BHLHE40 Works with GATA2 and GATA3 to Control the Development of Human Placental Cells

To find novel regulators of STB formation, a biotin‐labeled DNA segment containing the CGB promoter sequence was immobilized onto streptavidin‐coated magnetic beads to capture proteins from the primary first‐trimester villus nuclear extract (Figure1A). The eluted proteins were analyzed by liquid chromatograph/mass spectrometry (LC/MS) analysis. From high‐confidence candidates (FDR < 0.01), the top nine TFs were selected based on relative protein abundance ranking (quantified by chromatographic peak area) (Figure 1B). (A list of all candidates is provided in Table S1, Supporting Information). Next, single‐nucleus RNA sequencing (snRNA‐seq) datasets from first‐trimester trophoblasts^[27^] were analyzed to assess the expression level of these TFs in trophoblast subtypes. With unsupervised clustering based on canonical trophoblast markers, we found that among the nine TFs, only BHLHE40 exhibited statistically higher expression levels in STBs, compared to those in CTBs (Figure 1C). Early studies reported BHLHE40 expression in human placental tissues and its involvement in regulating EVT function.^[28, 29, 30^] However, the potential function of BHLHE40 in syncytialization, that is, STB differentiation, remains unexplored. Based on Uniform Manifold Approximation and Projection (UMAP) visualization analysis, BHLHE40 was enriched in STBs compared to EVTs or CTBs (Figure 1D). Moreover, single‐cell regulatory network inference (SCENIC) analysis predicted that BHLHE40 was a potential driver with an increased regulatory activity for STB identity (Figure S1A, Supporting Information). The analysis also showed that genes potentially targeted by BHLHE40 were enriched in STBs (Figure S1B, Supporting Information).

Next, we validated BHLHE40 expression patterns in human placental tissues. Immunofluorescence (Figure 1E) and immunohistochemistry (Figure S1C, Supporting Information) analyses of first‐trimester villous sections revealed predominant BHLHE40 localization in STBs, with moderate expression in CTBs. It has been reported that some miscarriage cases exhibited impaired STB differentiation, characterized by reduced HCG synthesis^[31, 32, 33, 34^] and syncytin‐1 expression^[35^] in placental villi. Analysis of human villi revealed reduced BHLHE40 expression in some spontaneous miscarriages compared to normal pregnancy controls (Figure S1D, Supporting Information).

In vitro models also demonstrated BHLHE40 upregulation after STB differentiation. Both mRNA (Figure 1F) and protein (Figure 1G) levels were significantly higher in STBs differentiated from human blastocyst‐derived TSCs than in undifferentiated TSCs. Additionally, given that cAMP signaling is a key pathway in trophoblast syncytialization^[36, 37^] and that cAMP response element like sequences were reported in the human BHLHE40 promoter,^[38, 39^] it seems that BHLHE40 could be a target of cAMP signaling. As expected, treatment of TSCs with db‐cAMP (1 µM) induced BHLHE40 expression (Figure S1E, Supporting Information). Furthermore, this induction was attenuated by cAMP pathway inhibitors H‐89 (PKA inhibitor; 8 µM) and ESI‐09 (EPAC antagonist; 8 µMm) (Figure S1F, Supporting Information). Together, these data suggested a potential role of BHLHE40 in STB differentiation.

To analyze the role of BHLHE40 in STB differentiation, Cleavage Under Targets and Tagmentation (CUT&Tag) ‐seq experiments were done to examine BHLHE40 targets in cultured TSCs and STBs. Peak distribution analysis revealed comparable genomic partitioning patterns of BHLHE40 in TSCs and STBs (Figure , Supporting Information). The BHLHE40 binding was predominantly in promoter regions, followed by introns and distal intergenic regions (Figure , Supporting Information), indicating that BHLHE40 regulated targeted gene expression by affecting both promoter and enhancer activity. Notably, BHLHE40 exhibited comparable global binding intensity between TSCs and STBs (Figure , Supporting Information). Thus, site‐specific variations in occupancy require to be further investigated. S2A S2A S2B

Using MAnorm2 method, differential binding analysis identified peaks with upregulated and downregulated signal intensity in STBs versus TSCs (Figure2A; Figure S2C, Supporting Information). The upregulated peaks in STBs showed preferential localization to transcriptional regulatory regions, with 62% occupying intron and distal intergenic regions and 35% occupying promoter regions (Figure S2D, Supporting Information). In functional annotation through Gene Ontology (GO) analysis, genes associated with the upregulated peaks were predicted to be involved in hormone responsiveness, placental development, and cell polarity (Figure 2B). In contrast, downregulated peaks showed no functional association with STB‐related processes (Figure S2E, Supporting Information). We next carried out a conjoint analysis of differential BHLHE40 sites with two core activation‐associated histone modifications, H3K27ac and H3K4me3, both of which undergo extensive reprogramming during early embryonic development and were linked to the regulation of key gene expression.^[40, 41, 42^] A modest global increase was found in H3K27ac and H3K4me3 signals following STB differentiation (Figure S2F, Supporting Information). Analysis revealed that 65.9% and 39.5% of upregulated BHLHE40 sites overlapped with H3K27ac and H3K4me3 binding sites in STBs, respectively (Figure 2C,D). Consistent with these findings, overlapped peaks between upregulated BHLHE40 and H3K27ac exhibited significant H3K27ac enrichment (Figure S2G, Supporting Information), whereas BHLHE40‐unique sites showed negligible signal. Similar differential enrichment patterns were observed for H3K4me3 at overlapping versus unique sites (Figure S2H, Supporting Information). In contrast, downregulated peaks in STBs had less overlaps with the H3K27ac (40.0%) and H3K4me3 (28.9%) sites (Figure S2I, Supporting Information). Furthermore, similarly increased signals were observed in STBs at BHLHE40 and H3K27ac binding loci, including key regulatory sites near CGB, HSD17B1, ERVW‐1, and ERVFRDF‐1 (Figure 2E). These results suggested that BHLHE40 chromatin occupancy correlated with histone modifications associated with active gene expression during STB differentiation.

To examine the importance of BHLHE40 in STB differentiation, we used CRISPR‐Cas9 to KO the BHLHE40 gene in TSCs. BHLHE40 disruption did not cause noticeable changes in self‐renewal or colony integrity of TSCs (Figure S3A), indicating that BHLHE40 was dispensable for the maintenance of cell stemness under the culture conditions. When wild‐type (WT) and BHLHE40‐KO TSCs were induced to undergo differentiation into STBs, reduced levels of CYP11A1 (related to steroidogenesis) and HCG α and β expression were detected by western blotting in STBs derived from BHLHE40‐KO TSCs (Figure3A). Consistently, parallel experiments using first‐trimester cytotrophoblast‐derived TSCs (TSCs^CT) showed decreased CYP11A1, CGA, and CGB expression in BHLHE40‐KO STBs (Figure S3B, Supporting Information). Immunocytochemistry also showed higher levels of HCG in the STBs from WT TSCs, whereas HCG levels were lower in the STBs from BHLHE40‐KO TSCs (Figure 3B). In RT‐PCR, genes encoding HCG (CGA) or related to STB morphology and function (CYP19A1, HSD17B, ERVV1, ERVV2, and ERVFRD) were expressed at much lower levels in STBs from BHLHE40‐KO TSCs, compared with those from WT TSCs (Figure 3C). This transcriptional downregulation was replicated in BHLHE40‐KO TSC^CT‐derived STBs for CGA, CYP11A1, ERVV2, ERVFRD, and HSD17B1 (Figure S3C, Supporting Information), indicating BHLHE40's essential role in regulating these markers across different TSC lines. Subsequently, we assessed the impact of BHLHE40 deficiency on syncytial fusion. WT STBs exhibited characteristic multinucleated syncytia, whereas BHLHE40‐KO STBs showed impaired fusion with predominant mononuclear CTB‐like cells and sparse multinucleated cells (Figure S3D, Supporting Information). In addition, reduced estrogen, progesterone, and HCG levels were found in the conditioned medium from BHLHE40‐KO TSC‐derived STBs, compared to those from WT TSC‐derived STBs (Figure 3D).

To examine the global transcriptomic changes, we performed RNA‐seq on WT and BHLHE40‐KO STBs, detecting 23910 expressed genes. Comparative analysis identified 122 significantly down‐regulated genes, many of which were STB markers, and 104 significantly up‐regulated genes in BHLHE40‐deficient STBs compared to WT controls (thresholds: |log2FC| > 1, p‐adj < 0.05) (Figure 3E). GO enrichment analysis of the down‐regulated genes indicated a close association with syncytium formation and hormone activity (Figure 3F). Reactome analysis also displayed term enrichments related to glycoprotein and steroid hormones (Figure S3E, Supporting Information). Particularly, the cAMP signaling pathway, crucial in trophoblast syncytialization,^[36^] was enriched in the KEGG analysis of those down‐regulated genes (Figure S3F, Supporting Information). Next, CUT&Tag‐seq experiments were performed to examine BHLHE40 targets in BHLHE40‐KO and WT STBs. Integrative analysis revealed direct BHLHE40 binding at 54.5% of RNA‐seq‐downregulated genes, including CG8 and ERVV‐2 (Figure S3G, Supporting Information), suggesting a possible direct role of BHLHE40 in regulating those genes that were important in STB differentiation and function, e.g., hormone secretion.

Furthermore, CUT&Tag profiling of H3K27ac and H3K4me3 modifications revealed global alterations in these two histone activation marks in BHLHE40‐KO versus WT STBs. Differential peak analysis identified global reductions in both marks, with H3K4me3 exhibiting greater sensitivity (27.59%) compared to H3K27ac (8.83%) (Figure 3G). Next, differential peak analysis of BHLHE40 CUT&Tag data identified genomic loci with reduced occupancy in BHLHE40‐KO versus WT STBs (Figure S3H, Supporting Information). These reduced peaks were designated as highly confident BHLHE40‐binding sites, mitigating non‐specific binding artifacts. These BHLHE40‐binding loci showed significant H3K27ac attenuation but minimal H3K4me3 changes upon BHLHE40‐KO (Figure 3H), indicating a correlation between BHLHE40 chromatin occupancy and H3K27ac deposition. Further investigations are important to determine whether BHLHE40 directly modulates H3K27ac modification. Collectively, these data indicate BHLHE40 as a potential factor of transcription activation during STB differentiation.

To elucidate the role of BHLHE40 in STB differentiation, a systematic analysis of the genome‐wide BHLHE40 binding landscape was conducted. De novo motif analysis of BHLHE40‐binding regions using the HOMER method identified consensus motifs for TFAP2C and GATA3, two TFs involved in trophoblast fate‐determination (Figure4A).^[23^] Previous studies have shown that TFAP2C can interact with YAP1 and TEAD4 to form a complex, regulating the transcription of trophoblast‐specific genes.^[43^] GATA3 has also been found to interact with other TFs, contributing to cell lineage determination.^[44^] To verify whether BHLHE40 directly interacts with TFAP2C and GATA3, we did co‐immunoprecipitation (Co‐IP) assays in STBs differentiated from TSCs in culture. In western blotting, both GATA3 and GATA2 (another GATA TF family member implicated in syncytialization) were detected among proteins co‐precipitated by an antibody against BHLHE40 (Figure 4B). Although TFAP2C immunoprecipitation was also observed, in comparison to the IgG control, its enrichment was much lower than that of GATA2/3 (Figure 4B). These protein‐protein interactions were verified in transfected HEK293T cells expressing HA‐tagged BHLHE40 and FLAG‐tagged GATA2 and GATA3 proteins, showing reciprocal interactions between BHLHE40 and GATA2/3 (Figure S4A,B, Supporting Information).

Next, CUT&Tag‐seq for GATA2 and GATA3 was performed in STBs and integrated with BHLHE40 chromatin occupancy data to determine whether BHLHE40 interacts with GATA2/3 to regulate gene expression. Genome‐wide co‐localization analysis identified co‐occupied loci between BHLHE40 and GATA2/3, some of which co‐localized at STB‐specific genes such as CGB, HSD17B1, and genes associated with cell fusion (Figure 4C). We analyzed GATA2 and GATA3 binding signals at genomic loci exhibiting increased BHLHE40 occupancy in STBs versus TSCs (red sites in Figure 2A). At these BHLHE40‐increased sites, we observed parallel enhancement of GATA2/3 binding signals following STB differentiation, consistent with the established protein‐protein interaction between BHLHE40 and GATA2/3(Figure 4D). Further analyses indicated that 74.6% (8794) and 69.2% (8152) of BHLHE40 peak regions overlapped with those of GATA2 and GATA3, respectively, in STBs (Figure S4C, Supporting Information). Based on GO analysis, those overlapping regions were related to hormone response, hormone signaling pathways, and cell polarity (Figure 4E,F). In contrast, unique BHLHE40 peaks that did not overlap with GATA2 or GATA3 sites (Figure S4D, Supporting Information) were not associated with functional enrichments for STB‐related processes (Figure S4E, Supporting Information). These results suggested that BHLHE40 probably interacted with GATA2/3 to regulate gene expression in STBs.

Additionally, enhanced H3K27ac and H3K4me3 signals were observed at BHLHE40‐GATA2 overlapping loci, compared to the BHLHE40 unique sites, which did not overlap with GATA2 loci (Figure 4G) in STBs. Similar results were found when BHLHE40 and GATA3 peak loci in STBs were analyzed (Figure 4H). Notably, GATA2 unique sites (Figure S4F, Supporting Information, left) and GATA3 unique sites (Figure S4F, Supporting Information, right), both of which did not overlap with BHLHE40, exhibited moderate H3K27ac and H3K4me3 signals, suggesting that BHLHE40 may be involved in H3K27ac and H3K4me3 deposition through GATA2/3 recruitment. Moreover, Integrative Genomics Viewer (IGV) visualization analysis confirmed spatial overlapping peaks of BHLHE40, GATA2, and GATA3 at CGB1/3/5/7/8, HSD17B1, ERVV1, and ERVV2 loci in STBs, exhibiting similar signatures of shape and localization (Figure 4I). Together, these findings indicated that BHLHE40 directly interacted with GATA2/3, which may regulate gene expression, differentiation, and function of STBs.

To examine the potential effects of BHLHE40 KO on GATA family binding landscape and chromatin remodeling, the CUT&Tag‐seq was performed for GATA2 and GATA3 in BHLHE40‐KO STBs. Differential binding analysis revealed 44.30% (70816) of GATA2 and 33.6% (27033) of GATA3 peaks exhibited significant signal reduction in BHLHE40‐KO STBs compared to WT STBs (Figure5A). Notably, the binding signals of GATA2/3 at these attenuated peak sites exhibited an upward trend following normal STB differentiation (Figure S5A,B, Supporting Information), which indicated that BHLHE40‐affected GATA2/3 loci likely represent important regulatory elements for STB differentiation. Among the reduced peaks, 605 GATA2 and 1144 GATA3 sites overlapped with BHLHE40‐binding loci (Figure 5B), suggesting that BHLHE40 deficiency might directly affect GATA2/3 binding profiles at these specific genomic regions in STBs. Further GO analysis indicated that these BHLHE40 and GATA2 (Figure 5C) or GATA3 (Figure 5D) loss overlapping sites were enriched in cellular processes associated with STB differentiation, e.g., responses to hormones, protein localization, and cell differentiation. Genomic distribution analysis showed BHLHE40‐dependent downregulated GATA2 peaks localized preferentially to intron (37%), distal intergenic (31%), and promoter (29%) regions, with analogous partitioning observed for GATA3 (Figure S5C, Supporting Information). These results suggested that BHLHE40 deficiency may alter GATA2/3 binding to promoters as well as enhancers at distal genomic regions.

To test this hypothesis, further CUT&Tag analysis was performed for the BHLHE40‐dependent decreased GATA2/3 peaks in BHLHE40‐KO STBs using WT STBs as a control. Normalized signal quantification indicated that the sites of GATA2 (Figure 5E) or GATA3 (Figure 5F) signal loss were accompanied by a marked reduction in H3K27ac and, to a lesser degree, H3K4me3 peaks (Figure 5E,F). Previously, endogenous retrovirus (ERV)‐derived enhancers were found as transcriptional regulators in human trophoblast syncytialization.^[14^] These genomic elements with long terminal repeat sequences contributed to the evolution of mammalian placentas.^[45, 46, 47^] We analyzed a set of ERV families that are located near STB genes and highly enriched at STB enhancers.^[14^] The GATA2 signal peaks at MER11D and MER84 increased when TSCs were differentiated into STBs, but such an increase was attenuated in the BHLHE40‐KO STBs (Figure 5G, top panel). Similar results were found in GATA3 at MER50 and MER84 (Figure 5G, bottom panel), suggesting that BHLHE40 played a unique role in GATA2/3 binding at these loci. GATA2/3 binding signals at remaining ERV family subtypes exhibited no significant changes upon BHLHE40 loss (Figure S5D, Supporting Information). Moreover, GATA2/3 and H3K27ac peaks at CGB1/5/8 and ERVV2 loci were significantly reduced in BHLHE40‐KO STBs, compared to control WT STBs (Figure 5H). These results suggested that BHLHE40 might alter GATA2 and GATA3 interactions with other genomic elements to regulate gene expression and differentiation of STBs.

TSC lines derived from human blastocysts (B7) and the first‐trimester cytotrophoblasts (CT11) were generously provided as a gift by Dr. Hiroaki Okae (Department of Informative Genetics, Environment and Genome Research Center Tohoku University Graduate School of Medicine). TSCs were cultured in Matrigel‐coated (354277, Corning) plates and maintained in Advanced DMEM/F‐12 medium with 0.2% fetal bovine serum (FBS), 0.3% bovine serum albumin (BSA; Sigma–Aldrich, A9418), 1% ITS‐X (Gibco, 51500), 50 ng mL⁻¹ recombinant human epidermal growth factor (EGF; Peprotech, AF‐100‐15), 2 µMm CHIR99021 (TargetMol, T2301), 0.5 µMm A83‐01 (PeproTech, 90943360), 5 µM Y‐27632 (TargetMol, T1870), 1 µMm SB431542 (TargetMol, T1726), 0.8 mMm valproic acid (TargetMol, T7064), 1.5 mg mL⁻¹ L‐ascorbic acid (TargetMol, T0928), 0.1 mMm β‐mercaptoethanol (Thermo Fisher Scientific, 21985023), and 0.5% penicillin‐streptomycin. For cell passages, TSCs were washed with phosphate‐buffered saline (PBS) and dissociated by TrypLE Express (Thermo Fisher Scientific) for 5 min, followed by seeding at a 1:3 to 1:4 ratio.

For STB differentiation, TSCs were plated on Matrigel‐coated plates. Upon cell adhesion, the culture medium was replaced with differentiation medium consisting of Advanced DMEM/F‐12 with 0.3% BSA, 4% KnockOut Serum Replacement (KSR; Thermo Fisher Scientific, 10828028), 1% ITS‐X, 5 µM Y‐27632, 1 µM dibutyryl cyclic AMP (db‐cAMP), 2 µM forskolin (FSK), and 0.5% penicillin‐streptomycin. Differentiated STBs were harvested after 48 h.

A single‐guide RNA (sgRNA) sequence (CAAGTGTACAAGTCAAGACG) targeting BHLHE40 was cloned into the piggyBac‐EF1A‐Cas9‐U6‐sgRNA‐PuroR plasmid. TSCs were transfected with 2 µg of the Cas9‐gRNA‐expressing plasmid and 1 µg piggyBac transposase‐expressing plasmid using polyethylenimine (PEI) according to the manufacturer's instructions. Selection of transfected cells was done using puromycin (5 µg mL⁻¹) starting three days post‐transfection. Surviving cells were dissociated into single cells and cultured at low density. Colonies from single cells were isolated, expanded, and verified by western blotting. Two independently verified BHLHE40‐KO TSC clones (designated KO‐1 and KO‐2) were utilized for subsequent experiments, with representative data from clone KO‐1 shown in figures. Phenotypic consistency was confirmed in both clones.

This study received approval from the Human Research Ethics Committee of Obstetrics and Gynecology Hospital, Fudan University (No. kyy2023‐17). Written informed consent was obtained from all participants prior to tissue collection. Human placental villous tissues were from first‐trimester pregnant women (gestational age: 6–10 weeks) aged 18–35 years, including those of elective termination with no prior miscarriage history and spontaneous miscarriages based on the absence of fetal cardiac activity and/or abnormal gestational sac development. Villous tissues were dissected from the maternal uterine tissue and rinsed with ice‐cold PBS to remove blood. For protein extraction, tissues were homogenized in radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris‐HCl, pH 7.4, 150 mM NaCl, 0.1% SDS, 1% sodium deoxycholate, and 1% Triton X‐100) with a protease inhibitor cocktail (1x; MCE) and nuclease (Beyotime, China) at a ratio of 100 mg tissue per 1 mL buffer. Homogenates were put on ice for 30 min and centrifuged at 12000 × g for 30 min at 4 °C. The supernatant was collected for further analyses.

A biotinylated DNA probe spanning the human CGB promoter region (1000 bp upstream of the transcription start site) was generated by PCR amplification using primers listed in Table (Supporting Information). The biotinylated probe was incubated with streptavidin‐coated magnetic beads at room temperature (RT) for 1 h. Nuclear extracts were prepared from 2–3 human first‐trimester villus using hypotonic buffer (20 mM Tris‐HCl, pH 7.5, 10 mM NaC1, and 0.1% NP‐40) with rotation at RT for 15 min and then lysed using buffer containing 20 mM Tris‐HCl, pH 7.5, 150 mM NaCl, and 0.5% Triton‐X100 with rotation at 4 °C overnight. The lysates were centrifuged and incubated with the aforementioned DNA‐bead complexes at 4 °C for 4 h with rotation. After incubation, sequential washes were done using low‐ (20 mM Tris‐HCl, pH 7.5, 150 mM NaCl, and 0.1% NP‐40) and high‐ (20 mM Tris‐HC1, pH 7.5, 300 mM NaCl, and 0.1% NP‐40) salt buffer (three washes per buffer condition) to remove nonspecific binding proteins. Captured proteins were eluted with Benzonase (Yeason, China) and subjected to SDS‐PAGE, followed by LC‐MS/MS analysis. Mascot and MaxQuant bioinformatics platforms were used to obtain the candidate list. S2

snRNA‐seq data from human villous tissue were from public datasets (https://Figureshare.com/s/ccec028e5a3ec5b20106↗).^[27^] Quality control and gene quantification were performed to retain high‐quality cells. The function FindVariableGenes (mean.function = FastExpMean, dispersion.function = FastLogVMR) from the Seurat package (v4.0.0) was used to identify the top hypervariable genes (HVGs). Principal component analysis (PCA) was conducted via the RunPCA function for dimensionality reduction. Cell clusters were generated based on HVG expression profiles using the FindClusters function and visualized via UMAP. Trophoblast subpopulations were annotated by computing correlations between cellular expression profiles and canonical trophoblast markers using the SingleR package (v1.4.1).^[74^]

SCENIC analysis was done following the established framework,^[75^] which integrates co‐expression networks, cis‐regulatory motif enrichment, and regulon activity scoring. The workflow began with co‐expression network construction using GRNBoost2, which computed pairwise associations between all TFs and genes based on the normalized snRNA‐seq expression matrix from trophoblast clusters after batch correction with Seurat V4. BHLHE40‐specific interactions were retained only if they met two statistical criteria: an importance score > 3 (quantifying association strength) and classification within the top 5% of all TF‐gene interactions to prioritize high‐confidence links. Next, motif enrichment validation was conducted using RcisTarget with the hg38 database, which scanned conserved motifs in ±10 kb promoter regions. Co‐expression modules were pruned to retain genes exhibiting both co‐expression with BHLHE40 and direct binding potential, as evidenced by a Normalized Enrichment Score (NES) > 3 for BHLHE40‐binding motifs and an area under the curve (AUC) threshold > 0.85. This dual‐evidence approach ensured that only genes with robust regulatory relationships were incorporated into high‐confidence regulons. Finally, regulon activity scoring was performed using AUCell, which quantified BHLHE40 regulon activity in each cell by ranking the expression of its validated target genes.

CUT&Tag was performed following the protocol by Henikoff.^[76^] Briefly, cells were dissociated using Accutase (A1110501, Gibco) at 37 °C for 5 min, washed, and incubated with Concanavalin A‐coated beads (Biolinkedin, L‐1017, China) at RT for 10 min. Cells were then incubated with the primary antibody at 4 °C overnight, followed by 1‐h incubation with an unconjugated secondary antibody (Jackson ImmunoResearch) and 1‐h incubation with pAG‐Tn5 (Novoprotein, M059‐YH01‐01A, China), preloaded with mosaic end adapters. Between incubations, beads bound to cells were washed with buffer. Tagmentation was performed by treating cell‐bound beads with MgCl₂ at 37 °C for 1 h, generating DNA fragments. The fragments were purified using DNA extraction beads (Beyotime, China) according to the manufacturer's protocol. The purified DNA fragments were amplified by PCR (21 cycles) using 5x Multiplex PCR Mix (Novoprotein, E086‐YSAA‐01A) to prepare the CUT&Tag library.

For data analysis, sequencing libraries were processed on a NovaSeq 6000 system, with raw reads filtered and trimmed using Trim Galore (v0.6.10) and aligned to the human reference genome (hg38) using Bowtie2 (v2.3.5).^[77^] Duplicate reads were removed using Sambamba (v0.6.6), and peak calling was performed with MACS2 (v2.2.6).^[77^] Differential peak analysis was conducted using MAnorm2 (v1.2.2) or the DiffBind package (v3.4.11).^[78^] Scatter plots representing differentially bound regions were generated using ggplot2 (v3.4.4) in R (v4.1.2). Motif analysis was conducted using HOMER (v4.9.8) with the findMotifsGenome.pl tool under default parameters. Peak file processing and co‐binding analysis were performed using bedtools (v2.27.1)^[79^] and DeepTools (v3.4.5). Peak annotation and functional enrichment analyses were carried out using the ChIPseeker (v1.30.3) and ClusterProfiler (v4.2.2) packages in R. Genome browser views of binding peaks were visualized using IGV (v2.17.4).

Total RNAs were extracted using the EZ‐press RNA Purification Kit (EZ Bioscience) according to the manufacturer's protocol. Complementary DNA (cDNA) synthesis was performed using a reverse transcription kit (Yeason, China), and cDNAs were diluted prior to analysis. qPCR was conducted using a SYBR Green‐based qPCR kit (Yeason, China) on an Applied Biosystems QuantStudio 6 instrument. Relative mRNA expression levels were quantified using β‐actin (ACTB) as a control, with gene expression changes calculated as fold‐change relative to ACTB. Primer sequences are listed in Table (Supporting Information). S2

All RNA samples were extracted using RNAiso reagent (TaKaRa) and exhibited RNA integrity numbers of 10.0, confirming optimal RNA integrity. Indexed libraries were prepared using the TruSeq Stranded Total RNA Library Preparation Kit (Illumina) following the manufacturer's protocol. Libraries were pooled and sequenced on a NovaSeq 6000 platform. Raw sequencing data were processed using HISAT2 (v2.0.4), including quality filtration, trimming, and alignment to the human reference genome. Gene expression levels were quantified using StringTie (v1.3.0). Differential gene expression analysis was performed using DESeq2 (v1.24), identifying differentially expressed genes (DEGs) based on p‐adj < 0.05 and |log2FC| > 1. Volcano plots were generated from DESeq2 results using the EnhancedVolcano package (v1.8.2) in R. Pathway enrichment analysis was performed using ClusterProfiler (v4.2.2).

Fresh human villi tissues obtained at 8 weeks of gestation were washed with PBS and fixed in 4% paraformaldehyde (Biosharp). Following paraffin embedding, tissues were sectioned into 3 µm‐thick slices. Sections were incubated at 60 °C for 2 h for tissue rehydration, followed by deparaffinization and dehydration through an ethanol gradient (50%, 75%, 85%, 95%, and 100%). Antigen retrieval was performed by heating the slides immersed in citrate buffer or Tris‐EDTA (pH 9.0) and cooling to RT. Endogenous peroxidase activity was quenched by incubation with 3% hydrogen peroxide for 10 min, followed by permeabilization using PBST for 20 min. To block nonspecific binding, sections were incubated with 5% normal goat serum for 30 min at RT, then probed with primary antibody (anti‐BHLHE40, ABclonal; A6534, diluted according to the manufacturer's instructions) at 4 °C overnight. The slides were incubated with secondary antibodies at RT for 30 min. Chromogenic detection was carried out using a DAB kit and counterstained with hematoxylin for 1 min. Following dehydration in an ascending ethanol gradient (50%, 75%, 85%, 95%, 100%) and xylene, the sections mounted with a neutral polymer medium and visualized under optical microscopy.

Cell lysates were prepared using RIPA buffer (50 mM Tris‐HCl, pH 7.4, 150 mM NaCl, 0.1% SDS, 1% sodium deoxycholate, and 1% Triton X‐100) with a protease inhibitor mixture (TargetMol, C0001) and Benzonase (Yeason, China). Protein concentrations were determined using a BCA assay, and 10 µg of each protein sample was separated by 7.5% or 10% SDS‐PAGE and transferred to PVDF membranes (0.45 µm pore size). The membranes were blocked with 5% skim milk in TBST at RT for 1 h, followed by incubation with primary antibodies at 4 °C overnight. The primary antibodies included: anti‐BHLHE40 (Abcam; ab259837), anti‐CGA (ABclonal; A1239), anti‐CGB (ABclonal; A12419), anti‐CYP11A1 (ABclonal; A23954), anti‐GATA2 (Abcam; ab109241), anti‐GATA3 (Abcam; ab199428), anti‐TFAP2C (Abcam; ab218107), anti‐DDDDK tag (Abcam; ab205606), anti‐HA tag (Abcam; ab9110), anti‐tublin (Beyotime; AF2827), anti‐GAPDH (Abcam; ab8245), and anti‐β‐actin (Abcam; ab8226), all of which were diluted according to the manufacturer's instructions. After washing, the membranes were incubated with horseradish peroxidase (HRP)‐conjugated secondary antibodies: goat anti‐rabbit IgG (Proteintech; SA00001‐2) or goat anti‐mouse IgG (Proteintech; SA00001‐1) at RT for 1 h. Protein bands were visualized using GE Healthcare Amersham Imager 600 system.

Cells were harvested from culture plates and lysed in hypertonic buffer (20 mM Tris‐HCl pH 7.5, 300 mM NaCl, and 0.1% Triton X‐100), with a protease inhibitor mixture and Benzonase (Yeason, China) for 10 min at RT with gentle rotation. The samples were centrifuged, and the supernatants were collected. Protein concentrations were determined using a BCA assay. 50 µg of each sample was reserved as input; lysate equivalent to 1 mg protein was incubated with protein A/G magnetic beads (Yeason) pre‐conjugated with primary antibodies, followed by overnight rotation at 4 °C. The primary antibodies used in this experiment were anti‐BHLHE40 (Santa Cruz; sc‐101023), anti‐DDDDK tag (Abcam; ab205606), and anti‐HA tag (Abcam; ab9110), with 1–2 µg antibody per 1 mg of total protein. To remove unbound proteins, the samples were loaded onto an Invitrogen magnetic separator to isolate the antibody‐bound protein complexes, followed by washing five times with TBST. Proteins were eluted using a 1*loading buffer and heated at 100 °C for 5 min. Eluted proteins were analyzed by western blotting.

For immunocytochemistry, cells were seeded in 24‐well plates and cultured with the indicated treatments. For imaging, cells were fixed in 4% paraformaldehyde for 10 min, permeabilized with 0.1% Triton X‐100 for 20 min, and blocked with goat serum at RT for 30 min. After overnight incubation with primary antibody (anti‐HCG, ABclonal; A27815 or anti‐E‐cadherin, Abcam; ab314063) diluted according to the manufacturer's instructions at 4 °C, the cells were incubated with secondary antibodies (YSFluor594 goat anti‐rabbit IgG, Yeason; 33112ES60 or YSFluor 488 rabbit anti‐mouse IgG, Yeason; 33906ES60) at RT for 30 min, followed by DAPI staining for 5 min. Human villous tissues were fixed with 4% paraformaldehyde, paraffin‐embedded, and sectioned into 3 µm slices. Antigen retrieval was performed on the tissue sections using Tris‐EDTA (pH 9.0), and the same immunofluorescence procedure used for the cells was followed. Anti‐BHLHE40 (Proteintech; 17895‐1‐AP) and anti‐HLA‐G (ABclonal; A24425) were used as primary antibodies.

Similar numbers of cells (5×10⁵) were seeded per well in six‐well plates and cultured for two days. After collecting the supernatant, cells were recounted, and data were normalized to cell counts. However, it is important to point out that cell counting might not be accurate for multinucleated cells, which is a limitation. Conditioned media were analyzed using human E2 (YOBIBIO U96‐1049E), progesterone (YOBIBIO U96‐3494E) and HCG (YOBIBIO U96‐1292E) ELISA kits according to the manufacturer's instructions. Briefly, standards of varying concentrations and supernatants were added to the wells and incubated at 37 °C for 50 min. After washing, a biotinylated antibody solution was added and incubated at 37 °C for 50 min. The wells were then incubated with SABC solution and TMB chromogenic solution at 37 °C for 30 and 20 min, respectively, with the final incubation performed in the dark. The reaction was terminated using a stop solution. Optical density (OD) values were measured at 450 nm in a microplate reader for 30 min. Sample concentrations were determined from the OD values and the standard curve.

The software GraphPad Prism (v8.0) was used for data analysis. All data were displayed as the mean ± standard error of the mean (SEM), with n values indicating the number of samples or independent experiments, as described in the figure legends. The data were tested for normality. Two‐tailed paired or unpaired Student's t‐test was used in a two‐group comparison, which was detailed in the figure legends. The Wilcoxon signed‐rank test was used in snRNA‐seq analysis. For all tests, a p‐value < 0.05 was considered statistically significant. Significance was shown using asterisks which were specified in legends.