Advertisement
Cell Stem Cell
This journal offers authors two options (open access or subscription) to publish research

Robust In Vitro Induction of Human Germ Cell Fate from Pluripotent Stem Cells

  • Author Footnotes
    12 Co-first author
    Kotaro Sasaki
    Footnotes
    12 Co-first author
    Affiliations
    Department of Anatomy and Cell Biology, Graduate School of Medicine, Kyoto University, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan

    JST, ERATO, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan
    Search for articles by this author
  • Author Footnotes
    12 Co-first author
    Shihori Yokobayashi
    Footnotes
    12 Co-first author
    Affiliations
    Department of Anatomy and Cell Biology, Graduate School of Medicine, Kyoto University, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan

    JST, ERATO, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan

    Center for iPS Cell Research and Application (CiRA), Kyoto University, 53 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606-8507, Japan
    Search for articles by this author
  • Tomonori Nakamura
    Affiliations
    Department of Anatomy and Cell Biology, Graduate School of Medicine, Kyoto University, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan

    JST, ERATO, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan
    Search for articles by this author
  • Ikuhiro Okamoto
    Affiliations
    Department of Anatomy and Cell Biology, Graduate School of Medicine, Kyoto University, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan

    JST, ERATO, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan
    Search for articles by this author
  • Yukihiro Yabuta
    Affiliations
    Department of Anatomy and Cell Biology, Graduate School of Medicine, Kyoto University, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan

    JST, ERATO, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan
    Search for articles by this author
  • Kazuki Kurimoto
    Affiliations
    Department of Anatomy and Cell Biology, Graduate School of Medicine, Kyoto University, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan

    JST, ERATO, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan
    Search for articles by this author
  • Hiroshi Ohta
    Affiliations
    Department of Anatomy and Cell Biology, Graduate School of Medicine, Kyoto University, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan

    JST, ERATO, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan
    Search for articles by this author
  • Yoshinobu Moritoki
    Affiliations
    Department of Anatomy and Cell Biology, Graduate School of Medicine, Kyoto University, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan

    JST, ERATO, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan

    Department of Nephro-Urology, Graduate School of Medical Sciences, Nagoya City University, Kawasumi, Mizuho-cho, Mizuho-ku, Nagoya 467-8601, Japan
    Search for articles by this author
  • Chizuru Iwatani
    Affiliations
    Research Center for Animal Life Science, Shiga University of Medical Science, Seta-Tsukinowa-cho, Otsu, Shiga 520-2192, Japan
    Search for articles by this author
  • Hideaki Tsuchiya
    Affiliations
    Research Center for Animal Life Science, Shiga University of Medical Science, Seta-Tsukinowa-cho, Otsu, Shiga 520-2192, Japan
    Search for articles by this author
  • Shinichiro Nakamura
    Affiliations
    Research Center for Animal Life Science, Shiga University of Medical Science, Seta-Tsukinowa-cho, Otsu, Shiga 520-2192, Japan
    Search for articles by this author
  • Kiyotoshi Sekiguchi
    Affiliations
    Institute for Protein Research, Osaka University, Osaka 565-0871, Japan
    Search for articles by this author
  • Tetsushi Sakuma
    Affiliations
    Department of Mathematical and Life Sciences, Graduate School of Science, Hiroshima University, 1-3-1 Kagamiyama, Higashi-Hiroshima, Hiroshima 739-8526, Japan
    Search for articles by this author
  • Takashi Yamamoto
    Affiliations
    Department of Mathematical and Life Sciences, Graduate School of Science, Hiroshima University, 1-3-1 Kagamiyama, Higashi-Hiroshima, Hiroshima 739-8526, Japan
    Search for articles by this author
  • Takahide Mori
    Affiliations
    Academia for Repro-Regenerative Medicine, 394-1 Higashi-Hinodono-cho, Ichijo-Shinmachi-Higashiiru, Kamigyo-ku, Kyoto 602-0917, Japan
    Search for articles by this author
  • Knut Woltjen
    Affiliations
    Center for iPS Cell Research and Application (CiRA), Kyoto University, 53 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606-8507, Japan

    Hakubi Center for Advanced Research, Kyoto University, Yoshida-Ushinomiya-cho, Sakyo-ku, Kyoto 606-8501, Japan
    Search for articles by this author
  • Masato Nakagawa
    Affiliations
    Center for iPS Cell Research and Application (CiRA), Kyoto University, 53 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606-8507, Japan
    Search for articles by this author
  • Takuya Yamamoto
    Affiliations
    Center for iPS Cell Research and Application (CiRA), Kyoto University, 53 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606-8507, Japan

    Institute for Integrated Cell-Material Sciences (iCeMS), Kyoto University, Yoshida-Ushinomiya-cho, Sakyo-ku, Kyoto 606-8501, Japan

    JST, CREST, 4-1-8 Honcho, Kawaguchi, Saitama, 332-0012, Japan
    Search for articles by this author
  • Kazutoshi Takahashi
    Affiliations
    Center for iPS Cell Research and Application (CiRA), Kyoto University, 53 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606-8507, Japan
    Search for articles by this author
  • Shinya Yamanaka
    Affiliations
    Center for iPS Cell Research and Application (CiRA), Kyoto University, 53 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606-8507, Japan
    Search for articles by this author
  • Mitinori Saitou
    Correspondence
    Corresponding author
    Affiliations
    Department of Anatomy and Cell Biology, Graduate School of Medicine, Kyoto University, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan

    JST, ERATO, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan

    Center for iPS Cell Research and Application (CiRA), Kyoto University, 53 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606-8507, Japan

    Institute for Integrated Cell-Material Sciences (iCeMS), Kyoto University, Yoshida-Ushinomiya-cho, Sakyo-ku, Kyoto 606-8501, Japan
    Search for articles by this author
  • Author Footnotes
    12 Co-first author
Open ArchivePublished:July 16, 2015DOI:https://doi.org/10.1016/j.stem.2015.06.014

      Highlights

      • Robust induction of hPGCLCs from primed hiPSCs occurs via incipient mesoderm-like cells
      • EpCAM and INTEGRINα6 are identified as markers for hPGCLC purification
      • hPGCLCs avoid activation of a somatic program and undergo epigenetic reprogramming
      • BLIMP1 stabilizes germline transcription and represses neuronal differentiation

      Summary

      Mechanisms underlying human germ cell development are unclear, partly due to difficulties in studying human embryos and lack of suitable experimental systems. Here, we show that human induced pluripotent stem cells (hiPSCs) differentiate into incipient mesoderm-like cells (iMeLCs), which robustly generate human primordial germ cell-like cells (hPGCLCs) that can be purified using the surface markers EpCAM and INTEGRINα6. The transcriptomes of hPGCLCs and primordial germ cells (PGCs) isolated from non-human primates are similar, and although specification of hPGCLCs and mouse PGCs rely on similar signaling pathways, hPGCLC specification transcriptionally activates germline fate without transiently inducing eminent somatic programs. This includes genes important for naive pluripotency and repression of key epigenetic modifiers, concomitant with epigenetic reprogramming. Accordingly, BLIMP1, which represses somatic programs in mice, activates and stabilizes a germline transcriptional circuit and represses a default neuronal differentiation program. Together, these findings provide a foundation for understanding and reconstituting human germ cell development in vitro.

      Graphical Abstract

      Introduction

      The germ cell lineage is a foundation for totipotency, perpetuating genetic as well as epigenetic information across generations in most multicellular systems. Accordingly, in humans, anomalies in the germline, which arises from primordial germ cells (PGCs) and forms either spermatozoa or oocytes through complex developmental pathways, lead to a variety of critical conditions, including infertility, impaired development/physiology, and a diverse array of genetic/epigenetic disorders within offspring. A precise understanding of the mechanism for germ cell development therefore bears significant implications not only in biology in general, but also in a broad range of human diseases.
      The mechanism for germ cell development in mammals has most extensively been studied using the mouse as a model organism, providing essential information practically applicable to all mammals, including humans. On the other hand, among diverse mammalian species, there exist significant differences in the precise mechanisms for germ cell development, which necessitates careful species-by-species studies for a precise understanding of germ cell development in a given species. In this regard, there is a critical lack of information as to the mechanism for germ cell development in humans, mainly due to the difficulties/limitations in accessing relevant experimental materials. This has been an impediment for the diagnosis/treatment of disorders arising from defects in human germ cells. It would therefore represent a promising breakthrough over these limitations if human germ cell development could be reconstituted in vitro from human pluripotent stem cells (hPSCs), including human embryonic stem cells (hESCs) (
      • Thomson J.A.
      • Itskovitz-Eldor J.
      • Shapiro S.S.
      • Waknitz M.A.
      • Swiergiel J.J.
      • Marshall V.S.
      • Jones J.M.
      Embryonic stem cell lines derived from human blastocysts.
      ) and human induced pluripotent stem cells (hiPSCs) (
      • Takahashi K.
      • Tanabe K.
      • Ohnuki M.
      • Narita M.
      • Ichisaka T.
      • Tomoda K.
      • Yamanaka S.
      Induction of pluripotent stem cells from adult human fibroblasts by defined factors.
      ).
      Accordingly, recent studies have demonstrated the reconstitution in vitro of the specification and development of the mouse germline by PSCs (
      • Hayashi K.
      • Ohta H.
      • Kurimoto K.
      • Aramaki S.
      • Saitou M.
      Reconstitution of the mouse germ cell specification pathway in culture by pluripotent stem cells.
      ,
      • Hayashi K.
      • Ogushi S.
      • Kurimoto K.
      • Shimamoto S.
      • Ohta H.
      • Saitou M.
      Offspring from oocytes derived from in vitro primordial germ cell-like cells in mice.
      ): mouse (m) ESCs/iPSCs with ground state pluripotency (
      • Ying Q.L.
      • Wray J.
      • Nichols J.
      • Batlle-Morera L.
      • Doble B.
      • Woodgett J.
      • Cohen P.
      • Smith A.
      The ground state of embryonic stem cell self-renewal.
      ) are induced into pre-gastrulation epiblast-like cells (EpiLCs), which are in turn induced into PGC-like cells (PGCLCs) with robust capacity for both spermatogenesis and oogenesis, and for the generation of offspring; these findings suggest a conceptual framework and feasibility for the reconstitution of human germ cell development in vitro. However, hESCs/iPSCs have been shown to have differentiation potential and other properties distinct from mESCs/iPSCs and bear a primed pluripotency with similarity to mouse epiblast stem cells (EpiSCs) (
      • Brons I.G.
      • Smithers L.E.
      • Trotter M.W.
      • Rugg-Gunn P.
      • Sun B.
      • Chuva de Sousa Lopes S.M.
      • Howlett S.K.
      • Clarkson A.
      • Ahrlund-Richter L.
      • Pedersen R.A.
      • Vallier L.
      Derivation of pluripotent epiblast stem cells from mammalian embryos.
      ,
      • Tesar P.J.
      • Chenoweth J.G.
      • Brook F.A.
      • Davies T.J.
      • Evans E.P.
      • Mack D.L.
      • Gardner R.L.
      • McKay R.D.
      New cell lines from mouse epiblast share defining features with human embryonic stem cells.
      ), which resemble post-gastrulation epiblasts and bear a limited, if any, potential for the germ cell fate (
      • Hayashi K.
      • Ohta H.
      • Kurimoto K.
      • Aramaki S.
      • Saitou M.
      Reconstitution of the mouse germ cell specification pathway in culture by pluripotent stem cells.
      ). It therefore remains unknown whether hESCs/iPSCs can be efficiently induced into a human germ cell fate through an appropriate developmental pathway, although a number of reports have shown that random differentiation of hESCs/iPSCs yields germ cell-like cells at low efficiency (reviewed by
      • Hayashi Y.
      • Saitou M.
      • Yamanaka S.
      Germline development from human pluripotent stem cells toward disease modeling of infertility.
      ).
      We here explored the possibility of inducing the germ cell fate from hiPSCs under a defined condition. We observed a robust induction, from multiple hiPSC lines with primed pluripotency, of a cellular state similar to hPGCs, which we designated hPGCLCs, based on a number of stringent criteria.

      Results

      Establishment of hiPSCs Bearing Double Germline Reporters

      We set out to establish hiPSC lines bearing reporters that mark hPGC specification. We selected BLIMP1/PRDM1 and TFAP2C/AP2γ as candidates that show expression upon hPGC specification, since Blimp1 and Tfap2c encode transcription factors (TFs) necessary and sufficient for mouse PGC specification (
      • Nakaki F.
      • Hayashi K.
      • Ohta H.
      • Kurimoto K.
      • Yabuta Y.
      • Saitou M.
      Induction of mouse germ-cell fate by transcription factors in vitro.
      ,
      • Ohinata Y.
      • Payer B.
      • O’Carroll D.
      • Ancelin K.
      • Ono Y.
      • Sano M.
      • Barton S.C.
      • Obukhanych T.
      • Nussenzweig M.
      • Tarakhovsky A.
      • et al.
      Blimp1 is a critical determinant of the germ cell lineage in mice.
      ,
      • Weber S.
      • Eckert D.
      • Nettersheim D.
      • Gillis A.J.
      • Schäfer S.
      • Kuckenberg P.
      • Ehlermann J.
      • Werling U.
      • Biermann K.
      • Looijenga L.H.
      • Schorle H.
      Critical function of AP-2 gamma/TCFAP2C in mouse embryonic germ cell maintenance.
      ), and since BLIMP1 and TFAP2C have been reported to be expressed in human fetal germ cells (
      • Eckert D.
      • Biermann K.
      • Nettersheim D.
      • Gillis A.J.
      • Steger K.
      • Jäck H.M.
      • Müller A.M.
      • Looijenga L.H.
      • Schorle H.
      Expression of BLIMP1/PRMT5 and concurrent histone H2A/H4 arginine 3 dimethylation in fetal germ cells, CIS/IGCNU and germ cell tumors.
      ,
      • Pauls K.
      • Jager R.
      • Weber S.
      • Wardelmann E.
      • Koch A.
      • Buttner R.
      • Schorle H.
      Transcription factor AP-2gamma, a novel marker of gonocytes and seminomatous germ cell tumors.
      ).
      We cultured two independent lines of male hiPSCs, 585A1 and 585B1, derived from peripheral mononuclear blood cells (PMBCs) (
      • Okita K.
      • Yamakawa T.
      • Matsumura Y.
      • Sato Y.
      • Amano N.
      • Watanabe A.
      • Goshima N.
      • Yamanaka S.
      An efficient nonviral method to generate integration-free human-induced pluripotent stem cells from cord blood and peripheral blood cells.
      ), under a feeder-free, defined condition with basic fibroblast growth factor (bFGF) on the E8 fragment of recombinant LAMININ511 (rLN511E8) (
      • Nakagawa M.
      • Taniguchi Y.
      • Senda S.
      • Takizawa N.
      • Ichisaka T.
      • Asano K.
      • Morizane A.
      • Doi D.
      • Takahashi J.
      • Nishizawa M.
      • et al.
      A novel efficient feeder-free culture system for the derivation of human induced pluripotent stem cells.
      ). The hiPSCs cultured under this condition allow single-cell passage, exhibit a homogeneous property, and show gene expression of a primed pluripotent state (
      • Nakamura T.
      • Yabuta Y.
      • Okamoto I.
      • Aramaki S.
      • Yokobayashi S.
      • Kurimoto K.
      • Sekiguchi K.
      • Nakagawa M.
      • Yamamoto T.
      • Saitou M.
      SC3-seq: a method for highly parallel and quantitative measurement of single-cell gene expression.
      ). Using a TALEN (transcription activator-like effector nucleases)-based strategy (
      • Sakuma T.
      • Hosoi S.
      • Woltjen K.
      • Suzuki K.
      • Kashiwagi K.
      • Wada H.
      • Ochiai H.
      • Miyamoto T.
      • Kawai N.
      • Sasakura Y.
      • et al.
      Efficient TALEN construction and evaluation methods for human cell and animal applications.
      ), we isolated several homologous recombinants bearing both the BLIMP1-2A-tdTomato (BT) and the TFAP2C-2A-EGFP (AG) alleles, and we selected one line, BTAG 585B1-868, which bears both recombinant alleles in a heterozygous fashion and a normal karyotype, for subsequent studies (Figure 1A, Figures S1A–S1D). The BTAG alleles activated tdTomato and EGFP upon BLIMP1 and TFAP2C expression, respectively (Figure S1E).
      Figure thumbnail gr1
      Figure 1Induction of BTAG(+) Cells Directly from hiPSCs
      (A) (Left) A phase-contrast image of BTAG 585B1-868 hiPSCs. Bar, 200 μm. (Right) FACS analysis for OCT3/4, SOX2, NANOG, TRA-1-60, and SSEA-4 expression in BTAG 585B1-868 hiPSCs.
      (B) Scheme for direct induction of BTAG(+) cells from hiPSCs.
      (C) Bright field (BF) and fluorescence (AG and BT) images of floating aggregates (3,000 cells/initial aggregate) of hiPSCs stimulated by BMP4, SCF, EGF, and LIF (left) or by no cytokines (right) for 2, 4, 6, and 8 days. Bar, 200 μm.
      (D) FACS analysis of BTAG expression during direct induction from hiPSCs by BMP4, SCF, EGF, and LIF (top) or by no cytokines (bottom) for 8 days. Boxed areas indicate BTAG(+) cells with their percentages.
      (E) Box plots for the numbers of BTAG(+) cells per aggregate. The average (horizontal line), 25th and 75th percentiles (box), and the maximum and minimum (error bars) of at least two independent experiments are shown.
      (F) Gene expression dynamics during BTAG(+) cell induction at days 0 (hiPSCs), 2 (whole aggregates), and 6 [BTAG(+) cells], as measured by qPCR. For each gene examined, the ΔCT from the average CT values of the two independent housekeeping genes Arbp and Ppia (set as 0) were calculated and plotted. The copy numbers of Arbp and Ppia were estimated (indicated in red) by the qPCR values of the spike-in RNAs. For each point, the average value from two independent experiments is shown on the log2 scale, with SDs. n.d., not detected.
      See also .

      Direct Induction of BTAG-Positive (+) Cells from hiPSCs

      Given that hiPSCs bear a primed pluripotency, we first explored whether the hiPSCs can be induced directly into hPGCLCs under a condition that induces EpiLCs into mPGCLCs (
      • Hayashi K.
      • Ohta H.
      • Kurimoto K.
      • Aramaki S.
      • Saitou M.
      Reconstitution of the mouse germ cell specification pathway in culture by pluripotent stem cells.
      ). The BTAG hiPSCs were dissociated into single cells and were cultured under a floating condition in GMEM + 15% knockout serum replacement (KSR) (GK15) with bone morphogenetic protein 4 (BMP4), stem cell factor (SCF), leukemia inhibitory factor (LIF), and epidermal growth factor (EGF) (3,000 cells/aggregate) in the presence of a ROCK inhibitor (Figure 1B) (
      • Watanabe K.
      • Ueno M.
      • Kamiya D.
      • Nishiyama A.
      • Matsumura M.
      • Wataya T.
      • Takahashi J.B.
      • Nishikawa S.
      • Nishikawa S.
      • Muguruma K.
      • Sasai Y.
      A ROCK inhibitor permits survival of dissociated human embryonic stem cells.
      ).
      The hiPSC aggregates did not show BT or AG expression at day 2 of stimulation by inspection under a fluorescent dissection microscope, and they appeared somewhat fragile (Figure 1C). However, notably, at day 4, some cells initiated BTAG expression, and at day 6, a distinct population of cells exhibited strong BTAG, which persisted at least until day 8 (Figure 1C). Consistently, fluorescence activated cell sorting (FACS) analyses revealed that at day 2, the entire aggregates shifted toward a state weakly positive for BTAG, and at day 4, ∼15% of the cells became BTAG(+) with BT and AG being upregulated at a similar kinetics, and at day 6–8, there appeared a distinct population of BTAG(+) cells (∼20%) (Figure 1D). The efficiency for the induction of the BTAG(+) cells was the highest when we started induction with 3,000 cells/aggregate and used more than 10% of KSR as the basal medium (Figure S1F), and the average number of BTAG(+) cells per aggregate at day 6 was around 200 (Figure 1E).
      We examined gene expression dynamics during the BTAG(+) cell induction by qPCR. The hiPSCs expressed high/middle levels of the key pluripotency genes POU5F1, NANOG, and SOX2 (∼1,000, ∼40, and ∼150 copies per cell, respectively), whereas they showed low/no expression of genes associated with naive pluripotency in mice (they did, however, express ZFP42 and PRDM14 at a relatively high level ([∼20–40 copies)) (Figure 1F). They showed no/very low expression of genes associated with PGCs, neuroectoderm, mesoderm, and endoderm (Figure 1F). At day 2 of stimulation, the aggregates upregulated key pluripotency genes, remained low for naive pluripotency genes, and initiated upregulation of genes for PGCs (BLIMP1 and TFAP2C, ∼20 and ∼150 copies, respectively), mesoderm (T, EOMES, SP5, and NODAL, ∼10–40 copies), and endoderm (GATA4 and SOX17, ∼a few copies) (Figure 1F). The BTAG(+) cells at day 6 highly upregulated POU5F1 and NANOG (∼4,000 and ∼600 copies, respectively), but became negative for SOX2 (Figure 1F), consistent with the observation that hPGCs lack SOX2 expression (
      • de Jong J.
      • Stoop H.
      • Gillis A.J.
      • van Gurp R.J.
      • van de Geijn G.J.
      • Boer Md.
      • Hersmus R.
      • Saunders P.T.
      • Anderson R.A.
      • Oosterhuis J.W.
      • Looijenga L.H.
      Differential expression of SOX17 and SOX2 in germ cells and stem cells has biological and clinical implications.
      ,
      • Perrett R.M.
      • Turnpenny L.
      • Eckert J.J.
      • O’Shea M.
      • Sonne S.B.
      • Cameron I.T.
      • Wilson D.I.
      • Rajpert-De Meyts E.
      • Hanley N.A.
      The early human germ cell lineage does not express SOX2 during in vivo development or upon in vitro culture.
      ). They exhibited high levels of early PGC markers such as BLIMP1, TFAP2C, and NANOS3 (∼150, ∼600, and ∼600 copies, respectively), as well as SOX17 (∼300 copies), but remained low/negative for DPPA3 (∼a few copies) and late PGC genes (DAZL and DDX4) (Figure 1F). They showed only modest/low levels of T and PRDM14 (∼50 and ∼10 copies, respectively), which are essential for PGC specification in mice (
      • Aramaki S.
      • Hayashi K.
      • Kurimoto K.
      • Ohta H.
      • Yabuta Y.
      • Iwanari H.
      • Mochizuki Y.
      • Hamakubo T.
      • Kato Y.
      • Shirahige K.
      • Saitou M.
      A mesodermal factor, T, specifies mouse germ cell fate by directly activating germline determinants.
      ,
      • Yamaji M.
      • Seki Y.
      • Kurimoto K.
      • Yabuta Y.
      • Yuasa M.
      • Shigeta M.
      • Yamanaka K.
      • Ohinata Y.
      • Saitou M.
      Critical function of Prdm14 for the establishment of the germ cell lineage in mice.
      ) (Figure 1F). Additionally, they upregulated genes for naive pluripotency (KLF4, TCL1B, and TFCP2L1,with ∼100, ∼100, ∼40 copies, respectively), mesoderm (EVX1 and MSX2, ∼70 and ∼150 copies, respectively), and endoderm (GATA4, ∼70 copies) (Figure 1F). These findings suggest that hPSCs bear the competence for the germ cell fate and the BTAG(+) cells correspond to early hPGCs.

      Robust Induction of BTAG(+) Cells via an Incipient Mesoderm-like State

      We noted that the aggregates directly induced from hiPSCs included a significant number of dead or dying cells, resulting in a relatively small yield of the BTAG(+) cells (Figure 1E). We reasoned that this induction strategy might not be optimal, so we explored a condition under which to induce hiPSCs into a more appropriate precursor. We found that by stimulating hiPSCs with Activin A (ACTA) and a WNT signaling agonist (CHIR99021 [CHIR];
      • Ying Q.L.
      • Wray J.
      • Nichols J.
      • Batlle-Morera L.
      • Doble B.
      • Woodgett J.
      • Cohen P.
      • Smith A.
      The ground state of embryonic stem cell self-renewal.
      ) in GK15 for around 2 days on a fibronectin-coated plate, hiPSCs were induced into flat epithelial cells with distinct cell-to-cell boundaries (Figures 2A–2C), which, upon floating aggregate formation and cytokine stimulation, activated BTAG efficiently as early as day 2 (∼30%–40%) and yielded BTAG(+) cells robustly (as much as ∼60% at day 4, from around 400 to 1,000 cells per aggregate at days 4–6) with lower occurrence of cell death (Figures 2D–2F and Figure S2A).
      Figure thumbnail gr2
      Figure 2Induction of BTAG(+) Cells from hiPSCs through iMeLCs
      (A) Scheme for BTAG(+) cell induction through iMeLCs.
      (B) A phase-contrast image of BTAG hiPSCs (top) and iMeLCs (bottom). Bar, 200 μm.
      (C) FACS analysis for OCT3/4, SOX2, and NANOG expression in iMeLCs (42 hr induction).
      (D) Bright field (BF) and fluorescence (AG and BT) images of floating aggregates (3,000 cells/initial aggregate) of iMeLCs stimulated by BMP4, SCF, EGF, and LIF (left) or by no cytokines (right) for 2, 4, 6, and 8 days. Bar, 200 μm.
      (E) FACS analysis of BTAG induction in aggregates of iMeLCs stimulated by BMP4, SCF, EGF, and LIF (top) or by no cytokines (bottom) for 8 days. Boxed areas indicate BTAG(+) cells with their percentages.
      (F) Box plots (as shown in E) demonstrating the percentages (left) and the numbers of BTAG(+) cells per aggregate (right).
      (G) Gene expression dynamics during BTAG(+) cell induction for 8 days, as measured by qPCR. The quantifications of gene expression levels (three independent experiments) were as shown in F. Filled circles: values for BTAG(+) cells; filled squares: values for BTAG(−) cells. n.d., not detected.
      (H) FACS analysis for OCT3/4, SOX2, and NANOG expression in d4 BTAG(+) cells.
      (I) IF of BLIMP1, TFAP2C, SOX2, and SOX17 expression in d8 BTAG(+) PGCLCs [EGFP(+), dotted lines] compared with those in hiPSCs. Bar, 10 μm.
      See also .
      The flat epithelial cells expressed genes for (naive) pluripotency at levels similar to hiPSCs, did not show genes for PGCs or endoderm, and modestly upregulated genes for mesoderm (T, EOMES, SP5, and MIXL1, with ∼a few, ∼30, ∼100, and ∼50 copies, respectively) (Figures 2C and 2G), indicating that they represent incipient mesoderm/primitive streak-like cells (which we designated as iMeLCs). The gene expression profiles of the BTAG(+) cells induced from iMeLCs were essentially identical to those of the BTAG(+) cells induced directly from hiPSCs (Figure 2G). FACS and immunofluorescence (IF) analyses confirmed the expression of OCT4, NANOG, BLIMP1, TFAP2C, and SOX17 and the repression of SOX2 in BTAG(+) cells (Figures 2H and 2I). These findings demonstrate that hiPSCs can first be induced into iMeLCs with upregulation of nascent mesodermal genes, which, in turn, are induced robustly and with a faster kinetics into the BTAG(+) cells, a state potentially similar to that of early hPGCs. The variation of induction efficiency would presumably depend on the conditions of hiPSCs/iMeLCs, including their passage numbers, (un)differentiated states, and viability upon passage/aggregate formation.
      We examined the relevant conditions/signaling pathways for BTAG(+) cell induction through iMeLCs. We found that the duration of stimulation with ACTA and CHIR is critical for hiPSCs to acquire a capacitated iMeL state, with a time of approximately 42–48 hr being optimal for the BTAG 585B1-868 hiPSCs (Figure S2B). Longer stimulation resulted in further upregulation of mesodermal/endodermal properties and depleted the capacity for BTAG(+) cell induction (Figures S2B and S2C). Both ACTA and CHIR/WNT3A were essential, but both BMP4 and bFGF were detrimental, for the iMeLC induction (Figures S3A–S3F). Notably, inhibition of FGF receptor (FGFR) signaling by a specific inhibitor (FGFRi, PD173074) during iMeLC induction led to more robust proliferation/survival of cells in the aggregates, including the BTAG(+) cells bearing appropriate gene expression, resulting in the generation of higher numbers (but not higher percentages) of the BTAG(+) cells per aggregate (Figures S3G and S3H). For BTAG(+) cell induction, BMP4 signaling through activin receptor-like kinase 2/3 (ALK2/3) was essential (Figures S4A and S4B). BMP2, but not BMP7 or BMP8A, replaced the role of BMP4 at an essentially identical concentration (Figure S4C). SCF, LIF, and EGF played additive roles for the maintenance, but not induction, of the BTAG(+) cells (Figure S4D). Thus, induction, proliferation, and survival of BTAG(+) cells involve signaling molecules similar to those for induction, proliferation, and survival of mPGCLCs (
      • Hayashi K.
      • Ohta H.
      • Kurimoto K.
      • Aramaki S.
      • Saitou M.
      Reconstitution of the mouse germ cell specification pathway in culture by pluripotent stem cells.
      ).

      BTAG(+) Cells as hPGCLCs

      We determined global transcription profiles of key cell types during BTAG(+) cell and mPGCLC induction and those of the gonadal PGCs of a non-human primate (a cynomolgus monkey; Macaca fascicularis) by using a highly quantitative RNA-sequencing (RNA-seq) technology (
      • Nakamura T.
      • Yabuta Y.
      • Okamoto I.
      • Aramaki S.
      • Yokobayashi S.
      • Kurimoto K.
      • Sekiguchi K.
      • Nakagawa M.
      • Yamamoto T.
      • Saitou M.
      SC3-seq: a method for highly parallel and quantitative measurement of single-cell gene expression.
      ). For cynomolgus monkey (cy) PGCs, we isolated cy embryos at E43, 50, and 51 (corresponding roughly to around E10.5–13.5 in mice), dissected out the gonads, generated single-cell cDNAs originated from PGCs [POU5F1(+)] (Figures S5A and S5B), and analyzed them by the RNA-seq (Figure S5E).

      BTAG(+) Cells Show a Transcription Profile Similar to That of cyPGCs

      Unsupervised hierarchical clustering (UHC) classified the cells during BTAG(+) cell induction into two large clusters, one with hiPSCs and iMeLCs, and the other with d2–d8 BTAG(+) cells (Figure 3A). Notably, d4 BTAG(+) cells induced from iMeLCs formed a sub-cluster with d6 BTAG(+) cells induced directly from hiPSCs (Figure 3A), suggesting that in floating aggregates, hiPSCs progress to BTAG(+) cells via a pathway similar to that for BTAG(+) cell induction through iMeLCs. Consistently, principal component analysis (PCA) revealed a directional and progressive transition of cellular properties during BTAG(+) cell induction (Figure 3B).
      Figure thumbnail gr3
      Figure 3BTAG(+) Cells as hPGCLCs
      (A) UHC of the transcriptomes (two independent experiments) of hiPSCs, iMeLCs, and BTAG(+) cells induced through iMeLCs (days 2, 4, 6, and 8) (black#) or directly from hiPSCs (day 6) (blue).
      (B) PCA of cells as in (A). Color codes for the cell types are indicated.
      (C) (Left) Scatter plots of the expression of genes upregulated (yellow) or downregulated (blue) (fold changes > 4, FDR < 0.01 by Welch t test) in cyPGCs compared to cyESCs or their human/mouse orthologs () in the indicated cell types. (Right) Histogram of the frequency distribution of DEGs between cyPGCs and cyESCs or their human/mouse orthologs plotted against their log2 fold changes in the indicated comparisons.
      (D) UHC of relevant cell types during BTAG(+) cell induction and of cyESCs (CMK9) and cyPGCs (E43, 50, 51, all female) based on the expression of genes (and their human orthologs) upregulated in cyPGCs compared to cyESCs (fold changes > 4, FDR < 0.01 by Welch t test). The gene expression level is represented by a heat map. For cyESCs and cyPGCs, the averaged expression levels of single-cell samples were shown. The GO functional terms and representative genes included are shown for each gene cluster.
      (E) (Top) Scatter plots of the expression of DEGs between hPGCs and H9 ESCs (fold changes > 4, p < 0.01 by ANOVA one-way test, up: yellow; down: blue) (
      • Irie N.
      • Weinberger L.
      • Tang W.W.
      • Kobayashi T.
      • Viukov S.
      • Manor Y.S.
      • Dietmann S.
      • Hanna J.H.
      • Surani M.A.
      SOX17 is a critical specifier of human primordial germ cell fate.
      ) in d6 BTAG(+) cells and iMeLCs. (Bottom) Histogram of the frequency distribution of DEGs between hPGCs and H9 ESCs plotted against log2 fold changes in the indicated comparison.
      (F) UHC of relevant cell types in Irie et al. (in red) (
      • Irie N.
      • Weinberger L.
      • Tang W.W.
      • Kobayashi T.
      • Viukov S.
      • Manor Y.S.
      • Dietmann S.
      • Hanna J.H.
      • Surani M.A.
      SOX17 is a critical specifier of human primordial germ cell fate.
      ) and the present study (in black) based on the expression of genes upregulated in hPGCs compared to H9 ESCs (fold changes > 4, p < 0.01 by ANOVA one-way test). The normalized gene expression level is represented by a heat map.
      (G) Scatter plots of the log2 fold changes of DEGs between iMeLCs and d4 BTAG(+) cells (fold changes > 3, n = 1,144) plotted against their log2 fold changes between pre-induced cells and d4 TNAP/NANOS3(+) cells (top) and of DEGs between pre-induced cells and d4 TNAP/NANOS3(+) cells (fold changes > 3, n = 1,165) plotted against their log2 fold changes between iMeLCs and d4 BTAG(+) cells (bottom). Pearson’s correlation coefficients (R) are shown.
      See also , , and .
      We compared the gene expression of the BTAG(+) cells with that of cyPGCs. To this end, we generated single-cell cDNAs from a line of cyESCs, CMK9 (
      • Suemori H.
      • Tada T.
      • Torii R.
      • Hosoi Y.
      • Kobayashi K.
      • Imahie H.
      • Kondo Y.
      • Iritani A.
      • Nakatsuji N.
      Establishment of embryonic stem cell lines from cynomolgus monkey blastocysts produced by IVF or ICSI.
      ) (Figures S5C and S5D), subjected these cDNAs to the RNA-seq analyses (Figure S5E), identified differentially expressed genes (DEGs) between cyESCs and cyPGCs, and determined their human orthologs (Table S1). Remarkably, scatter plot analyses showed that the genes up- or downregulated in cyPGCs were generally up- or downregulated in d6 BTAG(+) cells compared to iMeLCs/hiPSCs, but notably, this trend was much weaker in the comparison between mPGCLCs and EpiLCs (Figure 3C). Consistently, UHC analysis with genes upregulated in cyPGCs revealed that d2–d8 BTAG(+) cells were clustered with cyPGCs: the genes co-expressed between the BTAG(+) cells and cyPGCs were enriched with those bearing gene ontology (GO) terms such as “stem cell maintenance,” “reproductive developmental process,” and “gamete generation” (Figure 3D). A group of genes exclusively expressed in cyPGCs represented late PGC genes (the GO terms such as “sexual reproduction” and “sex differentiation”) (Figure 3D). Thus, the BTAG(+) cells bear gene expression profiles similar to those of early cyPGCs, and mPGCLCs show markedly different gene expression from cyPGCs, most likely due to the species difference.

      Comparison of BTAG(+) Cells with Gonadal hPGCs and NANOS3/TNAP(+) Cells by Irie et al.

      Recently, based on the concept of mPGCLC induction, Irie et al. reported the induction of TNAP/NANOS3(+) cells with a transcriptome similar to that of gonadal hPGCs at 7 weeks of gestation from hESCs/hiPSCs cultured with four kinase inhibitors (4i) (
      • Irie N.
      • Weinberger L.
      • Tang W.W.
      • Kobayashi T.
      • Viukov S.
      • Manor Y.S.
      • Dietmann S.
      • Hanna J.H.
      • Surani M.A.
      SOX17 is a critical specifier of human primordial germ cell fate.
      ). We compared transcriptional profiles of the BTAG(+) cells with those of hPGCs and TNAP/NANOS3(+) cells. Since the methods for cDNA preparation and the platforms for RNA-seq were different between the two studies, it was not possible to perform a strict comparison of their transcriptome data. We therefore first identified DEGs between H9 ESCs and gonadal hPGCs (Table S2) and examined their expression in our samples. Remarkably, genes up- or downregulated in hPGCs were almost exclusively up- or downregulated in the BTAG(+) cells compared to iMeLCs (Figure 3E). The heat map for the expression of genes upregulated in hPGCs in our samples and those of Irie et al. revealed that a majority of them, except mainly late PGC genes, were upregulated similarly in BTAG(+) cells and TNAP/NANOS3(+) cells (Figure 3F).
      We next identified the DEGs between pre-induced cells and TNAP/NANOS3(+) cells by Irie et al. and examined their expression in our cells; reciprocally, we identified the DEGs between iMeLCs and BTAG(+) cells and examined their expression in cells by Irie et al. This analysis revealed that the gene expression profiles of BTAG(+) cells and TNAP/NANOS3(+) cells are highly correlated (Figure 3G). Collectively, these findings indicate that the BTAG(+) cells bear similar transcription profiles to hPGCs and TNAP/NANOS3(+) cells. Combined with our analysis on cyPGCs, this led us to designate the BTAG(+) cells as hPGCLCs.

      A Unique Pathway for hPGCLC Specification

      The numbers of genes up- or downregulated between the cell-state transitions during hPGCLC induction were substantially smaller compared to those showing such regulation during mPGCLC induction (Figure 4A, Table S3, and Table S4). The genes upregulated during the hiPSC-to-iMeLC transition were enriched with those bearing GO terms such as “cell migration” and “pattern specification process,” whereas those downregulated were enriched with those for “chemical homeostasis” and “cell adhesion” (Figure 4B and Table S3). The genes upregulated during the iMeLC-to-d2 hPGCLC transition included potential regulators for hPGCLC specification (e.g., TFAP2C, PRDM1, SOX17, SOX15, KLF4, KIT, TCL1A, DND1, etc.) (Figures 3C and 4B and Table S3) and were enriched with genes for “stem cell maintenance” and “regulation of cell migration,” whereas those downregulated were enriched with genes for “pattern specification process” and “neuron development” (Figure 4B and Table S3). The gene expression changes between d2 and d4 hPGCLCs were relatively minor, with a few genes for “embryonic morphogenesis” downregulated in d4 hPGCLCs, and the gene expression profiles of the d6 and d8 hPGCLCs were essentially identical (Figures 4A and 4B and Table S3).
      Figure thumbnail gr4
      Figure 4A Pathway to hPGCLCs
      (A) Numbers of genes up- or downregulated (fold change > 2, FDR < 0.01 by ANOVA one-way test) between key stages/cell types during human (left) and mouse (right) PGCLC specification.
      (B) GO analysis of the genes shown in (A) during hPGCLC specification.
      (C) Numbers of genes showing the highest expression in each cell type (fold change > 2, compared to the other 3 cell types) during human (left) and mouse (right) PGCLC specification.
      (D) GO analysis of the human and mouse d2 PGCLC genes shown in (C).
      (E) Overlap between the human and mouse d2 PGCLC genes shown in (C).
      (F and G) Box plots (with the median value and 25th and 75th percentiles, top) and heat map (bottom) of the expression of human (F) and mouse (G) d2 PGCLC genes during human (left) and mouse (right) PGCLC specification.
      (H) Heat map of the normalized expression of selected genes associated with mouse naive pluripotency and the epiblast, mesoderm, and endoderm in mESCs, EpiLCs, EpiSCs, hiPSCs, and iMeLCs. Gene expression data by microarray analysis in
      • Hayashi K.
      • Ohta H.
      • Kurimoto K.
      • Aramaki S.
      • Saitou M.
      Reconstitution of the mouse germ cell specification pathway in culture by pluripotent stem cells.
      and data from the present study by RNA-seq analysis were used.
      See also and , , and .
      We identified genes expressed at the highest level in each cell type during hPGCLC and mPGCLC induction (Figure 4C). A key event for mPGC/PGCLC induction is acute and robust activation with subsequent repression of “a somatic mesodermal program” (
      • Kurimoto K.
      • Yabuta Y.
      • Ohinata Y.
      • Shigeta M.
      • Yamanaka K.
      • Saitou M.
      Complex genome-wide transcription dynamics orchestrated by Blimp1 for the specification of the germ cell lineage in mice.
      ,
      • Saitou M.
      • Barton S.C.
      • Surani M.A.
      A molecular programme for the specification of germ cell fate in mice.
      ,
      • Yabuta Y.
      • Kurimoto K.
      • Ohinata Y.
      • Seki Y.
      • Saitou M.
      Gene expression dynamics during germline specification in mice identified by quantitative single-cell gene expression profiling.
      ). Consistently, genes with the highest expression in d2 mPGCLCs (the d2 mPGCLC genes) were the largest in number (n = 756) and were enriched with such GO terms as “embryonic morphogenesis” and “pattern specification process” with highly significant p values, and, in a majority of cases, they were repressed in d4 mPGCLCs (Figures 4C and 4D and Table S4). In contrast, genes with the highest expression in d2 hPGCLCs (the d2 hPGCLC genes) were relatively small in number (104), exhibited enrichment in GO terms for “cellular component morphogenesis” and “neuron differentiation,” and gradually repressed in d4 and d6 hPGCLCs (Figures 4C and 4D and Table S3).
      Strikingly, the vast majority of the human orthologs of the d2 mPGCLC genes exhibited relatively constant expression (either not expressed or expressed at similar levels) during hPGCLC induction, while only a minority of the mouse orthologs of the d2 hPGCLC genes showed transient upregulation in d2 mPGCLCs (Figures 4E–4G, Table S3, and Table S4). Consequently, as few as 16 genes, not including the Hox genes, were shared in common between the d2 hPGCLC and mPGCLC genes (Figure 4E and Figure S5F). These findings revealed a distinct transcriptional program between hPGCLC and mPGCLC induction and a lack of prominent activation with subsequent repression of “a somatic program” during hPGCLC induction, even though hPGCLC and mPGCLC induction depend on identical signaling molecules.

      hiPSCs Bear a Property Intermediate between That of EpiLCs and EpiSCs

      We explored the relationship among hiPSCs, iMeLCs, mESCs, EpiLCs, and EpiSCs. Since global transcription states between human and mouse cells are not directly comparable due to the species difference, first, we selected representative genes in mice with key functions associated with naive pluripotency and the pre-gastrulation epiblast, mesoderm, and endoderm and examined their expression profiles among these cell types. Genes for naive pluripotency were highly expressed in mESCs, downregulated in EpiLCs, and further downregulated in EpiSCs, whereas genes for the epiblast were upregulated in both EpiLCs and EpiSCs (Figure 4H). Genes for the mesoderm and endoderm were generally low in mESCs and EpiLCs but were upregulated to some extent in EpiSCs (Figure 4H). Thus, mESCs, EpiLCs, and EpiSCs bear properties similar to the early epiblast, pre-gastrulation epiblast, and post-gastrulation epiblast, respectively. The expression patterns of genes for naive pluripotency in hiPSCs and iMeLCs were somewhat similar to those in EpiLCs (Figure 4H). Around half of the selected genes for the pre-gastrulation epiblast in mice exhibited strong expression in both hiPSCs and iMeLCs. Consistent with the qPCR analysis (Figures 1F and 2G and Figure S2C), some of the genes for the mesoderm—but not endoderm—were upregulated in iMeLCs, but not in hiPSCs (Figure 4H).
      Second, we classified the genes expressed at relevant levels among mESCs, EpiLCs, and EpiSCs (
      • Hayashi K.
      • Ohta H.
      • Kurimoto K.
      • Aramaki S.
      • Saitou M.
      Reconstitution of the mouse germ cell specification pathway in culture by pluripotent stem cells.
      ) into five classes (Figure S5G). We normalized the expression data in
      • Hayashi K.
      • Ohta H.
      • Kurimoto K.
      • Aramaki S.
      • Saitou M.
      Reconstitution of the mouse germ cell specification pathway in culture by pluripotent stem cells.
      and the present study, identified the human orthologs of the five classes of genes, and analyzed their expression in the relevant cell types, which revealed a trend of hiPSCs/iMeLCs exhibiting properties intermediate between those of EpiLCs and EpiSCs (Figure S5H): the averaged expression levels of the class I and V genes (high in EpiLCs and EpiSCs and low in EpiLCs and EpiSCs, respectively) in hiPSCs/iMeLCs were similar to those in EpiLCs and EpiSCs, and the averaged expression levels of the class II and IV genes (high and low in EpiSCs, respectively) in hiPSCs/iMeLCs were intermediate between those in EpiLCs and EpiSCs. Based on these observations, together with the analysis of the expression of key genes (Figure 4H), we propose that hiPSCs bear a property intermediate between that of EpiLCs and EpiSCs, and were able to take on a germ cell fate with a molecular pathway specific to humans.

      Epigenetic Properties of hPGCLCs

      We next evaluated the epigenetic properties of hPGCLCs. IF analyses revealed that compared to hiPSCs, d8 hPGCLCs exhibited lower histone H3 lysine 9 di-methylation (H3K9me2) and DNA methylation levels, whereas they showed variegated (some higher and others similar) levels of histone H3 lysine 27 tri-methylation (H3K27me3) (Figure S6A). We then examined the methylation states of differentially methylated regions (DMRs) of paternally (H19 and MEG3) and maternally (KCNQ1OT1 and PEG10) imprinted genes in hiPSCs and d8 hPGCLCs by bisulfite pyro-sequencing, which revealed that d8 hPGCLCs may initiate the imprint erasure of H19, but retain the imprints of MEG3, KCNQ1, and PEG10 (Figure S6B). Thus, the epigenetic properties of hPGCLCs are similar to those of mPGCLCs (
      • Hayashi K.
      • Ohta H.
      • Kurimoto K.
      • Aramaki S.
      • Saitou M.
      Reconstitution of the mouse germ cell specification pathway in culture by pluripotent stem cells.
      ,
      • Kurimoto K.
      • Yabuta Y.
      • Hayashi K.
      • Ohta H.
      • Kiyonari H.
      • Mitani T.
      • Moritoki Y.
      • Kohri K.
      • Kimura H.
      • Yamamoto T.
      • et al.
      Quantitative Dynamics of Chromatin Remodeling during Germ Cell Specification from Mouse Embryonic Stem Cells.
      ).
      We also examined the expression of epigenetic modifiers during hPGCLC induction and in cyPGCs (Figure S6C): among the molecules involved in DNA methylation, hPGCLCs repressed DNMT3B, but retained DNMT1 and UHRF1, and cyPGCs repressed DNMT3A/3B and UHRF1. Among the molecules involved in DNA demethylation, TET1 showed relatively constant expression, while other genes showed no/low expression in both hPGCLCs and cyPGCs. Both hPGCLCs and cyPGCs repressed EHMT2, a key enzyme for H3K9me2, and they expressed several H3K9me2 demethylases (KDM1A/3A/3B). Among the molecules involved in H3K27me3, EZH2 and SUZ12 showed constant expression during hPGCLC induction, while EED was repressed in hPGCLCs, whereas cyPGCs strongly expressed EZH2, EED, and SUZ12.

      Identification of Surface Markers for hPGCLCs

      We explored surface markers that distinguish hPGCLCs without the use of fluorescent reporters. Among a number of markers that we screened, the combination of EpCAM and INTEGRINα6 segregated the day-6 aggregates from BTAG hiPSCs into three distinct populations. Remarkably, the EpCAM-/INTEGRINα6-high population was nearly identical to the BTAG(+) hPGCLCs [∼98.9% of the EpCAM-/INTEGRINα6-high cells were BTAG(+)] and the other two populations were essentially negative/weak for BTAG (Figure 5A). As early as day 2 of induction, the EpCAM-/INTEGRINα6-high population was discernible (∼37%) and essentially identical to the BTAG(+) population, and thereafter it became more distinct and persisted until at least day 8 (Figure 5B).
      Figure thumbnail gr5
      Figure 5Surface Markers for hPGCLCs
      (A) (Left) FACS by EpCAM and INTEGRINα6 expression of d6 aggregates induced from BTAG hiPSCs through iMeLCs. (Right) FACS by BTAG of the three populations on the left classified by EpCAM and INTEGRINα6 expression.
      (B) (Top) FACS by EpCAM and INTEGRINα6 expression of cells during hPGCLC induction (until day 8) from BTAG hiPSCs through iMeLCs. Full: induction by BMP4, LIF, SCF, and EGF. P3 and P5 gates (boxed areas) indicate EpCAM/INTEGRINα6-high and -low/no cells, respectively. The percentages of cells in the P3 gate are shown. (Bottom) FACS showing the percentages of BTAG(+) cells in the P3 and P5 gates shown in the top panel.
      (C) FACS by EpCAM and INTEGRINα6 expression of cells during hPGCLC induction (until day 6) from 585A1 hiPSCs without fluorescent reporters through iMeLCs. Full: induction by BMP4, LIF, SCF, and EGF. The percentages of cells in the P3 gate are shown.
      (D) UHC of the transcriptomes (two independent experiments) of hiPSCs, iMeLCs, and BTAG(+) cells induced through iMeLCs (day 2, 4, 6, and 8) and EpCAM/INTEGRINα6-high cells (day 6) induced from 585A1 hiPSCs through iMeLCs (in red).
      (E) (Top) FACS by EpCAM and INTEGRINα6 expression of d6 aggregates induced from 1383D2 (left) and 1383D6 (right) through iMeLCs. iMeLCs were induced with FGFRi. (Middle and bottom) (Left) FACS by EpCAM and INTEGRINα6 expression of d4 aggregates induced from 201B7 through iMeLCs. iMeLCs were induced without (middle) or with (bottom) FGFRi. (Right) Comparison of expression levels of relevant genes (see , Primers Used in This Study) in the P3 and P5 populations with those in d4 BTAG(+) cells with correlation coefficients (R).
      We induced the 585A1 hiPSCs without the reporters into hPGCLCs. As early as day 2, there appeared an EpCAM-/INTEGRINα6-high population (∼21%), and this population was increased at days 4 (∼31%) and 6 (∼32%) (Figure 5C). To determine whether this population represented hPGCLCs, we isolated total RNA from the population at day 6 and analyzed its transcriptome. The UHC revealed that the d6 EpCAM-/INTEGRINα6-high cells were clustered tightly with d6 and d8 hPGCLCs, demonstrating that they were indeed the hPGCLCs (Figure 5D). The two newly derived hiPSC lines, 1383D2 and 1383D6 (XY, from PMBCs), were also induced robustly into EpCAM-/INTEGRINα6-high cells (Figure 5E). Finally, the 201B7 line, the first and one of the most frequently used hiPSC lines derived from human dermal fibroblasts (XX) (
      • Takahashi K.
      • Tanabe K.
      • Ohnuki M.
      • Narita M.
      • Ichisaka T.
      • Tomoda K.
      • Yamanaka S.
      Induction of pluripotent stem cells from adult human fibroblasts by defined factors.
      ), was induced robustly into EpCAM-/INTEGRINα6-high cells showing gene expression highly similar to that in hPGCLCs (Figure 5E). We conclude that hiPSCs with a primed pluripotency bear a robust competence for germ cell fate.

      A Critical Function of BLIMP1 in hPGCLC Specification

      We examined the role of BLIMP1 in hPGCLC specification. Since a knockout of BLIMP1 in BTAG hiPSCs might impair BT expression in an unpredictable fashion, we generated an hiPSC line bearing a TFAP2C-2A-EGFP allele (AG 585B1-17, 19) and replaced the exon 4 of the BLIMP1 gene with tdTomato (Figure 6A and Figures S7A and S7B). We isolated several clones in which BLIMP1 was heterozygously knocked out (BTAG; BLIMP1+/−) and those bearing one BLIMP1 allele with targeted replacement by tdTomato and another allele with frame-shift deletions (BLIMP1 homozygously knocked out clones: BTAG; BLIMP1−/−) (Figure 6A and Figure S7B).
      Figure thumbnail gr6
      Figure 6BLIMP1 Is Essential for hPGCLCs
      (A) Scheme of the targeting vector for BLIMP1 and the targeted allele and the deletions of the BLIMP1 locus.
      (B) IF analysis for AG (top, green) and BLIMP1 (bottom, red) expression in AG; BLIMP1+/+ or BTAG; BLIMP1−/− cells induced from hiPSCs (BLIMP1+/+: clone 1-7, 1-9; BLIMP1−/−: clone 1-7-5, 1-9-6) through iMeLCs. Bar, 10 μm.
      (C) Bright field (BF) and fluorescence (AG and BT) images of floating aggregates (3,000 cells/initial aggregate) induced from wild-type (left), BTAG; BLIMP1+/− (middle), or BTAG; BLIMP1−/− (right) hiPSCs through iMeLCs stimulated by BMP4, SCF, EGF, and LIF for 2, 4, and 6 days. Bar, 200 μm.
      (D) FACS analysis of the induction of (BT)AG(+) cells for 8 days from wild-type (top), BTAG; BLIMP1+/− (middle), and BTAG; BLIMP1−/− (bottom) hiPSCs through iMeLCs. Boxed areas indicate (BT)AG(+) cells with their percentages.
      (E) Numbers (two independent experiments with SDs) of (BT)AG(+) cells per aggregate induced from wild-type (wt: circle), BTAG; BLIMP1+/− (het: triangle), or BTAG; BLIMP1−/− (ko: square) hiPSCs through iMeLCs.
      (F) Gene expression profiles of d2 and d4 (BT)AG(+) cells induced from wild-type (wt) and BTAG; BLIMP1−/− (ko) hiPSCs, as measured by qPCR. The quantification of gene expression levels (two independent experiments) were as described in F.
      See also .
      We induced AG; BLIMP1+/+, BTAG; BLIMP1+/−, and BTAG; BLIMP1−/− hiPSCs into hPGCLCs. We confirmed that the AG(+) cells induced from the BLIMP1−/− hiPSCs lacked BLIMP1 (Figure 6B). The BLIMP1+/+ hiPSCs were robustly induced into AG(+) hPGCLCs and were maintained as hPGCLCs (d2: ∼69.3%; d4: ∼52.8%; d6: ∼32.2%; d8: ∼32.2%) (Figures 6C–6E). In contrast, although BLIMP1−/− hiPSCs were induced into BTAG(+) cells efficiently at day 2 (∼47.4%; BTAG/AG: ∼92%), the number of BTAG(+) cells declined sharply at day 4 (∼18.8%; BTAG/AG: ∼98%) and these cells had almost disappeared at day 6 (∼1.6%) (Figures 6C–6E), indicating that BLIMP1 is essential for the specification and maintenance of hPGCLCs. We obtained essentially the same results using the other independent BLIMP1−/− hiPSC line (data not shown). Notably, consistent with the dose-dependent function of Blimp1 in mice (
      • Ohinata Y.
      • Payer B.
      • O’Carroll D.
      • Ancelin K.
      • Ono Y.
      • Sano M.
      • Barton S.C.
      • Obukhanych T.
      • Nussenzweig M.
      • Tarakhovsky A.
      • et al.
      Blimp1 is a critical determinant of the germ cell lineage in mice.
      ), BLIMP1+/− hiPSCs exhibited a phenotype intermediate between those of the wild-type and BLIMP1−/− hiPSCs (d2: ∼49.6%; d4: ∼27.0%; d6: ∼8.1%; d8: ∼4.6%) (Figures 6C–6E).
      We isolated RNAs from d2 and d4 (BT)AG(+) cells induced from the BLIMP1+/+, BLIMP1+/−, and BLIMP1−/− hiPSCs, and we first analyzed the expression of key genes in BLIMP1+/+; AG(+) and BLIMP1−/−; BTAG(+) cells by qPCR (Figure 6F). This analysis revealed that, similarly to the wild-type cells, BLIMP1−/− cells upregulated POU5F1 and NANOG and repressed SOX2, indicating that regulation of key pluripotency genes in hPGCLCs is independent of BLIMP1. BLIMP1−/− cells upregulated TFAP2C normally, but they exhibited impaired upregulation of genes such as NANOS3, KLF4, TFCP2L1, and TCL1B. Interestingly, BLIMP1−/− cells failed to maintain T and MIXL1, whereas they failed to repress EVX1 and SP5, and apparently had no effects on EOMES and MSX2, indicating that BLIMP1 exerts differential effects on genes associated with mesoderm development. Similarly, BLIMP1−/− cells showed impaired repression of GATA4, but had no distinct influence on SOX17, GATA6, and FOXA2. Notably, BLIMP1−/− cells were unable to repress DNMT3B.
      We compared the transcriptome of BLIMP1−/−; BTAG(+) cells with that of wild-type and BLIMP1+/− hPGCLCs by RNA-seq (Figure S7C). UHC and PCA showed that although BLIMP1−/− cells acquired a property similar to that of d2 hPGCLCs, they failed to progress further toward the d4 PGCLC state (Figures 7A and 7B ). We explored genes up- or downregulated in d2 and d4 BLIMP1−/− cells in comparison to d2 and d4 PGCLCs, respectively. The genes upregulated in d2 BLIMP1−/− cells (104 genes) were enriched with genes for “neuron differentiation,” “gastrulation,” and “embryonic morphogenesis,” and those upregulated in d4 BLIMP1−/− cells (692 genes) were enriched with similar GO terms with much higher p values, indicating that BLIMP1 functions to repress such developmental programs in hPGCLCs (Figures 7C and 7D and Table S5). Consistent with the qPCR analysis (Figure 6F), among the genes involved in DNA (de)methylation, BLIMP1−/− cells failed to repress DNMT3B (Figure S7D). In contrast, the genes downregulated in BLIMP1−/− cells were smaller in number and were enriched with GO terms related to apoptosis and the cell cycle, suggesting mis-regulation of a fundamental cellular property in BLIMP1−/− cells (Figures 7C and 7D and Table S5).
      Figure thumbnail gr7
      Figure 7BLIMP1 Represses a Program for Neuron Differentiation
      (A) UHC of the transcriptomes of hiPSCs (two for BLIMP1+/+ [wt], one for BLIMP1+/− [het], and two for BLIMP1−/− [ko]) and d2 and d4 (BT)AG(+) cells induced from these hiPSCs.
      (B) PCA of cells as in (A). Color codes for the cell types are indicated.
      (C) Numbers of genes up- or downregulated (fold change > 2, p < 0.05 by Student’s t test) in BLIMP1−/− hiPSCs and d2 and d4 BTAG; BLIMP1−/− cells in comparison to their wild-type counterparts.
      (D) GO analysis of the genes shown in (C).
      (E) Box plots (with the median and 25th and 75th percentiles, top) and heat map (bottom) of the expression of genes upregulated (left) and downregulated (right) in the d2 BTAG; BLIMP1−/− cells shown in (C). Key genes with the GO terms listed in the top panel of (D) are indicated (# or ).
      (F) (Top) Expression dynamics of genes for “neuron differentiation” that show differential expression (FDR < 0.01 by ANOVA one-way test) categorized as indicated in hiPSCs and d2 and d4 hPGCLCs. Gene names for each category are shown and those upregulated in d2 BTAG; BLIMP1−/− cells are indicated in red. (Bottom) Expression dynamics of the genes shown above (top) in hiPSCs, and (BT)AG(+) d2 and d4 cells induced from wild-type (gray) and BLIMP1−/− (red) iPSCs.
      (G) A model for hPGCLC induction from hiPSCs through iMeLCs.
      See also and .
      The genes up- or downregulated in d2 BLIMP1−/− cells, including those for “neuron differentiation” and “embryonic morphogenesis,” exhibited differential expression patterns during hPGCLC induction; notably, in BLIMP1+/− cells, these genes were expressed at intermediate levels between wild-type and BLIMP1−/− cells (Figure 7E). Since genes for “neuron differentiation” and its related GO terms were highly enriched in genes upregulated in BLIMP1−/− cells, we explored the expression dynamics of such genes during hPGCLC induction and the impact of the BLIMP1 deficiency on their expression. These genes were classified into several expression categories and those upregulated in BLIMP1−/− cells were eminently included in both expression categories for upregulation (18/39, ∼46%) and downregulation (6/23, ∼26%) in d2 hPGCLCs (Figure 7F). It is also notable that the 15 genes without significant expression changes during hPGCLC induction were upregulated in BLIMP1−/− cells (Figure S7E). Thus, the key functions of BLIMP1 include the repression, in a dose-dependent fashion and to appropriate levels, of genes for “neuron differentiation” that are upregulated in d2 hPGCLCs, as well as the proper repression of such genes that are downregulated in d2 hPGCLCs.

      Discussion

      The precise definition of hPGCLCs involves an inherent difficulty due to the lack of information regarding the properties of early hPGCs. Nonetheless, we showed that the BTAG(+) cells bear properties highly similar to those of early cy and hPGCs (Figures 3C, 3D, and 3F) and defined them as hPGCLCs (Figure 7G). Our finding that hPGCLCs are robustly induced from hiPSCs in a primed pluripotent state, in particular through iMeLCs, is surprising, since EpiSCs with primed pluripotency show little, if any, competence for the germ cell fate (
      • Hayashi K.
      • Ohta H.
      • Kurimoto K.
      • Aramaki S.
      • Saitou M.
      Reconstitution of the mouse germ cell specification pathway in culture by pluripotent stem cells.
      ). However, it is important to note that hiPSCs/hESCs and EpiSCs do show differences in their properties (e.g., hiPSCs/hESCs express PRDM14, but EpiSCs do not [
      • Chia N.Y.
      • Chan Y.S.
      • Feng B.
      • Lu X.
      • Orlov Y.L.
      • Moreau D.
      • Kumar P.
      • Yang L.
      • Jiang J.
      • Lau M.S.
      • et al.
      A genome-wide RNAi screen reveals determinants of human embryonic stem cell identity.
      ,
      • Nakaki F.
      • Saitou M.
      PRDM14: a unique regulator for pluripotency and epigenetic reprogramming.
      ]) and it is difficult to strictly compare the properties between hiPSCs/hESCs and EpiSCs due to the species difference. We provided evidence that hiPSCs cultured under our conditions have a property intermediate between EpiLCs and EpiSCs (Figure 4H and Figure S5H). A definitive conclusion would require experiments such as an investigation into the properties of early post-implantation epiblasts in non-human primate models.
      We demonstrated that signaling requirements for the induction and proliferation/survival of hPGCLCs are similar to those for mPGCLCs, whereas downstream transcriptional programs are highly different between hPGCLCs and mPGCLCs, most likely due to the different regulatory structures of the genome between the two species (Figures 3C and 4). Importantly, moreover, unlike in the process for mPGC/PGCLC specification, transient activation and subsequent repression of the somatic mesodermal program are not eminent in hPGCLC specification, which, rather, appears to involve a more straightforward programming (Figure 4). Although a definitive conclusion would again require investigation in vivo in appropriate models, this might be a reflection of a potentially different mode of PGC specification arising from different time requirements and/or embryonic structures between mice and humans (and many other mammals) during the relevant period for PGC specification (
      • Hayashi Y.
      • Saitou M.
      • Yamanaka S.
      Germline development from human pluripotent stem cells toward disease modeling of infertility.
      ,
      • Saitou M.
      • Yamaji M.
      Germ cell specification in mice: signaling, transcription regulation, and epigenetic consequences.
      ). In this regard, it would be interesting to investigate whether T, a conserved mesodermal TF critical in mPGC specification (
      • Aramaki S.
      • Hayashi K.
      • Kurimoto K.
      • Ohta H.
      • Yabuta Y.
      • Iwanari H.
      • Mochizuki Y.
      • Hamakubo T.
      • Kato Y.
      • Shirahige K.
      • Saitou M.
      A mesodermal factor, T, specifies mouse germ cell fate by directly activating germline determinants.
      ), plays a similar role in hPGC specification.
      We demonstrated that BLIMP1 is essential for hPGCLC specification (Figures 6 and 7), as it is for mPGC/mPGCLC specification (
      • Kurimoto K.
      • Yabuta Y.
      • Ohinata Y.
      • Shigeta M.
      • Yamanaka K.
      • Saitou M.
      Complex genome-wide transcription dynamics orchestrated by Blimp1 for the specification of the germ cell lineage in mice.
      ,
      • Ohinata Y.
      • Payer B.
      • O’Carroll D.
      • Ancelin K.
      • Ono Y.
      • Sano M.
      • Barton S.C.
      • Obukhanych T.
      • Nussenzweig M.
      • Tarakhovsky A.
      • et al.
      Blimp1 is a critical determinant of the germ cell lineage in mice.
      ) (data not shown). The initial induction of BTAG(+) cells from BLIMP1−/− hiPSCs appeared normal, but BTAG(+) cells sharply declined thereafter and eventually disappeared (Figure 6). Contrary to a previous report (
      • Lin I.Y.
      • Chiu F.L.
      • Yeang C.H.
      • Chen H.F.
      • Chuang C.Y.
      • Yang S.Y.
      • Hou P.S.
      • Sintupisut N.
      • Ho H.N.
      • Kuo H.C.
      • Lin K.I.
      Suppression of the SOX2 neural effector gene by PRDM1 promotes human germ cell fate in embryonic stem cells.
      ), BLIMP1−/− cells repressed SOX2 in an essentially normal fashion, indicating that SOX2 repression is mediated by BMP4 signaling but is independent from BLIMP1 (Figure 6F). Consistent with a less prominent activation of the somatic mesodermal program during hPGCLC induction (Figure 4), the role of BLIMP1 in repression of such a program was variable and BLIMP1 was instead critical in activating genes potentially critical for PGCs and repressing genes for “neuron differentiation,” which may be an inherent/default program for differentiation from hiPSCs (Figure 7). Thus, the precise role of BLIMP1 appears to have diverged between mice and humans.
      Based on the concept of mPGCLC induction, two recent papers reported the induction of hPGCLCs from hESCs/hiPSCs (
      • Irie N.
      • Weinberger L.
      • Tang W.W.
      • Kobayashi T.
      • Viukov S.
      • Manor Y.S.
      • Dietmann S.
      • Hanna J.H.
      • Surani M.A.
      SOX17 is a critical specifier of human primordial germ cell fate.
      ,
      • Sugawa F.
      • Araúzo-Bravo M.J.
      • Yoon J.
      • Kim K.P.
      • Aramaki S.
      • Wu G.
      • Stehling M.
      • Psathaki O.E.
      • Hübner K.
      • Schöler H.R.
      Human primordial germ cell commitment in vitro associates with a unique PRDM14 expression profile.
      ). We showed that the transcriptional profiles of BTAG(+) cells are similar to those of TNAP/NANOS3(+) cells by Irie et al. (Figures 3E–3G). In contrast, cells by Sugawa et al. appear to exhibit different gene expression from BTAG(+) cells [e.g., fold enrichment ( × ) from hiPSCs, BTAG(+) cells versus KIT/TRA-1-81(+) cells in Figures 2D and 5C by Sugawa et al.: OCT4: × 4 versus × 0.25; NANOG: × 16 versus × 1; SOX2: undetectable versus ∼ × 0.4; BLIMP1: more than × 1,000 versus ∼ × 20; TFAP2C: × 250 versus × 1; NANOS3: more than × 1,000 versus ∼ × 30], suggesting that the cells by Sugawa et al. are not properly sorted or induced. Notably, the hPSCs in 4i by Irie et al. did not show consistent upregulation of genes for naive pluripotency, but rather exhibited upregulation of mesodermal markers, suggesting that the hPSCs in 4i are not in the postulated naive state, but rather in a type of peri-gastrulating epiblast-like state, similar to iMeLCs. The different competence of hPSCs for hPGCLC induction between the two studies might result from different culture conditions; i.e., we cultured hiPSCs under a feeder-free, defined condition on rLN511E8, whereas Irie et al. maintained hPSCs on feeders in different culture media. Our finding that hPGCLCs are induced robustly from a primed state through iMeLCs would necessitate a reconsideration of the definition of naive pluripotency and provide a critical opportunity to define the spectrum of human pluripotent states.
      The hPGCLCs do not show expression of late PGC markers such as DAZL and DDX4 even at day 12 of induction (Figure 1, Figure 2, Figure 3 and Figure S4E) and thus would correspond to early hPGCs. The late PGC genes have been shown to be repressed by H3K27me3 and H3K9me2 as well as by DNA methylation in mPGCLCs (
      • Kurimoto K.
      • Yabuta Y.
      • Hayashi K.
      • Ohta H.
      • Kiyonari H.
      • Mitani T.
      • Moritoki Y.
      • Kohri K.
      • Kimura H.
      • Yamamoto T.
      • et al.
      Quantitative Dynamics of Chromatin Remodeling during Germ Cell Specification from Mouse Embryonic Stem Cells.
      ) (data not shown) and exhibit strong upregulation upon aggregation with embryonic gonads, particularly with female embryonic gonads (
      • Hayashi K.
      • Ogushi S.
      • Kurimoto K.
      • Shimamoto S.
      • Ohta H.
      • Saitou M.
      Offspring from oocytes derived from in vitro primordial germ cell-like cells in mice.
      ). To further reconstitute human germ cell development in vitro, one of the key challenges would therefore be to explore whether it is possible to further differentiate hPGCLCs into those with a later hPGC phenotype. Such studies woud also highlight the mechanism of epigenetic reprogramming in the human germ cell lineage.

      Experimental Procedures

      All the animal experiments were performed under the ethical guidelines of Kyoto University and Shiga University of Medical Science. The experiments on the induction of hPGCLCs from hiPSCs were approved by the Institutional Review Board of Kyoto University and were performed according to the guidelines from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) of Japan. The experimental procedures for generation of the BTAG-knockin reporter lines; generation of BLIMP1-knockin/knockout hiPSCs; karyotyping and G-band analyses; FACS analysis; qPCR and RNA-seq analyses; mapping reads of RNA-seq and conversion to gene expression levels; comparison of gene expression among humans, cynomolgus monkeys, and mice; transcriptome analysis; immunofluorescence analysis; and bisulfite pyrosequencing are available in the Supplemental Information.

      Culture of hiPSCs

      The hiPSC lines (201B7, 585A1, and 585B1 were maintained under a conventional condition (Dulbecco’s modified Eagle medium [DMEM/F12; Life Technologies] supplemented with 20% [v/v] KSR [Life Technologies], 1% GlutaMax [Life Technologies], 0.1 mM nonessential amino acids, 4 ng/ml recombinant human bFGF [Wako Pure Chemical Industries], and 0.1 mM 2-mercaptoethanol) on mitomycin C (MMC)-inactivated SNL feeder cells (
      • Okita K.
      • Yamakawa T.
      • Matsumura Y.
      • Sato Y.
      • Amano N.
      • Watanabe A.
      • Goshima N.
      • Yamanaka S.
      An efficient nonviral method to generate integration-free human-induced pluripotent stem cells from cord blood and peripheral blood cells.
      ) and were subsequently adapted to a feeder-free condition (StemFit [Ajinomoto, Tokyo, Japan] medium on recombinant LAMININ511 [rLN511E8] [iMatrix-511, Nippi, Tokyo, Japan]-coated cell culture plates) (
      • Nakagawa M.
      • Taniguchi Y.
      • Senda S.
      • Takizawa N.
      • Ichisaka T.
      • Asano K.
      • Morizane A.
      • Doi D.
      • Takahashi J.
      • Nishizawa M.
      • et al.
      A novel efficient feeder-free culture system for the derivation of human induced pluripotent stem cells.
      ). The 1383D2 and 1383D6 lines were derived and cultured under the feeder-free condition. For the passage or the induction of differentiation, the cells were treated with a 1 to 1 mixture of TrypLE Select (Life Technologies) and 0.5 mM EDTA/PBS to enable their dissociation into single cells, and 10 μM of a ROCK inhibitor (Y-27632; Wako Pure Chemical Industries) was added for 24 hr after plating.

      Induction of iMeLCs and hPGCLCs

      The iMeLCs were induced by plating 1.0–2.0 × 105 hiPSCs maintained in StemFit onto a well of a human plasma fibronectin (Millipore, FC010)-coated 12-well plate in GK15 medium (GMEM [Life Technologies] with 15% KSR, 0.1 mM NEAA, 2 mM L-glutamine, 1 mM sodium pyruvate, and 0.1 mM 2-mercaptoethanol) containing 50 ng/ml of ACTA, 3 μM of CHIR, and 10 μM of a ROCK inhibitor (Y-27632; Wako Pure Chemical Industries). The hPGCLCs were induced by plating 3.0 × 103 iMeLCs or iPSCs into a well of a low-cell-binding V-bottom 96-well plate (Thermo, 81100574) in GK15 supplemented with 1,000 U/ml (5,000 U/ml for 201B7) of LIF (Millipore, #LIF1005), 200 ng/ml of BMP4, 100 ng/ml (200 ng/ml for 201B7) of SCF (R&D Systems, 455-MC), 50 ng/ml (250 ng/ml for 201B7) of EGF (R&D Systems, 236-EG), and 10 μM of the ROCK inhibitor. In some experiments, PD173074 (StemGent, #04-0008) or LDN193189 (StemGent, #04-0074) was added for iMeLC or hPGCLC induction, respectively. In other experiments, WNT3A (R&D Systems, 5036-WN) was used in place of CHIR, and BMP2 (R&D Systems, 355-BM), BMP7 (R&D Systems, 354-BP), or BMP8A (R&D Systems, 1073-BP) was used in place of BMP4.

      Preparation of Single-Cell cDNAs from cyESCs and cyPGCs

      cyESCs (CMK9) were the gift of Dr. Suemori (
      • Suemori H.
      • Tada T.
      • Torii R.
      • Hosoi Y.
      • Kobayashi K.
      • Imahie H.
      • Kondo Y.
      • Iritani A.
      • Nakatsuji N.
      Establishment of embryonic stem cell lines from cynomolgus monkey blastocysts produced by IVF or ICSI.
      ). They were cultured with conventional hESC medium (DMEM/F12 [Life Technologies] supplemented with 20% [vol/vol] of KSR [Life Technologies], 1 mM of sodium pyruvate [Life Technologies], 2 mM of GlutaMax [Life Technologies], 0.1 mM of nonessential amino acids [Life Technologies], 0.1 mM of 2-mercaptoethanol [Sigma-Aldrich], 1000 U/ml of ESGRO mouse LIF [Millipore], and 4 ng/ml of recombinant human bFGF [Wako Pure Chemical Industries]) on mouse embryonic feeders (MEFs). For isolating single cells for the single-cell mRNA 3-prime end sequencing (SC3-seq) analysis (
      • Nakamura T.
      • Yabuta Y.
      • Okamoto I.
      • Aramaki S.
      • Yokobayashi S.
      • Kurimoto K.
      • Sekiguchi K.
      • Nakagawa M.
      • Yamamoto T.
      • Saitou M.
      SC3-seq: a method for highly parallel and quantitative measurement of single-cell gene expression.
      ), cells were first detached as clumps with CTK solution (0.25% of Trypsin [Life Technologies], 0.1 mg/ml of Collagenase IV [Life Technologies], and 1 mM of CaCl2 [Nacalai Tesque]), incubated in 0.25% trypsin/PBS (Sigma-Aldrich) for around 10 min at 37°C, and dispersed into single cells in 1% (vol/vol) KSR/PBS.
      The technologies in cynomolgus monkeys for oocyte collection, intra-cytoplasmic sperm injection (ICSI), pre-implantation embryo culture, and transfer of pre-implantation embryos into foster mothers were reported previously (
      • Yamasaki J.
      • Iwatani C.
      • Tsuchiya H.
      • Okahara J.
      • Sankai T.
      • Torii R.
      Vitrification and transfer of cynomolgus monkey (Macaca fascicularis) embryos fertilized by intracytoplasmic sperm injection.
      ). Implanted embryos were monitored by ultrasound diagnosis and recovered by Caesarian section at embryonic days 43, 50, and 51. The gender of the embryos was determined by sex-specific PCR on genomic DNA isolated from somatic tissues (
      • Wilson J.F.
      • Erlandsson R.
      Sexing of human and other primate DNA.
      ). Genital ridges were dissected out and were dissociated into single cells by being incubated with 0.25% trypsin/PBS for around 10 min at 37°C followed by repeated pipetting. The resulting single cells were dispersed in 0.1 mg/ml of PVA/PBS (Sigma-Aldrich) and were processed for the SC3-seq analysis (
      • Nakamura T.
      • Yabuta Y.
      • Okamoto I.
      • Aramaki S.
      • Yokobayashi S.
      • Kurimoto K.
      • Sekiguchi K.
      • Nakagawa M.
      • Yamamoto T.
      • Saitou M.
      SC3-seq: a method for highly parallel and quantitative measurement of single-cell gene expression.
      ).

      Author Contributions

      K.S. and S.Y. conducted the overall experiments and analyzed the data. T.N., I.O., Y.Y., K.K., H.O., Y.M., and T.Y. contributed to the RNA-seq data. C.I., H.T., and S.N. contributed to the isolation of cynomolgus embryos. K.S. and M.N. contributed to the hiPSC culture. T.S., T.Y., and K.W. contributed to the TALEN experiments. T.M., K.T., S.Y., and M.S. conceived the project, and K.S., S.Y., and M.S. designed the experiments and wrote the manuscript.

      Acknowledgments

      We are grateful to K. Okita for providing us with hiPSCs (585A1 and 585B1), to H. Suemori for cyESCs (CMK9), and to J. Toga for rLN511E8. We also thank Y. Nagai, R. Kabata, T. Sato, and M. Kabata for their technical assistance. This work was supported in part by a Grant-in-Aid from MEXT , by JST-ERATO , and by the Academia for Repro-Regenerative Medicine .

      Accession Numbers

      The RNA-seq data in this study were deposited in the NCBI database (GEO accession number: GEO: GSE67259).

      Supplemental Information

      • Table S2. DEGs between H9ESCs and hPGCs (Irie et al., 2015), Pre-Induced Cells and TNAP/NANOS3(+) Cells (Irie et al., 2015), and iMeLCs and d4 BTAG(+) Cells and Their Expression during BTAG(+) or TNAP/NANOS3(+) Cell Induction, Related to Figure 3

      References

        • Aramaki S.
        • Hayashi K.
        • Kurimoto K.
        • Ohta H.
        • Yabuta Y.
        • Iwanari H.
        • Mochizuki Y.
        • Hamakubo T.
        • Kato Y.
        • Shirahige K.
        • Saitou M.
        A mesodermal factor, T, specifies mouse germ cell fate by directly activating germline determinants.
        Dev. Cell. 2013; 27: 516-529
        • Brons I.G.
        • Smithers L.E.
        • Trotter M.W.
        • Rugg-Gunn P.
        • Sun B.
        • Chuva de Sousa Lopes S.M.
        • Howlett S.K.
        • Clarkson A.
        • Ahrlund-Richter L.
        • Pedersen R.A.
        • Vallier L.
        Derivation of pluripotent epiblast stem cells from mammalian embryos.
        Nature. 2007; 448: 191-195
        • Chia N.Y.
        • Chan Y.S.
        • Feng B.
        • Lu X.
        • Orlov Y.L.
        • Moreau D.
        • Kumar P.
        • Yang L.
        • Jiang J.
        • Lau M.S.
        • et al.
        A genome-wide RNAi screen reveals determinants of human embryonic stem cell identity.
        Nature. 2010; 468: 316-320
        • de Jong J.
        • Stoop H.
        • Gillis A.J.
        • van Gurp R.J.
        • van de Geijn G.J.
        • Boer Md.
        • Hersmus R.
        • Saunders P.T.
        • Anderson R.A.
        • Oosterhuis J.W.
        • Looijenga L.H.
        Differential expression of SOX17 and SOX2 in germ cells and stem cells has biological and clinical implications.
        J. Pathol. 2008; 215: 21-30
        • Eckert D.
        • Biermann K.
        • Nettersheim D.
        • Gillis A.J.
        • Steger K.
        • Jäck H.M.
        • Müller A.M.
        • Looijenga L.H.
        • Schorle H.
        Expression of BLIMP1/PRMT5 and concurrent histone H2A/H4 arginine 3 dimethylation in fetal germ cells, CIS/IGCNU and germ cell tumors.
        BMC Dev. Biol. 2008; 8: 106
        • Hayashi K.
        • Ohta H.
        • Kurimoto K.
        • Aramaki S.
        • Saitou M.
        Reconstitution of the mouse germ cell specification pathway in culture by pluripotent stem cells.
        Cell. 2011; 146: 519-532
        • Hayashi K.
        • Ogushi S.
        • Kurimoto K.
        • Shimamoto S.
        • Ohta H.
        • Saitou M.
        Offspring from oocytes derived from in vitro primordial germ cell-like cells in mice.
        Science. 2012; 338: 971-975
        • Hayashi Y.
        • Saitou M.
        • Yamanaka S.
        Germline development from human pluripotent stem cells toward disease modeling of infertility.
        Fertil. Steril. 2012; 97: 1250-1259
        • Irie N.
        • Weinberger L.
        • Tang W.W.
        • Kobayashi T.
        • Viukov S.
        • Manor Y.S.
        • Dietmann S.
        • Hanna J.H.
        • Surani M.A.
        SOX17 is a critical specifier of human primordial germ cell fate.
        Cell. 2015; 160: 253-268
        • Kurimoto K.
        • Yabuta Y.
        • Ohinata Y.
        • Shigeta M.
        • Yamanaka K.
        • Saitou M.
        Complex genome-wide transcription dynamics orchestrated by Blimp1 for the specification of the germ cell lineage in mice.
        Genes Dev. 2008; 22: 1617-1635
        • Kurimoto K.
        • Yabuta Y.
        • Hayashi K.
        • Ohta H.
        • Kiyonari H.
        • Mitani T.
        • Moritoki Y.
        • Kohri K.
        • Kimura H.
        • Yamamoto T.
        • et al.
        Quantitative Dynamics of Chromatin Remodeling during Germ Cell Specification from Mouse Embryonic Stem Cells.
        Cell Stem Cell. 2015; 16: 517-532
        • Lin I.Y.
        • Chiu F.L.
        • Yeang C.H.
        • Chen H.F.
        • Chuang C.Y.
        • Yang S.Y.
        • Hou P.S.
        • Sintupisut N.
        • Ho H.N.
        • Kuo H.C.
        • Lin K.I.
        Suppression of the SOX2 neural effector gene by PRDM1 promotes human germ cell fate in embryonic stem cells.
        Stem Cell Reports. 2014; 2: 189-204
        • Nakagawa M.
        • Taniguchi Y.
        • Senda S.
        • Takizawa N.
        • Ichisaka T.
        • Asano K.
        • Morizane A.
        • Doi D.
        • Takahashi J.
        • Nishizawa M.
        • et al.
        A novel efficient feeder-free culture system for the derivation of human induced pluripotent stem cells.
        Sci. Rep. 2014; 4: 3594
        • Nakaki F.
        • Saitou M.
        PRDM14: a unique regulator for pluripotency and epigenetic reprogramming.
        Trends Biochem. Sci. 2014; 39: 289-298
        • Nakaki F.
        • Hayashi K.
        • Ohta H.
        • Kurimoto K.
        • Yabuta Y.
        • Saitou M.
        Induction of mouse germ-cell fate by transcription factors in vitro.
        Nature. 2013; 501: 222-226
        • Nakamura T.
        • Yabuta Y.
        • Okamoto I.
        • Aramaki S.
        • Yokobayashi S.
        • Kurimoto K.
        • Sekiguchi K.
        • Nakagawa M.
        • Yamamoto T.
        • Saitou M.
        SC3-seq: a method for highly parallel and quantitative measurement of single-cell gene expression.
        Nucleic Acids Res. 2015; 43: e60
        • Ohinata Y.
        • Payer B.
        • O’Carroll D.
        • Ancelin K.
        • Ono Y.
        • Sano M.
        • Barton S.C.
        • Obukhanych T.
        • Nussenzweig M.
        • Tarakhovsky A.
        • et al.
        Blimp1 is a critical determinant of the germ cell lineage in mice.
        Nature. 2005; 436: 207-213
        • Okita K.
        • Yamakawa T.
        • Matsumura Y.
        • Sato Y.
        • Amano N.
        • Watanabe A.
        • Goshima N.
        • Yamanaka S.
        An efficient nonviral method to generate integration-free human-induced pluripotent stem cells from cord blood and peripheral blood cells.
        Stem Cells. 2013; 31: 458-466
        • Pauls K.
        • Jager R.
        • Weber S.
        • Wardelmann E.
        • Koch A.
        • Buttner R.
        • Schorle H.
        Transcription factor AP-2gamma, a novel marker of gonocytes and seminomatous germ cell tumors.
        Int. J. Cancer. 2005; 115: 470-477
        • Perrett R.M.
        • Turnpenny L.
        • Eckert J.J.
        • O’Shea M.
        • Sonne S.B.
        • Cameron I.T.
        • Wilson D.I.
        • Rajpert-De Meyts E.
        • Hanley N.A.
        The early human germ cell lineage does not express SOX2 during in vivo development or upon in vitro culture.
        Biol. Reprod. 2008; 78: 852-858
        • Saitou M.
        • Yamaji M.
        Germ cell specification in mice: signaling, transcription regulation, and epigenetic consequences.
        Reproduction. 2010; 139: 931-942
        • Saitou M.
        • Barton S.C.
        • Surani M.A.
        A molecular programme for the specification of germ cell fate in mice.
        Nature. 2002; 418: 293-300
        • Sakuma T.
        • Hosoi S.
        • Woltjen K.
        • Suzuki K.
        • Kashiwagi K.
        • Wada H.
        • Ochiai H.
        • Miyamoto T.
        • Kawai N.
        • Sasakura Y.
        • et al.
        Efficient TALEN construction and evaluation methods for human cell and animal applications.
        Genes Cells. 2013; 18: 315-326
        • Suemori H.
        • Tada T.
        • Torii R.
        • Hosoi Y.
        • Kobayashi K.
        • Imahie H.
        • Kondo Y.
        • Iritani A.
        • Nakatsuji N.
        Establishment of embryonic stem cell lines from cynomolgus monkey blastocysts produced by IVF or ICSI.
        Dev. Dyn. 2001; 222: 273-279
        • Sugawa F.
        • Araúzo-Bravo M.J.
        • Yoon J.
        • Kim K.P.
        • Aramaki S.
        • Wu G.
        • Stehling M.
        • Psathaki O.E.
        • Hübner K.
        • Schöler H.R.
        Human primordial germ cell commitment in vitro associates with a unique PRDM14 expression profile.
        EMBO J. 2015; 34: 1009-1024
        • Takahashi K.
        • Tanabe K.
        • Ohnuki M.
        • Narita M.
        • Ichisaka T.
        • Tomoda K.
        • Yamanaka S.
        Induction of pluripotent stem cells from adult human fibroblasts by defined factors.
        Cell. 2007; 131: 861-872
        • Tesar P.J.
        • Chenoweth J.G.
        • Brook F.A.
        • Davies T.J.
        • Evans E.P.
        • Mack D.L.
        • Gardner R.L.
        • McKay R.D.
        New cell lines from mouse epiblast share defining features with human embryonic stem cells.
        Nature. 2007; 448: 196-199
        • Thomson J.A.
        • Itskovitz-Eldor J.
        • Shapiro S.S.
        • Waknitz M.A.
        • Swiergiel J.J.
        • Marshall V.S.
        • Jones J.M.
        Embryonic stem cell lines derived from human blastocysts.
        Science. 1998; 282: 1145-1147
        • Watanabe K.
        • Ueno M.
        • Kamiya D.
        • Nishiyama A.
        • Matsumura M.
        • Wataya T.
        • Takahashi J.B.
        • Nishikawa S.
        • Nishikawa S.
        • Muguruma K.
        • Sasai Y.
        A ROCK inhibitor permits survival of dissociated human embryonic stem cells.
        Nat. Biotechnol. 2007; 25: 681-686
        • Weber S.
        • Eckert D.
        • Nettersheim D.
        • Gillis A.J.
        • Schäfer S.
        • Kuckenberg P.
        • Ehlermann J.
        • Werling U.
        • Biermann K.
        • Looijenga L.H.
        • Schorle H.
        Critical function of AP-2 gamma/TCFAP2C in mouse embryonic germ cell maintenance.
        Biol. Reprod. 2010; 82: 214-223
        • Wilson J.F.
        • Erlandsson R.
        Sexing of human and other primate DNA.
        Biol. Chem. 1998; 379: 1287-1288
        • Yabuta Y.
        • Kurimoto K.
        • Ohinata Y.
        • Seki Y.
        • Saitou M.
        Gene expression dynamics during germline specification in mice identified by quantitative single-cell gene expression profiling.
        Biol. Reprod. 2006; 75: 705-716
        • Yamaji M.
        • Seki Y.
        • Kurimoto K.
        • Yabuta Y.
        • Yuasa M.
        • Shigeta M.
        • Yamanaka K.
        • Ohinata Y.
        • Saitou M.
        Critical function of Prdm14 for the establishment of the germ cell lineage in mice.
        Nat. Genet. 2008; 40: 1016-1022
        • Yamasaki J.
        • Iwatani C.
        • Tsuchiya H.
        • Okahara J.
        • Sankai T.
        • Torii R.
        Vitrification and transfer of cynomolgus monkey (Macaca fascicularis) embryos fertilized by intracytoplasmic sperm injection.
        Theriogenology. 2011; 76: 33-38
        • Ying Q.L.
        • Wray J.
        • Nichols J.
        • Batlle-Morera L.
        • Doble B.
        • Woodgett J.
        • Cohen P.
        • Smith A.
        The ground state of embryonic stem cell self-renewal.
        Nature. 2008; 453: 519-523