. and H.I.K.), University
of California, Cancer Research Coordinating Committee grant (to A.D.P.),
and the Jonsson Comprehensive Cancer Center grant (to A.D.P.). Flow cytom-
etry and cell sorting was performed in the UCLA Jonsson Comprehensive
Cancer Center (JCCC) and Center for AIDS Research Flow Cytometry Core
Facility, which is supported by National Institutes of Health awards CA-
16042 and AI-28697, and by the JCCC, the UCLA AIDS Institute, the David
Geffen School of Medicine at UCLA, and the UCLA Chancellor’s Office.
Received: November 5, 2009
Revised: August 22, 2010
Accepted: October 26, 2010
Published: January 6, 2011
REFERENCES
Abid, M.R., Schoots, I.G., Spokes, K.C., Wu, S.Q., Mawhinney, C., and Aird,
W.C. (2004). Vascular endothelial growth factor-mediated induction of manga-
nese superoxide dismutase occurs through redox-dependent regulation of
forkhead and IkappaB/NF-kappaB. J. Biol. Chem. 279, 44030–44038.
Behrens, M.M., Ali, S.S., and Dugan, L.L. (2008). Interleukin-6 mediates the
increase in NADPH-oxidase in the ketamine model of schizophrenia.
J. Neurosci. 28, 13957–13966.
Blanchetot, C., and Boonstra, J. (2008). The ROS-NOX connection in cancer
and angiogenesis. Crit. Rev. Eukaryot. Gene Expr. 18, 35–45.
Burch, P.M., and Heintz, N.H. (2005). Redox regulation of cell-cycle re-entry:
Cyclin D1 as a primary target for the mitogenic effects of reactive oxygen
and nitrogen species. Antioxid. Redox Signal. 7, 741–751.
Capela, A., and Temple, S. (2002). LeX/ssea-1 is expressed by adult mouse
CNS stem cells, identifying them as nonependymal. Neuron 35, 865–875.
Chen, C., Liu, Y., Liu, R., Ikenoue, T., Guan, K.L., Liu, Y., and Zheng, P. (2008).
TSC-mTORmaintains quiescence and function of hematopoietic stem cells by
repressing mitochondrial biogenesis and reactive oxygen species. J. Exp.
Med. 205, 2397–2408.
Chiarugi, P., and Fiaschi, T. (2007). Redox signalling in anchorage-dependent
cell growth. Cell. Signal. 19, 672–682.
Cell Stem Cell 8, 59–71, January 7, 2011 ª2011 Elsevier Inc. 69
Cell Stem Cell
Neural Stem Cell Redox Regulation
Chuikov, S., Levi, B.P., Smith, M.L., and Morrison, S.J. (2010). Prdm16
promotes stem cell maintenance inmultiple tissues, partly by regulating oxida-
tive stress. Nat. Cell Biol. 12, 999–1006.
Delgado-Esteban, M., Martin-Zanca, D., Andres-Martin, L., Almeida, A., and
Bolanos, J.P. (2007). Inhibition of PTEN by peroxynitrite activates the phos-
phoinositide-3-kinase/Akt neuroprotective signaling pathway. J. Neurochem.
102, 194–205.
Diehn, M., Cho, R.W., Lobo, N.A., Kalisky, T., Dorie, M.J., Kulp, A.N., Qian, D.,
Lam, J.S., Ailles, L.E., Wong, M., et al. (2009). Association of reactive oxygen
species levels and radioresistance in cancer stem cells. Nature 458, 780–783.
Doetsch, F., Garcıa-Verdugo, J.M., and Alvarez-Buylla, A. (1997). Cellular
composition and three-dimensional organization of the subventricular
germinal zone in the adult mammalian brain. J. Neurosci. 17, 5046–5061.
Garrido, A.M., and Griendling, K.K. (2009). NADPH oxidases and angiotensin II
receptor signaling. Mol. Cell. Endocrinol. 302, 148–158.
Goldstein, B.J., Mahadev, K., Wu, X., Zhu, L., and Motoshima, H. (2005). Role
of insulin-induced reactive oxygen species in the insulin signaling pathway.
Antioxid. Redox Signal. 7, 1021–1031.
Gregorian, C., Nakashima, J., Le Belle, J., Ohab, J., Kim, R., Liu, A., Smith,
K.B., Groszer, M., Garcia, A.D., Sofroniew, M.V., et al. (2009). Pten deletion
in adult neural stem/progenitor cells enhances constitutive neurogenesis.
J. Neurosci. 29, 1874–1886.
Groszer, M., Erickson, R., Scripture-Adams, D.D., Lesche, R., Trumpp, A.,
Zack, J.A., Kornblum, H.I., Liu, X., and Wu, H. (2001). Negative regulation of
neural stem/progenitor cell proliferation by the Pten tumor suppressor gene
in vivo. Science 294, 2186–2189.
Groszer, M., Erickson, R., Scripture-Adams, D.D., Dougherty, J.D., Le Belle, J.,
Zack, J.A., Geschwind, D.H., Liu, X., Kornblum, H.I., and Wu, H. (2006). PTEN
negatively regulates neural stem cell self-renewal by modulating G0-G1 cell
cycle entry. Proc. Natl. Acad. Sci. USA 103, 111–116.
Guo, S., Bragina, O., Xu, Y., Cao, Z., Chen, H., Zhou, B., Morgan, M., Lin, Y.,
Jiang, B.H., Liu, K.J., and Shi, H. (2008). Glucose up-regulates HIF-1 alpha
expression in primary cortical neurons in response to hypoxia through main-
taining cellular redox status. J. Neurochem. 105, 1849–1860.
Infanger, D.W., Sharma, R.V., and Davisson, R.L. (2006). NADPH oxidases of
the brain: Distribution, regulation, and function. Antioxid. Redox Signal. 8,
1583–1596.
Islam, O., Loo, T.X., and Heese, K. (2009). Brain-derived neurotrophic factor
(BDNF) has proliferative effects on neural stem cells through the truncated
TRK-B receptor, MAP kinase, AKT, and STAT-3 signaling pathways. Curr.
Neurovasc. Res. 6, 42–53.
Jang, Y.Y., and Sharkis, S.J. (2007). A low level of reactive oxygen species
selects for primitive hematopoietic stem cells that may reside in the low-
oxygenic niche. Blood 110, 3056–3063.
Kondo, T., and Raff, M. (2000). Oligodendrocyte precursor cells reprog-
rammed to become multipotential CNS stem cells. Science 289, 1754–1757.
Kunz, A., Anrather, J., Zhou, P., Orio, M., and Iadecola, C. (2007).
Cyclooxygenase-2 does not contribute to postischemic production of reactive
oxygen species. J. Cereb. Blood Flow Metab. 27, 545–551.
Kwon, J., Lee, S.R., Yang, K.S., Ahn, Y., Kim, Y.J., Stadtman, E.R., and Rhee,
S.G. (2004). Reversible oxidation and inactivation of the tumor suppressor
PTEN in cells stimulated with peptide growth factors. Proc. Natl. Acad. Sci.
USA 101, 16419–16424.
Kwon, C.H., Luikart, B.W., Powell, C.M., Zhou, J., Matheny, S.A., Zhang, W.,
Li, Y., Baker, S.J., and Parada, L.F. (2006). Pten regulates neuronal arboriza-
tion and social interaction in mice. Neuron 50, 377–388.
Lambeth, J.D., Kawahara, T., and Diebold, B. (2007). Regulation of Nox and
Duox enzymatic activity and expression. Free Radic. Biol. Med. 43, 319–331.
Lambeth, J.D., Krause, K.H., and Clark, R.A. (2008). NOX enzymes as novel
targets for drug development. Semin. Immunopathol. 30, 339–363.
Leslie, N.R. (2006). The redox regulation of PI 3-kinase-dependent signaling.
Antioxid. Redox Signal. 8, 1765–1774.
70 Cell Stem Cell 8, 59–71, January 7, 2011 ª2011 Elsevier Inc.
Leslie, N.R., Bennett, D., Lindsay, Y.E., Stewart, H., Gray, A., and Downes,
C.P. (2003). Redox regulation of PI 3-kinase signalling via inactivation of
PTEN. EMBO J. 22, 5501–5510.
Li, Z., Dong, T., Proschel, C., and Noble, M. (2007). Chemically diverse toxi-
cants converge on Fyn and c-Cbl to disrupt precursor cell function. PLoS
Biol. 5, 35.
Limoli, C.L., Rola, R., Giedzinski, E., Mantha, S., Huang, T.T., and Fike, J.R.
(2004). Cell-density-dependent regulation of neural precursor cell function.
Proc. Natl. Acad. Sci. USA 101, 16052–16057.
Liu, J., Cao, L., Chen, J., Song, S., Lee, I.H., Quijano, C., Liu, H., Keyvanfar, K.,
Chen, H., Cao, L.Y., et al. (2009). Bmi1 regulates mitochondrial function and
the DNA damage response pathway. Nature 459, 387–392.
Madhavan, L., Ourednik, V., and Ourednik, J. (2006). Increased ‘‘vigilance’’ of
antioxidant mechanisms in neural stem cells potentiates their capability to
resist oxidative stress. Stem Cells 24, 2110–2119.
Madhavan, L., Ourednik, V., and Ourednik, J. (2008). Neural stem/progenitor
cells initiate the formation of cellular networks that provide neuroprotection
by growth factor-modulated antioxidant expression. Stem Cells 26, 254–265.
Nam, H.S., andBenezra, R. (2009). High levels of Id1 expression defineB1 type
adult neural stem cells. Cell Stem Cell 5, 515–526.
Paik, J.H., Ding, Z., Narurkar, R., Ramkissoon, S., Muller, F., Kamoun, W.S.,
Chae, S.S., Zheng, H., Ying, H., Mahoney, J., et al. (2009). FoxOs cooperatively
regulate diverse pathways governing neural stem cell homeostasis. Cell Stem
Cell 5, 540–553.
Park, L., Zhou, P., Pitstick, R., Capone, C., Anrather, J., Norris, E.H., Younkin,
L., Younkin, S., Carlson, G., McEwen, B.S., and Iadecola, C. (2008). Nox2-
derived radicals contribute to neurovascular and behavioral dysfunction in
mice overexpressing the amyloid precursor protein. Proc. Natl. Acad. Sci.
USA 105, 1347–1352.
Pastrana, E., Cheng, L.C., and Doetsch, F. (2009). Simultaneous prospective
purification of adult subventricular zone neural stem cells and their progeny.
Proc. Natl. Acad. Sci. USA 106, 6387–6392.
Power, J., Mayer-Proschel, M., Smith, J., and Noble, M. (2002).
Oligodendrocyte precursor cells from different brain regions express divergent
properties consistent with the differing time courses of myelination in these
regions. Dev. Biol. 245, 362–375.
Renault, V.M., Rafalski, V.A., Morgan, A.A., Salih, D.A., Brett, J.O., Webb, A.E.,
Villeda, S.A., Thekkat, P.U., Guillerey, C., Denko, N.C., et al. (2009). FoxO3
regulates neural stem cell homeostasis. Cell Stem Cell 5, 527–539.
Romanko, M.J., Rothstein, R.P., and Levison, S.W. (2004). Neural stem cells in
the subventricular zone are resilient to hypoxia/ischemia whereas progenitors
are vulnerable. J. Cereb. Blood Flow Metab. 24, 814–825.
Ross, S.H., Lindsay, Y., Safrany, S.T., Lorenzo, O., Villa, F., Toth, R., Clague,
M.J., Downes, C.P., and Leslie, N.R. (2007). Differential redox regulation within
the PTP superfamily. Cell. Signal. 19, 1521–1530.
Smith, J., Ladi, E., Mayer-Proschel, M., and Noble, M. (2000). Redox state is
a central modulator of the balance between self-renewal and differentiation
in a dividing glial precursor cell. Proc. Natl. Acad. Sci. USA 97, 10032–10037.
Studer, L., Csete, M., Lee, S.H., Kabbani, N., Walikonis, J., Wold, B., and
McKay, R. (2000). Enhanced proliferation, survival, and dopaminergic differen-
tiation of CNS precursors in lowered oxygen. J. Neurosci. 20, 7377–7383.
Sundaresan, M., Yu, Z.X., Ferrans, V.J., Irani, K., and Finkel, T. (1995).
Requirement for generation of H2O2 for platelet-derived growth factor signal
transduction. Science 270, 296–299.
Tang, X., Falls, D.L., Li, X., Lane, T., and Luskin, M.B. (2007). Antigen-retrieval
procedure for bromodeoxyuridine immunolabeling with concurrent labeling of
nuclear DNA and antigens damaged by HCl pretreatment. J. Neurosci. 27,
5837–5844.
Tsai, P.T., Ohab, J.J., Kertesz, N., Groszer, M., Matter, C., Gao, J., Liu, X., Wu,
H., and Carmichael, S.T. (2006). A critical role of erythropoietin receptor in
neurogenesis and post-stroke recovery. J. Neurosci. 26, 1269–1274.
Tsatmali, M., Walcott, E.C., and Crossin, K.L. (2005). Newborn neurons
acquire high levels of reactive oxygen species and increased mitochondrial
proteins upon differentiation from progenitors. Brain Res. 1040, 137–150.
Cell Stem Cell
Neural Stem Cell Redox Regulation
Vivanco, I., Rohle, D., Versele, M., Iwanami, A., Kuga, D., Oldrini, B., Tanaka,
K., Dang, J., Kubek, S., Palaskas, N., et al. (2010). The phosphatase and tensin
homolog regulates epidermal growth factor receptor (EGFR) inhibitor
response by targeting EGFR for degradation. Proc. Natl. Acad. Sci. USA
107, 6459–6464.
Wang, Y., and Lou,M.F. (2009). The regulation of NADPHoxidase and its asso-
ciation with cell proliferation in human lens epithelial cells. Invest. Ophthalmol.
Vis. Sci. 50, 2291–2300.
Yilmaz, O.H., Valdez, R., Theisen, B.K., Guo, W., Ferguson, D.O., Wu, H., and
Morrison, S.J. (2006). Pten dependence distinguishes haematopoietic stem
cells from leukaemia-initiating cells. Nature 441, 475–482.
Yoneyama, M., Kawada, K., Gotoh, Y., Shiba, T., and Ogita, K. (2010).
Endogenous reactive oxygen species are essential for proliferation of neural
stem/progenitor cells. Neurochem. Int. 56, 740–746.
Zhang, J., Grindley, J.C., Yin, T., Jayasinghe, S., He, X.C., Ross, J.T., Haug,
J.S., Rupp, D., Porter-Westpfahl, K.S., Wiedemann, L.M., et al. (2006). PTEN
maintains haematopoietic stem cells and acts in lineage choice and leukaemia
prevention. Nature 441, 518–522.
Zhou, J., Blundell, J., Ogawa, S., Kwon, C.H., Zhang, W., Sinton, C., Powell,
C.M., and Parada, L.F. (2009). Pharmacological inhibition of mTORC1
suppresses anatomical, cellular, and behavioral abnormalities in neural-
specific Pten knock-out mice. J. Neurosci. 29, 1773–1783.
Cell Stem Cell 8, 59–71, January 7, 2011 ª2011 Elsevier Inc. 71
Cell Stem Cell
Article
Robo4 Cooperates with Cxcr4 to SpecifyHematopoietic Stem Cell Localizationto Bone Marrow NichesStephanie Smith-Berdan,1 Andrew Nguyen,1 Deena Hassanein,1 Matthew Zimmer,1 Fernando Ugarte,1 Jesus Ciriza,2
Dean Li,3 Marcos E. Garcıa-Ojeda,2 Lindsay Hinck,1 and E. Camilla Forsberg1,*1Institute for the Biology of Stem Cells, University of California Santa Cruz, Santa Cruz, CA 95064, USA2School of Natural Sciences, University of California Merced, Merced, CA 95343, USA3Department of Oncological Sciences, University of Utah, Salt Lake City, UT 84112, USA
*Correspondence: cforsber@ucsc.edu
DOI 10.1016/j.stem.2010.11.030
SUMMARY
Specific bone marrow (BM) niches are critical forhematopoietic stem cell (HSC) function during bothnormal hematopoiesis and in stem cell transplanta-tion therapy. We demonstrate that the guidancemolecule Robo4 functions to specifically anchorHSCs to BM niches. Robo4-deficient HSCs dis-played poor localization to BM niches and drasticallyreduced long-term reconstitution capability while re-taining multilineage potential. Cxcr4, a critical regu-lator of HSC location, is upregulated in Robo4�/�
HSCs to compensate for Robo4 loss. Robo4 deletionled to altered HSC mobilization efficiency, revealingthat inhibition of both Cxcr4- and Robo4-mediatedniche interactions are necessary for efficient HSCmobilization. Surprisingly, we found that WT HSCsexpress very low levels of Cxcr4 and respond poorlyto Cxcr4 manipulation relative to other hematopoi-etic cells. We conclude that Robo4 cooperates withCxcr4 to endow HSCs with competitive access tolimited stem cell niches, and we propose Robo4 asa therapeutic target in HSC transplantation therapy.
INTRODUCTION
The tremendous potential of stem cells to provide a complete
and permanent cure for a wide range of human disorders makes
progress in improving the safety and efficiency of cell-based
therapies a top priority in modern medicine. Successful hemato-
poietic cell transplantations have been performed for more than
50 years and have made HSCs the paradigm for stem cell
therapy. Still, the morbidity andmortality of hematopoietic trans-
plant recipients are unacceptably high and transplants are there-
fore reserved for patients with few other treatment options.
By investigating the molecular mechanisms of HSC interaction
with the bone marrow (BM) microenvironment, our goal is to
enable specific and efficient manipulation of both HSC mobiliza-
tion and engraftment.
Because mobilized peripheral blood (PB) is an increasingly
common source of HSCs, transplantation therapy involves
72 Cell Stem Cell 8, 72–83, January 7, 2011 ª2011 Elsevier Inc.
HSC movement both into and out of the BM. In mice, as well
as in humans, combined administration of cytoxan (cyclophos-
phamide) and G-CSF (Cy/G treatment) induces self-renewing
divisions of BM HSCs, resulting in an expansion of the HSC
pool followed by migration of HSCs to the blood stream (Morri-
son et al., 1997; Passegue et al., 2005; Wright et al., 2001).
More recently, AMD3100, an antagonist of the G protein-
coupled receptor Cxcr4, has been used to mobilize hematopoi-
etic cells (Broxmeyer et al., 2005; Liles et al., 2003; Watt and
Forde, 2008). In contrast to Cy/G, AMD3100-induced mobiliza-
tion is rapid, with increased numbers of progenitors detected
in the blood 1 hr after administration of a single dose of drug,
and thus does not involve cell expansion. Upon transplantation,
intravenously injected HSCs must find their way back to the BM
and engraft. Most likely, HSCs home in response to chemokines,
including the Cxcr4 ligand Sdf1 (also known as Cxcl12), followed
by adhesion to the niche by engaging in specific interactions with
cellular and matrix components. Engraftment of transplanted
HSCs requires partial or complete myeloablation to allow donor
HSCs access to HSC-supportive niches. The ability to long-term
engraft is a defining and unique property of HSCs and critically
important for both normal hematopoietic development and
transplantation therapy.
Sdf1 andCxcr4 play pivotal roles in HSC location and function.
Mice deficient in either Sdf1 or Cxcr4 die during late embryogen-
esis and lack BM hematopoiesis (Nagasawa et al., 1996; Zou
et al., 1998). As described above, the Cxcr4 antagonist
AMD3100 can be used to mobilize hematopoietic progenitors
from the BM to PB in mice and humans (Broxmeyer et al.,
2005; Watt and Forde, 2008), and Cxcr4-blocking antibodies
impair HSC engraftment (Peled et al., 1999). In addition, HSCs
actively migrate toward Sdf1 in transwell migration assays
(Lapidot, 2001; Wright et al., 2002), and recent data suggest
that HSCs specifically localize next to BM cells expressing
high levels of Sdf1 (Sugiyama et al., 2006). Thus, there is exten-
sive evidence supporting critical roles for Sdf1 and Cxcr4 in
regulating HSC location.
Surprisingly, however, deletion of Cxcr4 in adulthood results in
HSCs capable of homing and engraftment (Nie et al., 2008;
Sugiyama et al., 2006). In addition, many cells other than HSCs
express Cxcr4, making it unlikely that Cxcr4, alone, specifies
HSC location to stem-cell-supportive niches. In search of
HSC-specific receptors capable of specifying cell location, we
recently identified the single-transmembrane receptor Robo4
Cell Stem Cell
Robo4 Regulates HSC Location to Bone Marrow Niches
on HSCs by gene expression microarray analysis (Forsberg
et al., 2005). A subsequent report confirmed that Robo4 marks
long-term reconstituting HSCs (Shibata et al., 2009). Robo4,
like its family members Robo1-3, is capable of regulating cell
location by responding to the Slit family of secreted ligands
(Kaur et al., 2006; Park et al., 2003; Seth et al., 2005; Suchting
et al., 2005). Other than HSCs, Robo4 expression seems
restricted to endothelial cells, where it functions to regulate
blood vessel sprouting (Huminiecki et al., 2002; Park et al.,
2003). Robo4�/� mice, though grossly normal, have defects in
VEGF- and Slit-induced regulation of vascular integrity and
angiogenesis (Jones et al., 2008; London et al., 2010; Marlow
et al., 2010). Here, we show that Robo4 acts as an HSC-specific
adhesion molecule that cooperates with Cxcr4 to localize HSCs
to BM niches.
RESULTS
Robo4 Expression Is Restricted to HSCs TightlyAssociated with BM NichesOur previous gene expression microarray analysis showed that
Robo4 is expressed at higher levels by HSCs compared to
MPP, Cy/G-mobilized HSCs (M-HSCs), and leukemic HSCs
(L-HSCs) (Forsberg et al., 2005, 2010). We verified these results
by qRT-PCR and extended the analysis to include multiple BM
cell types representing the major hematopoietic progenitor pop-
ulations and lineages. We found that Robo4 is very selectively
expressed by HSCs and downregulated upon differentiation
and mobilization and in leukemogenesis (Figures 1A and 1B).
Substantial numbers of M-HSCs and L-HSCs are found in the
blood, spleen, and liver (Morrison et al., 1997; Passegue et al.,
2004), so Robo4 downregulation may facilitate exit from HSC
niches in the BM. Intriguingly, Robo4 transcripts were barely
detectable in fetal liver HSCs and increased significantly in BM
HSCs during fetal to adult development (Figure 1C), further
emphasizing the specificity of Robo4 expression to HSCs
located in the BM. Cell surface staining via a monoclonal
antibody specific for Robo4 (Figure 1E) showed that Robo4
protein is robustly expressed by all adult BM HSCs, with lower
levels on ST-HSCs and MPP, and absent from other hematopoi-
etic cell types (Figure 1D; for flow cytometry gating strategies
see Figure S1A available online), in agreement with the qRT-
PCR data (Figure 1A). Less than 1% of total nucleated BM cells
are Robo4 positive, so Robo4 is an excellent HSC-specific
marker.
Because different Robo receptors may be functionally redun-
dant, we also analyzed the expression of Robo1, 2, and 3.
Previous studies have reported that circulating hematopoietic
cells express Robo1 and respond to the Robo ligand Slit2 (Pra-
sad et al., 2007; Wu et al., 2001). In addition, it has been sug-
gested that Robo4 heterodimerization with Robo1 is required
for Robo4 response to Slits (Sheldon et al., 2009). However,
we did not detect robust expression for either Robo1, 2, or 3
in purified hematopoietic cell populations by using qRT-PCR
under conditions that readily detected these transcripts in brain
tissue (data not shown). Additionally, we were unable to detect
Robo1 on any BM or PB cell type, including HSCs, by flow cy-
tometry by means of a monoclonal antibody that detected
Robo1 on WT, but not Robo1�/�, brain cells (Figure S1B).
These data are consistent with a recent report (Shibata et al.,
2009) and suggest that Robo4 is the predominant Robo
receptor on hematopoietic cells. Importantly, Robo4 expression
is restricted to HSCs that maintain tight interactions with the
BM niche.
Reduced BM Interaction of HSCs Lacking Robo4To assess the functional role of Robo4 in vivo, we analyzed the
frequencies of hematopoietic cells in the BM, spleen, and blood
of Robo4-deficient mice. Strikingly, analysis of cell frequencies
in the BM under normal, nonstress conditions revealed that
Robo4�/� mice displayed a significant decrease in HSC
frequencies, whereas other cell types were not affected (Fig-
ure 2A). This decrease in HSC BM frequencies was mirrored
by a reproducible increase in HSC frequencies in PB (Fig-
ure 2B). HSC numbers in the spleen were not affected (Fig-
ure S2A). To test whether the decrease in HSC BM frequencies
reflects defects in HSC proliferation, we assayed proliferative
activity in vitro and in vivo. We detected no differences in the
cell cycle status of Robo4�/� HSCs or progenitors compared
to WT mice (Figures S2B and S2C). We also tested the
in vitro expansion rates of WT and Robo4�/� HSCs, and
whether the putative Robo4 ligand Slit2 elicits a proliferative
response on WT HSCs, without detecting significant differ-
ences (Figures S2D and S2E). Consistent with these data,
Robo4�/� HSCs were as able as WT HSCs to restore hemato-
poiesis after weekly injections of the cytotoxic agent 5-fluoro-
uracil (5-FU) (Figure S2F). Thus, loss of Robo4 does not
significantly impair HSC proliferative capacity. Lower HSC
frequencies in Robo4�/� BM may instead be explained by
reduced HSC retention in the BM. This is supported by the
HSC increase in PB in Robo4�/� mice (Figure 2B) and also
by downregulation of Robo4 in M-HSCs and L-HSCs (Figure 1B)
as mobilization and leukemia lead to higher numbers of HSCs
in the PB, spleen, and liver (Morrison et al., 1997; Passegue
et al., 2004).
Robo4–/– HSCs Display Poor BM Engraftment,but Normal Differentiation CapacityTo test whether Robo4 plays a role in HSC reconstitution of
hematopoiesis upon transplantation, we competitively trans-
planted 100 HSCs from WT and Robo4�/� mice into congenic
hosts and monitored PB cell readout for 16 weeks. Robo4�/�
HSCs performed as well as WT HSCs up to 3 weeks, but failure
to provide sustained hematopoietic expansion over time
resulted in a significant difference in PB cell readout beyond
6 weeks (Figure 2C). The ratios of mature myeloid, B, and
T cells were not significantly affected by the loss of Robo4
(Figure 2D). Interestingly, we detected no differences between
WT and Robo4�/� HSCs in in vivo spleen colony-forming assays
(CFU-S12) (Figure 2E), indicating that the impaired transplanta-
tion defect is specific for the BM. Indeed, analysis of the BM of
long-term reconstituted animals revealed significantly fewer
Robo4�/� HSCs compared to WT HSCs (Figure 2F). These
data show that Robo4�/� HSCs display a specific and signifi-
cantly impaired ability to engraft in the BM. However, the
Robo4�/� HSCs that do engraft are maintained over time and
produce normal ratios of mature cells.
Cell Stem Cell 8, 72–83, January 7, 2011 ª2011 Elsevier Inc. 73
Rob
o4 m
RN
A Le
vel
A B C
D
HSCST H
SCMPPCMPGMPMEPCLP
B Cell
sT C
ells
Myeloi
dEryt
hroid
WBM
HSCM-H
SCL-H
SC
******
****
**** ** ** ** ** **
*
*
Fetal HSC
L-E14
.5L-E
17.5
HSC
BM-E17
.5
WBM
Adult BM
Rob
o4 m
RN
A Le
vel
HSC ST-HSC MPP MyPro CLP Lin+
Robo4
**
E WT KLS Robo4-/- KLS
% M
ax
Robo4
% M
ax
Rob
o4 m
RN
A Le
vel
120
100
80
60
40
20
0
Figure 1. Robo4 Is Selectively Expressed by BM-Localized HSCs
(A) Relative levels of Robo4 transcripts in purified BM populations by qRT-PCR compared to HSCs. Data shown are from four independent experiments with
qPCR reactions performed in triplicate.
(B) Relative Robo4 mRNA levels by qRT-PCR in WT HSCs, mobilized HSCs (M-HSCs), and leukemic HSCs (L-HSCs).
(C) Quantitative RT-PCR revealed that Robo4 expression increases as HSCs (defined as ckit+Lin�Sca1+ cells) transition from fetal liver (L) to BM during devel-
opment.
(D) Cell surface Robo4 expression on BM subpopulations from WT mice, demonstrating highly selective Robo4 expression on HSCs.
(E) Flow cytometry plots of ckit+Lin�Sca1+ BM cells from WT and Robo4�/� mice demonstrating the specificity of the antibody for Robo4.
BM, bone marrow. Error bars represent SEM. *p < 0.005; **p < 0.0001. See also Figure S1.
Cell Stem Cell
Robo4 Regulates HSC Location to Bone Marrow Niches
HSCs Lacking Robo4 Mobilize Less Efficientlywith Cy/G TreatmentDecreased BM frequencies at steady-state (Figure 2A) and
impaired BM engraftment (Figures 2C and 2E) of Robo4�/�
HSCs suggest that Robo4 mediates adhesive interactions
between HSC and BM niches. Consequently, Robo4 downregu-
lation upon Cy/G-induced mobilization (Figure 1B) may be
necessary for efficient HSC relocation from BM to PB. We there-
fore hypothesized that Robo4�/� HSCs would be mobilized with
greater efficiency compared to WT HSC. To test this directly, we
subjected WT and Robo4�/� mice to the Cy/G injection
74 Cell Stem Cell 8, 72–83, January 7, 2011 ª2011 Elsevier Inc.
schedule of Figure 3A. As expected, WT mice displayed robust
increases in BM HSC numbers by day 2 (�20-fold; Figure 3B)
and high numbers of PB HSCs starting at day 2 with a further
significant increase by day 4 (Figure 3C). Robo4�/� HSCs in
the BM expanded to similar levels as WT HSCs (Figure 3B),
consistent with their normal in vitro proliferation rates and prolif-
erative capacity with in vivo 5-FU treatment (Figure S2).
However, contrary to our hypothesis that Robo4�/� HSCs
would relocate to the blood more efficiently because of weak-
ened niche interactions, we detected significantly fewer
Robo4�/� HSCs in the PB at day 2 (Figure 3C). This impaired
Robo4-/-
Robo4-/- Robo4
-/-
C D
FE
A BRobo4
-/-Robo4
-/-
Robo4-/-
Robo4-/-
Robo4-/-
Figure 2. Robo4–/– HSCs Displayed
Impaired BM Localization at Steady-State
and upon Transplantation
(A and B) HSC frequencies were significantly lower
in the BM (A) and higher in PB (B) inRobo4�/�mice
compared to WT mice. Other cell types were not
affected by Robo4 loss.
(C) Robo4�/� HSCs had drastically impaired long-
term reconstitution potential upon transplantation
compared to WT HSCs. Total donor-derived cells
in PB at the indicated time points after competitive
reconstitution with 100 WT and Robo4�/� HSCs
are shown.
(D) Relative lineage readout was not affected by
Robo4 deficiency. The ratios of mature B, T, and
myeloid cells in PB, BM, and spleen >16 weeks
after competitive transplantation of 100 WT and
Robo4�/� HSCs are shown.
(E)Robo4�/�HSCs gave rise to in vivo spleen colo-
nies with normal frequencies. Lethally irradiated
mice were transplanted with either 100 Robo4�/�
or WT HSCs. Twelve days after transplantation,
spleens were harvested for CFU-S analysis.
(F) The number of Robo4�/� HSCs and progenitor
cells in the BM of transplanted mice was signifi-
cantly lower thanWT cells at >16weeks posttrans-
plantation.
All data are from at least three independent exper-
imentswith at least threemiceper groupper exper-
iment (n R 9). Error bars represent SEM.
**p < 0.004; ***p < 0.0006. See also Figure S2.
Cell Stem Cell
Robo4 Regulates HSC Location to Bone Marrow Niches
mobilization was specific for HSCs, as MPP numbers in the PB
were similar between WT and Robo4�/� mice at all time points
(Figure 3D).
Sdf1 and Cxcr4 Are Upregulated to Compensatefor Loss of Robo4To determine whether upregulation of other cell surface recep-
tors accounts for the impaired HSC mobilization in Robo4�/�
mice, we compared the expression of potentially redundant
receptors in WT and Robo4�/� HSCs. We did not detect
compensatory increases in Robo1, Robo2, or Robo3 mRNA
levels in Robo4�/� HSCs (data not shown), and we failed to
detect cell surface Robo1 on either WT or Robo4�/� HSCs (Fig-
ure S1B and data not shown). Likewise, we detected no differ-
ences in the levels of Vcam1, CD31, or Esam1 (Figure S3A).
Because Cxcr4 has been suggested to retain HSCs in BMniches
by interaction with Sdf1-expressing cells, we assayed the effect
of Robo4 deficiency on Cxcr4 expression. Strikingly, we
observed a 3-fold increase in Cxcr4 transcript levels inRobo4�/�
mice (Figure 3E). Transcription did not appear to be regulated by
levels of histone H3 trimethylation of lysine 4 (H3K4Me3) and 27
(H3K27Me3) (Figures S3B and S3C). However, elevated Cxcr4
transcript levels were paralleled by increased cell surface levels
of Cxcr4 on HSCs, but not on MPP or myeloid progenitor cells
Cell Stem Cell 8, 72–
(Figure 3F). In addition, we observed an
increase in Sdf1 mRNA levels in BM
stromal cells inRobo4�/�mice (Figure 3G).
Interestingly, expression of Slit2 was not
affected by loss of Robo4 (Figure 3H). These results demon-
strated a specific upregulation of the Sdf1/Cxcr4 axis in
Robo4�/� BM.
Intriguingly, Cy/G treatment led to decreased Sdf1 expression
in BM stromal cells in both WT and Robo4�/� mice (Figure 3G).
In addition, Cxcr4 cell surface levels increased on BM HSCs,
but decreased on HSCs in PB upon Cy/G treatment (Figure 3I).
These results suggest that daily G injections eventually over-
come Cxcr4-mediated retention of HSC, and that only the high-
est Cxcr4-expressing HSCs remain in the BM by day 4. The
observation that Cy/G treatment affects Cxcr4 levels also
support our hypothesis that the elevated levels of Cxcr4 in
Robo4�/� HSCs accounts for their poor mobilization by day 2
(Figure 3C).
Inhibition of Cxcr4 Restores Cy/G-Induced HSCMobilization Efficiency in Robo4–/– MiceIf upregulation of Cxcr4 acts as a compensatory mechanism to
counteract the loss of Robo4, inhibition of Cxcr4-mediated
interaction with BM niche components should restore the mobi-
lization efficiency of Robo4�/� HSCs. To test this possibility
directly, we performed mobilization assays by using Cy/G
combined with the Cxcr4 inhibitor AMD3100 according to the
injection schedule of Figure 4A. BM and PB analysis of HSCs
83, January 7, 2011 ª2011 Elsevier Inc. 75
A
B C D
E F G H
I
Robo4-/-
Robo4-/-
Robo4
-/-
Robo4-/-
Robo4-/-
Figure 3. Robo4–/– HSCs Mobilized Less
Efficiently with Cy/G Treatment because of
Upregulation of Cxcr4
(A) Cy/G injection and tissue analysis schedule.
(B) HSC (ckit+Lin�Sca1+Flk2� cells) expansion in
the BM in response to Cy/G was normal in
Robo4�/� mice.
(C) Fewer Robo4�/� HSCs relocated to the PB at
day 2 of Cy/G treatment. No differences between
WT and Robo4�/� HSCs were observed at day 4.
(D) The number of MPP (ckit+Lin�Sca1+Flk2+ cells)mobilized to the blood was not affected by Robo4
deficiency.
(E) Cxcr4 mRNA levels were significantly higher in
Robo4�/� HSCs compared to WT HSCs.
(F) Robo4�/� HSCs displayed higher Cxcr4 cell
surface levels than WT HSCs by flow cytometry
analysis. No differences were observed for MPP
or myeloid progenitors.
(G) BM stromal (CD45�Ter119�) cells from
Robo4�/� mice expressed higher levels of Sdf1
than WT stromal cells. Cy/G treatment led to
downregulation of Sdf1 in both WT and Robo4�/�
stromal cells.
(H) Slit2 mRNA levels in BM stromal cells were not
affected by loss of Robo4.
(I) Cxcr4 cell surface levels increased on both WT
and Robo4�/� BM HSCs, but decreased on PB
HSCs upon Cy/G treatment.
Data represent at least three (B–G) or two (H and I;
n R 10) independent experiments with at least
three mice per cohort per experiment (B–D;
n R 9). Error bars represent SEM. *p < 0.05;
**p < 0.001. See also Figure S3.
Cell Stem Cell
Robo4 Regulates HSC Location to Bone Marrow Niches
in WT mice revealed no significant differences between treat-
ment with Cy/G alone or Cy/G plus AMD3100 (Figure 4B).
Strikingly, combined Cy/G and AMD3100 treatment of
Robo4�/� mice resulted in significantly better HSC mobilization
than Cy/G alone, restoring Robo4�/� HSC levels in the PB to
that of WT HSC (Figure 4B). This effect was unique to HSCs,
as there was no differential response between WT and
Robo4�/� MPP under these conditions (Figure 4C). These
results support our hypothesis that upregulation of Cxcr4
compensates for loss of Robo4-mediated interactions between
HSC and BM niches.
Differential Mobilization of Hematopoietic Stemand Progenitors by AMD3100We also investigated the effects of AMD3100 alone on HSC
mobilization in WT and Robo4�/� mice. Although progenitor
cell numbers increased robustly in the blood 1 hr after two
sequential AMD3100 injections, we found surprisingly few
circulating HSCs in WT mice (Figure 4D). These results were
76 Cell Stem Cell 8, 72–83, January 7, 2011 ª2011 Elsevier Inc.
consistent with different injection sched-
ules and routes (i.v., s.c.). Thus, MPP
and myeloid progenitors were mobilized
more efficiently with AMD3100 than
were HSCs.
We hypothesized that the relatively low
mobilization efficiency with AMD3100 is
due to HSC retention in BM niches by non-Cxcr4-mediated,
HSC-specific interactions such as Robo4 adhesion. Intriguingly,
the efficiency of AMD3100-induced HSC, but not progenitor,
mobilization was much greater in Robo4�/� mice compared to
WTmice (Figure 4E). In vitro colony-forming assays were consis-
tent with these data (Figure S4). This supports the hypothesis
that Robo4 acts to retain HSCs in the BM niche in collaboration
with Cxcr4, and that Cxcr4 upregulation compensates for Robo4
loss.
HSCs Express Relatively Low Levels of Cxcr4 andMigrate Less Efficiently toward Sdf1When investigating Cxcr4 expression (Figure 3F), we were
surprised to find very low Cxcr4 cell surface levels on WT HSCs.
Those results and the differential response of HSCs
and progenitors to AMD3100 (Figure 4D) prompted us to investi-
gate the relative importance of Cxcr4 for different BM subpopula-
tions. We first compared Cxcr4 expression levels by qRT-PCR. In
agreement with published literature, we found very high levels of
Cxcr4 transcripts in B cells (Figure 5A). HSCs also expressed
A
B C
D E
Robo4-/-
Robo4-/-
Robo4-/-
Robo4-/-
Figure 4. Robo4–/– HSCs Were More
Responsive to AMD3100 than Were WT
HSCs
(A) Injection and analysis schedule for (B) and (C).
PB was analyzed 1 hr after AMD3100 injections on
day 2.
(B)Robo4�/�HSCs, but not WTHSCs, weremobi-
lized more efficiently by Cy/G+AMD3100 than by
Cy/G alone.
(C) Mobilization of MPP was more efficient when
AMD3100 was added to the Cy/G treatment. No
differences were observed between WT and
Robo4�/� MPP.
(D) Hematopoietic progenitors were more effi-
ciently mobilized with AMD3100 compared to
HSCs. WT mice were subjected to two AMD3100
injections 1 hr apart, with PB analysis 1 hr after
the second injection.
(E) Robo4�/� HSCs were more efficiently mobi-
lized with AMD3100 compared to WT HSCs. No
differences were observed between WT and
Robo4�/� MPP or myeloid progenitors. Injection
and analysis schedule as in (D).
MPP, multipotent progenitors; MyPro, myeloid
progenitors (Lin�cKit+Sca1� cells). Error bars
represent SEM. Data represent at least three inde-
pendent experiments with at least three mice per
cohort per experiment (n R 9). *p < 0.03;
**p < 0.01. See also Figure S4.
Cell Stem Cell
Robo4 Regulates HSC Location to Bone Marrow Niches
Cxcr4 mRNA, although at lower levels than several other cell
types. A very similar pattern was observed when analyzing
Cxcr4 cell surface levels by flow cytometry (Figure 5B), revealing
that several cell types that aremore numerous than HSCs display
much higher levels of Cxcr4 (Figure 5C).
We therefore tested the functional consequences of differen-
tial Cxcr4 levels by comparing the in vitro migratory response
of different populations to Sdf1 (Aiuti et al., 1997). Although we
detected robust and reproducible HSC migration toward Sdf1,
cell types expressing higher levels of Cxcr4 (e.g., MPP, myeloid
progenitors, and B cells) migrated with significantly greater
efficiency (Figures 5C and 5D). These results suggest that the
Sdf1/Cxcr4 axis affects hematopoietic progenitor cells to
a greater extent than HSCs, consistent with the higher mobiliza-
tion efficiency of progenitors with AMD3100 in vivo (Figures
4B–4D).
Because Robo receptors on brain and endothelial cells are
capable of mediating migratory responses to Slit ligands, we
hypothesized that Slit2 might attract or repel HSCs. However,
we did not detect HSC migration toward Slit2 (data not shown)
under conditions where HSCmigration toward Sdf1 is readily de-
tected (Figure 5D). Because Slits can act as repellants (Park
et al., 2003; Seth et al., 2005), we also tested whether Slit2 in-
Cell Stem Cell 8, 72–
hibited HSC migration toward Sdf1.
Neither preincubation of HSC with Slit2
nor addition of Slit2 to Sdf1-containing
bottom wells had an effect on Sdf1-
induced HSCmigration (Figure S5A); like-
wise, migration of CD4+ T cells was not
affected (Figure S5B). We confirmed
that Slit2 was biologically active by
demonstrating inhibition of HL60 cell migration toward fMLP
(Figure S5C). Thus, Robo4 expression on HSCs does not trans-
late to detectable migratory responses in vitro.
Robo4 and Cxcr4 Cooperate to Localize HSCs to the BMupon TransplantationThe upregulation of Cxcr4 upon loss of Robo4 (Figures 3E and
3F) and the increased mobilization efficiency with AMD3100 in
Robo4�/� mice (Figure 4E) prompted us to investigate the role
of Cxcr4 and Robo4 on HSC localization to the BM upon trans-
plantation. We first tested whether preincubation with
AMD3100 was capable of inhibiting HSC migration toward
Sdf1 in transwell migration assays. Indeed, we detected
a dose-dependent decrease in migration of both WT and
Robo4�/� HSCs, with complete inhibition at 12.5 mM of
AMD3100 (Figure 6A; Figure S6).
We then transplanted untreated and AMD3100-treated HSCs
from WT and Robo4�/� mice into lethally irradiated recipients.
Three hours postinjection, BM, spleen, and PB were analyzed
for numbers of donor cells. In contrast to in vitromigration, where
AMD3100 completely abolished migration of HSCs toward Sdf1
(Figure 6A), AMD3100 was not expected to completely inhibit
homing in vivo because Cxcr4�/� HSCs are capable of BM
83, January 7, 2011 ª2011 Elsevier Inc. 77
CX
CR
4 m
RN
A L
evel
s
* * **
*
**
*
HSCMPPCMPGMPMEPCLP
B Cell
sT C
ells
Myeloi
dEryt
hroid
A
% C
XC
R4+
Cel
ls
HSCMPPCMPGMPMEPCLP
B Cell
sT C
ells
Myeloi
dEryt
hroid
C
MyPro
* ***
***
HSC
MPP
B Cell
s
% o
f Cel
ls M
igra
ting
D
%of
Max
CXCR4
B Cells
HSC
B Figure 5. HSCs Expressed Lower Levels of Cxcr4
and Migrated Less Efficiently toward Sdf1
Compared to More Mature Hematopoietic
Subpopulations
(A–C) HSCs expressed relatively low levels of Cxcr4 by (A)
qRT-PCR analysis and (B, C) flow cytometry cell surface
staining.
(D) Transwell migration assays revealed that HSC migra-
tion efficiency toward Sdf1 was lower than that of cells
expressing higher levels of Cxcr4.
Data represent at least three independent experiments.
Error bars represent SEM. *p < 0.03; **p < 0.0001;
***p < 0.00001. See also Figure S5.
Cell Stem Cell
Robo4 Regulates HSC Location to Bone Marrow Niches
engraftment (Nie et al., 2008; Sugiyama et al., 2006). Consistent
with this observation, AMD3100 preincubation of WT cells re-
sulted in a �2-fold reduction in donor cells localizing to the BM
(Figure 6B). Loss ofRobo4 led to a comparable decrease in trans-
planted cells in the BM (Figure 6B), a notable result because this
decrease occurred despite the elevated levels of Cxcr4 on
Robo4�/� HSCs (Figure 3F). Strikingly, treatment of Robo4-defi-
cient cells with AMD3100 resulted in a further decrease in BM
localization (Figure 6B), demonstrating that both Robo4 and
Cxcr4 function to localize HSCs to the BM upon transplantation.
Consistentwith thedecreasednumber of transplanted cells in the
BM for each condition, a reciprocal increase of donor cells was
detected in the bloodstream (Figure 6C). Interestingly, there
were no differences in localization to the spleen (Figure 6D),
supporting the BM-specific effects observed with Robo4�/�
HSCs in steady state, CFU-S, and multilineage reconstitution
assays (Figure 2; Figure S2). These data demonstrate that
Robo4 and Cxcr4, individually and together, regulate HSC local-
ization to the BM.
DISCUSSION
Robo4 Regulates HSC Interactions with BM NichesWe have identified Robo4 as a critical regulator of HSC locali-
zation to the BM. Robo4 expression was very low in fetal
HSCs residing in the liver, but increased during development
concurrent with the establishment of BM hematopoiesis (Fig-
ure 1C). Thus, Robo4 is very selectively expressed by adult
78 Cell Stem Cell 8, 72–83, January 7, 2011 ª2011 Elsevier Inc.
BM HSCs and downregulation occurs not
only during normal differentiation, but also
upon HSC mobilization and in leukemogenesis
(Figures 1A and 1B). Intriguingly, these pro-
cesses all involve alterations in cell location,
concomitant with a surge in proliferation.
Although we have not yet assessed the
functional role of Robo4 in leukemic transfor-
mation, its downregulation in L-HSCs is
consistent with the proposed tumor sup-
pressor functions of Robo receptors (Dallol
et al., 2002; Legg et al., 2008; Marlow et al.,
2008). Thus, downregulation of Robo4 may
be a prerequisite for HSC exit out of BM niches
regulating HSC function. Because very few BM
cells are Robo4 positive, our data suggest that
Robo4 is an excellent HSC-specific marker. It
will be interesting to investigate the utility of Robo4, alone
and in combination with other highly specific HSC markers
such as Esam1 (Ooi et al., 2009), in simplified HSC purification
protocols.
Consistent with its HSC-specific expression, Robo4 deletion
led to perturbations in HSC localization during steady-state (Fig-
ure 2A), in short-term homing (Figure 6) and long-term reconsti-
tution assays (Figures 2C and 2F), and upon mobilization with
both Cy/G and AMD3100 (Figures 3 and 4). These effects were
specific for BM localization, as spleen readouts and in vitro
HSC properties were not affected by Robo4 loss (Figures 2E
and 6D; Figure S2). Decreased Robo4�/� HSC frequencies in
BM at steady-state indicates that Robo4 stabilizes interactions
between HSC and BM niche components. Such a function is
consistent with the poor BM localization of Robo4�/� HSCs in
short-term homing assays and dramatically impaired long-term
engraftment. Importantly, the Robo4�/� HSCs that did engraft
had normal differentiation capacity (Figure 2D). Robo4 function
therefore appears restricted to regulating HSC interactions
with the BM niche and does not appear to affect cell fate choice.
Furthermore, Robo4�/� HSCs were more efficiently mobilized
with AMD3100 than were WT HSCs (Figure 4E), indicating that
Robo4 acts to retain HSCs in BM niches. In contrast to the
increased relocation to the blood with AMD3100, Cy/G-induced
HSC mobilization was impaired in Robo4�/� mice (Figure 3C).
Investigation of the underlying molecular mechanisms revealed
that Cxcr4 was upregulated in Robo4�/� HSCs (Figures 3E and
3F), suggesting that Cxcr4 can compensate for loss of Robo4.
WT Robo4-/-
BM PB Spleen
Perc
ent r
ecov
ery
- + - +AMD
* **
A
B C D
WT Robo4-/-
- + - +WT Robo4
-/-
- + - +AMD AMD
% o
f Cel
ls M
igra
ting
HSCAMD
MPP MyPro B Cells
***
**
****
***
**
****
*
Control SDF1 SDF1+AMD 0.25 µ µ µM SDF1+AMD 2.5 M SDF1+AMD 12.5 M
Perc
ent r
ecov
ery
Perc
ent r
ecov
ery
0.4
0.3
0.2
0.1
0.0
0.7
0.3
0.2
0.1
0.0
0.6
0.5
0.4
20
15
10
5
0
Figure 6. Combined Loss of Robo4 and
Cxcr4 Function Impaired HSC Localization
to the BM after Transplantation
(A) Preincubationof cellswith increasingamounts of
AMD3100 inhibited migration toward Sdf1 in vitro.
(B) Fewer HSCs localized to the BM 3 hr after trans-
plantation when Robo4 and/or Cxcr4 function was
blocked. CFSE-labeled cells from WT and
Robo4�/�micewithandwithoutAMD3100preincu-
bationwere injected i.v. into lethally irradiated recip-
ients, followed by tissue analysis for CFSE-positive
cells 3 hr later.
(C) A reciprocal increase of Robo4�/� and
AMD3100-treated HSCs was detected in PB 3 hr
after transplantation.
(D) No significant differences in localization to the
spleen were detected.
Data represent three independent experimentswith
three to fourmiceper cohort per experiment (nR9).
Error bars represent SEM. *p < 0.03; **p < 0.003;
***p < 0.0001. See also Figure S6.
Cell Stem Cell
Robo4 Regulates HSC Location to Bone Marrow Niches
Importantly, addition of AMD3100 to the Cy/G regimen restored
the mobilization efficiency to WT levels (Figure 4B). This
demonstrates that Cxcr4 and Robo4 act together to retain
HSCs in the BM. Developmental upregulation of Robo4 and
our finding that Robo4 tethers HSCs specifically to BM niches
provide a tantalizing explanation for how HSCs gain Cxcr4 inde-
pendence once seeded in the BM (Sugiyama et al., 2006; Nie
et al., 2008).
Slit2 Does Not Affect HSC Function In VitroThe role of Slits in Robo4 function has been debated, because
high-affinity, direct binding of Slit2 protein to Robo4 protein
is not detected (Suchting et al., 2005). However, Robo4
expression endows endothelial cells with migratory responses
to Slits (Kaur et al., 2006; Park et al., 2003), and Slit2-mediated
effects in the vasculature and mammary gland are Robo4
dependent (Jones et al., 2008; London et al., 2010; Marlow
et al., 2010). These observations have led to the concept that
a coreceptor enhances the affinity of Slit2 for Robo4. Proposed
coreceptors include Robo1 (Sheldon et al., 2009) and syndecans
(Hu, 2001; Johnson et al., 2004; Steigemann et al., 2004).
Because Robo1 is not expressed by HSCs (Figure S1B), synde-
cans are more likely coreceptor candidates in HSCs. Indeed, we
have previously reported differential regulation of syndecan
Cell Stem Cell 8, 72–
family members between HSCs and
progenitor cells (Forsberg et al., 2005).
To our knowledge, the functional conse-
quences of this differential expression
have not been investigated.
The lack of Slit2 effects onHSCprolifer-
ation and migration in vitro does not
preclude an important role for Slit2 on
HSC function in vivo. Indeed, if Robo4
acts to tether HSCs to BM niches, Slits
would be expected to have little impact
in solution. Instead, lack of Slit2 effects
in vitro supports a role for Slit/Robo
signaling in niche-dependent HSC func-
tion. Upregulation of Slit2 during hematopoietic stress (Shibata
et al., 2009) argues for a physiologically important role of Slit2
in HSC function. The relative importance of this role may be
amplified in stress situations, analogous to what has been
observed upon challenges to vascular integrity (Jones et al.,
2008; London et al., 2010; Marlow et al., 2010).
Differential Efficacy of Cxcr4 Manipulationon Hematopoietic Stem and Progenitor CellsCxcr4 is a well-established regulator of HSC localization to the
BM. Surprisingly, however, we found that HSCs express rela-
tively low levels of Cxcr4, both at the transcript and cell surface
protein levels. These results contrast those by Sugiyama and
colleagues, who reported higher Cxcr4 mRNA levels in HSCs
compared to MPP (Sugiyama et al., 2006), but are consistent
with a recent report assaying Cxcr4 expression and hematopoi-
etic cell migration (Sasaki et al., 2009). Importantly, we showed
that differential Cxcr4 expression had functional consequences,
as AMD3100-induced mobilization (Figure 4D) and migration
efficiency toward Sdf1 (Figure 5D) correlated with Cxcr4 expres-
sion levels (Figure 5). Our findings have important implications
for understanding the molecular mechanisms of HSC localiza-
tion next to Sdf1-expressing cells (Sugiyama et al., 2006).
Several cell types, far more numerous than HSCs, express
83, January 7, 2011 ª2011 Elsevier Inc. 79
Figure 7. Simplified Model of Robo4- and
Cxcr4-Mediated Control of HSC Migration,
Engraftment, and Mobilization
During developmental transition of HSC location
from fetal liver to BM, or upon transplantation,
HSCs home toward BM niches by the attractant
cues between Cxcr4 and stromal-derived Sdf1.
Adhesive interactions provided by both Cxcr4
and Robo4 promote stable interactions with the
niche with long-term engraftment as a result. B
cells and other cells expressing high levels of
Cxcr4 also home to the BM, but, similar to
Robo4�/� HSCs, fail to engage in stable niche
interactions. AMD3100-induced mobilization of
HSCs into the bloodstream is more efficient
when Robo4 is deleted, in spite of increased levels
of Cxcr4.
Cell Stem Cell
Robo4 Regulates HSC Location to Bone Marrow Niches
higher levels of Cxcr4 (Figure 5) and consequently respond
better to Sdf1 and AMD3100 (Figures 4D and 5D). This includes
myeloid progenitors, B, and T cells. Therefore, molecules other
than Cxcr4must specify location of HSCs to limited niche space.
Indeed, we show that Robo4 collaborates with Cxcr4 to provide
highly HSC-specific localization cues.
Because the molecular mechanisms mobilizing mouse and
human HSCs are remarkably similar, Robo4 cooperation with
Cxcr4 have potentially important clinical implications. A bolus
injection of AMD3100 alone does not yield sufficient numbers
of HSCs for an adult transplant. Therefore, alternative injection
protocols and combinatorial use with other mobilizing agents
have been explored, including continuous AMD3100 infusion,
and AMD3100 combined with G-CSF and integrin a4 inhibitors
(Bonig et al., 2009; Flomenberg et al., 2005; Liles et al., 2003).
A mobilizing agent specifically targeting HSCs, such as an
inhibitor of Robo4-mediated adhesion, may significantly boost
HSC yield.
Robo4 and Cxcr4 Employ Distinct MolecularMechanisms to Localize HSCs to the BMThe HSC phenotype upon Robo4 loss is similar to that of condi-
tional deletion or AMD3100-mediated inhibition of Cxcr4. For
example, deletion of Robo4 and AMD3100 treatment resulted
in similar decreases in HSC localization to the BM 3 hr postinjec-
tion (Figure 6B), and at steady state, HSC BM frequencies were
decreased upon either Robo4 (Figure 2A) or Cxcr4 (Sugiyama
et al., 2006) deletion. In addition, both Robo4�/� and Cxcr4�/�
HSCs display lower long-term engraftment but retained lineage
multipotency (Figures 2C and 2D; Nie et al., 2008; Sugiyama
et al., 2006). However, important differences distinguish the
mechanisms of receptor function. Cxcr4 expression endows
HSCs with an active migratory response toward Sdf1, but we
were unable to detect such effects with Slit2. Additionally,
Cxcr4 is expressed by many hematopoietic and nonhemato-
poietic cell types, whereas Robo4 expression is highly selective
for HSCs. Indeed, our functional data demonstrate highly HSC-
specific functions for Robo4.
In a simplified model, chemoattractants, including Sdf1, guide
HSCs to the BM (Figure 7). Once in the vicinity of HSC-
supportive niches, Cxcr4 and Robo4 together promote
80 Cell Stem Cell 8, 72–83, January 7, 2011 ª2011 Elsevier Inc.
HSC retention in the niche and stable engraftment. The highly
HSC-restricted Robo4 expression probably endows HSCs with
a competitive advantage to limited BM niche space compared
to cells expressing higher levels of Cxcr4, but not Robo4. Inhibi-
tion or loss of Cxcr4 results in fewer HSCs actively migrating
toward niches. Loss of Robo4, on the other hand, probably
results in equal, or because of Cxcr4 upregulation maybe even
greater, numbers of HSCs localizing close to niches. However,
BM localization is transient in the absence of Robo4 because
fewer HSCs engage in stable niche interactions. In both cases,
decreased long-term engraftment is observed. Because of these
dual cooperative adhesive cues, both Robo4- and Cxcr4-medi-
ated interactions with the niche have to be inhibited for efficient
HSC mobilization to the blood; thus, AMD3100-induced HSC
mobilization is more efficient in Robo4-deficient mice.
Receptor Redundancy in the Control of HSC FunctionUpregulation of Cxcr4 seems to partially compensate for Robo4
loss and attenuate the phenotype of Robo4�/�mice. This is sup-
ported by the inefficient HSCmobilization with Cy/G inRobo4�/�
mice (Figure 3C) and additive effects in BM homing experiments
(Figure 6B). Likewise, engraftment of Cxcr4�/� HSCs is likely
possible due to functional redundancy with Robo4 and other
adhesion receptors expressed by HSCs. Although we did not
detect upregulation of Vcam1, Esam1, or CD31 upon Robo4
deletion, these receptors are highly expressed by HSCs
(Figure S3A), and probably contribute to HSC localization (Kikuta
et al., 2000; Ooi et al., 2009; Ross et al., 2008). In the vasculature,
Robo4 intersects with pathways regulated by VE-cadherin and
VEGF receptors. Because VEGF signaling and the sinusoidal
endothelium affects hematopoietic reconstitution (Hooper
et al., 2009), Robo4may also affect hematopoiesis by its expres-
sion in endothelial cells. We recently reported increased defects
in angiogenesis under pathological conditions in Robo4�/� mice
(Jones et al., 2008) and we also found that Robo4 controls blood
vessel growth during mammary gland development (Marlow
et al., 2010). These reports demonstrated that Robo4 is dispens-
able under homeostatic conditions, but critically important
during tissue perturbation and remodeling. Mechanistically, it
is intriguing that the Sdf1/Cxcr4 axis is upregulated in Robo1�/�
mammary glands (Marlow et al., 2008). These results point to
Cell Stem Cell
Robo4 Regulates HSC Location to Bone Marrow Niches
conservation of molecular mechanisms across tissues and
between different Robo receptors.
Several molecules have been implicated in HSC homing and
engraftment, but the relationship between these factors and
how they work together to specify HSC location is unclear.
We recently proposed a ‘‘niche code hypothesis,’’ where HSC
location is specified by a combination of factors, much like the
histone code hypothesis dictates transcriptional outcome
(Forsberg and Smith-Berdan, 2009). This model takes into
account the contribution of multiple receptors in regulating
HSC location and function. Such receptor redundancy would
also allowHSCs to respond tomultiple types of cues to stimulate
production of the appropriate cell type. We have begun to
dissect this complex regulation by establishing a functional rela-
tionship between Robo4 and Cxcr4 in controlling HSC location.
A sophisticated understanding of the molecular cues from the
endogenous niche milieu that support HSC self-renewal will be
necessary to overcome our frustrating inability to expand and
generate transplantable HSCs ex vivo.
Therapeutic Potential of Manipulating Robo4 FunctionThe responsiveness of Robo receptors to soluble ligands
renders them optimal targets for manipulation by natural or
synthetic agonists and antagonists. A relevant precedence is
provided by the clinical utility of Cxcr4 antagonists in hematopoi-
etic cell mobilization. However, Cxcr4 is expressed by many
different cell types, including the brain, leading to significant
effects on non-HSC populations, and genetic Cxcr4 deletion is
embryonic lethal. In contrast, Robo4�/� mice are viable with
mild phenotype, and Robo4 expression is restricted to HSCs
and endothelial cells. Thus, pharmacologic manipulation of
Robo4 function will probably be safe and highly specific. Once
potent modulators of Robo4 function have been identified,
Robo4 is a potentially valuable clinical target to improve the
success of HSC transplantation therapy.
EXPERIMENTAL PROCEDURES
Mice
Mice were maintained by the UCSC animal facility according to approved
protocols. Robo4�/� mice were described previously (Jones et al., 2008;
London et al., 2010; Marlow et al., 2010). WT mice were generated from het/
het breeding of the Robo4�/� mice or purchased C57Bl6 mice from JAX
(Bar Harbor, Maine). Radiation was delivered as a split dose administered
3 hr apart with a Faxitron CP-160 X-ray instrument (Lincolnshire, IL).
Competitive Reconstitution Assays
HSC were isolated from Robo4�/� (Ly5.1) or WT (Ly5.1/5.2) donors by two
rounds of FACS and administered i.v. with whole bone marrow helper cells
(3e5 cells) from Ly5.2 congenic hosts. Recipient mice were bled at 3, 6, 9,
12, and 16 weeks posttransplant via the tail vein and peripheral blood was
analyzed for donor chimerism by means of antibodies to the Ly5.1 (Alexa488)
and Ly5.2 (Alexa680) alleles and the lineage markers B220 (APC-Cy7), CD3
(PE), Mac1 (PECy7), Ter119 (PECy5), and Gr1 (Pacific Blue) (eBioscience,
Biolegend, or BD Biosciences). Statistically significant differences for all
comparisons were calculated with two-tailed t tests, unless stated otherwise.
qRT-PCR
Quantitative RT-PCR was performed as described previously (Forsberg et al.,
2005, 2006), except reactions were conducted on a Corbett cycler with the
Quantace SensiMixPlus SYBR. Expression of b-actin was used to normalize
cDNA amounts between samples.
Modified Boyden Migration Assays
BM cells (lineage depleted by magnetic selection, when appropriate), were
preincubated at 37�C for 1 hr, then placed in the upper chamber of a transwell
insert (5 mm pore size). Bottom and/or top wells contained Sdf1 (100 ng/ml)
and/or Slit2, as indicated. Cells were allowed to migrate for 2 hr at 37�C before
harvesting and analysis by flow cytometry.
Cy/G and AMD3100 Mobilization
Mice were mobilized with cytoxan and G-CSF (Cy/G) as previously described
(Morrison et al., 1997). In brief, mice were injected i.p. with 200 mg/kg of
Cytoxan in HBSS (Sigma-Aldrich) on day �1, followed by two or four sequen-
tial daily s.c. injections of 200 mg/kg rhG-CSF (Humanzyme, Chicago, IL).
Tissueswere analyzed on day 2 or 4, as indicated (Figures 3A and 4A). A cohort
from each group was injected i.v. with 5 mg/kg of AMD3100 1 hr prior to sacri-
fice. For AMD3100 alone, mice were treated with two serial AMD3100 (5 mg/
kg) i.v. injections 1 hr apart. Peripheral blood, spleen, and bone marrow
were isolated 1 hr later and processed for cell counts and flow cytometry anal-
ysis to determine the numbers and frequencies of each cell population.
BM Homing Assays
BM cells were labeled with CFSE labeling dye (Invitrogen) for 5 min at rt,
followed by antibody labeling and isolation of cKit+/Linneg/Sca1+/CFSEhi cells
by two rounds of FACS. Sorted cells were split in two equal parts and incu-
bated with or without AMD3100 (12.5 mM) on ice for 30 min. Cells were
washed, pelleted by centrifugation, and resuspended in HBSS at 400,000
cells/ml. Hosts, lethally irradiated 24 hr prior to transplantation, were injected
i.v. with 40,000 cells in 100 ml. Three hours posttransplant, tissues were har-
vested from individual mice and analyzed for CFSE-labeled cells by flow
cytometry.
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental Procedures
and six figures and can be found with this article online at doi:10.1016/
j.stem.2010.11.030.
ACKNOWLEDGMENTS
We thank Dr. Andrew Leavitt for generously providing reagents. This work was
funded by University of California Santa Cruz start-up funds (E.C.F.); California
Institute for Regenerative Medicine (CIRM) Stem Cell Training Program
Awards (A.N., F.U., and J.C.); a UCSC Minority Access to Research Careers
Fellowship (D.H.); a postdoctoral fellowship from the Government of Navarra,
Spain (J.C.); and University of California, Merced start-up funds (M.E.G.-O.).
D.L. is supported by the DOD, AAF, JDRF, and NIH. L.H. was partially funded
by NIH (RO1CA-128902). E.C.F. is the recipient of a CIRMNew Faculty Award.
University of Utah has licensed intellectual property surrounding the Robo4
pathway to Navigen. Both the University of Utah and D.Y.L. have equity in
Navigen.
Received: June 16, 2010
Revised: September 14, 2010
Accepted: October 21, 2010
Published: January 6, 2011
REFERENCES
Aiuti, A., Webb, I.J., Bleul, C., Springer, T., and Gutierrez-Ramos, J.C. (1997).
The chemokine SDF-1 is a chemoattractant for human CD34+ hematopoietic
progenitor cells and provides a new mechanism to explain the mobilization of
CD34+ progenitors to peripheral blood. J. Exp. Med. 185, 111–120.
Bonig, H., Chudziak, D., Priestley, G., and Papayannopoulou, T. (2009).
Insights into the biology of mobilized hematopoietic stem/progenitor cells
through innovative treatment schedules of the CXCR4 antagonist AMD3100.
Exp. Hematol. 37, 402–415, e1.
Broxmeyer, H.E., Orschell, C.M., Clapp, D.W., Hangoc, G., Cooper, S., Plett,
P.A., Liles, W.C., Li, X., Graham-Evans, B., Campbell, T.B., et al. (2005).
Cell Stem Cell 8, 72–83, January 7, 2011 ª2011 Elsevier Inc. 81
Cell Stem Cell
Robo4 Regulates HSC Location to Bone Marrow Niches
Rapid mobilization of murine and human hematopoietic stem and progenitor
cells with AMD3100, a CXCR4 antagonist. J. Exp. Med. 201, 1307–1318.
Dallol, A., Forgacs, E., Martinez, A., Sekido, Y., Walker, R., Kishida, T.,
Rabbitts, P., Maher, E.R., Minna, J.D., and Latif, F. (2002). Tumour specific
promoter region methylation of the human homologue of the Drosophila
Roundabout gene DUTT1 (ROBO1) in human cancers. Oncogene 21, 3020–
3028.
Flomenberg, N., DiPersio, J., and Calandra, G. (2005). Role of CXCR4 chemo-
kine receptor blockade using AMD3100 for mobilization of autologous hema-
topoietic progenitor cells. Acta Haematol. 114, 198–205.
Forsberg, E.C., and Smith-Berdan, S. (2009). Parsing the niche code: The
molecular mechanisms governing hematopoietic stem cell adhesion and
differentiation. Haematologica 94, 1477–1481.
Forsberg, E.C., Prohaska, S.S., Katzman, S., Heffner, G.C., Stuart, J.M., and
Weissman, I.L. (2005). Differential expression of novel potential regulators in
hematopoietic stem cells. PLoS Genet. 1, e28.
Forsberg, E.C., Serwold, T., Kogan, S., Weissman, I.L., and Passegue, E.
(2006). New evidence supporting megakaryocyte-erythrocyte potential of
flk2/flt3+ multipotent hematopoietic progenitors. Cell 126, 415–426.
Forsberg, E.C., Passegue, E., Prohaska, S.S., Wagers, A.J., Koeva, M., Stuart,
J.M., and Weissman, I.L. (2010). Molecular signatures of quiescent, mobilized
and leukemia-initiating hematopoietic stem cells. PLoS ONE 5, e8785.
Hooper, A.T., Butler, J.M., Nolan, D.J., Kranz, A., Iida, K., Kobayashi, M.,
Kopp, H.G., Shido, K., Petit, I., Yanger, K., et al. (2009). Engraftment and
reconstitution of hematopoiesis is dependent on VEGFR2-mediated regener-
ation of sinusoidal endothelial cells. Cell Stem Cell 4, 263–274.
Hu, H. (2001). Cell-surface heparan sulfate is involved in the repulsive guid-
ance activities of Slit2 protein. Nat. Neurosci. 4, 695–701.
Huminiecki, L., Gorn, M., Suchting, S., Poulsom, R., and Bicknell, R. (2002).
Magic roundabout is a new member of the roundabout receptor family that
is endothelial specific and expressed at sites of active angiogenesis.
Genomics 79, 547–552.
Johnson, K.G., Ghose, A., Epstein, E., Lincecum, J., O’Connor, M.B., and Van
Vactor, D. (2004). Axonal heparan sulfate proteoglycans regulate the distribu-
tion and efficiency of the repellent slit during midline axon guidance. Curr. Biol.
14, 499–504.
Jones, C.A., London, N.R., Chen, H., Park, K.W., Sauvaget, D., Stockton, R.A.,
Wythe, J.D., Suh, W., Larrieu-Lahargue, F., Mukouyama, Y.S., et al. (2008).
Robo4 stabilizes the vascular network by inhibiting pathologic angiogenesis
and endothelial hyperpermeability. Nat. Med. 14, 448–453.
Kaur, S., Castellone, M.D., Bedell, V.M., Konar, M., Gutkind, J.S., and
Ramchandran, R. (2006). Robo4 signaling in endothelial cells implies attraction
guidance mechanisms. J. Biol. Chem. 281, 11347–11356.
Kikuta, T., Shimazaki, C., Ashihara, E., Sudo, Y., Hirai, H., Sumikuma, T.,
Yamagata, N., Inaba, T., Fujita, N., Kina, T., and Nakagawa, M. (2000).
Mobilization of hematopoietic primitive and committed progenitor cells into
blood in mice by anti-vascular adhesion molecule-1 antibody alone or in
combination with granulocyte colony-stimulating factor. Exp. Hematol. 28,
311–317.
Lapidot, T. (2001). Mechanism of human stem cell migration and repopulation
of NOD/SCID and B2mnull NOD/SCID mice. The role of SDF-1/CXCR4 inter-
actions. Ann. N Y Acad. Sci. 938, 83–95.
Legg, J.A., Herbert, J.M., Clissold, P., and Bicknell, R. (2008). Slits and
Roundabouts in cancer, tumour angiogenesis and endothelial cell migration.
Angiogenesis 11, 13–21.
Liles, W.C., Broxmeyer, H.E., Rodger, E., Wood, B., Hubel, K., Cooper, S.,
Hangoc, G., Bridger, G.J., Henson, G.W., Calandra, G., and Dale, D.C.
(2003). Mobilization of hematopoietic progenitor cells in healthy volunteers
by AMD3100, a CXCR4 antagonist. Blood 102, 2728–2730.
London, N.R., Zhu, W., Bozza, F.A., Smith, M.C., Greif, D.M., Sorensen, L.K.,
Chen, L., Kaminoh, Y., Chan, A.C., Passi, S.F., et al. (2010). Targeting Robo4-
dependent Slit signaling to survive the cytokine storm in sepsis and influenza.
Sci. Transl. Med. 2, ra19.
82 Cell Stem Cell 8, 72–83, January 7, 2011 ª2011 Elsevier Inc.
Marlow, R., Strickland, P., Lee, J.S., Wu, X., Pebenito, M., Binnewies, M., Le,
E.K., Moran, A., Macias, H., Cardiff, R.D., et al. (2008). SLITs suppress tumor
growth in vivo by silencing Sdf1/Cxcr4 within breast epithelium. Cancer Res.
68, 7819–7827.
Marlow, R., Binnewies, M., Sorensen, L.K., Monica, S.D., Strickland, P.,
Forsberg, E.C., Li, D.Y., and Hinck, L. (2010). Vascular Robo4 restricts proan-
giogenic VEGF signaling in breast. Proc. Natl. Acad. Sci. USA 107, 10520–
10525.
Morrison, S.J., Wright, D.E., and Weissman, I.L. (1997). Cyclophosphamide/
granulocyte colony-stimulating factor induces hematopoietic stem cells to
proliferate prior to mobilization. Proc. Natl. Acad. Sci. USA 94, 1908–1913.
Nagasawa, T., Hirota, S., Tachibana, K., Takakura, N., Nishikawa, S.,
Kitamura, Y., Yoshida, N., Kikutani, H., and Kishimoto, T. (1996). Defects of
B-cell lymphopoiesis and bone-marrow myelopoiesis in mice lacking the
CXC chemokine PBSF/SDF-1. Nature 382, 635–638.
Nie, Y., Han, Y.C., and Zou, Y.R. (2008). CXCR4 is required for the quiescence
of primitive hematopoietic cells. J. Exp. Med. 205, 777–783.
Ooi, A.G., Karsunky, H., Majeti, R., Butz, S., Vestweber, D., Ishida, T.,
Quertermous, T., Weissman, I.L., and Forsberg, E.C. (2009). The adhesion
molecule esam1 is a novel hematopoietic stem cell marker. Stem Cells 27,
653–661.
Park, K.W., Morrison, C.M., Sorensen, L.K., Jones, C.A., Rao, Y., Chien, C.B.,
Wu, J.Y., Urness, L.D., and Li, D.Y. (2003). Robo4 is a vascular-specific
receptor that inhibits endothelial migration. Dev. Biol. 261, 251–267.
Passegue, E., Wagner, E.F., and Weissman, I.L. (2004). JunB deficiency leads
to a myeloproliferative disorder arising from hematopoietic stem cells. Cell
119, 431–443.
Passegue, E., Wagers, A.J., Giuriato, S., Anderson, W.C., and Weissman, I.L.
(2005). Global analysis of proliferation and cell cycle gene expression in the
regulation of hematopoietic stem and progenitor cell fates. J. Exp. Med.
202, 1599–1611.
Peled, A., Petit, I., Kollet, O., Magid, M., Ponomaryov, T., Byk, T., Nagler, A.,
Ben-Hur, H., Many, A., Shultz, L., et al. (1999). Dependence of human stem
cell engraftment and repopulation of NOD/SCID mice on CXCR4. Science
283, 845–848.
Prasad, A., Qamri, Z., Wu, J., and Ganju, R.K. (2007). Slit-2/Robo-1 modulates
the CXCL12/CXCR4-induced chemotaxis of T cells. J. Leukoc. Biol. 82, 465–
476.
Ross, E.A., Freeman, S., Zhao, Y., Dhanjal, T.S., Ross, E.J., Lax, S., Ahmed, Z.,
Hou, T.Z., Kalia, N., Egginton, S., et al. (2008). A novel role for PECAM-1 (CD31)
in regulating haematopoietic progenitor cell compartmentalization between
the peripheral blood and bone marrow. PLoS ONE 3, e2338.
Sasaki, Y., Matsuoka, Y., Hase, M., Toyohara, T., Murakami, M., Takahashi,
M., Nakatsuka, R., Uemura, Y., and Sonoda, Y. (2009). Marginal expression
of CXCR4 on c-kit(+)Sca-1 (+)Lineage (-) hematopoietic stem/progenitor cells.
Int. J. Hematol. 90, 553–560.
Seth, P., Lin, Y., Hanai, J., Shivalingappa, V., Duyao, M.P., and Sukhatme, V.P.
(2005). Magic roundabout, a tumor endothelial marker: Expression and
signaling. Biochem. Biophys. Res. Commun. 332, 533–541.
Sheldon, H., Andre, M., Legg, J.A., Heal, P., Herbert, J.M., Sainson, R.,
Sharma, A.S., Kitajewski, J.K., Heath, V.L., and Bicknell, R. (2009). Active
involvement of Robo1 and Robo4 in filopodia formation and endothelial cell
motility mediated via WASP and other actin nucleation-promoting factors.
FASEB J. 23, 513–522.
Shibata, F., Goto-Koshino, Y., Morikawa, Y., Komori, T., Ito, M., Fukuchi, Y.,
Houchins, J.P., Tsang, M., Li, D.Y., Kitamura, T., and Nakajima, H. (2009).
Roundabout 4 is expressed on hematopoietic stem cells and potentially
involved in the niche-mediated regulation of the side population phenotype.
Stem Cells 27, 183–190.
Steigemann, P., Molitor, A., Fellert, S., Jackle, H., and Vorbruggen, G. (2004).
Heparan sulfate proteoglycan syndecan promotes axonal and myotube guid-
ance by slit/robo signaling. Curr. Biol. 14, 225–230.
Cell Stem Cell
Robo4 Regulates HSC Location to Bone Marrow Niches
Suchting, S., Heal, P., Tahtis, K., Stewart, L.M., and Bicknell, R. (2005). Soluble
Robo4 receptor inhibits in vivo angiogenesis and endothelial cell migration.
FASEB J. 19, 121–123.
Sugiyama, T., Kohara, H., Noda, M., and Nagasawa, T. (2006). Maintenance of
the hematopoietic stem cell pool by CXCL12-CXCR4 chemokine signaling in
bone marrow stromal cell niches. Immunity 25, 977–988.
Watt, S.M., and Forde, S.P. (2008). The central role of the chemokine receptor,
CXCR4, in haemopoietic stem cell transplantation: Will CXCR4 antagonists
contribute to the treatment of blood disorders? Vox Sang. 94, 18–32.
Wright, D.E., Cheshier, S.H., Wagers, A.J., Randall, T.D., Christensen, J.L.,
and Weissman, I.L. (2001). Cyclophosphamide/granulocyte colony-stimu-
lating factor causes selective mobilization of bone marrow hematopoietic
stem cells into the blood after M phase of the cell cycle. Blood 97, 2278–2285.
Wright, D.E., Bowman, E.P., Wagers, A.J., Butcher, E.C., and Weissman, I.L.
(2002). Hematopoietic stem cells are uniquely selective in their migratory
response to chemokines. J. Exp. Med. 195, 1145–1154.
Wu, J.Y., Feng, L., Park, H.T., Havlioglu, N., Wen, L., Tang, H., Bacon, K.B.,
Jiang Zh, Zhang Xc, and Rao, Y. (2001). The neuronal repellent Slit
inhibits leukocyte chemotaxis induced by chemotactic factors. Nature 410,
948–952.
Zou, Y.R., Kottmann, A.H., Kuroda, M., Taniuchi, I., and Littman, D.R. (1998).
Function of the chemokine receptor CXCR4 in haematopoiesis and in cere-
bellar development. Nature 393, 595–599.
Cell Stem Cell 8, 72–83, January 7, 2011 ª2011 Elsevier Inc. 83
Cell Stem Cell
Article
EGFR/Ras/MAPK Signaling MediatesAdult Midgut Epithelial Homeostasisand Regeneration in DrosophilaHuaqi Jiang,1,3 Marc O. Grenley,1 Maria-Jose Bravo,1 Rachel Z. Blumhagen,1 and Bruce A. Edgar1,2,*1Division of Basic Sciences, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue N., Seattle, WA 98109, USA2German Cancer Research Center (DKFZ)-Center for Molecular Biology Heidelberg (ZMBH) Alliance, Im Neuenheimer Feld 282, D-69120,
Heidelberg, Germany3Present address: Department of Developmental Biology, UT Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd., Dallas,
TX 75390, USA
*Correspondence: b.edgar@dkfz.de
DOI 10.1016/j.stem.2010.11.026
SUMMARY
Many tissues in higher animals undergo dynamichomeostatic growth, wherein damaged or aged cellsare replaced by the progeny of resident stem cells.To maintain homeostasis, stem cells must respondto tissue needs. Here we show that in response todamage or stress in the intestinal (midgut) epitheliumof adult Drosophila, multiple EGFR ligands andrhomboids (intramembrane proteases that activatesome EGFR ligands) are induced, leading to the acti-vation of EGFR signaling in intestinal stem cells(ISCs). Activation of EGFR signaling promotes ISCdivision andmidgut epithelium regeneration, therebymaintaining tissue homeostasis. ISCs defective inEGFR signaling cannot grow or divide, are poorlymaintained, and cannot support midgut epitheliumregeneration after enteric infection by the bacteriumPseudomonas entomophila. Furthermore, ISC prolif-eration induced by Jak/Stat signaling is dependentupon EGFR signaling. Thus the EGFR/Ras/MAPKsignaling pathway plays central, essential roles inISC maintenance and the feedback system thatmediates intestinal homeostasis.
INTRODUCTION
Homeostasis and regeneration in adult tissue has long fasci-
nated biologists and clinicians alike. The discovery of resident
somatic stem cells identified the source of the remarkable regen-
erating ability in some of adult human tissues, such as blood,
skin, hair, and the digestive tract (Fuchs, 2009). However, how
stem cells respond to tissue needs remains poorly understood
(Pellettieri and Sanchez Alvarado, 2007). In particular, how
stem cells are activated (for growth, proliferation, and differenti-
ation) to regenerate new tissues after tissue injury, stress, or
normal wear and tear is still unclear in most cases.
Homeostasis in the human small intestine and colon is medi-
ated by intestinal stem cells (ISCs) that reside in the crypts of
Lieberkuhn (Barker et al., 2007; Radtke and Clevers, 2005).
84 Cell Stem Cell 8, 84–95, January 7, 2011 ª2011 Elsevier Inc.
ISCs proliferate and differentiate to give rise to new functional
epithelial cells in order to replenish cell loss from the villi. This
dynamic process is intimately linked to the development of colo-
rectal carcinoma (CRC), the second leading cause of cancer
mortality in the western world (Radtke and Clevers, 2005).
Oncological studies have established a genetic model for CRC
development involving multiple steps: mutations in the Adeno-
matous polyposis coli (Apc) gene result in the activation of WNT
signaling, which promotes the formation of small adenomas in
the form of polyps. Subsequent mutations in KRAS, BRAF,
p53, MLH1, or TGF-b signaling promote the formation of carci-
nomas, and finally additional mutations drive tumor metastasis
(Vogelstein et al., 1988; Walther et al., 2009). Activation of
receptor tyrosine kinases, particularly the epidermal growth
factor receptor (EGFR), is believed to be an early event in the
development of colon adenomas. Ectopic activation of EGFR
signaling can cause intestinal and colonic hyperplasia, a likely
precursor to ademona formation (Calcagno et al., 2008;
Sandgren et al., 1990). Consistently, genetic studies have shown
that ectopic activation of the EGFR pathway can accelerate
tumor progression in the ApcMin/+ genetic background (Bilger
et al., 2008; Haigis et al., 2008; Phelps et al., 2009). Activating
mutations in KRAS (codon 12, 13, or 61, which permanently
lock it into the GTP-bound state) and BRAF (BRAFV600E) are
among the most common mutations found in colon cancer
samples (Andreyev et al., 1998; Fransen et al., 2004; Roth
et al., 2010). Furthermore, partial loss of function of EGFR
(Egfrwa2) severely impaired adenoma formation in Apcmin/+
mice (Roberts et al., 2002). Monoclonal antibodies against
EGFR (panitumumab or cetuximab) are effective in treating
CRC, provided that activating mutations in downstream KRAS
or BRAF are not present, further emphasizing the critical role
for EGFR signaling during CRC development (Amado et al.,
2008; Di Nicolantonio et al., 2008). Developmentally, neonatal
mice lacking EGFR function develop disorganized crypts in the
gastrointestinal tract (Threadgill et al., 1995). Despite these
many indications of its importance, the precise functions of
EGFR signaling in normal gut homeostasis in mammals are
poorly understood, making studies in model systems like
Drosophila potentially informative.
As in the human intestine, the Drosophila adult midgut epithe-
lium also undergoes rapid turnover, a dynamic process
mediated by thousands of intestinal stem cells (ISCs) (Micchelli
Figure 1. Drosophila EGFR Ligands Are
Induced in the Regenerating Adult Midgut
(A) RT-qPCR quantification of Drosophila EGFR
ligands (vn, spi, and Krn) and MKP3 (MAP kinase
phosphatase-3) mRNA expression in the regener-
ating midgut. The midgut was induced to regen-
erate by activating the JNK pathway in the ECs
(MyoIAts > HepAct, 24 hr or puc RNAi, 72 hr) or
inducing EC apoptosis (MyoIAts > Rpr, 24 hr) or
Pe infection (48 hr). Error bars indicate standard
deviation (STDEV) and p values (t test) are shown
in brackets.
(B–E) Expression of vn-lacZ reporter in control (B)
or regenerating posterior midguts (C–E). Two of
the four rows of circular visceral muscle cells
(VM) were shown.
(F and G) vn fluorescent in situ hybridization. The
strongest vn signals were in the nucleus (arrows)
of VMs (asterisks), most probably the loci of Vn
transcription.
(H and I) Krn fluorescent in situ hybridization. The
strongest Krn signals were in the nucleus of ECs
(arrows).
Inmock-infected control midguts, vn andKrnwere
expressed at low levels in the VM and ECs,
respectively (F, H).
Cell Stem Cell
EGFR Regulation of Drosophila ISCs
and Perrimon, 2006; Ohlstein and Spradling, 2006). In the fly
midgut epithelium, basally localized intestinal stem cells divide,
renew themselves, and give rise to progenitors called entero-
blasts (EBs). In contrast to transit amplifying cells in mammalian
intestinal crypts, Drosophila EBs appear not to proliferate, but
directly differentiate into two conserved cell types, the absorp-
tive enterocytes (ECs) and the secretory enteroendocrine
cells (EE). Genetic studies show that the Drosophila Notch and
WNT pathways play conserved roles in the self-renewal and
proliferation of ISCs (Bardin et al., 2010; Lee et al., 2009; Lin
et al., 2008; Ohlstein and Spradling, 2007). With this simple
model, we and others previously demonstrated a feedback regu-
latory mechanism for maintaining adult tissue homeostasis. In
this case, cell loss, damage, or stress in the midgut epithelium
triggers the expression of Unpaired (Upd) cytokines by differen-
tiated enterocytes, and these signals activate Jak/Stat signaling
in intestinal stem cells to promote their proliferation and differen-
tiation (Amcheslavsky et al., 2009; Apidianakis et al., 2009;
Biteau et al., 2008; Buchon et al., 2009a; Cronin et al., 2009;
Jiang et al., 2009). This feedback provides a truly homeostatic
mechanism for tissue maintenance in the Drosophila midgut
Cell Stem Cell 8, 84–
and may explain in general how stem
cells respond to tissue needs in other
organs and organisms.
In the present study we demonstrate
that, in response to gut epithelial damage
or stress in Drosophila, multiple EGFR
ligands and several rhomboids are
induced, and these activate the EGFR/
RAS/MAPK pathway in ISCs. In parallel
with Upd/Jak/Stat signaling, the activa-
tion of EGFR signaling promotes the
proliferation of ISCs and their subsequent
differentiation into mature midgut enterocytes, thus promoting
gut self-renewal.
RESULTS
Damage or Infection of the Midgut Induces EGFRSignalingTo test whether EGFR signaling is induced in the regenerating
Drosophila adult midgut, we assayed the expression of EGFR
ligands in whole midguts via RT-qPCR. We induced midgut
epithelium regeneration by expressing the cell death gene reaper
(Rpr), or activated JNKK (Drosophila HepAct), or RNAi against
puckered (puc; a feedback inhibitor of JNK signaling) in the en-
terocytes by means of the EC-specific-inducible Gal4 driver,
MyoIAts. Alternatively, we fed flies a pathogenic bacteria, Pseu-
domonas entomophila (Pe). As we showed previously, EC
apoptosis, JNK activation, and enteric Pe infection all induce
compensatory ISC proliferation and midgut epithelial regenera-
tion (Jiang et al., 2009). We found that three Drosophila EGFR
ligands, vein (vn), spitz (spi), and Keren (Krn), were induced in
these regenerating midguts (Figure 1A). Regenerating midguts
95, January 7, 2011 ª2011 Elsevier Inc. 85
Cell Stem Cell
EGFR Regulation of Drosophila ISCs
also induced the expression of MAP Kinase Phosphatase 3
(MKP3), a downstream target of Drosophila EGFR signaling (Fig-
ure 1A). We examined the expression pattern of vn by using the
vn-lacZ reporter. Weak expression was observed exclusively in
the visceral muscle cells (VM) of control midguts, similar to its
expression in the larval midgut (Figure 1B; Jiang and Edgar,
2009). vn-lacZ expression was highly induced in the VM of the
regenerating midgut (Figures 1C–1E). The induction of vn
expression in response to Pe infection was further confirmed
by vn fluorescent in situ hybridization (Figures 1F and 1G). The
strongest signals were found in the nuclei of circular and longitu-
dinal visceral muscle cells, appearing as intense foci, probably
the loci of vn transcription (Figures 1F and 1G). Similarly, the
activation of apoptosis and JNK signaling in the ECs also
induced vn expression in the VM (data not shown). However, in
the case of ectopic JNK activation (MyoIAts > HepAct), strong
vn induction was also observed in the ECs (Figures S1A and
S1B available online), where strong signals were also found in
the cytosol. Induction of vn in the ECs by HepAct is consistent
with the much higher vn induction in these midguts detected
by RT-qPCR (Figure 1A). Fluorescent in situ hybridization further
revealed thatKrnwas induced in the ECs in response toPe infec-
tion (Figures 1H and 1I). The strongest signal appeared as
intense foci in EC nuclei. In contrast, a reporter for spi (spi-
Gal4NP0261) was mainly expressed in small progenitor cells,
with low levels of expression also observed in some ECs (Figures
S1C and S1C0).Drosophila rhomboids encode intramembrane proteases that
cleave and activate some EGFR ligands, including Spi and Krn
(Urban et al., 2002). We quantified the expression of all seven
rhomboid-like genes in the midgut by RT-qPCR and observed
modest upregulation of rho, rho2, 4, and 6 in regenerating
midguts (Figure S2A). We also examined the expression of rho
with the rhoX81-lacZ reporter. rho-lacZ was weakly expressed
in the VM (data not shown) but not in the epithelial cells of
controls (Figure S2B). Although rho-lacZ expression in the VM
did not change after infection (data not shown), its expression
was induced in the ECs (Figures S2C–S2E). The induction of
rho in the ECs in response to Pe infection was confirmed by
in situ hybridization (Figures S2F and S2G).
The induction of multiple EGFR ligands and rhos in the midgut
was also detected when flies were infected with another patho-
genic bacteria, ECC15 (Buchon et al., 2009b). We reasoned that
the induction of these factors probably activates EGFR signaling.
To test this, we examined the activity of mitogen-activated
protein kinase (MAPK), a downstream effector of EGFR, by using
antibodies against the diphosphorylated, active form of MAPK,
termed dpERK (Gabay et al., 1997). Staining for dpERK in control
midguts revealed that MAPK was mainly active in ISCs but was
weak or absent in the EBs (Figure 2A; Figures S3A–S3A00). BriefPe infection (1 day) led to increased dpERK in both ISCs and
EBs (Figures 2B and 2B0), suggesting that Pe infection induced
the activation of MAPK in midgut progenitor cells. Interestingly,
MAPK activity in the progenitor cells decreased after 2 days of
Pe infection, and ectopic MAPK activity was observed in newly
formed pre-ECs (Figures 2C and 2C0). This downregulation in
progenitors is probably the result of increased expression of
MKP3, a negative regulator of MAPK (Figure 1A; Rintelen et al.,
2003). Consistent with the activation of MAPK in midgut progen-
86 Cell Stem Cell 8, 84–95, January 7, 2011 ª2011 Elsevier Inc.
itors, ectopic induction of strong EGFR ligands (MyoIAts > sSpi)
activatedMAPK only in the progenitor cells, but not in themature
ECs (Figures 2D and 2D0). However, activated Ras (esgtsF/O >
RasV12) led to strong cell-autonomous activation of MAPK in
both progenitors and large polyploid ECs (Figures 2E and 2E0).This suggests that differentiated ECs lack a critical component
of the EGFR pathway upstream of Ras and are therefore unable
to respond to EGFR ligands. One possibility is that ECs downre-
gulate EGFR as they differentiate.
EGFR Activates ISCs through RAS/RAF/MAPK SignalingWe previously reported that EGFR signaling drives the prolifera-
tion of adult midgut progenitors (AMPs) in the larval gut and
showed that VM-derived Vn is required for AMP proliferation
during early larval development (Jiang and Edgar, 2009).
By using an inducible visceral muscle driver, 24Bts, we overex-
pressed Vn specifically in adult VM and observed amild increase
of mitotic ISCs (Figure 3A). Thus VM-derived Vn is sufficient to
induce ISC proliferation. The mild effect on ISC proliferation is
probably because Vn is a weak EGFR ligand (Schnepp et al.,
1998). Next, we ectopically activated EGFR signaling in the
ISCs by expressing the strong EGFR ligands, sSpi or sKrn (Reich
and Shilo, 2002; Schweitzer et al., 1995), activated Egfr (lTOP)
(Queenan et al., 1997), or activated Ras (RasV12) (Karim and
Rubin, 1998) by using a lineage induction system, esgtsF/O. In
the esgtsF/O system, progenitor cells and all of their newborn
progeny express Gal4 and UAS-linked Gal4 targets, including
theUAS-GFPmarker (Jiang et al., 2009). We then examined their
effects on ISC proliferation. Activation of EGFR signaling
induced increased ISC division (Figure 3B), resulting in the
generation of many new midgut cells, including EC-like
GFP+ cells (Figures 3D–3F). Most of these large GFP+ cells
were positive for PDM-1, a marker for fully differentiated ECs
(Figures 3F–3F00). Therefore, EGFR/Ras signaling does not
suppress EC differentiation. In addition, we found that knocking
down Cbl, a negative regulator of EGFR signaling (Hime et al.,
1997; Meisner et al., 1997), by Cbl RNAi (esgtsF/O > Cbl RNAi),
also induced ISC proliferation (Figure 3B; Figure S4B). Prolonged
activation of EGFR signaling resulted in severely hyperplasic
midguts (Figure S8D).
We also induced EGFR ligands in mature ECs (MyoIAts > sSpi
or sKrn). This treatment similarly promoted ISC proliferation,
demonstrating that paracrine EGF signaling is able to activate
ISC division (Figure 3B). In fact, the source of ectopic EGFR
ligands did not seem to be important. No matter where Vn,
sSpi, or sKrn were induced (VMs, ECs, or progenitors), they
were always capable of inducing dramatic ISC proliferation
(data not shown).
To ask which downstream effectors of EGFR are responsible
for inducing ISC proliferation, we ectopically expressed
pathway-specific Ras variants (RasV12S35 orRasV12G37) in midgut
progenitor cells (Karim and Rubin, 1998). RasV12S35, which
specifically activates the MAPK pathway, was able to promote
ISC proliferation, whereas induction of RasV12G37, which prefer-
entially activates the PI3K/AKT pathway, had no effect on ISC
proliferation (Figure 3B). Activated Raf (Rafgof) also promoted
ISC proliferation (Figure 3B), and coexpressingMKP3 largely in-
hibited ectopic ISC proliferation induced by RasV12 (Figure 3B).
Furthermore, depleting Capicua (Cic) (esgtsF/O > Cic RNAi),
Figure 2. MAPK Is Activated in the Regenerating Midgut
The activity of Drosophila MAPK was assayed by anti-dpERK staining.
(A and B) MAPK activity in the mock-infected control midgut (A). MAPK activity
after infecting with Pe for 1 day (B). ISCs and EBs were marked by esgGal4-
driven GFP expression and indicated by arrowheads and arrows, respectively
(A, B).
(C) MAPK activity after infecting with Pe for 2 days. Differentiating ECs (pre-
ECs,medium nucleus) and newly formedmature ECs (large nucleus) were indi-
cated by arrowheads and arrows, respectively.
(D) MAPK activation induced by ectopic expression of sSpi (MyoIAts > sSpi).
(E) Cell-autonomous MAPK activation induced by activated Ras (esgtsF/O >
RasV12).
Cell Stem Cell
EGFR Regulation of Drosophila ISCs
a transcriptional repressor downstream of MAPK pathway
(Astigarraga et al., 2007), also induced ISC proliferation (Fig-
ure 3B; Figure S4C). We conclude that EGFR signaling induces
ISC proliferation specifically through Ras, Raf, and MAPK, rather
than via PI3K or another effector pathway.
EGFR Signaling Is Required for ISC Proliferationand Midgut RegenerationTo further explore the role of EGFR signaling in the midgut, we
generated mosaic ISC clones homozygous for rasDc40b, a null
allele (Schnorr andBerg, 1996), orEgfr (Egfrnull,Egfr[CO]) (Clifford
and Schupbach, 1989), or both ras and stat function (ras and
Stat92Edouble nullmutants, rasDc40b, stat397) (Silver andMontell,
2001) via theMARCMsystem (LeeandLuo, 2001).We thenquan-
tified the size of marked ISC clones at intervals after clone induc-
tion. Although the initial growth of ras and Egfrmutant ISC clones
was normal, their long-term proliferation was severely compro-
mised (Figures 4A–4E). For ras and statdoublemutant, the clones
were not only small, but also lacked ECs (Figure 4D), a phenotype
consistent with Jak/Stat’s critical role for ISC differentiation
(Beebe et al., 2010; Jiang et al., 2009). Consistent with the
EGFR pathway’s essential role in ISC proliferation, midgut
renewal after Pe infection was completely inhibited when EGFR
signaling was suppressed in the progenitor cells by Egfr RNAi
(Figures 4G–4J). Furthermore, prolonged EGFR suppression in
healthy animals (4 weeks) led to almost complete loss of entero-
blasts (esg+, Su(H)+) and�33% reduction of intestinal stem cells
(esg+, Su(H)�) (Figures 4F and 4I). In the short term, however,
EGFR suppression did not significantly alter the number of
ISCs, but probably only prevented their growth and division.
Interestingly, old ECs generated before the induction of lineage
marking were still present in these agedmidguts (�1month, Fig-
ure 4I), suggesting that EC loss were also partially inhibited.
Next we tested whether EGFR signaling is required for
compensatory ISC proliferation and midgut epithelium regener-
ation induced by Pe infection. We first examined the growth of
control ISC clones in Pe-infected midgut and observed large
ISC clones (�7 cells/clone) 2 days after clone induction (Fig-
ure 4E). However, the ISC clones lacking ras or Egfr function
were much smaller (�3 cells/clone). Like the long-term ras or
Egfr mutant ISC clones in noninfected midguts, these clones
did not grow even after the flies had recovered from Pe infection
for about a week (Figure 4E). Quantification of midgut mitotic
indices revealed that Pe-induced compensatory ISC prolifera-
tion was completely inhibited when Egfr or Raf was knocked
down (esgtsF/O > Egfr RNAi or Raf RNAi; Figure 4K). Further-
more, although Pe infection almost completely eliminated old
Cell Stem Cell 8, 84–95, January 7, 2011 ª2011 Elsevier Inc. 87
Figure 3. EGFR Signaling Promotes ISC
Proliferation and Midgut Growth
(A) Ectopic ISC proliferation induced by Vn. Vnwas
induced in the midgut via the inducible VM-
specific driver 24Bts.
(B) ISC proliferation induced by activated EGFR
signaling. Transgenes were induced in the midgut
for 2 days via the esgtsF/O or MyoIAts system.
Midguts were scored for PH3+ mitotic figures in
both (A) and (B). Error bars represent standard
deviation (STDEV) in (A) and (B).
(C–E) Adult midgut growth measured via the
esgtsF/O system. Both sSpi (D) and lTOP (E)
promoted significant new midgut cell formation.
(F) RasV12 also promoted the formation of new
mature midgut cells. Most of the newly formed
large polyploid midgut cells (GFP+, arrows) were
positive for mature EC marker, PDM-1.
Cell Stem Cell
EGFR Regulation of Drosophila ISCs
ECs and induced midgut epithelial regeneration in controls
(Figures 4L and 4M), suppression of EGFR signaling largely in-
hibited midgut epithelium regeneration (Figures 4N and 4O; Fig-
ure S5). In both cases, however, large numbers of progenitor
cells expressing these RNAis survived for the duration of the
experiment. In summary, EGFR signaling is required for ISC
proliferation during both normal midgut homeostasis and regen-
eration, such as that induced by Pe infection.
Multiple EGFR Ligands Function Redundantlyto Activate ISC ProliferationTo examine the function of EGFR ligands and rhomboid during
Drosophila midgut homeostasis and regeneration, we knocked
down spi, vn, and rho individually in the midgut via RNAi and
several midgut-specific drivers, including esgts, MyoIAts, and
24Bts. Inducing spi RNAi in midgut progenitors (esgts > spi
RNAi), vn RNAi in visceral muscle cells (24Bts > vn RNAi), or
rho RNAi in the ECs (MyoIAts > rho RNAi) all significantly knocked
down target gene expression (Figure S6A). In each case,
however, these RNAi-depleted midguts appeared to be normal,
even after long periods of gene knockdown (data not shown). We
88 Cell Stem Cell 8, 84–95, January 7, 2011 ª2011 Elsevier Inc.
then orally infected the flies with Pe and
quantified ISC proliferation. Pe infection-
induced ISC proliferation also appeared
normal in these RNAi-depleted midguts
(Figure 4Q; Figure S6B). Finally we exam-
ined the regenerative response in the
midguts of Krn (krn27-7-B, viable null), rho
(rhoA0544, viable partial loss-of-function),
and Star (Sd01624, viable partial loss-of-
function) mutants (Corl et al., 2009;
McDonald et al., 2006). In these cases
ISC proliferation induced by Pe infection
was also normal (Figure 4P; Figure S6B).
In further tests we quantified Pe-
induced ISC proliferation in spi and Krn
double mutants. In this case we found
that heterozygosity for spi in a Krn homo-
zygous mutant background (spiA14/+;
Krn27-7-B/Krn27-7-B) significantly reduced
Pe-induced ISC proliferation (Figure 4P). Our previous analysis
indicated that this double mutant does not affect the develop-
ment of the adult midgut progenitor (AMPs) in larvae (Jiang
and Edgar, 2009), and quantification of esg+ cells indicated
that these midguts had normal numbers of progenitor cells
(data not shown). Hence, the suppression of ISC mitotic
response suggests that spi and Krn function redundantly during
midgut epithelium regeneration. To test which cell types are the
source of spi expression, we knocked down spi expression with
RNAi, driven either by the esgts driver (progenitor-specific) or the
MyoIAts driver (EC-specific) in a Krnmutant background. Knock-
ing down spi in progenitor cells (esgts > spi IR, Krn27-7-B/Krn27-7-
B) but not ECs (MyoIAts > spi IR; Krn27-7-B/Krn27-7-B) significantly
reducedmidgutmitoses induced byPe ingestion (Figure 4Q).We
surmise that autocrine spi (from progenitor cells) and paracrine
Krn (from ECs) function redundantly to promote ISC proliferation
during midgut epithelium regeneration.
We next tested vein function, by using RNAi to deplete vn in
the visceral muscle of Krn mutant animals, via the 24Bts driver.
Simultaneous loss of Krn and vn (24Bts > vn IR, Krn27-7-B/
Krn27-7-B) significantly reduced the ISC proliferation (Figure 4Q),
Cell Stem Cell
EGFR Regulation of Drosophila ISCs
Cell Stem Cell 8, 84–95, January 7, 2011 ª2011 Elsevier Inc. 89
Cell Stem Cell
EGFR Regulation of Drosophila ISCs
suggesting that vn andKrn also have overlapping function during
midgut epithelium regeneration.
EGFRSignaling Is Required for ISCProliferation Inducedby Jak/Stat SignalingBecause both EGFR and Jak/Stat signaling are sufficient and
required for midgut epithelium regeneration and both pathways
are induced in the regenerating midgut (Figures 1–4; Buchon
et al., 2009a; Cronin et al., 2009; Gabay et al., 1997; Jiang
et al., 2009), we examined their epistatic relationship. We first
ectopically activated EGFR signaling and examined the expres-
sion of the Upd cytokines by RT-qPCR. When activated EGFR
ligand (MyoIAts > sKrn), activated Egfr (esgtsF/O > lTOP), or acti-
vated Ras (esgtsF/O > RasV12) were expressed in the midgut, all
three Upd cytokines were induced, along with downstream
target gene Socs36E (Figure 5A). Consistently, the upd-lacZ
reporter was induced in the midgut epithelial cells by RasV12
(Figures 5C and 5D). Similarly, when we ectopically activated
EGFR signaling (MyoIAts > RasV12), the upd3 reporter, upd3.1-
lacZ, was induced in the ECs (Figures 5E and 5F). Accordingly,
RasV12 expression in the ECs was capable of inducing ISC prolif-
eration (Figure 3B). The induction of cytokines and subsequent
activation of Jak/Stat signaling probably depends on the levels
of EGFR activation because the inductions by sKrn were much
lower than that by activated EGFR (lTOP) or RasV12 (Figure 5A).
Moreover ectopic expression of Vn (24Bts > Vn), a weak EGFR
ligand, did not induce cytokine expression (data not shown),
though it did promote mild ISC proliferation (Figure 3A).
We next asked what signals might induce Vn expression in the
visceral muscle. We observed increased nuclear STAT92E stain-
ing in the VM of Pe-infected midguts (Figures S7A and S7B),
suggesting that Jak/Stat signaling was activated in the VM.
Consistent with this, expression of the Jak/Stat reporter
10XSTAT-DGFP increased dramatically in the VM after Pe
infection (Figures S7C and S7D). Because the induction of vn
coincided with enhanced cytokine signaling in the VM, we spec-
ulated that it might be the result of Upds (cytokine) released from
the midgut epithelium. In testing this idea, we found that vn and
the vn-lacZ reporter could be induced in the VM in response to
EC-specific expression of Upd (MyoIAts > Upd) (Figures 5B,
5G, and 5H). Activating Jak/Stat signaling directly in the VM via
the expression of Drosophila Jak (24Bts > Hop) also induced
comparable vn expression (Figure 5B). These experiments
Figure 4. Drosophila EGFR Signaling Is Required for Midgut Homeosta
(A–D) MARCM analysis of ISC clones. Wild-type (A) and mutant ISC clones (B–D) w
of cells in each clone were indicated.
(E) Quantification of ISC clone sizes. The number of clones counted for each gen
(F) Quantification of progenitor cells in the posterior midguts of GFP and EGFR
Su(H)+) were indicated by squares, and presumed ISCs (esg+, Su(H)�) were indicat
(G–J) Midgut epithelium turnover assay. EGFR suppression inhibited midgut tu
depleted after long-term EGFR knockdown (I). In control midgut, GFP were pres
(J, esgtsF/O > GFP).
(K) Quantification of compensatory ISC proliferation induced by Pe infection. EGF
Raf RNAi.
(L–O) Midgut turnover in mock (L, N) or Pe-infected (M, O) animals. Midgut turno
(P and Q) Quantification of compensatory ISC proliferation in spi, vn, and Krnmuta
heterozygous background), spi RNAi knockdown in progenitors (esgts > spi IR) or E
repeats.
Error bars represent STDEV in (E), (F), (K), (P), and (Q).
90 Cell Stem Cell 8, 84–95, January 7, 2011 ª2011 Elsevier Inc.
indicate that midgut epithelium-derived cytokines can activate
Jak/Stat signaling and induce vn expression in the VM. However,
we found that Pe infection could induce vn upregulation in the
midguts of Jak mutants (hop25, partial loss-of-function) or
when statwasdepleted in the VM (24Bts > Stat RNAi; Figure S7E).
These data indicate that, although activated Jak/Stat signaling
can induce vn, Jak/Stat signaling is not required for vn induction
in response to Pe infection.
Further epistasis tests showed that when EGFR signaling was
activated in the background of reduced Jak/Stat signaling
(esgtsF/O > sKrn + Stat or Dome RNAi), its stimulatory effect
on ISC proliferation was not diminished (Figure 6A; Figures
S8D–S8F). Similar results were obtained when activated Egfr
(lTOP) or Ras (RasV12) was coexpressed with Stat or Dome
RNAi (data not shown). By using the MARCM technique, we
induced activated Ras in ISCs mutant for Stat (+RasV12, stat397)
and analyzed their clonal growth. Loss of Jak/Stat signaling
did not affect RasV12’s ability to drive the growth of large ISC
clones (Figures 6F and 6G). However, in a similar experiment,
clonal growth induced by the weak EGFR ligand, Vn, was largely
inhibited by loss of Stat (Figures 6C, 6D, and 6K). These data
suggest that the requirement of Jak/Stat signaling for ISC prolif-
eration probably depends on the levels of EGFR activation, such
that high-level EGFR activation is able to induce ISC proliferation
independent of Jak/Stat signaling, whereas ISC proliferation
induced by low-level EGFR activation (such as that induced by
Vn) is largely dependent on Jak/Stat signaling.
In further experiments we found that ISC proliferation induced
by ectopic Upd was completely inhibited when EGFR signaling
was downregulated in the ISCs (Figure 6A). Knocking down
Egfr or Ras completely inhibited the midgut hyperplasia pheno-
type that results from ectopic Upd expression (esgtsF/O >
Upd + Egfr or Ras RNAi; Figures S8G–S8I). Similar results were
obtained in a clonal setting, with the rasDc40b mutant allele
(Figures 6I–6K). Thus EGFR signaling is required for ISC prolifer-
ation induced by Jak/Stat signaling. However, activating Jak/
Stat and EGFR signaling simultaneously induced a much
higher ISC mitotic index than that induced by the activation
of either pathway alone (MyoIAts > Upd + sSpi; Figure 6A), indi-
cating that the two pathways can function synergistically to
induce ISC proliferation. Like the Jak/Stat signaling (Beebe
et al., 2010), EGFR signaling can also induce much higher
rate of ISC proliferation when Notch signaling is inhibited
sis and Regeneration
ere induced with the MARCM system and examined 8 days later. The number
otype were indicated inside each bar.
knockdown. Progenitor cells (esg+) were indicated by diamonds, EBs (esg+,
ed by triangles. Filled symbols, esgts >GFP; open symbols, esgts > EGFRRNAi.
rnover (H, esgtsF/O > Egfr RNAi). Furthermore, GFP+ progenitor cells were
ent in both progenitors and large polyploid cells (probably ECs) after 2 weeks
R signaling was suppressed in the progenitor cells by esgtsF/O-driven Egfr or
ver was assayed via the esgtsF/O system.
nts. We used viable Krn null mutant (Krn27-7-B), lethal spi null mutant (spiA14, in a
Cs (MyoIAts > spi IR), or vn RNAi knockdown in VMs (24Bts > vn IR). IR, inverted
Figure 5. Induction of EGFR and Jak/Stat
Signaling in the Midgut
(A) Activating EGFR signaling induced Jak/Stat signaling
in the midgut. The expression levels of Drosophila cyto-
kines (upds) and downstream target gene, Socs36E, in
the midgut were analyzed by RT-qPCR.
(B) Induction of vn expression in the midgut by Jak/Stat
signaling as quantified by RT-qPCR. Jak/Stat signaling
was activated in the VM by ectopic expression of Upd
in the ECs (MyoIAts > Upd) or Hop directly in the VM
(24Bts > Hop).
Error bars represent STDEV in both (A) and (B).
(C and D) Induction of the upd-lacZ reporter in the midgut
epithelium by activated Ras (esgtsF/O > RasV12, D).
(E and F) Induction of the Upd3.1-lacZ reporter in ECs by
activated Ras (MyoIAts > RasV12, F).
(G and H) Induction of the vn-lacZ reporter in the VM by
ectopic expression of Upd (MyoIAts > Upd, H).
Cell Stem Cell
EGFR Regulation of Drosophila ISCs
(esgtsF/O > sKrn + N IR; Figure 6A). Because Notch suppression
increases stem cell pools, this suggests that both pathways
primarily regulate ISC division, rather than ISC numbers.
Finally, we examined whether the induction of Upd/Jak/Stat
and EGFR signaling by Pe infection depended on each other.
We inhibited Pe-induced midgut epithelium regeneration by
knocking down Egfr (esgtsF/O > Egfr RNAi) or Stat (esgtsF/O >
Stat RNAi) and examined the expression of upds and Socs36E
or Egfr ligands and rhos by RT-qPCR. The induction of Jak/Stat
and EGFR signaling by Pe was normal in both cases (Figure 6L),
suggesting that these two signaling pathways can be induced
independently of each other by midgut damage (Figure 7).
DISCUSSION
EGFR Signaling Is Essential for ISC Growth and DivisionThese studies show that the EGFR pathway provides an
essential mitogenic signal for ISC proliferation during midgut
homeostasis and regeneration (Figure 4). Furthermore, ISC
proliferation induced by Jak/Stat signaling depends on
functional EGFR signaling (Figures 6A and 6H–6K; Figure S8G–
S6I). The critical role of EGFR signaling in the flymidgut is consis-
tent with its role during mammalian gut homeostasis and colo-
rectal cancer development. EGFR signaling is required for the
Cell Stem Cell
development, maintenance, and tumorigenesis
of mucosal epithelium in the mouse GI tract
(Roberts et al., 2002; Threadgill et al., 1995;
Troyer et al., 2001). Antibodies targeting EGFR
have been shown to be effective in treating
colorectal cancer provided there are no acti-
vating mutations in downstream signaling
components, such as KRAS or BRAF (Amado
et al., 2008; Di Nicolantonio et al., 2008).
Our data also demonstrate that EGFR
signaling is induced in response to damage in
the Drosophilamidgut and functions to promote
ISCproliferationduringmidgut epithelium regen-
eration (Figures1–3). In thiscapacity it is acentral
and essential component of the feedbackmech-
anism for adult tissue homeostasis that we
described previously (Figure 7; Jiang et al., 2009). Like EGFR
ligands in Drosophila, two mammalian EGFR ligands, epiregulin
and amphiregulin, have been reported to be upregulated in the
gut epithelium after damage (Lee et al., 2004; Nishimura et al.,
2008). Their expression is also increased in neoplastic lesions in
the colon, suggesting a possible role in colon cancer develop-
ment (Nishimura et al., 2008).
One of our more unexpected findings was that, whereas
differentiating immature cells (preECs) were often positive for
MAPK activity, fully differentiated midgut cells such as ECs
were not (Figures 2C and 2C0). A potential explanation for this
is that mature ECs lose EGFR or a downstream effector and
thereby become unresponsive to EGFR ligands. This is consis-
tent with our data showing that MAPK could be activated only
in progenitor cells (ICSs and EBs) even when activated EGFR
ligands (such as sSpi) were ectopically expressed at high levels
(Figures 2D and 2D0). A similar mechanism may confine the
activity of Jak/Stat signaling to the midgut progenitor cells
(Beebe et al., 2010; Buchon et al., 2009a; Jiang et al., 2009).
In this case Domeless, the receptor for the Upd cytokines, is
expressed in the midgut progenitor cells but not in their
progeny (Jiang et al., 2009). Switching off receptor expression
for cytokines or growth factors may be one way to ensure
that mature differentiated cells do not respond to these
8, 84–95, January 7, 2011 ª2011 Elsevier Inc. 91
Figure 6. Jak/Stat-Induced ISC Proliferation Requires EGFR
Signaling
(A) ISC proliferation induced by EGFR and Jak/Stat signaling. With
the exception of coexpressing sKrn and Upd in the ECs (MyoIAts >
Upd + sKrn), all the other ectopic expression experiments were per-
formed with the esgtsF/O driver. Midgut mitotic indices (PH3+) were
quantified after activating the transgenes for 2 days.
(B–J) ISC clonal assay. GFP-marked ISC clones were induced with
the MARCM system and analyzed 4 or 8 days later. The sizes of
the ISC clones were indicated. Vn-induced ISC proliferation is depen-
dent on Jak/Stat signaling (B–D). Activated Ras (RasV12)-induced ISC
proliferation is independent of Jak/Stat signaling (F, G). Some EB
clones overexpressingRasV12 underwent extra round of endoreplica-
tion (E). Upd-induced ISC proliferation is dependent on EGFR
signaling (H–J).
(K) Quantification of ISC clone sizes. The sizes of ISC clones were
measured 4 or 8 days after clone induction (ACI) via the MARCM
system.
(L) RT-qPCR analysis of the induction of Jak/Stat and EGFR signal-
ings by Pe infection in the absence of either pathway (esgtsF/O >
Stat or Egfr RNAi).
Error bars represent STDEV in (A), (K), and (L).
Cell Stem Cell
EGFR Regulation of Drosophila ISCs
92 Cell Stem Cell 8, 84–95, January 7, 2011 ª2011 Elsevier Inc.
Figure 7. UpdatedModel forMidgut Homeostasis and Regeneration
in Drosophila
Stressed or dying ECs induce the expression of fly cytokines (such as Upd3
and Upd2) and EGFs (such as Krn and Vn) in the midgut, which activate the
Jak/Stat and EGFR pathways in the midgut progenitor cells. Whereas EGFR
signaling functions mainly to promote ISC proliferation, Jak/Stat signaling
functions to promote both ISC proliferation and EB differentiation.
Cell Stem Cell
EGFR Regulation of Drosophila ISCs
mitogenic cues. Despite this failsafe mechanism, the expres-
sion of RasV12 was able to induce the cell-autonomous activa-
tion of MAPK (Figures 2E and 2E0) and the expression of Upd3
in the ECs (Figures 5E and 5F), leading to a non-cell-autono-
mous stimulation of ISC proliferation (Figure 3B). This suggests
that the downregulation of mitogen receptors upon differentia-
tion may be important to throttle EGFR)/Jak/Stat positive
feedback that might otherwise result in run-away signaling
and ISC proliferation.
As with the Upd cytokines, we know little about how the
Drosophila EGFR ligands are induced by stress or damage to
the midgut epithelium. In the case of the Upds, potential acti-
vating stress signals span a very wide range, including induced
apoptosis, autophagic cell death, JNK signaling, infection by
pathogenic bacteria, colonization by nonpathogenic enteric
bacteria, ingestion of detergents, oxidative stress inducers,
DNA damaging agents, and even physical ‘‘pinching’’ of the
epithelium (Amcheslavsky et al., 2009; Apidianakis et al., 2009;
Biteau et al., 2008; Buchon et al., 2009a; Cronin et al., 2009;
Jiang et al., 2009). The signals capable of activating the EGFR
ligands are likely to be just as diverse. Further genetic studies
in the fly should be able to determine whether these stress
responses are cell autonomous or a property of the epithelium
as a tissue and to identify the genes and pathways involved.
Given the critical roles of the mammalian Jak/Stat and EGFR
pathways in regulating tissue homeostasis and cancer develop-
ment, such studies should have some clinical relevance.
Is Visceral Muscle a Niche for ISCs?Expression of wingless (wg, a Drosophila Wnt) from the visceral
muscle (VM) has been reported to regulate ISC proliferation and
self-renewal, leading to the proposal that visceral muscle serves
as a niche for ISCs (Lin et al., 2008). However, although
DrosophilaWnt signaling appears to be required for ISC survival
(Lin et al., 2008), its role in promoting ISC self-renewal was not
confirmed in another independent study (Lee et al., 2009). In
addition, ISC proliferation induced by ectopic Wnt signaling is
much weaker than that induced by Jak/Stat or EGFR signaling
(Jiang et al., 2009; Lee et al., 2009; Lin et al., 2008). Thus,
although the role of VM-derived Wg in midgut homeostasis
and regeneration has not been rigorously tested, the data
suggest that other signaling systems play more critical roles.
Pertinent to the function of the visceral muscle, we discovered
that the EGFR ligand vnwas induced in the VM during gut regen-
eration (Figure 1), and that VM-derived Vn was capable of
inducing ectopic ISC proliferation (Figure 3A). This suggested
that the VM might serve as a part of the ISC niche by providing
a mitogenic signal. However, Pe-induced compensatory ISC
proliferation was not affected when we specifically downregu-
lated vn in the VM (Figure 4Q), suggesting that VM-derived Vn
is probably not by itself an essential EGFR ligand during midgut
epithelium regeneration. In fact, we also observed the induction
of two other EGFR ligands (spi and Krn) in midgut epithelial cells
during regeneration (Figure 1). Although the concurrent expres-
sion of multiple EGFR ligands complicated our efforts to identify
the exact role of each ligand, single and double mutant analysis
suggested that all three ligands have overlapping function in
activating EGFR signaling (Figures 4P and 4Q). Importantly,
a significant fraction of the mitogenic EGFR signals probably
come from the epithelium itself. Similarly, the Upd cytokines
are induced primarily in midgut epithelial cells (Buchon et al.,
2009a; Jiang et al., 2009). Moreover, the self-renewal and differ-
entiation of Drosophila intestinal stem cells are regulated by
Notch signaling, which occurs between the two daughter cells
produced after ISC division and is not known to directly involve
the VM (Bardin et al., 2010; Micchelli and Perrimon, 2006; Ohl-
stein and Spradling, 2006, 2007).
Therefore we propose that the most important component of
the niche for fly intestinal stem cells may be the midgut epithe-
lium itself. In this context it is interesting to note that an epithelial
niche has also been proposed for mouse intestinal stem cells
(Sato et al., 2009). The murine Lgr5+ ISCs reside at the bottom
of the crypts, juxtaposed directly with Paneth cells (Barker
et al., 2007). In vitro culture of individual Lgr5+ ISCs has demon-
strated that they can form self-organizing organoids in the
absence of mesenchymal cells. Lgr5+ ISCs are normally always
in contact with Paneth cells, which have been proposed to be
a niche for ISCs (Sato et al., 2009). Interestingly, EGF is one of
the factors required in the media to support the growth of intes-
tinal organoids (Sato et al., 2009). However, it is not yet clear
which cells are the endogenous source for EGFR ligands in the
mouse intestine or colon, nor which specific ligands are
expressed or functionally important. It is tempting to speculate
that Paneth cells, as a critical niche component, might be one
of the sources of mitogenic signals, such as EGFs and cytokines,
for mammalian intestinal stem cells.
EXPERIMENTAL PROCEDURES
Fly Genetics
See Supplemental Experimental Procedures for fly stocks used in this study.
Cell Stem Cell 8, 84–95, January 7, 2011 ª2011 Elsevier Inc. 93
Cell Stem Cell
EGFR Regulation of Drosophila ISCs
Upd3-lacZ Reporters
To generate upd3-lacZ reporters, four genomic PCR fragments (upd3.1-4, see
primer sequences in the Supplemental Experimental Procedures) covering the
original �4 kb upd3 promoter region (Agaisse et al., 2003) were digested with
BamHI/KpnI and cloned into the same restriction sites of pH-Pelican vector.
Transgenic lines were established through standard P-element-mediated
transformation.
RNA In Situ Hybridization in the Adult Midgut
RNA fluorescent in situ hybridization (FISH) in the midgut was performed as
described (Raj et al., 2008) with a few modifications. In brief, 40–48 20-mer
DNA oligos complementing the coding region of the target genes (vn, krn,
and rho) were designed with online software (http://www.singlemoleculefish.
com/designer.html). The oligos were synthesized with 30 amine modification
(Biosearch Technologies), then manually pooled and coupled with Alexa-
568, carboxylic acid, succinimidyl ester (Invitrogen A-20003). The labeled
oligos were purifiedwith HPLC (reverse phase C-18 column) and vacuumdried
and resuspended in 100 ml H2O. For RNA in situ hybridization, the midguts
were first dissected and fixed in 8% paraformalhyde overnight at 4�C, thenwashed with PBS and Triton X-100 (0.1%) for 3 times (15 min each). The
samples were further permeablized in 70% ethanol overnight at 4�C.The probes were used at dilution 1:2,000–10,000. The hybridization was
then performed according to the online protocol (http://www.
singlemoleculefish.com/protocols.html).
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental Procedures
and eight figures and can be found with this article online at doi:10.1016/j.
stem.2010.11.026.
ACKNOWLEDGMENTS
We thank Celeste Berg, Denise Montell, Gyeong-Hun Baeg, Erika Bach,
Jocelyn McDonald, Matthew Freeman, and the VDRC (Austria), NIG (Japan),
Bloomington (USA) Drosophila Stock Centers for fly stocks; the Moen’s lab
for confocal imaging; Xiaohang Yang for Pdm-1 antibody; David D. O’Keefe
for advice on anti-dpERK staining; and members of the B.A.E. lab for
comments. This work was supported by NIH grant R01 GM51186 to B.A.E.
Received: April 20, 2010
Revised: September 20, 2010
Accepted: October 25, 2010
Published online: December 16, 2010
REFERENCES
Agaisse, H., Petersen, U.M., Boutros, M., Mathey-Prevot, B., and Perrimon, N.
(2003). Signaling role of hemocytes in Drosophila JAK/STAT-dependent
response to septic injury. Dev. Cell 5, 441–450.
Amado, R.G., Wolf, M., Peeters, M., Van Cutsem, E., Siena, S., Freeman, D.J.,
Juan, T., Sikorski, R., Suggs, S., Radinsky, R., et al. (2008). Wild-type KRAS is
required for panitumumab efficacy in patients with metastatic colorectal
cancer. J. Clin. Oncol. 26, 1626–1634.
Amcheslavsky, A., Jiang, J., and Ip, Y.T. (2009). Tissue damage-induced intes-
tinal stem cell division in Drosophila. Cell Stem Cell 4, 49–61.
Andreyev, H.J., Norman, A.R., Cunningham, D., Oates, J.R., and Clarke, P.A.
(1998). Kirsten ras mutations in patients with colorectal cancer: themulticenter
‘‘RASCAL’’ study. J. Natl. Cancer Inst. 90, 675–684.
Apidianakis, Y., Pitsouli, C., Perrimon, N., and Rahme, L. (2009). Synergy
between bacterial infection and genetic predisposition in intestinal dysplasia.
Proc. Natl. Acad. Sci. USA 106, 20883–20888.
Astigarraga, S., Grossman, R., Dıaz-Delfın, J., Caelles, C., Paroush, Z., and
Jimenez, G. (2007). A MAPK docking site is critical for downregulation of
Capicua by Torso and EGFR RTK signaling. EMBO J. 26, 668–677.
94 Cell Stem Cell 8, 84–95, January 7, 2011 ª2011 Elsevier Inc.
Bardin, A.J., Perdigoto, C.N., Southall, T.D., Brand, A.H., and Schweisguth, F.
(2010). Transcriptional control of stem cell maintenance in the Drosophila
intestine. Development 137, 705–714.
Barker, N., van Es, J.H., Kuipers, J., Kujala, P., van den Born, M., Cozijnsen,
M., Haegebarth, A., Korving, J., Begthel, H., Peters, P.J., and Clevers, H.
(2007). Identification of stem cells in small intestine and colon by marker
gene Lgr5. Nature 449, 1003–1007.
Beebe, K., Lee, W.C., and Micchelli, C.A. (2010). JAK/STAT signaling coordi-
nates stem cell proliferation and multilineage differentiation in the Drosophila
intestinal stem cell lineage. Dev. Biol. 338, 28–37.
Bilger, A., Sullivan, R., Prunuske, A.J., Clipson, L., Drinkwater, N.R., and Dove,
W.F. (2008). Widespread hyperplasia induced by transgenic TGFalpha in
ApcMin mice is associated with only regional effects on tumorigenesis.
Carcinogenesis 29, 1825–1830.
Biteau, B., Hochmuth, C.E., and Jasper, H. (2008). JNK activity in somatic
stem cells causes loss of tissue homeostasis in the aging Drosophila gut.
Cell Stem Cell 3, 442–455.
Buchon, N., Broderick, N.A., Chakrabarti, S., and Lemaitre, B. (2009a).
Invasive and indigenous microbiota impact intestinal stem cell activity through
multiple pathways in Drosophila. Genes Dev. 23, 2333–2344.
Buchon, N., Broderick, N.A., Poidevin, M., Pradervand, S., and Lemaitre, B.
(2009b). Drosophila intestinal response to bacterial infection: activation of
host defense and stem cell proliferation. Cell Host Microbe 5, 200–211.
Calcagno, S.R., Li, S., Colon, M., Kreinest, P.A., Thompson, E.A., Fields, A.P.,
and Murray, N.R. (2008). Oncogenic K-ras promotes early carcinogenesis in
the mouse proximal colon. Int. J. Cancer 122, 2462–2470.
Clifford, R.J., and Schupbach, T. (1989). Coordinately and differentially
mutable activities of torpedo, the Drosophila melanogaster homolog of the
vertebrate EGF receptor gene. Genetics 123, 771–787.
Corl, A.B., Berger, K.H., Ophir-Shohat, G., Gesch, J., Simms, J.A., Bartlett,
S.E., and Heberlein, U. (2009). Happyhour, a Ste20 family kinase, implicates
EGFR signaling in ethanol-induced behaviors. Cell 137, 949–960.
Cronin, S.J., Nehme, N.T., Limmer, S., Liegeois, S., Pospisilik, J.A., Schramek,
D., Leibbrandt, A., Simoes, Rde.M., Gruber, S., Puc, U., et al. (2009). Genome-
wide RNAi screen identifies genes involved in intestinal pathogenic bacterial
infection. Science 325, 340–343.
Di Nicolantonio, F., Martini, M., Molinari, F., Sartore-Bianchi, A., Arena, S.,
Saletti, P., De Dosso, S., Mazzucchelli, L., Frattini, M., Siena, S., and
Bardelli, A. (2008). Wild-type BRAF is required for response to panitumumab
or cetuximab in metastatic colorectal cancer. J. Clin. Oncol. 26, 5705–5712.
Fransen, K., Klintenas, M., Osterstrom, A., Dimberg, J., Monstein, H.J., and
Soderkvist, P. (2004). Mutation analysis of the BRAF, ARAF and RAF-1 genes
in human colorectal adenocarcinomas. Carcinogenesis 25, 527–533.
Fuchs, E. (2009). The tortoise and the hair: slow-cycling cells in the stem cell
race. Cell 137, 811–819.
Gabay, L., Seger, R., and Shilo, B.Z. (1997). MAP kinase in situ activation atlas
during Drosophila embryogenesis. Development 124, 3535–3541.
Haigis, K.M., Kendall, K.R., Wang, Y., Cheung, A., Haigis, M.C., Glickman,
J.N., Niwa-Kawakita, M., Sweet-Cordero, A., Sebolt-Leopold, J., Shannon,
K.M., et al. (2008). Differential effects of oncogenic K-Ras and N-Ras on prolif-
eration, differentiation and tumor progression in the colon. Nat. Genet. 40,
600–608.
Hime, G.R., Dhungat, M.P., Ng, A., and Bowtell, D.D. (1997). D-Cbl, the
Drosophila homologue of the c-Cbl proto-oncogene, interacts with the
Drosophila EGF receptor in vivo, despite lacking C-terminal adaptor binding
sites. Oncogene 14, 2709–2719.
Jiang, H., and Edgar, B.A. (2009). EGFR signaling regulates the proliferation of
Drosophila adult midgut progenitors. Development 136, 483–493.
Jiang, H., Patel, P.H., Kohlmaier, A., Grenley, M.O., McEwen, D.G., and Edgar,
B.A. (2009). Cytokine/Jak/Stat signaling mediates regeneration and homeo-
stasis in the Drosophila midgut. Cell 137, 1343–1355.
Karim, F.D., and Rubin, G.M. (1998). Ectopic expression of activated Ras1
induces hyperplastic growth and increased cell death in Drosophila imaginal
tissues. Development 125, 1–9.
Cell Stem Cell
EGFR Regulation of Drosophila ISCs
Lee, T., and Luo, L. (2001). Mosaic analysis with a repressible cell marker
(MARCM) for Drosophila neural development. Trends Neurosci. 24, 251–254.
Lee, D., Pearsall, R.S., Das, S., Dey, S.K., Godfrey, V.L., and Threadgill, D.W.
(2004). Epiregulin is not essential for development of intestinal tumors but is
required for protection from intestinal damage. Mol. Cell. Biol. 24, 8907–8916.
Lee, W.C., Beebe, K., Sudmeier, L., and Micchelli, C.A. (2009). Adenomatous
polyposis coli regulates Drosophila intestinal stem cell proliferation.
Development 136, 2255–2264.
Lin, G., Xu, N., and Xi, R. (2008). Paracrine Wingless signalling controls self-
renewal of Drosophila intestinal stem cells. Nature 455, 1119–1123.
McDonald, J.A., Pinheiro, E.M., Kadlec, L., Schupbach, T., and Montell, D.J.
(2006). Multiple EGFR ligands participate in guiding migrating border cells.
Dev. Biol. 296, 94–103.
Meisner, H., Daga, A., Buxton, J., Fernandez, B., Chawla, A., Banerjee, U., and
Czech, M.P. (1997). Interactions of Drosophila Cbl with epidermal growth
factor receptors and role of Cbl in R7 photoreceptor cell development. Mol.
Cell. Biol. 17, 2217–2225.
Micchelli, C.A., and Perrimon, N. (2006). Evidence that stem cells reside in the
adult Drosophila midgut epithelium. Nature 439, 475–479.
Nishimura, T., Andoh, A., Inatomi, O., Shioya, M., Yagi, Y., Tsujikawa, T., and
Fujiyama, Y. (2008). Amphiregulin and epiregulin expression in neoplastic and
inflammatory lesions in the colon. Oncol. Rep. 19, 105–110.
Ohlstein, B., and Spradling, A. (2006). The adult Drosophila posterior midgut is
maintained by pluripotent stem cells. Nature 439, 470–474.
Ohlstein, B., and Spradling, A. (2007). Multipotent Drosophila intestinal stem
cells specify daughter cell fates by differential notch signaling. Science 315,
988–992.
Pellettieri, J., and Sanchez Alvarado, A. (2007). Cell turnover and adult tissue
homeostasis: from humans to planarians. Annu. Rev. Genet. 41, 83–105.
Phelps, R.A., Chidester, S., Dehghanizadeh, S., Phelps, J., Sandoval, I.T., Rai,
K., Broadbent, T., Sarkar, S., Burt, R.W., and Jones, D.A. (2009). A two-step
model for colon adenoma initiation and progression caused by APC loss.
Cell 137, 623–634.
Queenan, A.M., Ghabrial, A., and Schupbach, T. (1997). Ectopic activation of
torpedo/Egfr, a Drosophila receptor tyrosine kinase, dorsalizes both the
eggshell and the embryo. Development 124, 3871–3880.
Radtke, F., and Clevers, H. (2005). Self-renewal and cancer of the gut: two
sides of a coin. Science 307, 1904–1909.
Raj, A., van den Bogaard, P., Rifkin, S.A., van Oudenaarden, A., and Tyagi, S.
(2008). Imaging individual mRNA molecules using multiple singly labeled
probes. Nat. Methods 5, 877–879.
Reich, A., and Shilo, B.Z. (2002). Keren, a new ligand of the Drosophila
epidermal growth factor receptor, undergoes two modes of cleavage.
EMBO J. 21, 4287–4296.
Rintelen, F., Hafen, E., and Nairz, K. (2003). The Drosophila dual-specificity
ERK phosphatase DMKP3 cooperates with the ERK tyrosine phosphatase
PTP-ER. Development 130, 3479–3490.
Roberts, R.B.,Min, L.,Washington, M.K., Olsen, S.J., Settle, S.H., Coffey, R.J.,
and Threadgill, D.W. (2002). Importance of epidermal growth factor receptor
signaling in establishment of adenomas and maintenance of carcinomas
during intestinal tumorigenesis. Proc. Natl. Acad. Sci. USA 99, 1521–1526.
Roth, A.D., Tejpar, S., Delorenzi, M., Yan, P., Fiocca, R., Klingbiel, D., Dietrich,
D., Biesmans, B., Bodoky, G., Barone, C., et al. (2010). Prognostic role of
KRAS and BRAF in stage II and III resected colon cancer: results of the trans-
lational study on the PETACC-3, EORTC 40993, SAKK 60-00 trial. J. Clin.
Oncol. 28, 466–474.
Sandgren, E.P., Luetteke, N.C., Palmiter, R.D., Brinster, R.L., and Lee, D.C.
(1990). Overexpression of TGF a in transgenic mice: induction of epithelial
hyperplasia, pancreatic metaplasia, and carcinoma of the breast. Cell 61,
1121–1135.
Sato, T., Vries, R.G., Snippert, H.J., van de Wetering, M., Barker, N., Stange,
D.E., van Es, J.H., Abo, A., Kujala, P., Peters, P.J., and Clevers, H. (2009).
Single Lgr5 stem cells build crypt-villus structures in vitro without a mesen-
chymal niche. Nature 459, 262–265.
Schnepp, B., Donaldson, T., Grumbling, G., Ostrowski, S., Schweitzer, R.,
Shilo, B.Z., and Simcox, A. (1998). EGF domain swap converts a drosophila
EGF receptor activator into an inhibitor. Genes Dev. 12, 908–913.
Schnorr, J.D., and Berg, C.A. (1996). Differential activity of Ras1 during
patterning of the Drosophila dorsoventral axis. Genetics 144, 1545–1557.
Schweitzer, R., Shaharabany, M., Seger, R., and Shilo, B.Z. (1995). Secreted
Spitz triggers the DER signaling pathway and is a limiting component in embry-
onic ventral ectoderm determination. Genes Dev. 9, 1518–1529.
Silver, D.L., and Montell, D.J. (2001). Paracrine signaling through the JAK/
STAT pathway activates invasive behavior of ovarian epithelial cells in
Drosophila. Cell 107, 831–841.
Threadgill, D.W., Dlugosz, A.A., Hansen, L.A., Tennenbaum, T., Lichti, U., Yee,
D., LaMantia, C., Mourton, T., Herrup, K., Harris, R.C., et al. (1995). Targeted
disruption of mouse EGF receptor: effect of genetic background on mutant
phenotype. Science 269, 230–234.
Troyer, K.L., Luetteke, N.C., Saxon, M.L., Qiu, T.H., Xian, C.J., and Lee, D.C.
(2001). Growth retardation, duodenal lesions, and aberrant ileum architecture
in triple null mice lacking EGF, amphiregulin, and TGF-a. Gastroenterology
121, 68–78.
Urban, S., Lee, J.R., and Freeman, M. (2002). A family of Rhomboid intramem-
brane proteases activates all Drosophila membrane-tethered EGF ligands.
EMBO J. 21, 4277–4286.
Vogelstein, B., Fearon, E.R., Hamilton, S.R., Kern, S.E., Preisinger, A.C.,
Leppert, M., Nakamura, Y., White, R., Smits, A.M., and Bos, J.L. (1988).
Genetic alterations during colorectal-tumor development. N. Engl. J. Med.
319, 525–532.
Walther, A., Johnstone, E., Swanton, C., Midgley, R., Tomlinson, I., and Kerr,
D. (2009). Genetic prognostic and predictive markers in colorectal cancer. Nat.
Rev. Cancer 9, 489–499.
Cell Stem Cell 8, 84–95, January 7, 2011 ª2011 Elsevier Inc. 95
Cell Stem Cell
Short Article
Reprogramming Factor Expression InitiatesWidespread Targeted Chromatin RemodelingRichard P. Koche,1,2,3,7 Zachary D. Smith,1,4,5,7 Mazhar Adli,1,2 Hongcang Gu,1 Manching Ku,1,2 Andreas Gnirke,1
Bradley E. Bernstein,1,2,5,6 and Alexander Meissner1,4,5,*1Broad Institute, Cambridge, MA 02142, USA2Department of Pathology, Massachusetts General Hospital and Harvard Medical School, Boston, MA 02114, USA3Division of Health Sciences and Technology, MIT, Cambridge, MA 02139, USA4Department of Stem Cell and Regenerative Biology, Harvard University, Cambridge, MA 02138, USA5Harvard Stem Cell Institute, Cambridge, MA 02138, USA6Howard Hughes Medical Institute, Boston, MA 02114, USA7These authors contributed equally to this work
*Correspondence: alexander_meissner@harvard.edu
DOI 10.1016/j.stem.2010.12.001
SUMMARY
Despite rapid progress in characterizing transcrip-tion factor-driven reprogramming of somatic cellsto an induced pluripotent stem cell (iPSC) state,many mechanistic questions still remain. To gaininsight into the earliest events in the reprogrammingprocess, we systematically analyzed the transcrip-tional and epigenetic changes that occur during earlyfactor induction after discrete numbers of divisions.We observed rapid, genome-wide changes in theeuchromatic histone modification, H3K4me2, atmore than a thousand loci including large subsetsof pluripotency-related or developmentally regulatedgene promoters and enhancers. In contrast, patternsof the repressive H3K27me3 modification remainedlargely unchanged except for focused depletionspecifically at positions where H3K4 methylation isgained. These chromatin regulatory events precedetranscriptional changes within the correspondingloci. Our data provide evidence for an early, orga-nized, and population-wide epigenetic response toectopic reprogramming factors that clarify thetemporal order through which somatic identity isreset during reprogramming.
INTRODUCTION
Exposure to ectopic transcription factors has been established
as a robust way to shift somatic cells toward alternative somatic
states and to pluripotency (Graf and Enver, 2009). Ectopic
expression of four transcription factors, Oct4, Sox2, Klf4, and
c-Myc (OSKM), is capable of directing cells from any tissue
toward the formation of induced pluripotent stem cells (iPSCs)
in mouse and human (Hanna et al., 2010). Fully reprogrammed
iPSCs can contribute to all germ layers and can form complete,
fertile mice by tetraploid embryo complementation (Hanna et al.,
2010). Moreover, iPSCs are similar to their embryo-derived
96 Cell Stem Cell 8, 96–105, January 7, 2011 ª2011 Elsevier Inc.
counterparts on a molecular level, indicating a genome-wide
cascade of transcriptional and epigenetic changes that lead to
a stable, newly acquired state (Mikkelsen et al., 2008).
Despite the remarkable fidelity that governs the transition to
pluripotency, the overall frequency in which it occurs within
induced populations is low and requires an extended latency
of one or several weeks (Jaenisch and Young, 2008). Previous
studies and the general reprogramming timeline suggest
a requirement for secondary or stochastic events through which
certain cells acquire unique advantages that permit transition to
pluripotency (Hanna et al., 2009; Jaenisch and Young, 2008;
Meissner et al., 2007; Yamanaka, 2009). Therefore, the ectopic
expression of the current set of embryonic factors appears insuf-
ficient to completely reset the somatic nucleus alone and the
mechanism of action probably includes the activation of addi-
tional yet unidentified downstream effectors.
Recent evidence suggests that certain phases of the reprog-
ramming process may be more coordinated than previously
assumed. This includes live imaging analysis that demonstrates
conserved transitions within reprogramming populations (Smith
et al., 2010). Transcriptional profiling and RNAi screening in clon-
ally reprogramming populations have demonstrated that robust
silencing of somatic transcription factors and effectors as well as
activation of critical epithelial markers, govern the most imme-
diate definitive transition from fibroblast toward a ‘‘primed’’ or re-
programming amenable state; the output of somatic factor
repression or intermediate stabilizing signaling factors have
demonstrated improved iPSC colony generation that suggests
that this phase is an essential early step (Samavarchi-Tehrani
et al., 2010). Despite recent progress, the global nature and scale
of these early events as well as their impact on transcriptional
and epigenetic landscapes remain unknown.
To gain more insight into the early events during reprogram-
ming, we assayed global gene expression, chromatin state,
and DNA methylation in populations of induced fibroblasts that
have undergone a discrete number of divisions. We find that
dynamic transcription within the reprogramming population is
limited and restricted to promoters with pre-existing euchro-
matin. In contrast to the relative rarity of transcription changes,
we found that euchromatin-associated H3K4 methylation is
a predominant global early activating response and occurs in
Cell Stem Cell
Targeted Chromatin Remodeling during Reprogramming
the absence of transcriptional activation at corresponding loci.
Interestingly, these targets include the promoters of many
essential pluripotency-related and developmentally regulated
genes and describe a coherent shift in cellular identity. We
observe highly localized, coordinated depletion of repressive
chromatin (H3K27me3) exclusively at promoters where H3K4
methylation is gained. Finally, this targeted remodeling extends
to enhancers across the genome, which transition dramatically
from the somatic state, and represents an additional level
of cell state transition. Taken together, our results suggest
that early transcriptional dynamics are largely dependent on
pre-existing, accessible chromatin and that ectopic factor
induction initiates a concerted change in target chromatin
through which pluripotent targets are primed for subsequent
activation.
RESULTS
CFSE Labeling Enables Enrichment of Cells that HaveUndergone Discrete Numbers of Cell DivisionsTo further elucidate critical early steps in the reprogramming
process, we investigated responses to reprogramming factor
expression in cells that had undergone no cell division and cells
that had divided 1, 2, or more than 3 times. By using inducible
(OSKM) secondary mouse embryonic fibroblasts (MEFs), we
could ensure rapid and homogenous induction of the four factors
as described previously (Mikkelsen et al., 2008; Wernig et al.,
2008). We isolated doxycycline-induced cells that had under-
gone a defined number of cell divisions by combining the live
stain CFSE (carboxyfluorescein succinimidyl ester) and a serum
pulsing protocol. Four distinct fractions were enriched based
upon their mean proliferative number in a manner that ensures
that proliferation is the predominant experimental variable
(Figure 1A). All cells were collected in an arrested (serum-
starved) state except the final sample, which was allowed to
divide continuously under factor induction. We confirmed
that the relative fluorescence intensity remains unchanged in
the serum-starved control compared to a serum-starved,
doxycycline-induced population that remains exposed to the
reprogramming factors for 96 hr and experiences minimal
or no cell division (Figure 1A). Importantly, CFSE-labeled
cells that proliferated continuously for 96 hr (with a fluorescence
reduction indicating three or more divisions) show highly
similar global transcriptional attributes to populations that
had not undergone CFSE labeling or serum withdrawal,
demonstrating that this protocol does not interfere with the
general reprogramming process (Figures S1A and S1B available
online).
Transcriptional Dynamics of Early ReprogrammingPopulations Are Limited to Sites with Pre-existingH3K4 TrimethylationWe next used our discrete cell populations to investigate the
early gene expression and chromatin dynamics induced by the
four factors. Global mRNA expression profiles revealed contin-
uous trends across populations and a primary response to factor
induction that operates almost exclusively within accessible
H3K4me3 chromatin (Figure 1B, 97%, Fisher’s exact test p <
10�16). Upregulated (2-fold, t test p < 0.05) targets are predom-
inantly associated with promoter histone H3K4me3 in MEFs
prior to induction, and moreover are enriched 2.2-fold for loci
that are H3K4me3 within ESCs (Figure 1B). Repressed genes
(2-fold, t test p < 0.05) were enriched for H3K4me3 only
or H3K4me3/H3K27me3 (bivalent) promoters in MEFs, but
enriched 2.8-fold for the bivalent state in pluripotent cells
(Figure 1B). Both activated and repressed gene sets exhibited
preferential promoter binding for the induced factors, with an
asymmetric bias for enhanced expression among c-Myc-regu-
lated targets (9.5-fold increased likelihood, Fisher’s exact text
p < 10�16), consistent with its function in the transition to tran-
scriptional elongation as opposed to PolII recruitment/initiation
(Figure 1C; Rahl et al., 2010). These observations indicate that
early expression changes mediated by factor induction are in
large part constrained by pre-existing chromatin and may
operate only at promoters that are already in an open and acces-
sible state. Moreover, these changes occur immediately and
gradually increase with additional cell divisions (Figures S1C
and S1D). These data suggest that in the earliest phase of
reprogramming, fibroblast identity is predominantly perturbed
by transcriptional silencing of somatic targets and not the activa-
tion of pluripotency-associated targets of the reprogramming
factors.
Activating Chromatin Marks Are Targetedto Promoters prior to Transcriptional ActivationNext we investigated the consequences of ectopic factor activity
at the chromatin level by comparing the dynamics of functional
epigenetic markers to the more limited observations that could
be made when measuring transcriptional output alone. We
generated genome-wide chromatin maps for the three methyla-
tion marks on H3K4 (mono-, di-, and trimethylation) as well as for
H3K27 trimethylation and H3K36 trimethylation across the
isolated populations via ChIP-Seq (Mikkelsen et al., 2007). We
then focused our initial query on H3K4me2, because it is
a general marker of both promoter and enhancer regions and
is broadly amenable to genome-wide analysis (as opposed to
trimethylation that is exclusive to promoters) (Bernstein et al.,
2005; Heintzman et al., 2007). H3K27me3 was chosen as
a marker associated with transcriptional silencing, in particular
of developmental transcription factors (Bernstein et al., 2006;
Lee et al., 2006; Mikkelsen et al., 2007). Comparison with previ-
ously published data sets confirms that our serum-starvation
protocol does not induce significant chromatin changes in the
MEFs (Figures S1E and S1F), and ChIP followed by quantitative
PCR for representative loci confirms the trends observed in our
ChIP-Seq results (Figure S1G).
Surprisingly, H3K4me2 peaks exhibit dramatic changes at
more than 1500 genes and continuously increase with succes-
sive cell divisions (Figure 1D). The results highlight two striking
findings. First, H3K4me2 target loci do not correspond to
observed changes in gene expression (Figure 1E, chi square
test p > 0.1). Furthermore, changes in H3K4me2 are apparent
even in populations that have not yet divided based on CFSE
intensity (Mann-Whitney U test p < 10�16). Notably, these regions
are strongly enriched for pluripotency and developmentally
regulated targets, such as Sall4, Lin28, and Fgf4, which will not
become transcriptionally active until later stages of iPSC forma-
tion. These results provide insights into the reprogramming
Cell Stem Cell 8, 96–105, January 7, 2011 ª2011 Elsevier Inc. 97
A B
C
K4me3 biv K27me3 noneMEF:ES: K4me3
K4me3 biv K27me3 nonebivalent
Gen
e C
ount
184
Klf4
Oct4
Sox2 Myc
Klf4
Myc
Oct4
Sox2
Upregulated 2-fold Downregulated 2-foldCFSE Intensity
Rel
ativ
e E
nric
hmen
t
CFSE Intensity
Fac
tor
Indu
ctio
n
>3 2 1 0
Division
Sample
Coll
ectio
n
0 hr
96 hr
24 hr
48 hr
72 hr
Division
0
1
2
>3
Labe
ling
Serum
Withdraw
al
1030
50
Divisions Relative Intensity (A.U.)
0 (Control)
1
2
>3
0
46,887
47,057
28,967
14,010
4,593
(571)
(503)
(462)
(209)
(69)
(103)
(268)
(244)
100
80
60
40
20
0101 102 103 104 105
D
de novo H3K4me2 enhanced H3K4me2
loss of H3K4me2
E
MEF H3K4me2
Lin28Fgf4Sall4Pecam1
Rex1TdhAireFoxd3
Postn Mmp1b
NodalPax6Utf1Klf2Lin28b
EsrrbCdh1Neurog2Cbx2Onecut1
De novo H3K4me2
Enhanced H3K4me2
Loss of H3K4me2n=115
n=358
n=1083
0 5 10 15 20 25 30 35
>3
Div
H3K
4me2
020
4060
8010
0
x=y
x=y
10 100 1000 10000
1010
010
0010
000
MEF gene expression
>3
Div
gen
e ex
pres
sion
All genes∆H3K4me2 genes
Dynamic chromatin+expressionDynamic chromatin
Up Down012>3
012>3
Figure 1. Global Transcriptional and Epigenetic Dynamics during Early Induction of Reprogramming Factors(A) Schematic for enrichment of distinct proliferative cohorts by means of the live dye CFSE and serum pulsing under constant factor induction and time. After
96 hr of continued culture in doxycycline-supplemented medium, samples were scored via flow cytometry. Median fluorophore intensity was assessed as a rela-
tive metric for proliferative number and is shown on the right. Relative intensity is displayed in arbitrary units (A.U.).
(B) mRNA expression dynamics conditional on MEF/ES chromatin state progressing across cell division number (shown color coded in the inset) for up- and
downregulated genes. ESCH3K4me3-only loci and their respective states inMEFs are shown on the left, and ESC bivalent (H3K4me3/H3K27me3) loci are shown
on the right.
(C) Enrichment for Oct4, Sox2, Klf4, and c-Myc (OSKM) binding in promoter elements of dynamically regulated genes shows an asymmetric bias toward gene
activation within targets of the myc oncogene. Transcription factor binding taken from genome-scale profiling of embryonic stem cells (Kim et al., 2008; Marson
et al., 2008).
(D) Density plot of genes with dynamic H3K4me2 in reprogramming populations compared to control MEFs. Promoters exhibiting a dynamic shift in H3K4me2
(n z 1500) fall into three distinct classes: de novo (beige), enhanced (red), and loss (green). Representative genes from all three classes are highlighted on the
right.
(E) Expression data between starting state (control) and the >3 divisions induced population with dynamic H3K4me2 genes highlighted in red. Pie chart shows the
representation of genes that exhibit only H3K4me2 changes (pink) or both H3K4me2 and gene expression changes (red; n z 10%).
Cell Stem Cell
Targeted Chromatin Remodeling during Reprogramming
process and describe an unexpected chromatin-remodeling
response to the reprogramming factors that precedes transcrip-
tional activation of ESC-exclusive genes (Figure S2A). We
confirmed this observation with the transcriptionally associated
histone mark H3K36me3, which exhibits no enrichment at
98 Cell Stem Cell 8, 96–105, January 7, 2011 ª2011 Elsevier Inc.
identified loci across the early reprogramming phase or outside
of pluripotent cell types, and by RNA PolII occupancy at repre-
sentative promoters, which did not yield apparent enrichment
when compared to established iPSC lines (Figures S2B and
S2C). This suggests that complete chromatin remodeling to
Cell Stem Cell
Targeted Chromatin Remodeling during Reprogramming
transcriptional initiation is either unstable or not yet established
during this early phase.
For further analysis, we subdivided loci that gain H3K4me2
during early reprogramming into two classes: a set of ‘‘de
novo’’ H3K4me2 loci that have essentially undetectable
H3K4me2 levels in MEFs and a set of ‘‘enhanced’’ H3K4me2
loci whose H3K4me2 signals increase by a minimum of
2.5-fold relative to the MEF control (Figures 2A and 2B). In
both cases, the chromatin changes are reproducible across
the target loci and increase in magnitude with cell divisions,
suggestive of a progressive and coordinatedprocess (Figure 2C).
A third class of promoters was less represented but exhibited
a loss of promoter H3K4me2 that correlates with transcription-
ally silenced somatic determinants such as Postn (Figure 2D,
1.75-fold decrease in expression, nz 110 genes,Mann-Whitney
U test p < 0.02). Overall, the changes in promoter H3K4me2
occur rapidly and are primarily targeted to a set of loci that
function in early development or as active mediators of pluripo-
tency, including epigenetic reprogramming of the endogenous
Sox2, Klf4, and c-Myc promoters themselves (Figures S2D and
S2E). Moreover, promoters gaining H3K4me2 are significantly
enriched for targets of Oct4 and Sox2 (Figure 2E, Fisher’s exact
test p < 0.0009 and 0.00039 for Oct4 and Sox2, respectively).
We next investigated the positioning of the related histone
marks H3K4me1 and H3K4me3 to explore potential overlaps
with H3K4me2. Surprisingly, we find that H3K4me2 is exclusive
within the de novo promoter set, which is devoid of all forms of
H3K4 methylation in MEF controls and does not gain
H3K4me1 or H3K4me3 concurrently with H3K4me2 (Figure 2F).
Alternatively, the ‘‘enhanced’’ promoter set, which exhibits both
H3K4me2 and H3K4me3 within control populations, coordi-
nately increases both marks as induced populations continue
to proliferate (Figure 2F). These data emphasize the value of
H3K4me2 as a dynamic mark across promoters because it
detects nascent histone modification at de novo promoters,
which are under-enriched for these marks in MEFs, as well as
increased representation of pre-existing chromatin modifica-
tions within enhanced promoters that are augmented by ectopic
factor activity. Additionally, within pluripotent cells, H3K4me3 is
enriched at the vast majority of genes that gain H3K4me2 within
the early reprogramming phase. These H3K4me2-exclusive
promoters may therefore imply a decoupled and transiently
stable epigenetic mechanism that precedes complete remodel-
ing and gene activation.
The dynamic gain of H3K4 methylation occurs without
promoter-wide changes in somatically defined, repressive
H3K27me3 when inspected across the entirety of target
promoters (Figure S3A; Kolmogorov-Smirnov test p > 0.1). The
retention of somatic heterochromatin at the same promoters
highlights a possible barrier that prevents gene activation and
suggests that repressive modifications might be less dynamic
than H3K4me2.
Repressive H3K27me3 Is Lost Specifically at Siteswhere H3K4 Methylation Is GainedWe next investigated the positional context of H3K4me2 to
explore possible epigenetic or genetic determinants of the early
response to ectopic factor induction. EnhancedH3K4me2 peaks
occur directly at transcription start sites (TSS) in two distinct
promoter classes: those that will ultimately be activated at the
iPS cell stage and those that are not activated but are rather reset
to a poised bivalent state (Figure 3A, Figure S3B). The positional
gain of H3K4me2 is targeted to the TSS and does not display the
bimodality seen in ESCs/iPSCs that is associated with nucleo-
some depletion at the site of initiation (Figure 3B, shaded region).
We also examined chromatin changes at the subset of
promoters with H3K27me3 in MEFs. Here, we found that posi-
tional gain of H3K4me2 is accompanied by a corresponding
depletion of H3K27me3 (Figure 3C, Student’s t test p < 0.01).
Remarkably, this H3K27me3 reduction is present only within
the punctate boundaries of a sharply gained H3K4me2 peak
and does not spread to the surrounding regions, which retain
somatic levels of facultative, inhibitory heterochromatin as in
the starting state.
We also generated genome-wide DNA methylation data from
the 0, 1, and >3 division populations and compared them to
control and ESC promoters. As expected, themajority of regions
exhibiting dramatic H3K4me2 gain displayed promoter hypome-
thylation in all states (Figure 3B). Moreover, promoters with
the most dramatic shifts in chromatin state generally exhibit
higher CpG density and preferentially enrich for CpG islands
(82%, Fisher’s exact test p < 10�33). DNA methylation data
confirmed that these regions were consistently hypomethylated
across populations, including in the starting fibroblast state, an
expected epigenetic landscape that is generally characteristic
of CpG islands. Additionally, it is interesting to note that regions
with depletion of H3K4me2 were frequently associated with
transcriptional repression and a vast majority (95%, Fisher’s
exact test p < 10�41) corresponded to non-CpG island
promoters at which H3K4 methylation status is often predictive
of transcriptional activity. Taken together, these data suggest
that the plasticity of somatic chromatin to changes by reprog-
ramming factors is most amenable within certain boundaries in
part governed by genetic determinants, such as CpG density
and the targeting sequences for the reprogramming factors
themselves.
Enhancer Signatures Are Driven from a Somatictoward an ESC-like StateThe activity of reprogramming factors on target chromatin is not
restricted to the promoter regions and operates similarly within
intergenic regions (Figure 4A; Figure S4A). Nonpromoter inter-
vals enriched for H3K4me2 have been correlated to functional
enhancers genome-wide, the patterns of which are remarkably
variable across cell type and have been used as a high informa-
tion content signature of a given cell state (Heintzman et al.,
2007). We thus reasoned that nonpromoter H3K4me2 elements
that differ betweenMEFs and iPSCs could provide further insight
into the early dynamics of reprogramming. Unlike promoter
elements, which predominantly gain H3K4me2, epigenetic
signatures of enhancers are gained and lost as reprogramming
populations shift away from the somatic state (Figure 4B).
Moreover, enhancer dynamics are shifted rapidly; a majority of
intergenic H3K4me2 dynamics occur on or before a single cell
division (54% gained, 66% lost) and progress continuously
with division number (Figure S4B). Of the 11,228 H3K4me2
enhancers identified in the reprogramming populations, 46%
are shared with ESCs and 8,407 somatic exclusive enhancer
Cell Stem Cell 8, 96–105, January 7, 2011 ª2011 Elsevier Inc. 99
A
E F
B
C D
MEF 0 Div 1 Div 2 Div >3 Div ES MEF 0 Div 1 Div 2 Div >3 Div ES
05
1015
2025
H3K
4me2
sig
nal
020
4060
80H
3K4m
e2 s
igna
l
De novo H3K4me2 gain(n≈300 promoters)
enh. H3K4me2 gain(n≈1000 promoters)
20 kb
MEF
0 Div
>3 Div
2 Div
1 Div
ESC
Postn
28
28
28
28
28
28
H3K
4me2
MEF
0 Div
>3 Div
2 Div
1 Div
ESC
15
15
15
15
15
15
H3K
4me2
5 Kb
Aire
Klf4
Oct4
Sox2
Myc
(339)
(139)
(115)
(306)
0
4
8
12
16
20
TSS5Kb-5Kb
TSS5Kb-5Kb
TSS5Kb-5Kb
De novo H3K4me2 gain by >3 Div (n≈300)MEF control >3 Div ES
H3K4me2 gain by >3 Div (n≈1000)MEF control >3 Div ES
0
10
20
30
40
ChI
P-s
eq s
igna
lC
hIP
-seq
sig
nal H3K4me1
H3K4me2H3K4me3
TSS5Kb-5Kb
TSS5Kb-5Kb
TSS5Kb-5Kb
H3K4me1H3K4me2H3K4me3
Figure 2. H3K4 Dimethylation Increases at Pluripotency-Related Genes and Is Lost in Repressed Somatic Targets
(A) De novo H3K4me2 acquisition is continuous across cohorts and already visible before a single division (nz 300). Red line indicates median. Whiskers repre-
sent 2.5 and 97.5 percentile.
(B) Enhanced H3K4me2 at a subset of �1000 promoters over proliferative cohorts exhibit similar trends and approach expected ESC levels in dividing popula-
tions of reprogramming cells. Red line indicates median. Whiskers represent 2.5 and 97.5 percentile.
(C) ChIP-Seq tracks showing de novo H3K4me2 at the endogenous promoter of Aire as part of an orchestrated enrichment that is preferential for Oct4- and Sox2-
regulated promoters. Green bars on the bottom indicate CpG islands. Gray bar highlights the putative nucleosome-depleted region that is flanked by H3K4me2
within ESCs.
(D) H3K4me2 ChiP-seq map of the Postn locus, which is expressed in MEFs and silenced by >3 divisions, shows a loss of H3K4me2 levels at its promoter region
to ESC-like levels. The Postn locus represents 115 promoters for which H3K4me2 is lost during reprogramming factor induction.
(E) ESC transcription factor occupancy of genes demonstrating H3K4me2 enrichment show a predominance of Oct4 and Sox2 binding.
(F) Composite plots of H3K4 mono-, di-, and trimethylation distribution at de novo and enhanced promoter classes in control MEFs, after three divisions, and
within ESCs.
Cell Stem Cell
Targeted Chromatin Remodeling during Reprogramming
100 Cell Stem Cell 8, 96–105, January 7, 2011 ª2011 Elsevier Inc.
A C
B
10 kb12
12
12
12
12
12
12
12
12
12
12
12
Sall4
ESC
MEF
0 Div
>3 Div
2 Div
1 Div
CpG islands
0
5
10
15
20
25
30
02468
101214
H3K
4me2
H3K
27m
e3
MEF 0 Div 1 Div 2 Div >3 Div ES
0
0.2
0.4
0.6
Bivalent genes in ESCs
DN
Am
e
%GC
−0.2
−0.1
0
0.1
0.2
H3K
4me2
-K27
me3
Pea
rson
cor
rela
tion
TSS
5Kb-5Kb
0
10
20
Read D
ensity
H3K4me2
H3K27me3
5Kb-5Kb 5Kb-5Kb 5Kb-5Kb 5Kb-5Kb 5Kb-5Kb 5Kb-5Kb
H3K4me2
H3K27me3
NA
25
15
5
Figure 3. Chromatin Remodeling and Genetic Determinants Define the Early Reprogramming Phase
(A) The Sall4 locus exhibits a de novo gain of H3K4methylation at twoCpG islands (green bars). Gain of H3K4me2corresponds to a targeteddepletion of H3K27meth-
ylation within cycling cells that is limited to the site of H3K4methylation. Highlighted region displays theCpG island and the site of ESC-specific nucleosome depletion.
(B) General trends of epigenetic reprogramming events at ESC bivalent promoters (n = 688) within induced populations. Top: Composite plots of H3K4me2 gain within
ESCbivalent promoters compared against somatic andESCcontrols.Middle: Composite plot ofH3K27me3 levels stay constant except in themost proliferative cohort
(>3 divisions) where levels are inversely proportional to the gain in H3K4me2 and are subsequently depleted. Bottom: CpGmethylation values at regions of enhanced
H3K4me2 gain are predominantly hypomethylated across states as expected given the high CpG density of this promoter set (82%CpG islands). CpG density across
the promoters analyzed is highlighted and demonstrates the boundary of the dynamic changes in chromatin state. Scale ranges between 40% (white) and 80% (black)
GC content.
(C) Pearson correlation between H3K4me2 and H3K27me3 levels in 200 base pair sliding windows. Negative correlation between the two marks reaches significance
within 500 bp from the TSS. Histone mark enrichments for the promoter set are included as heat maps and emphasize this inverse relationship.
Cell Stem Cell
Targeted Chromatin Remodeling during Reprogramming
Cell Stem Cell 8, 96–105, January 7, 2011 ª2011 Elsevier Inc. 101
A B
C
D
3708
36661404
2450
59178958407MEF
>3 Div
ESC
0
2
4
6
0.2
0.4
0.6
0.8
0.2
0.4
0.6
0.8ESC
H3K
4me2
DN
A m
ethy
latio
n (f
ract
ion)
0
2
4
6ESC-exclusive enhancers
-5Kb 5Kb -5Kb 5Kb
MEF>3 DivESC
MEF>3 DivESC
Acquired ESC enhancers
12000 8000 4000 0 2000 4000 6000
MEF-specific ES-specificMEF0 Div1 Div2 Div>3 DivESC
Enhancer Shift
Number of enhancers marked by H3K4me2
0
2
4
6
8
All enhancers acquired by >3 Div (n≈3800)
H3K4me1H3K4me2H3K4me3
ChI
P-S
eq s
igna
l-5Kb 5Kb -5Kb 5Kb -5Kb 5Kb
MEF control 1 Div >3 Div
St14CpG island
H3K4me3
H3K4me2
H3K4me1
H3K27me3
H3K36me3
H3K4me3
H3K4me2
H3K4me1
H3K27me3
H3K36me3
H3K4me3
H3K4me2
H3K4me1
H3K27me3
H3K36me3
H3K4me3
H3K4me2
H3K4me1
H3K27me3
H3K36me3
20 kb E EP
Partial Gain of ES Chromatin States
ES
>3 Div
1 Div
MEF
MEFESCMEF
Figure 4. Global Epigenetic Dynamics during the Early Stage of Reprogramming Factor Induction Extends beyond Target Promoter Regions
to Putative Enhancers
(A) The CpG island promoter (P) (pink highlight) of the ESC-expressed St14 gene displays minimal H3K4 methylation in the somatic state and increases in
H3K4me2 with proliferation, concurrent with punctate loss of H3K27me3 at the CpG island (see also Figure 3A). The de novo K4me2 gain is accompanied by
gain of an intronic enhancer signature (E) (pink highlight). Expression levels for St14 are not detected until complete remodeling at later stages. Intergenic
enhancers (E) (pink highlight, right) are also gained and are progressively enriched for H3K4me1 and me2.
(B) Number of MEF-exclusive or ESC-exclusive putative enhancers that are gained or lost across division. The ‘‘ESC-specific’’ enhancer set does not include the
3708 enhancers that are shared between MEF, ESCs, and all reprogramming populations. Inset: Venn diagram of represented enhancers within reprogramming
cells against the starting somatic state and ESCs.
(C) Architecture and relationship of H3K4 methylation marks gained at newly acquired enhancer signatures called after >3 divisions as in (B). Enhancers gain
significant H3K4me1 in early proliferative cohorts followed by subsequent H3K4me2 enrichment.
(D) Composite plot of ESC H3K4me2 enhancer peaks gained in reprogramming populations demonstrate an equivalent CpG hypomethylation in somatic stem
cells and ESCs. Alternatively, ESC-specific enhancers that are not acquired after 96 hr of factor induction demonstrate differential and higher mean CpG meth-
ylation. Dashed lines highlight somatic CpG methylation in the acquired versus ESC-exclusive sets.
Cell Stem Cell
Targeted Chromatin Remodeling during Reprogramming
regions are depleted (Figure 4B). Intergenic analysis of additional
H3K4 methylation marks confirm the canonical architecture of
enhancer elements, with strong overlap of H3K4me1 and
H3K4me2 and relative lack of promoter-exclusive H3K4me3
(Figure 4C). Moreover, reprogramming induced enhancer signa-
tures appear to acquire stable H3K4 methylation sequentially,
first gaining H3K4me1 (Figure 4C, middle) followed by
H3K4me2 (Figure 4C, right). From this context, examination of
the epigenetic changes within intergenic regions provide
a unique opportunity to model enhancer dynamics; moreover,
102 Cell Stem Cell 8, 96–105, January 7, 2011 ª2011 Elsevier Inc.
genome-wide characterization of H3K4me2 confirms its value
as a highly informative epigenetic mark, being present in dispa-
rate promoter and intergenic contexts where H3K4me1 or
H3K4me3 are mutually exclusive (Figure S4D). Intergenic shifts
in H3K4me2 enrichment thus serve as a unique barcode for
cellular identity and sensitively measure the epigenetic changes
caused by reprogramming factor induction.
We incorporated genome-scale DNA methylation maps of
ESCs and MEFs (Meissner et al., 2008) with those generated
for our induced populations for use in our analysis of intergenic
Cell Stem Cell
Targeted Chromatin Remodeling during Reprogramming
H3K4me2. Genomic intervals that display rapid gain of
H3K4me2 tended to exhibit relatively lower DNA methylation
levels in MEFs (Figure 4D, left). In contrast, ESC enhancer
elements that are not activated after 96 hr of factor induction
have significantly higher DNA methylation levels in MEFs
(Figure 4D, right, Student’s t test p < 10�32). Interestingly, the
MEF-exclusive enhancers that are lost during reprogramming
display complete hypermethylation within ESCs, but not within
induced populations (Figure S4C). This suggests that ESC-like
DNA methylation patterns are not fully established until later
stages of reprogramming. The failure to re-establish DNA meth-
ylation at somatic intergenic H3K4me2 enhancers may, in part,
account for the instability/elasticity of reprogramming popula-
tions, which may traverse back toward a fibroblast-like state
upon premature removal of ectopic factor expression (Sama-
varchi-Tehrani et al., 2010).
The sensitivity of H3K4me2 enhancement to DNA methylation
is consistent with a model where DNA methylation and associ-
ated repressive chromatin structures limit the accessibility of
these elements to nuclear reprogramming (Mikkelsen et al.,
2008). Newly activated enhancers that are covered by
genome-scale CpG methylation assays exhibit lower methyla-
tion levels at the site of H3K4me2 gain and are generally hypo-
methylated in starting fibroblasts (Figure 4D). These data corrob-
orate changes in promoter histonemethylation, where H3K4me2
gain is restricted to sites of high CpG density, which are generally
hypomethylated (Meissner et al., 2008) and uniquely amenable
to rapid epigenetic reconfiguration (Xu et al., 2009).
DISCUSSION
To further advance our understanding of the transcription factor-
mediated reprogramming process, we isolated clonally induced
cells that had undergone defined cell divisions for genomic
characterization. Our data demonstrate a robust trend within
the early reprogramming population toward a primed epigenetic
state that clearly precedes transcriptional activation and
complete reprogramming. In addition to suggesting an early
coordinated response, our data highlight transcriptional
measurement as an incomplete descriptor of the cellular
response to reprogramming factor induction. Importantly, gain
of H3K4 methylation includes a broader array of notable targets
such as key pluripotency and early development genes. As we
report, these are particularly enriched for CpG island-containing
promoters. Moreover, at sites where H3K4me2 is dynamic,
somatic heterochromatin (marked by H3K27me3) is depleted
exclusively within the CpG island context but continues to be
present in the periphery. Re-establishment of H3K27me3 at
bivalent promoters is not observed and must pertain to a later
phase of iPSC generation (Pereira et al., 2010).
Our results provide a sensitive measurement of the somatic
response to transcription factor activity, which displays a greater
trend toward promoter-associated H3K4 methylated euchro-
matin and may represent a critical step toward transcriptional
activation. The continuous behavior of this trend as populations
divide clearly demonstrates unique underlying activity that is
likely to utilize the endogenous epigenetic machinery. The unex-
pected genome-wide extent of these events appears mostly
limited by sequence context and is most likely to occur within
CpG islands in which reprogramming factor regulatory motifs
are present. The scope through which promoters and enhancers
aremodified supports a deterministic model for the initial reprog-
ramming response, because the global events are at expected
targets and occur at a detectable frequency similar to what is
observed within pluripotent populations. This is further consis-
tent with more recent image-based data (Smith et al., 2010)
and provides an interpretation for the epigenetic response to
factor induction, inwhichgenome-wide remodelingoccurswithin
the majority of cells in the induced population, as opposed to
selectively within an exclusive subpopulation that will contribute
iPSCprogeny (Yamanaka, 2009). The immediate andprogressive
accumulation of euchromatin-associated marks at ESC-specific
promoters and enhancers suggests that a detectable majority of
cells in which the factors are induced undergo a certain level of
epigenetic reprogramming even in the absence of cell division;
these events are immeasurable by expression profiling alone
and have to date been largely overlooked.
Moreover, because these events precede detectable tran-
scription, it is likely that the chromatin dynamics observed at
the endogenous loci are a critical initial step in the transition to
molecular pluripotency. It is intriguing that the promoter
dynamics observed are initially restricted to areas of high CpG
density and especially CpG islands, whereas peripheral chro-
matin retains its original, somatic pattern. CpG islands are noted
for their plasticity and responsiveness to transcription factor
activity (Ramirez-Carrozzi et al., 2009). The periphery of these
regions behave inversely—they are less CpG rich and more
susceptible to DNA methylation and/or extended H3K27me3
spreading, marks that may stably maintain heterochromatin
domains in restricted cell types and may require transcriptional
activation to be completely depleted. Notably, it is in these
regions where somatic epigenetic artifacts might be observed
in iPSC characterization studies and a likely explanation could
be that these regions are generally less responsive to chromatin
remodeling. In our model, the type of mark, the developmental
history of its acquisition, and its distribution along target
promoter elements all contribute to the response observed.
At CpG-dense, hypomethylated transcription start sites, factor
expression is sufficient to induce the rapid redistribution of
H3K4me2 marks at the promoter that may signal or prime that
locus for transcriptional activation. This principle is recapitulated
at enhancer sites, where H3K4me2 gain is restricted to somati-
cally hypomethylated regions. As discussed earlier, factor induc-
tion alone is not sufficient for complete reprogramming. Instead,
the process probably depends on the presence of further chro-
matin remodeling complexes or transcriptional recruitment
elements that may be unavailable in somatic cells.
In conclusion, our data argue for an orchestrated response
that yields an epigenetically definable intermediate state in the
earliest stages of the reprogramming timeline. However, it
cannot as of yet be ascertained if the continuation to full pluripo-
tency is predetermined by existing effectors within a select
subpopulation or by stochastic activation of these players in
iPSC-forming lineages. It is also likely that these epigenetic re-
programming events describe the limiting effect of the four
factors (OSKM) themselves as they act within a population
where only a select subset will progress to endogenous target
activation; transition through this phase toward complete
Cell Stem Cell 8, 96–105, January 7, 2011 ª2011 Elsevier Inc. 103
Cell Stem Cell
Targeted Chromatin Remodeling during Reprogramming
reprogramming probably involves additional factors. Regard-
less, continued dissection of the reprogramming process prom-
ises for a comprehensive identification of a sufficient factor set
for complete and safe somatic to pluripotent reprogramming.
EXPERIMENTAL PROCEDURES
CFSE Labeling and Enrichment for Proliferative Cohorts
Mouse E13.5 fibroblasts were generated by blastocyst injection with doxycy-
cline-inducible Oct4, Sox2, Klf4, and c-Myc primary iPSCs as previously
described. Cells were passaged several times and serum starved with 0.5%
FBS-containing medium for 18 hr before CFSE labeling. Cells were labeled
with CFSE in 5 3 106 cell batches with 5 mM cellTrace CFSE (Invitrogen) in
PBS according to the manufacturer’s protocol and plated at 1 3 106 cells per
10 cm dish in 0.5% FBS for an additional 12 hr before the induction of OSKM-
reprogramming factors. Factors were induced with 2 mg/ml doxycycline-
supplementedmedium in either 0.5%or 15%FBS to control the relative number
of proliferation for 96 hr (see Figure 1A). In brief: our ‘‘no division’’ cohort was
culturedexclusively in0.5%FBS-containingmediumandeachsuccessiveprolif-
erative cohort was cultured in 15% FBS-containing medium containing doxycy-
clinemedium for 24 hr, 48 hr, and 96 hr. After serum pulsing, cellswere switched
back into 0.5%FBSmedium toquell further division; all sampleswere cultured in
doxycycline-supplementedmedium for the entire 96 hr. The relative proliferative
number for each cohort was ascertained with a BD LSR II fluorescent cytometer
against an uninduced, serum-starved control. RNA was collected with TRIzol
(Invitrogen) and cells were crosslinked with 1% formaldehyde.
ChIP-seq Library Preparation and RRBS
Generation of genome-wide sequencing libraries were performed with
�500,000 crosslinked samples as available input for a given antibody targeting
a covalent histonemodification. Sample sonication, chromatin immunoprecip-
itation, and library generation were performed as described (Mikkelsen et al.,
2007). RRBS libraries were generated on standardized 100 ng of genomic DNA
isolated by proteinase K digestion and phenol:chloroform extraction in accor-
dance with previously published methods (Gu et al., 2010). A refined protocol
with available antibodies and lot numbers used in this document are available
as Supplemental Information.
Analysis
Gene expression profiles were acquired with Affymetrix Mouse Genome 430
2.0 Arrays and Robust Multi-Array (RMA)-normalized with GenePattern
(http://www.broadinstitute.org/cancer/software/genepattern/). ChIP libraries
were sequencedwith the Illumina Genome Analyzer andmapped to themouse
mm8 genome as previously described (Mikkelsen et al., 2007). Description of
enrichment calculations, statistical analyses, and normalizations are available
as Supplemental Information. OSKM factor enrichment was performed with
previously published data and analysis (Kim et al., 2008; Marson et al., 2008).
ACCESSION NUMBERS
The data sets are available in the Gene Expression Omnibus (GEO) database
(http://www.ncbi.nlm.nih.gov/gds) under the accession number GSE26100.
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental Procedures,
four figures, and three tables and can be found with this article online at
doi:10.1016/j.stem.2010.12.001.
ACKNOWLEDGMENTS
We would like to thank Tarjei Mikkelsen for critical reading of the manuscript.
We would like to apologize to authors whose primary work we didn’t cite
because of space restrictions. B.E.B. is an early career scientist of the
HHMI. A.M. is a New Investigator of the Massachusetts Life Science Center
(MLSC) and Pew Scholar. This work was funded by the MLSC and Pew Char-
itable Trusts.
104 Cell Stem Cell 8, 96–105, January 7, 2011 ª2011 Elsevier Inc.
Received: July 2, 2010
Revised: October 22, 2010
Accepted: November 24, 2010
Published: January 6, 2011
REFERENCES
Bernstein, B.E., Kamal, M., Lindblad-Toh, K., Bekiranov, S., Bailey, D.K.,
Huebert, D.J., McMahon, S., Karlsson, E.K., Kulbokas, E.J., 3rd, Gingeras,
T.R., et al. (2005). Genomic maps and comparative analysis of histone modifi-
cations in human and mouse. Cell 120, 169–181.
Bernstein, B.E., Mikkelsen, T.S., Xie, X., Kamal, M., Huebert, D.J., Cuff, J., Fry,
B., Meissner, A., Wernig, M., Plath, K., et al. (2006). A bivalent chromatin struc-
ture marks key developmental genes in embryonic stem cells. Cell 125,
315–326.
Graf, T., and Enver, T. (2009). Forcing cells to change lineages. Nature 462,
587–594.
Gu, H., Bock, C., Mikkelsen, T.S., Jager, N., Smith, Z.D., Tomazou, E., Gnirke,
A., Lander, E.S., and Meissner, A. (2010). Genome-scale DNA methylation
mapping of clinical samples at single-nucleotide resolution. Nat. Methods 7,
133–136.
Hanna, J., Saha, K., Pando, B., van Zon, J., Lengner, C.J., Creyghton, M.P.,
van Oudenaarden, A., and Jaenisch, R. (2009). Direct cell reprogramming is
a stochastic process amenable to acceleration. Nature 462, 595–601.
Hanna, J.H., Saha, K., and Jaenisch, R. (2010). Pluripotency and cellular
reprogramming: Facts, hypotheses, unresolved issues. Cell 143, 508–525.
Heintzman, N.D., Stuart, R.K., Hon, G., Fu, Y., Ching, C.W., Hawkins, R.D.,
Barrera, L.O., Van Calcar, S., Qu, C., Ching, K.A., et al. (2007). Distinct and
predictive chromatin signatures of transcriptional promoters and enhancers
in the human genome. Nat. Genet. 39, 311–318.
Jaenisch, R., and Young, R. (2008). Stem cells, the molecular circuitry of plu-
ripotency and nuclear reprogramming. Cell 132, 567–582.
Kim, J., Chu, J., Shen, X., Wang, J., and Orkin, S.H. (2008). An extended tran-
scriptional network for pluripotency of embryonic stem cells. Cell 132, 1049–
1061.
Lee, T.I., Jenner, R., Boyer, L., Guenther, M., Levine, S., Kumar, R., Chevalier,
B., Johnstone, S., Cole, M., Isono, K., et al. (2006). Control of developmental
regulators by polycomb in human embryonic stem cells. Cell 125, 301–313.
Marson, A., Levine, S.S., Cole, M.F., Frampton, G.M., Brambrink, T.,
Johnstone, S., Guenther, M.G., Johnston, W.K., Wernig, M., Newman, J.,
et al. (2008). ConnectingmicroRNA genes to the core transcriptional regulatory
circuitry of embryonic stem cells. Cell 134, 521–533.
Meissner, A., Wernig, M., and Jaenisch, R. (2007). Direct reprogramming
of genetically unmodified fibroblasts into pluripotent stem cells. Nat.
Biotechnol. 25, 1177–1181.
Meissner, A., Mikkelsen, T.S., Gu, H., Wernig, M., Hanna, J., Sivachenko, A.,
Zhang, X., Bernstein, B.E., Nusbaum, C., Jaffe, D.B., et al. (2008). Genome-
scale DNA methylation maps of pluripotent and differentiated cells. Nature
454, 766–770.
Mikkelsen, T.S., Ku, M., Jaffe, D.B., Issac, B., Lieberman, E., Giannoukos, G.,
Alvarez, P., Brockman, W., Kim, T.K., Koche, R.P., et al. (2007). Genome-wide
maps of chromatin state in pluripotent and lineage-committed cells. Nature
448, 553–560.
Mikkelsen, T.S., Hanna, J., Zhang, X., Ku, M., Wernig, M., Schorderet, P.,
Bernstein, B.E., Jaenisch, R., Lander, E.S., and Meissner, A. (2008).
Dissecting direct reprogramming through integrative genomic analysis.
Nature 454, 49–55.
Pereira, C.F., Piccolo, F.M., Tsubouchi, T., Sauer, S., Ryan, N.K., Bruno, L.,
Landeira, D., Santos, J., Banito, A., Gil, J., et al. (2010). ESCs require PRC2
to direct the successful reprogramming of differentiated cells toward pluripo-
tency. Cell Stem Cell 6, 547–556.
Rahl, P.B., Lin, C.Y., Seila, A.C., Flynn, R.A., McCuine, S., Burge, C.B., Sharp,
P.A., and Young, R.A. (2010). c-Myc regulates transcriptional pause release.
Cell 141, 432–445.
Cell Stem Cell
Targeted Chromatin Remodeling during Reprogramming
Ramirez-Carrozzi, V.R., Braas, D., Bhatt, D.M., Cheng, C.S., Hong, C., Doty,
K.R., Black, J.C., Hoffmann, A., Carey, M., and Smale, S.T. (2009). A unifying
model for the selective regulation of inducible transcription by CpG islands and
nucleosome remodeling. Cell 138, 114–128.
Samavarchi-Tehrani, P., Golipour, A., David, L., Sung, H., Beyer, T., Datti, A.,
Woltjen, K., Nagy, A., and Wrana, J. (2010). Functional genomics reveals
a BMP-driven mesenchymal-to-epithelial transition in the initiation of somatic
cell reprogramming. Cell Stem Cell 7, 64–77.
Smith, Z.D., Nachman, I., Regev, A., and Meissner, A. (2010). Dynamic single-
cell imaging of direct Reprogramming reveals an early specifying event. Nat.
Biotechnol. 28, 521–526.
Wernig, M., Lengner, C.J., Hanna, J., Lodato, M.A., Steine, E., Foreman, R.,
Staerk, J., Markoulaki, S., and Jaenisch, R. (2008). A drug-inducible transgenic
system for direct reprogramming of multiple somatic cell types. Nat.
Biotechnol. 26, 916–924.
Xu, J., Watts, J.A., Pope, S.D., Gadue, P., Kamps, M., Plath, K., Zaret, K.S.,
and Smale, S.T. (2009). Transcriptional competence and the active marking
of tissue-specific enhancers by defined transcription factors in embryonic
and induced pluripotent stem cells. Genes Dev. 23, 2824–2838.
Yamanaka, S. (2009). Elite and stochastic models for induced pluripotent stem
cell generation. Nature 460, 49–52.
Cell Stem Cell 8, 96–105, January 7, 2011 ª2011 Elsevier Inc. 105
Cell Stem Cell
Resource
Dynamic Changes in the Copy Number of Pluripotencyand Cell Proliferation Genes in Human ESCsand iPSCs during Reprogramming and Time in CultureLouise C. Laurent,1,3,4,* Igor Ulitsky,6,7 Ileana Slavin,3,4 Ha Tran,3,4 Andrew Schork,2 Robert Morey,1,3,4 Candace Lynch,3,4
Julie V. Harness,8 Sunray Lee,9 Maria J. Barrero,10,11 Sherman Ku,5 Marina Martynova,12 Ruslan Semechkin,12
Vasiliy Galat,13,14 Joel Gottesfeld,5 Juan Carlos Izpisua Belmonte,10,11 Chuck Murry,15 Hans S. Keirstead,8
Hyun-Sook Park,9 Uli Schmidt,16 Andrew L. Laslett,17,18,19 Franz-Josef Muller,3,4 Caroline M. Nievergelt,2 Ron Shamir,7
and Jeanne F. Loring3,4
1Department of Reproductive Medicine2Department of Psychiatry
University of California, San Diego, La Jolla, CA 92093, USA3Department of Chemical Physiology4Center for Regenerative Medicine5Department of Molecular Biology
The Scripps Research Institute, La Jolla, CA 92037, USA6The Whitehead Institute, Cambridge, MA 02142, USA7Department of Computer Science, Tel Aviv University, Tel Aviv 69978, Israel8Department of Anatomy and Neurobiology, Sue and Bill Gross Stem Cell Center, University of California, Irvine, Irvine, CA 92697, USA9Modern Cell &Tissue Technologies (MCTT) Inc., Seoul 139-240, South Korea10The Salk Institute for Biological Studies, La Jolla, CA 92037, USA11Centro de Medicina Regenerativa de Barcelona, Barcelona E-08003, Spain12International Stem Cell Corporation, Oceanside, CA 92056, USA13Department of Pathology, Feinberg School of Medicine, Northwestern University, Chicago, IL 60611, USA14iPS and Human Stem Cell Core Facility, Northwestern University Children’s Memorial Research Center, Chicago, IL 60614, USA15Department of Pathology, University of Washington, Seattle, WA 98195, USA16Stem Cell Laboratory, Sydney IVF, Sydney, New South Wales 2000, Australia17Commonwealth Scientific and Industrial Research Organisation (CSIRO), Clayton, Victoria 3168, Australia18Australian Stem Cell Centre, Clayton, Victoria 3168, Australia19Department of Anatomy and Developmental Biology, Monash University, Clayton, Victoria 3168, Australia
*Correspondence: llaurent@ucsd.eduDOI 10.1016/j.stem.2010.12.003
SUMMARY
Genomic stability is critical for the clinical use ofhuman embryonic and induced pluripotent stemcells. We performed high-resolution SNP (single-nucleotide polymorphism) analysis on 186 pluripo-tent and 119 nonpluripotent samples. We reporta higher frequency of subchromosomal copy numbervariations in pluripotent samples compared tononpluripotent samples, with variations enrichedin specific genomic regions. The distribution ofthese variations differed between hESCs andhiPSCs, characterized by large numbers of duplica-tions found in a few hESC samples and moderatenumbers of deletions distributed across many hiPSCsamples. For hiPSCs, the reprogramming processwas associated with deletions of tumor-suppressorgenes, whereas time in culture was associated withduplications of oncogenic genes. We also observedduplications that arose during a differentiationprotocol. Our results illustrate the dynamic natureof genomic abnormalities in pluripotent stem cells
106 Cell Stem Cell 8, 106–118, January 7, 2011 ª2011 Elsevier Inc.
and the need for frequent genomic monitoring toassure phenotypic stability and clinical safety.
INTRODUCTION
The tremendous self-renewal and differentiation capabilities
of human pluripotent stem cells (hPSCs) make them potential
sources of differentiated cells for cell therapy. Cell therapies
are subject to rigorous safety trials, and high priority is placed
on demonstrating that the cells are nontumorigenic (Fox,
2008). Because genetic aberrations have been strongly associ-
ated with cancers, it is important that preparations destined for
clinical use are free from cancer-associated genomic alterations.
Human embryonic stem cell (hESC) lines have been shown to
become aneuploid in culture (Baker et al., 2007; Draper et al.,
2004; Imreh et al., 2006; Maitra et al., 2005; Mitalipova et al.,
2005), and the most frequent changes, trisomies of chromo-
somes 12 and 17, are also characteristic of malignant germ
cell tumors (Atkin and Baker, 1982; Rodriguez et al., 1993; Sko-
theim et al., 2002). Aneuploidies can be detected by karyotyping,
but less easily detectable subchromosomal genetic changes
may also have adverse effects. Small abnormalities have been
detected in hESCs by using comparative genomic hybridization
Cell Stem Cell
Genomic Instability of Human Pluripotent Cells
(CGH) and single-nucleotide polymorphism (SNP) genotyping
(Lefort et al., 2008; Narva et al., 2010; Spits et al., 2008). These
studies lacked sufficient resolution and power to identify cell
type-associated duplications and deletions. A recent study has
reported the use of gene expression data to detect genomic
aberrations in a large number of hESCs and hiPSCs (Mayshar
et al., 2010). However, the methods used could reliably detect
only relatively large (R10 megabase) aberrations, and the lack
of nonpluripotent samples for comparison precluded the authors
from determining which regions of genomic aberration were
specific to pluripotent stem cells.
In this study, we performed high-resolution SNP genotyping
on a large number of hESC lines, induced human pluripotent
stem cell lines (hiPSCs), somatic stem cells, primary cells, and
tissues. We found that hESC lines had a higher frequency of
genomic aberrations compared to the other cell types. Further-
more, we identified regions in the genome that had a greater
tendency to be aberrant in the hESCs when compared to the
other cell types examined. Recurrent regions of duplication
were seen on chromosome 12, encompassing the pluripo-
tency-associated transcription factor NANOG and a nearby
NANOG pseudogene, and on chromosome 20, upstream of
the DNA methyltransferase DNMT3B. Although the frequency
of genomic aberrations seen in the hiPSC lines was similar to
those of cultured somatic cells and tissues, we observed one
of the recurrent areas of duplication characteristic of hESCs in
one of the hiPSC lines.
Furthermore, comparison of 12 hiPSC lines generated from
the same primary fibroblast cell line identified genomic aberra-
tions that were present in the hiPSC lines and absent from the
original fibroblast line. Analysis of early- and late-passage
samples from these hiPSC lines allowed us to distinguish
between events that arose during the process of reprogramming
and those that accumulated during long-term passage. In
general, deletions tended to occur with reprogramming and
involve tumor-suppressor genes, whereas duplications accumu-
lated with passaging and tended to encompass tumor-
promoting genes. These results suggest that human pluripotent
stem cell populations are prone to genomic aberrations that
could compromise their stability and utility for clinical applica-
tions and that reprogramming and expansion in culture may
lead to selection for particular genomic changes.
RESULTS
High-resolution SNP genotyping (1,140,419 SNPs) was per-
formed on 324 samples, including 69 hESC lines (130 samples),
37 hiPSC lines (56 samples), 11 somatic stem cell lines
(11 samples), 41 primary cell lines (41 samples), and 20 tissue
types (67 samples), as well as samples of differentiated hESC
lines and mixtures of known ratios of a sample with a known
duplication with a sample without that duplication (Table S1
available online). Copy number variants for all samples were
identified in parallel with two algorithms, CNVPartition (Illumina,
Inc., Table S2A) and Nexus (Biodiscovery, Inc., Table S2B),
both of which have been demonstrated to be appropriate
for copy number variation (CNV) identification from SNP Geno-
typing data from Illumina microarrays (Kresse et al., 2010). The
concordance between these two algorithms was high (76.08%
C
for deletions, 98.60% for loss of heterozygosity (LOH), and
93.04% for duplications on the base-pair level) (Table S2C). A
subset of the CNV calls for both algorithms were validated via
qPCR. For the CNVPartition calls, 82% of 3-copy gains and
43% of 1-copy losses were confirmed. For Nexus, 68% of allelic
imbalance, 71%of copy number gain, 47%of copy number loss,
and 100% of loss of heterozygosity calls were confirmed
(Table S3, note that the allelic imbalance calls were judged to
be correct if the qRT-PCR result indicated either a significant
gain or a significant loss). Given the higher accuracy of the dupli-
cation calls in CNVPartition, and the ambiguity of the allelic
imbalance calls in Nexus, CNVPartition was subsequently used
as the primary algorithm. CNV calls that overlapped with
common CNVs observed in a reference set of 450 HapMap
samples (Conrad et al., 2010) were identified and removed
from subsequent analyses.
Figure 1 shows a map of the areas of CNV identified in all the
samples. Based on validation of the CNV calls by qRT-PCR,
which indicated that duplication calls were markedly more
accurate than deletion calls, we focused on duplications and
large deletions. We inspected the B-allele frequency (BAF) and
log R ratio (LRR) plots in order to combine adjacent areas of
CNV where appropriate; it is well appreciated that CNV calling
algorithms frequently break up large CNV events into multiple
calls. For example, the SIVF021 line was shown to have
a complete trisomy of chromosome 21 both by prenatal genetic
screening (PGS) of the embryo and karyotyping of the hESC line,
but CNVPartition and Nexus both call multiple noncontiguous
regions of CNV for this sample on chromosome 21 (Table S2).
A list of the regions mapped in Figure 1 is given in Table S4.
Large Regions of CNV in hESCs and hiPSCsSeveral hESC samples showed duplications of large regions: the
BG01 and BG01V samples both showed trisomy 12 and trisomy
17, but only the BG01 sample contained trisomy 3 and a deletion
of the long arm of chromosome 7. The MIZ13 sample also con-
tained trisomy 3. SIVF048 had a duplication of chromosome 5,
and the WA07P34MNP sample had a deletion of the same chro-
mosome (of note, this sample was from a directed differentiation
experiment from hESC to motor neuron progenitor). The FES29
sample had a duplication of the short arm, and a deletion of
the long arm, of chromosome 7. Large duplications of chromo-
somes 12, 17, and 20 were observed in multiple samples. A large
region of 2-copy LOH on chromosome 22 was identified for the
HFIB2IPS5 sample. In addition, large regions of 2-copy LOH
were identified on the X chromosome in several samples.
Because these samples were male, these calls corresponded
to duplications on the X chromosome; duplications of the entire
chromosome were identified for the BG01 hESC and the
TH1.60OCT4SOX2 hiPSC samples, and a large duplication of
the q-arm of the chromosome was found in the BG01V sample.
The aneuploidies in SIVF003 (chr16), SIVF011 (chr5), and
SIVF021 (chr21) were known prior to derivation from PGS. Aneu-
ploidies and large duplications of chromosomes 1, 12, 17, and X
have been previously reported to be common in hESCs (Baker
et al., 2007; Draper et al., 2004; Imreh et al., 2006; Mitalipova
et al., 2005).
In a recent publication (Narva et al., 2010), complex mosaic
aneuploidy was described in one of the lines we genotyped,
ell Stem Cell 8, 106–118, January 7, 2011 ª2011 Elsevier Inc. 107
Figure 1. Duplications and Large Deletions Identified by CNVPartition Mapped onto the Genome, for All Samples
The number and extent of regions of CNV regions are shown. Duplicated regions (3 or 4 copies) are shown in the dark bars, deleted regions (0 or 1 copy) are shown
in the light bars, and copy-neutral LOH regions are placed on the ideograms of the chromosomes. Where five or more samples of the same cell type have aber-
rations at the same region, the number of samples affected is indicated (e.g.,35,310). Regions for hESC samples are shown in red, regions for hiPSC samples
are shown in blue, and regions for non-PSC samples are shown in green. Some aneuploidies had been identified prior to hESC derivation and are indicated as
‘‘known from PGS.’’ Regions where the CNV is present in only a subpopulation of the cells in a sample are denoted ‘‘(sub).’’ The three regions of duplication on
chromosome 20 that arose in a subpopulation of the cells during differentiation of theWA07P96CMD7 sample are indicated. CNVs that overlap with the common
CNVs observed in 450 HapMap samples (Conrad et al., 2010) are indicated by an asterisk. See also Figure S1 and Tables S1–S4.
Cell Stem Cell
Genomic Instability of Human Pluripotent Cells
FES61. In our analysis, the B-allele frequency pattern from the
SNP genotyping data indicated that this line contained genetic
material from three male individuals (Figure S1), which makes
the data from this line uninterpretable for CNV analysis. We
therefore excluded this line from further analysis.
Recurrent Regions of CNV in hESCs and hiPSCsIn addition to these large duplications and deletions, we
observed multiple smaller regions of CNV, including both dele-
tions and duplications, which we examined to identify regions
108 Cell Stem Cell 8, 106–118, January 7, 2011 ª2011 Elsevier Inc.
of recurrent CNV in the human pluripotent stem cell samples.
As noted above, the validation rate for small duplications was
significantly higher than for small deletions, and therefore we
focused on duplications for our analyses. We ensured that the
recurrent regions identified were associated with the pluripotent
state rather than with high-frequency CNVs found in the human
population by comparing the CNVs found in the hPSC samples
with those found in the non-PSC samples, as well as a data
set identifying common CNVs via 450 HapMap samples (Fig-
ure 1; Table S2; Conrad et al., 2010).
Cell Stem Cell
Genomic Instability of Human Pluripotent Cells
In order to identify regions of recurrent duplication, we identi-
fied regions that were duplicated in multiple samples. Analyzing
all samples, and with Fisher’s exact test with a p value cutoff
of 0.05, yielded 152 regions where the duplications were distrib-
uted at a statistically significantly different rate between pluripo-
tent and nonpluripotent samples (Table S5). We then filtered for
regions where the fraction of pluripotent samples was >90%,
which yielded 18 regions. The two duplicated segments that fit
these criteria were located on chromosome 12 and chromosome
20 and are highlighted in Figure 2. The chromosome 12 region
was duplicated in 9 out of 69 hESC lines, with the smallest
common duplicated region encompassing NANOGP1 and
SLC2A3 (Figure 2A). NANOG itself is upstream of NANOGP1
and was duplicated in five lines. The chromosome 20 region
was identified in 7 out of 69 hESC lines and 1 out of 37 hiPSC
lines. In our manual curation of the data, we identified duplica-
tions of this region in two additional samples that CNVPartition
failed to detect. For one (WA07P96CMD7), the population was
mosaic and for the other (BG01P67), CNVPartition called dupli-
cations of regions flanking the recurrently duplication region
but missed the region itself. Six of the duplications we mapped
included the DNMT3B gene itself (Figure 2B). In two recent
publications, recurrent duplications were described in the
20q11.21 region of chromosome 20 in hESCs; these reports indi-
cated that several hESC lines had duplications in a region near
the pluripotency-associated gene DNMT3B, which codes for
a de novo DNA methyltransferase (Lefort et al., 2008; Spits
et al., 2008). Mutations in this region of chromosome 20 have
been noted in a number of cancers, suggesting that genetic
elements in this regionmay be associatedwith hyperproliferation
(Guan et al., 1996; Hurst et al., 2004; Koynova et al., 2007; Mid-
orikawa et al., 2006; Scotto et al., 2008; Tanner et al., 1996;
Tonon et al., 2005). We also found that 5 out of 69 hESC lines
and 1 out of 37 hiPSC lines had duplications in this region.
The occurrence of duplications near (but not including) the
pluripotency-associated genes NANOG and DNMT3B suggests
that the duplication of other genes in these regions are being
selected for in the cultures, or that an upstream control element
for these genes may be present in the duplicated regions. In
several cases, the duplication event was observed in only one
of multiple samples from the same cell line collected at different
times. In some instances, a more ‘‘severe’’ aberration was
present in an earlier passage sample from the same lab (see
SIVF019P53 and SIVF019P67 in Figure 2B), again reinforcing
the need for detailed records regarding the passage history of
cultures.
Comparison of CNVs in hESCs, hiPSCs, and Non-PSCsFor comparisons of the relative number and length of CNVs
among hESCs, hiPSC, and non-PSCs, we decided to eliminate
possible bias resulting from having multiple samples of some
of the cell lines. For such cell lines, we included the one sample
that had the largest number of total CNVs in our analysis. In addi-
tion, we removed hESC lines where preimplantation genetic
diagnosis on the embryo had demonstrated that there was an
aneuploidy.
Although there was considerable variation in the number of
regions of CNV among the samples, overall the average numbers
of regions of duplication and deletion were significantly higher
C
in the hiPSCs compared to the non-PSCs (Figure 3). The distribu-
tion of genomic aberrations across the hiPSC samples was
rather even. In contrast, the distribution among hESC samples
was highly skewed, so that although the average number of
regions of duplication was not significantly higher in the hESCs
than in the non-PSCs, it was clear that a subset of hESC samples
contained a very large number of duplications (Figure 3).
Not including calls on the X and Y chromosomes (the CNV
algorithms call a 1-copy deletion of the X for male samples and
a 0-copy deletion of the Y chromosome for female samples),
detected aberrations ranged in size from 0.7 to 1,791 kb
(0-copy deletion), from 0.6 to 12,875 kb (1-copy deletion), and
from 0.9 to 6,896 kb (3-copy duplication) (Figures S4A–S4E).
The average length of 3-copy duplications was higher in hESCs
and hiPSCs than in non-PSCs (Wilcoxon rank sum test p values =
1.42 3 10�15 and 5.32 3 10�5, respectively), suggesting that
either the incidence of large aberrations is higher in hPSC
cultures, there is positive selection for cells with large aberrations
in hPSC cultures, or there is negative selection against such cells
in non-PSC cultures.
Correlation between CNVs and Data Qualityor Culture ParametersThere was no correlation between the number of CNVs detected
in the samples and passage number, the quality of the SNP gen-
otyping data as measured by GenomeStudio genotyping call
rate, or the Nexus quality score (Figures S4F–S4H). We did not
observe a correlation between passage number or passage
method and the number of aberrations, even for samples
collected from the same cell line (Figures S4I–S4K). There were
several very early passage hESC and hiPSC samples with large
numbers of genomic aberrations, and the only noted association
between passage number and the number of aberrations was in
hiPSC lines that were meticulously cultured in a manner that
ensured a linear path from samples collected serially during
passage. In routine practice, the culture of any given line is highly
branched, and investigators frequently do not know the true rela-
tionship among the various cryopreserved stocks, frozen nucleic
acid samples, and live cultures for any given line. Our observa-
tions indicate that it is critical not only to record the passage
number, but also the ‘‘pedigree,’’ of each culture, in order to
be able to know with certainty whether a previous assessment
of the genomic stability of a line has any bearing on a current
culture of that line. It is important to note that these findings do
not exclude the possibility of an effect of culture conditions on
genomic stability, but indicate that experiments to assess such
an effect must be carefully designed and implemented.
Duplications of Pseudogenes of Pluripotency-Associated GenesInterestingly, we found a high frequency of duplications in
pseudogenes of the pluripotency-associated NANOG and
OCT4/POU5F1 genes, including NANOGP1 (Figure 2A). It has
been noted that genes active in early embryogenesis, such as
OCT4/POU5F1, NANOG, GDF3, and STELLA, tend to have
many pseudogenes (Booth and Holland, 2004; Elliman et al.,
2006; Liedtke et al., 2007; Pain et al., 2005). NANOG has
an unusually large number of pseudogenes (11) of which
NANOGP1 is the only unprocessed pseudogene, retaining the
ell Stem Cell 8, 106–118, January 7, 2011 ª2011 Elsevier Inc. 109
A Chromosome 12
GDF3 CLEC4C SLC2A14 SLC2A3 C3AR1 NECAP1
APOBEC1 DPPA3 NANOG NANOGP1 FOXJ2 CLEC4a
7.7 7.75 7.8 7.85 7.95 8.0 8.05 8.17.9 8.15 8.27.657.6million bp
POU5F1P3
HES2P28/55/82/114**
HES3P31/54/60**HUES13P21*
HUES7P21*WA01P51***
BG01P67/VP53**
WA09P77C1***
FES22P44*
Duplication in hESC lineDuplication hiPSC line
*one culture available for analysis **multiple cultures available for analysis, all contain duplication***multiple cultures available for analysis, only one culture contains duplication
B Chromosome 20
DEFB121 MYLK2 POFUT1 BPIL3
REM1 COX4I2 C20orf186
DEFB123 DUSP15 HCK mir-1825 C20orf185
DEFB119 FOXS1 C20orf160 BPIL1 C20orf114TPX2DEFB116 DEFB118 ID1 PDRG1 PLAGL2 C20orf112 COMMD7 SPAG4L CDK5RAP1
TTLL9DEFB124 KIF3B MAPRE1 C20orf71
HM13 BCL2L1DEFB115 XKR7 TM9SF4 ASXL1 DNMT3B C20orf70
PLUNC
29 29.5 30 30.5 31 31.5million bp
SIVF017HDP43***
WA07P34MNPD29***SIVF019P67***
CM14P87***ESIH3P114***
BG01P67*
SIVF001P41***
HDF51IPS11P33***
SIVF019P53***
WA07P96CMD7***
Early Passage: CM14P21
Late Passage: CM14P87
ESIH3P114**
Cell Stem Cell
Genomic Instability of Human Pluripotent Cells
110 Cell Stem Cell 8, 106–118, January 7, 2011 ª2011 Elsevier Inc.
0%
20%
40%
60%
80%
100%
120%
0 20 40 60Cum
ulat
ive
prec
ent o
f cel
l lin
es
Number of 0-copy events
0-copy events
0%
20%
40%
60%
80%
100%
120%
0 10 20 30 40 50Cum
ulat
ive
prec
ent o
f cel
l lin
esNumber of 1-copy events
1-copy events
0%
20%
40%
60%
80%
100%
120%
0 20 40 60 80Cum
ulat
ive
prec
ent o
f cel
l lin
es
Number of 3-copy events
3-copy events
D
Cell Type # samples total allelic losses 1-copy deletions 3-copy duplicationshESCs 64 4.98 7.56 4.88hiPSCs 35 4.63 9.00 3.43non-PSCs 69 3.75 6.04 1.87
total allelic losses 1-copy deletions 3-copy duplications0.425 0.288 0.1260.168 5.90E-06 1.54E-05
ComparisonhESC vs. non-PSChiPSC vs. non-PSC
Average number per sample
p value (by Wilcoxon rank sum test)E
A B C
hESC
hiPSC
non-PSC
Figure 3. Number of Regions of Duplication and Deletion, as Identified by CNVPartition
(A–C) Cumulative distribution function plots of the numbers of 0-copy (total allelic loss), 1-copy, and 3-copy, and total CNVs for each sample type (hESCs,
hiPSCs, and non-PSCs).
(D) Average number per sample of each type of CNV for the hESC, hiPSC, and non-PSC samples.
(E) Wilcoxon rank sum p values for each type of CNV, comparing hESC versus nonpluripotent and hiPSC versus nonpluripotent. Significant p values (<0.05) are
highlighted in red.
See also Figure S4.
Cell Stem Cell
Genomic Instability of Human Pluripotent Cells
exon-intron structure of the coding gene. Of the other NANOG
pseudogenes, NANOGP4 is in the region of chromosome 7
duplicated in the FES29P39 sample, and NANOGP8 is in the
region of chromosome 15 that was duplicated in a subpopulation
of the late-passage MIZ4P88 line (Figure 4A). NANOGP9
and NANOGP10 are on the X chromosome and were duplicated
in a subpopulation of the late-passage UC06P112 sample (Fig-
ure 4B). In terms of OCT4/POU5F1 pseudogenes, POU5F1P4
is located on chromosome 1, which was trisomic in the
WA07P95 sample; POU5F1P6 is located in a region of chromo-
some 3 that is duplicated in the SIVF002P17 and the MEL2P13
samples; and POU5F1P3 is located on chromosome 12, which
was trisomic in samples from five hESC lines (Figure 3). The
ESI051P37 sample is interesting, in that it possessed a large
deletion that encompasses the OCT4/POU5F1 and NANOGP3
genes. There is little known about the role that transcribed
Figure 2. Details of Regions of CNV on Chromosome 12 and Chromos
Chromosome 12 shown in (A) and chromosome 20 shown in (B). Areas of duplicat
of overlap between the hPSC samples are highlighted in pink. The pluripotency-as
vertical blue lines in (B) indicate the boundaries of the DNMT3B gene. The lower
some 20 arose during long-term passage of the hESC line CM14. See also Figur
C
pseudogenes may play in cellular function. In one report (Hirot-
sune et al., 2003), a pseudogene was shown to stabilize the tran-
script of its protein-coding homolog, although its mechanism of
action was unclear. It is intriguing to speculate that the pseudo-
genes of the pluripotency-associated genes may exert positive
or negative regulatory influence over these genes.
Dynamic Changes in Genomic Structurein hPSC PopulationsWe observed cases where duplications appeared and took over
hESCcultures. In theMIZ4 line, therewas evidence that a trisomy
of chromosome 15 had arisen in a subpopulation of cells
between passage 33 and passage 88 (Figure 4A). In the UC06
line, the subpopulation of cells that had a trisomy of the X chro-
mosome at passage 59 had taken over a larger proportion of the
population by passage 112 (Figure 4B). These instances
ome 20
ion are shown in red bars for hESC samples and blue for hiPSC samples. Areas
sociated genesNANOG andDNMT3B are highlighted by red ovals. The dashed
panel of (B) shows the BAF plots demonstrating that a duplication on chromo-
es S2 and S3 and Table S5.
ell Stem Cell 8, 106–118, January 7, 2011 ª2011 Elsevier Inc. 111
Figure 4. Dynamic Copy Number Changes
over Long-Term Passage
(A) BAF and LRR plots of chromosome 15 for
early- and late-passage samples of the MIZ4
hESC line. The early-passage plots show normal
autosomal BAF and LRR distributions, whereas
the late-passage plots indicate that a subpopula-
tion of the cells have a duplication of chromosome
15.
(B) BAF and LRR plots of the X chromosome for
early- and late-passage samples of the UC06
hESC line. There is a subtle widening of the band
of heterozygous SNPs in the BAF plot for the
early-passage sample, which has separated into
two distinct bands in the BAF plot for the late-
passage sample, indicating that the small subpop-
ulation of cells carrying a duplication of the X chro-
mosome in the early-passage population has
outcompeted the cells without the duplication
over long-term passage.
Cell Stem Cell
Genomic Instability of Human Pluripotent Cells
highlighted the need for improvedmethods for detecting CNVs in
mosaic populations of cells. We analyzed mixtures of cells,
where we varied the proportion of HDF51IPS11P33 cells, which
contain a duplication in chromosome 20, and the parental
HDF51 fibroblast line, which is genomically normal in this region.
By using CNVPartition, we were able to detect the presence of
112 Cell Stem Cell 8, 106–118, January 7, 2011 ª2011 Elsevier Inc.
the duplication when the percentage of
HDF51IPS11P33 cells was R70% of the
cells. However, calculating BAF distance
can be used to detect the presence of the
duplication when R20% of the popula-
tion is affected (Figure 5B; Figure S5A),
indicating that improvements in CNV
calling algorithms may be possible and
would be very useful.
Genomic Aberrations duringReprogramming and Passageof hiPSCshiPSCs present an ideal system for distin-
guishing between the effects of reprog-
ramming and passage on genomic
stability. They also confer the ability to
determine with certainty whether a given
alteration is new, because the parental
differentiated cells can also be analyzed.
Accordingly, we analyzed 3 samples
from a primary human fetal fibroblast
line, HDF51, and 12 independent hiPSC
clones generated from that line. For the
hiPSC clones, we collected samples at
early (passage 5–8), mid (passage 12–
15), and late (passage 25–34) passage
and analyzed at least the early- and
late-passage samples. In addition to
identifying duplications via CNVPartition,
we identified deletions by using a combi-
nation of CNVPartition and replicate error
detection, which identifies the discrepancies between SNP calls
from two samples (Table S6). Because all of the samples origi-
nated from the same individual, the replicate error detection rep-
resented a way of improving our confidence in our deletion calls.
Inspecting the duplication and deletion calls for the HDF51 and
HDF51IPS samples (Figure 6), we noticed that all 11 deletions
Day 2Differentiation
Day 7Differentiation
0
0.05
0.1
0.15
0.2
0.25
0.3
Mea
n D
ista
nce
BAF Distance for Heterozygous SNPS
100% duplication
50% duplication
30% duplication
A
B
1 2 3
Figure 5. Duplications on Chromosome 20
Arising over a 5 Day Period during Directed
Differentiation of hESCs to Cardiomyocytes
(A) The top two panels show the BAF and LRR
plots at day 2 of the differentiation protocol; the
bottom two panels show the plots at day 7. Three
segments showing different degrees of separation
of the ‘‘cloud’’ of BAF values for heterozygous
SNPs are labeled 1, 2, and 3.
(B) The BAF distance for heterozygous SNPs are
shown for regions duplicated on chromosome
20. The BAF distance for mixtures of known ratios
of HDF51 cells (which have no duplication on chro-
mosome 20) and HDF51IPS11P33 cells (which
have a duplication of the proximal portion of the
q-arm) are shown on the left (BAF and LRR plots
are shown in Figure S4A). The BAF distances
for the three partially duplicated segments
(corresponding to the segments labeled 1 (red),
2 (blue), and 3 (green) in [A]) are shown on the right
and have been used to estimate the percent of the
population carrying the duplication.
See also Figure S5.
Cell Stem Cell
Genomic Instability of Human Pluripotent Cells
appeared by the earliest passage time point, whereas 5 out of 6
duplications arose during the course of long-term passage. In
fact, some of the deletions receded from the population over
long-term passage, suggesting that they were positively
selected during reprogramming and negatively selected during
passage (Figure S6).
Of the seven duplicated regions that were present in an
HDF51IPS line, but absent from the parental HDF51s, six con-
tained the coding region and/or promoter region of at least one
gene. The overexpression of five of these genes (in red in Fig-
ure 6) were positively associated with tumorigenicity or cell
proliferation, whereas for one (FRS2, in green in Figure 6), low
expression has been associated with poor prognosis in non-
small cell lung cancer (Iejima et al., 2010). BCL2L1 (in orange
in Figure 6) has two isoforms, one of which is proapoptotic and
Cell Stem Cell 8, 106–11
the other is antiapoptotic (Boise et al.,
1993). All 12 deletion regions overlapped
at least one gene, and 5 of them con-
tained genes that have evidence of
tumor-suppressor activity.
It is notable that the presence of the
transduced copies of the reprogramming
factors did not confound our analysis by
appearing as duplications in the reprog-
ramming genes. This is due to the fact
that the transduced genes included only
the coding sequences (which have few
SNPs), and that to identify a CNV region
the CNV-calling algorithms require longer
stretches of consecutive SNPs to be
affected.
Genomic Aberrations Arisingduring DifferentiationThe most rapidly arising genomic aberra-
tions in our data set were identified
in samples from a directed differentiation experiment. Parallel
differentiations were performed with WA07 cells at passage
95 and 96, with samples collected from the undifferentiated
cells (WA07P95), on differentiation day 2 (WA07P95CMD2
andWA07P96CMD2), and differentiation day 7 (WA07P95CMD7
and WA07P96CMD7). Partial duplications of three segments
of chromosome 20 were found in the WA07P96CMD7
sample only (Figure 5A; Figure S5B), indicating that they
arose between day 2 and day 7 of differentiation. Comparing
the BAF plots for WA07P96CMD7 to those from mixtures
of known ratios of cells with and without a duplication of
a smaller region of chromosome 20 (Figure 5B; Figure S5A),
we estimated the percent of cells in the population carrying
duplications of the three segments to be 30%, 100%,
and 50%. This finding points out that differentiation can be
8, January 7, 2011 ª2011 Elsevier Inc. 113
2
9
11
17
18
19
6
21
22
15 HDF51HDF51IPS
Duplications Deletions
1P1
HDF51IPS clone # and passage #
5
1 3P34 2P34
WNT2B, RHOC, NRAS, others AKT3, ZNF238
12 3P34
MDM2, FRS2
14all13P8/14/34
4 2P34
13P8/14/34
3all all
14P8/27 5P6/34(sub) 7P5/3312P6/33(sub)
FHIT CADM2 NAALADL2
7all 14P27
CTAGE4
14P8/27 12P6/33(sub) 1P6/252P6/34
MAD1, MAD1L1, FTSJ2, NUDT1THSD7A CNTNAP2
86P5/33
VPS13B
10all
1P6/25
CTNNA3
1312P6/33(sub)
RASA3
166P5/33
WWOX
20 11P33
11P8/14/33
MACROD2
X1P6/252P6/34
DIAPH2
Pro-tumor geneTumor-suppressor gene
PSMA3
BCL2L1, PDRG1, POFUT1, others
Figure 6. Regions of Duplication and Deletion in the HDF51 Fibroblast Line and the HDF51IPS Lines
Duplicated regions (3 or 4 copies) (identified by CNVPartition) are shown in the dark bars above the ideogram of the chromosomes, and deleted regions (0 or 1
copy) (identified by both CNVPartition and replicate error analysis) are shown in the light bars below the ideograms of the chromosomes. The line number and
passage number of the HDF51IPS line (blue) are shown adjacent to each bar, with regions where the CNV is present in only a subpopulation of the cells in a sample
denoted ‘‘(sub).’’ HDF51 fibroblast line shown in green. The names of genes overlapping the regions of CNV are indicated.
See also Figure S6 and Table S6.
Cell Stem Cell
Genomic Instability of Human Pluripotent Cells
a highly selective process and that genomically aberrant
cells can rapidly take over a population undergoing differentia-
tion. We suggest that it is important to assess the genomic
114 Cell Stem Cell 8, 106–118, January 7, 2011 ª2011 Elsevier Inc.
normality of cells frequently, not only in the pluripotent state
but also at the endpoint of differentiation experiments or other
treatments.
Cell Stem Cell
Genomic Instability of Human Pluripotent Cells
Correlations between Genomic Aberrations and GeneExpressionTo determine whether the regions of frequent duplication in
hESCs might have common features linked to the pluripotent
phenotype, we used our large-scale mRNA expression data-
base, which contains gene expression levels for a large number
of pluripotent and nonpluripotent cell lines. We found that many
of the genes in the recurrently duplicated region on chromosome
12 were more highly expressed in human pluripotent cells
compared to multiple nonpluripotent cell types (Figures S2 and
S3A). There was not a statistically significant difference in the
expression of these genes between the hPSC samples that con-
tained duplications and those that did not. However, this result
could have been confounded by the differences in genetic back-
ground and culture conditions among the lines.
We therefore examined the expression of genes found within
areas of duplication in samples in which we had genetically
matched controls (Figure S3). There was higher expression of
many genes on chromosome 20 in theWA07P96CMD7 samples,
which had partial duplications of large stretches of this chromo-
some (shown in the BAF plot on the lower panel of Figure S3A),
compared to the WA07P95CMD7 samples, which were euploid
for this chromosome. One of the genes that was most highly
affected was DNMT3B, as seen on the panel on the right. We
noted that the higher expression was not restricted to the areas
involved in the duplications, indicating potential long-range
effects of chromosomal aberrations on gene expression. These
effects appeared to be weaker, but still present, on other chro-
mosomes (see chromosome 12 panel in Figure S6A). We
ensured that this effect was not simply due to variations in
gene expression between biological replicates by examining
the corresponding data for the samples collected at day 0 and
day 2 of the same experiment (upper two panels of Figure S3A).
We also had matched controls for the HDF51IPS lines, and we
did see correlation between gene expression and presence of
duplications for these samples as well (Figure S3B). These find-
ings suggest that duplications do result in increases in gene
expression, both at the site of duplication as well as at distant
sites, which can be detected when a genetically matched
sample is used for comparison. Even though these gene expres-
sion changes are not apparent when comparing samples from
unrelated cell lines, this is unlikely to be relevant, because
a cell containing a genomic aberration is competing in culture
with a population of otherwise genetically matched cells.
DISCUSSION
This study is the most comprehensive and highest-resolution
study of the genomic stability of hPSCs to date and includes
samples from a large number of both hESCs and hiPSCs, as
well as somatic stem cells, primary cell lines, and tissues for
comparison. In addition, we analyzed a primary HFF line and
12 hiPSC clones generated from it, collected at early and late
passage, which allowed us to distinguish between genomic
aberrations that arose during derivation versus long-term culture
of hiPSCs.
This study is unique in combining a sufficient numbers of
both pluripotent and nonpluripotent samples to detect cell-
type-specific recurrent genomic aberrations in a statistically
C
significant manner and a high-resolution analysis platform that
enables the detection of kilobase-length aberrations. A recently
published study using gene expression data to detect genomic
aberrations did not include nonpluripotent samples for
comparison and was limited to detection of duplications at least
10 megabases in length (Mayshar et al., 2010). In our results,
>90% of duplications in hPSCs and 100% of duplications in
non-hPSCs were <10 megabases (Figure S2, Table S2), indi-
cating that gene expression-basedmethods are unable to detect
small genomic aberrations. Moreover, the genomic locations as-
signed via gene expression data correspond to the location of
the coding sequences of the perturbed genes, rather than the
actual genomic coordinates of the genomic aberrations.
The results presented here indicate that hESC lines contain
numerous genomic aberrations, most of which would not be
detected by karyotyping or other microscopy-based methods.
Some regions of CNV occurred multiple times in unrelated
hESC and hiPSC lines, suggesting that certain changes may
be characteristic of self-renewing pluripotent cells. It should be
noted that it was not possible to establishwith certainty the stage
at which the genomic changes occurred in the hESC samples
for which there was not an earlier passage sample demon-
strating genomic normality; some of the abnormalities may
have been present in the embryos from which the cells were
derived. The analysis of hiPSCs does not suffer from this short-
coming, provided that the parental cells collected prior to
reprogramming are analyzed. It is also important to consider
other differences between hPSCs and cultured somatic cells.
In general, because they do not undergo senescence, the
hPSC lines had been in continuous culture longer than the
primary cell lines, so some of the genetic changes seen may
be a function of the selection pressures of cell culture in general,
rather than specific to pluripotent stem cell culture.
The relatively high frequency of duplications in hPSCs raises
the concern that these genetic aberrations may increase the
risk of oncogenesis. The recurrent regions of copy number
variation on chromosomes 12 and 20, which lie in close proximity
to known pluripotency genes, are particularly worrisome,
because a major issue in cell therapy is the elimination of plurip-
otent precursors in populations destined for transplantation.
Three out of the 10 duplications on chromosome 12, and 9
out of 10 duplications on chromosome 20, developed over the
course of long-term culture of hPSCs, raising the concern that
expansion of pluripotent cells may inevitably lead to increased
genetic abnormality. However, the NANOG and NANOGP1
duplications were seen in cell lines as early as passage 21
(HUES7), 21 (HUES13), and 28 (HES2), which suggests that
low passage number does not in itself ensure genetic integrity.
Our data indicate that the pattern of genomic aberrations in
hiPSCs and hESCs may differ slightly, but that both cell types
are prone to developing such changes, and that one of the two
most significant recurrent duplications seen in hESCs, on chro-
mosome 20, was also found in one of the hiPSC lines. The other
region of recurrent duplication, encompassing the NANOG/
NANOGP1 region of chromosome 12, was detected in a late-
passage hiPSC line by means of array CGH by Chin et al. (2009).
Our results and those of others (Lefort et al., 2008;Maitra et al.,
2005; Mayshar et al., 2010; Spits et al., 2008; Wu et al., 2008)
highlight the need for optimization of derivation and culture
ell Stem Cell 8, 106–118, January 7, 2011 ª2011 Elsevier Inc. 115
Cell Stem Cell
Genomic Instability of Human Pluripotent Cells
conditions that promote genetic stability of pluripotent stem
cells. These results also underscore the need to perform further
studies that include larger numbers of pluripotent cell lines and
careful phenotypic assessments in order to distinguish genetic
variations that are harmless from those that pose clinical risks.
The evidence for accumulation of genetic aberrations in culture
of existing hPSC lines makes it clear that new hPSC lines need
to be generated now and on a continuing basis, and emphasizes
the necessity of frequent assessments of genomic stability in
hPSC lines, both in the pluripotent state and when the cells are
subjected to other potentially selective conditions, such as
differentiation procedures.
EXPERIMENTAL PROCEDURES
Cell Culture
All cell types were derived and propagated as described in the references
listed in Table S1. This work was approved by the Embryonic Stem Cell
Research Oversight Committee at the University of California, San Diego,
which oversees pluripotent stem cell research at both UCSD and TSRI.
DNA Purification
Genomic DNA was purified with the DNeasy Blood & Tissue Kit (QIAGEN).
SNP Genotyping
SNP genotyping was performed on the Illumina OmniQuad version 1, which
interrogates 1,140,419 SNPs across the human genome. 1 mg input genomic
DNA (the yield from approximately 200,000 cells) was amplified and labeled
according to the manufacturer’s instructions. The DNA was quantified with
the PicoGreen reagent (Invitrogen, Inc.). The labeled product was then hybrid-
ized to the array and scanned on a BeadArray Reader (Illumina, Inc.). Genotyp-
ing calls were made with BeadStudio (Illumina, Inc.), via the standard cluster
files provided by the manufacturer. The GenCall threshold was set to 0.15,
and the call rates were between 0.979 and 0.999.
Copy Number Variation Assessment
For the SNP Genotyping data, data preprocessing was performed in
BeadScan (Illumina, Inc.). Data cleaning, SNP calling, and replicate error iden-
tification was performed in GenomeStudio (Illumina, Inc.). CNVPartition v2.4.4
(Illumina, 2008) was used as the primary CNV-calling algorithm for the results
presented in this paper. CNV regions were also identified with the SNPRank
Segmentation aligorithm in Nexus (Biodiscovery, Inc.) to assess concordance
between the two methods. The CNVPartition CNV score threshold was set at
50, with a minimum number of SNPs per CNV region of 10. The Nexus param-
eters included a significance threshold of 1 3 10�8 and a minimum number of
probes per segment of 10.
We chose to remove data from probes on the array that were designated as
‘‘CNV’’ probes prior to using CNVPartition and Nexus. We did this for two
reasons: first, the CNV probes were designed as monoallelic probes, and
hence provide no B-allele-frequency information, potentially reducing their
accuracy in calling duplications and deletions; second, we were interested
in detecting genomic aberrations that occurred with derivation and passage
of cell lines (and potentially with tissue-specific differentiation), rather than
CNVs that vary among individuals and are carried in the germline, which are
the ones targeted by the CNV probes.
Because the average spacing of SNPs on the SNP genotyping microarrays
used was 3 kb, the shortest detectable CNV regions were expected to be
approximately 30 kb. These two algorithms generally identified similar regions
of duplication (97% agreement on the individual SNP level for 3-copy duplica-
tions and 99% for 1-copy deletions) (Tables S1 and S2).
Overlap between CNV Calls and Common CNVs
An overlap was identified between the CNVs called by CNVPartition in our data
set and the common CNVs observed in 450 reference HapMap samples
(Conrad et al., 2010) when the common region of the CNVs exceeded both
20% of the CNV identified in our samples and 20% of the common CNV in
116 Cell Stem Cell 8, 106–118, January 7, 2011 ª2011 Elsevier Inc.
the reference set. The CNVs in our data set that overlap with common CNVs
are indicated by an asterisk in Figure 1 and were also removed from subse-
quent analyses.
Validation of CNV Calls
CNV calls for CNVPartition and Nexus were validated by performing qRT-PCR
for a subset of the CNV calls. TaqMan CNV assays (Life Technologies, Inc.)
were performed according to the manufacturer’s instructions. Assays were
performed in triplicate, with the HDF51IPS1P25 sample used as the reference.
The predicted copy number was calculated with the equation
CN= 2��2 ð�ðDeltaDeltaCtÞÞ
�:
Validation of SNP Calls
Because replicate errors could be identified only where samples were derived
from the same original cell line, replicate error calling was performed only for
the HDF51-derived lines. For these samples, SNP calls were validated by
performing qPCR for a subset of the loci where replicate errors were called.
TaqMan SNP assays (Life Technologies, Inc.) were performed according to
the manufacturer’s instructions. The HDF51P11 sample was used as the refer-
ence. There were 8 homozygous-to-homozygous replicate errors identified (0/
4 tested were confirmed), 313 homozygous-to-heterozygous replicate errors
(0/14 were confirmed), and 310 heterozygous-to-homozygous replicate errors
(11/11 were confirmed) (Table S3). These results indicate that the large
majority of apparent SNP mutations identified by replicate error analysis are
in fact due to SNP genotyping error; this result is not unexpected based on
reports that the discrepancy between SNP calls by sequencing and microar-
ray-based SNP genotyping is �0.1%–0.05% (Bentley et al., 2008). Based on
the average number of heterozygous and homozygous SNPs in the SNP gen-
otyping data (�20% heterozygous and 80% homozygous), we would have ex-
pected an excess of homozygous-to-heterozygous replicate error calls. The
reason for the larger than expected number of heterozygous-to-homozygous
calls was due to the fact that deletions and some duplications (when the
cluster separation is poor) appear to result in replicate error calls; this is also
the reason that heterozygous-to-homozygous replicate error calls are also
expected to be better validated.
Calculation of BAF Distance
For intervals of interest, homozygous SNPswere removed by eliminating SNPs
with BAF values >0.8 or <0.2. The heterozygous SNPswere separated into two
clusters, with the median BAF value of the heterozygous SNPs as a cutoff. The
‘‘AAB’’ cluster had BAF values < median BAF, and the ‘‘ABB’’ cluster had BAF
values > median BAF. The difference between the mean BAF for the AAB
cluster and the mean BAF for the ABB cluster was the BAF distance.
ACCESSION NUMBERS
The microarray data are available in the Gene Expression Omnibus (GEO)
database (http://www.ncbi.nlm.nih.gov/gds) under the accession number
GSE25925.
SUPPLEMENTAL INFORMATION
Supplemental Information includes six figures and six tables and can be found
with this article online at doi:10.1016/j.stem.2010.12.003.
ACKNOWLEDGMENTS
We would like to acknowledge all of the collaborators who contributed
samples to this study, including Eirini Papapetrou (Sadelain lab), Dongbao
Chen, Ralph Graichen, Jerold Chun, Martin Pera, James Shen, Scott
McKercher, Timo Otonkoski, and Sheng Ding. We would like to thank Gulsah
Altun for invaluable assistance. We would like to thank the NICHD Brain and
Tissue Bank for Developmental Disorders, Planned Parenthood of San Diego
and Riverside Counties, and Christopher Barry for generously providing
tissue specimens for this study. L.C.L. was supported by an NIH/NICHD
K12 Career Development Award and the Hartwell Foundation. J.F.L., I.S.,
Cell Stem Cell
Genomic Instability of Human Pluripotent Cells
H.T., C.L., and F.-J.M. are supported by CIRM (CL1-00502, RT1-01108,
TR1-01250, RN2-00931-1), NIH (R21MH087925), the Millipore Foundation,
and the Esther O’Keefe Foundation. I.U. was supported in part by a fellowship
from the Edmond J. Safra foundation in Tel Aviv University and by the Legacy
stem cell research fund. I.S. was supported by the PEW Charitable Trust.
H.-S.P. and S.L. were supported by a SCRC Grant (SC2250) of the 21st
Century Frontier Research Program funded by the Ministry of Education,
Science and Technology. M.J.B. was partially supported by grants RYC-
2007-01510 and SAF2009-08588 from the Ministerio de Ciencia e Innovacion
of Spain. Work in the laboratory of J.C.I.B. was supported by grants from MI-
CINN Fundacion Cellex, the G. Harold and Leila Y. Mathers Charitable Foun-
dation, and Sanofi-Aventis. C.M. was supported by NIH grants R01
HL64387, P01 HL094374, R01 HL084642, and P01 GM081719. V.G. was
partially supported by NHLBI, RC1HL100168. R. Shamir was supported in
part by the Israel Science Foundation (grant no. 802/08). A.L.L. was supported
by grants from the Australian Stem Cell Centre and from the Victoria-California
Stem Cell Alliance (TR101250) between CIRM and the state government of
Victoria, Australia. H.S.K. is the chairman of the scientific advisory board of
California Stem Cell, Inc. R. Semechkin and M.M. are employees and share-
holders of International Stem Cell Corporation.
Received: October 15, 2009
Revised: October 10, 2010
Accepted: December 7, 2010
Published: January 6, 2011
REFERENCES
Atkin, N.B., and Baker, M.C. (1982). Specific chromosome change, i(12p), in
testicular tumours? Lancet 2, 1349.
Baker, D.E., Harrison, N.J., Maltby, E., Smith, K., Moore, H.D., Shaw, P.J.,
Heath, P.R., Holden, H., and Andrews, P.W. (2007). Adaptation to culture of
human embryonic stem cells and oncogenesis in vivo. Nat. Biotechnol. 25,
207–215.
Bentley, D.R., Balasubramanian, S., Swerdlow, H.P., Smith, G.P., Milton, J.,
Brown, C.G., Hall, K.P., Evers, D.J., Barnes, C.L., Bignell, H.R., et al. (2008).
Accurate whole human genome sequencing using reversible terminator chem-
istry. Nature 456, 53–59.
Boise, L.H., Gonzalez-Garcıa, M., Postema, C.E., Ding, L., Lindsten, T., Turka,
L.A., Mao, X., Nunez, G., and Thompson, C.B. (1993). bcl-x, a bcl-2-related
gene that functions as a dominant regulator of apoptotic cell death. Cell 74,
597–608.
Booth, H.A., and Holland, P.W. (2004). Eleven daughters of NANOG.
Genomics 84, 229–238.
Chin, M.H., Mason, M.J., Xie, W., Volinia, S., Singer, M., Peterson, C.,
Ambartsumyan, G., Aimiuwu, O., Richter, L., Zhang, J., et al. (2009). Induced
pluripotent stem cells and embryonic stem cells are distinguished by gene
expression signatures. Cell Stem Cell 5, 111–123.
Conrad, D.F., Pinto, D., Redon, R., Feuk, L., Gokcumen, O., Zhang, Y., Aerts,
J., Andrews, T.D., Barnes, C., Campbell, P., et al; Wellcome Trust Case Control
Consortium. (2010). Origins and functional impact of copy number variation in
the human genome. Nature 464, 704–712.
Draper, J.S., Smith, K., Gokhale, P., Moore, H.D., Maltby, E., Johnson, J.,
Meisner, L., Zwaka, T.P., Thomson, J.A., and Andrews, P.W. (2004).
Recurrent gain of chromosomes 17q and 12 in cultured human embryonic
stem cells. Nat. Biotechnol. 22, 53–54.
Elliman, S.J., Wu, I., and Kemp, D.M. (2006). Adult tissue-specific expression
of a Dppa3-derived retrogene represents a postnatal transcript of pluripotent
cell origin. J. Biol. Chem. 281, 16–19.
Fox, J.L. (2008). FDA scrutinizes human stem cell therapies. Nat. Biotechnol.
26, 598–599.
Guan, X.Y., Horsman, D., Zhang, H.E., Parsa, N.Z., Meltzer, P.S., and Trent,
J.M. (1996). Localization by chromosome microdissection of a recurrent
breakpoint region on chromosome 6 in human B-cell lymphoma. Blood 88,
1418–1422.
C
Hirotsune, S., Yoshida, N., Chen, A., Garrett, L., Sugiyama, F., Takahashi, S.,
Yagami, K., Wynshaw-Boris, A., and Yoshiki, A. (2003). An expressed pseudo-
gene regulates the messenger-RNA stability of its homologous coding gene.
Nature 423, 91–96.
Hurst, C.D., Fiegler, H., Carr, P., Williams, S., Carter, N.P., and Knowles, M.A.
(2004). High-resolution analysis of genomic copy number alterations in bladder
cancer by microarray-based comparative genomic hybridization. Oncogene
23, 2250–2263.
Iejima, D., Minegishi, Y., Takenaka, K., Siswanto, A., Watanabe, M., Huang, L.,
Watanabe, T., Tanaka, F., Kuroda, M., and Gotoh, N. (2010). FRS2beta,
a potential prognostic gene for non-small cell lung cancer, encodes a feedback
inhibitor of EGF receptor family members by ERK binding. Oncogene 29,
3087–3099.
Illumina (2008). Technical Note: DNA Copy Number Analysis Algorithms.
http://wwwilluminacom/downloads/CNValgorithms_TechNotepdf.
Imreh, M.P., Gertow, K., Cedervall, J., Unger, C., Holmberg, K., Szoke, K.,
Csoregh, L., Fried, G., Dilber, S., Blennow, E., and Ahrlund-Richter, L.
(2006). In vitro culture conditions favoring selection of chromosomal abnormal-
ities in human ES cells. J. Cell. Biochem. 99, 508–516.
Koynova, D.K., Jordanova, E.S., Milev, A.D., Dijkman, R., Kirov, K.S.,
Toncheva, D.I., and Gruis, N.A. (2007). Gene-specific fluorescence in-situ
hybridization analysis on tissue microarray to refine the region of chromosome
20q amplification in melanoma. Melanoma Res. 17, 37–41.
Kresse, S.H., Szuhai, K., Barragan-Polania, A.H., Rydbeck, H., Cleton-Jansen,
A.M., Myklebost, O., and Meza-Zepeda, L.A. (2010). Evaluation of high-reso-
lution microarray platforms for genomic profiling of bone tumours. BMC Res
Notes 3, 223.
Lefort, N., Feyeux, M., Bas, C., Feraud, O., Bennaceur-Griscelli, A., Tachdjian,
G., Peschanski, M., and Perrier, A.L. (2008). Human embryonic stem cells
reveal recurrent genomic instability at 20q11.21. Nat. Biotechnol. 26, 1364–
1366.
Liedtke, S., Enczmann, J., Waclawczyk, S., Wernet, P., and Kogler, G. (2007).
Oct4 and its pseudogenes confuse stem cell research. Cell Stem Cell 1,
364–366.
Maitra, A., Arking, D.E., Shivapurkar, N., Ikeda, M., Stastny, V., Kassauei, K.,
Sui, G., Cutler, D.J., Liu, Y., Brimble, S.N., et al. (2005). Genomic alterations
in cultured human embryonic stem cells. Nat. Genet. 37, 1099–1103.
Mayshar, Y., Ben-David, U., Lavon, N., Biancotti, J.-C., Yakir, B., Clark, A.T.,
Plath, K., Lowry, W.E., and Benvenisty, N. (2010). Identification and classifica-
tion of chromosomal aberrations in human induced pluripotent stem cells. Cell
Stem Cell 7, 521–531.
Midorikawa, Y., Yamamoto, S., Ishikawa, S., Kamimura, N., Igarashi, H.,
Sugimura, H., Makuuchi, M., and Aburatani, H. (2006). Molecular karyotyping
of human hepatocellular carcinoma using single-nucleotide polymorphism
arrays. Oncogene 25, 5581–5590.
Mitalipova, M.M., Rao, R.R., Hoyer, D.M., Johnson, J.A., Meisner, L.F., Jones,
K.L., Dalton, S., and Stice, S.L. (2005). Preserving the genetic integrity of
human embryonic stem cells. Nat. Biotechnol. 23, 19–20.
Narva, E., Autio, R., Rahkonen, N., Kong, L., Harrison, N., Kitsberg, D.,
Borghese, L., Itskovitz-Eldor, J., Rasool, O., Dvorak, P., et al. (2010). High-
resolution DNA analysis of human embryonic stem cell lines reveals culture-
induced copy number changes and loss of heterozygosity. Nat. Biotechnol.
28, 371–377.
Pain, D., Chirn, G.W., Strassel, C., and Kemp, D.M. (2005). Multiple retropseu-
dogenes from pluripotent cell-specific gene expression indicates a potential
signature for novel gene identification. J. Biol. Chem. 280, 6265–6268.
Rodriguez, E., Houldsworth, J., Reuter, V.E., Meltzer, P., Zhang, J., Trent, J.M.,
Bosl, G.J., and Chaganti, R.S. (1993). Molecular cytogenetic analysis of i(12p)-
negative human male germ cell tumors. Genes Chromosomes Cancer 8,
230–236.
Scotto, L., Narayan, G., Nandula, S.V., Arias-Pulido, H., Subramaniyam, S.,
Schneider, A., Kaufmann, A.M., Wright, J.D., Pothuri, B., Mansukhani, M.,
and Murty, V.V. (2008). Identification of copy number gain and overexpressed
genes on chromosome arm 20q by an integrative genomic approach in
ell Stem Cell 8, 106–118, January 7, 2011 ª2011 Elsevier Inc. 117
Cell Stem Cell
Genomic Instability of Human Pluripotent Cells
cervical cancer: Potential role in progression. Genes Chromosomes Cancer
47, 755–765.
Skotheim, R.I., Monni, O., Mousses, S., Fossa, S.D., Kallioniemi, O.P., Lothe,
R.A., and Kallioniemi, A. (2002). New insights into testicular germ cell tumori-
genesis from gene expression profiling. Cancer Res. 62, 2359–2364.
Spits, C., Mateizel, I., Geens, M., Mertzanidou, A., Staessen, C., Vandeskelde,
Y., Van der Elst, J., Liebaers, I., and Sermon, K. (2008). Recurrent chromo-
somal abnormalities in human embryonic stem cells. Nat. Biotechnol. 26,
1361–1363.
Tanner, M.M., Tirkkonen, M., Kallioniemi, A., Isola, J., Kuukasjarvi, T., Collins,
C., Kowbel, D., Guan, X.Y., Trent, J., Gray, J.W., et al. (1996). Independent
118 Cell Stem Cell 8, 106–118, January 7, 2011 ª2011 Elsevier Inc.
amplification and frequent co-amplification of three nonsyntenic regions on
the long arm of chromosome 20 in human breast cancer. Cancer Res. 56,
3441–3445.
Tonon, G., Wong, K.K., Maulik, G., Brennan, C., Feng, B., Zhang, Y., Khatry,
D.B., Protopopov, A., You, M.J., Aguirre, A.J., et al. (2005). High-resolution
genomic profiles of human lung cancer. Proc. Natl. Acad. Sci. USA 102,
9625–9630.
Wu, H., Kim, K.J., Mehta, K., Paxia, S., Sundstrom, A., Anantharaman, T.,
Kuraishy, A.I., Doan, T., Ghosh, J., Pyle, A.D., et al. (2008). Copy
number variant analysis of human embryonic stem cells. Stem Cells 26,
1484–1489.
Cell Stem Cell
Erratum
In Vivo Fate Mapping and Expression AnalysisReveals Molecular Hallmarksof Prospectively Isolated Adult Neural Stem CellsRuth Beckervordersandforth, Pratibha Tripathi, Jovica Ninkovic, Efil Bayam, Alexandra Lepier, Barbara Stempfhuber,Frank Kirchhoff, Johannes Hirrlinger, Anja Haslinger, D. Chichung Lie, Johannes Beckers, Bradley Yoder, Martin Irmler,and Magdalena Gotz**Correspondence: magdalena.goetz@helmholtz-muenchen.deDOI 10.1016/j.stem.2010.12.016
(Cell Stem Cell 7, 744–758; December 3, 2010)
During the preparation of Figure 3, the authors inadvertently included amodified version of panel C in place of the isotype control data
intended to form panel B. The corrected version of the figure appears below. All other figure panels are the same as in the published
paper. Figure 3 in the online version of the paper has been replaced with this corrected version.
I
die
ncep
halo
n (n
on-n
euro
geni
c)
βIII tubulin GFAP O4
Diencephalon
SEZ
G
F
hGFAP-GFP prominin1-PE
hGFAP-GFP hGFAP-GFP prominin1-PE
H
0
20
40
60
80
hGFAP-GFP+prominin1+
hGFAP-GFP+only
prominin1+only
all negative
% o
f sin
gle
plat
ed c
ells
fo
rmin
g ne
uros
pher
es Secondary neurospheresPrimary neurospheres
prominin1+only
6.5%
hGFAP-GFP+
prominin1+
2.5%
hGFAP-GFP+only
17.1%
AC
hGFAP-GFP+ 11%
D E
WT Isotype control-PE
SE
Z (n
euro
geni
c)
Bprominin1+only
0%
hGFAP-GFP+
prominin1+
0%
hGFAP-GFP+only
0%
prominin1+only
0.2%
hGFAP-GFP+
prominin1+
0%
hGFAP-GFP+only
11%
Figure 3. FACS Analysis, Sorting, and Neurosphere-Forming Potential of the Sorted Cells
Cell Stem Cell 8, 119, January 7, 2011 ª2011 Elsevier Inc. 119
Cell Stem Cell
Editors’ Notes
History in the MakingLast month saw the Nobel committee award the 2010 prize for physiology or medicine to Robert Edwards for his pioneeringefforts to establish human in vitro fertilization (IVF).While a number of other groups added experimental tools that helped bringthe technique into modern clinical practice, in many ways Edwards can also lay claim to founding, at least intellectually, thehuman embryonic stem cell field. Numerous parallels exist between the public reception of and regulatory policies for IVF andhESCs, and in their Forum article, Gearhart and Coutifaris offer their take on the historical origins of both fields and of the polit-ical lessons that they feel hESC research proponents should bear inmind and aim to put into practice. In the debate over hESCresearch funding, some advocates claim that the availability of human iPSCs overcomes the need for continued hESC deri-vation. Loring and colleagues, however, describe that both categories of pluripotent cell lines are prone to subchromosomalgenomic aberrations. They emphasize that while hiPSCs and hESCs are biased towards different aberrations, the high rates ofchange in both cell types mean that frequent genomic monitoring will be required to assure clinical safety of any therapiesderived from pluripotent cells. The types of aberrations that arise in human pluripotent cells also seem to shift over time inculture, emphasizing the need to understand how culture conditions, including signaling molecules, regulate pluripotentcell-fate outcomes. Using a mouse ESC model system, Jin and coauthors shed light on the specific roles played by theNFAT and Erk signaling cascades in regulating the switch between self-renewal and lineage specification. Clarifying thesignals at play in mouse ESCs may also help improve protocols designed to support maintenance versus differentiation of
hESC and hiPSC populations. It has also been emphasized that workwith hESCs will be needed at least as long as the work to understandthe reprogramming process continues. To that end, Meissner andcolleagues use an inducible reprogramming system to track very earlyepigenetic changes that occur during the generation of mouse iPSCs.They find that histone methylation patterns are altered prior to geneexpression changes and, in doing so, offer insight into the temporal prog-ress of reprogramming in response to exposure to ectopic factors.The Impact of Age and StressThe specifics of the pathways activated in response to stress anddamage, and the outcome of those pathways on stem cells and theirprogeny, form a focus for three articles in this issue. Kornblum andcolleagues isolated neural progenitors from mouse brain and found that
this population is actually maintained and stimulated by reactive oxygen species (ROS) which act as second messengersin the PI3K/Akt signal cascade, unlike in other cell populations that typically translate ROS as a danger and damage stimulus.Clearly, the context of a given stress signal, and the identity and function of the cell type receiving that signal, will impact thespecific response made under different conditions. These themes are discussed in detail by Passague, Blanpain and coau-thors in their Review article, who also raise the topic of howDNA-damage-response pathwaysmay bemisused by, or perhapstargeted to eliminate, cancer stem cells. There are other situations when having insight into damage and stress responsesmight offer clinical insight. For example, Colman and colleagues describe the generation of human iPSCs derived fromHutch-inson-Gilford Progeria Syndrome patient fibroblasts. This lethal premature aging disease affects cells frommany tissues, and several differentiated progeny from themutant pluripotent cells, includingMSCs, display defective responses to DNA damage and stress. Theauthors use this model system to provide insight into the inner workingsof HGPS pathology, and these lines will likely help dissect cellularresponses involved in the mechanisms of aging as well.Signaling the NichePhysical damage and stress also relay important signals to stem cell pop-ulations, and Jiang and colleagues look at specific signaling cascadesactivated by this response in intestinal epithelium stem cells of theDrosophila midgut. They identify the EGFR/Ras/MAPK cascade asessential for promoting gut epithelial regeneration and highlight howthis stem cell population reads environmental cues in order to maintainor achieve homeostasis. External regulatory inputs on stem cell fateand function are often derived from the niche, which can include both
cellular and acellular components. For migratory stem cell populations, niche inputs need to reach a balance betweenstem cell anchoring and mobilization, and whether this balance is pushed to one side or the other will likely be determinedby circulating signals that originate from outside the niche. Forsberg and colleagues investigate how this balance is mediatedfor HSCs in the BM and show that Robo4, a guidance molecule, is needed for appropriate HSC recruitment, homing, andmobilization. Furthermore, Robo4 appears to cooperate with the chemokine receptor CXCR4, and modifying the pair maybe needed to efficiently mobilize HSCs from donors and also to improve the seeding of transplants HSCs back to the BMof a recipient.Cell Stem Cell 8, January 7, 2011 ª2011 Elsevier Inc. xi
top related