You are on page 1of 36

Supported by

A booklet from Science, produced by the

AAAS Office of Publishing and Member Services
2 SCIENCE Stem Cells
3 BD Biosciences

4 Embryonic Stem Cell Lines Derived from Human Blastocysts
James A. Thomson, Joseph Itskovitz-Eldor, Sander S. Shapiro, Michelle A. Waknitz,
Jennifer J. Swiergiel, Vivienne S. Marshall, Jeffrey M. Jones
Science 6 November 1998 282: 1145-1147

7 Dening the Epithelial Stem Cell Niche in Skin

Tudorita Tumbar, Geraldine Guasch, Valentina Greco,
Cedric Blanpain, William E. Lowry, Michael Rendl, Elaine Fuchs
Science 16 January 2004 303: 359-363; published online 11 December 2003

11 Endothelial Cells Stimulate Self-Renewal and Expand

Neurogenesis of Neural Stem Cells
Qin Shen, Susan K. Goderie, Li Jin, Nithin Karanth, Yu Sun,
Natalia Abramova, Peter Vincent, Kevin Pumiglia, Sally Temple
Science 28 May 2004 304: 13381340; published online 1 April 2004

14 MicroRNAs Modulate Hematopoietic Lineage Differentiation

Chang-Zheng Chen, Ling Li, Harvey F. Lodish, David P. Bartel
Science 2 January 2004 303: 83-86; published online 4 December 2003

18 Development of a Human Adaptive Immune System in Cord

Blood CellTransplanted Mice
Elisabetta Traggiai, Laurie Chicha, Luca Mazzucchelli, Lucio Bronz,
Jean-Claude Piffaretti, Antonio Lanzavecchia, Markus G. Manz
Science 2 April 2004 304: 104-107

22 Derivation of Oocytes from Mouse Embryonic Stem Cells

Karin Hbner, Guy Fuhrmann, Lane K. Christenson, James Kehler,
Rolland Reinbold, Rabindranath De La Fuente, Jennifer Wood, Jerome F. Strauss III,
Michele Boiani, Hans R. Schler
Science 23 May 2003 300: 1251-1256; published online 1 May 2003

28 Consent from Donors for Embryo and Stem Cell Research
Bernard Lo, Vicki Chou, Marcelle I. Cedars, Elena Gates, Robert N. Taylor,
Richard M. Wagner, Leslie Wolf, Keith R. Yamamoto
Science 15 August 2003 301: 921


Stem cells can follow a variety of blueprints directing cell differentiation.
But the final edifice is not built of stone, and differentiated cells show
remarkable plasticity. [Image from the cover of Science 25 February 2000]

tem cell research represents an area with great scien- been successfully used in human therapy for decades. Adult
tific and therapeutic promise. Because stem cells are stem cells have been reported elsewhere in the body, such as
unique in their ability to self-renew and to differenti- the mammary gland, prostate, lung, epidermis, intestine, and
ate into multiple cell types, they can be used to elucidate in umbilical cord blood. Claims have been made that adult so-
normal cell processes as well as to understand the diseased matic stem cells, such as hematopoietic stem cells, can switch
condition. In addition, stem cells from individuals with to another cell fate (termed plasticity) or revert to a less differ-
various diseases can be utilized for biotechnology advance, entiated state. However, in vivo the number of hematopoietic-
for example, in drug screening. There is a widespread hope derived cells found in nonhematopoietic tissues, for example,
that these cells will one day be used to treat human injury is quite small. In addition, cell fusion events, in which a stem
and disease. If, for example, stem cells can be directed to cell combines with a more differentiated cell, explain some
differentiate into specialized nerve cells, they could be used observations where implanted stem cells display differentia-
for the repair of a severed spinal cord or replacement of tion markers of another cell lineage. Providing examples of
neurons destroyed by neurodegenerative conditions such as reversion to a less differentiated state, studies employing so-
Parkinsons disease. Directed differentiation of stem cells matic cell nuclear transfer (SCNT) in animals such as mice,
into muscle cells could facilitate repair of an injured heart pigs, and sheep have shown that somatic cell nuclei can be
after cardiac infarct; stem cells differentiated into insulin- reprogrammed when placed inside an enucleated oocyte. Em-
producing cells could treat patients with diabetes. bryonic stem cells derived from SCNT blastocysts can give rise
Despite these grand purposes, the mention of stem cell re- to cells of all three germ layers. Furthermore, if SCNT em-
search evokes varied images and viewpointsboth in favor bryos are implanted in an animal uterus, development proceeds
of and opposed to the work. The complexity of this topic through the embryonic and fetal stages to produce viable ani-
ranges from debate over usage of simple terms to heated dis- mals; however, this method generally is inefcient and some
agreement about multifaceted moral, ethical, and political developmental defects have been reported.
issues. The latter concerns demand stringent oversight and Due to the high degree of conservation of molecular and
public dialogue regarding funding issues, informed consent, cellular mechanisms among organisms, research with animal
patenting, and the status of the embryo. The articles in this model systems can provide biological insight for mammalian
collection represent a small sampling of exciting advances stem cells. The flatworm Planaria displays a striking regener-
in stem cell biology as they have been presented in Science ative capability. Other invertebrate model systems, including
magazine over the past few years. the fruitfly Drosophila and nematode C. elegans, and vertebrate
The term stem cell does not pertain to a single, dened cell models, such as zebrafish and mouse, are revealing conserved
population. Instead, the description applies to a large number genes, signaling pathways, and differentiation mechanisms.
of very different cells: embryonic stem cells, germline stem Before stem cells can be applied in human therapy, many
cells, and adult somatic stem cells, including stem cells from hurdles must be jumped. In general, there are relatively few
umbilical cord blood and progenitor cells such as those in the stem cells in the body. Methods for cell enrichment and iso-
hematopoietic and neural lineages. lation are being developed. It is also necessary to work out
As an organism grows, the developmental potency, or the methods for directed differentiation of a particular cell type as
range of possible cell fates, gradually decreases for most needed for study in cell culture or for possible eventual use in a
cells. However, within the organism, a supply of cells is therapeutic setting. Special culture methods (utilizing growth
needed for tissue regeneration and repair. Stem cells t the factors, signaling molecules, microRNAs, etc.) can coax cells
bill. By denition, stem cells self-renew and have the abil- down a specific differentiation pathway. Other obstacles that
ity to progress down multiple cell lineages. The number of must be eliminated include the depletion of animal products
cell types that can be obtained is variable, dependent on the (as in media components, isolation techniques, or feeder lines),
intrinsic properties of the stem cell population and as inu- prevention of karyotypic abnormalities, methods of controlled
enced by the cells natural microenvironment, or niche, or cell growth, mitigation of transplant rejection, and demonstra-
by special cell culture conditions. Fertilized eggs show the tion of safety and efficacy.
greatest degree of developmental potency. They are totipo- The largest gains can be expected by studying multiple
tent, that is, they can form an entire organism. Embryonic branches of stem cell science and varied model systems. It is
stem cells also display great developmental potency. They anticipated that stem cell research will eventually lead to the
are pluripotent and can differentiate into cells of all three use of these cells in human therapy to treat injury and disease.
germ layers, as well as the germline. A major gain in the With the ethical sensitivity of this line of study, regulatory
human embryonic stem cell arena was realized in 1998 with oversight and dialogue among researchers and the public are
the isolation of embryonic stem cells from human blasto- essential. Despite the hopes associated with stem cell research,
cysts. This work enabled research on nontransformed hu- experts caution that human therapeutic application lies well in
man cells with unlimited growth in tissue culture. the future, perhaps on the order of a decade or more. Until the
Stem cells in adult somatic tissues are generally thought numerous obstacles are eliminated, these cells are providing a
to progress through strict tissue-specic cell fates. He- great resource to reveal the secrets of the normal and diseased
matopoietic stem cells, which are capable of forming the state and for biotechnology advance.
different types of blood cells in the body, have been most
thoroughly investigated. We know the detailed lineage of he- Beverly A. Purnell
matopoietic cells, and hematopoietic progenitor cells have Senior Editor, Science
Stem cell research is a rapidly expanding area of investigation,
with the ultimate goal to prevent, diagnose, and treat
human diseases, including heart disease, diabetes, cancer,
stroke, and neurological disorders, such as Parkinsons and
Alzheimers disease.

BD Biosciences has provided integrated, BD Biosciences was first to commercialize Stem Cell Source
high-value products and services for cell systems and reagents for characterizing From isolation to analysis
culture and analysis for more than two and isolating hematopoietic stem cells Thousands of proven reagents and
decades with hallmark brands such as for research use. This, fueled by research systems for cell analysis
BD Falcon and BD FACS. BD FACS developments, has led to the now
Advanced functional assays
brand instruments are considered the gold well-accepted therapeutic approaches to
standard in the field of flow cytometry leukemia and lymphoma treatment. It is Dependable innovation in cell culture
and, along with BD Pharmingen and one of our continuous goals to build our
BD Transduction Laboratories brand successes and maintain our leadership Over 20 years of cell separation
antibodies, we offer one of the largest as a provider of research and diagnostic leadership
collections of reagents for applications tools for somatic and embryonic stem
including flow cytometric analysis, western cell research.
blotting, immunofluorescence microscopy,
and ELISA. Please visit us online at
With the addition of new instrument to get an overview of our current
platforms such as the BD Pathway product portfolio for stem cell
high-content imager and BD CARV II investigation.
confocal imager, we now provide one
of the most comprehensive portfolios Part of the BD Biosciences core focus
of cell culture and cell analysis tools to has been to foster collaborations to
the research community. BD Biosciences develop or license new reagents and
- Discovery Labware and its BD Falcon, technologies. If you would like to
BD BioCoat, and BD brands offer contribute to our expanding portfolio
tools for researchers investigating the of stem cell research products and
mechanisms that dictate stem cell growth services, please contact one of our
and differentiation, as well as stem cell use BD Biosciences Marketing or Business
for tissue engineering research applications. Development Departments at

BD Biosciences
For Research Use Only. Not for use in diagnostic or therapeutic procedures. Not for resale. 2350 Qume Drive
All applications are either tested in-house or reported in the literature. See Technical Data Sheets for details.
Prices are subject to change without notice. BD flow cytometers are class I (1) laser products. San Jose, CA 95131
BD, BD Logo and all other trademarks are the property of Becton, Dickinson and Company. 2006 BD
XX karyotype after 6 months of culture and
Embryonic Stem Cell
Cell Lines
Lines has now been passaged continuously for
more than 8 months (32 passages). A period
Derived from Human
Human of replicative crisis was not observed for
any of the cell lines.

The human ES cell lines expressed high
levels of telomerase activity (Fig. 2). Telo-
merase is a ribonucleoprotein that adds telo-
James A. Thomson,* Joseph Itskovitz-Eldor, Sander S. Shapiro, mere repeats to chromosome ends and is
Michelle A. Waknitz, Jennifer J. Swiergiel, Vivienne S. Marshall, involved in maintaining telomere length,
Jeffrey M. Jones which plays an important role in replicative
life-span (7, 8). Telomerase expression is
Human blastocyst-derived, pluripotent cell lines are described that have normal highly correlated with immortality in human
karyotypes, express high levels of telomerase activity, and express cell surface cell lines, and reintroduction of telomerase
markers that characterize primate embryonic stem cells but do not characterize activity into some diploid human somatic cell
other early lineages. After undifferentiated proliferation in vitro for 4 to 5 lines extends replicative life-span (9). Dip-
months, these cells still maintained the developmental potential to form tro- loid human somatic cells do not express tel-
phoblast and derivatives of all three embryonic germ layers, including gut omerase, have shortened telomeres with age,
epithelium (endoderm); cartilage, bone, smooth muscle, and striated muscle and enter replicative senescence after a finite
(mesoderm); and neural epithelium, embryonic ganglia, and stratied squamous proliferative life-span in tissue culture (10
epithelium (ectoderm). These cell lines should be useful in human develop- 13). In contrast, telomerase is present at high
mental biology, drug discovery, and transplantation medicine. levels in germ line and embryonic tissues
(14). The high level of telomerase activity
Embryonic stem (ES) cells are derived define primate ES cells. expressed by the human ES cell lines there-
from totipotent cells of the early mamma- Fresh or frozen cleavage stage human fore suggests that their replicative life-span
lian embryo and are capable of unlimited, embryos, produced by in vitro fertilization will exceed that of somatic cells.
undifferentiated proliferation in vitro (1, 2). (IVF) for clinical purposes, were donated The human ES cell lines expressed cell
In chimeras with intact embryos, mouse ES by individuals after informed consent and surface markers that characterize undifferen-
cells contribute to a wide range of adult after institutional review board approval. tiated nonhuman primate ES and human EC
tissues, including germ cells, providing a Embryos were cultured to the blastocyst cells, including stage-specific embryonic an-
powerful approach for introducing specific stage, 14 inner cell masses were isolated, tigen (SSEA)3, SSEA-4, TRA-l-60, TRA-1-
genetic changes into the mouse germ line and five ES cell lines originating from five 81, and alkaline phosphatase (Fig. 3) (4, 5,
(3). The term ES cell was introduced to separate embryos were derived, essentially 15, 16). The globo-series glycolipid GL7,
distinguish these embryo-derived pluripo- as described for nonhuman primate ES cells which carries the SSEA-4 epitope, is formed
tent cells from teratocarcinoma-derived (5, 6 ). The resulting cells had a high ratio by the addition of sialic acid to the globo-
pluripotent embryonal carcinoma (EC) of nucleus to cytoplasm, prominent nucle- series glycolipid Gb5, which carries the
cells (2). Given the historical introduction oli, and a colony morphology similar to that SSEA-3 epitope (17, 18). Thus, GL7 reacts
of the term ES cell and the properties of of rhesus monkey ES cells (Fig. 1). Three with antibodies to both SSEA-3 and SSEA-4
mouse ES cells, we proposed that the es- cell lines (H1, H13, and H14) had a normal (17, 18). Staining intensity for SSEA-4 on the
sential characteristics of primate ES cells XY karyotype, and two cell lines (H7 and human ES cell lines was consistently strong,
should include (i) derivation from the pre- H9) had a normal XX karyotype. Each of but staining intensity for SSEA-3 was weak
implantation or periimplantation embryo, the cell lines was successfully cryopre- and varied both within and among colonies
(ii) prolonged undifferentiated prolifera- served and thawed. Four of the cell lines (Fig. 3, D and C). Because GL7 carries both
tion, and (iii) stable developmental poten- were cryopreserved after 5 to 6 months of the SSEA-4 and SSEA-3 epitopes and be-
tial to form derivatives of all three embry- continuous undifferentiated proliferation. cause staining for SSEA-4 was consistently
onic germ layers even after prolonged cul- The other cell line, H9, retained a normal strong, the relatively weak staining for
ture (4). For ethical and practical reasons,
in many primate species, including humans,
the ability of ES cells to contribute to the Fig. 1. Derivation of the
H9 cell line. (A) Inner
germ line in chimeras is not a testable cell massderived cells
property. Nonhuman primate ES cell lines attached to mouse em-
provide an accurate in vitro model for un- bryonic broblast feed-
derstanding the differentiation of human er layer after 8 days of
tissues (4, 5). We now describe human cell culture, 24 hours be-
lines that fulfill our proposed criteria to fore rst dissociation.
Scale bar, 100 m. (B)
H9 colony. Scale bar,
J. A. Thomson, M. A. Waknitz, J. J. Swiergiel, V. S. 100 m. (C) H9 cells.
Marshall, Wisconsin Regional Primate Research Cen- Scale bar, 50 m. (D)
ter, University of Wisconsin, Madison, WI 53715, Differentiated H9 cells,
USA. J. Itskovitz-Eldor, Department of Obstetrics and cultured for 5 days in
Gynecology, Rambam Medical Center, Faculty of the absence of mouse
Medicine, Technion, Haifa 31096, Israel. S. S. Shapiro embryonic broblasts,
and J. M. Jones, Department of Obstetrics and Gyne- but in the presence of
cology, University of Wisconsin, Madison, WI 53715, human LIF (20 ng/ml;
USA. Sigma). Scale bar, 100
*To whom correspondence should be addressed. m.

SSEA-3 suggests a restricted access of the Fig. 2. Telomerase ex-
antibody to the SSEA-3 epitope. In common pression by human ES
with human EC cells, the undifferentiated cell lines. MEF, irradiat-
ed mouse embryonic
human ES cell lines did not stain for SSEA-1, broblasts used as a
but differentiated cells stained strongly for feeder layer for the
SSEA-l (15) (Fig. 3). Mouse inner cell mass cells in lanes 4 to 18;
cells, ES cells, and EC cells express SSEA-1 293, adenovirus-trans-
but do not express SSEA-3 or SSEA-4 (17, formed kidney epithe-
19), suggesting basic species differences be- lial cell line 293; MDA,
breast cancer cell line
tween early mouse and human development. MDA; TSR8, quantita-
The human ES cell lines were derived tion control template.
by the selection and expansion of individ- Telomerase activity
ual colonies of a uniform, undifferentiated was measured with
morphology, but none of the ES cell lines the TRAPEZE Telomer-
was derived by the clonal expansion of a ase Detection Kit (On-
cor, Gaithersburg, Maryland). The ES cell lines were analyzed at passages 10 to 13. About 2000 cells
single cell. The uniform undifferentiated were assayed for each telomeric repeat amplication protocol assay, and 800 cell equivalents were
morphology that is shared by human ES loaded in each well of a 12.5% nondenaturing polyacrylamide gel. Reactions were done in triplicate with
and nonhuman primate ES cells and the the third sample of each triplet heat inactivated for 10 to 15 min at 85C before reaction to test for
consistent expression by the human ES cell telomerase heat sensitivity (lanes 6, 9, 12, 15, 18, 21, 24, and 27). A 36base pair internal control for
lines of cell surface markers that uniquely amplication efciency and quantitative analysis was run for each reaction as indicated by the
characterize primate ES and human EC arrowhead. Data were analyzed with the Storm 840 Scanner and ImageQuant package (Molecular
Dynamics). Telomerase activity in the human ES cell lines ranged from 3.8 to 5.9 times that observed
cells make it extremely unlikely that a in the immortal human cell line MDA on a per cell basis.
mixed population of precursor cells was
expanded. However, because the cell lines
were not cloned from a single cell, we Fig. 3. Expression of
cannot rule out the possibility that there is cell surface markers by
some variation in developmental potential H9 cells. Scale bar,
100 m. (A) Alkaline
among the undifferentiated cells, in spite of phosphatase. (B) SSEA-
their homogeneous appearance. 1. Undifferentiated cells
The human ES cell lines maintained the failed to stain for SSEA-
potential to form derivatives of all three 1 (large colony, left).
embryonic germ layers. All five cell lines Occasional colonies
produced teratomas after injection into se- consisted of non-
stained, central, undif-
vere combined immunodeficient (SCID) ferentiated cells sur-
beige mice. Each injected mouse formed a rounded by a margin
teratoma, and all teratomas included gut of stained, differentiat-
epithelium (endoderm); cartilage, bone, ed, epithelial cells
smooth muscle, and striated muscle (meso- (small colony, right).
derm); and neural epithelium, embryonic (C) SSEA-3. Some
small colonies stained
ganglia, and stratified squamous epithelium uniformly for SSEA-3
(ectoderm) (Fig. 4). In vitro, the ES cells (colony left of center),
differentiated when cultured in the absence but most colonies con-
of mouse embryonic fibroblast feeder lay- tained a mixture of
ers, both in the presence and absence of weakly stained cells
human leukemia inhibitory factor (LIF) and a majority of non-
stained cells (colony
(Fig. 1). When grown to confluence and right of center). (D)
allowed to pile up in the culture dish, the SSEA-4. (E) TRA-1-60.
ES cell lines differentiated spontaneously (F) TRA-1-81. Similar
even in the presence of fibroblasts. After results were obtained
H9 cells were allowed to differentiate for 2 for cell lines H1, H7,
weeks, both -fetoprotein (350.9 14.2 H13, and H14.
IU/ml) and human chorionic gonadotropin
(hCG, 46.7 5.6 mIU/ml) were detected in limited number of sectioned embryos and of human ES cells to specific lineages
conditioned culture medium, indicating to analogies drawn from the experimental could identify gene targets for new drugs,
endoderm and trophoblast differentiation embryology of other species (21). Although genes that could be used for tissue regen-
(20). the mouse is the mainstay of experimental eration therapies, and teratogenic or toxic
Human ES cells should offer insights mammalian embryology, early structures compounds.
into developmental events that cannot be including the placenta, extraembryonic Elucidating the mechanisms that control
studied directly in the intact human embryo membranes, and the egg cylinder all differ differentiation will facilitate the efficient,
but that have important consequences in substantially from the corresponding struc- directed differentiation of ES cells to spe-
clinical areas, including birth defects, in- ture of the human embryo. Human ES cells cific cell types. The standardized produc-
fertility, and pregnancy loss. Particularly in will be particularly valuable for the study tion of large, purified populations of eu-
the early postimplantation period, knowl- of the development and function of tissues ploid human cells such as cardiomyocytes
edge of normal human development is that differ between mice and humans. and neurons will provide a potentially lim-
largely restricted to the description of a Screens based on the in vitro differentiation itless source of cells for drug discovery and

Fig. 4. Teratomas were passaged by exposure to type IV collagenase
formed by the human (1 mg/ml; Gibco-BRL) or by selection of individual
ES cell lines in SCID- colonies by micropipette. Clump sizes of about 50
beige mice. Human ES to 100 cells were optimal. Cell lines were initially
cells after 4 to 5 karyotyped at passages 2 to 7.

7. C. B. Harley, Mutat. Res. 256, 271 (1991).
months of culture (pas-
8. , H. Vaziri, C. M. Counter, R. C. Allsopp, Exp.
sages 14 to 16) from
Gerontol. 27, 375 (1992).
about 50% conuent 9. A. G. Bodnar et al., Science 279, 349 (1998).
six-well plates were in- 10. L. Hayick and P. S. Moorhead, Exp. Cell Res. 25, 581
jected into the rear leg (1961).
muscles of 4-week-old 11. R. C. Allsopp et al., Proc. Natl. Acad. Sci. U.S.A. 89,
male SCID-beige mice 10114 (1992).
(two or more mice per 12. C. M. Counter et al., EMBO J. 11, 1921 (1992).
cell line). Seven to eight 13. C. M. Counter, H. W. Hirte, S. Bacchetti, C. B. Harley,
weeks after injection, Proc. Natl. Acad. Sci. U.S.A. 91, 2900 (1994).
the resulting teratomas 14. W. E. Wright, M. A. Piatyszek, W. E. Rainey, W. Byrd,
were examined histo- J. W. Shay, Dev. Genet. 18, 173 (1996).
logically. (A) Gutlike 15. P. W. Andrews, J. Oosterhuis, I. Damjanov, in Terato-
structures. Cell line H9. carcinomas and Embryonic Stem Cells: A Practical
Scale bar, 400 m. (B) Approach, E. Robertson, Ed. (IRL, Oxford, 1987), pp.
Rosettes of neural epi- 207248.
thelium. Cell line H14. 16. Alkaline phosphatase was detected with Vector Blue
substrate ( Vector Labs). SSEA-1, SSEA-3, SSEA-4,
Scale bar, 200 m. (C)
TRA-1-60, and TRA-1-81 were detected by immuno-
Bone. Cell line H14.
cytochemistry with specic primary monoclonal an-
Scale bar, 100 m. (D) tibodies and localized with a biotinylated secondary
Cartilage. Cell line H9. antibody and then an avidin or biotinylated horse-
Scale bar, 100 m. (E) radish peroxidase complex (Vectastain ABC system;
Striated muscle. Cell Vector Laboratories) as previously described (5). The
line H13. Scale bar, 25 ES cell lines were at passages 8 to 12 at the time
m. (F) Tubules inter- markers were analyzed.
spersed with struc- 17. R. Kannagi et al., EMBO J. 2, 2355 (1983).
tures resembling fetal 18. R. Kannagi et al., J. Biol. Chem. 258, 8934 (1983).
glomeruli. Cell line H9. 19. D. Solter and B. B. Knowles, Proc. Natl. Acad. Sci.
Scale bar, 100 m. U.S.A. 75, 5565 (1978).
20. hCG and -fetoprotein were measured by specic
radioimmunoassay (double AB hCG and AFP-TC kits;
Diagnostic Products, Los Angeles, CA). hCG assays
transplantation therapies. Many diseases, such References and Notes used the World Health Organization Third Interna-
as Parkinsons disease and juvenile-onset dia- 1. M. Evans and M. Kaufman, Nature 292, 154 (1981). tional Standard 75/537. H9 cells were allowed to
betes mellitus, result from the death or dysfunc- 2. G. Martin, Proc. Natl. Acad. Sci. U.S.A. 78, 7634 grow to conuence (day 0) on plates of irradiated
tion of just one or a few cell types. The replace- 3. A. Bradley, M. Evans, M. Kaufman, E. Robertson, Na-
mouse embryonic broblasts. Medium was replaced
ment of those cells could offer lifelong treat- daily. After 2 weeks of differentiation, medium in
ture 309, 255 (1984).
triplicate wells conditioned for 24 hours was assayed
ment. Strategies to prevent immune rejection of 4. J. A. Thomson and V. S. Marshall, Curr. Top. Dev. Biol. for hCG and -fetoprotein. No hCG or -fetoprotein
the transplanted cells need to be developed but 38, 133 (1998).
was detected in unconditioned medium.
5. J. A. Thomson et al., Proc. Natl. Acad. Sci. U.S.A. 92,
could include banking ES cells with defined 7844 (1995).
21. R. ORahilly and F. Muller, Developmental Stages in
major histocompatibility complex back- Human Embryos (Carnegie Institution of Washington,
6. Thirty-six fresh or frozen-thawed donated human
Washington, DC, 1987).
grounds or genetically manipulating ES embryos produced by IVF were cultured to the
22. G. Bain, D. Kitchens, M. Yao, J. E. Huettner, D. I.
cells to reduce or actively combat immune blastocyst stage in G1.2 and G2.2 medium (25).
Fourteen of the 20 blastocysts that developed Gottlieb, Dev. Biol. 168, 342 (1995).
rejection. Because of the similarities to hu- were selected for ES cell isolation, as described for 23. M. V. Wiles and G. Keller, Development 111, 259
mans and human ES cells, rhesus monkeys rhesus monkey ES cells (5). The inner cell masses (1991).
were isolated by immunosurgery (26), with a rabbit 24. M. G. Klung, M. H. Soonpaa, G. Y. Koh, L. J. Field,
and rhesus ES cells provide an accurate J. Clin. Invest. 98, 216 (1996).
antiserum to BeWO cells, and plated on irradiated
model for developing strategies to prevent (35 grays gamma irradiation) mouse embryonic 25. D. K. Gardner et al., Fertil. Steril. 69, 84 (1998).
immune rejection of transplanted cells and broblasts. Culture medium consisted of 80% Dul- 26. D. Solter and B. Knowles, Proc. Natl. Acad. Sci. U.S.A.
for demonstrating the safety and efficacy of beccos modied Eagles medium (no pyruvate, 72, 5099 (1975).
high glucose formulation; Gibco-BRL) supplement- 27. We thank the personnel of the IVF clinics at the
ES cell based therapies. Substantial ad- ed with 20% fetal bovine serum (Hyclone), 1 mM University of Wisconsin School of Medicine and at
vances in basic developmental biology are glutamine, 0.1 mM -mercaptoethanol (Sigma), the Rambam Medical Center for the initial culture
required to direct ES cells efficiently to and 1% nonessential amino acid stock (Gibco-BRL). and cryopreservation of the embryos used in this
After 9 to 15 days, inner cell mass derived out-
lineages of human clinical importance. growths were dissociated into clumps either by
study; D. Gardner and M. Lane for the G1.2 and
However, progress has already been made G2.2 media; P. Andrews for the NTERA2 cl.D1 cells
exposure to Ca2/Mg2-free phosphate-buffered
and the antibodies used to examine cell surface
in the in vitro differentiation of mouse ES saline with 1 mM EDTA (cell line H1), by exposure
markers; C. Harris for karyotype analysis; and
to dispase (10 mg/ml; Sigma; cell line H7), or by
cells to neurons, hematopoietic cells, and mechanical dissociation with a micropipette (cell Geron Corporation for the 293 and MDA cell pel-
cardiac muscle (2224). Progress in basic lines H9, H13, and H14) and replated on irradiated lets and for assistance with the telomerase TRAP
developmental biology is now extremely mouse embryonic broblasts in fresh medium. In- assay. Supported by the University of Wisconsin
dividual colonies with a uniform undifferentiated (UIR grant 2060) and Geron Corporation (grant
rapid; human ES cells will link this 133-BU18).
morphology were individually selected by micropi-
progress even more closely to the preven- pette, mechanically dissociated into clumps, and
tion and treatment of human disease. replated. Once established and expanded, cultures 5 August 1998; accepted 7 October 1998

LRCs initiate each new follicle, and that upon
Dening the Epithelial Stem Cell exit their progeny rapidly proliferate, change
biochemistry, and regenerate all differentiated
cell types.
Niche in Skin To determine whether bulge LRCs can
react to injury, we scratch-wounded 8-week-
Tudorita Tumbar, Geraldine Guasch, Valentina Greco, old chased mice. Within 24 to 48 hours,
Cedric Blanpain, William E. Lowry, Michael Rendl, Elaine Fuchs* GFPhigh cells were detected outside the bulge
(Fig. 1E). These were not scattered bulge
Many adult regenerative cells divide infrequently but have high proliferative cells, because they localized to infundibulum
capacity. We developed a strategy to uorescently label slow-cycling cells in and displayed underlying basement mem-
a cell typespecic fashion. We used this method to purify the label-retaining brane (antilaminin 5 immunoreactivity; not
cells (LRCs) that mark the skin stem cell (SC) niche. We found that these cells shown). Not seen in unwounded skin, GFP-
rarely divide within their niche but change properties abruptly when stimulated positive cells within infundibulum and epider-
to exit. We determined their transcriptional prole, which, when compared to mis expressed nuclear junB, a stress-response
progeny and other SCs, denes the niche. Many of the 100 messenger RNAs protein (Fig. 1E). Thus, in response to wound
preferentially expressed in the niche encode surface receptors and secreted stimuli, LRCs change their biochemistry, exit
proteins, enabling LRCs to signal and respond to their environment. the bulge, migrate, and proliferate to repopu-
late infundibulum and epidermis. The ability
Epidermis and its appendages undergo contin- these mRNAs are found in SCs of other tis- of LRCs to regenerate hair follicles and epi-
uous renewal and maintain reservoirs of mul- sues, whereas others specify the unique envi- dermis is a feature characteristic of bulge SCs
tipotent SCs whose descendants are organized ronment of the skin SC niche. (6, 7).
spatially and temporally. The epidermal basal To mark infrequently cycling cells of adult Immunofluorescence microscopy revealed
layer (BL) contains putative SCs in addition skin epithelium, we engineered transgen- that the zone harboring keratinocyte-specific,
to the transiently amplifying (TA) cells, which ic mice to express histone H2Bgreen fluo-
H2B-GFPbright LRCs was more restricted
give rise to terminally differentiating supra- rescent protein (GFP) (8) controlled by a tetra-
than that defined by known bulge-preferred
basal layers (13). The BL and the hair follicle cycline-responsive regulatory element (TRE).
markers, including K15, K19, 6-integrin,
outer root sheath (ORS) are contiguous and A tightly regulated TRE-mCMVH2B-GFP
founder animal was crossed with mice har- 1- integrin, CD34, S100A4, and S100A6
biochemically similar (fig. S1A). In the hair
boring a keratin 5 (K5) promoter driven tet (1114) (fig. S3). To further analyze the prop-
bulb, the dermal papilla (DP) maintains con-
repressorVP16 transgene (9), and offspring erties of LRCs, we prepared single-cell sus-
tact with matrix TA cells until they differenti-
ate to form the inner root sheath (IRS) and hair were selected for doxycycline (Tet)con- pensions from 8-week-old, chased transgenic
shaft. Follicles periodically undergo cycles of trolled regulation restricted to skin epithelium skins and performed fluorescence-activated
growth (anagen), destruction (catagen), and (Fig. 1A). Without Tet, backskin epithelial cell sorting (FACS). Of this population, 12%
rest (telogen). The zone between noncycling cells exhibited ~103 to 104 units of GFP fluo- displayed 10 to 104 units of GFP fluorescence,
and cycling segments is a SC niche, the ORS rescence (9, 10) (Fig. 1B). After feeding 4- with 1 to 2% exhibiting 103 to 104 units, rela-
bulge (4, 5). week-old mice Tet for 4 weeks to 4 months tive to background (Fig. 2A).
Multipotent epithelial SCs with high prolif- (chase), only bulge cells (1% total) retained Populations gated at 103 to 104 units
erative potential reside in the bulge (6, 7). The fluorescence at 103 units (Fig. 1B) (fig. S1). (GFPhigh) and 50 to 100 units (GFPlow) exclud-
bulge contains the majority of infrequently cy- Independent of cell surface markers, our ap- ed propidium iodide and exhibited surface
cling, label-retaining cells (LRCs), which can proach marks putative bulge SCs on the basis 4-, 1-, and 6-integrins, typical of BL/ORS
respond to anagen DP signals to regenerate of their slow-cycling properties. Although cells (Fig. 2B). GFPhigh cells were enriched in
the follicle. After wounding or transplantation, we used keratinocyte-specif ic Tet off VP16 CD34 (14), whereas GFPlow cells had more
bulge cells give rise to epidermis, follicles, mice, TRE-mCMV-H2B-GFP animals could CD71 (downregulated in bulge) (13). Two-col-
and sebaceous glands. Additionally, when dis- be used with mice expressing other promoter/ or analyses indicated that only ~30% of GFP+/
sected from rat whiskers and cultured, bulge enhancer-driven, Tet-regulatable activators/re- CD34+ (also 6+) cells were GFPhigh (Fig. 2B)
cells yield more colonies than other follicle pressors to isolate LRCs from other tissues. (fig. S4B). Semiquantitative fluorescence
segments (7). To track LRC fate during hair cycling, documented that GFPhigh cells corresponded
It is not known what features define this spe- we monitored GFP fluorescence intensities in fluorescence intensity to bulge cells, where-
cialized SC niche, what its interactions with relative to the proliferation-associated mark- as GFPlow fluorescence placed them outside
bulge LRCs are, and whether all LRCs are SCs. ers Ki67, phosphorylated histone H3 (P-H3; the niche (fig. S4A).
To begin to address these issues, we devised G /M) and basonuclin (BSN) (Fig. 1, B and
2 Although GFPhigh and GFPlow cells differed
a strategy based on the prediction that bulge C). Throughout the cycle, most LRCs re- in CD34/CD71 expression, they both ex-
SCs are uniquely both slow-cycling and ac- mained in the bulge as GFP-bright and Ki67, pressed BL/ORS keratins K5, K14, and K15
tive for a keratinocyte-specific promoter. With P-H3, and BSNlow cells (Fig.1, B and C). (1, 12), but not differentiation-specific K1
this strategy, we purify and characterize bulge During anagen, an abrupt switch in prolif- (1) (Fig. 2C). FACS by surface-4 yielded a
LRCs and related keratinocyte progeny in the eration markers and H2B-GFP intensity oc- larger pool of cells that were similar to GFPlow
BL and ORS. Analyses of their transcriptional curred at the transition between bulge and but with lower fluorescence (fig. S4B). Im-
profiles reveal the skin LRC mRNAs; some of new follicle downgrowth (large arrowheads). munofluorescence with six markers indicated
Overexposure verified that this downgrowth that these three FACS populations were >90%
Howard Hughes Medical Institute, Laboratory of
was largely GFP-positive deriving from bulge homogeneous (fig. S4B), and semiquantitative
Mammalian Cell Biology and Development, Rock- LRCs. Newly created GFP-positive popula- reverse transcription polymerase chain reac-
efeller University, New York, NY 10021, USA. tions included ORS (K5+) matrix (Lef1+/K5), tions (RT-PCRs) documented their distinctive
*To whom correspondence should be addressed. E- hair (AE13+/K5), and IRS (GATA-3+/K5) characteristics (fig. S4C). Cell cycle profiles
mail: (Fig. 1D). We conclude that only a few bulge showed that only 0.5% GFPhigh LRCs were

Fig. 1. System for marking slow-cycling SCs in vivo and monitoring their are retained in the bulge; their progeny rapidly divide, diluting H2B-GFP.
fate. (A) Strategy. (B to D) Skin sections of mice before and after 4-week (D) Early anagen II bulb overexposed for GFP and double-labeled (small
chase. Shown are epiuorescence of H2B-GFP (green) and 4,6- arrowheads) with Abs against each differentiation cell type. (E) Mice
diamidino-2-phenylindole (DAPI) (blue), and indirect immunouores- after chase were scratch-wounded and analyzed by immunouores-
cence with antibodies (Abs) indicated (Texas Red). The hair cycle stage is cence. Arrows denote likely directions of movements of GFP-positive
indicated on each set of after chase frames (see also g. S1, B to D, and LRCs and progeny. Abbreviations: Bu, bulge; DP, dermal papilla; Mx,
g. S2). Arrows (B) denote Ki67 sebaceous gland cells in telogen. matrix; hg, hair germ; Ep, epidermis; asterisk, hair shaft (autouorescent);
Arrowheads [(B) and (C)] denote transition zone between bulge and newly hf, hair follicle; Cx, cortex; ORS/IRS, outer/inner root sheaths; BM,
generated follicle downgrowth. Late anagen (Ki67 in red): GFP-bright cells basement membrane; In, infundibulum; W, wound. Scale bars, 50 m.

in G /M (Fig. 2D). Finally, although existing

2 ferred location (Fig. 3, B and C). Candidates to and transcription factors. This newfound rela-
methods did not permit long-term culturing of be involved in SC maintenance and/or activa- tion between bulge LRCs and other SC popu-
adult murine bulge cells (7, 14), GFPhigh LRCs tion, the shared LRC SC factors encompassed lations (15, 16) opens important avenues for
were highly enriched (>10 to 15) for cells proteins regulating cell growth and survival; future investigation.
forming colonies, some of which were >500 receptors able to sense and respond to growth As shown in table S2, 154 mRNAs were
cells (fig. S4D), consistent with the high pro- factors, hormones, and extracellular matrix; up-regulated by a factor of 2 in all four com-
liferative capacity documented for rat whisker
bulge (7).
Using microarray analyses, we obtained
transcriptional profiles for the three popula-
tions, and high-stringency analyses uncovered
the distinguishing features of LRCs (table S2
and Materials and Methods); ~4800 of 12,000
mRNAs were scored as present in each popu-
lation. When bulge LRCs were compared with
SC databases from hematopoietic (HSC),
embryonic (ESC), and neuronal (NSC) tis-
sues (15, 16), SCs were found to express 68%
of mRNAs present in LRCs and ~40% of
mRNAs up-regulated in LRCs relative to BL/
ORS (Fig. 3A). Moreover, overlap existed be-
tween HSCs and LRCs relative to their respec-
tive progenies (table S1).
The complete database with the raw
Affymetrix data files is available at www.
Up-regulated skin LRC mRNAs included
known SC markers such as stem cell fac- Fig. 2. Isolation and preliminary characterization of bulge LRCs and progeny. Animals were Tet-fed
tor (kit ligand), Dab2, ephrin tyrosine kinase for 4 weeks beginning at t 4 weeks. (A) FACS analyses of single-cell suspensions of skins. GFP
receptors Ephs), tenascin C (Tnc), inter- uorescence (FL) is in arbitrary units. (B) Two-color FACS analyses for GFP and ve surface markers.
leukin-11 receptor, Id binding protein2 6/CD34 data illustrate that GFPhigh LRCs represent only 30% of 6/CD34/K5-H2B-GFPpositive
cells. White, without primary Abs; red, with Abs. Percentages of total cells scoring positive are
(Idb-2), four-and-a-half lim domains (Fhl1), indicated. (C) GFPhigh FACS population analyzed by immunouorescence to illustrate homogeneity.
CD34, S100A6, and growth arrestspecific Figure S4B provides quantication of all three fractions screened for six markers. (D) Propidium
(Gas) proteins (1721). Immunofluorescence iodide (PI)FACS cell cycle proles by DNA content: G0 /G1 (n 1), G2/M (n 2), and S (n 1
and/or RT-PCR confirmed their bulge-pre- or 2). Percentage of total cells in G2/M is indicated.
Fig. 3. Transcriptional proling of bulge LRCs relative to other SCs. Duplicate called present in three other SCs (black), two other SCs (red), one other SC
mRNAs of bulge LRCs and their two progeny populations were amplied and (yellow), and no other SCs (white). (B) GFP epiuorescence and immuno-
hybridized to Affymetrix oligonucleotide chips. Files were analyzed by Mi- uorescence of skin sections from 8-week-old mice after 4-week chase. Abs
croarray Suite (MAS5.0) Affymetrix software followed by public database are against known SC markers found up-regulated in bulge LRCs. Abbrevi-
searches, functional annotation, and comparison with similar databases from ations are as in Fig. 1 legend; scale bars, 50 m. Lower magnication for
embryonic (ESC), neural (NSC), and hematopoietic (HSC) SCs (15, 16). (A) tenascin-C illustrates marker-specicity. (C) Semiquantitative RT-PCR.
Bar graphs: Percentage of mRNAs called present in skin LRCs (4839 total; left) Probes are for SC markers, up-regulated in skin LRCs. Mouse ESC mRNAs are
and of mRNAs increased in LRCs versus BL /ORS (154 total; right) and also shown for comparison. PCR was run for 29, 32, and 35 cycles.

parisons of bulge LRCs to GFPlow and 4-posi- thy of special mention. Consistent with their tors, including Srfp1, Dab2, Dkk3, Ctbp2,
tive BL/ORS progeny; Table 1 shows function- slow-cycling properties, skin LRCs expressed Tcf3, and Tle1, and the Wnt receptors Fzd3,
al classifications for a subset of these mRNAs. elevated transcripts encoding cell cycle F2d7, and Fzd2 (Table1) (fig. S5). Conversely,
With 25 primer sets, RT-PCR verified up-reg- regulatory proteins, and in particular, keratin- Wnt3a and Wnt3 were down-regulated >3 in
ulation of 24 putative LRC mRNAs relative to ocyte growth inhibitors implicated transform- LRCs. Consistent with Tcf3s repressor func-
at least one of the two progeny (fig. S5). Many ing growth factor (TGF) signaling (25). One tion (24), a Wnt-inhibited niche would explain
known bulge markers surfaced as up-reg- of these, LTBP-1, is necessary for latent TGF why at most stages in the hair cycle, the Wnt
ulated LRC transcripts, including CD34 (9), activation (26) and was strongly and specifi- reporter gene TOPGAL is silent in the bulge
S100A4 (5) (22), S100A6 (3) (22), Barx2 cally localized to bulge (Fig. 4B). Indicative (29).
(2) (23), and Tcf3 (3) (24) (Table 1). Im- of TGF receptor activation, nuclear phospho- Many bulge LRCup-regulated mRNAs
munofluorescence conf irmed their bulge- Smad2 immunoreactivity was more prevalent (43%) encoded secretory or integral mem-
preferred location relative to BL/ORS (fig. in bulge than progeny (Fig. 4B). Activated brane proteins (Fig. 5), which suggests the
S3). Most LRC mRNAs were specif ically TGF/phosphoSmad target genes and/or ability of skin LRCs to organize their
expressed in bulge relative to upper ORS and Smad interacting proteins were also included niche, communicate with neighboring cells,
BL. Some were exclusive for the bulge within in up-regulated LRC mRNAs (Table 1) (fig. and respond to their special environment. A
the skin. Some were present not only in bulge, S5) (27). Conversely, transcripts down-regu- case in point may be ephrin receptors ( Ephs)
but also other skin cells not analyzed here. An lated (163 total) encoded many proliferation- and their membrane-bound ligands (Efns),
example tektin2 (14), a putative microtubule- associated proteins, including Ki67 (3) (Fig. which signal bidirectionally in cell-cell com-
binding protein, increased in bulge relative 1, B and C), Cdc25C (2), and N-myc down- munication and tissue boundary formation
to BL/ORS and also present in arrector pili stream-regulated-like (2). (21). Although not specific, EfnB1, EphA4,
muscles (Fig. 4A). Additional up-regulated LRC mRNAs and EphB4 were expressed in LRCs (Fig. 4C).
Table 1 groups mRNAs into categories use- encoded members of the Wnt pathway, essen- EfnB1 was up-regulated further in matrix, a
ful in considering the properties of LRCs rel- tial for follicle morphogenesis and hair cycle compartment not analyzed here, and EphA4
ative to progeny cells. Several points are wor- activation (28, 29). These were inhibi- and EphB4 were also in a subset of lower ORS

Fig. 4. Implementation of array

analyses to examine characteristics
and dynamics of the skin SC niche.
GFP (green) and immunouores-
cence (red) of skin sections from
8-week-old mice (4-week chase).
Examples shown: (A) An mRNA
up-regulated 2 in LRCs relative
to epidermis/ORS. (B) Activated
(nuclear phospho-Smad2; arrow-
heads) or up-regulated (LTBP-1)
LRC factors involved in TGF sig-
naling. Quantication is at right
(graph). (C) Tissue polarity proteins
expressed in the SC niche. EphA4,
EphB4, and EfnB1 (right of EphB4);
boxed bulge in frame is also shown
with GFP colabeling. (D) Dynamics
of the niche during cycles of SC
activation (telogen/anagen transi-
tion). Abbreviations are as in Fig. 1
legend; APMu, arrector pili muscle;
asterisk denotes hair shaft
autouorescence. Scale bars, 50 m.
Table 1. Transcriptional proling of bulge LRCs relative to their BL/ORS progeny. Functional classication is shown for 66 mRNAs scored as increased in bulge
LRCs relative to BL/ORS progeny (full list in table S2). Average relative increase across the four comparisons is in parentheses; an mRNA that is present but not
increased is denoted P.

Category mRNAs

Known bulge factors Cd34 (9)*, S100a4 (5)*, S100a6 (3)*, Tcf3 (3)*, 1-integrin (P)*, 4-integrin (P)*, 6-integrin (P)*, Barx2 (2)*
Cell cycle Gas1 (growth arrest specic 1) (4), Ltbp1 (latent TGF binding protein) (8)*, Ltpb2 (10), Ltbp3 (3), TGF2 (3)*,
Inhbb (3), Ak1 (3)
TGF-induced factors Idb1 (2), Idb2 (8), Idb3 (2), Idb4 (4), Ctgf (8)*, Ltbp1 (8)*, Ltbp2 (10), Ltbp3 (3), Igfbp5 (6),* Igfbp7 (2)*,
Timp2 (5), 6-integrin (6)*, Tnc (3)*, EfnB1 (2)
Wnt signaling Sfrp1 (7), Dab2 (9), Dkk3 (5), Fzd2 (5), Fzd3 (3), Fzd7 (4), Ctbp2 (2), Fts (2), Tcf3 (3)*
Other signal transduction/ Stem cell factor (Kit-l) (2), EfnA4 (2), EfnB1 (2), EfnB2 (2), Bdnf (8)*, Tpst1 (3), Ptprk (3), Ppap2a (8),
cell-cell communication factors Gkap42-pending (6), Igfbp5 (6)
Extracellular matrix/basement Tnc (3)*, Col6a1 (3)*, Col18a1 (2)*, Mtn2 (2), Timp2 (4)*, Bgn (2), Agn (2), Sdc1 (2)*,
membrane proteins syndecan bp (2)
Nuclear proteins/ Idb1 (2), Idb2 (8), Idb3 (2), Idb4 (4), Nfatc1 (2), Osf (5), Pbx3 (4), Nb (3)*, Hist1h2bc (3), vdr (2)*,
transcription factors Mad4 (2)*, Max (2)*, LIM domain only 1 (2), Cited2 (2)
Cytoskeletal/cell adhesion Macf1 (ACF7) (3)*, Tekt2 (14), 6-integrin (6)*, Pdlim3 (15), Actn1 (4), Myo1b (4), Pfn2 (3), Krt2-6a (7)*,
factors Ndn (3)
*mRNAs previously reported in skin, irrespective of location.

, , , p
Fig. 5. Comparison of cellu- divide and are mobilized from their niche. ment 128, 2485 (2001).
lar localization of bulge LRC 19. B. A. Hocevar et al., EMBO J. 22, 3084 (2003).
In summary, we have uncovered a con-
mRNAs increased relative 20. A. V. Molofsky et al., Nature 425, 962 (2003).
to BL /ORS and mRNA stellation of distinguishing features of bulge 21. K. Kullander, R. Klein, Nature Rev. Mol. Cell Biol. 3,
present in LRCs. Left, in- LRCs relative to related keratinocyte progeny, 475 (2002).
creased in LRCs relative to which, together with their localization, likely 22. M. Ito, K. Kizawa, J. Invest. Dermatol. 116, 956 (2001).
BL /ORS (154 total); right, 23. G. Sander et al., J. Invest. Dermatol. 115, 753 (2000).
accounts for their special properties. Our 24. B. J. Merrill, U. Gat, R. DasGupta, E. Fuchs, Genes Dev.
present in skin LRCs (three findings suggest that the bulge SC niche is a 15, 1688 (2001).
pools of 150 mRNAs each 25. A. Iavarone, J. Massague, Mol. Cell. Biol. 19, 916 (1999).
growth and differentiationrestricted environ-
of the 4839 present were 26. Z. Isogai et al., J. Biol. Chem. 278, 2750 (2003).
analyzed). Black, expressed ment. The LRC-related changes that have thus 27. Y. C. Yang et al., Proc. Natl. Acad. Sci. U.S.A. 100,
sequence tags; red, intracel- far surfaced are already suggestive of a broad- 10269 (2003).
lular/cytosolic; blue, nucle- er interaction between environmental stimuli 28. T. Andl, S. T. Reddy, T. Gaddapara, S. E. Millar, Dev.
ar; gray, integral to mem- and the SC niche. Cell 2, 643 (2002).
brane; white, secreted. 29. L. Alonso, E. Fuchs, Genes Dev. 17, 1189 (2003).
30. T. Tumbar et al., data not shown.
References and Notes 31. V. A. Botchkarev, M. Yaar, B. A. Gilchrest, R. Paus,
1. E. Fuchs, S. Raghavan, Nature Rev. Genet. 3, 199 (2002). J. Invest. Dermatol. 120, 168 (2003).
2. C. S. Potten, R. J. Morris, J. Cell Sci. Suppl. 10, 45 (1988). 32. Y. Soma et al., Arch. Dermatol. Res. 293, 609 (2002).
cells in full anagen (Fig. 4C) (30). 3. I. C. Mackenzie, J. Invest. Dermatol. 109, 377 (1997). 33. K. Willert et al., Nature 423, 448 (2003).
Although most markers labeled bulge 4. G. Cotsarelis, T. T. Sun, R. M. Lavker, Cell 61, 1329 34. T. Reya et al., Nature 423, 409 (2003).
(1990). 35. We thank L. Degenstein, J. Fan, and L. Polak for help
LRCs irrespective of whether follicles were in 5. R. J. Morris, C. S. Potten, J. Invest. Dermatol. 112, 470 with mice; A. Glick for the K5Tetoff mice; S. Mazel and
anagen or telogen, Bdnf and TGF2 changed (1999). T. Shengelia for ow cytometry; and all who assisted
LRC expression with the hair cycle (31, 32). 6. G. Taylor, M. S. Lehrer, P. J. Jensen, T. T. Sun, R. M. with reagents (see supporting online material). Af-
Transient stimuli mediated by DP may be par- Lavker, Cell 102, 451 (2000).
fymetrix hybridizations were conducted by the HHMI
7. H. Oshima, A. Rochat, C. Kedzia, K. Kobayashi, Y.
ticularly important in influencing signaling Stanford Microarray facility. Supported by postdoctoral
Barrandon, Cell 104, 233 (2001).
pathways within the niche. Another example fellowships from the Life Sciences Foundation (T.T.),
8. T. Kanda, K. F. Sullivan, G. M. Wahl, Curr. Biol. 8, 377
(1998). Human Frontier Science Program (G.G., C.B.), EMBO
is 6-integrin, present only in early anagen (V.G.), Austrian Science Foundation (M.R.), BAEF and
9. I. Diamond, T. Owolabi, M. Marco, C. Lam, A. Glick,
and not telogen (Fig. 4D). v6 makes an at- J. Invest. Dermatol. 115, 788 (2000). NATO (C.B.), and NIH (W.E.L.). E.F. is an Investigator of
tractive candidate for SC activation/migration, 10. V. Vasioukhin, L. Degenstein, B. Wise, E. Fuchs, Proc. the Howard Hughes Medical Institute.
as it uses tenascin-C as ligand and is activated Natl. Acad. Sci. U.S.A. 96, 8551 (1999).
11. P. H. Jones, S. Harper, F. M. Watt, Cell 80, 83 (1995).
during skin wounding and tumorigenesis. Sim- 12. S. Lyle et al., J. Investig. Dermatol. Symp. Proc. 4, 296
ilarly, although not yet formally tested, DP-in- Supporting Online Material
duced changes in Wnt signaling could explain 13. H. Tani, R. J. Morris, P. Kaur, Proc. Natl. Acad. Sci.
Materials and Methods
U.S.A. 97, 10960 (2000).
why TOPGAL is transiently activated in the 14. C. S. Trempus et al., J. Invest. Dermatol. 120, 501 (2003). Figs. S1 to S5
bulge at early anagen (29). If signaling path- 15. M. Ramalho-Santos, S. Yoon, Y. Matsuzaki, R. C. Mul- Tables S1 and S2
ways (e.g., Wnts) are generally important in ligan, D. A. Melton, Science 298, 597 (2002).
16. N. B. Ivanova et al., Science 298, 601 (2002). 9 October 2003; accepted 21 November 2003
SC self-renewal, as they are in hematopoietic Published online 11 December 2003;
17. C. Sette, S. Dolci, R. Geremia, P. Rossi, Int. J. Dev. Biol.
SCs (33, 34), then differences in their status 44, 599 (2000). 10.1126/science.1092436
could have an impact on rates at which SCs 18. E. Garcion, A. Faissner, C. French-Constant, Develop- Include this information when citing this paper.

suggested that they form a niche for neural
Endothelial Cells Stimulate stem cells (2).
To examine a possible functional inter-
action, we cocultured neural and vascular
Self-Renewal and Expand cells (Fig. 1A). Neural stem cells from
mouse cerebral cortex from embryonic day
Neurogenesis of Neural Stem Cells 10 to 11 (E10-11) were plated at clonal
density on the base of culture wells. The
Qin Shen,1 Susan K. Goderie,1 Li Jin,1 Nithin Karanth,1 Yu Sun,1 upper transwell compartment was seeded
Natalia Abramova,1 Peter Vincent,2 Kevin Pumiglia,3 Sally Temple1* with purified vascular-associated or other
feeder cells: primary bovine pulmonary ar-
Neural stem cells are reported to lie in a vascular niche, but there is no direct tery endothelial (BPAE) cells, a mouse
evidence for a functional relationship between the stem cells and blood vessel brain endothelial (MbEND) cell line, vas-
component cells. We show that endothelial cells but not vascular smooth cular smooth muscle (VSM) cells, NIH3T3
muscle cells release soluble factors that stimulate the self-renewal of neural fibroblasts, or as a control, high-density
stem cells, inhibit their differentiation, and enhance their neuron production. age-matched cortical cells (CTX). CD31
Both embryonic and adult neural stem cells respond, allowing extensive pro- (platelet endothelial cell adhesion molecule
duction of both projection neuron and interneuron types in vitro. Endothelial c-1, PECAM-1) endothelial cells were nev-
coculture stimulates neuroepithelial cell contact, activating Notch and Hes1 to er found in the lower compartment when
promote self-renewal. These ndings identify endothelial cells as a critical BPAE or MbEND cells were plated in
component of the neural stem cell niche. the transwell upper compartment (Fig. 1B),
confirming that the feeder cells could not
Stem cell expansion and differentiation In the adult, neural stem cells lie close to migrate through the 0.4-m-diameter
are regulated in vivo by environmental fac- blood vessels: in the hippocampus (2), the membrane pores.
tors encountered in the stem cell niche (1). subventricular zone (SVZ) (3), and the As expected (6 ), embryonic stem cell
songbird higher vocal center (4 ). In the clones cocultured with CTX began produc-
developing central nervous system (CNS), ing neurons within a day. Most neuron
Center for Neuropharmacology and Neuroscience,
Center for Cardiovascular Sciences, 3Center for Cell
ventricular zone cells produce vascular en- production was over by 7 days, and growth
Biology and Cancer Research, Albany Medical College, dothelial growth factor, which attracts after this time was largely in glial lineages.
Albany, NY 12208, USA. vessel growth toward them (5). Thus, vas- Clones cocultured with BPAE or MbEND
*To whom correspondence should be addressed. E- cular cells are close to CNS germinal zones cells behaved differently (Fig. 1D and fig.
mail: throughout life (fig. S1), and it has been S2), growing into sheets of largely flat-

Fig. 1. Endothelial cell derived soluble factors

stimulate cortical stem cell expansion and delay
differentiation. (A) The coculture system. (B to G)
Results from E10-11 cortical stem cells. (B) CD31
stains endothelial cells in the transwells (top), but
no CD31 cells are detected below in the cortical
cell compartment (bottom). Scale bar, 25 m. (C)
Cortical stem cells generate larger, cohesive
clones of attened progeny with stronger junc-
tional -catenin staining in endothelial (bottom)
compared to cortical (top) coculture. Scale bar, 50
m. (D) Neural stem cell clones grown for 7 days
with endothelial (Endo) cells were sheet-like and
had more Nestin and LeX and fewer -
tubulin-III progeny than control clones. Scale
bar, 100 m. (E) Histogram of clone size (dened
by number of progeny) frequency at day 7 in
culture. (F) Mean clone size with different feeder
layers at 7 days in culture [analysis of variance
(ANOVA), * indicates P 0.01 by post-hoc
tests]. (G) Percentages of cells with progenitor
and neuronal markers per stem cell clone
(*, P 0.05, t test).

tened progeny that maintained tight cell- When the transwells were removed, stem cell clones growing in BPAE cocul-
cell contact (illustrated in Fig. 1C by strong endothelial-expanded stem cell clones con- tures contained a high percentage of neu-
junctional -catenin staining), with only a tinued to proliferate but also began to dif- rons, up to 64%, compared to clones grown
few immature, neuron-like cells appearing ferentiate (Fig. 2A), and within 4 days they in CTX coculture (Fig. 2, D and E), and
on top of the sheets. Neural stem cell clones produced -tubulin-III neurons (Fig. neuron production was prolonged (support-
grown with endothelial cells were larger, 2B), which were almost all microtubule- ing online text and fig. S3). Increased neu-
with more primitive progeny [expressing associated protein 2 (MAP-2). About 30% rogenesis from endothelial cocultured neu-
the progenitor markers Nestin and LeX (3)] of the neurons had acquired the later neu- ral stem cells did not occur at the expense
and fewer neurons (expressing -tubulin- ronal marker NeuN (7 ). The clones con- of gliogenesis: The percentage of glial
III), than were clones grown with CTX tained up to 10,000 progeny, and on av- fibrillary acidic protein (GFAP) astro-
(Fig. 1, D to G). Hence, endothelial factors erage 31% were neurons. In contrast, in cytes generated was similar, and although
facilitate expansion of cortical stem cell control CTX cocultures 4 days after trans- oligodendrocyte differentiation (indicated
clones and inhibit their differentiation. well removal, stem cell clones ranged up to by staining with the early oligodendrocyte
VSM and NIH3T3 cells also promoted neu- 4350 cells, and on average only 9% were marker O4) was reduced in BPAE cocul-
ral stem cell proliferation (Fig. 1F), but neurons. Similarly, E11 cortical cells cul- tures compared to CTX cocultures, the dif-
clones were less cohesive and included tured as neurospheres for 7 days then dif- ference could not account for the enhance-
more glial-like progeny than those in endo- ferentiated in adherent culture for 4 days ment of neuron generation (Fig. 2, C and
thelial coculture. produced only 7% neurons. Many more E). NIH3T3 cells enhanced oligodendro-
cyte generation. Coculture with VSM or
Fig. 2. Endothelial-expanded NIH3T3 cells reduced neurogenesis com-
stem cells show enhanced pared to CTX (Fig. 2E), showing that the
neuron production. (A) endothelial effect is cell-type specific.
Schematic of the experi- Endothelial cells stimulate proliferation
ment. (B to E) Results from
E10-11 cortical stem cells.
and neurogenesis of neural stem cells from a
(B) Endothelial-cocultured variety of embryonic CNS regions (7) and
stem cells (Endo) had from different stages. E15.5 cortical and adult
massive neuron production SVZ stem cells grown in endothelial cocul-
compared to control- ture generated sheets of LeX, Nestin cells.
cocultured cells (CTX), After differentiation, E15.5 endothelial-
shown by -tubulin-III ex-
pression at day 11 (D11).
expanded cortical cells and adult SVZ cells
Scale bar, 200 m. (C) Oli- produced more neurons compared to control
godendrocytes and astro- cells (Fig. 2, F and H).
cytes, shown by O4 (red) Neurosphere-expanded stem cells re-
and GFAP (green) labeling, sponded to endothelial factors. E15.5 cortical
respectively, are present in cells grown as neurospheres in fibroblast
both endothelial and CTX
cocultured clones at D14.
growth factor 2 (FGF2) for 7 days were
Scale bar, 100 m. (D) His- plated in adherent conditions and cocultured
togram of neuron genera- for 3 days with endothelial cells or with
tion per stem cell clone. (E) age-matched cortical cells, then differentiated
Enhancement of neuron by withdrawal of feeder cells for 4 days. Stem
production from stem cells cells exposed to endothelial factors produced
does not occur at the ex-
pense of glial cell produc-
22% neurons, compared to 2% neurons in
tion. Neuron production is control CTX cocultures (Fig. 2G).
specically enhanced by en- In vivo, most projection neurons are
dothelial cell coculture, born in the early embryonic period, where-
compared to coculture with as glia and interneurons arise later; adult
other cell types (*, P 0.05; stem cells are primed to generate interneu-
**, P 0.001; ANOVA,
Newman-Keuls test). (F to
rons (8, 9). To examine the neuron subtypes
H) Forebrain stem cells from generated from E10-11 cortical stem cells
older embryonic and adult expanded in endothelial coculture, differ-
stages show enhanced neu- entiated clones were stained for glutamic
rogenesis with endothelial acid decarboxylase (GAD67), a GABAer-
cell coculture. (F) Neurons gic marker typically expressed in interneu-
per E15.5 cortical cell ad-
herent clone (CTX) (*, P
rons, or Tbr1, an early pyramidal neuron
0.05, ANOVA, Newman- marker that preferentially labels projection
Keuls test). (G) Neurons per neurons (10) (Fig. 3A). More stem cell
E15.5 cortical neurosphere clones growing in BPAE coculture made
(NS) (*, P 0.001, ANOVA, Tbr1 projection neurons, compared to CTX-
Newman-Keuls test). (H) cocultured clones (Fig. 3B). BPAE-cocultured
Neurons per adult SVZ
adherent clone (*, P 0.01,
stem cells generated more Tbr neurons than
t test). neurosphere-expanded E10 cells that were sub-
sequently differentiated in adherent culture
(9.95% versus 2.41%). Thus, endothelial cell
coculture supports development of both projec-
tion neurons and interneurons.

Fig. 3. Endothelial- effector Hes1 was up-regulated, but Hes5
expanded E10 stem cell was not (Fig. 4C), consistent with involve-
clones retain the ability ment of Hes1 in neural stem cell self-
to generate Tbr1 pro-
jection neurons as well
renewal (17, 18).
as GAD interneurons. Our results identify endothelial cells as
(A) GAD (cytoplasmic, critical components of the neural stem cell
red), Tbr1 (nuclear mark- niche, as they secrete soluble factors that
er, red) and -tubulin-III maintain CNS stem cell self-renewal and
(green) staining. (B) His- neurogenic potential. Thus, although FGF2
tograms showing the
frequency of GAD and
promotes neural stem cell proliferation, it
Tbr1 neurons in stem cannot alone maintain their self-renewal;
cell clones. endothelial factors acting with FGF2
accomplish this.
Fig. 4. Endothelial factors stimu- In the presence of endothelial cells, a neu-
late self-renewal of neural stem ral stem cell undergoes symmetric, prolifera-
cells. (A) Comparison between tive divisions to produce undifferentiated
typical lineage trees reconstruct- stem cell sheets that maintain their multipo-
ed from time-lapse video record- tency and, upon endothelial cell removal,
ings of single E10 cortical stem
cells grown with endothelial cells generate neurons as well as astrocytes and
and those grown under control oligodendrocytes. No CD31 cells were de-
conditions. In endothelial cocul- tected in clones, showing that, at least under
ture, the cortical stem cell divid- these circumstances, neural stem cells do not
ed symmetrically and did not generate endothelial progeny.
make neurons during the record- Growth with endothelial cell derived fac-
ing period (all progeny were
Nestin as shown in the uores- tors may be an important tool for promoting
cence and phase images of the neural stem cell self-renewal and neurogen-
nal clone). In contrast, a corti- esis, allowing efficient production of neural
cal stem cell grown under con- stem cells and a variety of CNS neurons for
trol conditions generated an use in replacement therapies.
asymmetric lineage tree, gener-
ating neurons [-tubulin-III References and Notes
(red), designated as N in the lin- 1. F. Doetsch, Curr. Opin. Genet. Dev. 13, 543 (2003).
eage tree] as well as Nestin 2. T. D. Palmer, A. R. Willhoite, F. H. Gage, J. Comp.
progenitor cells (green). Neuro- Neurol. 425, 479 (2000).
nal progeny are numbered to 3. A. Capela, S. Temple, Neuron 35, 865 (2002).
show the match of cells in the 4. A. Louissaint Jr., S. Rao, C. Leventhal, S. A. Goldman,
Neuron 34, 945 (2002).
nal clone to the lineage tree. 5. G. Breier, U. Albrecht, S. Sterrer, W. Risau, Develop-
Arrows indicate neurons in the ment 114, 521 (1992).
eld that did not originate from 6. X. Qian et al., Neuron 28, 69 (2000).
this clone. (B) After treatment of 4-day-old cocultures with -secretase inhibitor II for 6 hours, 7. Q. Shen, L. Jin, S. K. Goderie, S. Temple, unpublished
-catenin staining is signicantly decreased and -tubulin-III staining signicantly increased in data.
BPAE cocultured clones, whereas there was no effect on CTX cocultured clones (ANOVA; *, P 8. J. O. Suhonen, D. A. Peterson, J. Ray, F. H. Gage,
0.01 by post-hoc tests). DMSO, dimethyl sulfoxide. (C) Hes1 is up-regulated after endothelial Nature 383, 624 (1996).
coculture, but Hes5 expression was similar to that in control coculture. Reverse transcription 9. D. G. Herrera, J. M. Garcia-Verdugo, A. Alvarez-Buylla,
Ann. Neurol. 46, 867 (1999).
polymerase chain reaction gel band densities were normalized to expression levels of glyceralde- 10. R. F. Hevner et al., Neuron 29, 353 (2001).
hyde phosphate dehydrogenase (GAPDH). 11. X. Qian, S. K. Goderie, Q. Shen, J. H. Stern, S. Temple,
Development 125, 3143 (1998).
12. A. Chenn, C. A. Walsh, Science 297, 365 (2002).
That projection neurons typical of the tical stem cells cocultured with endothelial 13. A. Chenn, C. A. Walsh, Cereb. Cortex 13, 599 (2003).
early embryo arise in E10-11 cocultures cells for 4 days generated more secondary 14. B. Lu, F. Roegiers, L. Y. Jan, Y. N. Jan, Nature 409, 522
after many cell divisions suggests that en- stem cell clones, neurospheres, and neuron- 15. S. Hitoshi et al., Genes Dev. 16, 846 (2002).
dothelial factors promote stem cell self- generating progenitor cells than did those 16. A. Chojnacki, T. Shimazaki, C. Gregg, G. Weinmaster,
renewal and inhibit the normal progression cocultured with CTX cells (fig. S4). S. Weiss, J. Neurosci. 23, 1730 (2003).
17. Y. Nakamura et al., J. Neurosci. 20, 283 (2000).
in which older stem cells preferentially pro- The most obvious effect of endothelial 18. T. Ohtsuka, M. Sakamoto, F. Guillemot, R. Kageyama,
duce glia or interneurons. We found few factors is that they promote neural stem cell J. Biol. Chem. 276, 30467 (2001).
Tbr1 neurons produced from E15.5 stem growth as epithelial sheets with extensive 19. We thank H. Singer for VSM cells, Y.-P. Hseuh for
Tbr1 antibody, and C. Fasano, Y. Wang, C. Butler, and
cells and none from adult SVZ cells, indi- junctional contacts (Fig. 1C), which could K. Kirchofer for help in manuscript preparation. Sup-
cating that endothelial factors are permis- promote self-renewal by influencing - ported by the National Institute of Neurological Dis-
sive, not instructive, for this fate: They catenin signaling pathways (12, 13), mode orders and Stroke and the New York State spinal cord
research program.
cannot reverse the restriction. of cell division (14 ), and Notch activation
Supporting Online Material
Supporting the hypothesis that endothelial (15 ). Indeed, stem cells cocultured with
factors promote stem cell self-renewal, time- endothelial cells and then exposed to - Materials and Methods
lapse video recording of dividing clones re- secretase inhibitor II, which inhibits SOM Text
veals that stem cells grown with endothelial Notch1 activation (16 ), showed a similar Figs. S1 to S7
References and Notes
cells underwent symmetric, proliferative di- extent of cell-cell contact, division, and
visions generating Nestin progeny, in con- differentiation to those in CTX cocultures 9 January 2004; accepted 19 March 2004
Published online 1 April 2004;
trast to the asymmetric division patterns seen (Fig. 4B and fig. S5). In neural stem cells 10.1126/science.1095505
in control conditions (6, 11) (Fig. 4A). Cor- cultured with endothelial factors, the Notch Include this information when citing this paper.

MicroRNAs Modulate Hematopoietic
Lineage Differentiation
Chang-Zheng Chen,1 Ling Li,1 Harvey F. Lodish,1,2*
David P. Bartel1,2*

MicroRNAs (miRNAs) are an abundant class of 22-nucleotide regulatory

RNAs found in plants and animals. Some miRNAs of plants, Caenorhabditis
elegans, and Drosophila play important gene-regulatory roles during de-
velopment by pairing to target mRNAs to specify posttranscriptional re-
pression of these messages. We identify three miRNAs that are specically
expressed in hematopoietic cells and show that their expression is dynam-
ically regulated during early hematopoiesis and lineage commitment. One
of these miRNAs, miR-181, was preferentially expressed in the B-lymphoid
cells of mouse bone marrow, and its ectopic expression in hematopoietic
stem/progenitor cells led to an increased fraction of B-lineage cells in both
tissue-culture differentiation assays and adult mice. Our results indicate
that microRNAs are components of the molecular circuitry that controls
mouse hematopoiesis and suggest that other microRNAs have similar reg-
ulatory roles during other facets of vertebrate development.

MicroRNAs (miRNAs) are ~22-nucleotide examined. miR-181 (9, 12, 21), miR-223 (9),
(nt) noncoding RNAs that can play impor- and miR- 142s (18) were carried forward for
tant roles in development by targeting the further analyses, because they, unlike miR-16
messages of protein-coding genes for cleavage and most of the other miRNAs cloned, were Fig. 1. Northern blots showing tissue expres-
or repression of productive translation (13). Ex- differentially or preferentially expressed in sion of four miRNAs cloned from mouse bone
amples include the lin-4 and let-7 miRNAs, hematopoietic tissues (Fig. 1). marrow (25) (g. S1). As loading controls,
which control the timing of Caenorhabdi- miR-181 was very strongly expressed in the blots were also probed for U6 small nuclear
tis elegans larval development (46); Bantam thymus, the primary lymphoid organ, which RNA. The lengths (in nucleotides) of RNA
miRNA, which regulates Drosophila tissue markers are indicated, as are the bands that
mainly contains T lymphocytes. It was also represent the mature miRNAs (miR) and pre-
growth by stimulating cell proliferation and strongly expressed in the brain and lung and sumed hairpin precursors (P).
preventing apoptosis (7); and miR-14, which was detectable in bone marrow and the spleen.
affects Drosophila fat metabolism and prevents miR-223 was nearly exclusively expressed in by the B220 surface antigen. In other differ-
apoptosis (8). Humans have between 200 and bone marrow, the primary hematopoietic or- entiated lineages, miR-181 expression did not
255 genes that encode miRNAs, an abundance gan, which consists of hematopoietic stem increase over that seen in Lin cells. Sorted
corresponding to almost 1% of the protein- cells and myeloid, erythroid, and lym- lineage cell populations are ~85% pure;
coding genes (9). Based on the evolutionary phoid cells at various differentiation stages. thus, some miRNA signals in the other lin-
conservation of many miRNAs among the miR-142s, whose gene is at the site of a trans- eages might be caused by residual B220+ cells.
different animal lineages, it is reasonable to location associated with an aggressive B cell miR-142s expression was lowest in the ery-
suspect that some mammalian miRNAs might leukemia (16, 18), was highly expressed in all throid (Ter-119+) and T-lymphoid (CD3e+)
also have important functions during develop- the hematopoietic tissues tested, with little or lineages and highest in B-lymphoid (B220+)
ment (1014). Moreover, genes for miR-142, no expression in nonhematopoietic tissues. and myeloid (Gr-1+ and Mac-1+) lineages,
miR-15, and miR-16 are at sites of transloca- Expression at embryonic day 13 in fetal liver, consistent with its ubiquitous expression in
tion breakpoints or deletions linked to human an embryonic hematopoietic organ, suggests bone, spleen, and thymus tissues (Figs. 1 and
leukemias (1518). However, no mammalian that miR-142 might also function in early he- 2). miR-223 expression was confined to my-
miRNAs have established functions (19). matopoietic development. eloid (Gr-1+ and Mac-1+) lineages, with barely
As a first step toward testing the idea Because the bone marrow, spleen, and thy- detectable expression in T- and B-lymphoid
that miRNAs might play roles in mam- mus each have specialized functions in adult and erythroid lineages (CD3e+, B220+, and
malian development, and more specifi- hematopoiesis and comprise largely differ- Ter-119+, respectively) (Fig. 2). This observa-
cally that some might regulate mamma- ent cell types, the differential expression of tion is consistent with miR-223 expression in
lian hematopoiesis, we cloned ~100 unique the miRNAs in these complex tissues sug- bone marrow but not in the spleen or thymus
miRNAs from mouse bone marrow, using the gested that individual hematopoietic cell types (Fig. 1). As observed for miR-181, expres-
protocol of Lau et al. (20). Most had already might differentially express the miRNAs. sion of miR- 223 and miR-142s was low in
been identified as vertebrate miRNAs, but When cells within bone marrow were sorted Lin cells relative to their preferred Lin+ cell
their expression in bone marrow had not been based on lineage markers, they were found populations, suggesting that these miRNAs
to differentially express the hematopoietic are also induced during lineage differentiation.
miRNAs (Fig. 2). In contrast, expression of For each of the miRNAs, specific expression
Whitehead Institute for Biomedical Research, Nine miR-16, an miRNA seen in broad range of tis- was validated by the reduction of correspon-
Cambridge Center, Cambridge, MA 02142, USA. sues, was more constant. dent miRNA expression in the reciprocal lin-
Department of Biology, Massachusetts Institute of
Technology, Cambridge, MA 02142, USA. Mature miR-181 expression in bone mar- eage-depleted cell populations (Fig. 2).
row cells was detectable in undifferentiated Differential expression of three miRNAs in
*To whom correspondence should be addressed. E-
mail: (H.F.L) and progenitor cells (Lin) and up-regulated in dif- specific hematopoietic lineages suggested that
edu (D.P.B.) ferentiated B lymphocytes, which are marked they might influence hematopoietic lin-

eage differentiation. To test this possibility, we where it is cleaved by Dicer to generate the and 2). In light of the recent report that Drosha
set out to ectopically express these miRNAs in mature miRNA (22). We reasoned that the miR- is responsible for the nuclear processing of
hematopoietic progenitor cells. A vector with 223(67) transcript was not cleaved by Dicer, miRNA primary transcripts (23), our results
the murine stem-cell retrovirus backbone and because it did not derive from a properly pro- can be explained by the idea that elements
a polymerase III (pol III) expression cassette cessed primary transcript, and that sequences needed for Drosha recognition reside within
was developed to efficiently express miRNAs flanking the hairpin precursor were needed for the sequences that flank the miR-223 pre-
in primary hematopoietic progenitor cells nuclear processing of the primary transcript. dicted hairpin.
(Fig. 3A). We first tried expressing a 67-nt To include the sequences needed for proper Having determined that sequences flanking
miRNA transcript that included the predicted processing, constructs with increasing genom- the hairpin were needed for detectable miRNA
miR-223 hairpin precursor. The miR-223(67) ic sequence flanking the 67-nt predicted hair- expression, we speculated that a general strat-
transcript was highly expressed, but no mature pin were generated. All constructs with at least egy for miRNA expression would be to use
~22-nt miRNA could be detected (Fig. 3B). 40 nt on each side of the 67-nt hairpin were ~270-nt primary transcripts that included the
Previous studies of miR-30 biogenesis indi- efficiently processed into the mature miRNA ~22-nt mature miRNA and 125 nt of genomic
cate that miRNA primary transcripts are first (Fig. 3B), with a heterogeneity pattern of sequence flanking each side of the miRNA.
cleaved in the nucleus to generate the hairpin 21- to 24-nt RNAs similar to that seen for the This strategy has proven successful for all
precursor, which is exported to the cytoplasm mature miR-223 in bone marrow cells (Figs. 1 13 of the miRNAs that we have attempted to
p y ( g ) p p y
observed for miR-181, expression of miR- express miRNAs in primary hematopoietic

Fig. 2. Northern analysis of miRNA expression

in hematopoietic lineages from mouse bone
marrow (g. S2). Antibodies to surface antigens
CD3e, B220, Gr-1, Mac-1, and Ter-119 were Fig. 3. A general strategy for the ectopic expression of miRNAs. (A) The retroviral construct for
used to purify mouse bone marrow cells of the miRNA expression. A pol III expression cassette containing the human H1 promoter (PH1), the
T, B, granulocyte, macrophage, and erythroid miRNA hairpin, and the T5 termination signal was placed in the U3 region of the viral 3 long
lineages, respectively, with magnetic-assisted terminal repeat (LTR). As a marker for infection, the vector also expressed the gene for GFP under
cell sorting. [Sorted cells were at least 85% control of the constitutive murine 3-phosphoglycerate kinase promoter (PPGK). This vector cong-
pure by subsequent uorescence-activated cell uration, which improves stable gene expression in primary cells, was termed double copy, because
sorting (FACS) analysis.] Both total RNA (5 g the process of retroviral reverse transcription and integration causes two copies of the miRNA
per lane) from the puried lineages (left) and expression cassette to be integrated into the host genome (29). (B) Summary of miR-223
total RNA (20 g per lane) from the cells expression and maturation from vectors designed to express successively longer miR-223 genomic
remaining after depletion of specic lineages fragments, each containing the miR-223 predicted hairpin and anking sequences. miR-223(67),
[CD3e, B220, Gr-1, Mac-1, and Ter-119 (87), (107), (147), (187), and (227) transcripts included the 67-nt miR-223 predicted hairpin at their
cells (right)] were analyzed. Total RNA from a center, with 0, 10, 20, 40, 60, or 80 nt of genomic sequence anking each side of the hairpin,
cell population depleted in Lin cells and thus respectively. Expression and maturation of miR-223 in transfected 293T cells was examined by
enriched for undifferentiated hematopoietic Northern analysis. High expression of the primary transcript was seen for each of the three
stem/progenitor cells was also analyzed (Lin- constructs that did not generate mature miRNA. (C to E) Expression of miR-223, miR-181, and
eage). The lengths (in nucleotides) of RNA miR-142s from 270-nt primary transcripts that included the 22-nt mature miRNA and 125 nt
markers are indicated, as are the bands that of genomic sequence anking each side of the miRNA, in transfected 293T cells or viral-transduced
represent the mature miRNAs and presumed mouse bone marrow cells (BM). (F) Expression of miR-30 from the 71-nt predicted miR-30 hairpin
hairpin precursors. For the loading control, or a 272-nt fragment with 125 nt of genomic sequence anking each side of the miRNA. Ethidium
blots were reprobed for U6 small nuclear RNA. staining of 5S ribosomal RNA served as the loading control.

Modest effects were also seen when analyzing
cells for myeloid lineage markers (fig. S3). In
contrast, ectopic expression of miR-30 had
little or no effect on the output of lymphoid
and myeloid cells, indicating that merely ex-
pressing an arbitrary miRNA does not influ-
ence lymphoid differentiation.
Because miR-181 ectopic expression
had the greatest effect in vitro, we examined
its effect in vivo. Mouse Lin bone marrow
cells were infected with either the retrovirus
that expressed miR-181 or the control vector
that expressed no miRNA and were then trans-
planted into lethally irradiated mice, where
they reconstituted all blood lineages. After 4.5
weeks, the lineage composition of peripheral
blood cells descending from infected stem/
progenitor cells (GFP+ cells) was examined
(25). As seen in vitro, miR-181 expression in
vivo led to a substantial increase in B-lym-
Fig. 4. Effect of miRNA ectopic expression on hematopoietic lineage differentiation in vitro. (A)
phoid (CD19+) cells, with the median fraction
Percentage CD-19 cells and (B) percentage Thy-1.2 cells among the differentiating hema-
topoietic progenitor cells ectopically expressing either no miRNA (vector), a nonhematopoietic of these cells in peripheral blood increasing to
miRNA (miR-30), or one of three hematopoietic microRNAs (miR-181, miR-223, and miR- 80% from the control value of 32% (Fig. 5A).
142s). The average of 12 culture replicates for each construct is shown, with error bars This increase was accompanied by a substan-
indicating the standard deviation. Statistically signicant differences from the vector control, tial (~88%) decrease in T-lymphoid (Thy-1.2+)
as determined by the Students t test, are indicated (*, P 0.01; **, P 0.0001; ***, P cells, particularly the CD8+ T cells, for which
107). In independent repetitions of this experiment, analogous changes relative to the control
the median percentage decreased from 16% to
vectors were observed with similar statistical signicance; however, the absolute percentages
of Thy-1.2 and CD-19 cells differed, perhaps because of the heterogeneity of the Lin bone 1.2% (Fig. 5, A and B). There were relatively
marrow cells from different groups of 5-uorouracilprimed mice. (C) Representative FACS small or insignificant decreases in CD4+ T
analyses of Thy-1.2 and CD-19 lineage marker expression, with antibodies conjugated to cells and myeloid lineage cells (Fig. 5, B and
allophycocyanin (APC) and phycoerythrin (PE), respectively, for the experiment in (A) and (B). C).
FACS plots were gated on GFP expression, which indicated the cells descending from infected Hematopoietic lineage differentiation,
progenitor cells. For each quadrant, the fraction of cells relative to the total number of GFP
the process of continuous development of
cells is shown as a percentage.
hematopoietic stem cells into at least eight
different blood lineages, is known to be
express, including the three hematopoietic cocktail of cytokines and growth factors (25,
controlled or modulated by complex mo-
miRNAs (Fig. 3, C to E). Mature miRNAs 26). Cells descending from infected progeni-
lecular events that simultaneously regulate
ectopically expressed in 293T cells or bone tor cells were distinguished on the basis of
the commitment, proliferation, apoptosis,
marrow cells had length distributions indis- the green fluorescent protein (GFP) marker
and maturation of hematopoietic stem/pro-
tinguishable from those of the endogenous carried by the vector, and differentiation of
genitor cells. The demonstration that certain
miRNAs, as shown for miR-142s (Fig. 3E). Lin cells to Lin+ cells was characterized by
miRNAs are differentially expressed in he-
miR-30 is unusual in that it can be expressed the expression of lineage-specific surface an-
matopoietic lineages in vivo and are able to
from transcripts in which its 71-nt hairpin tigens (25).
alter lineage differentiation provides solid evi-
is flanked by heterologous sequence (24). Ectopic expression of the hematopoi-
dence that miRNAs represent a class of mol-
Nonetheless, when expressed from our vector, etic miRNAs substantially altered lineage
ecules that regulate mouse hematopoiesis and,
miR-30 was much more efficiently processed differentiation (Fig. 4). Expression of
more broadly, mammalian development.
when presented in the context of its native miR-181 resulted in a doubling of cells in the
The ability of ectopically expressed
flanking sequence (Fig. 3F). B-lymphoid lineage with no significant
miR-181 to increase the fraction of B-lin-
To uncover the effects of ectopic expres- change in the T-lymphoid lineage, as mea-
eage cells in vitro and in vivo (Figs. 4 and 5)
sion of miRNAs on hematopoietic lineage sured by the fractions of cells that express the
coincides with its preferential expression in
differentiation, Lin hematopoietic progenitor Thy-1.2 or CD19 cell surface antigens, which
B-lymphoid cells in mouse bone marrow (Fig.
cells from mouse bone marrow were in- are markers for the T- and B-lymphoid lineag-
2), suggesting that miR-181 is a positive reg-
fected with viral vectors that expressed either es, respectively (Fig. 4, A and B). Ectopic ex-
ulator for B-cell differentiation. One explana-
miR-181, miR-223, miR-142s, miR-30 pression of miR-142s or miR-223 had oppo-
tion for the effect of miR-181 expression on
(a control miRNA), or no miRNA (25). site effectsa 30 to 40% increase in the T-lym-
the differentiation of both B cells (CD19+)
miR-30 was selected as a control because phoid lineage with little or no reduction in the
and cytotoxic T cells (CD8+), which are not
its expression was detectable in lung and B-lymphoid lineage. At the two extremes, the
developmentally linked during hematopoi-
kidney but not hematopoietic tissues. The ratio of T- to B-lineage cells ranged from
etic lineage commitment, is that miR-181
cells were then seeded onto S17 bone mar- about 1:1 to about 4:1 (Fig. 4) (when miR-181
acts independently in the two lineages, per-
row stromal cells and supplemented with a and miR-142s were expressed, respectively).
haps through the repression of different target

genes. Indeed, miR-181 is highly expressed exploration of the roles of hematopoietic miR- a promiscuous beginning, a so-called prim-
in the thymus, supporting the idea that it also NAs in modulating lineage differentiation, ing state in which many lineage-specific genes
modulates T cell development in this organ computational and molecular experiments are required for subsequent unique lineages are
(Fig. 1). The observation that the differentiation under way to determine their regulatory tar- coexpressed (27). Thus, selective gene silenc-
of myeloid and other lymphoid cell types was gets. If we assume a mode of regulation analo- ing might be a key event during subsequent
not totally blocked when the B-lymphoid lin- gous to that observed in invertebrates, miRNA hematopoietic lineage differentiation events.
eage increased suggests that miR-181, at least modulation of hematopoietic lineage differ- Clearly, progressive silencing of lineage-spe-
when considered singly rather than in combi- entiation supports the notion that the roles cific genes could be mediated by changes in
nation with other miRNAs, appears to func- of translational regulation in hematopoiesis the activation of master transcription factors
tion more as a lineage modulator than as a and, more broadly, vertebrate development or by chromatin remodeling. Our work adds
switch. might have been underappreciated. Studies to this list a set of hematopoietic-specific
In the known invertebrate examples, on the gene expression profiles of uncommit- miRNAs that presumably act by pairing to the
miRNAs repress the productive translation ted hematopoietic stem cells and intermediate mRNAs of their target genes to direct gene
of their mRNA targets (1). To facilitate further progenitor cells reveal that stem cells exhibit silencing processes critical for hematopoiesis.

Fig. 5. Effect of miR-181 ectopic

expression on hematopoietic
lineage differentiation in vivo.
(A) Percentage of T-lymphoid
(Thy-1.2) and B-lymphoid
(CD19) lineage cells in GFP
nucleated peripheral blood cells,
in mice reconstituted with bone
marrow cells transduced with
control (black) or miR-181 (red)
retroviral vectors. Box plots de-
scribe the distribution of indi-
vidual lineage composition from
all positively reconstituted re-
cipients (those with more than
1.0% GFP cells in peripheral
blood). The ends of the boxes
dene the 25th and 75th per-
centiles, a line indicates the me-
dian, and bars dene the 5th
and 95th percentiles. Individual
outliers are also shown. P values
were determined with the Mann-Whitney rank sum test. (B) T cell subtypes and Gr-1 double-negative cells (Mac-1 Gr-1) are nonmyeloid cells. (D)
marked by CD4 and CD8 surface antigens. (C) Neutrophils and monocytes Representative FACS analyses for the same experiment. Gating was on GFP
marked by Mac-1 and Gr-1 double-positive cells (Mac-1 Gr-1) and Mac-1 cells. For each quadrant, the fraction of cells relative to the total number of
positive and Gr-1 negative-to-low cells (Mac-1 Gr-1), respectively. Mac-1 GFP cells is indicated as a percentage.

References and Notes

1. V. Ambros, Cell 113, 673 (2003). 17. M. Lagos-Quintana, R. Rauhut, W. Lendeckel, T. Tus- Gilboa, Proc. Natl. Acad. Sci. U.S.A. 86, 3519 (1989).
2. B. Bartel, D. P. Bartel, Plant Physiol. 132, 709 (2003). chl, Science 294, 853 (2001). 30. We thank N. Lau and B. Reinhart for reagents and
3. J. F. Palatnik et al., Nature 425, 257 (2003). 18. M. Lagos-Quintana et al., Curr. Biol. 12, 735 (2002). advice on the cloning of endogenous Dicer products,
4. R. C. Lee, R. L. Feinbaum, V. Ambros, Cell 75, 843 19. The recent report that miR-23 regulates Hes1 during E. Weinstein for bioinformatic analysis, V. Carey for
retinoic acidinduced differentiation (28) has been help on statistical analysis, and members of the
5. B. Wightman, I. Ha, G. Ruvkun, Cell 75, 855 (1993).
called into question, because it mistakenly reported Lodish and Bartel laboratories and P. Zamore for
6. B. J. Reinhart et al., Nature 403, 901 (2000).
7. J. Brennecke et al., Cell 113, 25 (2003). the analysis of Homolog of ES1 rather than of Hes1. comments on the manuscript. Supported in part by
8. P. Xu, S. Y. Vernooy, M. Guo, B. A. Hay, Curr. Biol. 13, 20. N. C. Lau, L. P. Lim, E. G. Weinstein, D. P. Bartel, grants from NSF (H.F.L.) and NIH (D.P.B.) and by a
790 (2003). Science 294, 858 (2001). Donaldson, Lufkin, and Jenrette postdoctoral fellow-
9. L. P. Lim, M. E. Glasner, S. Yekta, C. B. Burge, D. P. 21. J. Dostie, Z. Mourelatos, M. Yang, A. Sharma, G. ship from the Cancer Research Institute (C.-Z. C.).
Bartel, Science 299, 1540 (2003). Dreyfuss, RNA 9, 631 (2003).
Supporting Online Material
10. A. E. Pasquinelli et al., Nature 408, 86 (2000). 22. Y. Lee, K. Jeon, J. T. Lee, S. Kim, V. N. Kim, EMBO J. 21,
11. V. Ambros, R. C. Lee, A. Lavanway, P. T. Williams, D. 4663 (2002).
Jewell, Curr. Biol. 13, 807 (2003).
Materials and Methods
23. Y. Lee et al., Nature 425, 415 (2003).
12. M. Lagos-Quintana, R. Rauhut, J. Meyer, A. Borkhardt, Figs. S1 to S3
24. Y. Zeng, B. R. Cullen, E. J. Wagner, RNA 9, 112 (2003).
T. Tuschl, RNA 9, 175 (2003). References and Notes
25. Materials and methods are available as supporting
13. L. P. Lim et al., Genes Dev. 17, 991 (2003).
material on Science Online. 25 September 2003; accepted 10 November 2003
14. E. G. Moss, L. Tang, Dev. Biol. 258, 432 (2003).
15. G. A. Calin et al., Proc. Natl. Acad. Sci. U.S.A. 99, 26. L. S. Collins, K. Dorshkind, J. Immunol. 138, 1082 Published online 4 December 2003;
15524 (2002). (1987). 10.1126/science.1091903
16. C. E. Gauwerky, K. Huebner, M. Isobe, P. C. Nowell, 27. T. Enver, M. Greaves, Cell 94, 9 (1998). Include this information when citing this paper.
C. M. Croce, Proc. Natl. Acad. Sci. U.S.A. 86, 8867 28. H. Kawasaki, K. Taira, Nature 423, 838 (2003)
(1989). 29. P. A. Hantzopoulos, B. A. Sullenger, G. Ungers, E.
( )

pression on some CD4+ thymocytes suggested
Development of a Human that human regulatory T cells might also be
generated via thymic development (fig. S3, A
and B) (20, 21).
Adaptive Immune System in Cord Mature T cells with a broad V repertoire
were detectable in thymus, spleen, mesenteric
Blood CellTransplanted Mice lymph nodes, and bone marrow (Fig. 1D; fig.
S3C). More than 40% of T cells displayed a
Elisabetta Traggiai,1* Laurie Chicha,1* Luca Mazzucchelli,2 nave phenotype as assessed by CD45RA/
Lucio Bronz,3 Jean-Claude Piffaretti,4 Antonio Lanzavecchia,1 CCR7 expression (22) (Fig. 1D). To test
Markus G. Manz1 whether human T cells in mouse secondary
lymphoid organs underwent similar numbers
Because ethical restrictions limit in vivo studies of the human hemato- of post-thymic cell divisions as T cells in hu-
lymphoid system, substitute human to small animal xenotransplantation man newborn blood, TCR-rearrangement ex-
models have been employed. Existing models, however, sustain only limited cision circles (TRECs) were measured. TREC
development and maintenance of human lymphoid cells and rarely produce levels were highest in mouse thymi, whereas
immune responses. Here we show that intrahepatic injection of CD34 somewhat reduced levels were detected in
human cord blood cells into conditioned newborn Rag2/c/ mice leads mouse spleen and lymph nodes, which cor-
to de novo development of B, T, and dendritic cells; formation of structured related closely with TREC levels in nontrans-
primary and secondary lymphoid organs; and production of functional im- planted human cord blood (fig. S3D). Thus,
mune responses. This provides a valuable model to study development and human T cells in mouse peripheral tissues
function of the human adaptive immune system in vivo. appeared to have undergone similar numbers
of divisions as T cells in cord blood. Together
Biomedical research in humans is restricted Rag2 c mice, a mutant strain that lacks B, these data imply that in the thymi of reconsti-
largely to in vitro assays that lack the com- T, and NK cells (17, 18), intrahepatically (i.h.) tuted Rag2 c mice human T cells with a
ponents and complexity of a living organism. with CD34+ cord blood cells (19). Mice were broad repertoire develop over at least 6 months
To overcome this limitation, substitute in vivo subsequently analyzed between weeks 4 and and home to secondary lymphoid organs with-
models have been developed in which human 26 of age, and human CD45+ hematopoietic out being massively activated.
hematopoietic cells and tissues are trans- cells were detected in all animals (Fig. 1A). The proliferative capacity of T cells in
planted into mice that are compromised in An increase in splenic and thymic cellularity response to mouse or human major histo-
their capacity to reject xenogenic grafts. En- was detectable, and all mice beyond 8 weeks compatibility complex (MHC) antigens was
graftment was first reported after transfer of developed mesenteric lymph nodes, several next tested using mixed lymphocyte reactions
mature human peripheral blood leukocytes in within size and cellularity of wildtype controls (MLRs). Human T cells isolated from mouse
severe combined immunodeficient mice (hu (Fig. 1B). lymph nodes and spleen proliferated vigor-
PBL-SCID mice) (1) and transplantation of Most CD19+ cells in bone marrow (BM) of ously when stimulated with human allogeneic
blood-forming fetal liver cells, fetal bone, fetal engrafted mice were negative for surface im- dendritic cells (DCs) but weakly, or not at all,
thymus, and fetal lymph nodes in SCID mice munoglobulin M (IgM) and CD20 expression, when stimulated with autologous human DCs
(SCID-hu mice) (2, 3). Subsequently, some whereas spleen, lymph node, and blood CD19+ (fig. S4A). This indicates that developing T
level of human hematopoietic development cells expressed these antigens (fig. S1A). This cells had undergone some level of selection on
was achieved by transplantation of blood-form- was consistent with generation of B cells in human MHC, possibly within the mouse thy-
ing cells in NOD/SCID, NOD/SCID2m , or BM and subsequent migration to spleen and mus. Reactivity to mouse DCs was, generally,
NOD/ SCIDc mice (47). However, transfer lymph nodes. Human IgM was detectable in low, possibly reflecting suboptimal xenogenic
of human cells in immunodeficient mice has, serum of young transplanted animals and in- cell interactions (fig. S4B). However, the
so far, not appeared to result in the de novo creased in most over time; IgG was detected in small difference in response to host BALB/c
formation of a functional human adaptive im- older animals, demonstrating class switching (Rag2 c ) versus mismatched C57BL/6
mune system (1, 716). of Ig isotypes (fig. S1B). Ig-producing cells DCs would also be consistent with the possi-
The liver contributes to perinatal hema- were located in BM and spleen and correlated bility that some thymic selection had occurred
topoiesis, and the hemato-lymphoid system closely with numbers of CD19+CD27+CD138+ on mouse MHC.
expands most significantly during the first plasma cells (fig. S1, C and D). Thus, in con- Human DCs can be divided into CD11c+
weeks of life. Thus, we reasoned that hu- trast to transplanted NOD/SCID mice where DCs and CD11cCD123+ plasmacytoid, type I
man hematopoietic stem and progenitor cells human B cells fail to produce Ig (13, 14), full interferon producing pre-DCs (23, 24). Both
transplanted into the liver of immunodeficient B cell maturation occurred in reconstituted subsets were present in bone marrow, spleen,
newborn mice might find better conditions Rag2 c mice. liver, and, at low levels, thymus and lymph
to engraft, expand, and reconstitute a human All thymi contained double-positive, as nodes of reconstituted mice (Fig. 2, A and B).
immune system. We transplanted newborn well as CD4 and CD8 single-positive, T cells Upon maturation ex vivo, CD11c+ cells dis-
in 1:1 to 4 :1 ratios, with thymi of young mice played typical DC morphology and became
containing fewer mature thymocytes than potent stimulators of allogeneic T cells (Fig. 2,
Institute for Research in Biomedicine (IRB), Via Vela thymi of older mice (Fig. 1C). T cell recep- C and D; fig. S5B). CD11c cells, stimulated
6, 6500 Bellinzona, Switzerland. 2Institute of Pathol- tor (TCR) was upregulated normally during ex vivo with influenza virus, produced high
ogy, Via in Selva 24, 6600 Locarno, Switzerland. 3De- transition from double-negative to single- amounts of -interferon (IFN) (Fig. 2E).
partment of Gynecology and Obstetrics, Ospedale
San Giovanni, 6500 Bellinzona, Switzerland. 4Institute positive stages (Fig. 1C), and typical thymic Therefore, reconstituted animals supported
of Microbiology, Via Mirasole 22A, 6501 Bellinzona, cortex and medulla structures were apparent development of both functional DCs and plas-
Switzerland. (fig. S2). Some thymi in animals beyond 25 macytoid pre-DCs.
*These authors contributed equally to this work. weeks still contained >70% double-positive Spleens from animals older than 16 weeks
To whom correspondence should be addressed. E- cells, indicating continuous T cell generation possessed white pulp-like structures, consist-
mail: over time. Interestingly, CD25 and Foxp3 ex- ing of central arterioles surrounded by human




Fig. 1. Human hematopoietic cell engraftment, organ enlargement, and T cell development in
transplanted animals. (A) Bar graphs represent percent of human CD45 cells in bone marrow
and spleen of consecutively analyzed mice (n 30) at indicated weeks after transplantation.
Bars are broken up into human CD45CD3 T cells, CD45CD19 B cells, and CD45CD19
CD3 cells. Letters X, Y, Z indicate numbers of CD34 cells transplanted per animal: X, 3.8 to
t 7 104; Y, 7 to 9 104; Z, 9 to 12 104. *Indicates animal of which organs are shown in
(B). (B) Representative photographs of spleen, thymus, and mesenteric lymph nodes of
19-week-old BALB/c, transplanted Rag2/c/, and Rag2/c/ control mice. (C) Con-
tour plots show representative human thymocyte staining proles of a young and older animal.
Histogram shows TCR expression on gated double-negative (DN), double-positive (DP), and
single-positive CD4 (CD4 SP) and CD8 (CD8 SP) thymocytes from the same animal. (D)
Representative mesenteric lymph node prole for CD4 and CD8, and CCR7 versus CD45RA,
expression on gated CD3 cells. For comparison, same staining on cord blood cells is shown.

T and B cells (Fig. 3, A and B). Although provide strong evidence that functional inter- vaccinations were started at 12 to 17 weeks of
B and T celldeficient mice lack follicular actions between immune cells were occurring, age, three of five mice produced measurable
dendritic cells (FDCs), immunoreconstitu- both within the engrafted human cell popula- anti-TT IgG antibodies (table S1) and memory
tion with mouse BM induces their formation, tions and across the species barrier. B cells could be detected in lymph nodes (ta-
likely from resident (hostderived) nonhema- To more directly test the immune response ble S2). Although antibody levels were lower
topoietic cells (25). This has been shown to after reconstitution, mice were vacci- than those achieved in human adults, anti-
be dependant on the presence of lymphotoxin nated with tetanus toxoid (TT) or were TT IgG to total IgG ratios compared closely
expressing B cells (25). As expected, no hu- infected with Epstein-Barr virus (EBV). In (table S1).
man FDCs were detectable in reconstituted three animals vaccinated with TT at 8 weeks, Engrafted animals were next infected with
animals, although mouse FDCs (FDC-M1+) no specific human IgG antibodies could be de- increasing doses of EBV and were analyzed
were induced (Fig. 3B). Furthermore, in some tected. It is possible that failure to respond at 5 to 10 weeks later (Fig. 4; table S3). In all
cases, typical germinal centerlike structures this stage may have been due to the immaturity animals, EBV was detectable by polymerase
could be observed (Fig. 3C). These findings of the immune system. In contrast, when TT chain reaction (PCR) in BM, spleen, and

Fig. 2. Functional human dendritic cell subsets develop in transplanted animals. (A) Identi- Fig. 3. Spleens of Rag2/c/ mice trans-
cation of human CD11cCD123 plasmacytoid pre-DC and CD11c DC in bone marrow, planted with human CD34 cord blood
spleen, and liver of an 11-week-old transplanted animal. (B) Histograms show plasmacytoid cells develop structured white pulp. (A) Focal
pre-DC and DC associated maker expression (solid lines) and isotype controls (dashed lines) on accumulation of T (CD3) and B (CD20)
CD11cCD123 and CD11c gated cells in bone marrow. (C) Sorted bone marrow CD11c cells around arterioles. (B) Close-up of (A)
cells, activated with lipopolysaccharide (LPS) and granulocyte-macrophage colony-stimulating showing white pulp containing predominant-
factor (GM-CSF) show typical dendritic cell morphology. (D) CD11c but not CD19 bone ly B cells (CD20) and T cells (CD3), as
marrow cells induce strong proliferation of allogeneic peripheral blood T cells. Histogram well as naive B cells (CD23) and mouse
shows overlay of CD3 gated cells stimulated with CD11c cells (solid line, closed histogram) FDCs (FDC-M1). Human FDCs (CD21)
or CD19 cells (dashed line). T cell proliferation was evaluated at day six of culture. One of were not detected. (C) Close-up showing
two experiments depicted. CFSE, carboxyuorescein diacetate succinimidyl ester. (E) a germinal centerlike structure with ac-
CD123BDCA-4 bone marrow cells produce high amounts of IFN upon viral stimulation. cumulation of highly proliferating
CD123BDCA-4CD45 (1) and CD123BDCA-4CD45 (3) cells from bone marrow of a CD20 bcl6ki67 B cells surrounded by
transplanted animal, and BDCA-4CD19CD14 cells from peripheral blood of a healthy adult CD20bcl6bcl2 B cells and CD3 T cells, in
(2) were sorted and stimulated overnight with inuenza virus (3.5 104 cells each). IFN in this case in a vaccinated animal. Representa-
supernatants was evaluated by enzyme-linked immunosorbent assay (ELISA). Representative tive analysis, 26 weeks [(A) and (B)] and 25
experiment of three. weeks (C) after reconstitution.

lymph node B cells (26). Two animals, in- B cells in vitro (Fig. 4C). These data suggest it remains to be clarified how human T cell se-
fected with the highest EBV doses, developed that upon in vivo challenge with moderate in- lection on mouse thymic background occurs,
LMP1+ B cell proliferation in spleen, liver, and fectious doses, T cell responses were initiated the T cells generated discriminate self from
kidney (Fig. 4A; table S3), whereas five ani- that could control EBV, at least for the time allogeneic MHC. Together, human cells were
mals infected with lower EBV doses did not periods observed. It will be important to next able to collaborate in forming lymphoid organ
develop obvious B cell proliferation. In the determine the type of EBV infection, as well structures and induce the differentiation of
latter, spleen T cells increased substantially as the duration, epitope, and MHC specificity mouse FDCs, thus providing evidence for an
(Fig. 4; table S3). Three of these five mice and of the T cell response. unexpectedly robust cellular interaction across
both mice with B cell proliferation developed Taken together, reconstitution of newborn xenogenic barriers. These findings provide
an inverse CD4+:CD8+ T cell ratio (Fig. 4B; Rag2 c mice with cord blood CD34+ cells a technically straightforward in vivo model
table S3). Moreover, CD8+ T cells from EBV- was found to lead to ortotopic de novo gen- with which it will be possible to characterize
infected animals proliferated strongly when eration and maturation of human DCs, B cells, development, homeostasis, and functional co-
stimulated with autologous EBV-transformed and T cells with a broad repertoire. Although operation of the human adaptive immune sys-
tem. Furthermore, the model should provide a
valuable tool to study pathogens that specifi-
cally target the human immune system and test
potential therapeutic interventions.

References and Notes

1. D. E. Mosier, R. J. Gulizia, S. M. Baird, D. B. Wilson,
Nature 335, 256 (1988).
2. J. M. McCune et al., Science 241, 1632 (1988).
3. J. M. McCune, Semin. Immunol. 8, 187 (1996).
4. L. D. Shultz et al., J. Immunol. 154, 180 (1995).
5. D. L. Greiner, R. A. Hesselton, L. D. Shultz, Stem Cells
16, 166 (1998).
6. O. Kollet et al., Blood 95, 3102 (2000).
7. M. Ito et al., Blood 100, 3175 (2002).
8. D. E. Mosier, Virology 271, 215 (2000).
9. J. Sandhu, B. Shpitz, S. Gallinger, N. Hozumi, J. Im-
munol. 152, 3806 (1994).
10. N. Delhem et al., J. Immunol. 161, 2060 (1998).
11. J. M. Carballido et al., Nature Med. 6, 103 (2000).
12. A. K. Palucka et al., Blood 102, 3302 (2003).
13. M. I. Rossi et al., J. Immunol. 167, 3033 (2001).
14. M. Ramirez, G. A. Rottman, L. D. Shultz, C. I. Civin,
Exp. Hematol. 26, 332 (1998).
15. T. C. Kerre et al., Blood 99, 1620 (2002).
16. H. Hiramatsu et al., Blood 102, 873 (2003).
17. J. P. Goldman et al., Br. J. Haematol. 103, 335
18. F. Mazurier et al., J. Interferon Cytokine Res. 19, 533
19. Materials and methods are available as supporting
material on Science Online.
20. S. Hori, T. Nomura, S. Sakaguchi, Science 299, 1057
21. M. S. Jordan et al., Nature Immunol. 2, 301 (2001).
22. F. Sallusto, D. Lenig, R. Forster, M. Lipp, A. Lanzavec-
chia, Nature 401, 708 (1999).
23. J. Banchereau, R. M. Steinman, Nature 392, 245
24. Y. J. Liu, Cell 106, 259 (2001).
25. Y. X. Fu, D. D. Chaplin, Annu. Rev. Immunol. 17, 399
26. M. G. Manz and J.-C. Piffaretti, unpublished data.
27. We thank M. Ito and the Central Institute for
Experimental Animals, Kawasaki, Japan, for their
provision of Rag2/c/ mice; the staff of ob-
stetrics, Ospedale San Giovanni, Bellinzona, for
help with cord blood collection; and M. Personeni
and L. Lunghi for technical assistance with histol-
ogy sample preparation. Supported in part by a
Novartis grant (M.G.M.) and the Swiss National
Science Foundation (3100-63885 to A.L. and 3100-
102221 to M.G.M.).
Fig. 4. EBV infection of CD34 cord blood reconstituted Rag2/c/ mice. (A) Spleen
histology showing CD20LMP1 B cell proliferation in an animal infected with high doses Supporting Online Material
of EBV ve weeks earlier. (B) Representative uorescence-activated cell sorter (FACS) anal-
ysis of spleen and lymph node T and B cells in a noninfected sibling control and in EBV infected DC1
animals that did and did not develop B cell proliferation. Plots show characteristic T cell Materials and Methods
increase and inversion of CD4:CD8 ratios. (C) Lymph node T cells of EBV-infected animals Tables S1 to S3
proliferate when stimulated ex vivo with autologous EBV-transformed B cells. Lymph node T Figs. S1 to S5
cells from noninfected siblings do not proliferate when stimulated under the same con- References
ditions. Proliferation was measured by CFSE dilution at day ve of cultures. Representative
experiment of two. 24 November 2003; accepted 27 January 2004

onic stem (ES) cells to generate all lineages
Derivation of Oocytes from of the embryo in vivo has been widely report-
ed in the literature, in striking contrast to the

Mouse Embryonic Stem Cells lack of data describing the derivation of germ
cells from ES cells in vitro. We attributed the
inability to demonstrate the derivation of
Karin Hubner,1 Guy Fuhrmann,3 Lane K. Christenson,4 germ cells from ES cells in culture to the lack
James Kehler,1 Rolland Reinbold,1 Rabindranath De La Fuente,2 of a suitable reporter system for the noninva-
Jennifer Wood,4 Jerome F. Strauss III,4 Michele Boiani,1 sive visualization of germ cell formation.
Hans R. Scholer1* Induction of germ cells in culture. Elu-
cidation of the various known regulatory el-
Continuation of mammalian species requires the formation and development ements within the germline-specific gene
of the sexually dimorphic germ cells. Cultured embryonic stem cells are gen-
erally considered pluripotent rather than totipotent because of the failure to 1
Germline Development Group, 2Female Germ Cell
detect germline cells under differentiating conditions. Here we show that Biology Group, Center for Animal Transgenesis and
mouse embryonic stem cells in culture can develop into oogonia that enter Germ Cell Research, School of Veterinary Medicine,
meiosis, recruit adjacent cells to form follicle-like structures, and later develop University of Pennsylvania, New Bolton Center, 382
West Street Road, Kennett Square, PA 19348, USA.
into blastocysts. Oogenesis in culture should contribute to various areas, in- 3
Centre de Neurochimie, Laboratoire de Neurobiolo-
cluding nuclear transfer and manipulation of the germ line, and advance studies gie du Developpement et de la Regeneration, FRE
on fertility treatment and germ and somatic cell interaction and differentiation. 2373 CNRS, 5 Rue Blaise Pascal, 67084 Strasbourg
Cedex, France. 4Center for Research on Reproduction
and Womens Health, School of Medicine, University
In the early mammalian embryo, the germ extraembryonic ectoderm (1, 2). Even during of Pennsylvania, 1349 Biomedical Research Building
line and soma are indistinguishable from each the specification period, precursor cells give 2/3, 421 Curie Boulevard, Philadelphia, PA 19104,
other. In the mouse, germ cell competence is rise to primordial germ cells and certain so- USA.
induced at embryonic day 6.5 in proximal matic cells, such as extraembryonic meso- *To whom correspondence should be addressed. E-
epiblast cells by signals emanating from the derm and allantois. The potential of embry- mail:

Fig. 1. Analysis of early stages of germ cell formation. (A) Schematic to 2.5 104 cells/cm2 were plated without feeder cells or growth factors in
representation of the Oct4 reporter gene (gcOct4-GFP) showing four ES cell medium Bar scale for (B) and (C), 75 m. (D) Expression analysis of
conserved sequences (CR1 to 4) in the 5 regulatory region with the four distinct cell populations of d7 cultures sorted by uorescence-activated
deleted area boxed. The conserved sequences overlap with two regu- cell sorting (FACS) using markers GFP and c-kit. RT-PCR results for Oct4, the
latory elements: a distal enhancer (DE, germ cell specic) and a Oct4-GFP reporter, c-kit, Vasa, DMC1, SCP3, and -actin are shown. (E)
proximal enhancer (PE, epiblast-specic) that have been described Scheme of differentiating early germ cells in vitro. Only positive markers
previously (5). (B) Phase contrast image of E14 ES cells transformed are presented. (F) FACS analysis of a d9 culture using the gcOct4-GFP
with gcOct4-GFP and growing on an embryonic broblast feeder layer reporter. The R3 population was further sorted in (G) on the basis of
(42). The clone presented here resulted in 3.5 dpc and 6.5 dpc mice c-kit staining, which shows that, at d9, only a minor fraction of
embryos that lacked GFP expression but showed specic expression in gcOct4-GFP cells are also positive for c-kit. R5 represents the
germ cells of 12.5 dpc fetuses. (C) Merged uorescent and phase population of c-kit negative cells, whereas R2 denotes cells in which
contrast image of gcOct4-GFP ES cells 7 days (d7) after 1 104 c-kit is weakly or moderately expressed.

Oct4 (3, 4 ) was instrumental in the develop- (Fig. 1D, lane 3) were found to express Oct4, formed by d12 (Fig. 2, A to C, represent a
ment of a system that can visualize the initial gcOct4-GFP, and Vasa, but not c-kit, and large colony). Large colonies in general
steps of germ cell formation in vitro. Com- thus may represent cells of an early postmi- exhibited reduced GFP expression but con-
parative analysis of the mouse, human, and gratory germ cell stage. GFP/c-kit cells did tained a high percentage of Vasa cells
bovine Oct4 genes highlighted three con- not express Oct4, gcOct4-GFP, or c-kit (Fig. 2B). Within these colonies, three dis-
served regions, CR2 to CR4, that lie within mRNA but expressed Vasa mRNA (Fig. 1D, tinct types of cells were identified: (i) cells
the known germ cell (DE) and epiblast (PE) lane 4). These cells likely represent postmi- only expressing GFP, (ii) cells expressing
enhancers (Fig. 1A) (5). To restrict expres- gratory germ cells that are about to enter both GFP and Vasa, and (iii) cells only
sion of an Oct4-based reporter to germ cells meiotic prophase I, because these cells did expressing Vasa. Strings of GFP and
during mouse development, we deleted CR2 not express the meiotic markers DMC1 and Vasa cells were also found in the cultures,
and CR3 from a genomic Oct4 fragment driv- SCP3 (9). GFP/c-kit cells seem therefore reminiscent of migratory germ cells (Fig.
ing an inserted green fluorescent protein to exhibit the same expression pattern as that in 2D) (15). GFP/Vasa cells were always
(GFP) rather than Oct4 (gcOct4-GFP, Fig. vivo, in which both Oct4 and c-kit are down- found in the vicinity of both GFP/Vasa
1A) (6 ). ES cells were transfected with regulated in female germ cells before the zygo- and GFP/Vasa cells and may correspond
gcOct4-GFP, and three positive clones were tene-pachytene stage of meiotic prophase I, to early postmigratory germ cells in vivo
expanded and tested for specific expression around 15.5 days postcoitum (dpc) (13). The (Fig. 2B). GFP-expressing cells had nuclei
in transgenic animals. Two transgenic lines fourth group (GFP/c-kit) was composed of with a more diffuse DNA staining, whereas
showed specific gcOct4-GFP expression in cells that were not part of the germ line (Fig. cells solely expressing Vasa were more
germ cells and no signal in blastocysts or 1D, lane 1) (6, 14). Further analyses are re- condensed, round, and physically separated
epiblast-stage embryos (7 ). quired to define the identity of these c-kit cells from each other (Fig. 2A), which is typical
The same gcOct4-GFP ES cells (Fig. 1B) that 2 days later were almost absent (d9) (Fig. 1, of postmigratory germ cells. Outside the
that had been used to generate transgenic F and G). colonies, all cells were negative for germ
mice were subsequently used to establish the Early germ cell differentiation. In our cell markers and therefore most likely rep-
conditions required to induce germ cell dif- cultures, colonies of variable size had resent somatic cell types.
ferentiation in vitro. ES cells were plated on
tissue culture dishes and maintained in ES
cell medium without any feeder cells or
growth factors besides the factors present in
the heat-inactivated serum. Expression of
gcOct4-GFP was detected in some cells after
4 days (d4), in 25% of all cells at d7 (Fig.
1C), and in 40% of all cells at d8 (not shown).
Day 7 cultures were sorted into four cell
populations on the basis of the levels of
expression of the gcOct4-GFP reporter and
endogenous c-kit, both being markers of ear-
ly germ cells (Fig. 1D). After sorting, we
further characterized cells by determining the
mRNA expression of Oct4, gcOct4-GFP, c-
kit, Vasa (a marker of postmigratory germ
cells) (8), and two meiosis-specific markers,
namely the synaptonemal complex protein 3
(SCP3) and the mouse homolog of the yeast
meiosis-specific homologous recombination
gene (DMC1) (9). The SCP3 protein is part of
the axial-lateral element of the synaptonemal
complex to which the chromatin loops are
attached and is an excellent marker for detec-
tion of the meiotic transition in mammals
because its expression is required for the
onset of the first meiotic division (10). The
DMC1 protein is supposed to function during
chromosome synapsis and homologous re-
combination events (11, 12).
Three of the sorted cell populations
(GFP/c-kit, GFP/c-kit, and GFP/c-kit)
contained cells at different stages of germ cell
development (Fig. 1, D and E). The GFP/
c-kit fraction expressed Oct4, gcOct4-GFP,
and c-kit but little Vasa mRNA (Fig. 1D, lane Fig. 2. Large d12 colony with different early germ cell stages. (A) Fluorescent image showing DNA
2). Because Vasa is a marker for postmigra- localization (Hoechst staining, blue), (B) merged uorescent image of gcOct4-GFP (green) and
immunoreactive Vasa protein (red), and (C) phase contrast image. Boxed area in (B) is enlarged in
tory germ cells until postmeiotic stages (8), (D) to show a string of gcOct4-GFP cells (arrow). (E) Merged uorescent GFP and phase contrast
these results suggest that these cells corre- image of the supernatant from a d12.5 culture. (F) Merged uorescent (green, GFP; red, Vasa) and
spond to premigratory or migratory primor- phase contrast image of one aggregate with a single GFP cell. All pictures represent whole mount
dial germ cells. Cells sorted as GFP/c-kit stainings. Bar scales, 50 m in (A) to (C) and (E) and 25 m in (F).
Individual cells or groups of cells detached In the mammalian female gonad, the de- es between the two cell populations, includ-
simultaneously from the large colonies, but in velopment of a meiotically competent oocyte ing increased expression of P450 aromatase
both cases these predominantly Vasa germ requires the interaction of somatic cells, (2.4-fold), CYP17 (1.4-fold), and StAR (3.5-
cells tended to form small aggregates in the which work in cooperation to synthesize es- fold) in d16F cells. These higher levels are
supernatant, with very few GFP cells (Fig. 2, tradiol. Androgen precursors produced by likely the result of an increased proportion of
E and F). It is likely that these cells detach theca cells are aromatized into estrogen by follicular to nonfollicular somatic cells in the
because of the loss of cell-cell contacts, as granulosa cells. Detection of estradiol in our replated culture. This is also supported by the
noted in the lower center area of the colony cultures, therefore, provides evidence for sevenfold increase in growth differentiation
(Fig. 2, A to C). A reduction in cell-cell adhe- functional activity of the somatic cells in factor (GDF-9) in the replated cells as com-
sion among these Vasa germ cells in culture is these follicle-like structures. We consistently pared with attached cells (d16 and d16F, Fig.
quite similar to that of postmigratory germ cells found estradiol in the culture medium by d12, 3D). Increased expression of the oocyte-
in vivo (8). Interestingly, the aggregates were with levels peaking around d20. Thereafter, specific cell marker GDF-9 in the replated
frequently attached to GFP/Vasa cells, which estradiol levels decreased (Fig. 3A). To de- cultures is an interesting phenomenon, be-
often resulted in a more compact structure than termine whether estradiol production was as- cause this growth factor is required for ovar-
that within the original colony (Fig. 2F). These sociated with expression of steroidogenic en- ian folliculogenesis (20, 21). The expression
tightly knit structures are quite similar to the zymes, we performed quantitative reverse of steroidogenic enzymes and follicular estro-
histotypic structures obtained after disaggrega- transcription polymerase chain reaction (RT- gen synthesis is typically thought to require
tion by mild trypsin treatment of male or female PCR) for key genes involved in estrogen gonadotropin support. However, expression
genital ridges (16, 17). Furthermore, cultures of biosynthesis, including steroidogenic acute of steroidogenic enzymes may occur in the
suspended ovarian cells of newborn mice and regulatory protein (StAR), P450 17-hydrox- absence of gonadotropins, because transfec-
rat can also yield primordial follicles (18). ylase-17/20 lyase (CYP17), and P450 aro- tion of ES cells with steroidogenic factor1
Formation of follicular structures. Ag- matase (Fig. 3G) (19). The mRNAs of P450 (SF-1), for example, induces steroidogenic
gregates were collected by centrifugation and aromatase, CYP17, and StAR were 4.3-, 3.3-, enzyme expression (22). In addition, the se-
cultured in new plates. Well-organized struc- and 1.8-fold greater, respectively, in d22 at- rum and possibly endogenously produced go-
tures formed, and some of these were morpho- tached cultures compared with those at d16. nadotropins in the cultures may provide suf-
logically similar to early ovarian follicles (Fig. Increased expression of these genes correlat- ficient hormonal support.
3, B and C) (6). During the next 2 weeks, ed with the observed increased concentration Characterization of oocytes. As early
advanced follicle-like structures formed both in of estradiol in the medium. The steroidogenic as d26 of culture, oocyte-like cells were re-
the master plate and aggregate cultures (Fig. 3, gene expression profiles for attached cells at leased from the vicinity of their companion
E, F, and H). The majority of these structures d16 were compared with those of replated somatic cells (Fig. 3I) and found floating in
degenerated upon further cultivation, and cells at d16F 4 days after replating. This the supernatant. The oocyte-like cells were
20% produced oocytes larger than 40 m. comparison illustrated pronounced differenc- enclosed in a coat resembling the zona pel-

Fig. 3. Formation and char-

acterization of follicle-like
structures. (A) Media estra-
diol levels ( pg/ml) from cul-
tures between d7 and d34.
(B and C) Representative
phase contrast images of
structures that are morpho-
logically similar to early pri-
mary or secondary follicles.
(D) Quantitative RT-PCR
analysis of GDF-9 in d16 and
d22 cultures and in d16F-
replated cultures. (E), (F),
and (H) depict aggregate
cultures and show that many
of the follicle-like structures
maintain their three-dimen-
sional organization. (F) is a
magnied representation of
(E) at a different area of the
same culture. (I) represents
an adherent culture; all other
panels show nonadherent
cells and structures. (G)
Quantitative RT-PCR analy-
sis of aromatase, CYP17, and
StAR in d16 and d22 cultures
and in d16F-replated cul-
tures. (I) Phase contrast im-
age of oocyte-like cells re-
leased from d26 culture. All
pictures are whole mount
stainings. Bar scales: 15 m
in (B), 30 m in (C), 500 m
in (E), 50 m in (F), 40 m in
(H), and 100 m in (I).

lucida, which was fragile and easily lost when (Fig. 5C), suggesting that these cells are in a had been omitted (not shown). In large germ
manipulated with a micropipette (compare stage before leptotene. Germ cells up to 25 cells (Fig. 5A, black arrow), SCP3/COR1
Fig. 4, B and C). Most of these cells were 50 m in size showed SCP3/COR1 staining that (Fig. 5, G to I) and SCP1/SYN1 (not shown)
to 70 m in size, which is in the size range of colocalized with the nucleus (Fig. 5, A, were localized predominantly in the cyto-
natural oocytes (23). However, some looked black-white arrow, and D to F). In contrast, plasm, which is indicative of a more ad-
swollen, reaching a size of 130 m, and had SCP3/COR1 was undetectable in accompa- vanced meiotic stage (28). Addition of
a thinned zona (Fig. 4C). They were GFP, nying smaller cells (Fig. 5, D to F) or in germ gonadotropins, i.e., pregnant mare serum go-
in accordance with Oct4 being reexpressed cell controls in which the primary antibody nadotropin and human chorionic gonadotro-
after birth in diplotene-arrested oocytes (Fig.
4D) (13). The large cells also exhibited cyto- Fig. 4. Characterization of oo-
plasmic staining for zona pellucida proteins 2 cytes. (A) RT-PCR analysis of cul-
and 3 (ZP2 and ZP3, respectively) (24 ) at a tures at three different time
location predominantly adjacent to or in the points, compared with an ovari-
cell membrane, which is distinct from that of an newborn tissue control (P).
Comparison of expression levels
the GFP signal (Fig. 4E, ZP2 shown in red; of oocyte-specic markers ZP1,
ZP3 not shown). To support the immunocy- ZP2, ZP3, and Fig by RT-PCR
tochemical analysis, we examined the cul- shows that at d30 only ZP1 re-
tures for mRNA expression of several oo- mains absent. -actin served as
cyte-specific markers, including ZP1, ZP2, an internal standard. (B) Phase
ZP3, a factor in the germ line (Fig), and contrast image of a small oo-
cyte-like cell with holding pi-
GDF-9 (20, 21, 24 27). Fig, a transcription pette (right). (C) Phase contrast
factor required for the expression of ZP1, image of grown oocyte-like cell
ZP2, and ZP3, was absent in the ES control with holding pipette (right) and
cells (not shown) and expressed at similar injection needle (left) for size
levels between d16 and d30 (Fig. 4A). As comparison. (D) Fluorescence
expected, expression of both ZP2 and ZP3 image of gcOct4-GFP, and (E)
ZP2 immunostaining of the same
was also observed; however, ZP1 expression cell shown in (C).
was not detectable (Fig. 4A, lanes 2 to 4).
This may indicate that factors required for
specific expression of ZP1 are not properly Fig. 5. Germ cells
expressed in our cultures. Because ZP1 showing meiotic mark-
ers. (A) Image of puta-
serves as a linker for ZP2 and ZP3 (28), its tive germ cells selected
absence may account for the fragile zona from cultures on the
observed on the ES-derived oocytes. Loss of basis of size of cells.
ZP1 has been reported to perturb but not The white arrow marks
impair folliculogenesis (28). a cell that is represen-
Expression of the murine oocyte-specific tative for the size of
cells stained in (B) and
GDF-9 in the d16 and d22 adherent cultures (C); the black-white ar-
and the increase in GDF-9 mRNA levels in row, cells representa-
the replated aggregate cultures (d16F; with a tive for (D) to (F); the
high oocyte/somatic cell ratio) provide addi- black arrow, germ cells
tional evidence for folliculogenesis (Fig. 3D). representative for (G)
Loss of GDF-9 expression (15.2-fold de- to (I). Bar scale, 25 m.
(B) SCP3 staining of a
crease) between d16 and d22 in the adherent germ cell derived from
cultures is also consistent with loss of oocyte ES cell cultures in com-
GDF-9 expression as follicular growth occurs parison with a germ
(20). cell derived from a day
Evidence of meiosis. DMC1 expression 15.5 embryonal ovary.
and lack of SCP3 expression as determined (C) The highly decon-
densed chromatin ap-
by RT-PCR at d16 indicated that the ES- pearance suggests that
derived germ cells were about to enter mei- the cell in (B) is in
osis (6 ). To further substantiate entry into a preleptotene stage.
meiosis, we mildly disaggregated d16 cul- (D to F) represent a
tures with trypsin, replated them, and collect- germ cell showing nu-
ed single cells of varying sizes after the ma- clear SCP3 staining (D).
(G to I) illustrate a
jority of the cell population had reattached germ cell that has pro-
(Fig. 5A). Expression and distribution of gressed to a diplotene-
SCP3/COR1 (Fig. 5, B, D, F, and I) or SCP1/ like nuclear congura-
SYN1 (not shown) within these cells were tion. (J) Chromatin con-
analyzed with the use of the respective anti- densation surrounding
bodies (29, 30). Germ cells indistinguishable the nucleolus in large
oocyte-like cells. Images of oocyte-like cell directly after hormone-induced extrusion from follicle (K)
in size from somatic cells showed SCP3/ and after xation (L) showing a structure resembling the polar body within the zona. (A), (G), (K), and
COR1 staining (Fig. 5, A, white arrows, and (L) are phase contrast images; (D) is a uorescent image for SCP3/COR1 protein (red); (E), (H), and ( J)
B) that was very similar to that of female are stained for DNA (blue); (B), (C), (F), and (I) are merged uorescent images showing SCP3/COR1
germ cells isolated from day 15.5 embryos protein in relationship to DNA.

pin, to isolated follicles resulted in extrusion predominantly to the nucleus [Fig. 5C in lecular mechanism by which genes such as
of oocytes and formation of a structure that (31)], are characteristic for a mouse morula. c-mos block the second meiotic metaphase
morphologically resembles a polar body, sug- Several structures were found in our cultures arrest (22, 33, 34 ).
gesting that some oocytes in the cultures can that resembled blastocysts (Fig. 6, H to L, Summary and conclusions. It is not sur-
complete the first phase of meiosis (Fig. 5, J, show a representative collection of five dif- prising that the derivation of oocytes and
K, and L). ferent blastocyst-like structures). blastocyst-like structures could be accom-
Blastocyst-like structures derived It is likely that the structures resembling plished with both female and male ES cells.
from ES cells. At about d43, structures are preimplantation embryos represent parthe- In the absence of appropriate SRY expression
found in the cultures that, with respect to notes, as suggested by the similarity of our in the gonads, male primordial germ cells
morphology (Fig. 6, A to M) and expression follicle outgrowths (Fig. 6A) to parthenoge- enter the female pathway and often undergo
of molecular markers (Fig. 6, M and N), are netically activated oocytes of the LT/Sv the first step of oogenesis, entering meiotic
similar to mouse preimplantation embryos. A mouse strain (32). Future experiments will arrest at prophase I [for review see (35)].
defined zona around the embryo was only address whether cleavage was induced by the Male-to-female sex reversal even happens in
detected in a few cases (for example, Fig. culture conditions. It is well known that the mammalian embryo when single genes
6B), which is likely a consequence of its agents such as ethanol or cations or thermic such as Sry, Sox9, or Fgf9 are not properly
fragile structure. A zona was identified after shock of the oocytes during the observations regulated or deleted (36 ). Because we detect-
it had detached from the embryo (compare may induce parthenogenesis with or without ed Sry expression by RT-PCR only in the
Fig. 6, C and D, showing the same embryo on extrusion of a second polar body. Addressing later stages of our cell cultures (not shown),
2 subsequent days). This embryo had defined these and related questions may provide a the gene may not be efficiently regulated in
features of a 16-cell morula (Fig. 6, E to G). better understanding of the physiological the areas where germ cells differentiated into
Nuclear staining by Hoechst, the cytoplas- problems of parthenogenesis in vivo. In ad- female germ cells.
mic-to-nuclear ratio, a compacted morpholo- dition, our system may contribute to a better The reports from 1978 cited above in the
gy, and Oct4 protein expression and distribu- understanding of the regulation of spontane- context of dissociation and reaggregation of
tion, which at this stage starts to be localized ous ovarian teratocarcinogenesis and the mo- gonadal cells are interesting also in this con-

Fig. 6. Characterization of preimplantation embryos. Cleavage stagelike

embryos and blastocyst-like structures around d43 of culture are shown.
(A) A two-cell stage embryo within an outgrown follicle-like structure. (B)
A oating three-cell stage embryo. (C and D) Attached embryo at a
24-hour interval; note the detaching zona pellucida in (D) (upper right). A
16-cell embryo stained for DNA [blue, (E)] or immunoreactive Oct4
protein [red uorescence, (F)]; (G) is a merged phase contrast and Oct4 protein image. Four blastomeres were removed before xation [see holes in
(F)]. (H) Blastocyst-like embryo with reduced cell-cell contact within the structure. (I and J) Blastocysts lacking a characteristic attened outer
trophoblast cell layer, instead nuclei bulge out. This appearance is also found in natural blastocysts when the zona pellucida is removed before cleavage
of the embryo (not shown). (K and L) Two blastocyst-like structures that are morphologically indistinguishable from natural blastocysts. (M)
Immunocytochemical localization of Troma-1 to the outer layer of cells (green) and Oct4 protein (red) to inner cells that are acentrically located in
the blastocyst-like structure, a staining pattern typical of natural blastocysts. (A) to (D) and (H) to (L) are phase contrast images; (M) is a merged
uorescent image of the embryo shown in (L). (N) RT-PCR analysis of d43 cultures and ES cells under nondifferentiating conditions shows differential
expression of Hand1, Pl-1, and TpBp in reference to -actin expression. Oct4 is expressed in both samples, which reects the critical role of Oct4 in
the inner cell mass of blastocysts and in ES cells. Mash2, although essential for extraembryonic development (42), is also expressed in ES cells. Bar
scales: 100 m in (A), 40 m in (G), 30 m in (H), 25 m in (I), 45 m in ( J), 30 m in (K), and 20 m in (L).

text (16, 17). Disintegration and reaggregation References and Notes 36. J. S. Colvin, R. P. Green, J. Schmahl, B. Capel, D. M.
of male mouse and rat testes resulted in long, 1. K. A. Lawson, W. J. Hage, Ciba Found. Symp. 182, Ornitz, Cell 104, 875 (2001).
(1994). 37. E. Simpson, D. Scott, P. Chandler, Annu. Rev. Immu-
twisting tubular structures as well as spherical
2. K. A. Lawson et al., Genes Dev. 13, 424 (1999). nol. 15, 39 (1997).
aggregates (16, 17). One type of aggregate was 3. M. Pesce, M. K. Gross, H. R. Scholer, Bioessays 20, 7 38. S. Tanaka, T. Kunath, A. K. Hadjantonakis, A. Nagy, J.
indistinguishable in appearance from isolated (1998). Rossant, Science 282, 2072 (1998).
medium-sized ovarian follicles in suspension 4. M. Pesce, H. R. Scholer, Mol. Reprod. Dev. 55, 4 39. E. W. Carney, V. Prideaux, S. J. Lye, J. Rossant, Mol.
(2000). Reprod. Dev. 34, 357 (1993).
and thus termed folliculoids. The yield of 5. V. Nordhoff et al., Mamm. Genome 12, 309 (200 40. J. C. Chisholm, E. Houliston, Development 101, 565
folliculoids in these cultures was markedly in- 6. Materials and methods are available as support (1987).
creased when treated with antibodies specific material on Science Online. 41. M. Boiani, S. Eckardt, H. R. Scholer, K. J. McLaughlin,
for H-Y, an antigen that is responsible for the 7. K. Hubner et al., in preparation. Genes Dev. 16, 1209 (2002).
8. Y. Toyooka et al., Mech. Dev. 93, 139 (2000). 42. F. Guillemot, A. Nagy, A. Auerbach, J. Rossant, A. L.
differentiation of the undetermined gonadal an- 9. S. Chuma, N. Nakatsuji, Dev. Biol. 229, 468 (200 Joyner, Nature 371, 333 (1994).
lage into a testis [for review see (37)]. Because 10. L. Yuan et al., Mol. Cell 5, 73 (2000). 43. We thank J. Dean for generously providing the mono-
even cells of natural gonads form such follicu- 11. K. Yoshida et al., Mol. Cell 1, 707 (1998). clonal ZP2 and ZP3 antibodies, P. Moens and B.
loids, the formation of similar structures from 12. D. L. Pittman et al., Mol. Cell 1, 697 (1998). Spyropoulos for SCP3/COR1 and SCP1/SYN1 anti-
13. M. Pesce, X. Wang, D. J. Wolgemuth, H. Scho bodies, S. Eckardt and J. McLaughlin for providing a
male ES cells seems not too farfetched to us. Mech. Dev. 71, 89 (1998). female ES cell line, T. Noce for providing the poly-
Two main arguments have been put forth 14. M. Saitou, S. C. Barton, M. A. Surani, Nature 418, 2 clonal Vasa antibody, E. Gonzales for helping with the
to explain why ES cells are not totipotent in (2002). immunocytochemistry, and A. Malapetsas for editing
15. R. Anderson, T. K. Copeland, H. Scholer, J. Heasm the manuscript. The Troma-1 hybridoma antibody,
vitro. Mouse ES cells have not been observed C. Wylie, Mech. Dev. 91, 61 (2000). developed by P. Brulet and R. Kemler, was obtained
to give rise to germ cells in vitro, nor have 16. M. T. Zenzes, U. Wolf, E. Gunther, W. Engel, Cyto- from the Developmental Studies Hybridoma Bank
they been shown to differentiate along the genet. Cell Genet. 20, 365 (1978). developed under the auspices of the National Insti-
17. S. Ohno, Y. Nagai, S. Ciccarese, Cytogenet. Cell Gen- tute of Child Health and Human Development (of
trophectodermal lineage. Expression of tro- NIH) and maintained by the University of Iowa,
et. 20, 351 (1978).
phectodermal markers, like Troma-1 protein 18. A. McLaren, D. Southee, Dev. Biol. 187, 107 (1997). Department of Biological Sciences. Supported, in
(Fig. 6M) or Pl-1, Hand-1, and TpBp mRNAs 19. L. K. Christenson, R. L. Stouffer, J. F. Strauss 3rd, part, by the Marion Dilley and David George Jones
J. Biol. Chem. 276, 27392 (2001). Funds, the Commonwealth and General Assembly of
(Fig. 6L) (38 40), together with the mor- Pennsylvania, grant NIH 1RO1HD42011-01 (K.H.,
20. J. Dong et al., Nature 383, 531 (1996).
phology of the structures observed, strongly 21. M. M. Matzuk, Mol. Cell. Endocrinol. 163, 61 (2000). M.B., J.K., R.R., and H.S.), grant HD06274 (L.C., J.W.,
suggests that a trophoblast is formed from ES 22. P. A. Crawford, Y. Sadovsky, J. Milbrandt, Mol. Cell. and J.S.), grant T32 HD07305 ( J.W.), and grant 4434
(G.F.) from the Association pour la Recherche sur le
cells in vitro. Notably, our blastocyst-like Biol. 17, 3997 (1997).
structures looked better than most of those 23. J. J. Eppig, M. J. OBrien, Biol. Reprod. 54, 197 (1996).
24. I. J. East, D. R. Mattison, J. Dean, Dev. Biol. 104, 49
obtained by nuclear transfer of somatic nu- (1984).
clei. In addition, they had the correct Oct4 25. O. Epifano, L. F. Liang, M. Familari, M. C. Moos Jr., J. Supporting Online Material
Dean, Development 121, 1947 (1995).
expression, which is in contrast to most
26. S. M. Soyal, A. Amleh, J. Dean, Development 127, Materials and Methods
clones, and yet the latter are generally con- 4645 (2000). SOM Text
sidered to represent blastocysts (41). We con- 27. L. Liang, S. M. Soyal, J. Dean, Development 124, 4939 Figs. S1 and S2
clude that mouse ES cells are capable of (1997).
28. T. Rankin, P. Talbot, E. Lee, J. Dean, Development 126, 12 February 2003; accepted 23 April 2003
differentiating into oocytes and form struc- 3847 (1999). Published online 1 May 2003;
tures very similar, if not identical, to blasto- 29. M. Tarsounas, R. E. Pearlman, P. B. Moens, J. Cell Sci. 10.1126/science.1083452
cysts, thereby demonstrating that these cells 112, 423 (1999). Include this information when citing this paper.
are actually totipotent even in vitro. Future 30. M. J. Dobson, R. E. Pearlman, A. Karaiskakis, B. Spy-
ropoulos, P. B. Moens, J. Cell Sci. 107, 2749 (1994).
experiments will reveal whether the oocytes 31. S. L. Palmieri, W. Peter, H. Hess, H. R. Scholer, Dev.
we have generated in culture from ES cells Biol. 166, 259 (1994).
can be fertilized, whether they have under- 32. L. C. Stevens, D. S. Varnum, Dev. Biol. 37, 369 (1974).
33. J. J. Eppig, K. Wigglesworth, D. S. Varnum, J. H.
gone a gender-specific resetting of the epige- Nadeau, Cancer Res. 56, 5047 (1996).
netic marks (imprinting), and whether they 34. N. Hashimoto et al., Nature 370, 68 (1994).
can be used as starting material to derive ES 35. C. Tilmann, B. Capel, Recent Prog. Horm. Res. 57, 1
cell lines after nuclear transfer.


Consent from Donors for

Embryo and Stem Cell Research
Bernard Lo,* Vicki Chou, Marcelle I. Cedars, Elena Gates,
Robert N. Taylor, Richard M. Wagner, Leslie Wolf, Keith R. Yamamoto

esearch on human embryos and embryonic stem cells holds great promise for
understanding human reproduction and development and for regenerative medi-
cine. The need for informed consent from research participants is well estab-
lished (1). Under U.S. federal regulations, persons who provide biological materi-
als for research are research subjects who need to provide consent (13). In research
involving human embryos, informed consent is particularly important because of the
diverse opinions and strong emotions that surround these issues (4, 5). Some potential
donors consider all such research to be unacceptable; others only support some forms
of research (68). A donor might consider infertility research acceptable but object
to research that could lead to stem cell lines, patenting, or commercial products (9,
10). Governmental bodies in several countries have considered the issue of consent for
embryo and stem cell research (5, 1113).
In the United States, federal regulations permit a waiver of informed consent for the
research use of anonymous biological material that cannot be linked to donors even
through a code (2, 14). However, people commonly place great emotional and moral
significance on their reproductive materials (15, 16). Using gametes or embryos for
certain kinds of research without consent, even after identifiers have been removed,
could be regarded as a wrong or offensive (17, 18).
The consent of the woman or couple in the assisted reproductive technology (ART)
program is clearly required for research with frozen embryos remaining after comple-
tion of infertility treatment (19, 2022). Frozen embryos may be created with sperm
or oocytes from donors who do not participate any further in assisted reproduction or
child rearing. Guidelines in the United Kingdom and Canada require consent from
gamete donors, as well as infertility patients, for research with frozen embryos (12,
23). However, current U.S. guidelines do not consider whether these gamete donors
also have an autonomy-based interest in the use of such embryos for research (20,
21). The argument against requiring such consent is that in donating to ART patients,
gamete donors have ceded their right to direct further usage of their gametes, particu-
larly when they received financial compensation. However, gamete donors who are
willing to help women and couples bear children may object to the use of their genetic

B. Lo, V. Chou, and L.Wolf are in the Program in Medical Ethics; M. I. Cedars, E. Gates, and
R. N. Taylor are in the Department of Obstetrics, Gynecology, and Reproductive Sciences;
R. M. Wagner is in the Human Subjects Protection Program; K. R. Yamamoto is in the Depart-
ment of Cellular and Molecular Pharmacology; all are at the School of Medicine at the Univer-
sity of California, San Francisco, CA 94143, USA.

*To whom correspondence should be addressed.


materials for research. In one study, 25% of women who donated oocytes for infertility
treatment did not want the embryos created to be used for research (18). Little is known
about the wishes of sperm donors concerning research. Moreover, if stem cell lines are
created, it is the donors genetic material that will be propagated indefinitely.
We suggest that gamete donors wishes should be determined and respected; in-
formed consent from both oocyte and sperm donors should be obtained for an embryo
to be used in research (19). Our position does not depend on ascribing personhood to
embryos or on classifying gametes or embryos per se as research subjects. There are lo-
gistical and practical differences in obtaining consent to embryo research from oocyte
and sperm donors. ART clinics can readily discuss donation for research with oocyte
donors during visits for oocyte stimulation and retrieval. However, most ART clinics
obtain donor sperm from sperm banks and typically have no direct contact with the do-
nors. Commercial agencies and nonacademic ART centers may be reluctant to include
research in their consent discussions with sperm donors. The current consent form used
by one of the largest sperm banks in the United States makes no provisions for use of
sperm other than for the primary purpose of causing pregnancy (24). Furthermore,
sperm is often donated anonymously, with strict confidentiality provisions, and frozen
sperm must be quarantined for 6 months (25). Thus, recontacting sperm donors may
be difficult or impossible and may violate donor privacy. Sperm banks and researchers
need to collaborate to change the consent process for future sperm donation to include
consideration of donation for research.
Basic and clinical scientists, ART clinicians, and leaders of research institutions
should together stimulate broad public discussion to create guidelines for informed
consent that protect donors while allowing important research to proceed.

References and Notes

1. 45 Code of Federal Regulations (C.F.R.) 46 (1991). health and ethical perspectives; available at
2. National Bioethics Advisory Commission, Research
involving human biological materials: Ethical issues index_e.shtml; updated 13 May 2003; accessed 30
and policy guidance (National Bioethics Advisory June 2003.
Commission, Rockville, MD, 1999). 14. 45 C.F.R. 46.101(b)(4) (1991).
3. 45 C.F.R. 46.102 (1991). 15. J. Robertson, Children of Choice: Freedom and the
4., Public backs stem cell research, 26 New Reproductive Technologies (Princeton Univ.
June 2001. Press, Princeton, NJ, 1994).
5. Commission of the European Communities, Report 16. New York State Task Force on Life and the Law,
on human embryonic stem cell research (Office of Assisted reproductive technologies: Recommend-
Publications for the EU, Luxembourg, 2003). ations for public policy (New York State Task Force
6. Center for Public Policy, Virginia Commonwealth on Life and the Law, New York, 1998).
University, VCU Life Sciences Survey (2002). 17. J. Feinberg, Harmless Wrongdoing (Oxford Univ.
7. C. A. McMahon et al., Hum. Reprod. 18, 871 (2003). Press, New York, 1988).
8. House of Lords, Report from the Select Committee on 18. A. Kalfoglou, G. Geller, Fertil. Steril. 74, 660 (2000).
Stem Cell Research (27 February 2002); available at 19. National Institutes of Health, Report of the Human Embryo Research Panel (National Institutes of
ldselect/ldstem/83/8301.htm. Health, Bethesda, MD, 1994).
9. M. J. Radin, Contested Commodities (Harvard Univ. 20. National Bioethics Advisory Commission, Ethical is-
Press, Cambridge, MA, 1996). sues in stem cell research (National Bioethics
10. A. Kimbrell, The Human Body Shop: The Cloning, Advisory Commission, Rockville, MD, 1999).
Engineering, and Marketing of Life (Regnery 21. Ethics Committee of the American Society for
Publishing, Washington, DC, 1997). Reproductive Medicine, Fertil. Steril. 78, 957 (2002).
11. Department of Health and Human Services, National 22. Calif. Health and Safety Code, prec. 125115 (2003).
Institutes of Health guidelines for research using human, 23. Human Fertilisation and Embryology Authority
pluripotent stem cells. Fed. Regist. 65, 51975 (2000). (HFEA), Code of Practice (Human Fertilisation and
12. Ad Hoc Working Group on Stem Cell Research, Embryology Authority, London, 5th ed., 2001).
Canadian Institutes for Health Research (CIHR), 24. California Cryobank, Donor consent agreement
Human pluripotent stem cell research: Recom- (1997), private communication, 1 February 2003.
mendations for CIHR-funded research (2002); avail- 25. American Society for Reproductive Medicine, Fertil.
able at Steril. 77, S2 (2002).
stem_cell/stem_cell_guidelines_e.shtml. 26. B. Lo, L. Wolf, and V. Chou are supported by the
13. CIHR, Human stem cell research: Opportunities for Greenwall Foundation.

BD Matrigel Matrix: A proven substrate for feeder-free culture of human
embryonic stem (hES) cells
By Suparna Sanyal, Ph.D. and Susan Qian, Ph.D.
BD Biosciences Discovery Labware, Billerica, MA, USA

Introduction laminin, fibronectin, and collagen IV as substrates for culturing various

Embryonic stem cells have an infinite capacity for self-renewal and are hES lines (WiCell Research Institute, Inc.) with MEF-conditioned media
pluripotent (i.e. can differentiate into any cell type represented by the (MEF-CM). BD Matrigel Matrix and laminin were equivalent in their
three germ layersectoderm, endoderm, and mesoderm). A great ability to support undifferentiated hES cell growth for extended periods
deal of hope is associated with the potential application of hES cells in of time. Cells grown under these conditions were found to have a
functional genomics, cell therapy and regenerative medicine. normal karyotype, demonstrated a stable proliferation rate, and high
Successful derivation of hES cells was first reported in 1998.1 In that telomerase activity. Cells expressed characteristic undifferentiated
study, hES cells were isolated from the inner cell mass of blastocysts hES cell markers and formed teratomas in severe combined immuno-
and plated onto mitotically inactivated murine embryonic fibroblast deficient (SCID) mice and differentiated into cells from all three
(MEF) feeder cells. Initial hES cell derivation and culturing techniques germ layers. Since then, BD Matrigel Matrix coupled with different
originated predominantly from methodology developed for murine feeder-conditioned media has been widely accepted as an alternative
ES (mES) cells. While MEF feeders have proved to be a robust surface method for culturing hES cells. More recently, others demonstrated
for long-term culture of hES cells, they also present a number of that hES cells could be propagated in an undifferentiated state on
limitations. Concerns about contamination of hES cells from animal- BD Matrigel Matrix using conditioned media from different types
derived pathogens makes this method of culturing unsuitable for of feeders.3-5 The major limitations of using conditioned media are
clinical applications. From a practical standpoint, it is inconvenient that it involves a tedious process of generating media from feeder
and tedious to grow and maintain two cell types. Furthermore, it cells and can vary in performance depending on the condition and
is difficult to transfect and genetically manipulate the compacted type of feeders used. Therefore, the logical next step in feeder-free
hES cell colonies on feeders. For all these reasons, efforts have been media development was focused on supplementing basal media
initiated to develop feeder-free conditions for culturing hES cells. with key factors necessary to maintain self-renewal of hES cells.

To date, many signaling pathway regulators have been identified that bFGF is known to play a central role in maintenance of hES cell
are thought to play an important role in mediating self-renewal and self-renewal.6 Recent studies illustrated that hES cells cultured on
differentiation of ES cells. These include bFGF, Wnt, Activin A/Nodal, BD Matrigel Matrix with unconditioned media (UM) supplemented
and BMP. Manipulation of specific signaling regulators has facilitated with knockout serum replacement and high concentrations of bFGF
development of feeder-free conditions required for propagating was sufficient for maintaining self-renewal.7-9 BMP activation has
undifferentiated hES cells. Recent studies have clearly delineated that been shown to promote differentiation of hES cells.10 Noggin and
hES cells can be grown in the absence of feeders as long as they other BMP antagonists have been found in hES media conditioned
are plated on an appropriate extracellular matrix (ECM) and cultured by MEFs and are thought to be important for reducing differentiation
in media either conditioned on feeder cells or supplemented with of hES.11 A combination of exogenous Noggin and bFGF appears
appropriate soluble factors. to have a synergistic effect in suppressing hES differentiation and
can be used to supplement unconditioned media and maintain
hES cells in an undifferentiated state for long periods of time on
Discussion BD Matrigel Matrix.11,12 Another signaling pathway deemed important
BD Matrigel Matrix is a reconstituted basement membrane isolated for maintaining self-renewal of hES cells is the Wnt pathway. A
from the Engelbreth-Holm-Swarm (EHS) mouse sarcoma. This matrix specific inhibitor of glycogen synthase kinase-3, 6-bromoindirubin-3-
is predominantly composed of laminin, collagen IV, entactin, and oxime (BIO), which activates the canonical Wnt pathway, maintained
heparin sulfate proteoglycan. In 2001, the first feeder-free hES culture an undifferentiated and pluripotent phenotype in hES cells in the
system was demonstrated.2 The authors compared BD Matrigel Matrix, absence of feeder-conditioned media on BD Matrigel Matrix.13

Table 1: Summary of BD Matrigel Matrix as ECM substrate for feeder-free human ES cell culture.

Cell lines or SR Media Characterization* Culture Duration** Reference

CM or Key components Marker Pluripotency Karyotype


H1, H7, H9 and H14 MEF-CM Up to 6 months 2***

H7, H9 HEF1-CM Up to 14 passages 3***
H1 and HES-NCL1 HES-df-CM Up to 14 passages 4***
SA 002, AS 038, SA121, SA167 k-VitroHES media Up to 35 passages 5
H7, H9 40 ng/ml of bFGF +/- other GFs 15 passages 7***
H1, H9 36 ng/ml of bFGF Up to 35 passages 8
H1, H9, H14 40 ng/ml of bFGF + 500 ng/ml of Noggin Up to 34 passages 11***
H1 NIH/3T3-Nog-CM, 40 ng/ml bFGF+500 ng/ml Noggin 7 passages 12***
H1, H9 >100 ng/ml of bFGF Up to 7 months 9
H1, H7, H9, H14 No TeSR1 media including 100 ng/ml of bFGF Up to 6 months 14

Abbreviations: SR: serum replacement; Mkr: Marker; Pluri: Pluripotency; Kary: Karyotype; TER: Teratoma formation in SCID mice; EB: Embryoid body formation; DF: Differentia-
tion to specic lineage; Ref.: Reference; MEF: Mouse embryonic broblast; CM: Conditioned media; HEF1-CM: Conditioned media from mitotically inactivated HEF1-hTERT cells
(human ES derived broblast, stably transfected with TERT); HES-df-CM: Conditioned media from hES cell derived broblast. k-VitroHES media: MEF conditioned VitroHES
media. *: Characterization may be reported on selected cell lines listed; **: The longest culturing time for one of the cell lines listed; ***: Table entry is modied from Table 1
in Ref. 17
Recently, BD Matrigel Matrix has been shown to support long-
term culture of hES cells with three independent chemically defined a) b)
media.14-16 One of the defined media (TeSR1) contains all human
supplements and was completely defined (except for human serum
albumin). TeSR1 media coupled with BD Matrigel Matrix has been
a successful combination for culturing different WiCell hES lines
for two-six months. In the same study, various ECMs (collagen
IV, fibronectin, laminin, and vitronectin) were combined to form
a human surface capable of supporting long-term proliferation
of hES cells using this media. Another defined media15, when
used with BD Matrigel Matrix, sustained the highest number
of undifferentiated hES cells compared to all other single ECM c) d)
substrates tested (collagen, fibronectin, and laminin). In this study,
a combination of fibronectin and collagen was required to match
the performance of BD Matrigel Matrix. A summary of BD Matrigel
Matrix applications for culturing hES cells is provided in Table 1.
There are a number of advantages in using BD Matrigel Matrix as a
feeder-free substrate for culturing hES cells. For instance, DNA, RNA,
and protein isolation is easier on the BD Matrigel Matrix surface due
to lack of potential contamination from the feeder cells. In addition, Figure 1. Representative phase contrast
e) images of H9 hES cells grown on a) MEF
genetic manipulation of hES cells requires efficient transfection capa- feeders b) BD Matrigel Matrix-coated
bilities. Transfection efficiencies of hES cells grown on a BD Matrigel plates. Immunofluorescent images
Matrix surface are improved compared to those grown on feeders. of H9 cells grown on BD Matrigel
Matrix-coated plates, stained positive
When cultured on BD Matrigel Matrix, hES cells form monolayer-like for undifferentiated markers c) Oct-
colonies, making individual cells more accessible for penetration with 3/4 d) SSEA (Green: hoechst nuclear
DNA or SiRNA.18-21 Although other ECM proteins have also been suc- stain, blue). Immunoflurescent images
were acquired on the BD Pathway
cessfully used for feeder-free cultures, several studies have suggested Bioimager using a montage feature to
that a combination of ECM proteins is required to perform as well as capture a larger portion of colonies. e)
BD Matrigel Matrix.16,17 Moreover, purified proteins are generally more Embryoid bodies formed upon plating
H9 cells cultured on BD Matrigel Matrix-
expensive and not as well tested for long-term performance as coated plates onto a low-attachment
BD Matrigel Matrix. surface in media supplemented
with FBS and without bFGF.
Despite the wide acceptance of BD Matrigel Matrix as a viable sub-
strate for long-term, feeder-free culturing of hES cells, it has a key
equivalent, and more often superior, performance compared to other
limitation. Since BD Matrigel Matrix is derived from the EHS tumor,
single ECMs tested. Two independent studies have demonstrated that
lot-to-lot variability in protein concentration and composition is inher-
a combination of ECMs was required to match the performance of
ent. In order to limit the amount of time researchers spend today
BD Matrigel Matrix as a substrate for maintaining integrity of hES
pre-screening lots of BD Matrigel Matrix and coating plates on a
cells in culture. BD Matrigel Matrix is a more affordable option
weekly basis, BD Biosciences has developed a stable, ready-to-use
compared to purified ECMs. The BD BioCoat Matrigel 6-well
BD BioCoat Matrigel 6-well Plate for ES culture. This culturing
Plate (Cat. No. 354671) provides lot-to-lot consistency needed
system was optimized using MEF-CM due to a lack of commercial
for standardized cultures. Many basic research questions can
media that is qualified and defined. Much of hES cell culturing success
be addressed by using recently developed media in conjunction
depends on methods used for ES cell derivation, dissociation, and cryo-
with the BD Matrigel Matrix-coated culture surface.
preservation. Moreover, protocols that have been validated over long
periods of time are essential. We have optimized a culturing protocol
using BD BioCoat Matrigel 6-well Plates which includes pre-qualified
1. Thomson, J.A., et al., Science 282:1145 (1998).
reagents such as BD human bFGF as a media supplement. Using this 2. Xu, C., et al., Nature Biotechnology 9:971 (2001).
protocol, H9 and H1 cell lines from WiCell have been cultured on the 3. Xu, C., et al., Stem Cell 22:972 (2004).
optimized BD BioCoat Matrigel Matrix surface for over 40 population 4. Stojkovic, P., et al., Stem Cells 23:306 (2005).
doublings with MEF-CM supplemented with 8 ng/ml bFGF (Figure 1). 5. Sjgren-Jansson, E., et al., Developmental Dynamics 233:1304 (2005).
hES cells typically spread out more on the BD BioCoat Matrigel surface 6. Amit, M., et al., Developmental Biology 227:271 (2000).
and form larger colonies (Figure 1b) compared to those cultured on 7. Xu, C., et al., Stem Cells 23:315 (2005a).
8. Wang, L., et al., Blood 105:4598 (2005a).
MEF feeders (Figure 1a). Cellular morphology was comparable to colo-
9. Levenstein, M.E., et al., Stem Cells 24:568 (2006).
nies cultured on freshly coated BD Matrigel Matrix Plates and consis- 10. Xu, R.H., et al., Nature Biotechnology 20:1261 (2002).
tent with previous reports.2,3 Cells grown on the newly optimized 11. Xu, R.H., et al., Nature Methods 2:185 (2005b).
BD BioCoat Matrigel Matrix surface stained positive for expression 12. Wang, G., et al., Biochemical and Biophysical Communications 330:934
of well-characterized, undifferentiated markers (Oct-3/4 and SSEA-4) (2005b).
(Figure 1c, d). H9 cells cultured continuously on this surface for 10 13. Sato, N., et al., Nature Medicine 10:55 (2004).
passages retained their ability to form embryoid bodies (Figure 1e) 14. Ludwig, T.E., et al., Nature Biotechnology, Brief Communications, January
suggesting that the ability of these cells to differentiate has not been 2006.
15. Lu, J., et al., PNAS 103:5688 (2006).
altered. Independent studies suggest that the BD Matrigel Matrix sur-
16. Yao, S., et al., PNAS 103:6907 (2006).
face will likely be compatible with a number of hES culturing media 17. Mallon, B.S., et al., Int. J. Biochem. & Cell Biol. 38:1063 (2006).
(Table 1) and that this surface may also be used to culture other types 18. Lakshmipathy, U., et al., Stem Cells 22:531 (2004).
of ES cells.22 19. Gerrard, L., et al., Stem Cells 23:124 (2005).
20. Ren, C.P., et al., Acta Biochim Biophys Sin (Shanghai) 37:68 (2005).
Conclusion 21. Zaehres, H., et al., Stem Cells 23:299 (2005).
In summary, BD Matrigel Matrix has served as an optimal surface for 22. Greenlee, A.R., et al., Toxicology In Vitro 19:389 (2005).
propagating hES cells, maintaining their self-renewal and pluripotency
characteristics for extended periods of time in culture. To the best of
our knowledge, it has worked with every type of media tested thus
far, feeder-conditioned or defined. BD Matrigel Matrix has provided an

We would like to thank Jennifer DeKarski, Jeff Partridge, and Jennifer Brown for their contributions in developing the BD BioCoat Matrigel 6-well plate for culturing ES cells.
Unless otherwise specified, all products are for research use only. Not intended for use in diagnostic or therapeutic procedures.
BD, BD logo, and all other trademarks are the property of Becton, Dickinson and Company. 2006 BD
The future of

is here.
Stem Cell Sorting on the BD FACSAria Cell Sorter
NEW 375 nm Near UV Laser Option

Spot the side-population "hummingbird."

Resembling a hummingbird, the so it can be purchased as an option
characteristic Hoechst 33342 prole for an instrument or as an upgrade
depicts the enrichment of stem cells to a current system. Call your BD
from a wide variety of tissues. With the sales representative or visit
introduction of the 375 nm near UV laser,
this application is now available on the for more information.
BD FACSAria cell sorter.
Shop online at
The small laser ts in the instrument
without changes to the optical bench,
BD Biosciences
2350 Qume Drive
Unless otherwise specied, all products are for Research Use Only. San Jose, CA 95131
BD, BD Logo and all other trademarks are the property of Becton, Dickinson and Company. 2006 BD.
You spend your whole life looking for an
answer. Why not spend a few minutes
looking at how we can help.

Stem Cell Source

From isolation to analysis

Stem Cell Source delivers:

Thousands of proven reagents and systems for cell analysis
Advanced functional assays
Dependable innovation in cell culture technology
Over 20 years of cell separation leadership

Contact us today at

to see how we can help you find answers.

BD Biosciences
2350 Qume Drive
BD, BD Logo and all other trademarks are the property of Becton, Dickinson and Company. 2006 BD. San Jose, CA 95131