Exploring the impacts of senescence on implantation and early embryonic development using totipotent cell-derived blastoids

Graphical abstract

Highlights

• •TBL-blastoids were generated using a novel three-dimensional system with totipotent blastomere-like cells (TBLCs).
• •TBL-blastoids could mimic the development and implantation of natural blastocysts with the similar process.
• •Senescence-related TBL-blastoids reflected the potential gene profiling of implantation failure of women with advanced age.

Abstract

Introduction

Advanced maternal age is associated with reduced implantation and pregnancy rates, yet the underlying mechanisms remain poorly understood, and research models are limited.

Objectives

Here, we aim to elucidate the impacts of senescence on implantation ability by employing blastoids to construct a novel research model.

Methods

We used a novel three-dimensional system with totipotent blastomere-like cells (TBLCs) to construct TBL-blastoids and established senescence-related embryo models derived from oxidative stress-induced TBLCs.

Results

Morphological and transcriptomic analyses revealed that TBL-blastoids exhibited characteristic blastocyst morphology, cell lineages, and a higher consistency in developmental rate. TBL-blastoids demonstrated the ability to develop into postimplantation structures in vitro and successfully implanted into mouse uteri, inducing decidualization and forming embryonic tissues. Importantly, senescence impaired the implantation potential of TBL-blastoids, effectively mimicking the impaired implantation ability and reduced pregnancy rates associated with advanced age. Furthermore, analysis of differentially expressed genes (DEGs) in human homologous deciduae revealed enrichment in multiple fertility-related diseases and other complications of pregnancy. The genes implicated in these diseases and the common DEGs identified in the lineage-like cells of the two types of TBL-blastoids and deciduae may represent potential targets for addressing impaired implantation potential.

Conclusion

These results unveiled that TBL blastoids are an improved model for investigating implantation and early postimplantation, offering valuable insights into pregnancy-related disorders in women with advanced age and potential targets for therapeutic interventions.

Introduction

The delay of women’s childbearing age is prevailing and plays an important role in the low fertility which is a worldwide issue [1], [2]. It has demonstrated that euploid blastocysts derived from women with AMA (women defined as ≥ 35 years old) often display impaired implantation, leading to a decreased pregnancy rate [3], [4]. Implantation is a pivotal process of embryonic survival and development. During this process, the trophectoderm (TE) of the blastocyst attaches the endometrium, followed by decidualization, which is a prerequisite for implantation and pregnancy reproduction in human being. Upon decidualization, the stromal compartment experiences a rapid remodelling process. Stromal cells proliferate and differentiate into a sponge-like cell mass [5], [6]. However, our insights of embryo-uterine interactions and implantation are restricted by the short of a suitable model and ethical limitations, thus hampering advancements in implantation outcomes for women with AMA.

Provided the conservation of gene expression profiles and pathways regulating implantation between humans and mice, the mouse model is plausible to explore the human implantation [7]. In recent years, mouse blastoids resembling blastocysts regarding morphology, cell lineages allocation, and transcriptome have been established [8], [9], [10], [11], [12], [13]. This model offers valuable insights into early embryonic development and implantation. Employing the in vitro culture (IVC) system, these blastoids can be cultured beyond the implantation stages [14] and mimicking certain aspects of the implantation process [13]. After transferring mouse blastoids to pseudopregnant mice,[11], [13] few successfully implanted and induced decidualization. These results indicate that the current blastoids have limited developmental potential in view of low construction efficiency and improper localization of cell lineages [9], [11], [12], [13]. Therefore, the establishment of a blastoid model with enhanced developmental potential is required to accurately mimic early embryonic development and implantation.

A novel type of mouse totipotent cell line, totipotent blastomere-like cells (TBLCs), can contribute to both embryonic and extraembryonic tissues in chimeric blastocysts [15]. TBLCs are generated by repressing the spliceosome and can be maintained long-term in vitro. At the aspect of transcriptome, TBLCs are similar to 2- and 4-cell blastomeres but differ from expanded or extended pluripotent stem cells (EPSCs) and 2-cell-like cells (2CLCs). When preforming chimeric experiments using 8-cell mouse embryos, TBLCs contribute to the inner cell mass (ICM) and TE, subsequently influencing the generation of epiblast (EPI) cells, extraembryonic ectoderm (ExE) cells, and ectoplacental cone (EPC) cells [15]. These features demonstrate that TBLCs have greater potential for simulating early embryonic development and implantation. Blastoids constructed employing TBLCs may possess enhanced developmental potential and improved implantation ability.

Several studies have revealed a maternal chronological age pattern in the transcriptome of blastocysts [16], [17], [18]. During early embryonic development, cellular senescence can significantly influence embryonic quality, viability, and developmental potential. Senescent cells exhibit cell cycle arrest, morphological changes, and dysfunction [19], [20]. Specific increases in reactive oxygen species (ROS) can induce cellular senescence [21], and hydrogen peroxide (H₂O₂) is also commonly employed to induce cellular senescence. Thus the construction of senescent blastoids by H₂O₂ treatment TBLCs may be instrumental in mimicking embryo implantation in women with AMA and elucidating the processes and molecular mechanisms underlying senescence-associated impaired implantation.

In this study, we successfully generated mouse blastoids using TBLCs, referred to as TBL-blastoids, which exhibited comparable morphology, cell lineage allocation, and transcriptome profiles to natural blastocysts. These findings illustrate that TBL-blastoids have greater potential for further development both in vivo and in vitro. Moreover, we derived aging TBL-blastoids from senescent TBLCs generated through oxidative stress and transferred them into pseudopregnant mice to mimic embryonic implantation in women with AMA, thereby exploring age-associated factors contributing to implantation disorders.

Material and methods

Ethics statement

All procedures related to animals were performed following the ethical guidelines of the Peking University Third Hospital (Approval no. M2023355).

Mice experiments

For generation of pseudopregnant surrogates, ICR female mice (8–12 weeks old) in oestrous and vasectomized ICR male mice were mated. Mice were housed in a 12 h light/12 h dark cycle in a temperature-controlled facility with free access to food and water.

Stem cells, culture materials and conditions

All stem cell lines were cultured on a layer of irradiated ICR mouse embryonic fibroblasts (MEFs or feeder cells) at 37℃ under 20 % O₂ and 5 % CO₂. All of the cell lines were routinely tested for mycoplasma contamination and all cell lines were mycoplasma-negative.

N2B27-2iL [22] medium was used to culture mESCs and miPSCs. N2B27-2iL medium consisted of N2B27 basal medium and 10 ng/mL hLIF (Peprotech, 300–05), 3 mM CHIR99021 (Tocris, 4423), and 1 mM PD0325901 (Selleck Chemicals, S1036). N2B27 basal medium was prepared using 1:1 mixture of DMEM/F12 (Thermo Fisher Scientific, 11330–032) and Neurobasal (Thermo Fisher Scientific, 21103–049) with 0.5x N2 (Thermo Fisher Scientific, 17502–048), 0.5x B27 (Thermo Fisher Scientific, 17504–044), 1x nonessential amino acids (Thermo Fisher Scientific, 11140–050), 2 mM L-GlutaMAX (Thermo Fisher Scientific, 35050–061), 0.1 mM 2-mercaptoethanol (Thermo Fisher Scientific, 21985–023), 5 % KnockOut Serum Replacement (Thermo Fisher Scientific, 10828–028). mEPSCs were cultured in N2B27 basal medium with 10 ng/mL LIF (Peprotech, 300–05), 3 mM CHIR99021 (Tocris, 4423), 2 mM (S)-(+)-dimethindene maleate (Tocris, 1425), and 2 mM minocycline hydrochloride (Santa Cruz Biotechnology, sc-203339) (termed as N2B27-LCDM).

Establishment of TBLCs

To prepare TBLCs, mESCs or mouse induced pluripotent stem cells (miPSCs) cultured in N2B27-2iL medium were treated with 1.5 nM PlaB (N2B27-2iLP) (Tocris, 6070). For the first passage, cells were plated on irradiated MEFs in N2B27-2iL medium. 24 h later, the medium was changed to N2B27-2iLP medium. After 5 passages, the induction process was usually completed.

To induce oxidative stress in TBLCs, TBLCs were treated with 1 mM H₂O₂ for 1 h to establish the senescent state followed by the replacement with fresh medium for another 24 h for quantitative real-time qPCR (qRT-PCR), immunofluorescence staining and blastoids generation.

Every two days, the medium was replaced with fresh medium. Cells were passaged as single cells every three to five days using TrypLE Express (Thermo Fisher Scientific, 12604–013).

Cell proliferation detection

Cells were dissociated into single cells using TrypLE Express. To dissociate irradiated MEFs, cell resuspension was transferred into a 0.1–0.2 % gelatine-coated plate and incubated at 37 °C for 40 min. Using the TC-10 counter (Bio-Rad, 1450001), the cell number was counted. Every 24 h, cells were counted until 96 h.

CCK-8 assay for cell proliferation

Cell Counting Kit-8 (CCK-8, Beyotime, C0037) was used to detect the cell proliferation of ctrl TBLCs and H₂O₂-treated TBLCs. Briefly, before every experiment, TBLCs were treated with 1 mM H₂O₂ for 1 h to establish the senescent state, followed by replacing the medium with fresh medium and cultured for another 24 h. Subsequently, ctrl TBLCs and H₂O₂-treated TBLCs were digested into single cells and diluted with N2B27-2iLP to a concentration of 3x10⁴ cells/mL. 100 µL of cell suspensions or cell medium (blank control) was seeded in 96-well plates per well for three biological replicates. The optical density (OD) values at a wavelength of 450 nm for 0 h, 24 h, 48 h, 72 h, and 96 h were tested and documented.

SA-β-gal staining for cellular senescence

Senescence-β-Galactosidase Staining Kit (Beyotime, C0602) was used to determine the cellular senescence of TBLCs. SA-β-gal-positive cells were photographed with a microscope. More than 20 cell colonies in 7 distinct fields were used for that TBLCs were aggregated to form cell colonies. The number of positive cell colonies within a field was counted to determine the percentage of SA-β-gal-positive cells.

Generation of TBL-blastoids

TBLCs were dissociated into single cells. To dissociate irradiated MEFs, cell resuspension was transferred into a 0.1–0.2 % gelatine-coated plate and incubated at 37 °C for 40 min.

The supernatant was collected, filtered through 40 µm cell strainers, and counted. 24-well AggreWell 400 plate (STEMCELL Technologies, 34415) was prepared following the manufacturer’s instructions. TBL-blastoid basal medium is prepared using 25 % TSC basal medium, 25 % N2B27 basal medium (see above), and 50 % KSOM. TBL-blastoid medium contained TBL-blastoid basal medium and 2 mM ROCK inhibitor Y-27632 (Selleck, S1049), 3 mM GSK3 inhibitor CHIR99021 (Tocris, 4423), 5 ng/mL BMP4 (Peprotech, 120-05ET), and 0.5 mM A83-01 (Axon Medchem, 1421). About 6,000 cells (five cells per microwell in 24-well AggreWell 400 plate) were resuspended in TBL-blastoid medium with and seeded into one well of the 24-well AggreWell 400 plate (Stem Cell Technologies, 07010). The day of cell seeding was Day 0 of the induction process. 24 h later (Day 1), the medium was removed and replaced with fresh TBL-blastoid medium without Y-27632 and supplemented with 200 nM LPA (Sigma-Aldrich, L7260). On Day 4, the medium was replaced with fresh TBL-blastoid medium without Y-27632 and LPA. From Day 5 but no more than Day 6, blastoids were manually picked up using a mouth pipette under a stereomicroscope for analysis and downstream experiments.

Tbl-blastoid transplantation in vivo

Using a mouth pipette, TBL-blastoids were manually picked up under a stereomicroscope and maintained in M2 droplets. The surrogate at 2.5 or 3.5 days post-coitum (dpc) was anaesthetized with pentobarbital sodium, and 20–25 TBL-blastoids were transferred into each uterine horn. At 7.5 dpc or 8.5 dpc, and the uterus was taken out and deciduae were dissected out, and then embryo-like structures were further dissected.

In vitro culture of TBL-blastoids beyond implantation

TBL-blastoids were transferred using a mouth pipette onto an 8-well µ-Slide (ibidi, 80826) and cultured in IVC1 medium. 20–25 TBL-blastoids were plated in one well. Within 24–48 h, TBL-blastoids attached to the plate. Then, the well was washed with PBS, and IVC2 medium was replaced. After two to four days, postimplantation embryo-like structures formed.

ELISA assay for detecting β-CG

According to the manufacturer’s instructions, ELISA for detecting β-CG were performed (Meimain, MM-0581 M2). Briefly, the supernatant of the culture medium was collected. Proteins of the standard and samples were incubated with monoclonal antibodies specific for β-CG precoated onto microplates for 30 min at room temperature, followed by five washes, HRP-conjugated antibodies against antigens were incubated for 30 min at room temperature. After another five washes, a substrate solution was added to the wells. Then, the stop solution was added, and the intensity of colour was measured as the optical density (OD) value at a wavelength of 450 nm indicated the intensity of colour.

Immunofluorescence staining

The samples were fixed with 4 % PFA for 15 min at room temperature and permeabilized with 0.2 % Triton X-100 for 15 min. Samples were then blocked with blocking buffer for 1 h at room temperature. Primary antibodies were added to samples and incubated at 4 °C overnight. After five washes with PBS containing 0.1 % Tween 20, fluorescence-conjugated secondary antibodies were added and incubated for 2 h at room temperature followed by three times with PBS containing 0.1 % Tween 20. Nuclei were counterstained with DAPI. Image acquisition was performed using a Zeiss LSM 710 and 880 confocal microscopes. Images were processed using Fiji (ImageJ, v2.0.0) or Zen (Zeiss).

The primary antibodies and dilutions used were rabbit anti-OCT4 (1:200; Abcam, ab19857), mouse anti-OCT4 (1:100; Santa Cruz Biotechnology, sc-5279), mouse anti-OCT4 (1:200; Cell Signaling Technology, 75463), mouse anti-SOX2 (1:200; Abcam, ab171380), mouse anti-CDX2 (1:100; Biogenex, MU392A), rabbit anti-CDX2 (1:200; Cell Signaling Technology, 3977), rabbit anti-GATA6 (1:100; Cell Signaling Technology, 5851), rabbit anti-GATA4 (1:200; Cell Signaling Technology, 36966), goat anti-GATA6 (1:200; R and D Systems, AF1700), rabbit anti-ZSCAN4 (1:1000, Millipore, AB4340), rabbit anti-TFAP2C (1:200; Abcam, ab203691), and rabbit anti-GATA2/3 (1:200; Abcam, ab182747).

The secondary antibodies used were Alexa Fluor 488-AffiniPure donkey anti-mouse IgG (H + L) (Jackson ImmunoResearch Labs, 715-545-151), Alexa Fluor 488-AffiniPure donkey anti-rabbit IgG (H + L) (Jackson ImmunoResearch Labs, 711-545-152), Alexa Fluor 555 donkey anti-mouse IgG (H + L) (Thermo Fisher Scientific, A-31570), Alexa Fluor 647-AffiniPure donkey anti-mouse IgG (H + L) (Jackson ImmunoResearch Labs, 715-605-151), and Alexa Fluor 647-AffiniPure donkey anti-rabbit IgG (H + L) (Jackson ImmunoResearch Labs, 711-605-152).

The cell fluorescence intensity was calculated by subtracting the mean grey value of the cell colony from the mean grey value of the blank area.

Total RNA extraction and quantitative real-time PCR

According to the manufacturer’s instructions, total RNA was extracted from cells using the RaPure Total RNA Mini Kit (Magen R4011-02). For mRNA qRT-PCR analysis, 1 mg RNA was converted into cDNA using HiScript II Q RT SuperMix (Vazyme, R223) for qRT-PCR. Then, the cDNA was quantified in duplicate or triplicate using 2x Realtime PCR Mix (Mei5 Biotech, MF013) and an Applied Biosystems StepOnePlus Real-Time PCR System (Thermo Fisher). Gene expression was normalized to that of β-actin. All primers used for qRT-PCR are listed in Table S1.

Generation of single-cell RNA-sequencing library

TBL-blastoids were manually picked up and about 500 TBL-blastoids were harvested and dissociated with TrypLE Express at 37 °C for 30 min with agitation. Dissociated cells were resuspended in DPBS containing 0.05 % BSA. Cell density was counted by a TC10 cell counter (Bio-Rad, 1450001). Chromium Single Cell B Chip (10X Genomics, PN-120262) and Chromium single cell controller (10X Genomics) were used to generate single-cell gel beads in the emulsion. Using the Chromium Single Cell 30 Reagent Kit v3 (10X Genomics, PN-1000092) and Chromium i7 Multiplex Kit (10X Genomics, PN-120262), the library was generated. The libraries were pooled and sequenced using NextSeq 500 (150 cycles, high output).

Analysis of single-cell RNA-Seq data

STAR v2.5.1b1 [23] and CellRanger v3.0.2 (10X Genomics) software were used for TBL-blastoids (5266 cells), H₂O₂-treated TBL-blastoids (4824 cells). A feature-barcode matrix was generated with the default setting for demultiplexing. scRNA-Seq data of E3.5 early blastocyst and EPS-blastoids were obtained from published datasets (GEO: GSE135701) [24]. The R package Seurat v4.3.0 [25] was used in R (version 4.1.0) to read and analyse the feature-barcode matrix. First, we filtered the cells in the TBL-blastoids, EPS-blastoids and H₂O₂-treated TBL-blastoids which had unique feature counts over 5000 according to quality control matrix plots (3408, 1832 and 2269 cells in the TBL-blastoids EPS-blastoids and H₂O₂-treated TBL-blastoids groups passed the filter, respectively). The cells in the blastocysts that had unique feature counts over 2500 were filtered (184 cells in the blastocysts passed the filter). The merge function was used to merge blastocysts, EPS-blastoids and TBL-blastoids, and then FindIntegrationAnchors and IntegrateData were used to integrate. The NormalizeData function of the ‘LogNormalize’ method was used to calculate the gene expression using the default settings. Nonlinear dimension reduction was performed by Seurat’s RunUMAP function and clustering with a resolution setting of 0.2. Featureplot function was used to visualize the lineage markers. Markers of the clusters were found by the FindMarkers function with a bimodal likelihood ratio test. The R package ComplexHeatmap v2.8.0[26] and pheatmap v1.0.12 were used to generate heatmap of marker genes. The differentially expressed marker genes were selected with a minimum threshold of 1 in absolute log_2-fold change and p.adjust (padj) of 0.05. Using DAVID [27], gene Ontology (GO) analysis was performed.

Bulk RNA-seq library generation and data analysis

Decidua-like tissues were collected for total RNA isolation using the TRIzol (Thermo Fisher Scientific, 15596026) method. PolyA(+) RNA was isolated for library construction with the VAHTS mRNA-seq V3 library Prep Kit for Illumina Smart-Seq2. Using paired-end sequencing on the Illumina HiSeq X Ten platform, samples were pooled and sequenced. Using the HISAT2 alignment program, sequencing reads were filtered and mapped to the mouse genome build mm10[28]. De novo transcriptome assembly and transcript and gene abundance calculations were performed using the StringTie assembler [29]. Differentially expressed genes (DEGs) were calculated using the R package DESeq2 v1.32.0[30]. With a minimum threshold of 1 in absolute log_2-fold change and p.adjust (padj) of 0.05, DEGs were selected. Using DAVID [27], GO analysis was performed. For homologous gene conversion, mouse genes were converted to human homologous genes with R package biomaRt [31], [32]. Disease-gene association analysis was performed using DOSE v3.18.3[33].

Quantification and statistical analysis

In the figure legends, quantification details on the number of biological replicates (n value) and data presentation are included. Values are shown as the mean, and error bars represent SD. Significant differences were considered when the P (or adjusted P) values were smaller than 0.05. Using GraphPad Prism 9 software or R package ggplot2[34] and enrichplot or other R packages, graphs were generated.

Data and code availability

Single-cell RNA-seq data have been deposited in the Gene Expression Omnibus (GEO) under accession numbers GSE GSE232718 (scRNA-seq data and bulk RNA-seq website: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE232718), and the single-cell RNA-seq of mouse blastocysts are under GEO accession GSE135701 (ref. [24]).

Results

Construction and characterization of blastoids from TBLCs

In this study, we induced mouse pluripotent stem cells (mPSCs) into TBLCs using a previously published method [15] (Fig. 1A, Fig. S1A). The obtained TBLCs exhibited higher expression levels of totipotent genes and reduced expression of pluripotent genes compared to mPSCs or mouse extended or expanded pluripotent stem cells (mEPSCs), as quantified using qRT-PCR and immunofluorescence staining (Fig. 1B-C and S1B). The cell proliferation of TBLCs showed no significant difference compared to that of mPSCs or mEPSCs in the first 72 h, but TBLCs exhibited slower growth than the other cell types at 96 h (Fig. S1C), consistent with previous findings indicating the limited proliferation ability of totipotent cells. These results indicate that the induced TBLCs possess characteristics resembling totipotent cells.

Fig. 1.

Generation of blastoids employing TBLCs. A. Phase-contrast image of TBLCs induced from mPSCs by the splicing inhibitor PlaB. Scale bar, 100 µm. B. qRT-PCR analysis of the expression levels of totipotent genes in mPSCs and TBLCs. Mean ± SD, n = 3 technical replicates of cells. ****p ≤ 0.0001; ns, no significance. C. Immunofluorescence staining of TBLCs for the pluripotent markers OCT4 and SOX2 and the totipotent marker ZSCAN4. Scale bar, 50 μm. D. Schematic of generating TBL-blastoids with TBLCs using the improved system (above). Phase-contrast images of TBL-blastoids on Days 4, 5 and 6. Scale bar, 100 µm. E. Representative images showing TBL-blastoids. Scale bar, 100 µm. F. Immunofluorescence staining of TBL-blastoids on Day 5 for the PrE marker GATA6, EPI marker OCT4, and TE marker CDX2. Scale bar, 20 μm. MAX, maximum intensity projection. G. UMAP plot showing that TBL-blastoids and blastocysts largely overlapped after integrative analysis using the Seurat R package and that TBL-blastoids contained three lineages (right). The cell percentages of three lineages in blastocysts or TBL-blastoids of scRNA-seq. H. Heatmap of cluster marker gene expression of TBL-blastoids on Day 5.

The Hippo pathway plays a crucial role in the specification of the first cell lineage during embryonic development [35], [36], [37]. Inactive Hippo pathway signalling in the outer cells of the embryo results in the translocation of YAP to the nucleus, where it binds to the transcription factor TEAD4 and activates the expression of trophectoderm (TE) lineage genes [35], [36], [37]. To construct blastoids using TBLCs, we optimized a three-dimensional (3D) induction system [13] by adding lysophosphatidic acid (LPA), an inhibitor of the Hippo pathway, from Days 1 to 4 (Fig. 1D). By Day 5, a typical blastocyst-like structure was observed, and we named these structures TBL-blastoids (Fig. 1E and S1D). Throughout the construction process, we collected cell aggregates and quantified lineage gene expression using immunofluorescence staining. During the first 2 days, the majority of cells in the aggregates uniformly expressed OCT4 (an epiblast (EPI) lineage marker), with minimal expression of CDX2 (a TE lineage marker) (Fig. S1E). From Days 3 to 5, a cavity emerged, and OCT4-positive cells were confined to the inner position of the cell aggregates, while the expression of CDX2 increased in the outer layer of the aggregates (Fig. S1E), consistent with the morphological changes during mouse early embryonic development. Quantification of Zscan4 expression using qRT-PCR and immunofluorescence staining revealed a dynamic pattern during TBL-blastoid construction: the expression level sharply decreased initially and then gradually increased. Ultimately, the formed blastoids exhibited high-level expression similar to that of TBLCs (Fig. S2A and S2B). These results aligned with previous findings demonstrating that Zscan4 is exclusively expressed in 2-cell embryos and embryonic stem cells (ESCs)[38]. We also observed that OCT4 was consistently expressed from Day 1 to 5 but gradually restricted to the inside of TBL-blastoids starting from Day 3 or 4 (Fig. S2B). These findings demonstrated the transition between totipotency and pluripotency during TBL-blastoid construction. Subsequently, we characterized the cell lineages within TBL-blastoids using immunofluorescence staining. The expression of CDX2 in the outer layer, along with the presence of OCT4-positive cells inside the blastoids and their proximity to CDX2-positive and GATA4/GATA6 (two primitive endoderm (PrE) markers)-positive cells, clearly indicated the presence of a cavity in the TBL-blastoids (Fig. 1F, S2C and S2D). These findings suggest that TBL-blastoids consist of three distinct lineage cells, resembling the cellular composition of natural blastocysts.

Fig. 2.

TBL-blastoids served as a good model to study implantation and early embryonic development. A. Immunofluorescence staining and bright-field images of EPS-blastoids (above) or TBL-blastoids (right) for the EPI marker OCT4 and the TE marker CK8. Scale bars, 20 μm. B. Representative images showing EPS-blastoids and TBL-blastoids. Scale bar, 100 µm. C. UMAP plot of EPS-blastoids and TBL-blastoids after integrative analysis using the Seurat R package, and the integrative cells were divided into four clusters, namely, ICM/EPI, TE, PrE (right) and intermediate (IM) clusters, which coexpressed markers of multiple cell lineages (left). The identities of each cluster were determined according to the expression of the lineage markers. D. The cell percent of four cell clusters in EPS-blastoids or TBL-blastoids determined by scRNA-seq, shown by stacked bar charts. E. Heatmap of IM marker gene expression in EPS-blastoids and TBL-blastoids. F. Volcano map showing the upregulated (red) and downregulated (blue) IM marker genes in the IM cluster. G. Gene Ontology analysis of biological functions for IM marker genes. The size of the bubble indicates the gene count of every term, the colour of the bubble indicates the p value, and the location indicates the fold enrichment. H. Representative images of attached TBL-blastoid-derived postimplantation embryo-like structures after delayed culture in the IVC system. Scale bar, 50 µm. I. Immunofluorescence staining of TBL-blastoid-derived postimplantation embryo-like structures for the EPI marker OCT4, VE marker GATA4 and ExE marker AP-2γ. Scale bar, 50 µm. J. Representative image showing the formation of decidua in the mouse uterus 2 days after TBL-blastoid transfer at 2.5 dpc (top right) and the dissected decidual-like structures (top left and middle). The image below shows the embryo-like tissues. The red arrowhead indicates the ovaries, and the black arrowhead indicates the decidua. The star indicates a red spot within the decidua. The black arrow indicates embryonic-like tissue. Scale bars, 1 mm and 250 µm (lower right). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Next, we performed single-cell RNA-seq (scRNA-seq) analysis of Day 5 TBL-blastoids. Integrative analysis of the transcriptome revealed significant overlap between cells of TBL-blastoids and blastocysts (Fig. 1G). Compared to blastocysts, TBL-blastoids contained a higher proportion of PrE and TE cells and a lower proportion of ICM/EPI cells (Fig. 1G). UMAP analysis identified four distinct clusters through both integrative analysis and separate analysis of TBL-blastoids (Fig. 1G, S2E and S2F). By evaluating the expression levels of lineage-specific genes, we defined one cluster as ICM/EPI, one as TE, and two as PrE (Fig. 1G, S2E and S2F). Moreover, the analysis of TBL-blastoids demonstrated that each cluster exhibited high expression of lineage-specific genes, indicating the clear separation of cell lineages within TBL-blastoids (Fig. 1H and S2G). Collectively, our findings demonstrate the successful construction of blastoids using TBLCs, and the resulting TBL-blastoids exhibit a transcriptome profile resembling that of natural blastocysts.

Tbl-blastoids are a better model to simulate embryo implantation in vitro

To compare the differences between TBL-blastoids and EPS-blastoids [13], we employed the previously established method to generate EPS-blastoids from mEPSCs [13]. Both TBL- and EPS-blastoids exhibited typical blastocyst morphology (Fig. 2A). Immunofluorescence staining of lineage-specific gene expression showed that OCT4-positive cells were located inside both types of blastoids, while CK8-positive TE lineage cells were located outside (Fig. 2B). Notably, the cavities of EPS-blastoids appeared more susceptible to collapse during physical manipulation (Fig. 2B).

To further characterize the differential gene expression patterns between TBL- and EPS-blastoids, we integrated scRNA-seq data from TBL-blastoids with published scRNA-seq data of EPS-blastoids. Based on the expression patterns of lineage-specific markers, ICM/EPI-, TE- and PrE-like cells were observed in both TBL- and EPS-blastoids (Fig. 2C). These clustered cells exhibited correct expression of lineage-specific genes (Fig. S3A). Furthermore, the expression levels of these lineage markers and the number of positive cells were significantly higher in TBL-blastoids than in EPS-blastoids (Fig. S3A). We also identified cells with coexpression of ICM/EPI and TE lineage markers, termed intermediate cells (IM cells), in both types of blastoids (Fig. 2C). Calculation of the proportion of clustered cells revealed a similar proportion of ICM/EPI-like cells between the two types of blastoids (Fig. 2D). However, TBL-blastoids contained a higher number of TE- or PrE-like cells than EPS-blastoids (Fig. 2D). Importantly, IM cells were rarely identified in TBL-blastoids, while over 40 % of cells in EPS-blastoids exhibited an intermediate state (Fig. 2D). This finding suggests that most cells differentiated correctly into lineage cells during the construction of blastoids using TBLCs. Integrative analysis with scRNA-seq data from blastocysts demonstrated that TBL-blastoids were more similar to natural blastocysts at the transcriptome level than EPS-blastoids (Fig. S3B and S3C). The expression of lineage markers further indicated the association of cell identity with cell lineage (Fig. S3D).

Fig. 3.

Oxidative stress induced cellular senescence of TBLCs and impaired the generation and developmental potential of TBL-blastoids A. Representative images showing the SA-β-gal activity of TBLCs (top left) or H₂O₂-treated TBLCs (lower left) and the percentage of SA-β-Gal-positive cells (right). The arrowhead indicates SA-β-gal-positive cells. Scale bars, 50 μm. B. qRT-PCR analysis of biomarkers of cellular senescence and totipotent genes in TBLCs or H₂O₂-treated TBLCs. Mean ± SD, n = 3 technical replicates. ****p ≤ 0.0001, ***p ≤ 0.001, **p ≤ 0.01, ns, no significance. C. Representative images of ctrl TBL-blastoids (left) or H₂O₂-treated TBL-blastoids (right) and quantification of the percentage of TBL-blastoid formation efficiency using ctrl TBLCs or H₂O₂-treated TBLCs. Scale bars, 50 µm. n = 3 biological replicates. D. Immunofluorescence staining of ctrl TBL-blastoids (above) or H₂O₂-treated TBL-blastoids (below) for the ICM/EPI marker OCT4, PrE marker GATA4 and TE marker CDX2. Scale bars, 50 μm. E. UMAP plot of ctrl TBL-blastoids and H₂O₂-treated TBL-blastoids after integrative analysis using the Seurat R package. F. Expression levels of markers of each blastocyst lineage in ctrl TBL-blastoids or H₂O₂-treated TBL-blastoids. G. Representative images showing structures of ctrl TBL-blastoids (above) or H₂O₂-treated TBL-blastoids (below) after 4 days of IVC culture. H. Quantification of the attachment ratio of ctrl TBL-blastoids or H₂O₂-treated TBL-blastoids. N = 6 biological replicates. I. Quantification of CG-β in the medium after TBL-blastoids attached to the plates on Day 2 or Day 4, and the value of the culture medium was set to zero. N = 3 biological replicates. J. Immunofluorescence staining of ctrl TBL-blastoid (above)- or H₂O₂-treated TBL-blastoid (below)-derived postimplantation embryo-like structures for the EPI marker OCT4, the VE marker GATA4 and the ExE marker AP-2γ. Scale bars, 50 µm.

We subsequently analysed the characteristics of IM cells by selecting IM-specific genes using the integrated data. We found that TE-like cells also highly expressed several IM-specific genes (Fig. 2E). Additionally, analysis of cell lineage markers showed that ICM/EPI and PrE markers were downregulated, while TE markers showed no significance in the IM cells (Fig. 2F). Interestingly, several imprinted genes, including H19, Rian, Mest and Peg3, were specifically expressed in IM cells. Maternally imprinted genes (H19 and Rian) were downregulated in IM cells, while paternally imprinted genes (Mest and Peg3) exhibited high expression (Fig. 2F)[39]. This finding suggests that the imbalanced expression of imprinted genes may contribute to the abundance of IM cells. Gene Ontology (GO) analysis of these IM-specific genes revealed paradoxical biological processes (Fig. 2G), indicating disorganized gene expression patterns in IM cells.

To mimic embryo implantation, we transferred TBL-blastoids into an in vitro culture (IVC) system [14], [40]. Most TBL-blastoids attached and underwent outgrowth, particularly in the TE-like region (Fig. 2H). From Days 6 to 8, structures resembling postimplantation embryos emerged and enlarged over time (Fig. 2H). Immunofluorescence staining results showed that OCT4-positive cells and AP-2γ (ExE lineage marker)-positive cells were located in the inner part, surrounded by GATA4 (visceral endoderm marker)-positive cells, resembling the in vitro structures of blastocysts (Fig. 2I). When TBL-blastoids were transferred to the uterus of pseudopregnant mice at 2.5 days post-coitum (dpc), TBL-blastoids successfully implanted and induced decidualization, and the implanted TBL-blastoids exhibited red spots, indicating increased vascular permeability (Fig. 2J). Dissecting the embryo-like tissue from the large deciduae of TBL-blastoids revealed degenerated-embryo-like structures (Fig. 2J), indicating that TBL-blastoids possessed the ability to initiate implantation in vivo.

Overall, these results demonstrate that TBL-blastoids can develop into structures resembling postimplantation embryos and serve as a superior model for studying implantation and early postimplantation processes. This finding is supported by their morphological and transcriptomic similarity to blastocysts, as well as their performance in in vitro culture and in vivo implantation experiments.

Oxidative Stress-Induced cellular senescence of TBLCs and impaired generation and postimplantation developmental potential of TBL-Blastoids

To investigate the mechanisms underlying decreased oocyte quality and impaired embryo implantation rates in women with AMA, we attempted to establish senescence-related embryo models using TBL-blastoids. Previous studies have shown that oxidative stress can induce cellular senescence, known as stress-induced premature senescence (SIPS), with hydrogen peroxide (H₂O₂) being a commonly used inducer of SIPS [41]. Therefore, we treated TBLCs with H₂O₂ (referred to as H₂O₂-treated TBLCs) and examined the characteristic features of cellular senescence·H₂O₂-treated TBLCs exhibited smaller cell colonies and significantly reduced cell viability, characterized by less compacted colonies and enlarged cell bodies, than untreated TBLCs (Fig. S4A). Following H₂O₂ treatment, senescence-associated β-galactosidase (SA-β-gal) activity was activated (Fig. 3A and S4B). CCK-8 assays demonstrated that the growth rate of the H₂O₂-treated TBLCs peaked at 24 h post-treatment and then slightly decreased (Fig. S4C), coinciding with the presence of dead cells observed under microscopy, indicating that cellular senescence affects cell proliferation in TBLCs. Subsequently, we assessed the impact of senescence on the totipotent features of TBLCs. The expression of senescence-associated secretory phenotype (SASP)-related genes, including p16, p21 and TNF-α, was significantly upregulated in the H₂O₂-treated TBLCs, whereas the expression of totipotent genes was downregulated (Fig. 3B). Similar results were obtained through quantification of SASP-related and totipotent gene expression using immunofluorescence staining (Fig. S4D), indicating that H₂O₂treatment impairs the totipotency of TBLCs.

Fig. 4.

Impacts of senescence on the in vivo developmental potential of TBL-blastoids and identification of potential genes. A. Representative image showing the formation of decidua in the mouse uterus 2 days after ctrl TBL-blastoid (top right) or H₂O₂-treated blastoid (bottom right) transfer at 2.5 dpc and the dissected decidual-like structures (top left and bottom right). Scale bars, 2 mm. B. Heatmap of decidual-associated gene expression in ctrl decidual-like structures or H₂O₂-treated decidual-like structures. Genes in the red block are decidual markers in mice, and genes in the yellow block are changed genes in natural pregnancy decidualization. Genes in the orange block are upregulated by molecular signals from the conceptus, while those in the blue block are downregulated by molecular signals from the conceptus. C. Treeplot showing the association of homologous DEGs in humans with human diseases and hierarchical clustering. D. Enrichment plot representation of the GSEA results obtained from homologous DEGs. Each disease is presented by a node with a different size, proportional to the number of genes, and the connecting line represents gene overlap between diseases. E. Heatmap of homologous DEG expression differences in every disease. Genes involved in multiple diseases are marked in red. F. Comparison of overlapping DEGs between ctrl decidual-like structures and H₂O₂ decidual-like structures among DEGs of EPI, PrE or TE between ctrl TBL-blastoids and H₂O₂-treated TBL-blastoids. G. Gene Ontology analysis of enriched biological functions for common DEGs. The size of the bubble indicates the gene count of every term, the colour of the bubble indicates the p value, and the location indicates the fold enrichment. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

We then generated TBL-blastoids using H₂O₂-treated TBLCs (referred to as H₂O₂-treated TBL-blastoids) and compared them to normal TBL-blastoids (referred to as ctrl TBL-blastoids). We observed that H₂O₂ treatment significantly impaired the efficiency of TBLCs in generating blastoids (Fig. 3C). Through immunofluorescence staining of lineage-specific genes, we found that fewer H₂O₂-treated TBL-blastoids developed a cavity, along with incorrect cell type allocation, compared to ctrl TBL-blastoids (Fig. 3D). This finding suggests that senescent TBLCs struggle to form blastocyst-like structures. To further assess the effect of H₂O₂ during TBL-blastoid construction, we performed scRNA-seq. By integrating clustering and cell lineage gene expression, we identified three major cell clusters corresponding to blastocyst cell lineages (EPI, PrE, and TE) (Fig. 3E). IM cells, characterized by mixed expression of lineage genes, were also identified in H₂O₂-treated TBL-blastoids, in contrast to ctrl TBL-blastoids (Fig. 3E). Moreover, H₂O₂ treatment resulted in the downregulation of lineage gene expression in the three defined clusters (Fig. 3F) and a significant increase in the proportion of IM cells (Fig. S4E). These findings indicate that H₂O₂ treatment disrupts TBL-blastoid construction. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis of differentially expressed genes (DEGs) between H₂O₂-treated and ctrl TBL-blastoids revealed enrichment of senescence and apoptosis-related pathways in H₂O₂-treated TBL-blastoids, along with downregulation of pluripotency and oxidative phosphorylation-related pathways (Fig. S4F), suggesting that senescence induced by H₂O₂ treatment impairs the totipotency of TBLCs, subsequently compromising blastoid construction.

To validate the impacts of senescence on the in vitro developmental potential, we transferred H₂O₂-treated TBL-blastoids to an IVC system. Compared to ctrl TBL-blastoids, H₂O₂-treated TBL-blastoids showed lower attachment, exhibited lower growth rates, and showed vulnerability of derived EPI-like cells (Fig. 3G and 3H). Quantification of the levels of CG-β (a protein supporting the implantation) secreted by TE showed a reduction in H₂O₂-treated TBL-blastoids, whereas it was drastically increased in ctrl TBL-blastoids (Fig. 3I). The expression of lineage markers in H₂O₂-treated TBL-blastoids was downregulated and mislocalized, as detected by immunofluorescence staining (Fig. 3J). Collectively, these results demonstrate that senescence profoundly impairs the developmental potential of TBL-blastoids in vitro.

The impacts of senescence on the in vivo developmental potential of TBL-Blastoids and the identification of candidate related genes

To assess the effects of cellular senescence on in vivo developmental potential, we transplanted H₂O₂-treated TBL-blastoids into the uterus of pseudopregnant mice. We observed that a significantly lower number of H₂O₂-treated TBL-blastoids formed decidual tissues (referred to as H₂O₂ deciduae) with low efficiency than ctrl TBL-blastoids (Fig. 4A), consistent with the findings from the in vitro culture experiments mentioned earlier. We next performed RNA-seq on the obtained H₂O₂ deciduae. Two well-known decidualization markers, Prl8a2 and Alpl [42], [43], were both downregulated in H₂O₂-treated deciduae compared to ctrl deciduae (Fig. 4B). Additionally, the expression levels of the authentic decidualization-associated genes, except Dio2 and Serpinf1, were downregulated in H₂O₂-treated deciduae (Fig. 4B). Furthermore, we compared the expression of a set of genes influenced by molecular signals from the conceptus between the H₂O₂-treated and ctrl deciduae [44]. We found that these two types of deciduae exhibited opposing expression patterns of these genes, including Cpan6, Aldh3a2, Ear2, Ltf, and Inhbb (Fig. 4B). This finding suggests that senescence influences the biology of decidualization at the molecular level. We identified 6654 DEGs in H₂O₂-treated deciduae compared to ctrl samples (Fig. S5A). Functional annotation analysis revealed that the upregulated DEGs were enriched in multiple classes, including apoptosis, negative regulation of cell proliferation, neural and myoblast specification, development and differentiation of germ cells and response to sex hormones (Fig. S5B). However, the downregulated DEGs were enriched in cell proliferation, organ development and embryonic development (Fig. S5B). These results indicate that cellular senescence may restrict embryonic development and induce miscarriage.

To gain insights into how senescence influences implantation and early development, we explored the association of DEGs with diseases using the DOSE toolkit [33]. We also analysed the mouse-to-human homologous DEGs to obtain further information on disease ontology (DO) and the influence of senescence on embryonic development and implantation in women with AMA [31], [32]. We found that the top diseases associated with homologous DEGs included infertility, ovarian failure, menopausal symptoms and other fertility-related diseases, consistent with the common disorders of women with AMA (Fig. 4C and S5C–E). Additionally, the homologous DEGs were enriched in adrenal dysfunction, endocrine abnormalities, premature coronary artery atherosclerosis and chronic active hepatitis (Fig. 4C and S5C). We focused on fertility-related diseases for further analysis, and an enrichment map was used to interpret the general functional themes of the DEGs. We identified a tight network of declined reproductive functions, including primary physiologic amenorrhea, primary hypogonadism and decreased fertility in females (Fig. 4D). These results indicate that cellular senescence may have major impacts on fertility-related diseases, which might be associated with complications of pregnancy or factors contributing to decreased implantation. To identify potential genes involved in cellular senescence-related impacts on fertility and embryonic development, we analysed the expression of genes mediating multiple fertility-related diseases, including CYP17A1, CYP19A1, FSHR, NR5A1, AMH, BMP15, and LEP (Fig. 4E, S5A and B). These genes may represent potential candidates contributing to senescence-impacted fertility-related diseases in humans and play a role in impaired implantation.

Since implantation is closely related to the quality of TBL-blastoids, we analysed the common DEGs of H₂O₂-treated TBL-blastoids compared to ctrl TBL-blastoids and their derived deciduae (Fig. 4F). The most common DEGs were expressed in TE-like cells (Fig. 4F), suggesting that ageing primarily influences TE and, consequently, placental functions. Among these DEGs, three cell lineage markers in blastocysts and imprinted genes were also detected (Fig. S6C)[39]. Maternally expressed genes were highly expressed in normal deciduae, whereas paternally expressed genes were upregulated in H₂O₂-treated deciduae (Fig. S6C). Many common DEGs were highly expressed in TE-like cells of TBL-blastoids (Fig. S6D), indicating that downregulation of these genes may be involved in decidual differences and impaired implantation and development in vivo. Functional annotation of the common DEGs revealed their enrichment in placental development, metabolism, and haematopoietic processes (Fig. 4G). We hypothesized that senescence might influence placental function and thus maternal adaptations to pregnancy [45]. and diseases associated with tumour and immune system dysfunctions, which are also related to placental invasion and supporting pregnancy (Fig. S6E). Overall, senescence of TBL-blastoids primarily affects TE-like cells, followed by placental invasion and maternal adaptations to pregnancy.

Discussion

Constructing blastoids using totipotent cells is a promising approach for generating embryo models. Compared to other blastoids derived from PSCs, the production efficiency, cavitation ratio and implantation rate of TBL-blastoids has shown significant improvement. Moreover, TBL-blastoids exhibit a higher degree of molecular similarity to natural blastocysts. In summary, TBL-blastoids provide a superior model for studying early embryonic development and are readily to manufacture due to their consistent development rate and morphology. Meanwhile, the successful generation of blastoids also highlights the totipotency of TBLCs. Unlike pluripotency, there is no universally accepted standard for defining totipotency. Therefore, utilizing totipotent cells to construct blastoids offers a valuable strategy to characterize totipotent-like cells and investigate genome imprinting in blastoids [46], [47], [48], [49]. This approach allows for a better understanding of the unique properties and developmental potential of totipotent cells. However, the induced population of TBLCs is heterogeneous, which may contribute to the incomplete representation of the expression characteristics of totipotent and pluripotent genes. Previous studies have shown that only a subset of induced TBLCs exhibits high expression of Zscan4c [15], [50]. Isolating these cells with enhanced totipotent features for blastoid construction and induction of aging-related embryonic models could potentially yield results closer to natural blastocysts.

In many successfully established blastoid induction systems, the use of Hippo pathway inhibitor as inducers of trophoblast-like cells (TBLCs) has been demonstrated, including in human blastoid induction systems [51], [52]. This indicates the importance and conservation of the Hippo pathway in the development of human and mouse embryos. We also successfully differentiated TBLCs to trophoblast-like cells using Hippo pathway inhibitor in the blastoid induction. Additionally, we treated 2D-cultured TBLCs with LPA and found that it can transform TBLCs into an epithelial-like cell and induce the expression of trophoblast markers. However, we have not yet conducted long-term culture of the induced cells or performed in-depth molecular-level analysis.

In the future research, the investigation of the similarities and differences between LPA-induced trophoblast-like cells and in vivo trophoblast cells can be conducted to improve the construction of blastoids and unravel the mechanisms of TBL-blastoid formation. By comparing the transcriptomes, epigenetics, and functional characteristics of these cells, we can gain a better understanding of their relationship with in vivo trophoblast cells and provide more accurate models for blastocyst research.

We have continuously detected the expression levels of the Zscan4 during the formation of TBL-blastoids. Immunofluorescence and qRT-PCR results suggested a trend of decreased expression followed by an increase in Zscan4 expression. Previous studies have shown that Zscan4 is only expressed in late 2-cell stage embryos, ESCs and iPSCs [38,53]. We speculate that during the early stages of blastoids formation, TBLCs are capable of recapitulating the expression features of Zscan4. Subsequently, some cells may regain the characteristics of mPSCs and re-express Zscan4. Another hypothesis is that the heterogeneity of induced TBLCs cells and the limited proportion of highly totipotent cells. In the induction process, residual low totipotent cells readily enter into the pluripotent state.

TBL-blastoids derived from totipotent cells exhibit morphological and developmental characteristics comparable to natural blastocysts. Moreover, the integrative analysis of transcriptomic data from natural blastocysts and TBL-blastoids reveals that TBL-blastoids consist of three cell types that correspond to the cell lineages in natural blastocysts, exhibiting comparable transcriptomic characteristics. One study has summarized the proportions of different lineage cells in natural blastocysts and various types of blastoids [54]. However, the analysis of cell numbers and ratios indicates that TBL-blastoids have an excessively high proportion of PrE-like cells, while the proportions of ICM/EPI and TE-like cells are relatively low. This suggests the need for optimization of the concentrations and treatment duration of cytokines and small molecules, as well as an increase in the proportion of highly totipotent cells within the TBLCs cell population. However, unlike other blastoids, TBL-blastoids do not contain undefined cell types according to the transcriptomic analysis. The high efficiency of construction and transcriptional similarity observed in TBL-blastoids suggest that they are a superior model for embryonic development.

Although we have demonstrated the developmental potential of TBL-blastoids to initiate post-implantation development both in vitro and in vivo, we have not observed further development of TBL-blastoids when extended the retention time in the uterine. In fact, they have even exhibited phenomena similar to miscarriage. Currently, there is no research that has successfully developed blastoids to term, which may be a major obstacle in blastoid research. However, optimizing protocols to extend the simulation of embryonic development in TBL-blastoids would have significant implications. Additionally, coculturing TBL-blastoids with self-renewing endometrial epithelial organoids (EEOs) would help to provide more details about the mechanisms underlying implantation.

To obtain TBLCs with evident aging features, we treated TBLCs with H₂O₂ to induce SIPS·H2O2 is the major agent signalling at low physiological levels in the nanomolar range. It interacts with specific protein targets, engaging in metabolic regulation and stress responses to facilitate cellular adaptation to a changing environment and stress [55]. When exposed to high levels of ROS, cells often undergo apoptosis or enter a senescent state. Cellular senescence refers to the irreversible growth arrest and loss of proliferative capacity in cells. Under conditions of oxidative stress, cells can enter a senescent state without telomere shortening, known as SIPS, which may be a protective mechanism to prevent further damage to the organism. Experiments examining the senescent features of H₂O₂-treated TBLCs after 24 h of normal culture and the induction of blastoids revealed that short-term treatment with high concentrations of H₂O₂ upregulated the expression of senescence-specific markers in TBLCs. The activity of SA-β-Gal also increased. Moreover, blastoids derived from H₂O₂-treated TBLCs showed upregulation of genes related to the senescent pathway. Based on these results and previous researches, we hypothesize that short-term, sublethal treatment with high concentrations of H₂O₂ induces a senescent state in TBLCs, resulting in the generation of blastoids with senescent characteristics. However, when transferred to an H₂O₂-free culture medium and cultured for a period of time, there was no significant differences when assessing the proliferative capacity of control TBLCs and H₂O₂-treated TBLCs. In fact, the cell count of H₂O₂-treated TBLCs was even higher than that of ctrl TBLCs after 24 h of passaging (Fig. S4C). This discrepancy observed in the proliferation assay conducted after a period of continued culture following H₂O₂ treatment may indicate an adaptive response of TBLCs driven by H₂O₂ treatment, aimed at recovering their proliferative capacity, similar to the sharp increase in H₂O₂-induced adaptation after exercise [56].

Previous study has shown that prolonged treatment of mESCs with H₂O₂ at concentrations higher than 150 µM does not induce cellular senescence but instead leads to apoptosis. In contrast, short-term treatment with sublethal concentrations of H₂O₂ in mESCs does not exhibit any senescent features or express senescence marker genes. Instead, they undergo transient cell cycle arrest followed by the restoration of cell proliferation rates without affecting self-renewal and pluripotency [57]. Researches have also shown that mESCs treated with 1 mM H₂O₂ for 0–2 h increase intracellular ROS levels and cell damage [58], and treated with 1 mM H₂O₂ for 4 h induces apoptosis [59]. Our results indicate that treatment of TBLCs with 1 mM H₂O₂ for 1 h upregulates the expression of senescent-specific marker genes and increases the activity of SA-β-Gal. Moreover, blastoids derived from H₂O₂-treated TBLCs also show upregulation of genes related to the cellular senescent pathway. In the future, more experiments and investigations are needed to explore the effects of specific ROS at different concentrations and treatment times on aging and apoptosis in TBLCs.

In addition to their involvement in cellular senescence and apoptosis, ROS also play a role in the maintenance and conversion of pluripotency and totipotency. There are differences in the intracellular ROS levels between mESCs and another pluripotent-like cell type called 2 cell-like cells (2CLCs), with higher ROS levels observed in 2CLCs, which are involved in regulating totipotency [60]. In fact, ROS is a generic term encompassing a diverse range of oxidant molecules with distinct properties and biological functions. These functions span from signaling roes to potentially causing cellular damage. It’s essential to identify the specific ROS, its concenntrations and signalling targe that play a key role. Therefore, there are still many aspects to explore regarding the association between specific ROS and TBLCs.

The construction of senescent blastoids using senescent TBLCs provides a model to simulate the developmental processes of embryos of AMA, offering insights into the pathological mechanisms and potential strategies to improve pregnancy outcomes. The quality and developmental potential of embryos are significantly influenced by parental age. Owing to cellular senescence, many embryos cultured in vitro are arrested at specific stages of development, resulting in failed embryonic development [61], [62]. Cellular senescence constitutes a mechanism to prevent the further development of low-quality embryos and the transmission of abnormal genomes [63], [64]. Thus, cellular senescence plays a complex role in the fertility of women with AMA. The aging TBL-blastoids we constructed also exhibited characteristics such as abnormal morphology, incorrect cell lineage composition, and decreased in vivo and in vitro developmental potential. By analyzing the transcriptomic data of control and aging TBL-blastoids, we found that H₂O₂-treated TBL-blastoids were enriched in pathways related to cellular senescence and apoptosis, while pluripotency and oxidative phosphorylation-related pathways were downregulated. In terms of in vivo and in vitro developmental potential assays, H₂O₂ deciduae showed adownregulation of decidualization markers. By using the DOSE toolkit to investigate gene-disease associations, we identified enriched diseases related to fertility and pregnancy complications. TBL-blastoids derived from senescent TBLCs likely influence the maternal uterus, leading to the upregulation of genes associated with fertility-related diseases and impacting implantation and development. Future studies could employ old mice as surrogates or senescent EEOs to explore the interplay between TBL-blastoids and the maternal uterus under the internal environment. Additionally, investigating the role of enriched complications of pregnancy could provide further insights into their relationship with fertility-related diseases, AMA, and pregnancy outcomes.

Several genes associated with endocrine and sex hormones play crucial roles in successful implantation [61,63]. Maternal hormones interplay with signalling molecules to promote implantation, and hormone receptors are important for pregnancy outcomes [5]. Genes involved in steroidogenesis, such as CYP17A1, CYP19A1, and FSHR, are crucial for female fertility [62], [65]. NR5A1 and BMP15 are involved in ovarian steroidogenesis and granulosa cell function, and their expression is related to follicular development and hormone synthesis [64,65]. LEP (leptin) plays a critical role in placental development, implantation, and trophoblast proliferation [66], [67], [68]. These genes associated with fertility-related diseases and hormone regulation represent potential targets for improving implantation and pregnancy outcomes.

In conclusion, cellular senescence plays a complex role in the fertility of women with AMA. To develop novel therapeutic strategies and improving pregnancy outcomes, it is crucial to explore the mechanisms and signalling pathways of cellular senescence in female reproductive ageing. The employment of TBL-blastoids derived from totipotent cells presents a valuable model for studying early embryonic development and investigating the pathological mechanisms underlying poor infertility of women with AMA. Moreover, coculturing TBL-blastoids with EEOs offers a platform for studying implantation and hormone regulation. Further research is required to understand the interaction between TBL-blastoids and the maternal uterus, as well as the implications of gene-disease associations and hormone-related genes in fertility-related diseases and implantation.

Conclusion

We established a novel blastoid, TBL-blastoids, using totipotent cells-TBLCs. In terms of morphology, cell lineages and transcriptome, TBL-blastoids are closer to blastocysts than other blastoids. TBL-blastoids are also capable of recapitulating early embryonic development, especially implantation. Therefore, TBL-blastoid is a high-qualified model to explore implantation. Furthermore, we generated senescence TBL-blastoids to explore the impacts of aging on implantation ability and underlying mechanisms. Senescence impaired the developmental potential of TBL-blastoids in vitro and in vivo. we identified some candidates involving in this effect, which might also contribute to the impaired implantation in embryos of AMA women.

Y.F., Y. Y., R.L. and Z.M. developed the concepts and designed the experiments. Y.L. and C.A. joined the experiment design, performed the experiments and wrote the manuscript. K.Z performed the experiments and data analysis. P.Z. joined the manuscript draft and revision, and RNA-seq data analysis. H.P. guided the RNA-seq data analysis. D.L., H.L., Q.G., and W.W. joined the data discussion. All authors have read and approved the final manuscript.

CRediT authorship contribution statement

Yuxin Luo: Investigation, Visualization, Writing-original draft, Writing-review & editing. Chenrui An: Visualization, Writing-original draft. Ke Zhong: Investigation, Visualization. Ping Zhou: Data Curation. Dan Li: Methodology. Hui Liu: Methodology. Qing Guo: Methodology. Wei Wei: Methodology. Hen Pan: Validation, Formal analysis. Zheying Min: Supervision. Rong Li: Conceptualization, Funding acquisition, Resources. Yang Yu: Conceptualization, Funding acquisition, Supervision, Writing-original draft, Writing-review & editing. Yong Fan: Conceptualization, Funding acquisition, Project administration, Writing-review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: