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 found

upstream 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