Article

Chimeric contribution of human extended pluripotent stem cells to monkey embryos ex vivo

Graphical abstract

Highlights

d Generation of human-monkey chimeric embryos ex vivo with

hEPSCs

d hEPSCs differentiated into hypoblast and epiblast lineages

d scRNA-seq analyses revealed developmental trajectories of

human and monkey cells

d The approach may allow for enhancing chimerism between

evolutionarily distant species

Tan et al., 2021, Cell 184, 2020–2032 April 15, 2021 ª 2021 Elsevier Inc. https://doi.org/10.1016/j.cell.2021.03.020

Authors

Tao Tan, Jun Wu, Chenyang Si, …,

Weizhi Ji, Yuyu Niu,

Juan Carlos Izpisua Belmonte

Correspondence [email protected] (T.T.), [email protected] (J.W.), [email protected] (W.J.), [email protected] (Y.N.), [email protected] (J.C.I.B.)

In brief

Human cells, in the form of extended

pluripotent stem cells, have the ability to

contribute to both embryonic and extra-

embryonic lineages in ex-vivo-cultured

monkey embryos.

ll

ll

Article

Chimeric contribution of human extended pluripotent stem cells to monkey embryos ex vivo Tao Tan,1,4,* Jun Wu,2,4,5,* Chenyang Si,1,4 Shaoxing Dai,1,4 Youyue Zhang,1,4 Nianqin Sun,1 E Zhang,1 Honglian Shao,1

Wei Si,1 Pengpeng Yang,1 Hong Wang,1 Zhenzhen Chen,1 Ran Zhu,1 Yu Kang,1 Reyna Hernandez-Benitez,2

Llanos Martinez Martinez,3 Estrella Nuñez Delicado,3 W. Travis Berggren,2 May Schwarz,2 Zongyong Ai,1 Tianqing Li,1

Concepcion Rodriguez Esteban,2 Weizhi Ji,1,* Yuyu Niu,1,* and Juan Carlos Izpisua Belmonte2,6,* 1State Key Laboratory of Primate Biomedical Research, Institute of Primate Translational Medicine, Kunming University of Science and Technology, Kunming, Yunnan 650500, China 2Gene Expression Laboratory, Salk Institute for Biological Studies, La Jolla, CA 92037, USA 3Universidad Católica San Antonio de Murcia (UCAM), Campus de los Jerónimos, No 135 12, Guadalupe 30107, Spain 4These authors contributed equally 5Present address: Department of Molecular Biology, University of Texas Southwestern Medical Center, Dallas, TX 75390, USA 6Lead contact

*Correspondence: [email protected] (T.T.), [email protected] (J.W.), [email protected] (W.J.), [email protected] (Y.N.), [email protected] (J. C.I.B.)

https://doi.org/10.1016/j.cell.2021.03.020

SUMMARY

Interspecies chimera formation with human pluripotent stem cells (hPSCs) represents a necessary alternative to evaluate hPSC pluripotency in vivo and might constitute a promising strategy for various regenerative medicine applications, including the generation of organs and tissues for transplantation. Studies using mouse and pig embryos suggest that hPSCs do not robustly contribute to chimera formation in species evolutionarily distant to humans. We studied the chimeric competency of human extended pluripotent stem cells (hEPSCs) in cynomolgus monkey (Macaca fascicularis) embryos cultured ex vivo. We demonstrate that hEPSCs survived, proliferated, and generated several peri- and early post-implantation cell lineages in- side monkey embryos. We also uncovered signaling events underlying interspecific crosstalk that may help shape the unique developmental trajectories of human and monkey cells within chimeric embryos. These re- sults may help to better understand early human development and primate evolution and develop strategies to improve human chimerism in evolutionarily distant species.

INTRODUCTION

Pluripotent stem cells (PSCs) are capable of indefinite self-

renewal in culture and generating all adult cell types (De Los An-

geles et al., 2015; Hackett and Surani, 2014; Rossant and Tam,

2017; Wu and Izpisua Belmonte, 2016). PSCs have recently been

harnessed for interspecies organogenesis via blastocyst

complementation, a technique that holds potential to provide

large quantities of in-vivo-generated human cells, tissues, and

organs for regenerative medicine applications, including organ

transplantation (Suchy and Nakauchi, 2018; Wu et al., 2016).

One of the requirements for successful interspecies blastocyst

complementation with human PSCs (hPSCs) is their ability to

contribute to chimera formation. The chimeric competency of

hPSCs has been systematically tested in several animal species

(Wu et al., 2016), but despite sustained efforts from different lab-

oratories, the general consensus is that hPSCs do not consis-

tently and robustly contribute to chimera formation when the

host animal has a high evolutionary distance from humans

(e.g., mice and pigs; Wu et al., 2016, 2017). This is the case

2020 Cell 184, 2020–2032, April 15, 2021 ª 2021 Elsevier Inc.

even when human cell apoptosis is inhibited (Das et al., 2020;

Huang et al., 2018; Wang et al., 2018). Xenogeneic barriers be-

tween hPSCs and evolutionarily distant host animal species

have been suggested to account for limited chimerism (Wu

et al., 2016, 2017), though the use of hPSCs for chimera studies

in host species evolutionarily close to humans remains unex-

plored to date.

Cultured PSCs reflect the continuum of pluripotency that is

seen in vivo, and different cell culture formulations result in

distinct pluripotency states in vitro (Morgani et al., 2017; Smith,

2017; Weinberger et al., 2016). hPSCs in different pluripotency

states exhibit distinct transcriptional, epigenetic, and metabolic

features. They also differ in their chimeric potential when intro-

duced into animal embryos (De Los Angeles, 2019; Harvey

et al., 2019; Zhang et al., 2018). Recently, we and others identi-

fied human extended PSCs (hEPSCs) that demonstrated

improved chimeric capability in mouse conceptuses (Gao

et al., 2019; Yang et al., 2017b). To date, however, the chimeric

competency of hEPSCs has not been determined in other

species.

ll Article

Leveraging a recently developed culture system that enables

cynomolgus monkey (monkey for short) embryos to develop

up to 20 days ex vivo (Ma et al., 2019; Niu et al., 2019), we per-

formed microinjection of hEPSCs into monkey blastocysts and

examined their contribution to cultured monkey embryos at

different time points (Figure 1A). We found that hEPSCs could

integrate into the inner cell masses (ICMs) of late monkey blasto-

cysts and contributed to both embryonic and extra-embryonic

lineages in peri- and early post-implantation stages during pro-

longed embryo culture. We also determined the differentiation

trajectory of hEPSCs within cultured monkey embryos by sin-

gle-cell RNA sequencing (scRNA-seq) analysis.

RESULTS

Generation of human-monkey chimeric blastocysts in vitro

To determine the chimera competency of hPSCs in a closely

related non-human primate species, we used a well-character-

ized hEPSC line generated by cellular reprogramming, iPS1-

EPSCs, which demonstrated improved chimerism in embryonic

day 10.5 (E10.5) mouse conceptuses over other reported hPSCs

(Yang et al., 2017b). Consistent with the previous report, iPS1-

EPSCs exhibited a dome-shaped colony morphology and ex-

pressed the core pluripotency transcription factors OCT4,

NANOG, and SOX2 (Figure S1A). To generate human-monkey

chimeric embryos, early blastocysts from cynomolgus monkeys

(6 days post-fertilization [d.p.f.6]) were injected with 25 iPS1-

EPSCs labeled with tdTomato (TD). Injected embryos were first

cultured to the late blastocyst stage (d.p.f.7) for analysis. The

proliferation of hEPSCs within monkey blastocysts was evident

(Video S1). In total, TD+ iPS1-EPSCs were detected in all

d.p.f.7 monkey blastocysts (100%, n = 132) (Figure S1B).

Chimeric contribution of hEPSCs to peri- and post- implantation monkey embryos We next took advantage of a recently established prolonged em-

bryo culture system that supports ex vivo primate (human and

monkey) embryogenesis to the gastrulation stage (Deglincerti

et al., 2016; Ma et al., 2019; Niu et al., 2019; Shahbazi et al.,

2016; Xiang et al., 2020; Zhou et al., 2019). In this embryo culture

system, the zona pellucida is removed and the denuded blasto-

cysts are allowed to attach to the culture dish for further develop-

ment. We used this system to trace the fate of hEPSCs in

monkey embryos at peri- and post-implantation stages. Similar

to noninjected controls (92.31%, n = 104) (Niu et al., 2019),

most embryos injected with hEPSCs attached at approximately

d.p.f.10 (92.79%, n = 111). After attachment, the injected em-

bryos continued to grow, and an embryonic disc became visible

at approximately d.p.f.11 as seen with controls (Figure 1B). TD+

human cells were found in the embryonic disc of more than half

of the injected embryos at d.p.f. 9, but this ratio progressively

declined at approximately one-third by d.p.f.13 (Figure 1C). To

determine whether introduced hEPSCs may affect embryo

development, we evaluated and compared the developmental

status of injected and control embryos (Niu et al., 2019). We

found that the developmental ratios of injected embryos were

slightly lower than that of controls (Figure 1D). Furthermore,

similar to controls, the developmental ratios of injected embryos

dropped sharply at approximately d.p.f.15, which may reflect the

limitation of the 2D attachment embryo culture system

(Figure 1D).

To study the developmental potential of hEPSCs in monkey

embryos, we performed immunofluorescence (IF) studies using

antibodies specific for several embryonic and extra-embryonic

lineages. Analyses were performed between d.p.f.9 and

d.p.f.19. At the peri-implantation stage (d.p.f.9), on average,

10.2 ± 7.2 (n = 9) iPS1-EPSCs were found incorporated into

the ICM (the number of TD+OCT4+ cells found within or close

to NANOG+ or OCT4+ monkey cells) (Figure 2A). Although these

cells expressed OCT4, only a few of them expressed NANOG

(Figure S1C). Supporting their epiblast (EPI)-like identity, we

did not observe GATA6 (a hypoblast [HYP] marker) expression

in these cells (Figure S1C). In addition, TD+GATA4+ HYP-like

cells were also detected (Figure 2B). In contrast to the results re-

ported in mice (Yang et al., 2017b), only a few iPS1-EPSCs were

detected in the trophectoderm (TE) layer of monkey blastocysts

and expressed TE (TE or trophoblast) marker genes (e.g.,

TFAP2C and CK7) (Figure 2C).

At d.p.f.11, TD+OCT4+cells were also detected in the EPI layer

of monkey embryos, and these cells rarely expressed T+ (also

known as Brachyury), a marker of gastrulation. By contrast, T+

monkey cells were detected at the dorsal amnion of the embryos

(Figure 2D). In addition, TD+ human cells expressing a HYP

marker, platelet-derived growth factor receptor-alpha

(PDGFRa), were found intermingled within monkey HYP cells

(Figure 2E). OCT4+ human cells expressing COL6A1, a marker

of extra-embryonic mesenchyme cells (EXMCs), were found

outside of the EPI layer, suggesting ongoing differentiation of

hEPSCs toward EXMCs (Figure S1D). At d.p.f.13, hEPSCs

were detected beneath the EPI layer and expressed an endo-

derm marker, SOX17, suggesting that they have initiated gastru-

lation (Figure 2F). Interestingly, we found that from d.p.f.13

onward human cells tended to group together and segregate

from the monkey EPI layer. These human cells appeared to un-

dergo differentiation into gastrulating cells as evidenced by the

gained expression of OTX2 (Martyn et al., 2018; Vincent et al.,

2003) while maintaining OCT4 expression (Figure S1E). Overall,

we found that hEPSCs exhibited reasonable contribution to the

EPI (with the highest contribution of 7.08% observed at

d.p.f.15), relatively lower contribution to the HYP (with the high-

est contribution of 4.96% observed at d.p.f.19), and limited

contribution to the TE in peri- and post-implantation monkey em-

bryos (Figure 2G).

Transcriptional landscape of human-monkey chimeric embryos To further delineate the developmental trajectory of human-

monkey chimeric embryos, scRNA-seq analysis was carried

out to profile the transcriptomes of human and monkey cells at

different developmental stages. Following embryo dissociation,

single human (TD+) and monkey (TD�) cells were manually collected using fluorescence microscopy and subjected to

scRNA-seq. In total, we sequenced 227 human and 302 monkey

cells isolated from chimeric embryos at different time points dur-

ing ex vivo culture (d.p.f.9–d.p.f.17; Table S2). TD expression and

Cell 184, 2020–2032, April 15, 2021 2021

Figure 1. Generation and developmental capability of human-monkey ex-vivo chimeras

(A) Schematic of the generation and analyses of chimeric embryos derived from blastocyst injection of hEPSCs (created with BioRender.com). hEPSCs, human

extended pluripotent stem cells; EPI, epiblast; HYP, hypoblast; TE, trophectoderm; EXMC, extra-embryonic mesenchyme cell; GAS, gastrulating cell; IF,

immunofluorescence.

(B) Representative bright-field images of hEPSC-injected monkey embryos cultured in vitro until d.p.f.19 (n = 111 embryos for d.p.f.9; n = 91 embryos for d.p.f.11;

n = 60 embryos for d.p.f.13; n = 38 embryos for d.p.f.15; n = 12 embryos for d.p.f.17 and n = 3 embryos for d.p.f.19). Scale bar, 100 mm. Yellow dotted lines indicate

ICM (d.p.f.9) or embryonic disc (d.p.f.11 to d.p.f.19).

(C) Histogram showing the percentages of host monkey embryos containing human cells within ICM or embryonic disc (yellow dotted lines in B).

(D) Dynamics of developmental ratios of chimeric (n = 126, d.p.f.8) and non-chimeric monkey embryos (n = 104, data from Niu et al., 2019). Embryos without clear

embryonic disc structure and/or appear dead were excluded from the analysis.

See also Figure S1 and Video S1.

ll

2022 Cell 184, 2020–2032, April 15, 2021

Article

Figure 2. hEPSCs contribute to chimera

formation in peri- and post-implantation

monkey embryos

(A) Representative IF images showing integration

of TD-positive hEPSCs into ICM of host monkey

embryos at d.p.f.9 (n = 7). The embryos were

stained for OCT4 (green) and NANOG (gray). Scale

bar, 250 mm. Bottom, enlargements of the insert

(white dotted line) in the top panel. Arrows indicate

TD-positive hEPSCs expressing OCT4 and

NANOG. Yellow dotted line indicates ICM. Scale

bar, 50 mm.

(B) Representative IF images showing hEPSCs

differentiated into HYP-like cells within host

monkey embryos at d.p.f.9. The embryos were

stained for GATA4 (gray) and NANOG (green).

Scale bar, 250 mm. Bottom, enlargements of the

insert (white dotted line) in the top panel. Arrow

indicates a TD-positive hEPSC expressing

GATA4. Yellow dotted line indicates ICM. Scale

bar, 100 mm.

(C) Representative IF images showing integration

of TD-positive hEPSCs into TE of host monkey

embryos at d.p.f.9 (n = 3). The embryos were

stained for TFAP2C (gray) and CK7 (green). Scale

bar, 100 mm. Bottom, enlargements of the insert

(white dotted line) in the top panel. Arrow indicates

a TD-positive hEPSC expressing TFAP2C and

CK7. Scale bar, 50 mm.

(D) Representative IF images showing incorporation

of hEPSCs into EPI of host monkey embryo at

d.p.f.11 (n = 3). The embryos were stained for T

(gray) and OCT4 (green). Scale bar, 250 mm. Bot-

tom: enlargements of inserts (white dotted line) in

the top panel. Notably, hEPSCs rarely express T

(arrow), a marker of mesoderm, whereas the

expression of T is detected in monkey cells

(arrowhead). Yellow dotted line indicates EPI. Scale

bar, 25 mm.

(E) Expression of HYP marker, PDGFRa, in hEPSCs

at d.p.f.11 (n = 2). The embryos were stained for

PDGFRa (green). Scale bar, 50 mm. Bottom: en-

largements of the insert (white dotted line) in the top

panel. Arrow indicates a tdTomato-positive hEPSC

expressing PDGFRa. Yellow dotted line indicates

EPI. Scale bar, 10 mm.

(F) IF images of sections of monkey embryos at

d.p.f.13 (n = 5) staining for SOX17 (green).

Arrows indicate tdTomato-positive hEPSCs expressing SOX17. Scale bar, 50 mm. Yellow dotted line indicates EPI.

(G) Levels of chimerism of hEPSCs within EPI, HYP, and TE. EPI cells expressed only OCT4, and HYP cells expressed GATA6 and/or GATA4, whereas TE expressed

CK7 (a total of 25 embryos and 17,938 cells were analyzed). EPI, epiblast; HYP, hypoblast; TE, trophectoderm; TD, tdTomato; DAPI, 40,6-diamidino-2-phenylindole. See also Figure S1.

ll Article

the ratio of reads mapped to the human or cynomolgus monkey

genomes were used to further confirm each cell’s species of

origin (Figures S2A and S2B). After stringent filtering, 200 human

and 272 monkey cells were used for further analyses (Table S2).

On average, 9,798 genes (transcripts per million [TPM] > 0) and

27,936,953 reads were detected per cell. There was no statistical

difference in the number of genes and reads detected between

human and monkey cells (Figure S2C). For comparison, we

also included published scRNA-seq datasets containing cells

from monkey and human embryos in the analyses (Nakamura

et al., 2016; Niu et al., 2019; Xiang et al., 2020; Zhou et al.,

2019). To avoid batch-specific systematic variations of scRNA-

seq caused by integration of different datasets, we used an ‘‘an-

chors’’ method that is recommended for batch-effect removal

(Stuart et al., 2019) (Figure S2C).

We first performed t-distributed stochastic neighbor embed-

ding (t-SNE) analysis on the scRNA-seq datasets. Based on

the expression of known lineage markers, we identified four ma-

jor cell clusters that were present in all samples (both chimeric

and control embryos): EPI, HYP, TE, and EXMC (Figures 3A,

3B, and S2D). These cells also expressed lineage-specific

markers that showed conservation between humans and

Cell 184, 2020–2032, April 15, 2021 2023

A

C

D

B Figure 3. Transcriptional landscape of hu- man-monkey chimeric embryos

(A) t-SNE plot of cells from chimeric and control

non-chimeric embryos. Cells were identified as

EPI, HYP, TE, and EXMC. Cells were colored by

different species origins and datasets.

(B) Expression of lineage-specific marker genes of

EPI, HYP, and TE exhibited on t-SNE plots. A

gradient of gray, yellow, and red indicates low to

high expression.

(C) The phylogenetic tree shows the cluster of EPI,

HYP, and TE cells from chimeric embryos at

different stages (d.p.f.9, 11, 13, 15, and 17). Cells

are highlighted by species origins (human or

monkey), different stages (d.p.f.9, 11, 13, 15, and

17), and different cell types (EPI, HYP, and TE).

(D) Bar plot showing the distribution of cells from

different origins in the four lineages (EPI, EXMC,

HYP, and TE). EPI, epiblast; HYP, hypoblast; TE,

trophectoderm; EXMC, extra-embryonic mesen-

chyme cell.

See also Figure S2 and Table S2.

ll Article

monkeys identified in a previous study (Figure S2E) (Zhou et al.,

2019). The presence of these cell types in chimeric embryos sug-

gests that the development of host embryos was by and large

unaffected by the injected hEPSCs (Ma et al., 2019; Niu et al.,

2019), which is consistent with the morphological analysis

(Figures 1B and 1D). Interestingly, phylogenetic tree analysis

(based on gene expression levels) revealed that while most mon-

key cells within chimeric embryos (chimeric monkey cells for

short) segregated into distinct cell-type-specific clusters (EPI,

HYP, and TE), chimeric human HYP- and TE-like cells clustered

with EPI-like cells (Figure 3C). Thus, it seems that the chimeric

monkey cells exhibited a more faithful lineage segregation than

the introduced hEPSCs. In agreement with IF results, very few

human TE-like cells were identified in the scRNA-seq data (Fig-

ures 3A and 3D) and were therefore excluded from subsequent

analyses. Together, these results demonstrate that hEPSCs

can differentiate into several peri- and early post-implantation

cell types after being introduced into monkey early blastocysts

followed by ex vivo embryo culture.

Transcriptome dynamics of hEPSCs during human- monkey chimera development We next investigated the transcriptional kinetics of chimeric hu-

man and monkey cells. We first constructed a force-directed K-

nearest neighbor graph (SPRING) (Weinreb et al., 2018) based on

transcriptomic properties of all cells (see STAR Methods). All

cells bifurcated into three branches: EPI, HYP, and TE (Fig-

ure 4A). The correlations of gene expression patterns between

chimeric and control (human and monkey) embryos were deter-

2024 Cell 184, 2020–2032, April 15, 2021

mined (Figure 4B). Similar correlation co-

efficients were obtained when chimeric

human cells were compared to control

human (0.460) or monkey (0.459) cells

(Figure 4B, right panels). Intriguingly,

when compared to control embryos,

chimeric monkey cells exhibited higher

correlation coefficients than chimeric human cells (Figure 4B,

left panels). We next generated lineage-specific correlation

matrices. We found that chimeric human EPI-like cells were

similar to EPI cells in human embryos, whereas chimeric human

HYP- and EXMC-like cells shared the highest correlation coeffi-

cients with chimeric monkey HYP cells and EXMCs, respectively

(Figure 4C). Of note, we also found that chimeric monkey EPI

cells and EXMCs more resembled chimeric than control human

cells. Interestingly, chimeric human EPI-like cells were found to

gradually gravitate toward the chimeric monkey EPI cells, with

R2 values increasing from 0.363 (pre-implantation EPI [Pre_EPI])

to 0.464 (post-implantation late EPI [PostL_EPI]) and then to

0.693 (gastrulating [Gast] cells) (Figure 4C). Taken together,

these results demonstrate that the monkey embryonic microen-

vironment exerts influence on the transcriptional states of human

cells and vice versa.

As monkey cells exhibited transcriptomic changes in the pres-

ence of human cells, we next sought to delineate the develop-

mental dynamics of monkey cells within chimeric embryos. We

first identified differentially expressed genes (DEGs) between

chimeric and control monkey embryos. Comparisons of EPI

cells, HYP cells, and EXMCs revealed that 424, 7, and 241 genes

were downregulated, whereas 5, 2, and 13 genes were upregu-

lated, respectively, in cells from chimeric compared to control

embryos, although the expression levels of lineage marker genes

remained comparable (Figures 4D and S3A). Gene Ontology

(GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG)

enrichment analyses identified a number of signaling pathways

that were up- and downregulated in chimeric monkey EPI cells,

A

C

D

F G

E

B

(legend on next page)

ll

Cell 184, 2020–2032, April 15, 2021 2025

Article

ll Article

HYP cells, and EXMCs. For example, the Hippo and transforming

growth factor b (TGF-b) signaling pathways were downregulated

in chimeric monkey EPI cells and EXMCs (Figure 4E),

respectively.

Having shown that the transcriptomic profiles of monkey EPI

cells were altered in chimeric embryos, we next investigated

whether the monkey EPI embryonic niche was also affected by

human cells. To this end, CellPhoneDB (v2.0.1) (Vento-Tormo

et al., 2018) was applied to identify potential cellular interactions

between EPI and other lineages (HYP and EXMC) in both

chimeric and control embryos (see STAR Methods). We sought

to identify interactions that were specific to chimeric or control

embryos. We found more ligand-receptor interactions in

chimeric embryos when compared to control embryos (e.g.,

117 [chimeric] versus 10 [control] specific ligand-receptor inter-

actions were detected in monkey EPI cells) (Figures 4F and S3B;

Table S3). KEGG analysis was performed to reveal specific

signaling pathways that were enriched within chimeric and

control embryos. We found several signaling pathways (e.g.,

phosphatidylinositol 3-kinase [PI3K]-Akt and mitogen-activated

protein kinase [MAPK] signaling pathways) that were strength-

ened in chimeric embryos and new signaling pathways (e.g.,

WNT signaling pathway) that were specifically enriched in

chimeric embryos (Figure 4G). Using the same method, we

also determined human and monkey cell-cell interactions within

chimeric embryos and identified distinct ligand-receptor interac-

tions in EPI cells, such as FGF5-FGFR4, NOTCH4-JAG2,

WNT2B-FZD4, WISP3-SORL1, and PLXNB2-PTN (Figures S3C

and S3D; Table S3). Taken together, our results suggest that

cell-cell interactions are reinforced within the chimeric embryos

and potentially lead to activation of additional signaling

pathways.

Chimeric human EPI-like cells display a distinct developmental trajectory EPI development is characterized by pluripotency transitions that

may exhibit different dynamics between species. As proper EPI

specification and differentiation are critical for chimera formation

and development, we examined the lineage allocation of human

EPI-like cells within the chimeric embryos and compared it with

the datasets of in-vitro-cultured human and monkey embryos

Figure 4. Developmental trajectory of human-monkey chimeric embry

(A) The differentiation trajectory of chimeric cells and control non-chimeric cells

colored by cell lineages.

(B) Overall similarity under the four comparisons (chimera-monkey versus control

versus control non-chimeric human [‘‘Chimera-Human vs. Control-Human’’], c

Control-Monkey’’], and chimera-human versus control non-chimeric monkey [‘‘C

(C) Heatmap of the correlation coefficients among different cell origins under cor

(D) Histogram showing the numbers of DEGs for different lineages (EPI, HYP, and

embryos. The red and blue colors represent up- and down-regulated genes, res

(E) GO and KEGG enrichment analyses of the DEGs in (D). Red and blue represen

respectively.

(F) The Venn diagrams showing the overlap of the ligand-receptor interactions bet

non-chimeric (Control-Monkey) monkey cells.

(G) Comparison of KEGG pathways enriched by the specific interactions betwee

cells in (F). EPI, epiblast; EXMC, extra-embryonic mesenchyme cell; HYP, hypobla

gastrulating cell; DEG, differentially expressed gene.

See also Figure S3 and Table S3.

2026 Cell 184, 2020–2032, April 15, 2021

(Nakamura et al., 2016; Niu et al., 2019; Zhou et al., 2019). Human

EPI-like cells were identified at pre-implantation, post-implanta-

tion, and gastrulating stages, and at each stage, they expressed

distinct markers (Figures 5A and S4A–S4C). A Sankey diagram

also showed the same developmental dynamics of human EPI-

like cells (Figure S4D). We next determined the relationship be-

tween hEPSCs (Yang et al., 2017b), chimeric human EPI-like

cells, EPI cells from control human and monkey embryos (Niu

et al., 2019; Zhou et al., 2019), and human and monkey PSCs

(primed and naive) (Chan et al., 2013; Chen et al., 2015; Gafni

et al., 2013; Theunissen et al., 2014). We observed that hEPSCs

were more similar to early post-implantation EPI (PostE_EPI)

and PostL_EPI cells from human and monkey embryos, respec-

tively, as well as human and monkey naive PSCs. Chimeric hu-

man PostL_EPI-like cells showed higher correlation coefficients

to primed PSCs than naive PSCs (Figure S4E). To further investi-

gate the transcriptional kinetics of hEPSCs (Yang et al., 2017b),

chimeric humanEPI-like cells, andhost monkeyEPI cells, we per-

formed RNA velocity (La Manno et al., 2018) and Slingshot ana-

lyses (Street et al., 2018) (Figure 5B). We observed two distinct

patterns of RNA velocity vectors; chimeric human PostL_EPI-

like cells bore long vectors, and gastrulating-like cells bore short

vectors. In contrast, host monkey PostL_EPI cells lacked long

vectors, and gastrulating cells bore long vectors (Figure 5B, left

two panels). These results imply that development is delayed

for chimeric human EPI-like cells. Slingshot analysis revealed

that after injection into monkey blastocysts, hEPSCs followed

the EPSC to PostL_EPI to gastrulation developmental trajectory

(Figure 5B, right panel). To further delineate the developmental

trajectory of chimeric human EPI-like cells, we mapped all EPI-

related human and monkey reads to a consensus genome and

aligned EPI development trajectories between species using a

previously reported method (Kanton et al., 2019). In agreement

with the RNA velocity analysis, we found that chimeric human

EPI-like cells differentiated more slowly than EPI cells from host

monkey, control monkey, and human embryos (Figure 5C). These

results suggest that the specification and/or differentiation of

hEPSCs toward the EPI lineages was less efficient than embry-

onic cells.

As mentioned above, global transcriptional profiles of chimeric

human EPI-like cells more resembled monkey EPI cells within the

os

. The differentiation trajectory was reconstructed using SPRING. Cells were

non-chimeric human [‘‘Chimera-Monkey vs. Control-Human’’], chimera-human

himera-monkey versus control non-chimeric monkey [‘‘Chimera-Monkey vs.

himera-Human vs. Control-Monkey’’]). Cells are colored by cell origins.

responding lineages (EPI, EXMC, HYP, Pre_EPI, PostL_EPI, and Gast).

EXMC) in the host monkey cells compared to cells from control non-chimeric

pectively.

t enrichment p values (�log10 transformed) for up- and downregulated DEGs,

ween EPI-EPI, EPI-HYP, and EPI-EXMC in host (Chimera-Monkey) and control

n host (Chimera-Monkey) and control non-chimeric (Control-Monkey) monkey

st; Pre_EPI, pre-implantation EPI; PostL_EPI, post-implantation late EPI; Gast,

A

C

D

F

E

B

Figure 5. Developmental trajectory of chimeric human EPI-like cells within monkey embryos

(A) t-SNE plot of EPI cells at different sublineages. Cells are colored by cell origins and designated as ICM, Pre_EPI, PostE_EPI, PostL_EPI, and Gast. Cells of

‘‘Control-Human 1’’ derive from dataset ‘‘Human-1’’ (Zhou et al., 2019). Cells of ‘‘Control-Monkey 1’’ derive from dataset ‘‘Monkey-1’’ (Niu et al., 2019). Cells of

‘‘Control-Monkey 2’’ derive from dataset ‘‘Monkey-2’’ (Nakamura et al., 2016). Lower right: single cells are colored according to embryonic stages.

(B) RNA velocity analysis of chimeric human (chimera-human) and host monkey (chimera-monkey) cells, respectively (left two panels). The amplitude and di-

rection of the vector reflects a transcriptional trajectory. Slingshot analysis of chimeric human EPI-like cells and hEPSCs (third panel). The curves and arrows

indicate the potential development trajectory. Cells are colored by cell lineages.

(C) Pseudotime alignment between control and chimeric cells. Left and right panels show the pseudotime alignment of chimeric cells with control human and host

monkey cells, respectively.

(legend continued on next page)