Single-cell analysis reveals essential lncRNAs regulating human trophoblast lineage differentiation

Abstract

The human placenta sustains pregnancy through intricate trophoblast lineage dynamics that are critical for fetal development and pregnancy success. While studies on protein-coding genes (PCGs) have advanced our understanding of placental biology, the regulatory roles of noncoding RNAs, particularly long noncoding RNAs (lncRNAs), in trophoblast lineage specification and function remain poorly understood. Here, we profile single-cell lncRNA dynamics across human placental development, revealing distinct cell-type- and gestational stage-specific expression profiles. Integrated analysis revealed that lncRNAs modulate histone modification levels at the regulatory regions of target genes via cis-action, thereby regulating the expression of key genes essential for trophoblast differentiation. Functional studies by using in vivo and in vitro models fully identify ECEL1P2–ALPP, SEMA3B–AS1–SEMA3B, and MYCNUT/MYCNOS–MYCN as pivotal regulatory axes driving cytotrophoblast self-renewal, syncytiotrophoblast fusion, and epithelial–mesenchymal transition, respectively, which are essential for trophoblast identity and function. Notably, dysregulation of lncRNA–PCG pairs in pathological pregnancies underscores the clinical relevance of these noncoding networks. Together, our findings uncover an unappreciated layer of lineage-specific noncoding regulation, providing mechanistic insight and potential biomarkers for placental development and associated disorders.

Graphical abstract

Keywords

placenta / lncRNA / protein coding gene / cis-regulation / trophoblasts

Introduction

With the continuous advancement of genomics, it has become clear that, beyond traditional protein-coding genes (PCGs), a vast number of noncoding transcripts play crucial roles in gene regulatory networks. Among these, long noncoding RNAs (lncRNAs) not only outnumber PCGs but also contribute to essential biological functions, including chromosomal stability, cell differentiation, and disease pathogenesis [1]. Compared with mRNAs, lncRNAs exhibit distinct features in transcriptional regulation and processing: their promoters typically recruit fewer transcription factors and epigenetic marks, resulting in generally lower expression levels but stronger cell type specificity [2–5]. Functionally, lncRNAs can fold into specific secondary or tertiary structures that allow them to interact with proteins, participate in chromatin remodeling and transcriptional regulation. Moreover, they can act as cis- or trans-regulatory elements to modulate gene expression over short or long genomic distances [6–8].

At the genomic level, lncRNAs are frequently positioned near or embedded within PCGs, forming diverse spatial configurations. According to the classification proposed by Luo et al. [9], lncRNA–PCG pairs can be categorized into six configurations: antisense head-to-head (XH), antisense tail-to-tail (XT), intronic antisense (XI), exonic overlapping (XO), sense upstream (SU), and sense downstream (SD). These arrangements suggest that physical proximity may shape transcriptional regulation. Indeed, deletion of lncRNA loci often alters the expression of nearby genes, changes that may be attributed to their embedded DNA regulatory elements or the transcriptional process itself [10–13]. However, whether lncRNAs exert genuine cis-regulatory effects through their RNA transcripts remains unresolved and requires further validation. Beyond these mechanistic insights, accumulating evidence from diverse developmental systems highlights the indispensable role of lncRNAs in cell fate decisions. For instance, lncRNAs have been shown to regulate zygotic genome activation (ZGA) during early embryogenesis, to control hematopoietic stem cell emergence, and modulate germ cell differentiation [14–16]. These findings emphasize that lncRNAs are not merely transcriptional noises but critical regulators of development processes.

The human placenta is a transient yet indispensable organ that sustains pregnancy by mediating nutrient and oxygen exchange, endocrine signaling, and immune tolerance at the maternal–fetal interface [17, 18]. Originating from the trophectoderm after implantation [19], placental trophoblasts include cytotrophoblasts (CTBs) that function as progenitors, syncytiotrophoblasts (STBs) that facilitate maternal–fetal exchange, and extravillous trophoblasts (EVTs) that remodel uterine spiral arteries to ensure adequate blood flow [20–22]. The differentiation of CTBs into STBs and EVTs is essential for placental development and maternal–fetal communication [23]. Dysregulation of trophoblast development can result in severe pregnancy complications, including preeclampsia (PE), characterized by defective trophoblast invasion and vascular remodeling [24], and gestational diabetes mellitus (GDM), where maternal hyperglycemia disrupts placental homeostasis [25].

Recent technological advances have significantly improved our understanding of placental biology. The establishment of human trophoblast stem cell (hTSC) lines provides a robust platform for investigating trophoblast proliferation and differentiation [19]. In addition, trophoblast organoids (TOs) allow for physiologically relevant modeling of self-organization and early developmental processes, including embryo implantation [26]. At the molecular level, single-cell transcriptomic analyses have mapped trophoblast heterogeneity and lineage trajectories, revealing key PCG regulatory networks that govern cell fate specification [27]. Although individual lncRNAs, such as H19, MEG3, and TUG1, have been implicated in regulating trophoblast proliferation, invasion, and syncytialization [28–30], the global landscape and functional significance of lncRNAs during early human trophoblast fate decisions remain largely unexplored.

In this study, we systematically analyzed lncRNA expression patterns across human placental development, utilizing single-cell transcriptomic data spanning from the first trimester to term. This resource enables the systematic identification of functionally relevant lncRNAs associated with trophoblast differentiation and maturation. Through molecular validation, we uncovered essential roles for ECEL1P2, SEMA3B–AS1, MYCNUT, and MYCNOS in driving trophoblast stemness maintenance, syncytialization, and invasion, mediated by their ability to act in cis to remodel histone signatures at the regulatory regions of target genes. Collectively, our study reveals an essential layer of lncRNA-mediated regulation in trophoblast lineage specification, offering new insights into human placental development and its associated disorders.

Results

lncRNA expression patterns distinguished by trophoblast subtype and gestational stage

To comprehensively investigate the roles of lncRNAs in human placental trophoblast self-renewal and lineage differentiation, we analyzed single-nucleus RNA sequencing (snRNA-seq) data from six first-trimester (6–9 weeks) and six term (38–39 weeks) placental samples sourced from publicly available datasets [27] (Figs. 1A, 1B and S1A–D). Given that lncRNAs are frequently enriched in the nucleus [31], the use of snRNA-seq enables robust detection of nuclear-retained and nascent lncRNA transcripts, making it particularly suitable for systematic lncRNA profiling. Transcriptomic analysis revealed that lncRNAs constituted 3.47% of all the transcripts, a proportion that remained remarkably consistent across major trophoblast subtypes and different gestational stages (Figs. 1C and S1E). This proportional stability was further corroborated by a re-analysis of published bulk RNA-seq datasets (Fig. 1C) [19], underscoring the consistent and biologically relevant representation of lncRNAs throughout trophoblast development.

Quantification of transcript counts revealed that the number of lncRNAs was much lower than that of mRNAs. While mRNAs exhibited more pronounced fluctuations across cell types and developmental stages (Fig. 1D), the overall expression trends of lncRNAs paralleled those of mRNAs during trophoblast maturation and differentiation. To further dissect the dynamics underlying these expression patterns, we employed RNA velocity analysis. This revealed that the initiation of differentiation, particularly in early-gestation EVTs, was characterized by high kinetic activity to support invasive and migration plasticity [32]. In contrast, the endpoints of trophoblast differentiation displayed a unified trend: a concurrent decline in both transcript abundance and velocity length (Figs. 1D and S1F). This synchronization reflects a global entry into transcriptional quiescence, marking the transition from high transcriptional activity to a stable functional state.

Furthermore, we identified numerous lncRNAs whose expression was restricted to specific trophoblast subtypes and gestational stages (Fig. 1E). To explore the expression characteristics of lncRNAs during human placental development, we classified lncRNAs into six clusters and delineated their dynamic expression landscape in human trophoblasts (Figs. 1F and S1G). Notably, late-gestation CTBs and EVTs were characterized by a particularly abundant set of cell-type-specific lncRNAs. UMAP visualization further highlighted the differential enrichment of key lncRNAs across trophoblast subtypes (e.g. MYCNOS, SEMA3B–AS1, ARHGEF3–AS1, and PAPPA2–AS1) (Fig. 1G). Importantly, the expression patterns of several PCGs closely resembled those of their neighboring lncRNAs, which displayed lineage-specific expression (Figs. 1F and S1H). Gene ontology analysis of PCGs neighboring lncRNAs strongly suggested that they were involved in stage and cell-type-specific regulatory functions, including epithelial morphogenesis regulation in CTBs, immune response in EVTs, and cell fusion and cell metabolism in STBs (Fig. 1H). These results support a model in which lncRNAs may act as cis-regulators and highlight that lncRNAs are closely coordinated with the regulatory networks underlying human trophoblast differentiation.

lncRNAs regulate neighboring PCGs involved in human trophoblast development

To further explore the potential regulatory relationships between lncRNAs and neighboring PCGs in the human placenta, we divided the genome into 5-kb windows and analyzed the coexpression patterns within these regions. Notably, the lncRNA–PCG pairs exhibited significant correlations in expression (Fig. S2A), suggesting potential cis-regulatory interactions. Moreover, we categorized lncRNA–PCG pairs in trophoblasts into six distinct locus configurations according to previous descriptions [9] (Fig. 2A). Among these, XH was the most frequent configuration (33.59%), whereas XO was the rarest (2.19%). The other configurations occurred at comparable frequencies, ranging from 12.08% to 19.98%. The occurrence of all six distributions in the extraembryonic lineage indicates a universal genomic organizational pattern. Notably, among these configurations, both antisense and sense lncRNAs were consistently positively correlated with the expression of their nearest PCGs (Fig. 2B); this correlation was further corroborated across three independent external datasets [33–35] (Fig. S2B–D, Table S1), revealing a pattern that differed from that observed in embryonic cells [14, 16].

Pseudotime reconstruction of the trophoblast lineage revealed three major trajectories during placental development: the maturation of CTB, differentiation from CTBs to EVTs, and differentiation from CTBs to STBs (Fig. S2E). Similarly, the expression dynamics of lncRNA–mRNA pairs with different genomic configurations tended to be positively correlated along these developmental trajectories (Fig. 2C). For example, the XH-type lncRNA NR2F2–AS1was highly expressed in CTBs with the self-renewal regulator NR2F2 [36] (Fig. 2C), suggesting that NR2F2–AS1 plays a role in regulating CTB stemness. In addition, during EVT differentiation, both XT-type GATA2–AS1–GATA2 and SU-type MYCNUT–MYCN were upregulated, suggesting that lncRNAs and PCGs may synergistically contribute to trophoblast invasion [34, 37] (Fig. 2C). Moreover, the transmembrane protein SEMA6A and related XI-type SEMA6A–AS1 (Fig. 2C), which were previously reported to exhibit an inverse expression relationship in hepatocellular carcinoma [38], were synchronously upregulated during STB differentiation, suggesting a specific interaction pattern during human trophoblast development.

We subsequently identified 21 high-confidence lncRNA–PCG regulatory pairs that were active during trophoblast developmental transitions (Fig. 2D), all of which were strongly positively correlated (Fig. S2F). To validate the potential cis-regulation acted by lncRNAs, we selected three lncRNAs that were active according to the three major trophoblast developmental trajectories discussed above (Fig. S2E) and explored the effects of lncRNA knockdown on the transcription of related PCGs using a well-established in vitro hTSC model (Fig. S3A). Consistent with their enrichment in CTB in vivo (Fig. 2D), the lncRNA ECEL1P2 and its neighboring gene ALPP were actively co-expressed in in vitro hTSCs, distinguishing the trophoblast lineage from human embryonic stem cells (hESCs) (Fig. S3B). The knockdown of ECEL1P2 led to decreased ALPP mRNA (Fig. 2E). In addition, the lncRNA SEMA3B–AS1 and its neighboring gene SEMA3B were both upregulated during STB differentiation (Figs. 2D and S3C), and SEMA3B–AS1 deficiency significantly downregulated the expression of SEMA3B (Fig. 2F). Interestingly, in addition to SU-type MYCNUT–MYCN, MYCN also showed a close relationship with another antisense lncRNA MYCNOS, and these genes were highly expressed during EVT differentiation (Figs. 2D and S3D). Consistently, both MYCNUT and MYCNOS deficiency impaired the expression of MYCN (Fig. 2G and 2H). Mechanistically, chromatin immuno precipitation-qPCR (ChIP-qPCR) analysis revealed that knockdown of these lncRNAs significantly disrupted the local epigenetic landscape at their respective target loci (Fig. S4). The diverse and complex alterations observed among active (H3K9ac, H3K4me3) and repressive (H3K27me3) marks indicate that these lncRNAs are essential for organizing and maintaining locus-specific chromatin states, likely acting as critical structural or regulatory components within the local chromatin architecture. These findings provide epigenetic evidence for lncRNA-mediated cis-regulatory control through the remodeling of local histone signatures in the human placental trophoblast lineage.

lncRNA deficiency impairs trophoblast self-renewal ability

Differential expression analysis revealed extensive lncRNA reprogramming during trophoblast development (Fig. S5A). Further functional enrichment analysis of the neighboring PCGs of these differential lncRNAs strongly linked them to placental developmental programs, including stemness maintenance regulation during CTB maturation, oxidative stress response during EVT differentiation, and metabolic regulation during STB differentiation (Fig. S5B). To evaluate their relevance to placental pathologies, we analyzed the expression of these lncRNA–PCG regulatory pairs in placental tissues from patients with pathological pregnancies, including those with PE and GDM [39, 40]. Compared with healthy controls, approximately one-quarter of these pairs exhibited significant dysregulation (Fig. S5C), highlighting the potential involvement of these lncRNA–PCG axes in placental development and pregnancy health.

To further determine and elucidate the regulatory roles of specific lncRNAs, we investigated the functional consequences of lncRNA deficiency using stem cell models (Fig. 2E–H). We first focused on ECEL1P2, an antisense lncRNA embedded within the ALPP locus and enriched in stem-like CTBs (Figs. 2D and S3B). While key transcriptional regulators such as TFAP2C and ELF5 were not significantly affected (Fig. S6A), ECEL1P2 knockdown led to the attenuation of epithelial traits, accompanied by loss of CDH1 and EGFR expression (Figs. 3A, 3B and S6B). In addition, ECEL1P2 deficiency resulted in reduced hTSC proliferation (Fig. 3C), which coincided with the significant downregulation of TP63 (Figs. 3D and S6B), a master regulator of CTB self-renewal [41]. Colony formation assay further confirmed a significant decline in both the number and size of colonies after ECEL1P2 knockdown (Fig. 3E–G). This was further supported by decreased expression of the proliferation marker MKI67 (Fig. S6C and S6D), indicating compromised self-renewal capacity.

To assess cell viability and elucidate mechanisms underlying growth defects, we performed further functional analyses of ECEL1P2-deficient hTSC. EdU incorporation assays validated the reduced proliferative activity (Figs. 3H and S6E), while cell cycle profiling revealed G0/G1 arrest and a decreased proportion of cells in the M phase (Fig. S6F). These changes were accompanied by upregulation of the cell cycle inhibitor CDKN1C and downregulation of CDK1, a key cell cycle promoter (Fig. S6G and S6H). These findings suggest a cell cycle blockade and aberrant cellular state. Further examination revealed elevated expression of TP53 and γH2A.X (Figs. 3I and S6I), marker of DNA damage and cellular stress. However, the stress response machinery appeared to be functionally impaired in ECEL1P2-deficient cells (Fig. S6J), suggesting a failure to adequately resolve stress and maintain homeostasis. This was supported by CCK-8 assays (Fig. 3J), which confirmed significantly reduced cell viability.

Notably, consistent with findings in 2D cultures, 3D TOs derived from ECEL1P2-deficient hTSCs exhibited severely impaired self-renewal, with fewer self-organized structures and the emergence of an abnormal vacuolated-like morphology (Figs. 3K and S6K). Moreover, the in vivo lesions from ECEL1P2-deficient hTSCs displayed prominent cytoplasmic vacuolization and nuclear pyknosis (Fig. 3L), suggesting increased cellular stress and apoptosis in a more physiological context. Together, these results suggest that ECEL1P2 is a critical regulator of the self-renewal regulatory network in hTSC. Its deficiency disrupts cell cycle progression, compromises proliferative capacity, and triggers apoptotic responses, which ultimately impair trophoblast stem cell maintenance.

Specific lncRNAs are essential for trophoblast differentiation

As SEMA3B and MYCN have been implicated in STB and EVT differentiation, respectively, we next investigated whether their associated regulatory lncRNAs influenced trophoblast lineage specification (Fig. 2D).

During STB differentiation, knockdown of SEMA3B–AS1 led to impaired syncytialization, characterized by the persistence of mononuclear cells with enlarged internuclear spaces resembling vacuoles (Fig. S7A). qRT-PCR analysis further supported this phenotype, showing significantly reduced expression of fusion-related genes (ERVV1, ERVV2, ERVW1, and ERVFRD1) (Fig. 4A), which are essential for the syncytialization process [37]. Consistently, immunofluorescence staining revealed a decrease in SDC1 expression (Fig. 4B), and quantitative analysis confirmed a significant reduction in the formation of multinucleated syncytia (Fig. 4C). Using a dual-fluorescence reporter assay, we further validated the impaired syncytialization after SEMA3B–AS1 knockdown (Figs. 4D, 4E and S7B). Moreover, knockdown of SEMA3B–AS1 markedly compromised hormone-producing capacity (Fig. 4F and 4G). ELISA analysis further revealed an approximately 50% reduction in human chorionic gonadotropin (hCG) secretion (Fig. 4H). These findings were further corroborated by 3D STB differentiation models and TOs, both confirmed that loss of SEMA3B–AS1 significantly impaired syncytialization and functional maturation of STBs (Figs. 4I, 4J and S7C–F). Together, these results highlight SEMA3B–AS1 as a critical regulator of STB formation and endocrine function during placental development.

During EVT differentiation, individual knockdown of MYCNUT or MYCNOS perturbed epithelial-to-mesenchymal transition (EMT) and downregulated the expression of canonical EVT markers (Fig. S8A and S8B). Interestingly, despite being transcribed in opposite orientations, knockdown of either lncRNA not only reduced the expression of the target gene MYCN but also altered the expression of the other lncRNA (Fig. S8C), suggesting they may exert similar functions in regulating EVT formation. To further investigate this potential synergy, we performed a double knockdown of both lncRNAs (Fig. S8D), which led to a consistent and significant delay in EMT during EVT differentiation (Fig. 5A). Notably, qRT-PCR and immunofluorescence staining revealed a failure to exit the epithelial state, as evidenced by sustained high levels of epithelial cell adhesion molecule (EPCAM) and cadherin-1 (E-cadherin) expression in the knockdown cells (Figs. 5B and S8E). Interestingly, although prolonged stimulation eventually induced an EVT-like morphology (Fig. S8F), functional maturation remained compromised. Specifically, expression of HLA-G, a key regulator of immune tolerance at the maternal–fetal interface, was markedly suppressed in the knockdown group (Fig. 5C). Moreover, MMP2, a crucial effector of cell invasion and migration, was greatly downregulated (Fig. 5D).

Collectively, these data demonstrate that the SEMA3B–AS1–SEMA3B and MYCNUT/MYCNOS–MYCN pairs play critical and distinct roles in regulating STB and EVT differentiation, respectively.

Discussion

The placenta is essential for reproductive success, with the establishment of a robust trophoblast transcriptional program forming the foundation for its functional capacity. While previous studies have extensively characterized the roles of mRNAs and their encoded proteins in regulating trophoblast development, the contributions of noncoding RNAs, particularly lncRNAs, to human trophoblast fate and placental development remain elusive. Here, we integrated single-cell transcriptomic data to systematically characterized the dynamic expression landscape and functional regulatory networks of lncRNAs during human placental trophoblast development. Notably, given the predominant nuclear localization of lncRNAs, the use of snRNA-seq enabled more sensitive detection of these nuclear transcripts. Genomic colocalization analysis combined with functional validation revealed that lncRNAs frequently exert strong cis-activating effects on neighboring PCGs, and play critical roles in regulating human trophoblast self-renewal and lineage specification.

Emerging evidence suggests that lncRNAs often act as cis-elements that regulate adjacent PCGs through coexpression or mutual activation [9]. Consistently, we found that in human trophoblasts, lncRNAs are positively coexpressed with neighboring PCGs, regardless of their genomic configurations, trophoblast subtypes, and gestational stages. Moreover, both lncRNAs and mRNAs exhibited synchronous temporal expression pattern during trophoblast maturation and differentiation, suggesting coordinated regulation or functional coupling. These observations contrast with findings in preimplantation embryos, where lncRNAs are often activated prior to PCGs during zygotic genome activation (ZGA) [15]. The broader developmental span between early and late gestation in postimplantation trophoblasts may warrant finer temporal resolution in future studies.

Functionally, knockdown of specific lncRNAs in hTSCs resulted in reduced transcription of neighboring PCGs, implying that lncRNAs directly promote transcriptional activation rather than posttranscriptional regulation in this context. Prior studies have suggested that lncRNAs contribute to the organization of chromatin state and epigenomic remodeling [42, 43]. Consistent with this regulatory framework, we propose that the identified cis-activating lncRNAs such as ECEL1P2 and MYCNOS may function as molecular scaffolds that recruit histone methyltransferases (e.g. MLL complex via WDR5 interaction) or acetyltransferases to deposit active marks (H3K4me3/H3K9ac) and facilitate enhancer–promoter communication [44–46]. Conversely, for the SEMA3B–AS1 axis, the lncRNA likely operates via a “decoy” mechanism to antagonize the recruitment of Polycomb repressive complexes (PRC2), thereby preventing the deposition of silencing H3K27me3 marks [47, 48]. Our data suggest that these lncRNAs exert positive cis-regulatory effects, essential for maintaining chromatin state at target gene promoters. Notably, human trophoblasts are characterized by unique epigenetic features, including widespread partially methylated domains (PMDs) and a progressive global loss of histone modifications during gestation [49, 50]. In this context, it is plausible that these lineage-specific lncRNAs contribute to the local preservation of active chromatin states at essential developmental loci, thereby sustaining gene expression amidst the unique epigenetic landscape of the placenta.

Consistent with the notion that lncRNAs are highly cell-type-specific [37], our study revealed trophoblast subtype and gestational stage-dependent reprogramming of the lncRNA transcriptome. Functional studies involving lncRNA knockdown demonstrated that their biological significance in both trophoblast maturation and lineage differentiation. The strong positive cis-regulatory influence of lncRNAs suggested their involvement in human trophoblast lineage differentiation, usually mirroring the functions of their neighboring PCGs [8]. For example, ALPP, a placental alkaline phosphatase expressed on the fetal side of the placenta and essential for nutrient transfer [51], was downregulated following knockdown of its antisense lncRNA ECEL1P2, leading to impaired hTSC self-renewal. However, the precise contributions of PCG loss and the hierarchical regulatory relationships among lncRNAs, PCGs, and downstream effectors require further clarification. Moreover, recent studies have showed that lncRNAs with accessible chromatin states are enriched for binding sites of key trophoblast transcription factors, including GATA3, TEAD4, and TFAP2C [52], implicating a model in which these pioneer factors initiate chromatin remodeling, trigger lncRNA transcription, and modulate downstream gene networks. In line with this, a recent study proposed the existence of a TP63–MYCNUT–MYCN axis contributes to EVT differentiation [34], further emphasizing the intertwined roles of transcription factors and lncRNAs in coordinating the trophoblast developmental program.

In summary, our study characterizes the lineage-specific trajectories of lncRNA expression in human placental trophoblasts, and functionally demonstrates that lncRNAs serve as critical regulators of stemness maintenance and differentiation, largely by acting in cis to orchestrate promoter chromatin states of neighboring genes. The dysregulation of lncRNA–mRNA regulatory pairs in placental pathologies (PE and GDM) emphasizes the clinical relevance of these noncoding networks. Our findings offer new insight into the epigenetic and transcriptional control of human placental development, and suggest promising avenues for identifying novel biomarkers and therapeutic targets for pregnancy-associated disorders (Fig. 6).

Research limitations

While this study provides a comprehensive analysis of the dynamic expression patterns and functional roles of lncRNAs in human placental trophoblast development, several limitations should be acknowledged. First, although our transcriptomic data are derived from in vivo placental tissues, the experimental validations were primarily performed using in vitro models, including hTSC and TOs. While these are well-established and physiologically relevant systems, they may not fully recapitulate the complexity of the in vivo microenvironment. Second, the lncRNA−PCG regulatory relationships identified here were largely inferred from coexpression and genomic proximity analyses. Third, our study primarily focused on cis-acting lncRNAs, leaving the potential roles of trans-acting lncRNAs or lncRNA–protein interactions largely unexplored. These aspects warrant further investigation to fully delineate the multilayered regulatory functions of lncRNAs in placental biology.

Methods

snRNA-seq processing and analysis

Raw snRNA-seq reads were processed with Cell Ranger (v8.0.1) using the GENCODE (v32) gene annotation. Downstream analyses were performed in R using Seurat (v5.1.0) Quality control followed Wang [27], retaining nuclei with ≥ 3,500 UMI counts and ≤ 50,000 UMIs, expressing 1,500–5,000 genes, and with mitochondrial gene content < 5%. Potential doublets were removed using Scrublet (v0.2.3) in Python with a threshold of 0.5.

Normalized expression matrices were used for feature selection and dimensionality reduction. The top 2,000 highly variable genes were selected for principal component analysis (PCA). Statistically significant principal components were determined via JackStraw analysis (500 replicates), and the top 30 PCs were used for subsequent analyses. Batch effects were corrected using Harmony (v1.2.3), followed by nonlinear dimensionality reduction with UMAP. Clustering resolution was optimized using clustree and set to 0.7. Cluster-specific marker genes were identified with Seurat’s FindAllMarkers function.

Correlation analysis between lncRNAs and protein-coding genes

Correlation analysis between lncRNAs and PCGs in single-cell RNA-seq data was performed as follows. Single-cell expression matrices were first smoothed using the R package MAGIC (v3.0.0) to impute dropouts and reduce technical noise. Subsequently, Pearson correlation coefficients between each lncRNA and its respective PCGs were calculated using the R function cor.test ().

Pseudotime trajectory analysis

Pseudotemporal ordering was performed using Monocle2 (v2.30.1). UMI count matrices of early trophoblast subtypes (early CTBs, STBs, EVTs, and late CTBs) extracted from Seurat objects were converted into Monocle2 CellDataSet objects. Genes expressed in fewer than 10 cells were excluded. Size factors and dispersion estimates were calculated using Monocle2 default settings. Ordering genes were selected based on Seurat differential expression results, including genes with log₂ (fold change) ≥ 2, adjusted P-value < 0.05, and within the top 20% of expression. Both PCGs and lncRNAs were included. DDRTree (v0.1.5) was used for dimensionality reduction, and cells were ordered along developmental trajectories using the orderCells function.

RNA velocity analysis

To estimate the transcriptional kinetics and directional flow of cell differentiation, RNA velocity analysis was performed using the velocyto and scVelo framework. First, spliced and unspliced expression matrices were generated from the Cell Ranger output BAM files using the velocyto command-line tool (v0.17.17) with the standard run10x logic. The resulting loom files were merged with the Seurat-processed metadata and converted into AnnData format for downstream analysis in Python using scVelo (v0.3.3). Genes were filtered based on minimum shared counts (min_shared_counts = 20) and top likelihood variability. The dynamical model was employed to solve the full transcriptional kinetics of splicing and degradation rates. Velocity vectors were projected onto the pre-computed UMAP embedding. Velocity length—quantified as the length of the velocity vector in the high-dimensional gene expression space—was calculated to reflect the rate of differentiation and transcriptional activity of individual cells.

GO enrichment analysis

Functional enrichment analysis was performed with ClusterProfiler (v4.10.1).

Bulk RNA-seq processing and differential expression analysis

The transcriptional expression matrix was obtained from Okae [18]. Normalization and differential expression analyses were performed using DESeq2 (v1.42.1).

Experimental model and study participant details

Specific-pathogen-free (SPF) BALB/c nude mice were maintained within the animal facility at Tongji University (Shanghai, China). The animals were housed under standard environmental conditions, which included a 12 h light/12 h dark cycle, and were provided ad libitum access to standard chow and water.

Culture of hTSC

hTSC were generated following the published protocol [18]. Briefly, a 35 mm dish was coated with 5 μg/mL collagen IV (Sigma-Aldrich, C7521) at 37 ℃ for at least one hour and a density of 8 × 10⁴–10 × 10⁴ cells were seeded onto the precoated dish. Cells were then cultured in 2 mL of standard TS medium: DMEM/F12 (Thermo Fisher Scientific, 11320033) supplemented with 0.1 mM 2-mercaptoethanol (Thermo Fisher Scientific, 21-985-023), 0.2% Fetal Bovine Serum (FBS) (Gibco, 16000-044), 0.5% Penicillin-Streptomycin (Thermo Fisher Scientific, 15140163), 0.3% bovine serum albumin (BSA) (Bovine Albumin, 199896), 1% ITS-X supplement (BasalMedia, S452J7), 1.5 μg/mL L-ascorbic acid (Sigma-Aldrich, A4544), 50 ng/mL EGF (PeproTech, AF-100-15), 2 μM CHIR99021 (Selleck Chemicals, S2924), 0.5 μM A83-01 (Selleck Chemicals, S7692), 1 μM SB431542 (Selleck Chemicals, S1067), 0.8 mM valproic acid (VPA) (Sigma-Aldrich, PHR1061-1G), and 5 μM Y27632 (Selleck Chemicals, S1049) at 37 ℃ in 5% O₂. When cells reached about 80% confluence, they were dissociated with TrypLE (Thermo Fisher Scientific, 12604021) for 10–15 min at 37 ℃ and passaged at 1:3–1:4 split ratio. The TSC medium was replaced every day.

Plasmid and cell line construction

To construct hTSC-membrane-EGFP/mCherry (hTSC^EGFP+/hTSC^mCherry+) cell lines, the plasmids FUGW-membrane-EGFP or FUGW-membrane-mCherry were introduced by lentivirus infection. After 48 h of infection, the EGFP- or mCherry-positive cells were sorted and collected for further culture.

To establish stable lncRNA-knockdown hTSC lines, we utilized the pLKO.1-TRC lentiviral vector system. The plasmid was linearized by double digestion with AgeI (New England Biolabs, R3552L)/EcoRI (New England Biolabs, R3101S) restriction endonuclease and purified for subsequent cloning. Using an online website DSIR (biodev.extra.cea. fr/DSIR/DSIR.html) to design multiple shRNA sequence targets, annealed to synthesize oligonucleotides, and connected to the pLKO.1-TRC vector. For lentiviral production, the constructed shRNA plasmids were co-transfected with psPAX2 and pMD2.G into HEK293T cells using Vigofect (Vigorous, T001) transfection reagent, following the manufacturer’s protocol. Viral supernatants were collected at 48–72 h posttransfection, filtered (0.45 μM), and concentrated by centrifugation (4800 rpm, 30 min, 4 ℃). For hTSC infection, cells were seeded at about 50% confluency and transduced with shRNA lentivirus for 6–8 h. Then the medium was replaced, and 1 μg/mL puromycin was added for stable selection. An empty pLKO.1-TRC vector was used as a negative control to account for off-target effects. The sequences of the oligonucleotides are provided in Table S2.

Differentiation of hTSC

Differentiation of hTSC toward EVT and STB was performed as previously described [18]. Briefly, for induction of EVT (iEVT), a density of 7.5 × 10⁴ hTSC was seeded in a 35 mm dish precoated with 1 μg/mL collagen IV. Cells were cultured in 2 mL of EVT medium [DMEM/F12 supplemented with 0.1 mM 2-mercaptoethanol, 0.5% Penicillin–Streptomycin, 0.3% BSA, 1% ITS-X supplement, 100 ng/mL NRG1 (Sino Biological, 13499-H08H-100), 7.5 mM A83-01, 2.5 mM Y27632, and 4% KnockOut Serum Replacement (KSR) (Thermo Fisher Scientific, 10828028)], and Matrigel (Corning, 356231) was added to the medium containing the cells at a final concentration of 2%. On the third day, replace the fresh EVT medium without NRG1 and add Matrigel to a final concentration of 0.5%. On the sixth day, the cells were reached about 80% confluence, they were dissociated with TrypLE for 15–20 min at 37 ℃ and passaged at 1:2 split ratio. Suspended the cells with EVT medium lacking NRG1 and KSR, add Matrigel to a final concentration of 0.5%, and cultured for another two days.

For induction of 2D-STB (iSTB-2D), a density of 1 × 10⁵ hTSC was seeded in a 35 mm dish precoated with 2.5 μg/mL collagen IV. Cells were cultured in 2 mL of STB-2D medium [DMEM/F12 supplemented with 0.1 mM 2-mercaptoethanol, 0.5% Penicillin–Streptomycin, 0.3% BSA, 1% ITS-X supplement, 2.5 μM Y27632, 2 μM forskolin (Selleck Chemicals, S2449), and 4% KSR]. The STB-2D medium was replaced at Day 3, and the cells were collected at Day 6 for analysis.

For induction of 3D-STB (iSTB-3D), a density of 2.5 × 10⁵ hTSC was seeded in a 6 cm Petri dish. Cells were cultured in 3 mL of STB-3D medium [DMEM/F12 supplemented with 0.1 mM 2-mercaptoethanol, 0.5% Penicillin–Streptomycin, 0.3% BSA, 1% ITS-X supplement, 2.5 μM Y27632, 50 ng/mL EGF, 2 μM forskolin, and 4% KSR]. After 3 days of culture, add an equal amount of fresh STB-3D medium into the dish. The cells were collected by using 40 μM filter to remove dead cells at Day 6. Both iEVT and iSTB were cultured under the condition of 37 ℃ and 5% O₂.

To quantitatively assess cell fusion and STB formation, 1 × 10⁵ hTSC^EGFP+ and 1 × 10⁵ hTSC^mCherry+ cells were thoroughly mixed and co-seeded into a 35 mm dish containing STB-2D differentiation medium. The medium was replaced with fresh media after 24 h. After further culture 48 h, cells were dissociated with TrypLE for 15–25 min and analyzed by flow cytometry. The double-positive cells were defined as fused STBs.

TOs formation and cultivation

For the generation of TOs, cells were dissociated with TrypLE for 10–15 min and collected by centrifugation at 1100 rpm for 4 min. Subsequently, 3 × 10³ cells were resuspended in 25 μL of ice-cold Matrigel and mixed gently. Pipette the drop into the center of each well of a 48-well tissue culture plate and place in the incubator to set for at least 15 min. TOs were cultured in 250 μL of TO medium (TOM): Advanced DMEM/F12 (Thermo Fisher Scientific, 12634010), 1×N2 (Invitrogen, 17502048), 1×B27-VA (Thermo Scientific, 12587010), 1.25 mM N-acetyl-L-cysteine (Sigma-Aldrich, A9165), 2 mM L-glutamine (Millipore Sigma, TMS-002-C), 50 ng/mL EGF, 1.5 μM CHIR99021, 80 ng/mL Recombinant human R-spondin-1 (PeproTech, 120-38), 100 ng/mL Recombinant human FGF-2 (PeproTech, 100-18B-50), 50 ng/mL HGF (Sino Biological, 10463-HNAS-100), 500 nM A83-01, 2.5 μM prostaglandin E2 (Selleck, S3003), 2 μM Y27632, 100 μg/mL primocin (Invivogen, ANT-PM-1). The TOM was replaced every 3 days.

Measurement of hCG

To measure hCG concentration in the differentiation of iSTB-2D and iSTB-3D, the medium was collected at Days 3 and 6. To measure hCG concentration in the TO formation, the medium was collected at Days 9 and 13. Then the medium was centrifuged at 1100 rpm for 5 min to remove cell debris. The supernatant was harvested and then they were processed for hCG level by Shanghai Enzyme-Linked Biotechnology Co., Ltd.

Cell proliferation assay

Cell growth and viability were evaluated utilizing the Cell Counting Kit-8 (CCK-8). Cells were plated into 96-well microtiter plates at a density of 2 × 10³ cells per well and cultured over consecutive days. At a consistent time point each day, 10 μL of the CCK-8 reagent (MCE, HY-K0301) was introduced into each well, followed by a 2 h incubation period. Absorbance was subsequently measured at 450 nm using a Spark^® microplate reader (Tecan Trading AG, Switzerland) to quantify cell viability.

Flow cytometry

For epithelial characteristics markers analysis, cells were dissociated by TrypLE and treated with APC-conjugated anti-CDH1 antibody (BioLegend, 147311) or APC-conjugated anti-EGFR antibody (BioLegend, 352906) for 30 min on ice. Then washed with Dulbecco’s phosphate-buffered saline (DPBS) (Servicebio, G4200) containing 1% BSA for three times and determined by CytoFLEX S (Beckman Coulter, USA).

Cell cycle assay

Cell cycle assay was performed using propidium iodide (PI) (Beyotime, ST511) in accordance with the manufacturer’s instruction. Briefly, harvested cells were fixed overnight in 70% ethanol at 4 ℃. After fixation, the cells were washed three times with DPBS and subsequently stained in the dark with a solution containing RNase A (500 μg/mL) (TIANGEN, RT405) and PI (50 μg/mL) for 30 min at 37 ℃. Following staining, the cells were centrifuged at 1100 rpm for 5 min, resuspended in DPBS, and subjected to flow cytometry analysis.

Colony formation assay

hTSC were seeded in 12-well plates precoated with 7.5 μg/mL collagen IV (5 × 10³ cells per well) and cultured for 96 h. After the supernatant was removed and gently washed three times with DPBS, the cells were fixed with formaldehyde for 10 min. Subsequently, the cells were stained with crystal violet (Beyotime, C0121) for 10 min and washed with DPBS. After natural air-drying, the culture plate was scanned, the colony number and colony size were calculated using ImageJ.

EdU assay

EdU (10 μM) was added to the medium during the period of vigorous cell division and treated for 2 h. Then cells were harvested for fixation, permeabilization, and the Click reaction using Azide 647 (Beyotime, C0081S). 4′,6-diamidino-2-phenylindole (DAPI) (Beyotime, C1002) was used for DNA count according to the manufacturer’s recommendations. EdU-positive cells were quantified by confocal microscopy (ZEISS LSM880) or flow cytometry.

Quantitative RT-PCR

Total RNA was isolated using RNAiso plus (Takara, 9109) and reverse-transcribed using 5× All-In-One RT Master Mix (ABM, G490) according to the manufacturer’s recommendations. Quantitative RT-PCR was performed using SYBR Premix Ex Taq II (Takara, RR820B) and signals were detected with QuantStudio3 (Thermo Fisher Scientific, A28136). Data were normalized to GAPDH expression with 2^−ΔΔCT method (n = 3). Details of the primers are provided in Table S2.

Immunofluorescence

Cells were seeded on Matrigel-coated glass slides and fixed with 4% paraformaldehyde (PFA) (Servicebio, G1101) for 10 min at room temperature. After three times wash with TBST (Epizyme, PS103), cells were permeabilized with 0.5% Triton X-100 (Sigma-Aldrich, T8532) for 15 min. The samples were blocked with 2.5% BSA at 25 ℃ for 1 h and then incubated overnight at 4 ℃ with the following primary antibodies: TP53 (Santa Cruz, sc126), γH2A.X (Abcam, ab81299), Ki67 (Cell Signaling Technology, 9129 T), TP63 (Abcam, ab124762), SDC1 (Proteintech, 67155-1-Ig), CGB (Abcam, ab9582), E-Cadherin (Cell Signaling Technology, 31955), HLA-G (Abcam, ab52454), MMP2 (Cell Signaling Technology, 40994), or EPCAM (Invitrogen, MA512436). Cells were then washed three times with TBST and incubated with secondary antibodies for 1 h at room temperature. The nuclei were stained with DAPI. Confocal imaging was performed with Zeiss LSM 880 confocal microscopes and analyzed with Zeiss Zen Blue Edition.

Western blot

Cells were lysed in RIPA lysis buffer (Beyotime, P0013B) with protease inhibitor cocktail (Roche, 04693132001), then denatured in loading buffer by heating at 95 ℃ for 10 min. Denature protein were separated by 10% SDS–PAGE gels (Epizyme, PG112) in Tris-glycine running buffer (Epizyme, PS105) and transferred onto PVDF membranes (Millipore Sigma, IPVH00010). The membrane was blocked with 5% BSA in TBST for 1 h at room temperature and then incubated overnight at 4 ℃ with the following primary antibodies: CDK1 (Santa Cruz, sc-54), GAPDH (Proteintech, 60004). Following extensive washes with TBST, the membrane was incubated with an HRP-conjugated goat anti-mouse IgG secondary antibody (Epizyme, LF101) for 1 h at room temperature. Signal detection was performed using a SuperSignal West Pico PLUS (Thermo Scientific, 34580) and visualized with a ChemiDoc MP Imaging System. The intensity of the bands was quantified using ImageJ software.

Engraftment of hTSC into BALB/C nude mice

5 × 10⁶ hTSC were dissociated with TrypLE for 10–15 min and resuspended in 100 μL of a mixture (Matrigel and DMEM/F12 containing 0.3% BSA and 1% ITS-X supplement at 1:2 split ratio). The suspensions were subcutaneously injected into 6- to 8-week-old male BALB/C Nude mice. Lesions were collected a week after injection and fixed with 4% PFA overnight at 4 ℃. The lesions were sent to Wuhan Servicebio Technology Co., Ltd for anti-KRT7 (Abcam, ab68459) immunohistochemistry and scanned with Olympus VS200.

Chromatin immunoprecipitation-qPCR

Following a previously established protocol [53]. ChIP-qPCR assays were conducted utilizing 1 × 10⁴ cells per reaction. Briefly, harvested cells were resuspended in a nuclear extraction buffer, and the chromatin was sheared by treatment with MNase for 10 min at 25 ℃. The sheared chromatin samples were then subjected to overnight immunoprecipitation at 4 ℃ with 1 μg of specific antibodies conjugated to 10 μL of Protein A Dynabeads (Life Technologies). Both the immunoprecipitated (IP) DNA and input DNA were isolated via phenol–chloroform extraction and subsequent ethanol precipitation. Prior to quantitative PCR analysis, the purified DNA was diluted 3-fold. The following primary antibodies were utilized: anti-IgG (Vazyme, Ab207-01), anti-H3K4me3 (Cell Signaling Technology, 9751S), anti-H3K27me3 (Diagenode, C15410195), and anti-H3K9ac (Abcam, ab4441). All primer sequences employed for this assay are provided in Table S2.

Research ethics

All in vivo animal procedures and husbandry practices adhered strictly to the Health Guide for the Care and Use of Laboratory Animals. The animal experimental protocols were formally reviewed and approved by the Biological Research Ethics Committee of Tongji University.

The collection of human placental tissue for hTSC derivation was approved by the Ethics Committee of Shanghai First Maternity and Infant Hospital (Approval No. KS23268). Written informed consent was obtained from all enrolled subjects prior to their participation, and the study was conducted in accordance with the principles of the Declaration of Helsinki.

Statistical analysis

Statistical analyses were performed using GraphPad Prism (version 10.3.0) or R (version 4.3.1). Data are presented as mean ± SD. Group comparisons were conducted using Student’s t-test or two-way ANOVA, while Wilcoxon rank-sum tests were applied for snRNA-seq data. Sample sizes are detailed in the figure legends. P < 0.05 was considered statistically significant (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant).

References

[1] Djebali S, Davis CA, Merkel A et al. Landscape of transcription in human cells. Nature 2012;489:101-8.

[2] Melé M, Mattioli K, Mallard W et al. Chromatin environment, transcriptional regulation, and splicing distinguish lincRNAs and mRNAs. Genome Res 2017;27:27-37.

[3] Marques AC, Ponting CP. Intergenic lncRNAs and the evolution of gene expression. Curr Opin Genet Dev 2014;27:48-53.

[4] Andersson R, Gebhard C, Miguel-Escalada I et al. An atlas of active enhancers across human cell types and tissues. Nature 2014;507:455-61.

[5] Qin T, Zhang H, Zou Z. Unveiling cell-type-specific mode of evolution in comparative single-cell expression data. J Genet Genomics 2026;53:28-42.

[6] Rinn JL, Kertesz M, Wang JK et al. Functional demarcation of active and silent chromatin domains in human HOX loci by noncoding RNAs. Cell 2007;129:1311-23.

[7] Zhao J, Sun BK, Erwin JA et al. Polycomb proteins targeted by a short repeat RNA to the mouse X chromosome. Science 2008;322:750-6.

[8] Engreitz JM, Haines JE, Perez EM et al. Local regulation of gene expression by lncRNA promoters, transcription and splicing. Nature 2016;539:452-5.

[9] Luo S, Lu JY, Liu L et al. Divergent lncRNAs regulate gene expression and lineage differentiation in pluripotent cells. Cell Stem Cell 2016;18:637-52.

[10] Lai F, Orom UA, Cesaroni M et al. Activating RNAs associate with mediator to enhance chromatin architecture and transcription. Nature 2013;494:497-501.

[11] Anderson KM, Anderson DM, McAnally JR et al. Transcription of the non-coding RNA upperhand controls Hand2 expression and heart development. Nature 2016;539:433-6.

[12] Yin Y, Yan P, Lu J et al. Opposing roles for the lncRNA haunt and its genomic locus in regulating HOXA gene activation during embryonic stem cell differentiation. Cell Stem Cell 2015;16:504-16.

[13] Paralkar VR, Taborda CC, Huang P et al. Unlinking an lncRNA from its associated cis element. Mol Cell 2016;62:104-10.

[14] Zhou J, Xu J, Zhang L et al. Combined single-cell profiling of lncRNAs and functional screening reveals that H19 is pivotal for embryonic hematopoietic stem cell development. Cell Stem Cell 2019;24:285-98.e5.

[15] Chen K, Liu W, Zhu J et al. Pivotal role for long noncoding RNAs in zygotic genome activation in mice. Sci China Life Sci 2024;67:958-69.

[16] Wang N, He J, Feng X et al. Single-cell profiling of lncRNAs in human germ cells and molecular analysis reveals transcriptional regulation of LNC1845 on LHX8. Elife 2023;12.

[17] Gude NM, Roberts CT, Kalionis B et al. Growth and function of the normal human placenta. Thromb Res 2004;114:397-407.

[18] Burton GJ, Fowden AL, Thornburg KL. Placental origins of chronic disease. Physiol Rev 2016;96:1509-65.

[19] Okae H, Toh H, Sato T et al. Derivation of human trophoblast stem cells. Cell Stem Cell 2018;22:50-63.e6.

[20] Xiao Z, Yan L, Liang X et al. Progress in deciphering trophoblast cell differentiation during human placentation. Curr Opin Cell Biol 2020;67:86-91.

[21] Roberts RM, Fisher SJ. Trophoblast stem cells. Biol Reprod 2011;84:412-21.

[22] Latos PA, Hemberger M. From the stem of the placental tree: trophoblast stem cells and their progeny. Development 2016;143:3650-60.

[23] Turco MY, Moffett A. Development of the human placenta. Development 2019;146.

[24] Roberts JM, Gammill HS. Preeclampsia: recent insights. Hypertension 2005;46:1243-9.

[25] Catalano PM, Shankar K. Obesity and pregnancy: mechanisms of short term and long term adverse consequences for mother and child. BMJ 2017;356:j1.

[26] Rivron NC, Frias-Aldeguer J, Vrij EJ et al. Blastocyst-like structures generated solely from stem cells. Nature 2018;557:106-11.

[27] Wang M, Liu Y, Sun R et al. Single-nucleus multi-omic profiling of human placental syncytiotrophoblasts identifies cellular trajectories during pregnancy. Nat Genet 2024;56:294-305.

[28] Xu Y, Ge Z, Zhang E et al. The lncRNA TUG1 modulates proliferation in trophoblast cells via epigenetic suppression of RND3. Cell Death Dis 2017;8:e3104.

[29] Chen Y, Ding H, Wei M et al. MSC-Secreted exosomal H19 promotes trophoblast cell invasion and migration by downregulating let-7b and upregulating FOXO1. Mol Ther Nucleic Acids 2020;19:1237-49.

[30] Zhang J, Liu X, Gao Y. The long noncoding RNA MEG3 regulates Ras-MAPK pathway through RASA1 in trophoblast and is associated with unexplained recurrent spontaneous abortion. Mol Med 2021;27:70.

[31] Jonas K, Calin GA, Pichler M. RNA-Binding proteins as important regulators of long non-coding RNAs in cancer. Int J Mol Sci 2020;21.

[32] Ma J, Liu Y, Guo Z et al. The diversity of trophoblast cells and niches of placenta accreta spectrum disorders revealed by single-cell RNA sequencing. Front Cell Dev Biol 2022;10:1044198.

[33] Li H, Peng H, Hong W et al. Human placental endothelial cell and trophoblast heterogeneity and differentiation revealed by single-cell RNA sequencing. Cells 2022;12.

[34] Ounadjela JR, Zhang K, Kobayashi-Kirschvink KJ et al. Spatial multiomic landscape of the human placenta at molecular resolution. Nat Med 2024;30:3495-508.

[35] Wei P, Dong M, Bi Y et al. Identification and validation of a signature based on macrophage cell marker genes to predict recurrent miscarriage by integrated analysis of single-cell and bulk RNA-sequencing. Front Immunol 2022;13:1053819.

[36] Chen Y, Siriwardena D, Penfold C et al. An integrated atlas of human placental development delineates essential regulators of trophoblast stem cells. Development 2022;149.

[37] Arutyunyan A, Roberts K, Troulé K et al. Spatial multiomics map of trophoblast development in early pregnancy. Nature 2023;616:143-51.

[38] Yu SM, Zhang M, Li SL et al. Long noncoding RNA semaphorin 6A-antisense RNA 1 reduces hepatocellular carcinoma by promoting semaphorin 6A mRNA degradation. World J Gastroenterol 2025;31:102527.

[39] Yang Y, Guo F, Peng Y et al. Transcriptomic profiling of human placenta in gestational diabetes mellitus at the single-cell level. Front Endocrinol (Lausanne) 2021;12:679582.

[40] Zhou W, Wang H, Yang Y et al. Trophoblast cell subtypes and dysfunction in the placenta of individuals with preeclampsia revealed by single‑cell RNA sequencing. Mol Cells 2022;45:317-28.

[41] Knöfler M, Haider S, Saleh L et al. Human placenta and trophoblast development: key molecular mechanisms and model systems. Cell Mol Life Sci 2019;76:3479-96.

[42] Nojima T, Proudfoot NJ. Mechanisms of lncRNA biogenesis as revealed by nascent transcriptomics. Nat Rev Mol Cell Biol 2022;23:389-406.

[43] Herman AB, Tsitsipatis D, Gorospe M. Integrated lncRNA function upon genomic and epigenomic regulation. Mol Cell 2022;82:2252-66.

[44] Fok ET, Davignon L, Fanucchi S et al. The lncRNA connection between cellular metabolism and epigenetics in trained immunity. Front Immunol 2018;9:3184.

[45] Aho ER, Wang J, Gogliotti RD et al. Displacement of WDR5 from chromatin by a WIN site inhibitor with picomolar affinity. Cell Rep 2019;26:2916-28.e13.

[46] Liu S, Jiao B, Zhao H et al. LncRNAs-circRNAs as rising epigenetic binary superstars in regulating lipid metabolic reprogramming of cancers. Adv Sci (Weinh) 2024;11:e2303570.

[47] Bonitto K, Sarathy K, Atai K et al. Is there a histone code for cellular quiescence? Front Cell Dev Biol 2021;9:739780.

[48] Bohnsack JP, Teppen T, Kyzar EJ et al. The lncRNA BDNF-AS is an epigenetic regulator in the human amygdala in early onset alcohol use disorders. Transl Psychiatry 2019;9:34.

[49] Zhang B, Kim MY, Elliot G et al. Human placental cytotrophoblast epigenome dynamics over gestation and alterations in placental disease. Dev Cell 2021;56:1238-52.e5.

[50] Schroeder DI, Blair JD, Lott P et al. The human placenta methylome. Proc Natl Acad Sci USA 2013;110:6037-42.

[51] Bellazi L, Germond S, Dupont C et al. A sequence variation in the promoter of the placental alkaline phosphatase gene (ALPP) is associated with allele-specific expression in human term placenta. Placenta 2010;31:764-9.

[52] Kong X, Li R, Chen M et al. Endogenous retrovirus HERVH-derived lncRNA UCA1 controls human trophoblast development. Proc Natl Acad Sci USA 2024;121:e2318176121.

[53] Wang M, Zhao K, Liu M et al. BMP4 preserves the developmental potential of mESCs through Ube2s- and Chmp4b-mediated chromosomal stability safeguarding. Protein Cell 2022;13:580-601.

RIGHTS & PERMISSIONS

The Author(s) 2026. Published by Oxford University Press on behalf of Higher Education Press.