Corrected 19 October 2021. See full text.

RES EARCH

◥ spermatogenesis upon transplantation into

REVIEW SUMMARY testes. Furthermore, the oogenic fate determi-
nation pathway has been reconstituted with-
GAMETOGENESIS out the use of ovarian somatic cells, revealing

Mammalian in vitro gametogenesis

the mechanisms for epigenetic reprogramming,
oogenic fate determination, and meiotic entry.
Human PSCs (hPSCs) have been induced
Mitinori Saitou* and Katsuhiko Hayashi* into human PGCLCs (hPGCLCs), and the mech-
anism for hPGC(LC) specification has been
clarified and shown to involve key TFs, their
BACKGROUND: Mammalian germ cells differ- (iPSCs) generated from somatic cells through hierarchy of actions, and their regulatory
entiate into oocytes in females and spermato- transcription factor (TF)–induced reprogram- wiring, all of which are divergent from those
zoa in males. An oocyte and a spermatozoon ming. This process, referred to as in vitro game- for mPGC(LC) specification. hPGCLCs have
fuse to form a zygote and then develop to togenesis (IVG), is not only providing a basis for been shown to undergo epigenetic reprogram-
create a new individual, thereby transmitting further exploration of the mechanisms of germ ming and differentiate into early oocytes or
their genetic and epigenetic information to cell development but also creating prospects for prospermatogonia by culturing with mouse
the next generation. During development, innovative medical applications. embryonic ovarian or testicular somatic cells
germ cells undergo epigenetic reprogram- (xenogeneic rOvaries or rTestes), respectively,
ming and programming to acquire totipo- ADVANCES: Mouse PSCs (mPSCs) have been establishing a foundation for human IVG.
tency upon fertilization, maintain genome induced into mouse primordial germ cell– More recently, mPSCs have been induced into
information with high fidelity, and, paradoxi- like cells (mPGCLCs), the founding germ cell mouse embryonic ovarian somatic cell–like
cally, create genome diversity through meiotic population. Oogenesis has been reconstituted cells fully capable of supporting mPGCLC-

Downloaded from https://www.science.org on December 05, 2023

recombination. by culturing mPGCLCs with mouse embry- based oogenesis, paving a way for generating
Intensive efforts have been made to explore onic ovarian somatic cells [reconstituted ovaries corresponding cells in humans and other spe-
the molecular and systems-level mechanisms (rOvaries)], and the resultant oocytes have cies, including endangered species, and hence
underlying these distinctive functions that germ produced offspring. This system has revealed for robustly promoting their IVG.
cells acquire during their development. On the key mechanisms for oocyte development, in-
basis of the knowledge obtained through such cluding the TFs engendering oocyte-like growth OUTLOOK: The proof-of-concept mouse IVG
endeavors, notable progress has been made in mouse ESCs. The mPGC-to-spermatogonia and the basic framework of human IVG
over the past decade on research aiming at development has been reconstituted by cul- have been established, creating opportuni-
reconstituting germ cell development in vitro turing mPGCLCs with mouse embryonic tes- ties for exploring the mechanisms of mouse
using pluripotent stem cells (PSCs), including ticular somatic cells [reconstituted testes and human germ cell development, includ-
embryonic stem cells (ESCs) derived from (rTestes)], and the resultant spermatogonia ing those that have been both technically
preimplantation embryos and induced PSCs have propagated in vitro and contributed to and ethically difficult to address. IVG tech-
nology can be applied to other animals, in-
cluding endangered species, enabling species
preservation. Realizing human IVG will re-
quire further advances but will create possi-
bilities for diagnosing and modeling infertility,
exploring its remedies, and improving ar-
tificial reproductive technologies, thereby
advancing reproductive medicine. Human
PGCLCs IVG could also be extended for reproduc-
tive application, but before that would be
deemed permissible, analyses must include
appropriate “normality” assessments of the
iPSCs
IVG-derived animal models—ideally primate
models—and genetic and epigenetic assess-
In vitro Spermato- Immature
ments of hPSCs and the resultant gametes.
gonia oocytes
gametogenesis Once all technological concerns have been
resolved, it will be crucial to hold society-wide
Germline
discussions about whether to use IVG-derived
Soma Oocytes gametes for human reproduction, because
Die away Sperm
in one such an application would change our un-
generation derstanding of human origins and the conti-

Genetic/epigenetic integrity?
nuity of life.
▪
Ethical issues? Basic mechanisms The list of author affiliations is available in the full article online.
Modeling infertility *Corresponding author. Email: saitou@anat2.med.kyoto-u.ac.jp
Improving ART (M.S.); hayashik@hgs.med.kyushu-u.ac.jp (K.H.)
Cite this article as M. Saitou, K. Hayashi, Science 374,
eaaz6830 (2021). DOI: 10.1126/science.aaz6830
In vitro gametogenesis. IVG aims to recreate germ cell development in vitro using PSCs, including iPSCs
from somatic cells. Human IVG realization requires further efforts and will create new opportunities in READ THE FULL ARTICLE AT
reproductive biology research and medicine. ART, assisted reproductive technologies. https://doi.org/10.1126/science.aaz6830

Saitou et al., Science 374, 47 (2021) 1 October 2021 1 of 1

 Corrected 19 October 2021. See full text.
RES EARCH

◥ spermiogenesis to form haploid spermatozoa

REVIEW (10). Notably, in mice, but not in humans, a
primary SSC line, referred to as germline stem
GAMETOGENESIS cells (GSCs), with a robust capacity for self-

Mammalian in vitro gametogenesis

renewal in vitro and for spermatogenesis upon
transplantation into testes, has been derived,
prompting investigations into the properties
Mitinori Saitou1,2,3* and Katsuhiko Hayashi4,5* of SSCs and GSCs (11). Testicular somatic
cells, particularly Sertoli cells, play pivotal
Germ cells differentiate into sexually dimorphic gametes, oocytes, and spermatozoa, which unite to form roles in supporting male germ cell develop-
new individuals. Accordingly, germ cell development entails intricate regulations of genome functions ment (4, 8, 10).
for genetic and epigenetic inheritance. The past decade has seen considerable advances in in vitro The sexually dimorphic composition of the
gametogenesis (IVG), which aims to recreate germ cell development from pluripotent stem cells (PSCs) sex chromosomes has key effects on germ cells
in culture. Mouse PSCs can be induced into functional oocytes and spermatozoa, whereas human PSCs in a cell-intrinsic manner, although it primar-
can be induced into early oocytes and prospermatogonia, promoting mechanistic understanding of ily affects the sex of the somatic cells. The XX
mammalian germ cell development. The prospect for inducing human gametes with appropriate composition is beneficial for the growth and
functions has been heightened, and such advances will create possibilities in reproductive medicine, development of oocytes (12), whereas XY is
including modeling infertility to explore remedies. The use of IVG-derived gametes for human critical for the development of spermatogonia
reproduction will require careful legal and ethical discussions. as well as for the spermatogenesis (13). The
general principles of germ cell development

M
are well conserved among mammals. How-
ammalian germ cells develop via intri- ation to demethylate DNA genome-wide and ever, there are critical species differences in

Downloaded from https://www.science.org on December 05, 2023

cate and precise processes to generate alter histone modification, leading to the era- the time scale required for key developmen-
sexually dimorphic gametes, oocytes sure of parental imprints and X chromosome tal transitions, developmental synchronicity,
and spermatozoa. Gametes fuse and reactivation in females (2, 3). and the signaling, transcriptional, and meta-
develop to form new individuals that In the ovaries of XX embryos, oogonia re- bolic properties of developmentally homolo-
inherit genetic and epigenetic information ceive signals from nascent granulosa cells (so- gous cells (1, 2, 5) (Fig. 1).
from the parents. Anomalies in germ line de- matic cells that surround developing oogonia)
velopment result in infertility and genetic or or the ovarian environment to differentiate IVG in mice
epigenetic disorders in the offspring. The past into primary oocytes and enter into meiotic Mouse PGC-like cells (mPGCLCs)
decade has seen great progress in research to prophase (4–7) (Fig. 1). The oocytes proceed During development, mouse PGCs (mPGCs)
generate germ cells in vitro from pluripotent into the diplotene stage of meiotic prophase originate from the most proximal epiblast at
stem cells (PSCs), including embryonic stem and form a tight complex with granulosa cells, around embryonic day 6.5 (E6.5) as Blimp1
cells (ESCs) and induced pluripotent stem cells creating primordial follicles. At this stage, most (also known as Prdm1)–positive cells in re-
(iPSCs) that are derived from somatic cells. of the primordial follicles pause their devel- sponse to bone morphogenetic protein 4
Here, we will discuss the current state of mam- opment. They resume follicle growth, that is, (BMP4) from the extraembryonic ectoderm
malian in vitro gametogenesis (IVG) research oocyte growth coupled with expansion of the (14–17) (Figs. 1 and 2). Proto-oncogene protein
(1), key challenges driving the research forward, granulosa cell layers, around puberty in re- WNT3, which is activated in the epiblast by
and critical considerations associated with the sponse to hormonal cues and develop into BMP signaling, is also essential for mPGC
applications of IVG research. fully matured antral follicles. Mature oocytes specification (17) (Fig. 2). The three transcription
possess the maternal genome and epigenome factors (TFs) BLIMP1, PRDM14, and TFAP2C
Germ cell development in mammals (6, 7). Upon ovulation and fertilization with are essential for mPGC specification, which
Germ cells arise as primordial germ cells (PGCs) spermatozoa, the antral follicles complete involves repression of a somatic program, re-
from pluripotent cells during early mamma- their first and second meiotic divisions, respec- acquisition of the pluripotency network, and
lian development (1, 2) (Fig. 1). PGCs migrate tively, to form a zygote. Ovarian somatic cells, epigenetic reprogramming (16, 18, 19) (Fig. 3).
through the developing hindgut endoderm particularly granulosa cells, exert an essential The epiblast cells from around E5.25 to E6.25
and mesentery and colonize the embryonic function in promoting female germ cell de- show competence to differentiate into mPGCs
gonads, where they proliferate as oogonia in velopment (4, 6, 7). in response to BMP4 (17).
females or gonocytes in males and initiate sex- In the testes of XY embryos, nascent Sertoli There are at least two distinct states of
ually dimorphic development. In contrast to cells (somatic cells that surround developing pluripotency in culture: naïve pluripotency
somatic cells, PGCs undergo epigenetic repro- gonocytes and form seminiferous tubules) or and primed pluripotency (20) (Fig. 2). Mouse
gramming during their migration and prolifer- the testicular environment signal gonocytes ESCs (mESCs) derived from the inner cell mass
to differentiate into prospermatogonia, which of the blastocysts bear naïve pluripotency and
1
Institute for the Advanced Study of Human Biology (WPI-ASHBi), undergo mitotic arrest and acquire the paternal contribute to chimeras and germ cells, and
Kyoto University, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, epigenome, including imprints and a defense they are typically cultured with inhibitors for
Japan. 2Department of Anatomy and Cell Biology, Graduate
School of Medicine, Kyoto University, Yoshida-Konoe-cho,
against transposon activities (3–5, 8, 9) (Fig. 1). fibroblast growth factor and mitogen-activated
Sakyo-ku, Kyoto 606-8501, Japan. 3Center for iPS Cell After birth, prospermatogonia differentiate into protein kinase (FGF/MEK) and glycogen syn-
Research and Application (CiRA), Kyoto University, 53 spermatogonia or spermatogonial stem cells thase kinase 3b (GSK3b) in the presence of
Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606-8507, Japan.
4 (SSCs). These cells essentially halt their develop- leukemia inhibitory factor (LIF) (a culture
Department of Developmental Stem Cell Biology, Faculty of
Medical Sciences, Kyushu University, Maidashi 3-1-1, Higashi-ku, ment until puberty, at which time they initiate condition known as 2i+LIF) (21–23). mESCs
Fukuoka 812-8582, Japan. 5Department of Germline Genetics, robust waves of spermatogenesis, the process show properties similar to those of the peri-
Graduate School of Medicine, Osaka University, 2-2 Yamada- by which a mitotic stem cell gives rise to hap- implantation epiblast (~E4.5) (21, 24) (Fig. 2).
oka, Suita, Osaka 565-0871, Japan.
*Corresponding author. Email: saitou@anat2.med.kyoto-u.ac.jp loid gametes, consisting of mitotic amplifi- In contrast, mouse epiblast stem cells (mEpiSCs)
(M.S.); hayashik@hgs.med.kyushu-u.ac.jp (K.H.) cations followed by meiotic divisions and derived from the postimplantation epiblast

Saitou et al., Science 374, eaaz6830 (2021) 1 October 2021 1 of 9

 Corrected 19 October 2021. See full text.
RES EARCH | R E V I E W

Primary oocytes Antral Oocytes (MII)

(Prophase I)
Pm Pr
Sc

ExE Al
ICM Mesoderm
Ovary Granulosa

SG SC
R-ST El-ST
PGCs SG (TypeB)
(TypeA)
TE Pro-SG Sertoli
Epiblast VE Migrating PGCs PGCs in
genital ridges PGCs in gonads Testis Spermatozoa
Mouse
Pro-SG

Fertilization 2cell Blastocyst E5.0 E6.5 E8.5 E10.5 E12.5 E13.5 Birth Puberty
1week
Fertilization 2cell Blastocyst 2wks 3wks 5wks 8wks 10wks 20wks Birth Puberty
Oogonia/Primary oocytes
(Prophase I) Antral Oocytes (MII)
Amnion
Human PGCs in Pm Pr Sc
PGCs
york sac
ICM Epiblast Al

Ovary Granulosa
SG SC R-ST El-ST
York sac
Pro-SG SG (TypeB)
Hypoblast (TypeA)
TE ExMC Migrating PGCs Sertoli
PGCs in gonads Testis Spermatozoa
Gonocytes/Pro-SG
(Mitotic arrest)

Downloaded from https://www.science.org on December 05, 2023

Oogonia/Gonocytes/Pro-SG Sertoli cells Interstitial cells (Male) Myoid cells
Primary oocytes Granulosa cells Interstitial cells (Female) Epithelial cells

Fig. 1. Scheme for germ cell development in mice and humans. Schematic granulosa cell layers), and antral follicles (multiple granulosa cell layers with antral
representations of germ cell development in mice (top) and humans (bottom). cavity). In testes, Sertoli cells surround gonocytes and prospermatogonia and
Embryonic days (E) in mice and developmental weeks (wks) in humans are form seminiferous tubules, in which male germ cell development progresses both
shown, with 1-week periods marked with symbols, as indicated. In ovaries, during development and in adults. ICM, inner cell mass; TE, trophectoderm; ExE,
granulosa cells surround oogonia and primary oocytes, creating primordial extraembryonic ectoderm; Al, allantois; PGCs, primordial germ cells; Pm, primordial
follicles (a single squamous granulosa cell layer). They undergo follicle growth, follicles; Pr, primary follicles; Sc, secondary follicles; MII, meiosis II; Pro-SG,
that is, oocyte growth coupled with expansion of the granulosa cell layers, prospermatogonia; SG, spermatogonia; SC, spermatocyte; R-ST, round spermatids;
and develop into primary (a single cuboidal granulosa cell layer), secondary (several El-ST, elongating spermatids.

at ~E5.5 to E6.5 show primed pluripotency iPSCs (miPSCs) (28, 29) (Fig. 4), establishing chromosome trisomy, such as XXY (Klinefelter
with a biased differentiation potential and con- a foundation for IVG. syndrome), preferentially lose the extra sex
tribute little, if any, to chimeras under conven- The state exhibited by mEpiLCs or pregas- chromosome, and the resultant euploid XY
tional conditions (25, 26). mEpiSCs self-renew trulating mouse epiblast has been proposed miPSCs differentiate into mPGCLCs and
in the presence of activin and FGF and exhibit to represent a formative state, an immediately contribute to spermatogenesis and offspring,
properties similar to the epiblast after E6.5 precursory state for multilineage differenti- providing a strategy for overcoming infer-
(27) (Fig. 2). ation, including germ cell differentiation, in tility with abnormal sex chromosome con-
mEpiSCs are refractory to differentiation response to external cues (30) (Fig. 3). Con- stitutions (40).
into germ cells (28). A strategy to induce mESCs sistently, mEpiLCs have a distinctive epigenome
into competent epiblast-like cells (mEpiLCs) with abundant bivalent histone modifications In vitro oogenesis
was explored, leading to the finding that an in key developmental genes, which are poised On the basis of organ and reaggregation cul-
~2-day stimulation by activin and FGF induces for timely expression (31, 32). Recently, by ture of embryonic ovaries (41–43), a method-
mESCs into mEpiLCs bearing properties sim- mainly modulating activin and WNT signal- ology for rOvary culture has been established
ilar to the epiblast at E5.5 to E6.0 (28) (Figs. 2 ing, a formative state has been stably captured (44) (Fig. 4). rOvaries are cultured under an
and 4). In response to BMP4, mEpiLCs robustly in vitro (33, 34). air–liquid interface condition for in vitro de-
differentiated into mPGC-like cells (mPGCLCs) The mPGCLC induction system has been velopment (IVD; ~3 weeks), followed by in vitro
with properties similar to mPGCs at the migrat- useful in analyzing the mechanisms of mPGC growth (~2 weeks). During IVD, mPGCLCs dif-
ing stage (~E9.5). mPGCLC induction faithfully specification (35–39) (Fig. 5). For example, the ferentiate into oogonia and then into oocytes
recapitulated the mPGC specification process finding that T (also known as TBXT), a down- with an entry into the prophase of meiosis I
in vivo. Male mPGCLCs contributed to sper- stream effector of WNT signaling, activates (MI) and form oocytes at the secondary follicle
matogenesis when transplanted into testes of Blimp1 and Prdm14 illuminates a mechanism stage. Under the in vitro growth conditions,
neonatal mice lacking endogenous spermato- that bridges the external signaling and in- they differentiate into fully grown oocytes at
genesis (28); female mPGCLCs contributed to ternal transcription network for mPGC(LC) the antral follicle stage, which can be induced
oogenesis when aggregated with mouse em- specification (35) (Fig. 3). Homeobox protein into oocytes at the meiosis II (MII) stage through
bryonic ovarian somatic cells (reconstituted OTX2, a TF essential for anterior neuroecto- in vitro maturation (Fig. 4). The MII oocytes
ovaries, or rOvaries) and transplanted under derm specification, has been shown to act as from both mESCs and miPSCs can be fertilized
the ovarian bursa of immune-deficient mice a repressor of mPGC(LC) specification, whose to form offspring (44), establishing a proof-of-
(29); the resultant spermatozoa and fully absence leads to an extended competence of concept for IVG.
grown oocytes produced fertile offspring the epiblast and mEpiLCs to adopt the germ Compared with in vivo oogenesis, however,
(28, 29) (Fig. 4). These processes were rep- cell fate (38). Moreover, it has been shown that the fidelity of the in vitro oogenesis procedure
licated using both male and female mouse miPSCs generated from infertile mice with sex is compromised. The developmental rate of

Saitou et al., Science 374, eaaz6830 (2021) 1 October 2021 2 of 9

 Corrected 19 October 2021. See full text.
RES EARCH | R E V I E W

Naive Formative PGC specification Primed

pressure yielded dormant oocytes, demon-
strating a critical involvement of biophysical
Rodent Implantation BMP4 mechanisms for primordial follicle recruit-
ment (51, 53).
Mouse A P
Embryo The in vitro oogenesis system also revealed
development
WNT3 that XO and XY oocytes in rOvaries with XX
Rat gonadal somatic cells exhibit impaired devel-
(mESC/iPSC) bFGF+activin A bFGF opment owing to frequent chromosomal mis-
2i+LIF/3i
(rESC/iPSC) (mEpiLC (transient state)) (mEpiSC)
Ying et al.
Buehr et al.
Hayashi et al. Tesar et al. pairing during the prophase of MI, followed
Brons et al.
PSC state Li et al. AloXR (activin A+XAV939+BMS493) by oocyte apoptosis (54). The impaired de-
(mFSCs)
Kinoshita et al. velopment was more profound in XY oocytes,
Conv. (mESC/iPSC) FTW (bFGF+activin A+CHIR) partially owing to an inhibitory effect on
(mXPSCs)
Yu et al. oogenesis of a Y-linked translational initiator,
Eif2s3y, that promotes spermatogenesis (54, 55).
Primate This system offers a strategy for circumvent-
Implantation WNT3A
BMP4 ing the infertility caused by sex chromosome
Embryo
development A P disorders, for example, screening of chemicals
that nullify a meiotic checkpoint by sex chro-
Human 5i/L/A (hESC/iPSC) 4i (hESC/iPSC) mosome heteroploidy in oocytes.
Theunissen et al. Gafni et al.
Cynomolgus AloXR (activin A+XAV939+BMS493)
Moreover, a set of TFs sufficient for oocyte
t2i/L/Gö (hESC/iPSC)
monkey Takashima et al. (hFSCs) growth has been identified: simultaneous ex-
Kinoshita et al.
pression of NOBOX, FIGLA, LHX8, TBPL2,

Downloaded from https://www.science.org on December 05, 2023

PSC state FTW (bFGF+activin A+CHIR)
(hXPSCs) SOHLH1, STAT3, SUB1, and DYNLL1, each of
Yu et al.
which is essential for oocyte growth, induced
Conv. (hESC/iPSC) female mESCs into oocyte-like cells when ag-
AITS+IF20 (cyESC) gregated with ovarian somatic cells, and no-
Sakai et al.
tably, such cells were capable of fertilization,
Others demonstrating that oocyte growth per se can
(Rauber’s layer species) Implantation
BMP2/4 be independent from preceding germ cell de-
Rabbit Embryo velopment (56) (Fig. 5), which extends a pre-
development A P
WNT3A vious finding that oocyte growth is dissociable
from meiosis (57). This work paves the way for
Pig Epiblast Hypoblast/Primitive endoderm Amnion Nascent Mesoderm
Trophectoderm Anterior Visceral Endoderm Primordial Germ Cells Extra-embryonic Mesenchyme the TF-based generation of developmentally
competent ooplasm.
Fig. 2. Divergence of early development and spectrum of pluripotency in vivo and in vitro in various
mammals. Schematic illustrations of the development of a portion of preimplantation and early In vitro spermatogenesis
postimplantation embryos in rodents (mice and rats, top), primates (humans and monkeys, middle), and A methodology for the reaggregation culture
other representative mammals (rabbits and pigs, bottom). PSCs cultured under various conditions in vitro are of male mPGCLCs with embryonic testicular
aligned with embryonic cells bearing the closest similarity in their transcriptomes. See text and specified somatic cells [reconstituted testes (rTestes)]
references for details. Conv., conventional ESC culture media with serum and LIF; bFGF, basic fibroblast has been explored (58) (Fig. 4). The rTestes
growth factor; AloXR, low activin A, XAV939, and RAR inhibitor; mFSC, mouse formative stem cells; FTW, FGF, cultured under an air–liquid interface condi-
TGFb, and WNT; mXPSC, mouse chimera and PGC dual-competent PSCs; AITS+IF20, AlbuMax, insulin, tion self-organized seminiferous tubule-like
transferrin, selenoprotein, IWR1, bFGF 20 ng/ml; A, anterior; P, posterior. structures, in which mPGCLCs differentiated
into gonocytes, prospermatogonia, and then
in vitro–derived two-cell embryos to term (~3%; and in vitro oogenesis will be an important spermatogonia (~3 weeks). Compared with
mESC-derived: 1.2 to 3.9%; miPSC-derived: challenge. in vivo development, in vitro development was
0.3 to 0.9%) is ~1/20 to 1/100 that of in vivo Notwithstanding these limitations, the protracted and less efficient, and mPGCLC-
embryos (44). In vitro oogenesis is charac- in vitro oogenesis system has been instrumen- derived spermatogonia failed to undergo mei-
terized by a higher incidence of homologous tal in clarifying key mechanisms for oocyte osis, although some progressed to the prophase
chromosome mispairing, lower mitochondrial development. In the ovary, most oocytes re- of MI after a prolonged culture (~8 weeks).
DNA amounts, and precocious resumption main dormant as primordial follicles, ensuring The potential of mPGCLC-derived spermato-
of the first meiotic division. These defects oocyte storage and reproductive longevity (6, 7). gonia was demonstrated by the derivation of
could stem from both an inadequate quality In rOvaries, however, few dormant oocytes GSC-like cells (GSCLCs) in culture (58) (Fig. 4).
of mPSCs and suboptimal in vitro culture (primordial follicles) are formed, and almost GSCLCs were similar to GSCs in morphology,
conditions. Regarding the former, mPSCs— all oocytes develop alongside somatic cells to expansion capacity, and transcriptome and
particularly female mPSCs—cultured under form secondary follicles (29, 44). Transcriptome contributed to spermatogenesis in adult testes,
the 2i+LIF condition show epigenetic abnor- analyses revealed that oocytes in primordial with the resultant spermatozoa giving rise to
malities, including genome-wide DNA demeth- follicles in vivo are enriched with genes for fertile offspring. The spermatogenesis efficiency
ylation (45, 46), and a prolonged culture leads response to hypoxia and extracellular matrix, of GSCLCs, however, was substantially lower
to their compromised developmental poten- which is consistent with the fact that they than that of GSCs (3 out of 15 GSCLC lines
tial (47, 48). iPSCs bear additional epigenetic are located around the ovarian cortex bearing contributed to spermatogenesis, and the three
abnormalities derived from somatic cells and an avascular and collagen-rich histology contributing lines showed ≦~20% spermato-
reprogramming errors (49, 50). Thus, improve- (51, 52). Accordingly, rOvary culture under a genesis colonies, compared with ~100% sper-
ments of culture conditions for female mPSCs hypoxic condition or with extra-hydrostatic matogenesis by GSCs) owing to an aberrant

Saitou et al., Science 374, eaaz6830 (2021) 1 October 2021 3 of 9

 Corrected 19 October 2021. See full text.
RES EARCH | R E V I E W

tiation (65). A replication-coupled genome-wide

Fig. 3. Signaling and
DNA demethylation, which erases methyla-
transcriptional network
tions on key genes for oogenesis and spermato-
for PGC specification
Mouse genesis, and BMP signaling that induces the
Human in mice and humans.
BMP oogenic programs fulfill the mechanisms for
Schematic representations
the oogenic fate licensing. The TF-induced
GATA3 of the signaling and
GATA2 oocyte-like growth in female mESCs (56), which
transcriptional network
bear a globally demethylated epigenome simi-
for PGC specification
WNT lar to that of E13.5 germ cells (45, 46), is consist-
in mice (left) and humans
ent with this idea. Uncovering the mechanism
T EOMES (right). Arrows indicate
for implementing the spermatogenic fate re-
regulatory relationships
mains a key challenge.
among signaling and
transcription factors. IVG in humans
Modulate
Pluripotency Origin of the human germline: Lessons from
Blimp1 SOX17
cynomolgus monkeys
A translation of the strategies used in mouse
Germ cell IVG into humans could be one of the reason-
development able means for realizing human IVG. For this
Prdm14 BLIMP1
purpose, two key issues require careful consid-
eration. hPSCs are more similar to mEpiSCs
Epigenetic than to mESCs in morphology, signaling, and

Downloaded from https://www.science.org on December 05, 2023

reprogramming
transcriptional and epigenetic profiles (20, 66).
Tfap2c TFAP2C Also the origin of human germ cells has been
intractable, as it occurs early in postimplanta-
Somatic
differentiation tion, a stage ethically and technologically diffi-
cult to analyze (Figs. 1 and 2).
Studies using cynomolgus monkeys, a pri-
mate closely related to humans, have provided
epigenome programming. Nonetheless, GSCLC- taining their transcriptome, mPGCLCs exhibit insights into these two issues (67, 68). Single-
derived offspring appear grossly normal, sug- genome-wide DNA demethylation in a manner cell transcriptome analyses revealed that epi-
gesting that appropriate epigenotypes are consistent with a replication-coupled, passive blast cells in cynomolgus monkeys change their
selected during spermatogenesis or early de- demethylation, reaching a level equivalent to transcriptome considerably upon implantation
velopment or that animals indeed bear epi- germ cells at E13.5 (~5%), which bear the lowest and thereafter, while generating gastrulat-
mutations whose physiological manifestations DNA methylation levels throughout the germ- ing cells (~E13), maintain a relatively stable
escape detection (58). This study demonstrates line cycle and initiate either an oogenic or a transcriptome, and cynomolgus monkey ESCs
an in vitro induction of SSC activity from mPSCs spermatogenic program (46, 60). This work (cyESCs) bear a transcriptome highly similar
and points to the importance of epigenetic re- demonstrates that genome-wide DNA demeth- to that of postimplantation late (E16 or E17)
programming and programming for generat- ylation and germ cell sex determination are and early (E12 or E13) epiblast cells (67) (Fig.
ing an appropriate paternal epigenome, whose genetically dissociable. 2). Regarding the expression of a gene set
more faithful reconstitution warrants further Accordingly, it has been demonstrated that (~500 genes) that characterizes the develop-
investigation. BMP signaling is a primary mechanism that mental transitions of epiblast cells from the
There are several reports claiming to have induces mPGC(LC)s into the oogenic pathway, naïve state to the primed state, hPSCs are high-
achieved the generation of haploid spermatid- including meiotic entry (61, 62). In mPGC(LC)s ly similar to cyESCs and hence to the post-
like cells from mPSCs in vitro [see (59)], but that have undergone epigenetic reprogram- implantation late or early epiblast (67) (Fig. 2).
these studies did not recapitulate the in vivo ming, BMP (BMP2 and BMP5 are expressed in A recent report on single-cell transcriptomes of
processes and did not perform an appropri- granulosa cells) and its downstream effector, a human gastrulating embryo (E16 to E19) also
ate evaluation of the key intermediate events, ZGLP1, a conserved TF with GATA-like zinc showed that hPSCs represent a postimplanta-
including epigenetic reprogramming and fingers, activate key oogenic programs, such tion epiblast state (69). At the same time, ef-
programming. The in vitro reconstitution of as those for chromatin modification, meiotic forts have been made to realize the derivation of
the meiotic divisions for spermatogenesis re- cell cycle, and folliculogenesis. On the other naïve hPSCs. Two culture systems called 5i/L/A
mains a critical challenge. hand, retinoic acid (RA), which has been re- (chemical inhibitors for MEK, GSK3b, RAF,
garded as a key for female germ cell specifi- SRC/LCK and ROCK, plus LIF and activin) and
mPGCLC differentiation under cation [see (4) for review], assists in the overall t2i/L/Gö (chemical inhibitors for MEK, GSK3b
defined conditions oogenic program maturation as well as the and PKC, plus LIF) convert primed hPSCs
Attempts have been made to promote mPGCLC PGC program repression. mPGCLCs induced into those bearing transcriptional and epi-
differentiation under defined conditions free by BMP or ZGLP1 progress at least up to the genetic features of preimplantation epiblast
from gonadal somatic cells. It has been shown pachytene stage of meiotic prophase (61, 62). (70, 71) (Fig. 2). These cultures, however, are
that mPGCLCs are propagated to ~50-fold Consistently, recent reports have shown that not optimal, because they show relatively
in 1 week in the presence of forskolin and in mice lacking all RA receptor isoforms or all weak transcriptional concordance with the
rolipram, which stimulate cyclic adenosine RA synthesizing genes, female germ cells still preimplantation epiblast (67, 69) and exhibit
monophosphate signaling, on m220 feeders enter into meiosis (63, 64). epigenetic and karyotypic instability (70, 71).
that provide a membrane-bound form of stem A concept has been introduced that germ Nonetheless, these cells, but not primed hPSCs,
cell factor (60). During expansion, while main- cells require “licensing” for their sex differen- could be induced into the trophectoderm

Saitou et al., Science 374, eaaz6830 (2021) 1 October 2021 4 of 9

 Corrected 19 October 2021. See full text.
RES EARCH | R E V I E W

Derivation of mouse Derivation of

mouse oocytess Derivation of
PGCLCs (female) mouse spermatozoa
Hikabe et al.
Hayashi et al. Derivation of ESCs/iPSCs Derivation of Ishikura et al.
ESCs/iPSCs mouse spermatid
-like cells mouse germline ESCs
stem cell-like cells
Derivation of mouse Zhou et al.
PGCLCs (male) Ishikura et al.
Hayashi et al. ESCs EpiLCs ESCs
EpiLCs
EpiLCs
ESCs/iPSCs

GSCLCs
EpiLCs PGCLCs EpiLCs
PGCLCs
EpiLCs PGCLCs

Reaggregation with
Reaggregation with E12.5 ovarian soma
E12.5 ovarian soma PGCLCs PGCLCs Expansion in vitro

PGCLCs Injection

Co-culture with newborn Reaggregation with

testicular soma E12.5 testicular soma
Transplantation Transplantation
into testes into ovaries Reaggregation with
E12.5 testicular soma
Spermatozoa

Spermatozoa Round spermatid Oocytes GSCLCs

Oocytes

Mouse

2011 2012 2015 2016 2018 2020 2021

human Rhesus monkey Cynomolgus human iPSCs
ESCs/iPSCs human iPSCs human iPSCs ESCs monkey ESCs

Downloaded from https://www.science.org on December 05, 2023

iMeLCs
iMeLCs iMeLCs PGCLCs
Pre-induction iMeLCs PGCLCs
Human /4i culture PGCLCs PGCLCs
Sakai et al.
PGCLCs PGCLCs Transplantation Derivation of Reaggregation with
Cynomolgus/ Reaggregation into testes cynomolgus E12.5 mouse
Rhesus monkey with E12.5 mouse
monkey PGCLCs
testicular soma
ovarian soma
Irie et al. Sasaki et al.

Derivation of human PGCLCs

Early oocytes
Pro-spermatogonia
-like cells
Immature spermatogonia
Hwang et al.
Sosa et al.
Derivation of rhesus Derivation of human
Yamashiro et al. monkey PGCLCs pro-spermatogonia
Derivation of human
oogonia/primary oocytes

Fig. 4. Progress of IVG research in mice and humans. A research timeline of the progress of key IVG research in mice (top) and humans and monkeys (bottom).
Top right inset image shows elongated spermatids expressing the Ddx4-RFP transgene (red) and peanut agglutinin (green) counterstained with DAPI (white). Bottom
inset image shows early oocytes expressing DDX4 (magenta) and SCP3 (cyan).

EpiLCs PGCLCs
Blimp1, Prdm14, Tfap2c Transplantation into testes
Nakaki et al.
Spermatozoa
EpiLCs PGCLCs
Nanog
Murakami et al. Primary oocytes
PGCLCs
Zglp1 Oocyte-like cells
(w/o meiosis)
ESCs Nagaoka et al.
Nobox, Figla, Lhx8, Tbpl2
Skip Sohlh1, Stat3, Sub1, Dynll1
Hamazaki et al.
Mouse Epiblast

ICM Epiblast/Amnion
PGCs Primary oocytes Primordial
(Prophase I) follicles MII oocytes
ESCs (4i) PGCLCs
SOX17, BLIMP1
Kobayashi et al.
Human
Oogonia
iMeLCs PGCLCs
GATA3 or GATA2
SOX17, TFAP2C Reaggregation
Kojima et al. with E12.5 mouse
ovarian soma

Fig. 5. TF-based IVG research in mice and humans. Schematic illustrations of the progress of key TF-based IVG research in mice (top) and humans (bottom). The
cell types in which indicated TFs are expressed and the resultant cell types are shown, with an alignment with in vivo cell types bearing the closest similarity.

Saitou et al., Science 374, eaaz6830 (2021) 1 October 2021 5 of 9

 Corrected 19 October 2021. See full text.
RES EARCH | R E V I E W

lineage, indicating that they capture the po- petence for germline differentiation (28). hPSCs demethylation, which might result in trans-
tential of preimplantation embryonic cells appear to bear an extended primed pluripo- generational epigenetic inheritance.
(72, 73). Accordingly, recent studies showed tency competent to generate all the cell types Human female fetal germ cells (FGCs) can
that under appropriate three-dimensional cul- that emerge from the epiblast after implan- be classified into four cell types [mitotic (hPGCs
ture conditions, naïve hPSCs self-organize into tation, including the amnion and germ cells. and oogonia), RA-responsive, meiotic, and
human blastocyst-like structures with epiblast-, Accordingly, recent studies have reported a oogenesis], whereas male FGCs can be classified
trophectoderm-, and hypoblast-like cell differ- potential reconstitution of amnion differen- into at least three [migrating (hPGCs), mitotic
entiation (74, 75). tiation from hPSCs (83). (gonocytes), and mitotic arrest (fetal pros-
A histological study showed that cynomolgus With the hPGCLC induction system, the permatogonia)] (5, 100). RA-responsive FGCs
monkey PGCs (cyPGCs) that express key mark- mechanism of hPGC(LC) specification has been emerge starting at around week 11 and express
ers [SOX17 and TFAP2C (see below)] originate explored (79, 80, 84–88). It has been shown genes such as ZGLP1, STRA8, and ANHX at
in the amnion, which differentiates from the that SOX17, which is known as a key TF for high levels while repressing early hPGC genes,
epiblast after implantation (68) (Fig. 2). cyPGCs endoderm differentiation but is dispensable yet they show low-level expression of meiotic
are observed in the amnion from around E11 to for PGC specification in mice, acts as one of genes including SYCP1, SPO11, and PRDM9;
E17, initially dorsally and later posteriorly, and the most upstream TFs for hPGC(LC) specifi- meiotic FGCs that express such meiotic genes at
increase their numbers and migrate in the cation, and BLIMP1 functions downstream of high levels emerge starting at around week 14,
hindgut. The amnion itself expresses not only SOX17 (79, 84) (Fig. 3). SOX17 appears to be suggesting that it should take ~3 weeks for
BMP4 and its effectors, but also mesodermal critical in PGCs not only among primates RA-responsive FGCs to mature into meiotic
markers, including T, fulfilling the properties (68, 89, 90) but also in evolutionarily distant FGCs. In mice, the corresponding transition
of cells destined toward the germ cell fate (68). mammals such as pigs (91), suggesting that the occurs much more quickly (in under ~2 days),
Experiments involving an extended culture of mouse may have evolved a distinct strategy for and the RA-responsive state has not been rec-
cynomolgus monkey or human preimplanta- PGC specification. Indeed, hPGCLC specifica- ognized as a distinct entity. Consistent with

Downloaded from https://www.science.org on December 05, 2023

tion embryos have shown that cyPGCs or hu- tion involves key TFs, their hierarchy of ac- the histological observations (105, 106), even
man PGCs (hPGCs) indeed originate from an tions, and their regulatory network distinct at week 26, all four female FGC types coexist in
amnion-like structure (76, 77). Although the from mPGC(LC) specification (84, 86–88) individual female embryos, whereas two male
possibility that cyPGCs and hPGCs originate (Fig. 3), and the whole transcriptomes for FGC types coexist in single male embryos
from the epiblast has not been excluded, these hPGCLC specification diverge from those for (100), demonstrating substantial develop-
findings represent an evolutionary divergence mPGC(LC) specification (80, 90, 92). mental asynchrony of human FGCs.
of developmental pathways among mammals. Moreover, whereas mPGC(LC) specification An aggregate culture of hPGCLCs with mouse
It seems reasonable to expect that primates, is directly coupled with epigenetic reprogram- embryonic ovarian somatic cells [xenogeneic
including humans, adopt a strategy for germ ming (31, 46, 60, 93), hPGCLC specification is rOvary (xrOvary)] has been used to induce
cell specification that reflects their structur- genetically dissociable from epigenetic reprog- hPGCLCs into human germ cells at later de-
al characteristics, which are different from ramming (94). Upon mPGC(LC) specification, velopmental stages (81) (Fig. 4). In xrOvaries,
those in mice and other mammals. The hPGC BLIMP1, PRDM14, and TFAP2C repress key hPGCLCs interact with mouse granulosa cells
transcriptome obtained at the earliest stage machineries for de novo and maintenance DNA and induce maturation of their characteristics,
(E16 to E19) may shed new light on their origin methylation (Dnmt3a, Dnmt3b, and Uhrf1) and albeit at slower speed and with lower efficiency
and properties (69). create a state with little DNA methylation ac- than hPGCs in vivo: At around 70 days in cul-
tivity, leading to a replication-coupled, pas- ture, hPGCLC-derived cells differentiate into
hPGCLC induction and mechanism of sive genome-wide DNA demethylation upon oogonia or gonocyte-like cells, expressing key
hPGC(LC) specification mPGC(LC) proliferation (31, 46, 60, 93, 95, 96). genes including DAZL, DDX4, and PIWIL2,
Methodologies to induce human PGCLCs In contrast, although hPGCLC specification and notably, at around 120 days, they acquire
(hPGCLCs) from hPSCs have been explored. accompanies a down-regulation of DNMT3A/ a state highly similar to RA-responsive FGCs
hPSCs cultured under a 4i condition [chemical 3B, it maintains UHRF1, which recruits DNMT1 (Fig. 4). Accordingly, they undergo genome-
inhibitors for MEK, GSK3b, p38, and c-Jun into replication foci, at a substantial level, and wide DNA demethylation, imprint erasure, and
N-terminal kinase (JNK)] (78), which represent hPGCLCs retain their genome-wide DNA meth- partial X reactivation (81). These findings es-
a perigastrulating epiblast-like state (67, 79) ylation profiles and levels upon their prolif- tablish a foundation for human IVG. Similarly,
(Fig. 2), are induced into hPGCLCs in response eration (94), highlighting another layer of TF it has been shown that an aggregation culture
to BMP signaling (79) (Fig. 4). Alternatively, network divergence between mouse and human of hPGCLCs with mouse embryonic testicular
hPSCs cultured under a conventional condi- PGC(LC) specification. somatic cells [xenogeneic rTestis (xrTestis)]
tion are first induced into incipient mesoderm- differentiates hPGCLCs into gonocytes and
like cells (iMeLCs) and then into hPGCLCs by Human early oocyte and then into prospermatogonia in a time frame
BMP signaling (80) (Fig. 4). hPGCLCs in both prospermatogonia induction similar to that of the differentiation of hPGCLCs
studies show gene expression properties sim- Advances in omics technologies have created to oogonia to RA-responsive FGCs (82).
ilar to early hPGCs and to each other (79, 80). rapid progress in our understanding of the The mechanism underlying hPGCLC matu-
Subsequent studies have shown that hPGCLCs transcriptome and epigenetic dynamics as- ration into early oocytes and prospermatogonia
have the potential to undergo epigenetic repro- sociated with human germ cell development in xrOvaries and xrTestes, respectively, is un-
gramming, a hallmark of germ cells, and differ- (5, 82, 97–104). Human germ cells use a dis- known. The defined culture system for hPGCLC
entiate into early oocytes or prospermatogonia tinctive gene regulatory network and, over a expansion should serve as a platform to explore
(81, 82) (Fig. 4), indicating that hPGCLCs are period of a few months, undergo epigenetic such mechanisms, including a mechanism to
a bona fide in vitro counterpart of hPGCs (see reprogramming with a cell-to-cell variation. induce epigenetic reprogramming (94). The
below). This includes genome-wide DNA demethyla- impact of the X chromosome states in female
These studies show that the primed plu- tion, with imprint erasure and X chromosome hPSCs on the hPGCLC specification and matura-
ripotency of hPSCs is distinct from that of reactivation in females and with evolution- tion requires investigation. hPSCs typically bear
mEpiSCs, because the latter show little com- arily young transposons relatively resistant to one active and one inactive X chromosome

Saitou et al., Science 374, eaaz6830 (2021) 1 October 2021 6 of 9

 Corrected 19 October 2021. See full text.
RES EARCH | R E V I E W

(Xa and Xi, respectively). However, the Xi Concomitantly, the development of organ alies. Thus, IVG-derived animals should be
in hPSCs is unstable, with the expression of (ovary or testis) culture systems that allow the scrutinized with respect to many parameters,
XIST (X-inactive specific transcript), a long maturation of immature germ cells in humans including gene expression, epigenetic proper-
noncoding RNA important for X inactivation, and other species represents a critical chal- ties, behavior, longevity, and disease suscep-
being progressively repressed upon hPSC pas- lenge. Meanwhile, transplantation-based strat- tibility. Given that some epigenetic mutations
sages, leading to a partial derepression of Xi egies could be a first option to demonstrate are heritable (125), these evaluations should be
(an X-eroded state, Xe) (107–109). There is a the feasibility of generating PGCLC-derived performed for several generations. It is also
trend in which female hPSCs show lower effi- gametes in primates and other species, par- essential to create IVG-derived animal models
ciency for hPGCLC induction than do male ticularly in males (89, 90, 119–121). Efforts to more relevant to human physiology, such as
hPSCs (92, 110), and the degree of the X chro- expand SSCs in vitro (GSCs) in humans and IVG-derived macaques, and to perform similar
mosome reactivation in oogonia might re- other species are also critical. normality assessments.
flect the original X chromosome state in hPSCs An ultimate goal of IVG research will be to Concomitantly, the genetic and epigenetic
and influence the maturation of oogonia into reconstitute the entirety of gametogenesis quality of hPSCs and the resultant gametes
oocytes (81). using defined factors only. The processes of require careful assessment. It has become in-
mPGCLC specification, propagation, epigenetic creasingly evident that somatic cells accumu-
Key challenges reprogramming, and female sex determination late de novo mutations widely (126, 127) and
IVG has established itself as a key research have been recapitulated (60–62), and hPGCLCs that iPSCs acquire additional genetic and epi-
area in biomedical science, one facing a num- have been propagated to ~106-fold under de- genetic mutations during their derivation
ber of salient challenges in coming years. fined conditions (94). Moreover, key TFs driv- (49, 50), although most such mutations can
Although in vitro oogenesis in mice has been ing the corresponding respective processes as be neutral. In contrast, germ cells exhibit a
achieved, PSC-based in vitro spermatogenesis well as oocyte growth have been determined mutation rate that is about 1/10 that of somatic
remains a key challenge. There have been re- (36, 56, 62, 88). With further understanding cells (128, 129), with SSCs bearing a higher

Downloaded from https://www.science.org on December 05, 2023

ports describing an organ culture of immature of the mechanisms for germ cell development, fidelity for genome replication and oocytes
testicular fragments in mice, by which sperma- including those for later processes such as sper- entering into the meiotic prophase early during
togonia and SSCs go through proper meiotic miogenesis in males and full oocyte growth in their development (130, 131). Nonetheless,
divisions and develop into spermatids with full females, the extent to which the gametogenesis studies have shown that offspring of aged
developmental potential (111). With this system, processes are recapitulated under defined con- fathers who accumulate more mutations may
GSCs can be used as an appropriate donor (112). ditions would expand further. be susceptible to certain diseases, including
The most recent version of this system uses a As an alternative source for haploid cells, autism spectrum disorders (130, 131). SSC
multilayered microfluidic chamber, where sem- haploid ESCs have been generated from andro- clones bearing mutations advantageous for
iniferous tubules spread into a sheetlike struc- genetic or parthenogenetic (bearing only the their expansion can dominate a testicular niche
ture, allowing an equal distribution of nutrients paternal or maternal genome, respectively) and produce spermatozoa with the same mu-
and oxygen across the tissue masses (113). Ac- haploid eggs in mice and humans, providing tations, resulting in offspring with the corre-
cordingly, a potential strategy for an mPSC- a platform to explore the function of genes in sponding disorders (132, 133). Human iPSC
based in vitro spermatogenesis could be to relevant developmental pathways [see (122, 123) (hiPSC)–derived gametes, which are not orig-
combine an improved rTestis-based GSCLC for review]. On the other hand, mouse offspring inally protected by the germline mechanisms,
induction with an organ culture of GSCLC- derived from such cells often die prematurely could bear more mutations than in these ex-
transplanted testes. Indeed, using such strategy, in part because of aberrant imprints, indicating amples. A potential strategy to circumvent this
a recent study has demonstrated the in vitro genetic or epigenetic vulnerability of haploid problem may be to reserve embryonic or neo-
reconstitution of whole male germ cell devel- mESCs (122, 123). Further studies are required natal cells, such as cord blood cells, which
opment from mouse PSCs (Fig. 4) (114). for appropriate applications of haploid mouse would accumulate fewer mutations, as donors
To further extend IVG in humans, use of a and human ESCs. for future hiPSC generation. Clearly, more
xenogeneic system would not be ideal, because studies are needed to assess the impact of
species differences would preclude efficient Critical considerations genetic and epigenetic mutations on human
human gametogenesis in such systems. Greater The realization of human IVG will create pos- physiology and diseases.
success would be expected if human fetal or sibilities in reproductive medicine, such as With the exception of the germ cell origins,
neonatal gonadal somatic cells were used to opportunities for diagnosing and modeling IVG seeks to recapitulate all key genetic and
form human rOvaries or rTestes. The isola- infertility and genetic and epigenetic disor- epigenetic events during germ cell development
tion and use of adequate amounts of human ders, exploring their remedies, and improving in vivo, including epigenetic reprogramming
fetal gonadal somatic cells at an appropriate culture methodologies for assisted reproduc- and programming and meiotic recombination.
developmental stage, however, represents a tive technologies (124). If deemed legally and In the future, when all the technological con-
technical and ethical challenge. Similarly, using ethically permissible, the use of IVG-derived cerns have been resolved, it will be crucial to
gonadal material from the same species would human gametes for producing offspring would hold wide discussions about whether to use
also improve efficiency for IVG in endangered require careful assessments and intensive studies. IVG-derived gametes for human reproduc-
animals (115). A strategy to fundamentally re- It would be essential that the “normality” of the tion, because such an application involves
solve this limitation is to induce such cells from IVG-derived animal models be strictly assessed, somatic cells as human origins and changes
PSCs. Indeed, in line with the advancements although thus far such mouse models appear our understanding of the continuity of life
in our understanding of the mechanism of grossly normal (28, 29, 44, 58–60, 93). It should through germ cells.
gonadal development (116, 117), fetal ovarian be noted that many of the mouse embryos gen-
somatic cell–like cells fully capable of supporting erated from oocytes via in vitro oogenesis died Conclusions
oogenesis have been induced from mPSCs (118). prenatally owing to abnormal development, The salient functions of germ cells are to pre-
This study will serve as a basis to supply cells and even those born apparently normally bore serve genetic and epigenetic information with
equivalent to embryonic gonadal somatic cells heavier placentae (44), suggesting that the high fidelity and also, somewhat paradoxically,
in humans and other species in coming years. surviving animals may have cryptic anom- to create diversity in this information through

Saitou et al., Science 374, eaaz6830 (2021) 1 October 2021 7 of 9

 Corrected 19 October 2021. See full text.
RES EARCH | R E V I E W

distinctive mechanisms as a driving force for 20. J. Nichols, A. Smith, Naive and primed pluripotent states. 43. K. Morohaku et al., Complete in vitro generation of fertile
evolution. With proof-of-concept studies in Cell Stem Cell 4, 487–492 (2009). doi: 10.1016/ oocytes from mouse primordial germ cells. Proc. Natl. Acad.
j.stem.2009.05.015; pmid: 19497275 Sci. U.S.A. 113, 9021–9026 (2016). doi: 10.1073/
mice and a growing extension to humans, IVG 21. Q. L. Ying et al., The ground state of embryonic stem cell pnas.1603817113; pmid: 27457928
has great potential for future applications. self-renewal. Nature 453, 519–523 (2008). doi: 10.1038/ 44. O. Hikabe et al., Reconstitution in vitro of the entire cycle of
On the other hand, the genetic as well as epi- nature06968; pmid: 18497825 the mouse female germ line. Nature 539, 299–303 (2016).
22. M. Buehr et al., Capture of authentic embryonic stem cells doi: 10.1038/nature20104; pmid: 27750280
genetic integrities of gametes and developing from rat blastocysts. Cell 135, 1287–1298 (2008). 45. E. Habibi et al., Whole-genome bisulfite sequencing of
germ cells generated through IVG still appear doi: 10.1016/j.cell.2008.12.007; pmid: 19109897 two distinct interconvertible DNA methylomes of mouse
substantially lower than those of gametes and 23. P. Li et al., Germline competent embryonic stem cells derived embryonic stem cells. Cell Stem Cell 13, 360–369 (2013).
from rat blastocysts. Cell 135, 1299–1310 (2008). doi: 10.1016/j.stem.2013.06.002; pmid: 23850244
germ cells generated under in vivo contexts, doi: 10.1016/j.cell.2008.12.006; pmid: 19109898 46. K. Shirane et al., Global landscape and regulatory principles of
indicating that there is much to be learned 24. T. Boroviak, R. Loos, P. Bertone, A. Smith, J. Nichols, The DNA methylation reprogramming for germ cell specification by
from investigations into germ cell development ability of inner-cell-mass cells to self-renew as embryonic mouse pluripotent stem cells. Dev. Cell 39, 87–103 (2016).
stem cells is acquired following epiblast specification. doi: 10.1016/j.devcel.2016.08.008; pmid: 27642137
in vivo in diverse organisms.
Nat. Cell Biol. 16, 513–528 (2014). doi: 10.1038/ncb2965; 47. M. Yagi et al., Derivation of ground-state female ES cells
pmid: 24859004 maintaining gamete-derived DNA methylation. Nature 548,
25. P. J. Tesar et al., New cell lines from mouse epiblast share 224–227 (2017). doi: 10.1038/nature23286; pmid: 28746308
RE FE RENCES AND N OT ES defining features with human embryonic stem cells. 48. J. Choi et al., Prolonged Mek1/2 suppression impairs the
1. M. Saitou, H. Miyauchi, Gametogenesis from pluripotent stem Nature 448, 196–199 (2007). doi: 10.1038/nature05972; developmental potential of embryonic stem cells. Nature
cells. Cell Stem Cell 18, 721–735 (2016). doi: 10.1016/ pmid: 17597760 548, 219–223 (2017). doi: 10.1038/nature23274;
j.stem.2016.05.001; pmid: 27257761 26. I. G. Brons et al., Derivation of pluripotent epiblast stem cells pmid: 28746311
2. W. W. Tang, T. Kobayashi, N. Irie, S. Dietmann, M. A. Surani, from mammalian embryos. Nature 448, 191–195 (2007). 49. U. Ben-David, N. Benvenisty, The tumorigenicity of human
Specification and epigenetic programming of the human doi: 10.1038/nature05950; pmid: 17597762 embryonic and induced pluripotent stem cells. Nat. Rev. Cancer
germ line. Nat. Rev. Genet. 17, 585–600 (2016). doi: 10.1038/ 27. Y. Kojima et al., The transcriptional and functional properties 11, 268–277 (2011). doi: 10.1038/nrc3034; pmid: 21390058
nrg.2016.88; pmid: 27573372 of mouse epiblast stem cells resemble the anterior primitive 50. S. Bar, N. Benvenisty, Epigenetic aberrations in human
3. H. J. Lee, T. A. Hore, W. Reik, Reprogramming the streak. Cell Stem Cell 14, 107–120 (2014). doi: 10.1016/ pluripotent stem cells. EMBO J. 38, e101033 (2019).
methylome: Erasing memory and creating diversity. Cell Stem j.stem.2013.09.014; pmid: 24139757 doi: 10.15252/embj.2018101033; pmid: 31088843

Downloaded from https://www.science.org on December 05, 2023

Cell 14, 710–719 (2014). doi: 10.1016/j.stem.2014.05.008; 28. K. Hayashi, H. Ohta, K. Kurimoto, S. Aramaki, M. Saitou, 51. S. Shimamoto et al., Hypoxia induces the dormant state in
pmid: 24905162 Reconstitution of the mouse germ cell specification oocytes through expression of Foxo3. Proc. Natl. Acad. Sci.
4. C. Spiller, P. Koopman, J. Bowles, Sex determination in the pathway in culture by pluripotent stem cells. Cell 146, U.S.A. 116, 12321–12326 (2019). doi: 10.1073/pnas.1817223116;
mammalian germline. Annu. Rev. Genet. 51, 265–285 519–532 (2011). doi: 10.1016/j.cell.2011.06.052; pmid: 31147464
(2017). doi: 10.1146/annurev-genet-120215-035449; pmid: 21820164 52. Y. Feng et al., CLARITY reveals dynamics of ovarian
pmid: 28853925 29. K. Hayashi et al., Offspring from oocytes derived from in follicular architecture and vasculature in three-dimensions.
5. L. Wen, F. Tang, Human germline cell development: from the vitro primordial germ cell–like cells in mice. Science 338, Sci. Rep. 7, 44810 (2017). doi: 10.1038/srep44810;
perspective of single-cell sequencing. Mol. Cell 76, 320–328 971–975 (2012). doi: 10.1126/science.1226889; pmid: 28333125
(2019). doi: 10.1016/j.molcel.2019.08.025; pmid: 31563431 pmid: 23042295 53. G. Nagamatsu, S. Shimamoto, N. Hamazaki, Y. Nishimura,
6. M. A. Edson, A. K. Nagaraja, M. M. Matzuk, The mammalian 30. A. Smith, Formative pluripotency: The executive phase in a K. Hayashi, Mechanical stress accompanied with nuclear
ovary from genesis to revelation. Endocr. Rev. 30, 624–712 developmental continuum. Development 144, 365–373 rotation is involved in the dormant state of mouse oocytes.
(2009). doi: 10.1210/er.2009-0012; pmid: 19776209 (2017). doi: 10.1242/dev.142679; pmid: 28143843 Sci. Adv. 5, eaav9960 (2019). doi: 10.1126/sciadv.aav9960;
7. E. A. McGee, A. J. Hsueh, Initial and cyclic recruitment of 31. K. Kurimoto et al., Quantitative dynamics of chromatin pmid: 31249869
ovarian follicles. Endocr. Rev. 21, 200–214 (2000). remodeling during germ cell specification from mouse 54. N. Hamada et al., Germ cell-intrinsic effects of sex
pmid: 10782364 embryonic stem cells. Cell Stem Cell 16, 517–532 (2015). chromosomes on early oocyte differentiation in mice.
8. G. Manku, M. Culty, Mammalian gonocyte and spermatogonia doi: 10.1016/j.stem.2015.03.002; pmid: 25800778 PLOS Genet. 16, e1008676 (2020). doi: 10.1371/
differentiation: Recent advances and remaining challenges. 32. Y. Xiang et al., Epigenomic analysis of gastrulation identifies a journal.pgen.1008676; pmid: 32214314
Reproduction 149, R139–R157 (2015). doi: 10.1530/ unique chromatin state for primed pluripotency. Nat. Genet. 55. Y. Yamauchi, J. M. Riel, Z. Stoytcheva, M. A. Ward, Two Y
REP-14-0431; pmid: 25670871 52, 95–105 (2020). doi: 10.1038/s41588-019-0545-1; genes can replace the entire Y chromosome for assisted
9. S. Chuma, T. Nakano, piRNA and spermatogenesis in mice. pmid: 31844322 reproduction in the mouse. Science 343, 69–72 (2014).
Philos. Trans. R. Soc. London Ser. B 368, 20110338 (2013). 33. M. Kinoshita et al., Capture of mouse and human stem cells doi: 10.1126/science.1242544; pmid: 24263135
doi: 10.1098/rstb.2011.0338; pmid: 23166399 with features of formative pluripotency. Cell Stem Cell 28, 56. N. Hamazaki et al., Reconstitution of the oocyte transcriptional
10. M. D. Griswold, Spermatogenesis: The commitment to 453–471.e8 (2021). doi: 10.1016/j.stem.2020.11.005; network with transcription factors. Nature 589, 264–269
meiosis. Physiol. Rev. 96, 1–17 (2016). doi: 10.1152/ pmid: 33271069 (2021). doi: 10.1038/s41586-020-3027-9; pmid: 33328630
physrev.00013.2015; pmid: 26537427 34. L. Yu et al., Derivation of intermediate pluripotent stem cells 57. G. A. Dokshin, A. E. Baltus, J. J. Eppig, D. C. Page, Oocyte
11. M. Kanatsu-Shinohara, T. Shinohara, Spermatogonial stem cell amenable to primordial germ cell specification. Cell Stem Cell differentiation is genetically dissociable from meiosis in mice.
self-renewal and development. Annu. Rev. Cell Dev. Biol. 29, 28, 550–567.e12 (2021). doi: 10.1016/j.stem.2020.11.003; Nat. Genet. 45, 877–883 (2013). doi: 10.1038/ng.2672;
163–187 (2013). doi: 10.1146/annurev-cellbio-101512-122353; pmid: 33271070 pmid: 23770609
pmid: 24099084 35. S. Aramaki et al., A mesodermal factor, T, specifies mouse 58. Y. Ishikura et al., In vitro derivation and propagation of
12. T. Taketo, The role of sex chromosomes in mammalian germ germ cell fate by directly activating germline determinants. spermatogonial stem cell activity from mouse pluripotent
cell differentiation: Can the germ cells carrying X and Dev. Cell 27, 516–529 (2013). doi: 10.1016/ stem cells. Cell Rep. 17, 2789–2804 (2016). doi: 10.1016/
Y chromosomes differentiate into fertile oocytes? Asian J. j.devcel.2013.11.001; pmid: 24331926 j.celrep.2016.11.026; pmid: 27926879
Androl. 17, 360–366 (2015). pmid: 25578929 36. F. Nakaki et al., Induction of mouse germ-cell fate by 59. Q. Zhou et al., Complete meiosis from embryonic stem
13. P. S. Burgoyne, The role of the mammalian Y chromosome in transcription factors in vitro. Nature 501, 222–226 (2013). cell-derived germ cells in vitro. Cell Stem Cell 18, 330–340
spermatogenesis. Development 101 (suppl.), 133–141 (1987). doi: 10.1038/nature12417; pmid: 23913270 (2016). doi: 10.1016/j.stem.2016.01.017; pmid: 26923202
doi: 10.1242/dev.101.Supplement.133; pmid: 3503711 37. K. Murakami et al., NANOG alone induces germ cells in 60. H. Ohta et al., In vitro expansion of mouse primordial germ
14. K. A. Lawson et al., Bmp4 is required for the generation of primed epiblast in vitro by activation of enhancers. Nature cell-like cells recapitulates an epigenetic blank slate. EMBO J.
primordial germ cells in the mouse embryo. Genes Dev. 13, 529, 403–407 (2016). doi: 10.1038/nature16480; 36, 1888–1907 (2017). doi: 10.15252/embj.201695862;
424–436 (1999). doi: 10.1101/gad.13.4.424; pmid: 10049358 pmid: 26751055 pmid: 28559416
15. M. Saitou, S. C. Barton, M. A. Surani, A molecular programme 38. J. Zhang et al., OTX2 restricts entry to the mouse germline. 61. H. Miyauchi et al., Bone morphogenetic protein and retinoic
for the specification of germ cell fate in mice. Nature 418, Nature 562, 595–599 (2018). doi: 10.1038/s41586-018-0581-5; acid synergistically specify female germ-cell fate in mice.
293–300 (2002). doi: 10.1038/nature00927; pmid: 12124616 pmid: 30283136 EMBO J. 36, 3100–3119 (2017). doi: 10.15252/
16. Y. Ohinata et al., Blimp1 is a critical determinant of the 39. J. Tischler et al., Metabolic regulation of pluripotency and embj.201796875; pmid: 28928204
germ cell lineage in mice. Nature 436, 207–213 (2005). germ cell fate through a-ketoglutarate. EMBO J. 38, e100615 62. S. I. Nagaoka et al., ZGLP1 is a determinant for the oogenic
doi: 10.1038/nature03813; pmid: 15937476 (2019). doi: 10.15252/embj.201899518; pmid: 30257965 fate in mice. Science 367, eaaw4115 (2020). doi: 10.1126/
17. Y. Ohinata et al., A signaling principle for the specification of 40. T. Hirota et al., Fertile offspring from sterile sex chromosome science.aaw4115; pmid: 32054698
the germ cell lineage in mice. Cell 137, 571–584 (2009). trisomic mice. Science 357, 932–935 (2017). doi: 10.1126/ 63. N. Vernet et al., Meiosis occurs normally in the fetal ovary of
doi: 10.1016/j.cell.2009.03.014; pmid: 19410550 science.aam9046; pmid: 28818972 mice lacking all retinoic acid receptors. Sci. Adv. 6, eaaz1139
18. M. Yamaji et al., Critical function of Prdm14 for the 41. J. J. Eppig, M. J. O’Brien, Development in vitro of mouse (2020). doi: 10.1126/sciadv.aaz1139; pmid: 32917583
establishment of the germ cell lineage in mice. Nat. Genet. oocytes from primordial follicles. Biol. Reprod. 54, 197–207 64. A. A. Chassot et al., Retinoic acid synthesis by ALDH1A
40, 1016–1022 (2008). doi: 10.1038/ng.186; pmid: 18622394 (1996). doi: 10.1095/biolreprod54.1.197; pmid: 8838017 proteins is dispensable for meiosis initiation in the mouse
19. S. Weber et al., Critical function of AP-2 gamma/TCFAP2C in 42. K. Hashimoto, M. Noguchi, N. Nakatsuji, Mouse offspring fetal ovary. Sci. Adv. 6, eaaz1261 (2020). doi: 10.1126/
mouse embryonic germ cell maintenance. Biol. Reprod. 82, derived from fetal ovaries or reaggregates which were cultured sciadv.aaz1261; pmid: 32494737
214–223 (2010). doi: 10.1095/biolreprod.109.078717; and transplanted into adult females. Dev. Growth Differ. 34, 65. Y. Lin, M. E. Gill, J. Koubova, D. C. Page, Germ cell-intrinsic
pmid: 19776388 233–238 (1992). doi: 10.1111/j.1440-169X.1992.tb00012.x and -extrinsic factors govern meiotic initiation in mouse

Saitou et al., Science 374, eaaz6830 (2021) 1 October 2021 8 of 9

 Corrected 19 October 2021. See full text.
RES EARCH | R E V I E W

embryos. Science 322, 1685–1687 (2008). doi: 10.1126/ 90. Y. Sakai et al., Induction of the germ cell fate from pluripotent 114. Y. Ishikura et al., In vitro reconstitution of the whole male
science.1166340; pmid: 19074348 stem cells in cynomolgus monkeys. Biol. Reprod. 102, 620–638 germ-cell development from mouse pluripotent stem cells.
66. J. A. Thomson et al., Embryonic stem cell lines derived from (2020). doi: 10.1093/biolre/ioz205; pmid: 31724030 Cell Stem Cell 10.1016/j.stem.2021.08.005 (2021).
human blastocysts. Science 282, 1145–1147 (1998). 91. T. Kobayashi et al., Principles of early human development 115. T. B. Hildebrandt et al., Embryos and embryonic stem cells
doi: 10.1126/science.282.5391.1145; pmid: 9804556 and germ cell program from conserved model systems. Nature from the white rhinoceros. Nat. Commun. 9, 2589 (2018).
67. T. Nakamura et al., A developmental coordinate of 546, 416–420 (2017). doi: 10.1038/nature22812; pmid: 28607482 doi: 10.1038/s41467-018-04959-2; pmid: 29973581
pluripotency among mice, monkeys and humans. Nature 537, 92. S. Yokobayashi et al., Clonal variation of human induced 116. I. Stévant et al., Dissecting cell lineage specification and sex
57–62 (2016). doi: 10.1038/nature19096; pmid: 27556940 pluripotent stem cells for induction into the germ cell fate. fate determination in gonadal somatic cells using single-cell
68. K. Sasaki et al., The germ cell fate of cynomolgus monkeys is Biol. Reprod. 96, 1154–1166 (2017). doi: 10.1093/biolre/ transcriptomics. Cell Rep. 26, 3272–3283.e3 (2019).
specified in the nascent amnion. Dev. Cell 39, 169–185 iox038; pmid: 28453617 doi: 10.1016/j.celrep.2019.02.069; pmid: 30893600
(2016). doi: 10.1016/j.devcel.2016.09.007; pmid: 27720607 93. H. Ohta et al., Cyclosporin A and FGF signaling support the 117. K. Sasaki et al., The embryonic ontogeny of the gonadal
69. R. C. V. Tyser, E. Mahammadov, S. Nakanoh, L. Vallier, proliferation/survival of mouse primordial germ cell-like cells somatic cells in mice and monkeys. Cell Rep. 35, 109075
A. Scialdone, S. Srinivas, A spatially resolved single cell atlas in vitro. Biol. Reprod. 104, 344–360 (2021). doi: 10.1093/ (2021). doi: 10.1016/j.celrep.2021.109075; pmid: 33951437
of human gastrulation. bioRxiv 2020.07.21.213512 [Preprint] biolre/ioaa195; pmid: 33079185 118. T. Yoshino et al., Generation of ovarian follicles from mouse
(2020). https://doi.org/10.1101/2020.07.21.213512. 94. Y. Murase et al., Long-term expansion with germline potential pluripotent stem cells. Science 373, eabe0237 (2021).
70. T. W. Theunissen et al., Systematic identification of culture of human primordial germ cell-like cells in vitro. EMBO J. 39, doi: 10.1126/science.abe0237; pmid: 34437124
conditions for induction and maintenance of naive human e104929 (2020). doi: 10.15252/embj.2020104929; 119. Z. Liu et al., Generation of macaques with sperm derived
pluripotency. Cell Stem Cell 15, 471–487 (2014). doi: 10.1016/ pmid: 32954504 from juvenile monkey testicular xenografts. Cell Res. 26,
j.stem.2014.07.002; pmid: 25090446 95. S. Seisenberger et al., The dynamics of genome-wide DNA 139–142 (2016). doi: 10.1038/cr.2015.112; pmid: 26369429
71. Y. Takashima et al., Resetting transcription factor control methylation reprogramming in mouse primordial germ cells. 120. A. P. Fayomi et al., Autologous grafting of cryopreserved
circuitry toward ground-state pluripotency in human. Cell Mol. Cell 48, 849–862 (2012). doi: 10.1016/j. prepubertal rhesus testis produces sperm and offspring.
158, 1254–1269 (2014). doi: 10.1016/j.cell.2014.08.029; molcel.2012.11.001; pmid: 23219530 Science 363, 1314–1319 (2019). doi: 10.1126/science.
pmid: 25215486 96. S. Kagiwada, K. Kurimoto, T. Hirota, M. Yamaji, M. Saitou, aav2914; pmid: 30898927
72. C. Dong et al., Derivation of trophoblast stem cells from naïve Replication-coupled passive DNA demethylation for the 121. C. Ramathal et al., Fate of iPSCs derived from azoospermic
human pluripotent stem cells. eLife 9, e52504 (2020). erasure of genome imprints in mice. EMBO J. 32, 340–353 and fertile men following xenotransplantation to murine
doi: 10.7554/eLife.52504; pmid: 32048992 (2013). doi: 10.1038/emboj.2012.331; pmid: 23241950 seminiferous tubules. Cell Rep. 7, 1284–1297 (2014).
73. J. K. Cinkornpumin et al., Naive human embryonic stem cells 97. S. Gkountela et al., DNA demethylation dynamics in the doi: 10.1016/j.celrep.2014.03.067; pmid: 24794432
can give rise to cells with a trophoblast-like transcriptome human prenatal germline. Cell 161, 1425–1436 (2015). 122. A. Wutz, Haploid mouse embryonic stem cells: Rapid

Downloaded from https://www.science.org on December 05, 2023

and methylome. Stem Cell Reports 15, 198–213 (2020). doi: 10.1016/j.cell.2015.05.012; pmid: 26004067 genetic screening and germline transmission. Annu. Rev.
doi: 10.1016/j.stemcr.2020.06.003; pmid: 32619492 98. W. W. Tang et al., A unique gene regulatory network resets the Cell Dev. Biol. 30, 705–722 (2014). doi: 10.1146/
74. L. Yu et al., Blastocyst-like structures generated from human human germline epigenome for development. Cell 161, 1453–1467 annurev-cellbio-100913-012920; pmid: 25288120
pluripotent stem cells. Nature 591, 620–626 (2021). (2015). doi: 10.1016/j.cell.2015.04.053; pmid: 26046444 123. I. Sagi, N. Benvenisty, Haploidy in humans: An evolutionary
doi: 10.1038/s41586-021-03356-y; pmid: 33731924 99. F. Guo et al., The transcriptome and DNA methylome landscapes and developmental perspective. Dev. Cell 41, 581–589 (2017).
75. X. Liu et al., Modelling human blastocysts by reprogramming of human primordial germ cells. Cell 161, 1437–1452 (2015). doi: 10.1016/j.devcel.2017.04.019; pmid: 28633015
fibroblasts into iBlastoids. Nature 591, 627–632 (2021). doi: 10.1016/j.cell.2015.05.015; pmid: 26046443 124. T. Ishii, M. Saitou, Promoting in vitro gametogenesis research
doi: 10.1038/s41586-021-03372-y; pmid: 33731926 100. L. Li et al., Single-cell RNA-seq analysis maps development of with a social understanding. Trends Mol. Med. 23, 985–988
76. Y. Niu et al., Dissecting primate early post-implantation development human germline cells and gonadal niche interactions. (2017). doi: 10.1016/j.molmed.2017.09.006; pmid: 29032005
using long-term in vitro embryo culture. Science 366, eaaw5754 Cell Stem Cell 20, 858–873.e4 (2017). doi: 10.1016/ 125. E. Heard, R. A. Martienssen, Transgenerational epigenetic
(2019). doi: 10.1126/science.aaw5754; pmid: 31672917 j.stem.2017.03.007; pmid: 28457750 inheritance: Myths and mechanisms. Cell 157, 95–109
77. D. Chen et al., Human primordial germ cells are specified 101. Á. Vértesy et al., Parental haplotype-specific single-cell (2014). doi: 10.1016/j.cell.2014.02.045; pmid: 24679529
from lineage-primed progenitors. Cell Rep. 29, 4568–4582.e5 transcriptomics reveal incomplete epigenetic reprogramming 126. S. Jaiswal et al., Age-related clonal hematopoiesis associated
(2019). doi: 10.1016/j.celrep.2019.11.083; pmid: 31875561 in human female germ cells. Nat. Commun. 9, 1873 (2018). with adverse outcomes. N. Engl. J. Med. 371, 2488–2498
78. O. Gafni et al., Derivation of novel human ground state naive doi: 10.1038/s41467-018-04215-7; pmid: 29760424 (2014). doi: 10.1056/NEJMoa1408617; pmid: 25426837
pluripotent stem cells. Nature 504, 282–286 (2013). 102. F. Fang et al., A PAX5-OCT4-PRDM1 developmental switch 127. A. Yokoyama et al., Age-related remodelling of oesophageal
doi: 10.1038/nature12745; pmid: 24172903 specifies human primordial germ cells. Nat. Cell Biol. 20, 655–665 epithelia by mutated cancer drivers. Nature 565, 312–317
79. N. Irie et al., SOX17 is a critical specifier of human primordial (2018). doi: 10.1038/s41556-018-0094-3; pmid: 29713018 (2019). doi: 10.1038/s41586-018-0811-x; pmid: 30602793
germ cell fate. Cell 160, 253–268 (2015). doi: 10.1016/ 103. J. Guo et al., Single-cell analysis of the developing human 128. B. Milholland et al., Differences between germline and somatic
j.cell.2014.12.013; pmid: 25543152 testis reveals somatic niche cell specification and fetal mutation rates in humans and mice. Nat. Commun. 8, 15183
80. K. Sasaki et al., Robust in vitro induction of human germ cell germline stem cell establishment. Cell Stem Cell 28, 764–778. (2017). doi: 10.1038/ncomms15183; pmid: 28485371
fate from pluripotent stem cells. Cell Stem Cell 17, 178–194 e4 (2021). doi: 10.1016/j.stem.2020.12.004; pmid: 33453151 129. L. Ségurel, M. J. Wyman, M. Przeworski, Determinants of
(2015). doi: 10.1016/j.stem.2015.06.014; pmid: 26189426 104. T. Chitiashvili et al., Female human primordial germ cells mutation rate variation in the human germline. Annu. Rev.
81. C. Yamashiro et al., Generation of human oogonia from display X-chromosome dosage compensation despite Genomics Hum. Genet. 15, 47–70 (2014). doi: 10.1146/
induced pluripotent stem cells in vitro. Science 362, 356–360 the absence of X-inactivation. Nat. Cell Biol. 22, 1436–1446 annurev-genom-031714-125740; pmid: 25000986
(2018). doi: 10.1126/science.aat1674; pmid: 30237246 (2020). doi: 10.1038/s41556-020-00607-4; pmid: 33257808 130. A. Kong et al., Rate of de novo mutations and the importance
82. Y. S. Hwang et al., Reconstitution of prospermatogonial 105. T. G. Baker, A quantitative and cytological study of germ cells of father’s age to disease risk. Nature 488, 471–475 (2012).
specification in vitro from human induced pluripotent stem in human ovaries. Proc. R. Soc. London Ser. B 158, 417–433 doi: 10.1038/nature11396; pmid: 22914163
cells. Nat. Commun. 11, 5656 (2020). doi: 10.1038/ (1963). doi: 10.1098/rspb.1963.0055; pmid: 14070052 131. J. M. Goldmann et al., Parent-of-origin-specific signatures of
s41467-020-19350-3; pmid: 33168808 106. T. Fukuda, C. Hedinger, P. Groscurth, Ultrastructure of developing de novo mutations. Nat. Genet. 48, 935–939 (2016).
83. Y. Zheng et al., Controlled modelling of human epiblast and germ cells in the fetal human testis. Cell Tissue Res. 161, 55–70 doi: 10.1038/ng.3597; pmid: 27322544
amnion development using stem cells. Nature 573, 421–425 (1975). doi: 10.1007/BF00222114; pmid: 1149138 132. E. Giannoulatou et al., Contributions of intrinsic mutation rate
(2019). doi: 10.1038/s41586-019-1535-2; pmid: 31511693 107. S. Mekhoubad et al., Erosion of dosage compensation impacts and selfish selection to levels of de novo HRAS mutations in the
84. Y. Kojima et al., Evolutionarily distinctive transcriptional and human iPSC disease modeling. Cell Stem Cell 10, 595–609 paternal germline. Proc. Natl. Acad. Sci. U.S.A. 110, 20152–20157
signaling programs drive human germ cell lineage specification (2012). doi: 10.1016/j.stem.2012.02.014; pmid: 22560080 (2013). doi: 10.1073/pnas.1311381110; pmid: 24259709
from pluripotent stem cells. Cell Stem Cell 21, 517–532.e5 (2017). 108. C. Vallot et al., Erosion of X chromosome inactivation 133. G. J. Maher et al., Selfish mutations dysregulating RAS-MAPK
doi: 10.1016/j.stem.2017.09.005; pmid: 28985527 in human pluripotent cells initiates with XACT coating and signaling are pervasive in aged human testes. Genome Res. 28,
85. D. Chen et al., The TFAP2C-regulated OCT4 naive enhancer is depends on a specific heterochromatin landscape. Cell Stem 1779–1790 (2018). doi: 10.1101/gr.239186.118; pmid: 30355600
involved in human germline formation. Cell Rep. 25, 3591–3602. Cell 16, 533–546 (2015). doi: 10.1016/j.stem.2015.03.016;
e5 (2018). doi: 10.1016/j.celrep.2018.12.011; pmid: 30590035 pmid: 25921272 AC KNOWLED GME NTS
86. A. Sybirna et al., A critical role of PRDM14 in human primordial 109. S. Patel et al., Human embryonic stem cells do not change We thank the members of our laboratory for their input on this study.
germ cell fate revealed by inducible degrons. Nat. Commun. 11, 1282 their X inactivation status during differentiation. Cell Rep. 18, We apologize to the colleagues whose primary works are not cited
(2020). doi: 10.1038/s41467-020-15042-0; pmid: 32152282 54–67 (2017). doi: 10.1016/j.celrep.2016.11.054; pmid: 27989715 because of space constraints. Funding: This work was supported
87. M. Pierson Smela, A. Sybirna, F. C. K. Wong, M. A. Surani, 110. D. Chen et al., Germline competency of human embryonic by a Grant-in-Aid for Specially Promoted Research from JSPS
Testing the role of SOX15 in human primordial germ cell fate. stem cells depends on eomesodermin. Biol. Reprod. 97, (17H06098), a JST-ERATO grant (JPMJER1104), a grant from HFSP
Wellcome Open Res. 4, 122 (2019). doi: 10.12688/ 850–861 (2017). doi: 10.1093/biolre/iox138; pmid: 29091993 (RGP0057/2018), and grants from the Pythias Fund and Open
wellcomeopenres.15381.1; pmid: 31583280 111. T. Sato et al., In vitro production of functional sperm in Philanthropy Project to M.S., and by a Grant-in-Aid for Scientific
88. Y. Kojima et al., GATA transcription factors, SOX17 and cultured neonatal mouse testes. Nature 471, 504–507 (2011). Research (A) from JSPS (17H01395) and a Grant-in-Aid for Scientific
TFAP2C, drive the human germ-cell specification program. doi: 10.1038/nature09850; pmid: 21430778 Research on Innovative Areas from JSPS (18H05545) to K.H.
Life Sci. Alliance 4, e202000974 (2021). doi: 10.26508/ 112. T. Sato et al., In vitro production of fertile sperm from murine Competing interests: M.S. and K.H. are inventors on patent
lsa.202000974; pmid: 33608411 spermatogonial stem cell lines. Nat. Commun. 2, 472 (2011). applications relating to induction of germ cells from pluripotent stem
89. E. Sosa et al., Differentiation of primate primordial germ doi: 10.1038/ncomms1478; pmid: 21915114 cells filed by Kyoto University and Kyushu University, respectively, and
cell-like cells following transplantation into the adult gonadal 113. K. Kojima et al., Neonatal testis growth recreated in vitro by are founders of Houjou, Inc.
niche. Nat. Commun. 9, 5339 (2018). doi: 10.1038/ two-dimensional organ spreading. Biotechnol. Bioeng. 115,
s41467-018-07740-7; pmid: 30559363 3030–3041 (2018). doi: 10.1002/bit.26822; pmid: 30144353 10.1126/science.aaz6830

Saitou et al., Science 374, eaaz6830 (2021) 1 October 2021 9 of 9

 Mammalian in vitro gametogenesis
Mitinori Saitou and Katsuhiko Hayashi

Science 374 (6563), eaaz6830. DOI: 10.1126/science.aaz6830

Reconstituting reproduction in culture

Research on in vitro gametogenesis (IVG) aims to reconstitute germ cell development, oogenesis and
spermatogenesis, in culture. Saitou and Hayashi review some of the recent developments in mammalian IVG.
Advances in methods and culture conditions in mice to generate mature oocytes and spermatocytes from pluripotent
stem cells have informed similar studies with nonhuman primate and human cells, but differences among species
are clear. IVG has great potential for reproductive medicine, including novel diagnosis and modeling of infertility. The

Downloaded from https://www.science.org on December 05, 2023

realization of human IVG requires further intensive efforts, but as technical hurtles are overcome, careful consideration
must be given to the potential application of methods for reproductive purposes. —BAP

View the article online

https://www.science.org/doi/10.1126/science.aaz6830
Permissions
https://www.science.org/help/reprints-and-permissions

Use of this article is subject to the Terms of service

Science (ISSN 1095-9203) is published by the American Association for the Advancement of Science. 1200 New York Avenue NW,
Washington, DC 20005. The title Science is a registered trademark of AAAS.

Copyright © 2021 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim
to original U.S. Government Works