JEG-3 (HTB-36™, American Type Culture Collection) is an epithelial-like choriocarcinoma cell line with an EVT-like phenotype (KRT7⁺ GATA3⁺ TFAP2C⁺ HLAG⁺) (19). It was cultured in DMEM/F12 supplemented with 10% fetal bovine serum (FBS).

The formation of trophoblast organoids has been described previously (6, 18). Briefly, 1 × 10⁴ JEG-3 cells (WT or KO) were resuspended in 30 μl of growth-factor-reduced Matrigel (Corning, USA) and then inoculated into 48-well plates with basic trophoblast organoid culture medium in a humidified incubator at 37°C with 5% CO₂. Small trophoblast organoids were visible after 7 days.

The organoids were transferred into 24-well plates or 35-mm dishes and induced to differentiate into ST or EVT. The ST medium contained forskolin (2μM), whereas the EVT medium was a sequential treatment with EVT medium containing NRG1 (100 ng/ml) for 8 days followed by EVT medium without NRG1 for 3 days. Either the trophoblast organoids culture medium or the ST/EVT differentiation medium was constructed according to the formulas described by Turco et al. and Okae et al.

HLA-G knockout JEG-3 cells were generated using CRISPR/Cas9-based gene editing technology (20). As described previously, small guide RNAs were synthesized by the Shanghai Sheng Gong Company (21) ( Supplemental Table 1 ); they were then cloned into the sg-RNA expression vector pSpCas9(BB)-2A-Puro (Addgene, USA). For transfection, a NEPAGENE (Nepa Gene, Japan) electroporator with a poring pulse (voltage, 150 V; pulse length, 7.5 ms) was used. After transfection, JEG-3 cells were selected using puromycin (2 µg/ml). The efficiency of HLA-G knockout was validated by polymerase chain reaction (PCR) and western blot analysis.

Genomic DNA was extracted from gene-edited JEG-3 cells using a QIAamp DNA Mini Kit (Qiagen, USA). A nanodrop ND-2000 spectrophotometer was used to determine DNA quality and concentration. The PCR utilized Premix Taq and the following PCR protocol, 94°C for 30 s, 54.3°C for 30 s, 72°C for 1 min 10 s; 35 cycles. The PCR product was run for 30 min at 150 V in 2% agarose. Subsequently, densitometric analysis was performed using a ChemiDoc image analyzer under UV radiation.

At Sangon Biotech (Shanghai) Co., Ltd., HiSeq XTen sequencers were used to perform an RNA sequencing experiment (lllumina, San Diego, CA, USA). After ensuring the integrity of the sequenced data using FastQC (Version 0.11.2), the raw reads were filtered and mapped to the reference genome using HISAT2 (Version 2.0) with the default parameters. The alignment outcomes were analyzed using RAeQC (Version 2.6.1). DESeq2 (Version 1.12.4) was employed to identify differentially expressed genes (DEGs) between the specified groups. Genes were considered as significantly differentially expressed if q-val was <0.05 and |FoldChange| was >2. Subsequently, the functional analysis of DEGs, including Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment, was performed to determine which DEGs were significantly enriched in specific GO terms or pathways using R (Version 4.2.0) and Rstudio (Version 2022.02.3). GO terms and KEGG pathways with a false discovery rate (q-val) of <0.05 were deemed significant. Transcripts per million (TPM) was used for specific gene set expression, thereby eliminating the influence of gene lengths and sequencing discrepancies and enabling direct comparison of gene expression between two samples.

Total RNA was extracted using TRIzol reagent (Gibco, USA) in accordance with the manufacturer’s instructions and then reverse transcribed using a Prime-Script RT Reagent Kit (TaKaRa, Japan). Quantitative PCR was conducted using an ABI 7500 Real-Time PCR system (Life Tech, USA) using SYBR Green Reagents (TaKaRa, Japan). In addition, glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used to normalize the gene expression levels. Subsequently, the relative gene expression levels were calculated using the threshold cycle (CT) method (2−^△△CT method). All the primers used for real-time PCR are listed in Supplementary Table 1 .

Total proteins extracted from JEG-3 cells or JEG-3-ORGs were lysed in lysis buffer containing a protease inhibitor mixture and quantified using Bradford reagent (Solarbio, China). The lysates were separated by SDS-polyacrylamide gel electrophoresis and then transferred to a polyvinylidene fluoride membrane. The membranes were incubated overnight at 4°C with primary antibodies against human HLA-G (1,1,000, Abcam, UK), PAPPA2 (1,1,000, Abcam, UK), and β-tubulin (1,5,000, proteintech, China); they were incubated subsequently with horseradish peroxidase-conjugated rabbit or mouse secondary antibodies (1,10,000, Abcam, UK). After treatment with enhanced chemiluminescence reagent (Meilunbio, China), a ChemiDoc image analyzer was used for densitometric analysis. β-tubulin served as the loading control.

The organoids were washed with cell recovery solution (Corning, USA) before being fixed for 30 min in 4% paraformaldehyde. After permeation and blocking, primary antibodies against human GATA3 (1,100, Abcam, UK), TFA2A (1,100, Abcam, UK), TFA2C (1,100, Abcam, UK), beta-hCG (1,500, Abcam, UK), E-cadherin (1,250, Abcam, UK), HLA-G (1,50, Abcam, UK), and PAPPA2 (1,100, Abcam, UK) were added and incubated at 4°C overnight. The following day, the organoids were sequentially incubated with fluorescence-labeled secondary rabbit or mouse antibody. Phalloidin and DAPI were added accordingly. Subsequently, the stained organoids were imaged using a ZEISS LSM 700 Confocal Laser Scanning Microscope and ZEN imaging software.

Cell proliferation was measured using cell counting kit-8 reagent (GlpBio, USA) with 2-(2-methoxy-4-nitrophenyl)-3-(4-nitrophenyl)-5-(2, 4-disulfophenyl)-2H-tetrazolium sodium salt (WST-8). Briefly, 1×10⁴ JEG-3 cells in each group were inoculated into a 96-well plate with 100 μl of culture medium per well. After 24 h of cultivation, 10 μl of WST-8 was added to each well and incubated at 37°C for 1 h. A colorimetric analysis was performed at a wavelength of 450 nm.

Invasion assays were conducted using Corning BioCoat Matrigel invasion chambers (BD, USA) with 8.0 μm PET membranes in 24-well plates (Corning, USA). Culture medium (600 μl) containing 10% FBS was added to the lower compartment, while 200 μl of serum-free medium containing 1×10⁵ JEG3 cells (HLA-G KO cells or negative control cells) was added to the upper compartment. After culturing for 24 h, the Transwells were fixed with methanol and stained with crystal violet. Finally, the invaded cells were counted under a light microscope.

Matrigel was polymerized on 24-well culture plates by coating them with undiluted Matrigel (300 μl/well) at 37°C for 30 min. HUVEC cells (5×10⁴) were pre-labeled with Green fluorescence CMFDA for 30 min before being seeded into each well to form endothelial cellular networks. After 4 h, Red fluorescence CMPTX dye-labeled JEG-3 cells (5×10⁴) with or without HLA-G were added to each well and incubated for 24 h. Finally, red and green fluorescence images were captured using a ZEISS Axio Observer fluorescence microscope.

Initially, JEG-3 (1×10⁵) cells were labeled with Red fluorescence CMPTX dyes and seeded into a 12-well plate. Subsequently, NK92 cells (1×10⁶) were added to each well. After 24 h of culture, co-culture cells were harvested, blocked with human IgG in RPMI1640 with 1% FBS, and stained with Alexa Fluor^® 700-conjugated anti-human CD45 (1,200 BioLegend, US), APC-conjugated anti-human HLA-G (1,200 BioLegend, US), and 7-AAD staining solution (1,400, BioLegend, USA) for 30 min. Subsequently, the data were collected using a Deflex B5-R3-V5 Flow Cytometer (405/488/638) and analyzed using the FlowJo_v10.6.2 software. All the antibodies used for this study are described in Supplementary Table 2 .

In conjunction with EVT differentiation from JEG-3 organoids (HLA-G KO group), the sequential EVT medium was supplemented with recombinant HLA-G protein (OriGene, USA) at varying concentrations (0.025 µg/ml, 0.05 µg/ml, and 0.1 µg/ml) as described previously (22). On days 0, 3, 6, and 9, the EVT-differentiated JEG-3 organoids (HLA-G KO group) were harvested for RNA analysis.

SPSS Statistics version 25.0 (IBM) was used for the statistical analysis. Using GraphPad Prism 8.0, the data were graphically represented as mean ± standard deviations (SD). In addition, the independent t-test and one-way analysis of variance were used to determine the differences between two and multiple groups, respectively. Each experiment was conducted a minimum of three times. P <0.05 was regarded as the statistical significance threshold.

The cytotrophoblast organoid medium containing activators (EGF, HGF, FGF2, CHIR99021, and R-spondin-1) and inhibitors (A83-01 and Y-27632) targeting the Wnt, MAPK, cAMP/AKT, TGF-beta, and ROCK signaling pathways was applied to the formation of JEG-3 organoids. JEG-3 cells were successfully reprogrammed into trophoblast organoids (JEG-3-ORGs) following a 7-day culture ( Figures 1A , 2A ). It has intercellular vacuoles and a three-dimensional growth pattern ( Figure 2B ). JEG-3-ORGs expressed trophoblast markers GATA3, TFAP2A, and TFAP2C, just like organoids derived from the human primary trophoblast ( Figure 2C ). To examine the potential for the differentiation of JEG-3-ORGs, organoids were directed towards ST and EVT. During ST/EVT differentiation, the presence of beta-hCG/HLA-G-positive cells at the periphery of JEG-3-ORGs indicated the ST-like or EVT-like nature of the cells ( Figures 2D, E ).

To further characterize the identity of JEG-3-ORGs, a comparison of their transcriptome profiles with those of JEG-3 cells was performed. JEG-3-ORGs and JEG-3 cells formed two clusters via hierarchical clustering ( Figure 3A ), indicating their distinct molecular identities. Using differentially expressed gene (DEG) analysis, 1,141 upregulated and 3,093 downregulated genes were identified in the JEG-3-ORGs relative to the parental JEG-3 cells ( Figure 3B ). JEG-3-ORGs were in a cytotrophoblast-like state, as indicated by a heatmap of marker gene expression. With the exception of DAO and ITGA1, the majority of EVT markers (such as MMP2, ITGA5, PAPPA2, and ERBB2) were downregulated. Except for GDF15, genes belonging to the pregnancy-specific glycoprotein family (PSG) were upregulated as ST markers ( Figure 3C upper). Nevertheless, few TSC and CTB markers were altered (e.g., TP63, EGFR, POU5F1, ITGA6, KRT7, and TFAP2C). Only TE-related genes (e.g., CDX2 and EOMES) were downregulated ( Figure 3C below). HLA-G and HLA-C were downregulated in anticipation. ELF5 was highly expressed, which was consistent with its reported hypomethylated status, whereas ELF3 was weakly expressed, indicating that it was hypermethylated. This change in marker genes indicates that the EVT-like fate of JEG-3 cells in TO-ORGs medium has been predominantly reversed to a CTB-like state. On the basis of the transcriptomic data, we also investigated the strongly involved pathways regulating the organoid formation in JEG-3 cells and found that the DEGs during reprogramming were enriched in GO terms converging on development and metabolic processes ( Figure 3D ), which were consistent with those reported in the primary trophoblast ORGs (5). The top 20 KEGG-enriched signaling pathways included the “MAPK”, “PI3K-Akt”, and “Wnt” pathways ( Figure 3E ). These three pathways also cover most of the activators and inhibitors for the trophoblast organoid medium. In addition, the crosstalk for the top eight enriched development-relevant pathways, represented by a chord diagram, showed that the JEG-3-ORGs were in the germinative stage, with downregulated bHLH transcription factors (such as HAND1, ID1, and MYC) and TGF-beta signaling (e.g., SMAD4) ( Figure 3F ). Altogether, JEG-3-ORGs can be produced in the published trophoblast organoid medium and, to a certain extent, have similar properties to human primary TB-ORGs for further trophoblast functional study.

To generate a JEG-3 cell line devoid of HLA-G, the CRISPR/Cas9 system was used. As the extracellular domain α1 encoded by exon 2 is the common peptide for seven HLA-G protein isoforms (G1 to G7), two small guide RNAs (sg-RNAs) targeting exon 2 and one sg-RNA targeting exon 3 ( Figure 1B ) were designed and used in conjunction to ensure the deletion of nearly all known HLA-G isoforms. After electroporation and drug selection, a successful HLA-G knockout JEG-3 cell line (JEG-3 KO cells) was established, and its efficacy was first validated by PCR and western blotting ( Figure S1A ). Additional functional validation was performed on the well-established functions of HLA-G, including the inhibition of natural killer cell cytotoxicity and the promotion of endothelial cell angiogenesis. When co-cultured with HUVEC, JEG-3 KO cells exhibited impaired tube remodeling ( Figure 4A ); when co-cultured with NK92, JEG-3 KO cells exhibited increased cytotoxicity compared with wild-type JEG-3 cells ( Figure 4B ). Overall, the results demonstrated the successful generation of an HLA-G JEG-3 knockout cell line using CRISPR/Cas9, which could be used for future functional studies of HLA-G.

In addition to examining the effect of HLA-G on proliferation and invasion ability in JEG-3 KO cells using wild-type JEG-3 as a control, we examined the effect of HLA-G on proliferation in JEG-3 KO cells. The CCK8 proliferation assay revealed no significant difference between the JEG-3 cells of the wild-type (WT) and KO genotypes ( Figure S1B ). Likewise, the Transwell invasion assay revealed no significant difference between the KO and WT cells ( Figure S1C ). The mRNA expression of invasion-related genes (MMP2, MMP9, TIMP1, and TIMP2) was also unaffected by the absence of HLA-G ( Figure S1D ). To compare the global molecular profile, we sequenced the transcriptomes of both JEG-3 KO and wild-type cells. Analysis of differential gene expression revealed only a few significant genes, indicating that HLA-G knockout has a minimal effect on the functions of JEG-3 cells ( Figure S1E ).

In this section, we used the CRISPR/Cas9 system to generate a validated molecular and functional HLA-G knockout JEG-3 cell line. Transcriptome and functional assay analysis of the HLA-G KO model demonstrated that HLA-G knockout had a minimal effect on the proliferation and invasion of JEG-3 cells. However, when co-cultured with NK92 and HUVEC, the effect of HLA-G knockout on cytotoxicity and angiogenesis was significant, indicating that HLA-G is a molecule that has an effect that is dependent on its coupling receptors, which may in turn indirectly affect JEG-3 cells.

With JEG-3 KO cells, we attempted to generate trophoblast organoids in the same manner as JEG-3 WT cells ( Figure 1A ). Interestingly, the JEG-3 KO cells-derived ORGs (KO-ORGs) exhibited similar morphology to WT-ORGs ( Figures 5A, B ). The KO-ORGs also expressed GATA3, TFAP2A, and TFAP2C ( Figure 5C ) and had the ability to differentiate into ST cells, as indicated by beta-hCG ( Figures 5D upper, 5E ). Additionally, upon directed EVT differentiation, migration was observed in KO-ORGs ( Figure 5D below). As HLA-G, the commonly used EVT marker, was the investigated target that was knocked out in our study, PAPPA2, another traditional EVT marker, was used instead to confirm the differentiation of EVTs from organoids ( Figure 5F ). When comparing differentiated EVT cells from KO-ORGs and WT-ORGs, there was no significant difference in PAPPA2 protein expression ( Figure 5G ).

The formation of JEG-3-ORGs without HLA-G was further characterized by comparing the transcriptomic profiles of JEG-3 KO cells and their organoids. Two distinct clusters were revealed by hierarchical clustering ( Figure S2A ). Between these two groups, a total of 3,675 DEGs (1,227 upregulated genes and 2,448 downregulated genes) were identified ( Figure S2B ). Except for DAO, PSG2, PSG6, and PAPPA2, the heatmap of ST- and EVT-characteristic genes exhibited similar changes to the formation of organoids from JEG-3 WT cells ( Figure S2C upper). With the exception of EGFR, GATA3, and TFAPAPA, clusters of TSC and CT markers, which were mentioned previously in WT-ORGs, also exhibited similar changes ( Figure S2C below). Nonetheless, an integrated GO term analysis revealed that the formation of KO-ORGs could be approximately summarized as the biological process of development and signal pathway response ( Figure S2D ), indicating that JEG-3 KO cells were activated by stress during reprogramming. The top 20 KEGG-enriched pathways for the formation of KO-ORGs are also involved in the “MAPK” and “PI3K-Akt” signaling pathways, except for the “Wnt” pathway ( Figure S2E ). Additionally, several regulatory components of the mutual signaling pathway were somewhat distinct in KO-ORGs ( Figure S2F ). In MAPK signal, FASLG, rather than FAS, TP53 and SRF, is regulated in KO-ORGs. In addition, JAG1 and JAG2, not the PSENEN, were regulated in KO-ORGs in response to the Notch signal. Overall, the findings demonstrated that JEG-3-ORGs could be generated from HLA-G knockout JEG-3 cells in the identical culture medium.

Comparison of the KO-ORGs and WT-ORGs via transcriptome sequencing revealed no significant difference. The similarity was demonstrated by a volcano plot ( Figure 6A ). The heatmap for the TSC- and CT-classified gene clusters revealed that their undifferentiated status was comparable ( Figure 6B ). The similar expression of mRNA for TP63, VGLL1, and ITGA6 further supported their identity ( Figure 6C ).

Although similar, a comparison of the different DEGs during reprogramming in the presence and absence of HLA-G could be instructive. The 2,184 DEGs shared between KO-ORGs and WT-ORGs enriched 33 KEGG pathways, which overlapped with the individual analyses ( Figure 6D ). In accordance with the regulatory components of the culture medium, the “MAPK” and “PI3K-Akt” signaling pathways were the dominant developmentally significant pathways, irrespective of the HLA-G ( Figure 6E ). Interestingly, in our JEG-3ORGs, the Rap1 signaling pathway ranked first. In addition, it comprises one of the planar cell polarity-related pathways. To trace how the trophoblast was still able to develop in the absence of HLA-G, a PPI network was constructed using the individual DEGs [1,417 genes (WT group) and 859 genes (KO group)] ( Figure 6F ). The relationship of the top 20 hub gene clusters between the absence of HLA-G (e.g., HIST2H2AC, CDK1, CXCR4, and CIRH1A) and the presence of HLA-G (e.g., CRAT, DAO, TP53, BMP4, and BCL2L11) may be a trigger for compensation mechanisms of HLA-G-relative trophoblast development. Even though DEGs were unequal between HLA-G KO and WT, HLA-G was not required for the formation of JEG-3 -ORGs.

HLA-G is gradually expressed in EVT during early pregnancy and gradually decreases as the pregnancy progresses (13). This leads us to the question of whether HLA-G facilitates the generation of EVTs. Under the cytotoxicity of Cas9 protein and the stress of electrotransfection, the compact structure of trophoblast organoids renders them fragile. Therefore, genetic modification of all TB-ORGs remains a difficult task. Instead, HLA-G knockout TB-ORGs were derived directly from JEG-3 cells that are susceptible to CRISPR-Cas9-mediated editing. First, we directly compared the differentiation process between WT-ORGs and KO-ORGs on the basis of their previously attested nearly equivalent identities ( Figure 1A ). The mRNA expression of conventional EVT identification markers (MMP2, ITGA5, CDH3, and IGF2) following EVT differentiation exhibited no significant changes, indicating that HLA-G knockout had no effect on EVT production ( Figure 7A ). In addition to the knockout, recombinant HLA-G protein was overexpressed through exogenous addition ( Figure 1A ). During differentiation between the KO and HLA-G addition groups, the mRNA expression of EVT-recognized markers (MMP2, ITGA5, ITGA1, ERBB2, MCAM, NOTCH2, PAPPA2, and TCF4) did not change significantly at any concentration (0.025 µg/mL, 0.05 µg/mL, and 0.1 µg/mL) ( Figure 7B ). Briefly, HLA-G played an inconsequential role in EVT differentiation based on JEG-3-ORGs.