Research Article

In vitro culture of cynomolgus monkey embryos beyond early gastrulation

See allHide authors and affiliations

Science  15 Nov 2019:
Vol. 366, Issue 6467, eaax7890
DOI: 10.1126/science.aax7890

In vitro development of monkey embryos

Owing to technical and ethical limitations, the molecular and cellular mechanisms underlying primate gastrulation are far from clear (see the Perspective by Tam). Two independent studies used an in vitro culture system to study cynomolgus monkey embryo postimplantation development up to and beyond gastrulation (day 9 to day 20). Niu et al. observed in vivo morphogenetic events and used single-cell RNA sequencing and single-cell chromatin accessibility to study the distinct cell lineages in developing embryos. Ma et al. also observed that key events of in vivo early development were recapitulated in their system, and single-cell RNA-sequencing analysis revealed molecular signatures of postimplantation cell types. These systems will help elucidate the dynamics and regulation of gastrulation in primates, including possible relevance to human development.

Science, this issue p. eaaw5754, p. eaax7890; see also p. 798

Structured Abstract

INTRODUCTION

Gastrulation is a landmark event in development that involves a complex series of molecular, physical, and energetic remodeling transitions in early embryogenesis. Processes vary among species, leading to the diversity of animal forms on Earth. A dearth of primate embryo samples at the gastrulation stage has limited our understanding of this critical event in primates. Recently, human embryos were grown in culture for 12 to 13 days. Many governments and international organizations have recommended that human embryos should not be allowed to grow beyond 14 days in vitro. Therefore, it is expected that analysis of nonhuman primate embryo model systems will elucidate mechanisms underlying gastrulation and hopefully shed light on human development and the processes associated with defects and disease that arise during early development.

RATIONALE

Monkeys have long been considered a reliable animal model with which to study human physiological and pathological events because of their high degree of similarity to humans in both genomic and morphological characteristics. Therefore, we developed a system that supports the growth of cynomolgus monkey embryos in vitro for up to 20 days postfertilization (d.p.f.). Histological and immunofluorescent staining, combined with single-cell RNA-sequencing (RNA-seq) analysis, demonstrated that these in vitro–cultured (IVC) monkey embryos developed beyond early gastrulation and recapitulated key events of early primate in vivo postimplantation development.

RESULTS

We cultured cynomolgus monkey blastocysts with a mature blastocoel (d.p.f. 7 to 8) using the IVC system. At d.p.f. 13 to 14, a bilaminar disc–like structure appeared in ~27.7 ± 3.2% of the IVC embryos (n = 167, 26 experiments). At d.p.f. 15 to 16, the disc-like structure was clearly observable under the optical microscope. Some of the embryos successfully developed to d.p.f. 20 in vitro. Hematoxylin and eosin (H&E) and immunofluorescent staining confirmed that the IVC embryos recapitulated the major hallmarks of in vivo early postimplantation development: segregation of the epiblast (OCT4+, NANOG+) and hypoblast (GATA6+) lineages; formation of the amniotic and yolk sac cavities; appearance of the presumptive primordial germ cells (SOX17+, TFAP2C+, BLIMP1+); establishment of the anterior-posterior axis (asymmetric localization of OTX2+ cells); and gastrulation itself (gastrulating cells: T+/OCT4+, VIMENTIN+/T+/OCT4+). Furthermore, single-cell RNA-seq analysis showed that the monkey IVC embryos were similar to their in vivo counterparts in gene expression profiles and cell types, including presumptive parietal trophoblasts, extraembryonic mesenchyme cells, postimplantation early and late epiblasts (E-EPI and L-EPI, respectively), visceral and yolk sac endoderm, early primordial germ cells (E-PGC), early gastrulating cells (E-Gast), late gastrulating cells 1 and 2 (L-Gast1 and L-Gast2, respectively), early amnion cells (E-AM), and late amnion cells 1 and 2 (L-AM1 and L-AM2, respectively).

CONCLUSION

We have established an IVC system that can support the development of cynomolgus monkey embryos beyond early gastrulation in vitro. The IVC embryos recapitulate numerous key events of in vivo early postimplantation development of primate embryos. Single-cell RNA-seq analysis reveals the molecular signatures of several cell types of primate early postimplantation embryos, including amnion cells. The monkey IVC system provides a platform for future studies of molecular signatures and mechanisms of early embryogenesis that are specific to primates with potential relevance to human diseases that arise during early development.

Monkey embryos grow in vitro beyond early gastrulation.

IVC embryos were stained with antibodies for OCT4 (green) and GATA6 (red) (d.p.f. 13 to 14 or 19) or H&E (d.p.f. 20). Single-cell transcriptome analysis revealed the similarities among cell types in the in vitro and in vivo monkey early embryos and the developmental trajectory of epiblast derivatives. Scale bars, 100 μm.

Abstract

Gastrulation is a key event in embryonic development when the germ layers are specified and the basic animal body plan is established. The complexities of primate gastrulation remain a mystery because of the difficulties in accessing primate embryos at this stage. Here, we report the establishment of an in vitro culture (IVC) system that supports the continuous development of cynomolgus monkey blastocysts beyond early gastrulation up to 20 days after fertilization. The IVC embryos highly recapitulated the key events of in vivo early postimplantation development, including segregation of the epiblast and hypoblast, formation of the amniotic and yolk sac cavities, appearance of the primordial germ cells, and establishment of the anterior-posterior axis. Single-cell RNA-sequencing analyses of the IVC embryos provide information about lineage specification during primate early postimplantation development. This system provides a platform with which to explore the characteristics and mechanisms of early postimplantation embryogenesis in primates with possible conservation of cell movements and lineages in human embryogenesis.

The mammalian zygote develops to the blastocyst stage through several cleavage divisions in the oviduct before the blastocyst implants into the maternal uterus for further development (1). After implantation, gastrulation occurs, which is defined by the formation of the primitive streak (PS); the activation of a set of molecular markers, including T/Brachyury and Otx2 (2, 3); and a succession of morphogenetic rearrangements, resulting in the formation of the primary germ layers and a fundamental body plan (4, 5). Gastrulation has been very difficult to observe and analyze because it happens within the uterus. Therefore, in vitro culture (IVC) systems serve as a powerful tool to solve this problem. Mouse blastocysts have been cultivated in vitro up to the egg cylinder stage (610). In fact, by combining an IVC system with live imaging, anterior-posterior (A-P) axis formation and egg cylinder morphogenesis have been characterized in mouse embryos cultured in vitro (11, 12).

Despite the advances made with mouse embryo IVC systems, murine development events do not fully reflect primate embryogenesis because the morphogenesis of early postimplantation primate embryos is very different from that in rodent embryos (13, 14). Instead of forming a cuplike egg cylinder as in the mouse embryo, epiblast (EPI) cells in the primate embryo are flattened to form a bilaminar disc. Human embryos have been cultured in vitro for 12 to 13 days to initiate the formation of the EPI cells, amnion, and yolk sac (15, 16), and this system has been used to investigate the exit of EPI cells from the naïve pluripotent state in human embryos (17). Because of ethical regulations worldwide, culture of human embryos has to be stopped before 14 days postfertilization (d.p.f.), so human gastrulation-related events have not been explored in vitro (14, 18). Monkeys have been considered a reliable animal model with which to study human physiology and pathology because of their evolutionary, genomic, and physiological similarities to humans (13). The establishment of an IVC system for monkey postimplantation development would substantially improve our understanding of primate and human early embryonic development and related diseases. In particular, the blastocysts of the marmoset, a New World primate species, have been previously cultured for the study of postimplantation primate embryonic development. However, the resultant embryoids were not correctly organized (19). In the present study, we have established a system that allows cynomolgus monkey blastocysts to grow in vitro up to d.p.f. 20. Our IVC monkey embryos largely recapitulate critical events of in vivo early postimplantation development at the structural, cellular, and molecular levels and develop beyond early gastrulation.

In vitro culture of monkey embryos until d.p.f. 20

Cynomolgus monkey (Macaca fascicularis) blastocysts were obtained from an in vitro fertilization (IVF) procedure as previously reported (Fig. 1A and fig. S1, A and B) (20). Blastocysts at d.p.f. 7 to 8 with a mature blastocoel (>75% volume of the embryo) were selected for further culture (Fig. 1B and fig. S1, A and B). After removing the zona pellucida, the monkey blastocysts (n = 167) were cultured in conventional Matrigel-coated petri dishes containing media adapted from mouse blastocyst cultural media (Fig. 1, fig. S2, and tables S1 and S2) (7, 10). After 2 days of culture, at d.p.f. 9 to 10, ~67.3 ± 3.5% of embryos started to attach onto the dish, whereas the remainder attached onto the dish at d.p.f. 11 to 12 (table S3). The culture medium was changed daily after the embryos attached onto the dish (table S2). At d.p.f. 9 to 10, the embryos gradually enlarged, whereas the trophoblastic cells adjacent to the inner cell mass (ICM) firmly attached onto the Matrigel. At d.p.f. 11 to 12, the trophoblastic cells migrated onto the Matrigel surface and rapidly proliferated to form a flat and roughly circular sheet of outgrowth cells. By contrast, the ICM remained as a spherical and dense cell mass in the center of the outgrowth. At d.p.f. 13 to 14, a disclike structure appeared in ~27.7 ± 3.2% of the IVC embryos (n =167, 26 experiments; Fig. 1B, fig. S3, and table S3). At d.p.f. 15 to 16, the disclike structure was clearly observable under the optical microscope (Fig. 1B). Embryos that attached onto the dish at d.p.f. 8 to 9 had a higher chance of successfully forming disclike structures than those that attached onto the dish at d.p.f. 10 to 11 (35.4 ± 3.9%, n = 115, versus 5.9 ± 2.8%, n = 52; p < 0.01), suggesting that successful implantation during a critical time window might be important for further embryonic development. Although many IVC embryos started to show worsening morphology after d.p.f. 15, a portion of embryos with the bilaminar disc–like structure survived and developed to d.p.f. 16 to 17 (n = 14), and some even up to d.p.f. 19 to 20 (n = 4) (fig. S3).

Fig. 1 Establishment of the IVC system for monkey embryos.

(A) Time scheme of superovulation, IVF, and one-cell to gastrulation transition in the IVC system. rhFSH, recombinant human follicle-stimulating hormone; rhCG, recombinant human chorionic gonadotrophin. (B) Representative bright-field images of cynomolgus monkey embryos growing at the indicated stages of development. Blue dashed line indicates the ICM in d.p.f. 7 to 10 IVC embryos. Black dashed line indicates a portion of the expanded trophoblast cells in d.p.f. 11 to 14 IVC embryos. Scale bars, 100 μm. (C and D) H&E staining of the sections of IVC embryos with optically visible disclike structure at d.p.f. 15 [n = 1 (C)] and d.p.f. 20 [n = 1 (D)]. Scale bars, 100 μm.

IVC embryos with the disclike structures were fixed and analyzed by hematoxylin and eosin (H&E) staining for study of their overall morphology. The d.p.f. 15 IVC embryos contained tightly packed, columnar EPI-like cells (EPILCs) and epithelial amnion-like cells (AMLCs), together forming the amnion sac–like cavity (AMSLC) (Fig. 1C). Adjacent to EPILCs, we also observed a secondary yolk sac–like cavity (SYSLC), which was surrounded by irregular visceral endoderm-like cells (VELCs) and squamous parietal endoderm–like cells (PELCs) (Fig. 1C). These structures in the IVC embryos continued growing between d.p.f. 15 and 20, with the rostral end of the discs appearing thicker than the caudal end at d.p.f. 20 (Fig. 1D). Notably, the diameter of the embryonic discs increased nearly three times from ~0.166 ± 0.017 mm at d.p.f. 15 to 16 to ~0.469 ± 0.019 mm at d.p.f. 19 to 20, resembling the size changes of in vivo monkey embryos at similar stages (21, 22). These histological results suggest that the IVC embryos were similar to those of monkey embryos in vivo (22).

IVC monkey embryos recapitulate the key events of early postimplantation development

To determine whether the key cellular events of monkey early embryogenesis also occur in the IVC embryos, we performed immunofluorescent staining on IVC embryos at different time points using the following lineage markers: OCT4 and NANOG for EPI cells; GATA6 for hypoblast cells; and SOX17, TFAP2C, and BLIMP1 for primordial germ cells (PGCs) (23, 24). In d.p.f. 7 to 8 IVC embryos, the segregation of EPI cells and hypoblast progenitor cells was observable, and OCT4+/NANOG+ EPI cells formed a spherical cluster (Fig. 2A and fig. S4A). In d.p.f. 9 to 10 IVC embryos, OCT4+/NANOG+ EPI cells started to surround a proamniotic cavity–like lumen, whereas GATA6+ hypoblast cells clustered to one side of the EPI cells (Fig. 2A and fig. S4B). In d.p.f. 11 to 12 IVC embryos, the amniotic cavity–like lumen expanded while GATA6+ hypoblast cells were clustered to one side, marking the preparative step for yolk sac formation (Fig. 2A and fig. S4C). SOX17+/TFAP2C+ cells were detected at the amnion adjacent to the trophoblasts, beneath the EPI cells, or within the visceral endoderm (VE) (Fig. 2B and fig. S5). In d.p.f. 13 to 14 IVC embryos, EPI cells maintained robust OCT4 and NANOG expression, whereas amniotic cells expressed a relatively low level of OCT4 and NANOG. Simultaneously, the amniotic cavity further expanded, and the GATA6+ hypoblast cells formed the yolk sac cavity (Fig. 2A and fig. S4D). The amniotic cavity and the yolk sac cavity delineated a disc structure (Fig. 2A and figs. S4D, S5, and S7, A and B). At d.p.f. 14 to 16, SOX17+/TFAP2C+ and TFAP2C+/BLIMP1+ cells appeared beneath the posterior EPI cells or within the VE (Fig. 2, B and C, and figs. S5 and S6). Therefore, these IVC embryos are similar to monkey embryos in vivo (24). The mRNA expression of the lineage markers was further validated in d.p.f. 15 IVC embryos by reverse transcription polymerase chain reaction (RT-PCR) (Fig. 2D and table S4). Moreover, these d.p.f. 15 IVC embryos expressed other important ontogenic genes, such as SOX2, GATA4, and the mesodermal transcription factor and gastrulation regulator T/Brachyury (25) (Fig. 2D). Together, these data suggest that IVC embryos develop beyond the bilaminar disc stage and recapitulate the major hallmarks of monkey in vivo early postimplantation development: segregation of the EPI cells and hypoblasts, formation of the amniotic and yolk sac cavities, the appearance of presumptive PGCs, and probably the initiation of gastrulation.

Fig. 2 Recapitulation of monkey early postimplantation development in IVC embryos.

(A) Monkey blastocysts were cultured until d.p.f. 7 to 8 (n = 2), d.p.f. 9 to 10 (n = 2), d.p.f. 11 to 12 (n = 3), and d.p.f. 13 to 14 (n = 5) and fixed for whole-mount staining with antibodies for OCT4 (green), GATA6 (red), and NANOG (white). Shown are representative images reconstructed with the representative confocal z-images (d.p.f. 7 to 8, plane nos. 11 to 29; d.p.f. 9 to 10, plane nos. 12 to 39; d.p.f. 11 to 12, plane nos. 10 to 37; and d.p.f. 13 to 14, plane nos. 13 to 42). White dashed line and green dashed line indicate the proamniotic (d.p.f. 9 to 12) and amniotic (after d.p.f. 13) cavity and the yolk sac cavity, respectively. White line and red line indicate the EPI cells and amniotic cells in the d.p.f. 13 to 14 embryo, respectively. Scale bars, 50 μm. (B) Monkey blastocysts were cultured until d.p.f. 10 (n =1), d.p.f. 11 to 12 (n = 3), and d.p.f. 15 to 16 (n = 4) and then fixed for whole-mount staining with antibodies for SOX17 (green) and TFAP2C (red). The images were reconstructed with the representative confocal z-images (d.p.f. 10, plane nos. 13 to 38; d.p.f. 11 to 12, plane nos. 18 to 44; and d.p.f. 15 to 16, plane nos. 12 to 45). White arrowheads indicate SOX17+ and TFAP2C+ cells. White dashed line indicates the proamniotic and amniotic cavities at the indicated stages. Scale bars, 50 μm. (C) Monkey blastocysts were cultured until d.p.f. 14 (n = 1) and fixed for whole-mount staining with antibodies for BLIMP1 (green), TFAP2C (red), and OCT4 (white). Shown are representative images reconstructed with the representative confocal z-images (plane nos. 29 to 46). White arrowheads indicate BLIMP1+ and TFAP2C+ cells. White dashed line indicates the amnion cavity. Scale bar, 50 μm. (D) Monkey embryos (n = 2) were cultured until d.p.f. 15 and the discs in the embryos were harvested for RT-PCR of the indicated genes. PCR without template cDNAs was included as the negative control (NC).

Gastrulation is initiated in monkey IVC embryos

In mammals, gastrulation starts with PS formation and is marked by the migration of gastrulating cells (4). When we examined a representative d.p.f. 14 IVC embryo stained with OCT4 and GATA6, we observed that some OCT4+ cells appeared between the EPI cells and hypoblasts, especially at the presumptive caudal end of the disc, suggesting that these cells were gastrulating cells (Fig. 3, A and B; fig. S7, A and B; and movie S1). We reconstructed the images of this embryo at 90° orientation and found a putative PS at the caudal end of the embryo disc (Fig. 3, C and D, and fig. S7C). To further confirm the gastrulating cells, we performed coimmunostaining on the IVC embryos with antibodies against OCT4, GATA6, and T, the marker of gastrulating cells (23). At d.p.f. 11 to 12, T+/OCT4+ cells were detected at the amnion (Fig. 3E and fig. S8A). In d.p.f. 13 to 14 IVC embryos, the majority of T+/OCT4+ cells were still detected at the dorsal amnion, but a few began to appear between the EPI cells and the VE (Fig. 3E and fig. S8B). In d.p.f. 15 to 16 IVC embryos, only a few T+/OCT4+ cells were still located at the amnion; most were detected between the EPI cells and the VE (Fig. 3E, fig. S8C, and movie S2). At d.p.f. 19, the number of T+ cells increased considerably (Fig. 3E, fig. S8D, and movie S3). These observations suggest that the T+/OCT4+ cells might be gastrulating cells and could be a mixture of nascent gastrulating cells that originated from the EPI or cells that might have migrated from the dorsal amnion, consistent with previous reports (23, 24).

Fig. 3 Gastrulation in monkey IVC embryos.

(A) One representative d.p.f. 14 IVC embryo stained with antibodies for OCT4 (green) and GATA6 (red). Shown is an image reconstructed with the representative confocal z-images (plane nos. 206 to 220). Arrowheads indicate presumptive gastrulating cells. Scale bar, 60 μm. (B) Diagram based on the image in (A). AC, amniotic cavity. (C) The image in (A) was reconstructed at a 90° pitch. Scale bar, 60 μm. (D) Diagram based on the image in (C). (E) Monkey blastocysts were cultured until d.p.f. 11 to 12 (n = 2), d.p.f. 13 to 14 (n = 2), d.p.f. 15 to 16 (n = 2), and d.p.f. 19 (n = 1) and fixed for whole-mount staining with antibodies for GATA6 (green), T (red), and OCT4 (white). Shown are images reconstructed with representative confocal z-images (d.p.f. 11 to 12, plane nos. 24 to 48; d.p.f. 15 to 16, plane nos. 34 to 55; and d.p.f. 19, plane nos. 18 to 43). T+/OCT4+ cells are indicated with a white dashed line at the indicated stages. Scale bars, 50 μm.

Gastrulation is also characterized by the epithelial–mesenchymal transition (EMT) of gastrulating cells and the establishment of an A-P axis (4). We costained the IVC embryos with antibodies for T, OCT4, and VIMENTIN, a marker for EMT (26). The results showed that some T+/OCT4+ cells expressed VIMENTIN in d.p.f. 13 to 14 and d.p.f. 16 IVC embryos (Fig. 4A and fig. S9). We also stained the IVC embryos with antibodies for OTX2 and EOMES, two important transcription factors for mouse germ layer specification (27, 28). The protein expression of OTX2 and EOMES was observed in the visceral endoderm of d.p.f. 13 IVC embryos (Fig. 4, B and C, and fig. S10, A to C). Notably, OTX2 was expressed strongly at the anterior region but weakly at the posterior region in the visceral endoderm of d.p.f. 15 to 16 IVC embryos (Fig. 4B and fig. S10B). This asymmetrical distribution of OTX2 cells in d.p.f. 15 to 16 IVC embryos indicates the establishment of an A-P axis. We also observed neural crest–like, forebrain-like, and neural groove–like structures in d.p.f. 19 IVC embryos (Fig. 4D; fig. S11, A and B; and movies S3 and S4) (29), similar to human embryos at embryonic days 17 (E17) to E19 [Carnegie stage 8 (29, 30)]. Taken together, these lines of evidence suggest that the IVC monkey embryos have developed beyond early gastrulation.

Fig. 4 EMT and the establishment of the A-P axis in IVC embryos.

(A) Monkey d.p.f. 13 to 14 (n = 2) and d.p.f. 16 (n = 1) IVC embryos with the disc were stained with antibodies for T (green), VIMENTIN (red), and OCT4 (white). Shown are images reconstructed with representative confocal z-images (d.p.f. 13 to 14, plane nos. 19 to 27; d.p.f. 16, plane nos. 206 to 223). White dashed line indicates the amniotic cavity at the indicated stages. Scale bars, 50 μm. (B) Monkey blastocysts were cultured until d.p.f. 13 (n = 1) and d.p.f. 15 to 16 (n = 2) and fixed for whole-mount staining with antibodies for OTX2 (green) and OCT4 (white). White and yellow arrowheads indicate the anterior and posterior region in VE, respectively. Shown are images reconstructed with representative confocal z-images (d.p.f. 13, plane nos. 34 to 45; d.p.f. 15 to 16, plane nos. 58 to 68). White dashed line indicates the amniotic cavity at the indicated stages. Scale bars, 50 μm. (C) Monkey blastocyst was cultured until d.p.f. 13 (n = 1) and fixed for whole-mount staining with antibodies for EOMES (green) and OCT4 (white). Shown are images reconstructed with representative confocal z-images (plane nos. 28 to 45). White dashed line indicates the amniotic cavity at the indicated stages. Scale bar, 50 μm. (D) Images from Fig. 3E at d.p.f. 19 reconstructed by Imaris 9.0.2 with representative whole z-stack images. The 3D image was rotated 90° around the y-axis and then clockwise rotated at 45° orientation around the z-axis (movie S4). NPFLS, neural plate of forebrain-like structure; NGLS, neural groove-like structure; Meso, mesoderm. Scale bar, 100 μm.

IVC monkey embryos’ molecular signatures are similar to in vivo counterparts

Previous studies using single-cell RNA-sequencing (RNA-seq) analysis delineated the cells of the monkey early postimplantation embryos as postimplantation early or late EPI (postE-EPI or postL-EPI, respectively); gastrulating cells 1, 2a, and 2b (Gast1, 2a and 2b); visceral/yolk sac endoderm (VE/YE); extraembryonic mesenchyme (EXMC); postimplantation parietal trophectoderm (paTE); and early PGC (E-PGC) (23, 24). To identify the various cell types of the IVC monkey embryos and to understand their gene expression dynamics, we collected 2016 single cells from nine IVC embryos at six developmental stages (d.p.f. 11, n = 1; d.p.f. 12, n = 2; d.p.f. 13, n = 2; d.p.f. 14, n = 2; d.p.f. 16, n = 1; and d.p.f. 17, n = 1) and performed single-cell RNA-seq using a modified Smart-seq2 technique (fig. S12, A and B) (31). Among these 2016 cells, 1453 single cells with expression of >2000 genes were selected for further analyses (fig. S12C and table S5). We compared these single cells with those from their in vivo counterparts (23). t-Distributed stochastic neighbor embedding (t-SNE) analysis revealed that cells from IVC and in vivo embryos were clustered into six cell types as described previously (23), including postE-EPI, postL-EPI, VE/YE, Gast, EXMC, and paTE (Fig. 5A and fig. S13, A and B).

Fig. 5 Classification of key cell types in the IVC embryos by single-cell RNA-seq analysis.

(A) Visualization of major classes of cells from the IVC embryos (d.p.f. 11, 12, 13, 14, 16, and 17; red, n = 1453) and the archived in vivo embryos [E13, E14, E16, and E17; cyan, n = 211; (23)] by t-SNE analysis. AM/Gast, amnion/gastrulating cells. (B) Key cell types identified from IVC and in vivo embryos at the indicated development stages by t-SNE analysis. After excluding paTE and EXMC, 10 clusters were identified and marked with different colors at specific stages [in vitro, n = 588; in vivo, n = 167; (23)]. (C) Heat map of DEGs in the 10 clusters of the postimplantation embryos. The colors from magenta to yellow indicate relative expression levels from low to high. Right, representative genes and key gene ontology (GO) enrichment analysis results (table S7).

After excluding the paTE and EXMC cells, the 755 cells [IVC, n = 588; in vivo, n = 167 (23)] were reclassified into 10 clusters, including the previously annotated postE-EPI, postL-EPI, VE/YE, E-PGC, and Gast (23, 24). We annotated gastrulating cells into three distinct clusters, the early gastrulating cells (E-Gast) and late gastrulating cells 1 and 2 (L-Gast1 and L-Gast2). We also annotated early amnion cells (E-AM) and late amnion cells 1 and 2 (L-AM1 and L-AM2) (Fig. 5B, fig. S14, and tables S5 to S7). The cell clusters that appeared in d.p.f. 11 to 14 IVC embryos were similar to those in the E13 to E14 embryos in vivo, whereas the clusters in d.p.f. 16 to 17 IVC embryos were similar to those in E16 to E17 embryos in vivo (Fig. 5B). We annotated these cell clusters on the basis of their differentially expressed genes (DEGs) and the timing of their appearance during development (early stages: in vivo E13 to E14/in vitro d.p.f. 11 to 14; late stages: in vivo E16 to E17/in vitro d.p.f. 16 to 17; Fig. 5, B and C).

The postE-EPI, postL-EPI, VE/YE, and E-PGC highly expressed genes that were enriched in numerous developmental pathways, such as stem cell population maintenance in postE-EPI, neuron differentiation in postL-EPI, extracellular structure organization in VE/YE, and mRNA metabolic processes in E-PGC (Figs. 5C and 6A and figs. S15 and S16). The representative genes that were validated by our immunostaining experiments in these cell clusters included NANOG and OCT4 in postE-EPI, OTX2 in VE/YE, BLIMP1 and TFAP2C in E-PGC, and SOX17 in VE/YE and E-PGC (Figs. 2, A and B, and 4B and fig. S5). We also observed high or low levels of marker gene expression (OTX2, GATA6, and IHH) in the VE/YE cluster, suggesting early or later VE/YE cell identities (Fig. 6A and fig. S16B).

Fig. 6 Gene expression analysis of single cells from the postimplantation embryos.

(A to F) t-SNE plots showing the expression patterns of ontogenic markers in VE/YE (A), E-Gast (B), L-Gast1 (C), L-Gast2 (D), E-AM (E), and L-AM (F) cells (red dashed lines) from the IVC and in vivo embryos. (G) PCA on the single cells of postimplantation embryos [IVC, 588 cells; in vivo, 167 cells (23)]. (H) PCA on the single cells of postimplantation embryos) after excluding VE/YE [IVC, 487 cells; in vivo, 149 cells (23)]. (I) PAGA analysis on 636 single cells from IVC and in vivo embryos. The boldness of the line indicates the degree of the relationship between clusters.

Our platform and the number of cells involved in this study allowed us to reannotate the gastrulating cells as E-Gast, L-Gast1, and L-Gast2. E-Gast cells were mostly from d.p.f. 11 to 14 IVC embryos, whereas L-Gast1 and L-Gast2 were mainly from d.p.f. 16 to 17 IVC embryos (Fig. 5B and fig. S14, A and B). E-Gast cells expressed CDX1/2 and T and were enriched with genes in the Nodal signaling pathway, such as NODAL and FOXH1 (Figs. 5C and 6B and fig. S17A). L-Gast1 cells were enriched for genes in the gastrulation signatures, such as T, MIXL1, and EOMES (Figs. 5C and 6C and fig. S17B). L-Gast2 cells strongly expressed mesodermal development–related transcription factors such as FOXF1, HAND1, and MESP1 (Figs. 5C and 6D and fig. S17C) (32, 33). The expression levels of gastrulating cell markers (T, EOMES, NODAL, LEF1, and MIXL1) were different in these clusters of cells (Fig. 6, A to C, and fig. S17, A and B). In addition, both in vitro and in vivo postE-EPI, postL-EPI, E-Gast, L-Gast, and VE/YE cells expressed representative genes in the Wnt signaling pathway in heterogeneous patterns, probably reflecting the shared importance of Wnt signaling in many subsets of cells during gastrulation (fig. S18) (4). The heterogeneous expression patterns of the gastrulation markers and key regulators of important developmental pathways in these clusters further suggest the progressive establishment of an A-P axis in the IVC embryos (2, 11).

We also observed three clusters of cells exhibiting OCT4+/BMP4+/SOX2low/– (fig. S19), representing potential monkey amnion cells (24). We annotated these three clusters as E-AM cells, which primarily appeared in d.p.f. 11 to 14 IVC embryos and E13 to E14 embryos in vivo, and late amnion cells 1 and 2 (L-AM1 or L-AM2), which were primarily present in d.p.f. 16 to 17 IVC embryos and E16 to E17 embryos in vivo (Fig. 5B and fig. S14) (24). The E-AM cells highly expressed embryonic genes such as HOXD3, WNT6, and TFAP2A (Figs. 5C and 6E and fig. S20), whereas the L-AM1 and L-AM2 cells highly expressed genes involved in biological adhesion, such as SPNS2, PDZRN4, and NGF (Figs. 5C and 6F and fig. S20).

We next examined the transitions between cell lineages. Principal component analysis (PCA) indicated a unidirectional progression in the EPI developmental trajectory, suggesting an orderly and progressive specification of EPI derivatives (Fig. 6G). By contrast, PCA indicated a scattered distribution of VE/YE cells (Fig. 6G), suggesting complex and heterogeneous origins of VE/YE cells. To clarify our understanding of the EPI developmental trajectory of EPI derivatives, we excluded the VE/YE cells and performed PCA and partition-based graph abstraction (PAGA) analyses. These analyses revealed a close relationship between E-Gast and postE-EPI or E-AM cells, as well as a close similarity between L-Gast1 and postL-EPI or L-AM1 cells, supporting the idea that gastrulating cells originate from EPI and/or amnion cells (Fig. 6, H and I). These relationships were well inosculated with force-directed graph-drawing analysis (fig. S21). Single-cell analysis of genome-wide gene expression profiles suggests that the IVC embryos have cell types with gene signatures that are very similar to their in vivo counterparts, further supporting our morphological and cellular observations that our IVC monkey embryos could develop beyond early gastrulation.

Concluding remarks

In this study, we developed an IVC system that supports the growth and development of cynomolgus monkey embryos for up to d.p.f. 20 without maternal contribution. We provide several lines of evidence to demonstrate that these IVC embryos recapitulate most of the key events of early postimplantation development as seen in their in vivo counterparts: (i) formation of a clear, bilaminar, disclike structure delineated by the amniotic cavity and yolk sac cavity; (ii) timely occurrence of several hallmark events, including segregation of the EPI and hypoblast, formation of the amniotic and yolk sac cavities, specification of the presumptive PGCs, and the establishment of the A-P axis; (iii) initiation of gastrulation, as demonstrated by the formation of the PS, the appearance of gastrulating cells, and the EMT process of gastrulating cells; and (iv) similarities in cell types and gene expression profiles between IVC embryos and their in vivo counterparts at the single-cell and genome-wide levels.

We analyzed the IVC embryos by single-cell RNA-seq and obtained insights into primate early embryogenesis. We identified primate amnion cells and characterized their molecular signatures at different developmental stages. We also reannotated primate gastrulating cells as E-Gast, L-Gast1, and L-Gast2 cells to represent the shifting identities of these dynamic cells during primate early embryogenesis.

This study provides a platform technology and a data resource with which to further investigate the characteristics and key regulators of early postimplantation embryogenesis in primates. In combination with CRISPR-Cas9–mediated gene editing and cell lineage tracing, the monkey IVC system could help to elucidate the mysterious dynamics of early embryonic development in primates, with possible relevance to human development.

Summary of materials and methods

The experiments on the cynomolgus monkey and the mouse were performed according to the guidance of the ethics committees of the Institute of Zoology and Kunming Institute of Zoology, Chinese Academy of Sciences (CAS). An in vitro culture system for monkey blastocysts based on the culture system for mouse postimplantation embryos was established. H&E staining, RT-PCR, and whole-mount embryo immunostaining were used to examine the histology, key cellular events, and cell lineages of early postimplantation development in the IVC monkey embryos. To investigate the profiles of gene expression and cell types, single cells were isolated from the IVC monkey embryos. The libraries of single-cell transcriptomes were constructed using the Smart-seq2 method and sequenced with the Hiseq4000 system. The single-cell data were analyzed by the Seurat version 3.0 R package and the Scanpy version 1.4.4 Python package. Details of the materials and methods are available in the supplementary materials.

Supplementary Materials

science.sciencemag.org/content/366/6467/eaax7890/suppl/DC1

Materials and Methods

Figs. S1 to S21

Tables S1 to S7

References (3443)

Movies S1 to S4

References and Notes

Acknowledgments: We are grateful to Q. Zhou, W. Li, F. Gao, and N. Jing at the CAS, and N. Plachta at A*STAR in Singapore, for helpful discussions; F. Tang and X. Fan at Peking University for helping to establish the single-cell RNA-seq platform; S. Ng (CAS) for critical reading of the manuscript; and S. Li and X. Zhu of the imaging platform of CAS for their outstanding support. Funding: This work was supported by the Strategic Priority Research Program of the CAS (XDA16020700), the National Key R&D Program of China (2018YFC1004500, 2017YFC1001400), the National Natural Science Foundation of China (31571533, 31590832), and the exchange program of State Key Laboratory of Genetic Resources and Evolution, Kunming Institute of Zoology, CAS (GREKF17-10). Author contributions: H.M. designed and developed the monkey embryo culture system, drafted the manuscript, and performed data analyses. J.Z. and H.W. performed embryo culture experiments, whole-mount staining of IVC monkey embryos, imaging, and analyses. X.J. performed single-cell experiments and analyzed data. Y.X., L.W., and Z.-A.Z. optimized the culture system for mouse blastocysts. L.W., X.H., and B.Z. contributed to monkey maintenance. L.L., P.Z., and H.W. initiated and supervised the study, analyzed data, and wrote the manuscript. All authors commented on the manuscript. Competing interests: The authors declare no competing interests. Data and materials availability: All data are available in the main text or the supplementary materials or from the corresponding authors upon request.

Stay Connected to Science

Navigate This Article