Two-step oligoclonal development of male germ cells
Hiroo Uenoa,1, Brit B. Turnbullb, and Irving L. Weissmana,1
aDepartments of Pathology and Developmental Biology, Institute of Stem Cell Biology and Regenerative Medicine, Ludwig Institute at Stanford University,
andbDepartment of Health Research and Policy, Division of Biostatistics, Stanford University, Stanford, CA 94305
Contributed by Irving L. Weissman, October 30, 2008 (sent for review June 16, 2008)
During mouse development, primordial germ cells (PGCs) that give
rise to the entire germ line are first identified within the proximal
epiblast. However, long-term tracing of the fate of the cells has not
been done wherein all cells in and around the germ-cell lineage are
identified. Also, quantitative estimates of the number of founder
PGCs using different models have come up with various numbers.
Here, we use tetrachimeric mice to show that the progenitor
numbers for the entire germ line in adult testis, and for the
initiating embryonic PGCs, are both 4 cells. Although they prolif-
erate to form polyclonal germ-cell populations in fetal and neo-
natal testes, germ cells that actually contribute to adult spermat-
ogenesis originate from a small number of secondary founder cells
that originate in the fetal period. The rest of the ‘‘deciduous’’ germ
cells are lost, most likely by apoptosis, before the reproductive
period. The second ‘‘actual’’ founder germ cells generally form
small numbers of large monoclonal areas in testes by the repro-
ductive period. Our results also demonstrate that there is no
contribution of somatic cells to the male germ cell pool during
development or in adulthood. These results suggest a model of
2-step oligoclonal development of male germ cells in mice, the
second step distinguishing the heritable germ line from cells
selected not to participate in forming the next generation.
stem cells ? testis ? premordial germ cells ? apoptosis chimeras
accurately to the next generation. It was initially proposed (1), and
later shown that they usually originate as a very small founding
population that is segregated from somatic cells early in develop-
ment, at least in organisms where the overall body plan is estab-
lished early (2, 3). However, retroviral marking of mouse early
embryos at the 4- to 16-cell stage suggests that at least 3 cells
contribute to the germ line and are set aside before somatic tissue
monoclonal. In species such as Drosophila and Caenorhabditis
elegans, germ plasm, a unique cluster of mRNAs and organelles, is
a good marker to follow the origin and development of early germ
cells (3). However, there is no visible germ plasm in mammals,
making it difficult to follow early populations of germ cells (5). It
has been shown that inner cell mass (ICM) or epiblast cells earlier
and germ cells (6–8).
Tissue nonspecific alkaline phosphatase (TNAP) has long been
used as a marker for candidate primordial germ cells (PGCs) (9,
10). Also, many markers for PGCs have been identified, including
SSEA-1 (11), Oct3/4 (12), Stella/PGC-7 (13, 14), Fragilis (14), and
Blimp-1 (15). By using these markers, earlier candidate PGCs that
cannot be identified with TNAP could be detected. It has been
(14), and that progenitors for PGCs emerge as Blimp-1-positive
cells in the epiblast as early as E6.25 (15). At ?E7.5, 30–50
of the developing allantois, posterior to the primitive streak. They
proliferate rapidly, and at ?E8.5, they start migrating. At E9.5,
?120 TNAP-positive cells are identified at the wall of the hindgut
and split to the left and right genital ridges; ?4,000 TNAP-positive
cells are observed at E12.5 (16). Male and female germ cells start
differentiating within the genital ridges at ?E13.5 (9, 16). In male
he germ line is made up of a highly protected and strictly
regulated group of cells that transmit genetic information
mice, the proliferating PGCs are enclosed by Sertoli cell precursors
by E13.5 to form testicular cords, a primitive form of seminiferous
Although these scenarios are well established, many questions
are unanswered. First, it is unclear whether all of the candidate
PGCs contribute to adult spermatogenesis. Also it has been dem-
sis in the testis during the prepubertal period, although it has not
phenomenon remains unclear; however, inhibiting the apoptosis by
overexpression of Bcl-2 (18, 19) or disruption of Bax (20) are both
known to cause male infertility, indicating that programmed death
of some fraction of germ cells is essential for establishment of
the visualization and categorization of all cells in a tissue (21), we
have analyzed lineage relationships between the earliest PGCs and
adult germ cells that maintain spermatogenesis.
Chimeric Testes in Tetrachimeric Mice. In our method, 3 kinds of
Rosa26 knock-in fluorescent ES cells are injected into uncolored
blastocysts to form tetrachimeras (21). We applied the method to
an analysis of development of male germ cells. In each single-color
mouse generated from our knock-in ES cells, all of the germ cells
[supporting information (SI) Fig. S1] and somatic cells (data not
shown) express fluorescent markers without down-regulation.
The R1 ES cells (22) we used in this study were generated from
our R1 cells have a normal male karyotype, 40XY (Fig. 1A).
Depending on the karyotype of the host blastocyst, 4 situations are
possible for the resulting chimeric mice. (i) If the host blastocyst is
male, the resulting chimeric mouse is male. If the host blastocyst is
female, 3 situations are possible. (ii) If contribution of ES cell-
in the urogenital ridges, is sufficiently large, chimeric mice develop
into males (23). In this case, host-derived 40XX PGCs can develop
into prospermatogonia (24, 25), however, they die within the first
few days postpartum (24, 26). In the 40XY/40XX chimeric testes,
40XX-derived oocytes that are nonfunctional are observed in
seminiferous tubules (26, 27). (iii) If contribution of ES-derived
(23). (iv) Other than these 3 possibilities, they can develop into
male mice, cases i and ii above, for the analysis of testis germ cells,
and we checked the karyotype of host cells by FISH analysis with
a Y chromosomal probe to distinguish cases i and ii (Fig. S2). The
mice whose testes we analyzed and the results of Y chromosome
FISH are summarized in Dataset S1.
Author contributions: H.U. and I.L.W. designed research; H.U. performed research; H.U.
The authors declare no conflict of interest.
whom correspondencemay beaddressed.E-mail: firstname.lastname@example.org
This article contains supporting information online at www.pnas.org/cgi/content/full/
© 2008 by The National Academy of Sciences of the USA
January 6, 2009 ?
vol. 106 ?
no. 1 ?
Spermatogenesis in Adult Testis Is Maintained by a Small Number of
Monoclonal Germ-Cell Clusters.Weshowexamplesofchimerictestes
older than 4 weeks in Fig. 1 and Fig. 2. We surveyed 32 chimeric
testes and found that (i) germ cells within seminiferous tubules
generally form large single-color domains, and that (ii) in most
cases, germ cells have fewer colors than the surrounding mesen-
chymal tissues (Fig. 1B; Dataset S1). We verified by the germ-cell
marker TRA98 that cells forming single-color patches within
seminiferous tubules are germ cells (Fig. 1C). However, the inter-
stitial tissues between seminiferous tubules (including vessels, fi-
broblasts, myoid cells, and Leydig cells) generally have all of the
colors that the mouse has (Fig. 1 B and C). In a magnified picture,
Sertoli cells that have projections and attach to the wall of the
basement membrane were observed at the periphery of the semi-
niferous tubules, and they generally had all of the colors that the
also checked the color of epithelial cells in the epididymis and
confirmed that they always had all of the colors (Fig. 1F).
By untangling seminiferous tubules, such single-color domains
tend to span multiple tubules (Fig. S3 D–H). Also, single-color
patches of the germ cells of the same color tend to be present
through adjacent serial sections (Fig. 1B; Fig. S3); 1 tetrachimeric
other 2 or 3 color chimeric testes, the germ cells with particular
Testis Germ Cells Are Generally Composed of Fewer Colors than the
Whole Mouse in Chimeric Mice. We show examples of sections of
colors of pups generated from chimeric male mice, and we verified
that they matched the colors of the testis germ cells in the cases we
used are heterozygous, half of the resulting pups were nonfluores-
cent, as expected (Fig. S4). We verified that fluorescent markers
distribute equally in whole mice (Fig. 2E) and in testes (Fig. 2F).
It should be noted that left and right testes always had the same
combination of colors of germ cells (Fig. 2 A–D; Dataset S1),
consistent with the hypothesis that they originate from the same
population of progenitors and that the selection of colors occurs
before they split to bilateral genital ridges. Two models
to explain the finding are discussed in SI Results (see also Fig. S5).
Embryonic PGCs Are Composed of Fewer Colors than the Whole
Embryo. To test whether the germ line is initiated from a small
number of ‘‘founder’’ PGCs, as in the ‘‘oligoclonal development’’
model (Fig. S5A), we next examined whether ?4 color PGCs are
found in embryonic PGCs of fully tetrachimeric embryos. If our
hypothesis is correct, we should generally observe ‘‘fewer color’’
to identify candidate PGCs (Fig. S6). At the early bud stage, the
signal for Stella was less bright and the shape of Stella-positive cells
was not as characteristic as in later PGCs (Fig. S6A). The Stella-
positive cells exist in a tighter cluster in early- to mid-bud stage, but
they start scattering after the late-bud stage (Fig. S6 B–F). These
Note that the interstitial cells have green, blue, and red cells. (C) Single-color (green) patches of germ cells occupying seminiferous tubules in a tetrachimeric mouse.
epididymal epithelial cells from a tetrachimeric mouse. (Scale bars, 1 mm in B; 100 ?m in C–F.)
Germ cells in testis form large single-color areas after reproductive period. (A) Normal male karyotype of R1 ES cells used in this study. The image of the SKY
www.pnas.org?cgi?doi?10.1073?pnas.0810325105 Ueno et al.
colors than the whole embryo (Fig. S6; Dataset S1). These results
further support our hypothesis that germ cells generally originate
from a restricted number of progenitors.
Estimation of Progenitor Numbers for Adult Testis Germ Cells and
Embryonic PGCs. We then statistically estimated the number of
progenitors for male germ cells from the frequency of 1- to 3-color
estimate, it is important to define what the progenitors are for the
tissue. As described, it has been shown that ICM or epiblast cells
earlier than E5.5 can contribute to any lineage of somatic and germ
cells (6, 8); we suppose that a subset of clonogenic ICM or epiblast
cells are founder cells in our progenitor number estimate. For this
analysis, we made several assumptions. (i) We selected chimeric
mice with balanced colors wherein the contribution of ES cells is
?75% by analysis of the hair colors. Under this condition, our
results best matched the model that contribution of host ICM-
derived nonfluorescent cells is twice as much as each ES clone
(R:C:G:NF ? 1:1:1:2) from the results of endodermal organs (see
Materials and Methods). (ii) From the background described above,
if the host blastocyst is female, host-derived nonfluorescent germ
cells cannot survive in the testis, which matches with our results
(Dataset S1). In these cases, it is not possible retrospectively to
know whether host-derived cells had contributed to the germ line
and had died, or they could not have contributed to the germ
line from the beginning. Because these two possibilities affect the
fate in adulthood. (iv) As discussed above, we assume that germ
cells develop exclusively from PGCs and are not supplied from
Under these assumptions, our results show that germ cells are
generated by ?4 progenitors, estimating from both adult testis and
embryonic PGCs (Table 1). Changing the ratios of contribution of
host ICM-derived cells as described did not considerably affect the
results (data not shown).
testes germ cells in a tetrachimeric mouse. (B) Two-color (ECFP and nonfluorescent) testes germ cells in a tetrachimeric mouse. (C) Three-color (mRFP1, ECFP, and
nonfluorescent) testes germ cells in a tetrachimeric mouse. (D) Three-color (mRFP1, ECFP, and EGFP chimeric germ cells) testes germ cells in a tetrachimeric mouse. (E)
Difference of distribution of mRFP1 (R)-, ECFP (C), and EGFP (G)-expressing cells in adult male chimeric mice was not significant. (F) Difference of distribution of R-, C-,
Table 1. Estimates of the progenitor numbers for germ cells
nN ? 1
N ? 2
N ? 3
N ? 4
N ? 5
N ? 6
N ? 7Estimate
Testis, germ cells*
Testis, germ cells†
Testis, Leydig cells
Epidydimis, epithellial cells
The posterior probabilities for the progenitor numbers are shown. Boldface indicates the most probable values. OR, out of range.
*Estimate assuming that all testes with female blastocysts had NF germ cells, but they died.
†Estimate assuming that all testes with female blastocysts did not have NF germ cells from the beginning (see text).
Ueno et al.
January 6, 2009 ?
vol. 106 ?
no. 1 ?
Formation of Monoclonal Cell Clusters of Germ Cells in Testis Suggests
Secondary Selection of Germ Cells. By surveying PGCs during
development, we noticed that PGCs are observed as a mixed
population with different colors during migration from allantois to
ridges and testicular cords (Fig. 3 B and C). Therefore, we analyzed
the mechanism by which large single-color patches in adult testes
are formed. Interestingly, the single-color patches of testis germ
cells are not observed in neonatal or prepubertal testes (Fig. 4). In
adult testes during the reproductive period, spermatogonia on the
transverse sections of seminiferous tubules are always the same
color within a tubule (Fig. 1C and data not shown). However, in
testes between postnatal day (P)0 and P14, a mixture of ?1 color
of germ cells is frequently observed within a seminiferous tubule
(Fig. 4 A–E). At 2 weeks, single-color areas of germ cells start to
form (Fig. 4C). Although at ?P14 examination at a high magnifi-
cation each seminiferous tubule generally has a mixture of 2 colors
of germ cells, the proportions of the 2 colors are biased to 1 color
in such areas (Fig. 4 D and E). However, interestingly, different
colors are selected in distinct regions of the testis (Fig. 4 C and E).
long-term germ stem cells (28, 29). In the area where 1 color is
selected, NGN3-positive cells tend to have the selected color (Fig.
4F). However, clusters of apoptotic germ cells were frequently
observed at this stage of testis, and each cluster tends to have the
color(s) not selected in the area (Fig. 4G and data not shown). The
apoptotic cells were observed during neonatal or prepubertal
periods (data not shown), indicating that the apoptosis occurs
gradually. The finding suggests that a small fraction of germ cells
present at birth are long-term stem cells and their progeny form
clones are lost by apoptosis. From these findings, we propose that
germ cells that actually contribute to adult reproduction are prog-
eny of a small fraction of PGCs, which we call secondary founder
germ cells. The secondary founder germ cells represent the same
colors initially found in the 4 founder PGCs, and so are not
contributed by other candidate founders.
The next question is when the secondary founder germ cells are
selected. Interestingly, even before local selection of particular
color-expressing spermatogonia is initiated, NGN3-positive cells
that appear at ?P3 already form single-color domains (Fig. 4H and
data not shown). This finding means that each same-color NGN3-
positive cluster derives from cluster-initiating common progenitors
in which the selection occurs at an earlier stage. As shown in Fig. 3,
during migration from allantois and proliferation in genital ridges,
PGCs are observed as a mixed population. Because PGCs are
enclosed and separated by Sertoli cells by E13.5 (17), and our data
indicate that single-color patches generally span multiple tubules
occur after they seed onto genital ridges, but at least some before
PGCs are enclosed and separated by Sertoli cell precursors by
E13.5. One possibility is that a subset of PGCs find or compete for
a spermatogonial stem cell niche, and only those cells that inhabit
the niche self-renew.
Our model of 2-step oligoclonal development of male germ cells
is shown in Fig. 5. We discuss the number of the secondary founder
cells in Discussion.
Oligoclonal Development of the Germ Line. By a series of studies,
Surani and colleagues (15) have shown that the progenitors for
PGCs can be identified as Blimp-1-positive epiblast cells as early as
E6.25, and it seems that they initially appear as a small number of
cells. Although the fate of the Blimp-1-positive cells in the epiblast
has not been examined after birth, the results strongly suggest that
the fate decision of germ cells occurs before the prestreak stage of
showed contribution of 4 different clones to testis germ cells
(Dataset S1), indicating that at least 4 different clones can contrib-
ute to the whole germ line. By our statistical method, we showed
that 4 founder cells contribute to both embryonic PGCs and adult
testis germ cells, estimated by the fate of the cells. Although these
are estimates, the standard errors are quite small. Also, because
germ cells are a highly regulated cell type during evolution, their
mouse variance of progenitor numbers is not likely. We cannot yet
determine when these 4 cells become determined to the germ-line
lineage, nor whether they retained any somatic lineage potential,
which could be determined by other fate-mapping studies.
The Significance of Secondary Oligoclonal Development. The impor-
tant question is why a small fraction of PGCs are selected as the
lost, most likely by apoptosis, before the reproductive period. It has
been shown that inhibiting the apoptosis of germ cells during this
stage causes male infertility (18, 19, 33). Although the role of
apoptosis has been considered to adjust the ratio of germ cells and
of different colors. (A) Migrating PGCs observed in the
hindgut region of an E8.5 tetrachimeric embryo. Red
indicate PGCs. (B) PGCs in the genital ridge of an E12.5
tetrachimeric embryo. (Lower) A magnified image of a
white box shown in Upper Left. Arrows indicate fluores-
cent PGCs. Dashed lines indicate locations of TNAP-
positive cells. (A Right and B Right) TNAP staining of the
of an E13.5 tetrachimeric embryo. Arrows indicate fluo-
www.pnas.org?cgi?doi?10.1073?pnas.0810325105Ueno et al.
the niche in testes, we have further shown that selected cells and
removed cells are determined before quite an early stage of
formation of genital ridges. Interestingly, it has been reported that
the peak of apoptosis of germ cells is observed not only at 10–13
days postnatally but also at ?E13 (34), when the secondary
selection of germ cells initiates in our model.
There are several potential merits of the second selection of
the germ line. First, abnormal germ cells could be removed
during the selection. Second, in our model, the number of cell
divisions required for the ‘‘actual’’ germ cells could be reduced,
compared with those for the rest of the germ cells, which
potentially decreases the risk of mutations. Although the signif-
icance of the ‘‘deciduous’’ germ cells that cannot contribute to
reproduction is unclear, evidence has accumulated that germ
cell–Sertoli cell interaction is required for Sertoli cell maturation
(35), which is needed in preparation for spermatogenesis in the
reproductive period. Our findings could imply that not all
spermatogonial progenitors are equal: epigenetic and genetic
sources of variation could contribute to future ‘‘winners’’ in the
competition. Epigenetic expression of nontissue genes in the
GFP RFP TRA98 GFP RFP TRA98
GFP CFP RFP
CFP RFP TRA98
example of a chimeric testis at P7. TRA98-positive (green) germ cells show a mixture of blue and red cells (colored arrowheads). (C–E) An example of a green and red
chimeric testis at P14. In this example, single-color patches of different colors have started forming in distinct areas of a testis. (D and E) Magnified pictures of areas
in white rectangles shown in C, in which germ cells are stained with TRA98 (white). (D) In the region, red germ cell patches are forming, although a small number of
are observed. (F) NGN3-positive cells form a green patch in the region where green germ cells are selected. (G) TUNEL staining of the region where green germ cells
are selected. TUNEL-positive apoptotic cells (purple) are exclusively observed in red germ cells. (H) NGN3-positive germ cells form single-color patches (colored
arrowheads) in green and blue chimeric testis of a P3 tetrachimeric mouse. (Right) Magnified picture of the white rectangle shown in Left. (Scale bars, 250 ?m in C;
100 ?m in A and B; 50 ?m in D–H.)
‘Deciduous’ germ cells
removed by apoptosis
‘Actual’ long-term germ stem cells
testis germ cells. Male germ line starts from 4 cells. They
proliferate, migrate, and split to bilateral genital ridges.
small number of a second founder population that ini-
tially seed onto genital ridges. The rest of germ cells
(deciduous germ cells) are removed, most likely by apo-
ptosis, before the reproductive period.
Ueno et al.
January 6, 2009 ?
vol. 106 ?
no. 1 ?
thymus is regulated by the AIRE gene, and assures the negative
selection of nascent T cells with autoimmune T cell antigen
receptors; AIRE mutant mice are male sterile and have excess
apoptosis (36). Genetic variation can occur also, with the
movement of transposable elements, regulated by microRNA
species (37); loss of the regulatory microRNA also leads to male
sterility. Further studies are needed to investigate these issues.
clones derived from tissue stem cells also occurs in at least some
somatic tissues. Elsewhere we show, by using tetrachimeric mice,
that the whole endoderm is generated from ?16 cells, and that
these clonal subsets are initially intermixed randomly as to colors.
that allows the outgrowth of very large clones that can form
I.L.W., unpublished work). We speculate that clonal selection of
tissue stem cells, whether germ line or soma, might be a constant
feature in development and homeostasis. It is critical to test this
postulate in other tissues as well, and to test whether such clonal
selections result from capture of limiting numbers of stem cell
niches. Regulation of stem cells by such niches could have an
important role in preventing the emergence of neoplastic and
aberrant stem cell clones (38).
Several issues are also discussed in SI Discussion.
Materials and Methods
Cell Culture and Blastocyst Injection of Fluorescent ES Clones and Generation of
Chimeric Embryos and Adult Mice. Culture of mouse R1 ES cells and generation
of Rosa 26 knock-in clones were described (21). Mice were bred and maintained
at Stanford University Research Animal Facility in accordance with Stanford
University guidelines. The generation of tetrachimeric embryos and mice was
performed as described (21).
Histological Analysis and Immunostaining. Organsfromadultmiceandembryos
with Hoechst 33342 (H33342). Anti-Stella (Abcam) and anti-NGN3 (a gift from S.
antibody (a gift from Y. Nishimune, Osaka University, Osaka) were used for
sciences). For TNAP staining, frozen sections were permeabilized with 0.1%
9.5/100 mM NaCl/10 mM MgCl2) containing nitro-blue tetrazolium (NBT) and
5-bromo-4-chloro-3-indolyl phosphate (BCIP) (Promega). TUNEL staining was
performed with In Situ Cell Death Detection Kit (Roche).
SKY Analysis. SKY analysis was done with SKY paint kit (Applied Spectral Imag-
ing) according to the manufacturer’s protocol.
likelihood assuming a multinomial model (H.U., B.B.T., and I.L.W., unpublished
Estimation of the ratio of probabilities of RFP-expressing (R), CFP-expressing (C),
GFP-expressing (G), and nonfluorescent (NF) cells. Toestimatetheratio,weassume
equal probabilities for R, C, and G cells and maximize the likelihood simulta-
neously over the ratio and the number of progenitors. That is, we find the pair
(ratio, number of progenitors) for which the probability of the observed data is
greatest. For testis, this ratio is ?1:1:1:2.
Calculation of posterior probabilities of numbers of progenitors. We use Bayes’ rule
with a uniform prior distribution, indicating no knowledge of the number of
progenitors a priori, which is equivalent to normalizing the likelihoods of the
possibilities to sum to 1.
Relationship between number of colors lost and age. The proportional odds
colors lost and age. We reject the null hypothesis if the P value is ?0.05.
Comparison of R, C, and G frequency. A?2testisperformedtotestforadifference
in color frequencies. The null hypothesis of equal proportions is rejected if the P
value is ?0.05.
ACKNOWLEDGMENTS. We thank P. Soriano (Mount Sinai School of Medicine,
for the mRFP1 cDNA; Y. Nishimune (Osaka University, Osaka) for the TRA98
antibody; S. Kim (Stanford University) for the NGN3 antibody; T. Nakano (Osaka
T. Doyle for help with chromosomal analysis; M. Kwan for technical advice; D. B.
Escoto and A. Mosley for animal management; and L. Jerabek for laboratory
management. H.U. was supported by a Floren Family Fund gift. B.B.T. was
supported by a Stanford Cancer Center fellowship. This work was supported by
grants from the U.S. Public Health Service National Institutes of Health (I.L.W.)
and the Smith Family Fund.
1. Weismann A (1892) The Germ-Plasm. A Theory of Heredity, Translated from German
(Gustav Fischer, Jena, Germany).
182:92–110; discussion 110–120.
3. Wylie C (1999) Germ cells. Cell 96:165–174.
4. Soriano P, Jaenisch R (1986) Retroviruses as probes for mammalian development: Alloca-
tion of cells to the somatic and germ cell lineages. Cell 46:19–29.
5. Hogan BL (2001) Primordial germ cells as stem cells. Stem Cell Biology, 189–204.
cell mass cells by blastocyst injection. J Embryol Exp Morphol 52:141–152.
7. Gardner RL, Lyon MF, Evans EP, Burtenshaw MD (1985) Clonal analysis of X-chromosome
8. Lawson KA, Meneses JJ, Pedersen RA (1991) Clonal analysis of epiblast fate during germ
layer formation in the mouse embryo. Development 113:891–911.
in the mouse embryo. Anat Rec 118:135–146.
10. Ginsburg M, Snow MH, McLaren A (1990) Primordial germ cells in the mouse embryo
during gastrulation. Development 110:521–528.
Neerl Scand 24:103–110.
12. Rosner MH, et al. (1990). A POU-domain transcription factor in early stem cells and germ
cells of the mammalian embryo. Nature 345:686–692.
13. Sato M, et al. (2002). Identification of PGC7, a new gene expressed specifically in preim-
plantation embryos and germ cells. Mech Dev 113:91–94.
14. Saitou M, Barton SC, Surani MA (2002) A molecular programme for the specification of
germ cell fate in mice. Nature 418:293–300.
15. Ohinata Y, et al. (2005). Blimp1 is a critical determinant of the germ cell lineage in mice.
16. Heath JK (1978). Mammalian primordial germ cells, in Development in Mammals, ed
Johnson MH (Elsevier, Amsterdam), Vol 3, pp 267–298.
17. McLaren A (1998) Gonad development: Assembling the mammalian testis. Curr Biol
18. Rodriguez I, Ody C, Araki K, Garcia I, Vassalli P (1997) An early and massive wave of
germinal cell apoptosis is required for the development of functional spermatogenesis.
EMBO J 16:2262–2270.
19. Furuchi T, Masuko K, Nishimune Y, Obinata M, Matsui Y (1996) Inhibition of testicular
germ cell apoptosis and differentiation in mice misexpressing Bcl-2 in spermatogonia.
20. Knudson CM, Tung KS, Tourtellotte WG, Brown GA, Korsmeyer SJ (1995) Bax-deficient
mice with lymphoid hyperplasia and male germ cell death. Science 270:96–99.
21. Ueno H, Weissman IL (2006) Clonal analysis of mouse development reveals a polyclonal
origin for yolk sac blood islands. Dev Cell 11:519–533.
23. Tilmann C, Capel B (2002) Cellular and molecular pathways regulating mammalian sex
determination. Recent Prog Horm Res 57:1–18.
24. McLaren A (1995) Germ cells and germ cell sex. Philos Trans R Soc London Ser B 350:229–
25. Adams IR, McLaren A (2002) Sexually dimorphic development of mouse primordial germ
cells: Switching from oogenesis to spermatogenesis. Development 129:1155–1164.
26. Isotani A, et al. (2005). Genomic imprinting of XX spermatogonia and XX oocytes recov-
ered from XX7XY chimeric testes. Proc Natl Acad Sci USA 102:4039–4044.
27. McLaren A (1980) Oocytes in the testis. Nature 283:688–689.
28. Yoshida S, et al. (2004) Neurogenin3 delineates the earliest stages of spermatogenesis in
the mouse testis. Dev Biol 269:447–458.
29. Nakagawa T, Nabeshima Y, Yoshida S (2007) Functional identification of the actual and
potential stem cell compartments in mouse spermatogenesis. Dev Cell 12:195–206.
30. de Rooij DG (1998) Stem cells in the testis. Int J Exp Pathol 79:67–80.
mouse spermatogenesis following transplantation. Biol Reprod 69:1872–1878.
that lacks the self-renewing spermatogonia stage. Development 133:1495–1505.
33. Coultas L, et al. (2005) Concomitant loss of proapoptotic BH3-only Bcl-2 antagonists Bik
and Bim arrests spermatogenesis. EMBO J 24:3963–3973.
34. Wang RA, Nakane PK, Koji T (1998) Autonomous cell death of mouse male germ cells
during fetal and postnatal period. Biol Reprod 58:1250–1256.
35. Weinbauer GF, Wessels J (1999) ‘Paracrine’ control of spermatogenesis. Andrologia
36. Schaller CE, et al. (2008). Expression of Aire and the early wave of apoptosis in spermat-
ogenesis. J Immunol 180:1338–1343.
37. Carmell MA, et al. (2007). MIWI2 is essential for spermatogenesis and repression of
transposons in the mouse male germline. Dev Cell 12:503–514.
J Am Med Assoc 294:1359–1366.
www.pnas.org?cgi?doi?10.1073?pnas.0810325105Ueno et al.