Derivation of human triploid trophoblast stem cells

Abstract

Purpose

Human trophoblast stem cells (hTSCs) are counterparts of the precursor cells of the placenta and are valuable cell models for the study of placental development and the pathogenesis of placental diseases. The aim of this work was to establish a triploid human TSC (hTSC^3PN) derived from the tripronuclear embryos, which are clinically discarded but readily available, for potential applications in basic placental research and disease modeling.

Methods

Eighteen tripronuclear human zygotes from IVF were collected and cultured for 5–6 days. Five high-quality blastocysts were harvested and were individually cultured in hTSC medium. Finally, two hTSC lines were established after 10 days and could be passaged stably.

Results

The karyotyping analysis showed that hTSC^3PN contained three sets of chromosomes. And the hTSC^3PN exhibited typical features of hTSCs, with the ability to differentiate into two trophoblast lineages: extravillous cytotrophoblasts (EVTs) and syncytiotrophoblasts (STs). In addition, the hTSC^3PN can mimic some vital features of trophoblast, including hormone secretion and invasion. Further studies showed that the proliferation and differentiation of hTSC^3PN were reduced compared with normal hTSCs, which may be related to the disturbed metabolic signaling in hTSC^3PN.

Conclusions

We established the triploid hTSC lines derived from tripronuclear embryos, which provides a potentially useful research model in vitro to study human placental biology and diseases.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10815-022-02436-w.

Introduction

Triploidy, a somewhat common chromosomal abnormality in human pregnancy, occurs in up to 3% of conceptuses [1]. Around 8–10% of spontaneous abortions are caused by triploidy [2, 3] and two-thirds of such cases are partial hydatidiform mole (PHM) [4]. PHM is a gestational trophoblastic disease characterized by abnormal placentas with a triploid diandric monogynic genome [5]. The PHM develops without inner cell mass and exhibits abnormal triploid trophoblast cells [4]. However, with lack of a suitable disease model in vitro, the pathology of PHM has been poorly understood.

Trophoblast stem cells (TSCs) are faithful counterparts of cytotrophoblast (CT) in the first-trimester placenta, which can differentiate into extravillous cytotrophoblasts (EVTs) and syncytiotrophoblasts (STs). The establishment of TSC lines of mice [6], rats [7], rhesus monkeys [8], and humans [9] has provided an ideal in vitro model to study placental biology and pregnancy complications. In addition, the recent development of haploid TSC (haTSCs) from haploid embryonic stem cells and parthenogenetic mouse blastocysts has further expanded this powerful tool for studying placentation [10, 11]. Tripronuclear (3PN) zygotes occur at a rate of 5.0–8.1% d during conventional in vitro fertilization (IVF) [12]. Although they are not suitable for transfer, 3PN embryos could be used for research purposes, such as genome editing. The triploid embryonic stem cell lines derived from tripronuclear human zygotes had been established to analyze the development of the 3PN embryo. However, triploid TSCs have not been obtained so far.

Here, we established two triploid hTSC lines (hTSC^3PN) from 3PN embryos. The karyotyping analysis showed that hTSC^3PN contained three sets of chromosomes. Similar to normal 2PN hTSCs, the hTSC^3PN grow as typical colonies and can be maintained in the long term in the hTSC culture condition. Importantly, the hTSC^3PN can be differentiated into EVT and ST. Intriguingly, the transcriptome analysis reveals that the hTSC^3PN has similar transcriptional features to normal 2PN hTSCs and PHM. Our study presents the first detailed analysis of hTSC^3PN, which could be potentially useful to investigate placental diseases, such as PHM.

Materials and methods

This study was approved by the Ethics Committee of Sun Yat-sen Memorial Hospital of Sun Yat-sen University (2020 Reproductive Ethics No. 15). The 3PN embryos were obtained from the patients who received assisted reproductive technology (ART) treatment at the center for reproductive medicine of Sun Yat-sen Memorial Hospital. The consent of the patients was obtained before collecting the embryos.

Acquisition of 3PN embryos and the establishment of hTSC

Oocytes were retrieved 36 h after the hCG was stimulated and were fertilized with sperm in 1 ml of G-IVF medium (Vitrolife, 10136) for 16–18 h. The zygotes were scored according to the Istanbul consensus [13]. Three PN zygotes were collected and cultured in G1-plus medium (Vitrolife, 10128) for 3 days and then cultured in G2-plus medium (Vitrolife, 10132) for 2 or 3 days at 37 °C in 6% CO₂, 5% O₂, and 89% N₂. Blastocyst quality is graded according to the Istanbul consensus [13]. High quality is defined as inner cell mass (ICM) and trophectoderm (TE) both containing many cells, ICM being compacted and tightly adhered together, and TE cells forming a sticky epithelium [13]. High-quality blastocysts were collected on D5 or D6 and cultured in hTSC medium [9] individually in four-well plates, which were pretreated with collagen IV (CORNING) for at least 30 min, at 37°C in 6% CO₂, 5% O₂, and 89% N₂ for 7 days. The hTSC medium was made up of DMEM/F12(GIBCO), 0.2%FBS (GIBCO), 0.1mM 2-mercaptoethanol, 0.5% penicillin-streptomycin, 1.5 mg/ml l-ascorbic acid, 1% ITS-X supplement (GIBCO), 50 ng/ml hEGF (CST), 0.5 mM A83-01 (CST), 1 mM SB431542 (CST), 2 mM CHIR99021 (CST), 0.8 mM VPA (CST), and 5 mM Y27632(CST). If the cells adhered to the plate wall, half of the medium was changed every 2 days until a clone was formed. About day 10, a larger cell cluster was formed, then the cells were digested and passaged. The cell clone was washed by PBS (GIBICO) for 3 times and digested in TrypLE (GIBCO) for 3 min, and then collect the cells, were centrifuged, seeded into two holes of a 6-well plate which was pre-coated with 7.5 μg/ml Col IV. The passage is in a ratio of 1:3 on a six-well plate during expansion culture.

The differentiation of hTSCs into ST and EVT

The cells were passaged routinely once every 4–5 days. After expansion, the hTSCs were collected for ST and EVT differentiation tests. To differentiate hTSCs to EVT, cells from one well of a 6-well plate were dissociated with TrypLE (GIBCO) for 3 min and then split into single cells. Those cells were then seeded in a 6-well plate pre-coated with 2 μg/ml Col IV and cultured in 2 ml DMEM/F12 supplemented with 0.1 mM 2-mercaptoethanol, 0.2% Primocin, 0.3% BSA, 1% ITS-X supplement, 7.5 mM A83-01, 2.5 mM Y27632, 100 ng/ml NRG1(CST), 4% KnockOut Serum Replacement (KOSR) (GIBCO), and 2% Matrigel (CORNING). The culture medium was changed every 2 days. On the fourth day, the cells began to shuttle. The cells were cultured for 6 days, using TrypLE to digest the EVT cells and then followed by extraction for RNA The immunofluorescence analysis was performed at this time.

To differentiate hTSCs to ST, cells from one well of a 6-well plate were dissociated with TrypLE (GIBCO) for 3 min and then dissociated into single cells. The single cells were then seeded in a 6-well plate pre-coated with 15 μg/ml Col IV and cultured in 2ml DMEM/F12 supplemented with 1% ITS-X supplement, 2 mM forskolin, 2.5 mM Y27632, and 4% KOSR. On the second day, the cells began to fuse into multinucleated syncytia, and over time, the proportion of fused cells increased. The cells were cultured for 4 days, followed by extraction for RNA and immunofluorescence analysis.

RNA isolation

Total RNA was purified using TRIzol reagent (Thermo). First-strand cDNA was synthesized from the total RNA using PrimeScript RT Mix (Takara), and real-time PCR (qPCR) was performed using qPCR system and SYBR Green PCR Master Mix (Thermo). The amount of the target mRNA was determined using the △△Ct method with GAPDH as the internal control. PCR primer sequences used for qPCR are listed in Table 1.

Table 1.

The primers used for qPCR

Primers	Target	Forward/reverse primer (5′–3′)
House-keeping genes (qPCR)	GAPDH	GAGTCCACTGGCGTCTTCAC/TTCACACCCATGACGAACAT
Trophoblast stem cell markers	CDX2	GACGTGAGCATGTACCCTAGC/GCGTAGCCATTCCAGTCCT
	GATA2	GACCACTCATCAAGCCCAAG/CACAGGCGTTGCAGACAG
	GATA3	ACTACGGAAACTCGGTCAGG/GGTAGGGATCCATGAAGCAG
	DLX3	GAGCCTCCTACCGGCAATAC/TCCTCCTTCACCGACACTG
	KRT18	TGATGACACCAATATCACACGA/GGCTTGTAGGCCTTTTACTTCC
	TEAD4	CTCCACGAAGGTCTGCTCTT/GTCCATTCTCATAGCGAGCATA
	ITGA6	AGCCTCTTCGGCTTCTCG/TTGGCTCTCTGCAGTGGAA
	CLDN10	TTGATCCTCTCTTTGTTGAGCA/AAAGCAAAATATGACACCACCA
	ELF5	ACTACCCTGCCTTTGAGCATC/ACACATGGCGCTTAGTCCAG
Syncytiotrophoblast markers	SDC1	AGGATGGAGGTCCTTCTGC/CCGAGGTTTCAAAGGTGAAGT
	INHA	CGAGGAAGAGGAGGATGTCTC/GCAGCTGACTTGTCCTCACA
	CGA	TGTCGGTGTTTCTGCATGTT/GCTGGGAGAAGAATGGGTTT
	CGB	AGCACTTTGCTCGGGTCACGG/TGGTCCAGCGCCAAGGGTGA
	ERVV1	CCTCTGTTCCAGAAGGGAACT/TGGGGAATAGCTCCTACCTTG
	ERVV2	GAAACGCAGCCCACTGATAG/GGATGGGTTATTAGAGAAGTGCTG
	ERVW1	CTTCCTCTCATTCTTAGTGCCC/CCAATGCCAGTACCTAGTGC
	ERVFRD	ACCGCCATCCTGATTTCCC/GAGGCTGGATAAGCTGTCCC
	PSG1	CCTCCCAGCCCCTTCC/GGTGTAGGTTCCTGCATCCTT
	PSG9	GACTTGTCCTGCTTCACGGA/ATGGCAGGGACCAGAGACTT
Extravillous cytotrophoblast markers	FN1	GGGAGAATAAGCTGTACCATCG/TCCATTACCAAGACACACACACT
	MMP2	ATAACCTGGATGCCGTCGT/AGGCACCCTTGAAGAAGTAGC
	HLA-G	CCACCACCCTGTCTTTGACTAT/ACGTCCTGGGTCTGGTCCT

Immunostaining

The cells were rinsed with PBS, fixed with 4% paraformaldehyde (PFA) for 15 min, permeabilized with 0.3% Triton X-100 for 10 min, and blocked with 2% FBS for 1.5 h at room temperature. The following primary antibodies were used: anti-GATA2 (1:100), anti-GATA3 (1:100), anti-HLA-G (1:100), anti-E-cadherin (1:100) (Abcam), and anti-SDC1 (1:100) (Abcam); Primary antibodies were incubated overnight in a 4° refrigerator. Then used PBS was washed 3 times. Alexa Fluor 555– and Alexa Fluor 488–conjugated secondary antibodies were used and incubated for about 2 h. Nuclei were stained with DAPI (Abcam).

Karyotyping

The hTSC^3PN at passage 8 were treated with 25 μg/ml Colchicine for 60 min at 37 °C. Cells were trypsinized and then resuspended in 0.075 mol/l KCL solution for 15 min at 37 °C, fixed with Carnoy’s solution. After these procedures, the cells were dropped onto slides. Staining with Giemsa and then G-banding was performed by trypsinizing the cells for 8 min. After that, metaphases were analyzed.

HTSC transplantation

HTSC transplantations were performed as published [9]. Briefly, 1×10⁷/ml hTSC^3PN cells were resuspended in 200 μl of 1:1 mixture of Matrigel and DMEM/F12 containing 1% ITS-X supplement and 0.3% BSA, and subcutaneous injection in the hind limb of 8-week-old NOD-SCID mice. It took about 4 days to form a pimple. Seven days post-injection, blood was collected for hCG measurement and lesions were collected for immunohistochemistry staining or hematoxylin-eosin staining.

Hematoxylin-eosin (HE) staining

Lesions were dissected, fixed in 4% paraformaldehyde overnight, and frozen sectioned. The sections were passing through decreasing concentrations of alcohol (100%, 90%, 80%, 70%) and water, respectively. Then, the sections were stained in hematoxylin for 4 min, dipped into 1% acid alcohol for a few seconds, and washed in water for 3 min. Then, dehydration was done in increasing concentration of alcohols. Rinse with water for 15–20 min and 0.5% eosin solution for 2–5min; rinse with distilled water for 2s, 80% ethanol 2 s, 90% ethanol 2 s, anhydrous ethanol 5–10s, anhydrous ethanol 10–30s, xylene carbolic acid in seconds, xylene 2 min, and xylene 2 min. Observe whether the staining was clear under microscope, neutral resin sealing sheet.

Measurement of hCG

HTSCs and hTSC^3PN were seeded at a density of 2×10⁴/ well of 12-well plates, and the supernatant was collected 48 h later. The hCG level was measured by Runda hCG test kit (A85264). The kit was also used to test the hCG of mouse serum. Mouse blood was collected by eyeball sampling method, blood was waited to coagulate and the centrifuged to take the supernatant, and then serum hCG was detected.

RNA-seq

Total RNA of hTSC and hTSC^3PN were extracted, and used for library construction.

The reads were aligned to the reference genome (UCSC hg38) using Hisat2 [14] with the Refseq gene annotation. Using DESeq2 to identify the differentially expressed genes (DEGs). Enrichment analyses of DEGs were performed using the Clusterprofiler [15] for GO and KEGG pathways. The expression matrix of the GSE138250 dataset (of the control and HM groups) was downloaded from the GEO database [16]. The edgeR package [17] was used to analyze the differentially expressed genes of the sample counts. p-value < 0.05 with a change ≥ twofold were set to filter the expressed genes based on statistically significant differences.

Results

Establishment of the trophoblast stem cell lines derived from triploid embryos

We attempted to establish the hTSC^3PN following the protocol that was previously used to derive normal hTSCs from human blastocysts (Fig. 1A). In this study, 18 3PN zygotes (Fig. 1B) from IVF were collected and cultured for 5–6 days. Five high-quality blastocysts were harvested and individually cultured in the hTSC medium [9]. Finally, hTSC-like colonies arose after 10 days of culture from two blastocysts and could be stably passaged further (Fig. 1C, supplement 2). Thus, we established two hTSC lines from two 3PN blastocysts. The karyotyping analysis confirmed that hTSC^3PN contained three sets of chromosomal constitutions. However, chromosomes no. 3, no. 5, no. 7, and no. 20 are composed of 4 chromosomes, respectively (73, XYY) (Fig. 1D). This phenomenon was also observed in triploid human embryonic stem cells (hESCs), with randomly increased or decreased numbers [18, 19]. HTSC^3PN continued to proliferate to P31 until now.

Fig. 1.

Establishment of hTSC^3PN line. A Flow chart of establishing the hTSC^3PN line. Human tripronuclear zygotes developed into blastocysts and then were induced into hTSC^3PN. B Human tripronuclear zygote (the scale bars indicate 50μm.). C Blastocyst for hTSC^3PN establishment (the scale bars indicate 100 μm). D The karyotype analysis of hTSC^3PN. E Similarities in RNA-seq between normal hTSCs and hTSC^3PN cells. hTSC1-2 were induced by normal karyotype blastocysts in this experiment. F Immunofluorescence staining of GATA2, GATA3, and ITGA6 in hTSC^3PN cells (the scale bars indicate 100 μm). F The qPCR results of hTSC markers between normal hTSC and hTSC^3PN cells. The data were normalized to GAPDH expression. The error bars indicate SD (n = 6). *p < 0.05, **p<0.01, ***p<0.001, ****p < 0.0001

For comparison, we also established one normal hTSC line from the human 2PN blastocyst (supplement 1). Both normal hTSCs and hTSC^3PN were subjected to RNA-seq analysis. Our analysis revealed that the transcriptome of hTSC^3PN was comparable to that of normal hTSCs with a similarity of about 90% (Fig. 1E). Both RT-qPCR (Fig. 1F) and immunostaining (Fig. 1G) analyses confirmed the marker genes (e.g., GATA2, GATA3, ITGA6) of hTSCs were highly expressed in hTSC^3PN. These data suggest despite the abnormal chromosomal constitutions, the hTSC^3PN may share the majority of molecular features with normal hTSCs.

Differentiation potential of the hTSC^3PN

To assess the differentiation potential of hTSC^3PN, we first induced the hTSC^3PN into EVT (EVT^3PN) and ST (ST^3PN) in following the reported methods [9]. For EVT^3PN differentiation, the marker genes of EVTs such as FN1, MMP2, and HLA-G were highly expressed in the EVT^3PN (Fig. 2A–C). On the other hand, the hTSC^3PN can aggregate and fuse into large syncytium upon induction, suggesting ST^3PN formed (Fig. 2D). RT-qPCR results showed that expression levels of ST marker genes, including SDC1, INHA, CGB, CGA, ERVV1, ERVV2, ERVW1, and ERVFRD, were increased in ST^3PN when compared to hTSC^3PN (Fig. 2E). Consistently, the significant secretion of hCG was detected in the culture media from ST^3PN (Fig. 2F). In addition, the immunofluorescence staining of E-cadherin showed the ST^3PN exhibited a confluent state with fused nuclei (Fig. 2G). Moreover, immunofluorescent staining of SDC1 and hCG conformed that they were expressed in these syncytia (Fig. 2H). Taken together, those results suggest that the hTSC^3PN holds the potential to differentiate into EVT and ST trophoblasts.

Fig. 2.

HTSC^3PN differentiate into EVT and ST. A Bright field of EVT induced by hTSC^3PN cells (scale bar, 100 μm and 20 μm). B The expression levels of EVT-specific marker genes (FN1, MMP2, and HLA-G) in EVT^3PN cells. Data were presented as mean ± SD (n = 3). C Immunofluorescence staining of HLA-G in EVT^3PN cells (scale bar, 100 μm and 10 μm). D Bright field of ST^3PN induced by hTSC^3PN cells. E The expression levels of ST-specific marker genes (SDC1, INHA, CGB, CGA, ERVV1, ERVV2, ERVW1, and ERVFRD) in ST^3PN cells. Data were presented as mean ± SD (n = 3). F Levels of hCG secreted by control (culture medium), hTSC^3PN cells, and ST^3PN cells. Data were presented as mean ± SD (n = 3). G Immunofluorescence staining of E-cadherin in ST^3PN cells (scale bar, 100 μm and 10 μm). H Immunofluorescence staining of SDC1, hCG in ST^3PN cells (scale bar, 100 μm and 10 μm). *p < 0.05, **p<0.01, ***p<0.001, ****p < 0.0001

Evaluation of engrafting potential of hTSC^3PN in vivo

To further evaluate the in vivo engrafting potential of hTSC^3PN, we subcutaneously injected the normal hTSCs and hTSC^3PN cells into non-obese diabetic (NOD)-severe combined immunodeficiency (SCID) mice. Seven days later, the teratoma was isolated for analysis. The H&E staining showed that the hTSC^3PN-derived teratoma exhibited similar morphology to that of normal hTSCs (Fig. 3A). In addition, immunohistochemical staining revealed that ITGA6 and hCG were expressed in these teratoma tissues (Fig. 3B, C), confirming that the injected cells were derived from hTSCs. Notably, we detected higher hCG levels in the serum of NOD-SCID mice injected with the hTSC^3PN cells, compared to the non-injection control group (Fig. 3D). Collectively, these data suggest that the hTSC^3PN can be successfully transplanted subcutaneously and maintain some functional properties of trophoblasts [9].

Fig. 3.

Evaluation of hTSC^3PN in NOD-SCID mice. A HE staining of hemorrhagic lesions derived from hTSCs and hTSC^3PN. The arrows show clearly the multinucleated cell. (scale bar, 100 μm). B Immunofluorescence staining of ITGA6 in the hTSCs and hTSC^3PN-derived lesion from NOD-SCID mice (scale bar, 100 μm). C Immunofluorescence staining of hCG in the hTSCs and hTSC^3PN-derived lesion from NOD-SCID mice (scale bar, 100 μm). D HCG levels of the NOD-SCID mice without injection cells and hTSC^3PN cells injected groups (n=1)

HTSC^3PN cells are deficient in proliferation and differentiation

Next, we further explored the difference between the hTSC^3PN and normal hTSCs. We found that either clone formation or cell proliferation of hTSC^3PN was decreased compared to normal hTSCs (Fig. 4A–C). In consistence with this, the immunohistochemical staining revealed that the expression of Ki67, a proliferation marker, was decreased in hTSC^3PN (Fig. 4D). The expression of ST markers such as INHA, CGA, CGB, and ERVV1 (Fig. 4E) and the secretion of hCG (Fig. 4F) were reduced in ST^3PN compared with the ST derived from normal hTSCs, suggesting the capacity of hTSC^3PN to differentiate into ST could be compromised. Similarly, the transcriptional expression of FN1, MMP2, and HLA-G was decreased in hTSC^3PN derived EVT^3PN (Fig. 4G). Taken together, these results indicate that despite the similar morphology to that of normal hTSCs, hTSC^3PN exhibits some deficiencies in proliferation and differentiation.

Fig. 4.

HTSC^3PN cells are deficient in proliferation and differentiation. A Clone formation experiments of normal hTSCs and hTSC^3PN. A purple spot is a clone formed by a single cell. B A bar graph showing the number of clones. Data were presented as mean ± SD (n = 3). C Statistics of the cell numbers in 48h and 72h. Data were presented as mean ± SD (n = 3). D Immunofluorescence staining of Ki67 in hTSCs and hTSC^3PN (scale bars, 20 μm). E The qPCR results of ST lineage markers. Data were presented as mean ± SD (n = 3). F The testing results of hCG levels secreted by control (culture medium), ST derived from hTSCs (ST) and ST derived from hTSC^3PN (ST^3PN). Data are presented as mean ± SD (n = 3). G The qPCR results of EVT lineage markers gens in EVT induced by different cell lines (EVT derived from normal hTSCs; EVT^3PN derived from hTSC^3PN). Data were presented as mean ± SD (n = 3). The above experiments were repeated three times independently. *p < 0.05, **p<0.01, ***p<0.001, ****p < 0.0001

Different molecular signaling between normal hTSCs and hTSC^3PN

Considering the functional variations, we next tried to investigate the underlying molecular mechanisms by comparing the transcriptome between hTSC^3PN and normal hTSCs. The hierarchical cluster analysis was performed on the common differentially expressed genes (DEGs). OCT4 is a master transcription factor along with other factors that regulates the pluripotency of pluripotent stem cells, and POU class 5 homeobox 1 pseudogene 3 (POU5F1P3) is a pseudogene of OCT4 [20]. CPED1 is involved in alternative splicing [21]; its upregulated may cause the disorder of RNA splicing. The DEG results indicated that POU5F1P3, CPED1, and docking protein 5(DOK5) were the top three genes with the largest increase in hTSC^3PN. Wnt family member 5B(WNT5B) is a member of the WNT family and is involved in cell migration, cell differentiation, and cell proliferation in many cell types [22–24]. Family with sequence similarity 50 member B (FAM50B) is imprinted genes expressed in human tissues except for the ovary. It is reported that FAM50B may influence cell proliferation [25]. FAM133A, FAM50B, and WNT5B were the three genes with the most decreased expression and these genes (Fig. 5A). GSEA-KEGG analysis revealed that the Herpes simplex virus 1 infection and the PI3K-Akt signaling pathway were activated in hTSC^3PN cells. However, the pathways related to steroid hormone biosynthesis the oxidative phosphorylation were inhibited in hTSC^3PN (Fig. 5B–D). These results imply differential signaling pathways consist in normal hTSCs and hTSC^3PN.

Fig. 5.

Signaling pathway differs between hTSCs and hTSC^3PN. A Hierarchical clustering of gene expression when hTSC^3PN compared to hTSCs. The red dots represent highly expressed genes and blue dots represent low expressed genes. B Bubble plot of GSEA-KEGG signal pathway enrichment analysis. The size of the bubble represents the number of genes enriched in the signal pathway, and the intensity of the color represents statistical significance. C The enrichplot of steroid hormone biosynthesis. D The enrichplot of the oxidative phosphorylation

The application potential of hTSC^3PN in modeling PHM

Next, we interrogated whether the hTSC^3PN cell line could be used as a potential in vitro model of HM (PHM). Limited to the scarcity of clinical samples, we only compared the transcriptome of HM placenta with that of hTSC^3PN cells (supplement 3A). We found that HM and gestational age–matched healthy placenta could be used to cluster the transcriptome of hTSC^3PN and normal hTSCs (supplement 3B). The DEGs between normal hTSCs and hTSC^3PN were taken intersection with the DEGs between HM and healthy control placentas. These genes were significantly enriched in GO categories of female pregnancy, cell-cell junction, and metal ion transmembrane transporter activity (supplement 3C). The PSG family genes including PSG1, PSG2, PSG3, PSG4, PSG6, PSG9, and PSG11 are mainly expressed in ST [26] and presented lower expression in HM (supplement 3D), suggesting the ST function may be impaired in HM. Surprisingly, the expression of PSG family was decreased in hTSC^3PN when compared with normal hTSCs (supplement 3E). Further studies were taken to explore the PSG expression between the ST derived from normal hTSCs and hTSC^3PN. As expected, the qPCR results showed that PSG1, PSG2, PSG3, PSG4, PSG6, PSG9, and PSG11 were dramatically decreased in ST derived from hTSC^3PN (supplement 3F). Of note, P57, an imprinted gene maternally expressed in PHM but not in HM placenta [27–29], was also expressed in hTSC^3PN (supplement 3G). It has been reported that P57 could be used as a biomarker to improve the diagnosis of these two types [27–29]. Altogether, these data suggest the hTSC^3PN exhibits some molecular features of PHM.

Discussion

Triploidy may lead to implantation failure, miscarriage, and other pregnancy diseases such as PHM [1]. PHM is a gestational trophoblastic disease characterized by an abnormal placenta with a triploid diandric monogynic genome [30]. Clinical prevention and treatment need a better understanding of human placentation. Recent advance in hTSC development has provided unprecedented opportunities to study human placental biology and pregnancy diseases.

In this study, we successfully established the triploid hTSC cell line, hTSC^3PN, from triploid blastocysts. The hTSC^3PN line exhibited morphology similar to that of normal hTSCs and shared some molecular characteristics between these two hTSC lines. Importantly, the hTSC^3PN maintained the ability to differentiate into EVT and ST cells with hCG secretion in vitro, albeit at lower efficiency. Furthermore, the hTSC^3PN cells injected into NOD-SCID mice mimicked many key features of trophoblast cell invasion during implantation. Altogether, these data indicate that the hTSC^3PN cell line presented here can recapitulate some canonical features of hTSCs, which could be acted as a potentially useful tool to study placental diseases, such as PHM.

Despite the similarities, the hTSC^3PN exhibits abnormal phenotypes compared to normal hTSCs, we found that both proliferation and differentiation potential of hTSC^3PN were compromised, which might be due to the following reasons. Firstly, it has been thought that polyploid cells endoreplicate their DNA without cell dividing [31]. Secondly, the RNA-seq results revealed that genes associated with cellular metabolism signaling pathways, such as steroid hormone biosynthesis and oxidative phosphorylation, are all disturbed in hTSC^3PN when compared to normal hTSCs. These pathways have been implicated in cell metabolism and cell proliferation [32, 33]. Herpes simplex virus 1 infection signaling pathway is related to the immune system [34]. During pregnancy, the placenta maintains an immunotolerant environment to protect the fetus [35]. Abnormalities in immune-related signaling pathways have been associated with abnormal development of the placenta or fetus [36]. PI3K-Akt signaling pathway affects the development of trophoblast cells [37], and its disorder is associated with preeclampsia [38]. The dysregulation of these signaling pathways in hTSC^3PN may lie behind abnormal phenotypes of this cell line.

The genomes of placental cells are prone to exhibit abnormality in copy number. It has been reported that the incidence of mosaic placentas is higher than the incidence of mosaic fetuses and the placentas frequently exhibit chromosomal aberration [39]. Interestingly, the different responses of embryonic and extra-embryonic cells to aneuploid damage have been documented [40]. In agreement with these phenotypes, the evidence provided by Tim et al. has suggested that mosaicism is a vital feature of placental development [41]. Previous research has also shown that euploid cells in the extra-embryonic tissues present a lower rate of apoptosis than aneuploid cells in the embryonic epiblast in mouse embryos [42, 43]. Baharvand and colleagues have reported that the triploid hESCs can spontaneously differentiate to the trophoblast lineage and can secrete hormone [44]. Our results confirmed that in the case of chromosomal abnormalities, hTSC^3PN can still display the basic characteristics of hTSCs, including expression of hTSC marker genes and differentiation into EVT and ST with a deficient efficiency. The accumulating pieces of evidence have shown that hTSCs have a high tolerance for chromosomal abnormalities. These data support that the hTSC^3PN could be a valuable in vitro model to study the chromosomal mosaicism of the trophoblast lineage [45].

In 2017, HMol1-2C, HMol1-3B, and HMol3-1B were established from primary cultures of HM explants by introduction of hTERT and other genes such as CDK (CDK4R24C), cyclin D1, p53C234, MYC, and HRAS [46]. These cells exhibit characteristics of cytotrophoblastic cells and could differentiate into STs or EVTs with compromised functions [46]. Based on this research, PHM cell line may have the following features: characteristics of cytotrophoblastic cells; potential to differentiate into ST or EVT; triploid karyotype and expression of P57 (which can be used to distinguish PHM from HM). Owing to the lack of clinical samples of PHM, we compared the transcriptome of hTSC^3PN with HM placenta and found some shared transcriptome features between hTSC^3PN and HM: similar activated genes and signaling pathways, such as lowly expressed PSGs, FAM50B, and WNT5B.

In summary, we successfully established the hTSC^3PN cell lines derived from human 3PN embryos, which provides a promising research tool for molecular and functional characterization of human placentation in 3PN embryos. Furthermore, this cell line is potentially valuable for understanding the pathogenesis of aneuploidy-related diseases such as PHM.

Declarations

Ethics approval and consent to participate

The study protocol and all subjects who participated in this study were approved by the Ethics Committee of Sun Yat-sen Memorial Hospital of Sun Yat-sen University (2020 Reproductive Ethics No. 15), and informed consent was obtained from all patients prior to their participation in accordance with institutional and national guidelines.

Conflict of interest

The authors declare no competing interests.