Generation of human oogonia from induced pluripotent stem cells in culture
The human germ-cell lineage originates as human primordial germ cells (hPGCs). hPGCs undergo genome-wide epigenetic reprogramming and differentiate into oogonia or gonocytes, precursors for oocytes or spermatogonia, respectively. Here, we describe a protocol to differentiate human induced pluripotent stem cells (hiPSCs) into oogonia in vitro. hiPSCs are induced into incipient mesoderm-like cells (iMeLCs) using activin A and a WNT pathway agonist. iMeLCs, or, alternatively, hPSCs cultured with divergent signaling inhibitors, are induced into hPGC-like cells (hPGCLCs) in floating aggregates by cytokines including bone morphogenic protein 4. hPGCLCs are aggregated with mouse embryonic ovarian somatic cells to form xenogeneic reconstituted ovaries, which are cultured under an air–liquid interface condition for ~4 months for hPGCLCs to differentiate into oogonia and immediate precursory states for oocytes. To date, this is the only approach that generates oogonia from hPGCLCs. The protocol is suitable for investigating the mechanisms of hPGC specification and epigenetic reprogramming.
During mammalian development, germ cells are induced from their pluripotent precursors as primordial germ cells (PGCs), before the formation of embryonic gonads (Fig. 1 ) 1 , 2 . Along with the progression of embryonic development, PGCs undergo migration through the developing hindgut and mesentery, eventually colonizing sexually uncommitted, incipient embryonic gonads, by which time they mature as oogonia in females or gonocytes in males 1 , 2 . A key event that occurs in such cells is epigenetic reprogramming, most notably, genome-wide DNA demethylation, which results in imprint erasure and X chromosome reactivation in females 1 , 2 , 3 . Oogonia and gonocytes are almost identical in terms of morphology, gene expression and epigenetic profiles, except that X-linked genes are more highly expressed in oogonia due to X reactivation 1 , 2 , 3 .
Fig. 1: A scheme for human germ-cell development.
A schematic representation of human germ-cell development, including the developmental time lines and key developmental events 1 , 2 . After implantation of blastocysts, human germ-cell lineage is assumed to be specified at around week 2 (2wk) of development as PGCs (top panel). The origin of human PGCs is currently unknown, but can be the amnion or the epiblast 29 , 30 . Human PGCs undergo migration through developing hindgut and mesentery and colonize the genital ridges, the primordia of either testes or ovaries, from around 6wk, to differentiate as gonocytes/pro-spermatogonia or oogonia, respectively. Gonocytes differentiate into fetal spermatogonia during the embryonic period and then into spermatogonia at around birth. Spermatogonia form spermatogonial stem cells, and after puberty, they initiate spermatogenesis, generating spermatocytes, spermatids, and then sperm throughout the life. Oogonia differentiate into oocytes, initiate meiosis and form primordial follicles with granulosa cells during the embryonic period. After puberty, primordial follicles initiate their growth, developing into primary, secondary and antral follicles, which are ovulated for fertilization. A diagram representing the dynamics of genome-wide DNA methylation (5-methylcytosine: 5mC) levels during germ cell development is shown in the middle panel. Genome-wide 5mC levels decrease during pre-implantation development and increase around the peri-implantation period. The PGCs originally bear a high genome-wide 5mC level. During their migration period and in the genital ridges, PGCs undergo genome-wide DNA demethylation, which leads to erasure of parental imprints in both sexes and reactivation of X chromosomes in females. This is likely coupled with histone modification remodeling. During the sex differentiation period, male and female germ cells acquire sex-specific epigenetic programming, including parental imprints, the detail of which is currently unknown. The developmental processes of germ cells recapitulated by in vitro experiments 11 , 12 , 20 are also shown (bottom panel).
Full size image
To generate xrOvaries, hPGCLCs and mouse embryonic ovarian somatic cells are aggregated at a ratio of 5,000:50,000–75,000 and cultured for 2 d under a floating aggregate condition (Steps 75–77 of the main procedure). The resulting aggregates (xrOvaries) are transferred by a glass capillary to a Transwell culture system for an air–liquid interface culture (Steps 78–83 of the main procedure). The culture can be continued for up to ~5 months with half the medium changed every 3 d. hPGCLC-derived cells are considered to initiate genome-wide DNA demethylation along with their induction and the onset of xrOvary culture, and typically start the expression of genes characteristic for oogonia/gonocyte development (including DAZL) from ~35 d of culture 11 . hPGCLC-derived cells continue genome-wide DNA demethylation, which reaches a level of ~20% after ~10 weeks of xrOvary culture. At this point, they express high levels of oogonia/gonocyte genes, such as DAZL, DDX4, MAEL, PIWIL1 and SPATA22, and can be defined as oogonia/gonocytes 11 . Upon prolongation of the xrOvary culture, hPGCLC-derived oogonia further erase their DNA methylation, which reaches a level of ~10% after ~4 months of xrOvary culture, and further up-regulate genes for oogonia. At this time point, up to ~25% of hPGCLC-derived cells acquire an AG−VT+ state and initiate the expression of genes characteristic for RA-responsive FGCs, including STRA8, REC8, ANHX, SMC1B and ASB9 11 .
Applications of the protocol
Until fairly recently, the mechanism for human germ-cell development had been investigated only in histological studies and a limited number of ex vivo culture studies. Recent advances in single-cell genomics/epigenomics, however, have begun to reveal transcriptomic and epigenomic landscapes as well as their dynamics associated with human germ-cell development at an unprecedented resolution 10 , 25 , 26 , 27 . Alongside these efforts, methods for recapitulating the human germ-cell specification pathway in vitro using hPSCs as starting materials have been developed 12 , 20 , 21 , creating opportunities for addressing the molecular mechanisms underlying human germ-cell specification in culture.
Studies exploiting these new technologies have shown that key transcription factors, along with their hierarchy of actions for hPGCLC specification, are divergent from those in mouse PGC/mPGCLC specification 12 , 20 , 22 . Most notably, SOX17—which has long been known as a key regulator for endoderm specification, but not for PGC specification, in mouse studies—has been identified as one of the most upstream regulators for hPGCLC specification 12 . Conversely, T—which has long been known as a key transcription factor for mesoderm development and has been shown to up-regulate Blimp1 and Prdm14 and to be essential for mouse PGC/mPGCLC specification 28 —is dispensable for hPGCLC specification 22 . Such divergence in the mechanism for germ-cell specification between humans and mice has been partly validated by studies using cynomolgus monkeys and other mammalian models 29 , 30 . The hPGCLC specification system is expected to yield further insights into the mechanism of human germ-cell specification.
The mechanisms for epigenetic reprogramming and oogonia/gonocyte differentiation in humans are unknown and remain one of the key frontiers in germ cell biology. Such mechanisms would diverge in various ways from the mechanisms in mice, as in the case of PGC specification. The xrOvary system that we have developed offers new opportunities to uncover these fundamentally important, divergent mechanisms. For example, using the xrOvary system, one can study the functions of key candidate genes potentially involved in these processes with gain- and loss-of-function experiments, combined with single-cell genomics/epigenomics analyses. Moreover, given that RA-responsive FGCs are in the process of entering into the meiotic prophase, the xrOvary system, with further improvements, could serve as a platform to explore the mechanism of meiotic entry in human oocytes, and such investigations would lead to delineation of the mechanism underlying the generation of genetic diversity in the human female germ line.
Comparison with other methods
The generation of oogonia from hPGCLCs has been achieved only using our method involving xrOvaries (described here), but the induction of the hPGCLCs themselves has been achieved either by using iMeLCs (also described here) or by using hPSCs cultured with a cocktail of inhibitors for four kinases (mitogen-activated protein kinase/extracellular signal-regulated kinase, glycogen synthase kinase 3, p38 and c-Jun N-terminal kinase) known as the 4i hPSCs 12 , 31 . Originally, Gafni et al. reported that the 4i hPSCs show properties of naïve hPSCs 31 , but subsequently, Irie et al. found that the 4i hPSCs exhibit a primitive streak-like state, show a gene-expression property similar to iMeLCs and are induced directly into hPGCLCs using a method very similar to that for hPGCLC induction from iMeLCs 12 . The efficiency of hPGCLC induction from iMeLCs appears to be similar to that of hPGCLC induction from 4i hPSCs—both inductions have efficiencies in the range of 20–60%—and the hPGCLCs induced from iMeLCs or 4i hPSCs show similar gene-expression properties 12 , 20 . Thus, either method may be suitable for hPGCLC induction, although the method involving iMeLCs (described here) might be more convenient for many researchers, since it starts with hPSCs cultured under conventional conditions 11 , 13 , 20 , 21 , 22 .
The hPGCLC induction system from hiPSCs uses a defined condition, with the yield of hPGCLCs being up to ~106 cells per typical experiment. The system is compatible with any type of gain- and loss-of-gene-function experiments, including genome-wide CRISPR/CAS9 screening, and thus it is highly suitable for analyzing gene functions associated with hPGC specification. However, it should be noted that the hPGCLC state reflects hPGCs just after their specification and can be maintained only up to ~10 d under a floating aggregate condition 11 , 20 , 21 , 22 , 29 . It would therefore be difficult with the PGCLC induction system to examine the gene-mutation effects that may manifest later in PGC development.
The development of the xrOvary system has made it possible to analyze the functions of genes involved in processes such as hPGC survival, epigenetic reprogramming and oogonia/gonocyte differentiation. Before proceeding to such analyses, however, it is important to note that it takes 10–16 weeks for the epigenetic reprogramming and oogonia/gonocyte differentiation to occur. In other words, this is a time-consuming process. Second, we note that a majority of hPGCLCs fail to survive in xrOvaries. In the case of an xrOvary starting with ~5,000 hPGCLCs, only ~500 cells will survive, successfully undergo epigenetic reprogramming and differentiate into oogonia/gonocytes after ~10 weeks of culture 11 . It is therefore necessary to take into account that any investigation into the mechanisms of epigenetic reprogramming and oogonia/gonocyte differentiation will require expertise for analyzing small numbers of cells (e.g., ~5,000 cells for ~10 xrOvaries). Finally, because oogonia/gonocyte differentiation in xrOvaries would take place, at least to a certain extent, in an asynchronous fashion, single-cell analyses would be an ideal strategy. Clearly, there is a need for a more defined system (e.g., one that can induce epigenetic reprogramming and oogonia differentiation under defined conditions without the use of embryonic gonadal somatic cells) 18 , 32 . Such a system would further facilitate the mechanistic analysis of human germ-cell development.
It appears that use of the xrOvary system might not be an efficient approach for the further differentiation of hPGCLC-derived oogonia and RA-responsive FGCs into human primary oocytes with the progression of meiotic prophase I, or for the subsequent formation of primordial/primary follicles 11 . This is because the mouse embryonic ovarian somatic cells might not support such differentiation processes of human cells, particularly under suboptimal culture conditions. In the future, therefore, conditions for the differentiation of hPGCLC-derived oogonia/RA-responsive FGCs into human primary oocytes must be identified, and this will require the use of human fetal ovarian somatic cells to form human reconstituted ovaries. Such an endeavor will be key to achieving human gametogenesis in vitro.
The efficiency of hPGCLC induction varies substantially depending on the hiPSC/hESC lines used (from ~1,000 hPGCLCs per iMeLC aggregate), and this variation is reflected in the gene-expression states of the iMeLCs. The expression levels of EOMES, MIXL1 or T in the iMeLCs are positively correlated with hPGCLC induction efficiency, whereas the expression levels of CDH1, SOX3 or FGF2 are negatively correlated, suggesting that hiPSC/ESC lines have different properties that influence their responsivity to iMeLC induction 13 , 21 . Furthermore, hPGCLC induction from female hPSCs is somewhat inefficient and difficult 13 , 21 , presumably due to the effects of X chromosome erosion during female hPSC culture 33 , 34 . It is thus necessary to screen a number of hPSC cell lines for the induction efficiency of hPGCLCs and to use an appropriate cell line for the experiments of interest. The optimal hPGCLC induction efficiency of a given hPSC line should be determined by examining several iMeLC induction times (e.g., 42, 48 and 54 h) as well as the effect of adding an inhibitor of fibroblast growth factor receptor (FGFR) signaling during iMeLC induction 20 , 21 .
The hPGCLC induction efficiency also depends on the lot of knockout serum replacement (KSR) used for the hPGCLC induction, most likely due to lot-to-lot variation in KSR quality/components. It is therefore important to select an appropriate lot of KSR before performing the experiments of interest. Typically, we purchase small aliquots of several lots of KSR, compare their impacts on hPGCLC induction efficiency and secure a sufficient amount of an appropriate lot.
Up to the oogonia/gonocyte stage, female and male germ cells bear essentially identical properties, with the exception of the X-linked gene dosage 1 , 2 , 3 , and both male and female hiPSCs acquire an oogonia/gonocyte-like state in xrOvaries with appropriate gene expression and epigenetic reprogramming 11 . We therefore consider that either male or female hPSCs can be used for oogonia/gonocyte induction in xrOvaries, depending on the purpose of the experiments.
We typically use hiPSCs bearing the reporter alleles for hPGCLCs (BT+AG+), oogonia (AG+VT+/AG−VT+) and RA-responsive FGCs (AG−VT+) 11 . When such reporters are not readily available, cell-surface markers such as EpCAM and INTEGRINα6 can be used for isolating hPGCLCs 11 , 20 , 21 , 22 . In addition, markers such as tissue non-specific alkaline phosphatase, KIT and CD38, alone or in appropriate combinations, have been used successfully to isolate human oogonia/gonocytes from aborted embryos 10 , 25 , 26 , 27 , and we therefore assume that such markers could also be used to isolate hPGCLC-derived oogonia, although it might be difficult to specifically isolate RA-responsive FGCs.
For the isolation of embryonic ovarian somatic cells to generate xrOvaries, we use the ICR strain, mainly because it is easy to breed and constantly produces a relatively large number of embryos of a relatively larger size (~5 × 104 embryonic ovarian somatic cells per embryo at embryonic day (E) 12.5 (see also Anticipated results)). Whether embryonic ovarian somatic cells of other frequently used strains such as C57BL6 provide a similar function to support the differentiation of hPGCLCs into oogonia remains to be examined. To sort out mouse germ cells from ovarian somatic cells, we use MACS, rather than FACS. MACS can be applied to a larger number of cells, yields cells of interest more quickly and causes less damage to such cells than FACS. MACS results in a less pure cell population, but a minor contamination of mouse germ cells in xrOvaries does not impact the differentiation of hPGCLCs into oogonia.
See Limitations and Anticipated results for the approximate yields of hPGCLCs and oogonia in a typical experiment to appropriately estimate the requirements for a given experimental setup of interest (e.g., the numbers of iMeLC aggregates and xrOvaries to generate per experiment).
Mice: pregnant female ICR mice at 12.5 d.p.c. for collection of fetal ovarian somatic cells
All the animal experiments must be performed in accordance with the ethical guidelines of the institution. All animal experiments in this protocol were performed under the ethical guidelines of Kyoto University (approval no. MedKyo19001)
Use mice bred in a specific pathogen-free environment under appropriate breeding conditions. We use pregnant female mice of the ICR strain (8–10 weeks old and 30–40 g body weight at mating) bred under a condition defined by Japan SLC, Inc.
hiPSC lines with PRDM1 (BLIMP1)-tdTomato: TFAP2C-EGFP (BTAG; RRID: CVCL_UP50) or TFAP2C-EGFP: DDX4 (VASA)-tdTomato (AGVT; RRID: CVCL_YE15) reporter alleles. See Experimental design for selecting hPSC lines for oogonia induction
The experiments on the induction of human oogonia from hiPSCs must be performed in accordance with the guidelines of the institution. All experiments using hiPSCs in this protocol were approved by the Institutional Review Board of Kyoto University and were performed according to the guidelines of the Ministry of Education, Culture, Sports, Science, and Technology of Japan.
Glasgow minimal essential medium (GMEM; Invitrogen, cat. no. 11710-035)
Knockout serum replacement (KSR; Invitrogen, cat. no. 10828-028)
PGCLC induction efficiency may vary between KSR lots. An optimal lot should be selected in advance (see Experimental design).
l-glutamine, 200 mM (l-Glu; Invitrogen, cat. no. 25030-081)
MEM non-essential amino acid (NEAA; Invitrogen, cat. no. 11140-050)
Sodium pyruvate solution, 100 mM (PyrNa; Invitrogen, cat. no. 11360-070)
Penicillin-streptomycin (Pen/Strep; Invitrogen, cat. no. 15140-122)
2-Mercaptoethanol, 55 mM in Dulbecco’s phosphate-buffered saline (DPBS) (2-ME; Invitrogen, cat. no. 21985-023)
Hazardous. Avoid breathing fumes. Wear glasses, gloves and other appropriate protection and handle with care.
Dulbecco’s modified Eagle medium (DMEM; Invitrogen, cat. no. 10313-021)
Fetal bovine serum, non-USA origin (FBS; Sigma, cat. no. F7524)
HEPES, 1 M (Invitrogen, cat. no. 15630-080)
GlutaMax (Invitrogen, cat. no. 35050-061)
DMEM/F-12 (Invitrogen, cat. no. 11330-032)
Minimal Essential Medium Eagle with alpha modification, nucleosides and GlutaMAX (αMEM; Invitrogen, cat. no. 32571-036)
l-ascorbic acid (Sigma-Aldrich, cat. no. A4403)
Enzymes, growth factors and chemicals
TrypLE Select (Invitrogen, cat. no. 12563-011)
TrypLE Express (Invitrogen, cat. no. 12604-021)
DNase I (Sigma-Aldrich, cat. no. DN25)
Y-27632 (Tocris, cat. no. 1254)
Activin A, recombinant human (Peprotech, cat. no. 120-14)
CHIR99021 (Tocris, cat. no. 4423)
Toxic. Avoid contact and inhalation. Wear a mask and gloves and handle with care.
PD173074 (Stemgent, cat. no. 04-0008)
BMP4, recombinant human (R&D Systems, cat. no. 314-BP-010)
SCF, recombinant mouse (R&D Systems, cat. no. 455-MC-010), or recombinant human (R&D Systems, cat. no. 255-SC-001MG)
EGF, recombinant human (R&D Systems, cat. no. 236-EG-200)
Leukemia inhibitory factor, recombinant human, 10 µg/ml (LIF; Millipore, cat. no. LIF1010)
Antibodies and immunostaining reagents
Allophycocyanin (APC)-conjugated anti-human CD326 (EpCAM) antibody (BioLegend, cat. no. 324207; RRID: AB_756081)
Brilliant Violet 421-conjugated anti-human/mouse CD49f antibody (BioLegend, cat. no. 313623; RRID: AB_2562243)
Anti-SSEA-1 (CD15) MicroBeads, human and mouse (Miltenyi Biotec, cat. no. 130-094-530; RRID: AB_2814656)
Anti-CD31 MicroBeads, mouse (Miltenyi Biotec, cat. no. 130-097-418; RRID: 2814657)
Normal donkey serum (Jackson Immunoresearch Laboratories, cat. no. 017-000-121; RRID: AB_2337258)
Rat anti-GFP antibody (Nacalai Tesque, cat. no. 04404-84; RRID: AB_10013361)
Rabbit anti-DDX4 antibody (Abcam, cat. no. ab13840; RRID: AB_443012)
Goat anti-FOXL2 antibody (Novus, cat. no. NB-100-1277; RRID: AB_2106188)
Rabbit anti-SCP3 antibody (Abcam, cat. no. ab15093; RRID: AB_301639)
AlexaFluor 488-conjugated donkey anti-rat IgG (Life Technologies, cat. no. A21208; RRID: AB_2535794)
AlexaFluor 568-conjugated donkey anti-rabbit IgG (Life Technologies, cat. no. A10042; RRID: AB_2534017)
AlexaFluor 647-conjugated donkey anti-goat IgG (Life Technologies, cat. no. A21447; RRID: AB_2535864)
AlexaFluor 488-conjugated donkey anti-mouse IgG (Life Technologies, cat. no. A31571; RRID: AB_162542)
DAPI (Nacalai Tesque, cat. no. 11034-56)
VECTASHIELD (Vector Laboratories, cat. no. H-1000; RRID: AB_2336789)
Other reagents and chemicals
Phosphate-buffered saline without potassium chloride, pH 7.2 (PBS; Nacalai Tesque, cat. no. 11480-35)
Distilled water (Nacalai Tesque, cat. no. 06442-95)
iMatrix-511 (Nippi, cat. no. 892014)
0.5-mol/l EDTA solution, pH 8.0 (Nacalai Tesque, cat. no. 06894-14)
Human plasma fibronectin, 1 mg/ml (Millipore, cat. no. FC010)
Dimethyl sulfoxide (DMSO; Sigma-Aldrich, cat. no. D8418)
Trypsin-EDTA, 0.5% (wt/vol) (Invitrogen, cat. no. 15400-054)
BSA fraction V, 7.5% (wt/vol) solution (Invitrogen, cat. no. 15260-037)
BD FACSflow Sheath Fluid (BD Biosciences, cat. no. 342003)
CELLOTION (ZENOAQ, cat. no. CB051)
Trypan blue stain solution, 0.5% (wt/vol) (Nacalai Tesque, cat. no. 29853-34)
CELLBANKER 1 plus (ZENOAQ, cat. no. CB021)
Paraformaldehyde (Nacalai Tesque, cat. no. 26126-25)
OCT (Optimal Cutting Temperature) compound (Sakura Finetek, cat. no. 4583)
Sucrose (Nacalai Tesque, cat. no. 30404-45)
Polyoxyethylene (20) sorbitan monolaurate (equivalent to Tween-20; Wako, cat. no. 166-21115)
Optical microscope (Olympus, cat. no. CKX41 or equivalent)
Phase contrast microscope
Full size image
Dilute 9.6 µl/well of iMatrix-511 with 1.5 ml/well of PBS, pipette the solution into one well of a six-well culture plate and then incubate for ≥1 h at 37 °C. Just before plating the iPSCs, add 750 µl/well of iPSC passaging medium into the iMatrix-511–coated well, aspirate the solution and pipette 1.5 ml/well of iPSC passaging medium.
Dilute 10 µl/well of human plasma fibronectin with 600 µl/well of PBS, pipette the solution into one well of a 12-well plate and then incubate for ≥1 h at 37 °C. Add 300 µl/well of GK15 and aspirate the solution just before plating the iPSCs.
Full size table
Steps 1–13, maintenance of human iPSCs: ≥1 h for coating the plate and 30 min for passage
Steps 14–15, medium replacement during iPSC culture: ~10 min each time
Steps 16–24, iMeLC differentiation: ≥1 h for coating the culture plate, 30 min for treating cells and 2 d for incubation
Steps 25–35, PGCLC differentiation: 30 min for treating cells and 6 d for incubation
Steps 36–47, dissociation and FACS sorting of hPGCLCs: 30 min for dissociating cells and 1 h for cell sorting
Steps 48–74, collection of fetal ovarian somatic cells from mouse female embryos: 2–4 h
Steps 75–77, generation and culture of xrOvaries: 30 min
Steps 78–83, culture of xrOvaries on culture insert membranes: 30 min
Steps 84–95, dissociation of xrOvaries and FACS sorting of human oogonia: 30 min for dissociating xrOvaries and 30 min for cell sorting
Steps 96–113, immunostaining of xrOvaries to determine the efficiency of oogonia differentiation: 2–3 d
hiPSCs cultured under the feeder-free condition form dense, round steric colonies (Fig. 3 ). When dissociating hiPSCs from an iMatrix-coated plate, it may be necessary to use a scraper to detach the cells, since hiPSCs adhere strongly to the laminin-coated plate (Step 8). The iMeLCs show a flat, epithelial morphology with distinct cell-to-cell boundaries (Fig. 3 ). Approximately 2 × 105 hiPSCs yield 1.5–3 × 105 iMeLCs during the 2 d of the differentiation culture. When iMeLCs are induced into hPGCLCs under a floating aggregate condition, a small core-like structure emerges 1 d after the aggregate formation, and this structure becomes firmer each day. When BTAG or AGVT reporter hiPSC-derived iMeLCs are used for hPGCLC differentiation, intense AG and BT signals become apparent from around culture d4 onward under a fluorescence dissection microscope. Depending on the hiPSC lines used, at d6 of hPGCLC induction, 20–50% of all living cells in the aggregates show AG (and BT) positivity in a FACS analysis, and 1,000–3,000 reporter positive cells can be obtained per aggregate 11 .
Approximately 5 × 104 fetal ovarian somatic cells can be isolated from one female embryo at E12.5. Typically, per experiment, we use 10 pregnant ICR female mice for ovarian somatic cell isolation and cryopreservation. Depending on the number of embryos and their sex ratio, 10 pregnant females usually yield 2–3 × 106 fetal ovarian somatic cells, which allow the formation of ~40 xrOvaries.
For the generation of xrOvaries, hPGCLCs and embryonic ovarian somatic cells are subjected to a 2-d floating culture for their reaggregation, and then to a gas–liquid interface culture. Prolonged continuation of the floating culture often results in an apparently unhealthy appearance of xrOvaries and poor survival of hPGCLC-derived cells 11 . During the gas–liquid interface culture, hPGCLC-derived cells are well maintained in xrOvaries until around ag21, with 1,000–5,000 reporter-positive cells surviving per one xrOvary. However, the number of hPGCLC-derived cells in xrOvaries decreases with the extension of the culture period, with 100–1,000 reporter-positive cells per xrOvary after ag35. The proportion of the reporter-positive cells in an xrOvary is somewhat variable, ranging from 0.1% to 5% of all living cells of the xrOvary. When performing this protocol using the AGVT reporter hiPSC line, the properties of the AG+ (but VT−) hPGCLCs gradually change, and the hPGCLCs reach the AG+VT+ state (Fig. 8 )—which represents an oogonia state—at around ag77. When the culture is continued until around ag120, some hPGCLC-derived cells acquire an AG− but VT+ state (Fig. 8 ), which represents an RA-responsive FGC state 11 . Under the conditions reported here, the gas–liquid interface culture of xrOvaries can be extended up to ag200, but the differentiation of hPGCLC-derived cells does not progress beyond the RA-responsive FGC state. At around this time point, mouse ovarian somatic cells begin to show a deteriorating phenotype, and the structure of xrOvaries cannot be properly maintained.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
The data generated or analyzed during this study are included in this published article.
We thank the members of our laboratory for their helpful input on this study. We are grateful to Y. Nagai, N. Konishi, E. Tsutsusmi and M. Kawasaki for their technical assistance. This work was supported by a Grant-in-Aid for Specially Promoted Research from JSPS (17H06098), a JST-ERATO Grant (JPMJER1104), a grant from HFSP (RGP0057/2018) and grants from the Pythias Fund and Open Philanthropy Project to M.S.
Present address: Department of Biomedical Sciences, University of Pennsylvania School of Veterinary Medicine, Department of Pathology and Laboratory Medicine, University of Pennsylvania Perelman School of Medicine, Philadelphia, PA, USA