Single-cell RNA sequencing reveals abnormal fluctuations in human eight-cell embryos associated with blastocyst formation failure

Abstract

Infertility has become a global health issue, with the number of people suffering from the disease increasing year by year, and ART offering great promise for infertility treatment. However, the regulation of early embryonic development is complicated and a series of processes takes place, including the maternal-to-zygotic transition. In addition, developmental arrest is frequently observed during human early embryonic development. In this study, we performed single-cell RNA sequencing on a biopsied blastomere from human eight-cell embryos and tracked the developmental potential of the remaining cells. To compare the sequencing results between different eight-cell embryos, we have combined the research data of this project with the data previously shared in the database and found that cells from the same embryo showed a higher correlation. Additionally, the transcriptome of embryos with blastocyst formation failure was significantly different from developed embryos, and the gene expression as well as cell signaling pathways related to embryonic development were also altered. In particular, the expression of some maternal and zygotic genes in the failed blastocyst formation group was significantly altered: the overall expression level of maternal genes was significantly higher in the failed blastocyst than the developed blastocyst group. In general, these findings provide clues for the causes of human embryonic arrest after the eight-cell stage, and they also provide new ideas for improving the success rate of ART in clinical practice.

Introduction

As the prevalence of infertility increases worldwide, ART has become one of the most important clinical treatments for infertility. However, a series of adverse medical outcomes, including high miscarriage rate, high birth defects rate and abnormal early embryonic development, was also related to ART (Boulet et al., 2016). Clinical practice and scientific research have found that the development of in vitro human embryos is inefficient, and only 40–60% successfully developed into blastocysts, thus posing a challenge to improving the success rate of ART (Gardner et al., 2000; French et al., 2010).

The developmental arrest of preimplantation embryos is a common phenomenon in the process of ART, and the mechanisms of blastocyst formation failure after human eight-cell embryos are still under investigation. During early embryonic development, the transcription and expression of specific genes in the embryo undergo a series of tremendous changes (Wu et al., 2018; Godini and Fallahi, 2019). Previous studies have shown the causes for embryonic arrest are very complex, and factors such as embryo culture environment (Wale and Gardner, 2016), chromosomal abnormalities (Ambartsumyan and Clark, 2008), the transition from meiosis to mitosis (Clift and Schuh, 2013) and maternal-to-zygotic transition (MZT) (Sha et al., 2020a) can all lead to abnormal embryonic development or even miscarriage.

Embryonic development is initially in a transcriptionally quiescent state and is guided by maternal proteins and RNAs in the oocyte cytoplasm (Xia et al., 2019). Fertilized embryos during early development, however, require the restart of gene transcription, and zygotic genome activation (ZGA) is one of the key events responsible for the transition from maternal to zygotic regulation (Schulz and Harrison, 2019). In this process, the embryo begins to synthesize its own mRNAs and proteins, and no longer relies on the maternal factors for subsequent development (Jukam et al., 2017). Studies have confirmed that, as one of the events of embryonic development, defects in maternal mRNA degradation and the ZGA process are associated with early developmental arrest of in vitro fertilized human embryos (Lee et al., 2014; Eckersley-Maslin et al., 2018). Given the results of previous studies, when a human embryo reaches the eight-cell stage, abnormal MZT may be one of the reasons for blastocyst formation failure. With advances in technology, we are able to identify highly resolved gene expression patterns and observe the development of human embryos (Potter, 2018). Previously, we have systematically analyzed the transcriptome of each stage from the oocyte and each subsequent embryonic development stage using single-cell high-throughput sequencing analysis and obtained a landscape of gene expression during human early embryonic development (Yan et al., 2013).

Human embryos cultured in vitro are prone to arrest at any stage, and particular problems arise at the first cleavage division, as well as at blastocyst transition (Zamora et al., 2011; Qi et al., 2014). The transition to blastocyst failed in various ways after the eight-cell stage, such as cell lysis and fragmentation, and the collapse of cavitation, which coincides with the time of maternal factor degradation and embryonic genome activation (Artley et al., 1992). To explore the reasons behind blastocyst formation failure, fertilized embryos were obtained through ICSI and a single cell from the eight-cell embryo was biopsied through micromanipulation for single-cell RNA sequencing (scRNA-seq). Subsequently, the embryos were tracked to observe their developmental outcomes and determine whether they have the potential to grow into blastocysts after the biopsy. The results showed that the gene expression of embryos in the blastocyst formation failed group was significantly different from that of the developed embryo group: the expression levels of maternal genes were increased, and the signal pathways related to embryonic development were also altered. Our study provides clues for the causes of abnormal human blastocyst formation and new ideas for possible clinical improvements in the success rate of ART.

Materials and methods

Sample acquisition and informed consent

The volunteers involved in this project were healthy couples who have naturally conceived and given birth to healthy offspring, and husbands have normal semen parameters and indicators according to the standards proposed by the World Health Organization (Cooper et al., 2010). The human experiments conducted in this study were approved by the Reproductive Research Ethics Committee of Peking University Third Hospital (License number: 2013SZ018), all embryos were obtained with written informed consent signed by the donor couples, confirming that donor couples voluntarily donated gametes and embryos only used for scientific research on human early embryonic developmental mechanisms, without financial payments.

A total of three couple donors were included in this study and the average age of the women was 32.6 years, the BMI was 23.59 ± 0.55 kg/m². The ovarian stimulation of the donors followed routine clinical programs, and oocytes were obtained through vaginal puncture under ultrasound guidance following standard clinical procedures (Huang et al., 2009) at the Reproductive Medical Center of Peking University Third Hospital. Mature metaphase II stage oocytes were screened for ICSI, and embryos were obtained through ICSI using a micromanipulation system (Nikon Instruments, Tokyo, Japan). After 67–68 h of ICSI, one of the blastomeres of the eight-cell embryo was biopsied. We defined all embryos donated by each couple as one batch, and finally obtained three batches of embryos.

Embryo culture at different developmental stages

Embryo culture was carried out according to the procedures described previously (Yan et al., 2013). Briefly, embryos after ICSI were cultured in 20 μl G1 medium (Vitrolife, Västra Frölunda, Sweden) covered with mineral oil (Sigma Aldrich, MO, USA) at 37°C, with a 6% CO₂ and 100% humidity environment inside the embryo incubator (Thermo Fisher, MA, USA) until the eight-cell stage on the third day. A total of 25 surviving eight-cell embryos were collected from three batches, and a single blastomere was randomly biopsied for each eight-cell embryo by laser-assisted micromanipulation. The remaining cells of the embryos were cultured in 20 μl G2 medium (Vitrolife, Västra Frölunda, Sweden) covered with mineral oil until the blastocyst stage. After the biopsy, the embryo’s developmental status was recorded and photographed at a fixed time each day until the sixth day. An overview of the experimental design is shown in Supplementary Fig. S1.

Morphology and quality assessment of the eight-cell embryos and blastocysts

The embryos were evaluated for their developmental potential at the eight-cell and blastocyst stage. At the cleavage stage, we considered important indicators including blastomere size, cell morphology, cavitation, fragmentation and embryo densification (de los Santos et al., 2014). Next, the cell morphology for blastocyst quality was considered according to Gardner’s blastocyst scoring system, as previously reported (Minasi et al., 2016). Briefly, a grade was given according to the degree of blastocyst expansion and hatching status (numbers 1 to 6, 1 and 2 represent the primary status, 3 and 4 represent the expanding status, 5 and 6 represent the hatching status). We conducted a second evaluation to assess the quality of inner cell mass (ICM) and trophectoderm (TE) for fully expanded blastocysts. For ICM, grades A, B and C are used to label tightly grouped cells, loosely packed cells, and minimally present cells, respectively. For TE, grades A, B and C represent cells forming a cohesive epithelium, a loose epithelium and scattered cells that do not form an epithelium, respectively.

Single-cell RNA-seq library preparation and sequencing

To reduce the false positive rate when screening differentially expressed genes (DEGs), we sequenced the transcriptome of each single-cell twice, according to a previously published MR-seq (measure a single-cell transcriptome repeatedly) method (Yang et al., 2017), and then we combined the results of the two measurements. In brief, a mouth pipette was used to quickly transfer a single blastomere into the lysis buffer. The cell lysate mixture was split into two equal parts for reverse transcription, and a poly-A tail was added to the 3′ end of the first-strand cDNA using terminal deoxynucleotidyl transferase. The reverse transcription product was then used as a template to perform 10–12 PCR cycles to amplify the single-cell cDNA library, as described in our previous study (Tang et al., 2010), and we performed real-time quantitative PCR analysis on a set of housekeeping genes to verify the quality of the amplified single-cell cDNA. About 200 ng cDNAs for every single cell were sheared into short fragments using a Covaris S2 ultrasonicator (COVARIS INC, MA, USA). The adaptors were then added and a sequencing library was constructed following the manufacturer’s protocols for the TruSeq DNA Library Preparation Kit (Illumina, CA, USA).

RNA sequencing data processing

All the single-cell RNA-seq raw data were preprocessed to remove the Illumina adaptor sequences and 3′poly-A sequences (Zhou et al., 2016), and the RNA-seq reads were then aligned using hg19 as the reference under STAR v.2.7.0e in a per-sample two-pass mode (Tukiainen et al., 2017). The featureCounts tool (Liao et al., 2014) was used to obtain the read counts for each gene, and the Fragments Per Kilobase of exon model per Million mapped fragments (FPKM) was calculated (Supplementary Table SI) according to the formula: read counts/(mapped reads (Millions) * exon length (KB)), and the raw data have been deposited in the Genome Sequence Archive (GSA) database as HRA001473.

Sequencing depth and quality control

The RNA-seq library was sequenced on the Illumina Sequencing Platform. The original data generated by the sequencer was converted into sequence data by base calling, and then passed through the standard quality control (QC) process to remove all unqualified readings. Next, an RNA-seq quality control software, RSeQC (version: v3.0.0) were used to calculate the transcript integrity number (TIN) to assess the integrity of sample RNA, according to the standard process (Wang et al., 2016). Samples with varying degrees of degradation would be removed from subsequent analysis. In addition, we filtered the expressed genes by selecting genes with FPKM >1 in at least one cell and heterogeneous genes among the different eight-cell embryos were deleted. Finally, a total of 16 867 qualified genes were obtained for subsequent analysis.

Analysis of heterogeneity among the eight-cell embryos

The dataset was downloaded from the Gene Expression Omnibus database (GSE36552) to conduct inter-embryonic and intra-embryonic heterogeneity analysis (Yan et al., 2013). Only genes with FPKM >1 in at least one cell were analyzed, and a value of 1 was added to the expression level of each gene before log₂ transformation. The principal component analysis (PCA) method was used to cluster 20 single cells selected from three different embryos. After calculating the coefficient of variation (CV) for each gene of the three embryos, the CV values were arranged in descending order. By selecting the top 30% CV genes for each embryo and then compare for overlapping, a total of 152 genes with high variability between the eight-cell embryos was obtained. Then, the Corrplot package (version: 0.84, https://Github.Com/Taiyun/Corrplot) was used to visualize the inter-embryonic and intra-embryonic correlation coefficient, through the Pearson correlation coefficient.

Copy number variation and embryo sex using scRNA-seq data

We inferred the copy number variation (CNV) status of each cell based on the average expressions of a genomic-ordered list of genes following the method reported in the previous study (Patel et al., 2014). Briefly, the expression of each gene was normalized to the average expression of 50 upstream and downstream genes, and the genes were reordered according to their positions in the genome. Subsequently, the normalized gene expression values were further centered to infer the CNV of each chromosome. Subsequently, according to the expression levels of X- and Y-chromosome-specific genes, the sex of the embryo could also be roughly inferred.

Cell cycle estimation using scRNA-seq data

We next analyzed genes related to cell cycle regulation, including the core set of 43 G1/S and 54 G2/M genes reported in a previous study (Li et al., 2017). Furthermore, the expression levels of genes related to cell cycle regulation in both developed and blastocyst formation failed groups were analyzed based on the average FPKM.

Acquisition and analysis of differentially expressed genes

The DESeq2 package (Love et al., 2014) was used to find the DEGs, and 202 DEGs (Supplementary Table SII) were selected for log₂ (fold change) values >1 or < −1 with statistical significance (P-value < 0.01). Subsequently, DEGs were used to perform Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis (Supplementary Tables SIII and SIV). A significant difference between the two groups would be identified if the P-value was <0.05.

A list of maternal and zygotic genes was acquired by analyzing our previous data (Yan et al., 2013): 8498 maternal genes and 6576 zygotic genes were individually intersected with 202 DEGs and identified 92 maternal and 67 zygotic genes that were differentially expressed.

Correlation between differentially expressed genes and embryonic quality

The P-value for each gene was corrected through multiple tests to avoid false positives, and the value of Padj was calculated. We selected 19 DEGs with a Padj value <0.2, and the correlation between these DEGs and embryonic development was calculated based on embryo quality using the Pearson correlation coefficient. Furthermore, the Search Tool for the Retrieval of Interacting Genes (STRING, version: 11.0, http://string-db.org/) was used to determine the genetic interactions related to embryonic quality, and the interactions among the selected genes were illustrated in our results and using the medium confidence (>0.400) for the minimum required interaction score. Subsequently, the DevOmics database (http://devomics.cn/) (Yan et al., 2021) was used to obtain gene expression levels (RNAseq) during human preimplantation embryo development.

Statistical analysis

R packages (R, Vienna, Austria) were applied for bioinformatics analysis, and the experimental data were analyzed using SPSS statistical software (IBM, version 25.0, NY, USA). Statistical data were presented as mean ± SEM, and Student’s t-test with two tails was performed to determine the significance. If the P-value was <0.05, the difference was considered to be significant.

Results

Developmental potential of biopsied embryos

A total of 25 human eight-cell embryos were collected from three batches of sample collection and the remaining cells were cultured for another 3 days after blastomere biopsy to track the development potential of the embryo. It was found that out of the 25 cultured embryos, 13 failed to form blastocysts, while the other 12 embryos successfully grew into blastocysts, and the blastocyst formation rate in this study was about 48%, which indicated that the micromanipulation and laser-assisted single-cell biopsy did not significantly affect the embryonic developmental potential, which is consistent with the blastocyst formation rate of about 50% in the case of embryo non-biopsy (Qi et al., 2014). Representative embryo morphology of the two groups after blastomere biopsy is shown in Fig. 1A, and the embryo quality was evaluated from Day 3 to Day 6 (Supplementary Table SV).

Subsequently, we analyzed the scRNA-seq data, and calculated the median TIN and the number of expressed genes in all samples. Three embryos with the number of expressed genes less than 6000 or the median TIN less than 25 were considered to have poor RNA quality (Fig. 1B and C). To ensure the reliability of data analysis, these three samples were excluded in the subsequent analysis.

Heterogeneity between different human eight-cell embryos and cells in the same embryo

To evaluate the heterogeneity of eight-cell human embryos, we re-analyzed previously generated data (Yan et al., 2013), where the embryonic transcriptome during different stages of development was systematically analyzed by scRNA-seq. RNA-seq data of 20 single cells of three eight-cell stage human embryos were analyzed using PCA through an unsupervised approach, and the 20 cells were divided into three subgroups through PCA dimensionality reduction, suggesting that heterogeneity existed among the three embryos (Fig. 2A).

Furthermore, we used the Pearson correlation coefficient to analyze the correlation of single-cell RNA-seq data between embryos and within the same embryo. We found the average Pearson correlation coefficient between embryos was 0.8020 ± 0.0065, and the coefficient within the same embryo was 0.9529 ± 0.0027, and the inter-embryo analysis showed a significantly lower correlation coefficient compared with intra-embryos (Fig. 2B, P < 0.0001). The details of cells of the same eight-cell embryo are shown in Fig. 2C–E and show extremely high similarities. Comparatively, high heterogeneity was detected among the three eight-cell embryos, indicating potential transcriptome differences among the embryos, while the cells within the embryos were highly similar in terms of RNAseq data.

Copy number variation, embryo sex and cell cycle between embryos

In order to detect genomic deletions and rearrangements suggested by the sequencing data, we analyzed the chromosome copy number of all samples. The results suggested that four embryos had autosomal or sex chromosomal partial deletions (Supplementary Fig. S2A). Of these embryos, three failed to develop into blastocysts and another one has postponed the process of blastocyst formation. By analyzing the ratio of specific gene expressions and the total expression level of the X and Y chromosomes (Supplementary Fig. S2B and C), we found a greater X-chromosome specific gene expression ratio and a low Y chromosome total expression level in nine embryos, suggesting that these nine embryos were female and the other 13 were male. Seven of the 13 male embryos developed to the blastocyst stage, while four of the nine female embryos developed into blastocysts.

During embryonic development, cell mitosis is kept in an active state. Therefore, the expression levels of cell cycle-regulating genes in the developed and arrested embryos were analyzed, but no significant differences were found (P = 0.761 for the G1/S phase and P = 0.626 for the G2/M phase, n = 11) between the developed and arrested embryos (Supplementary Fig. S3A and B).

Blastocyst formation failed embryos showed a large difference in gene expression levels

Firstly, we performed dimensionality reduction analysis on embryos based on gene expression. The t-distributed stochastic neighbor embedding (t-SNE) results showed that, whether embryos can successfully develop to the blastocyst stage or not, they have similar transcriptome expression at eight-cell stage (Supplementary Fig. S4A). To further explore the reasons behind human blastocyst formation failure, we characterized the 202 DEGs between the arrested and the developed group (Fig. 3A). The results indicated that 126 genes were upregulated and 76 genes were downregulated (Supplementary Fig. S4B). Furthermore, we used DEGs to determine whether the embryonic development outcomes were correlated with these genes. The t-SNE results showed that the use of DEGs can better distinguish the two groups of the embryos, with respect to their developmental outcomes (Fig. 3B). This finding indicated that these DEGs might play an important role in regulating the development of embryos after the eight-cell stage.

To explore the expression levels of maternal and zygotic genes in DEGs, we individually intersected the genes expressed in the oocytes and the genes activated during the ZGA stage with the 202 DEGs (Fig. 3C). The expression levels of the maternal genes in the blastocyst formation failed group was significantly higher than the developed group (P < 0.0001, Fig. 3D).

Subsequently, the obtained DEGs underwent comparison and gene ontology (GO) analysis. The relating biological processes, molecular functions and cellular components of the DEGs were annotated (Supplementary Fig. S4C). GO enrichment analysis results suggested that the genes involved in cell division regulation, and cell bipolarity establishment and maintenance in the blastocyst formation failed group were significantly different from those in the developed group. Next, the results of KEGG enrichment analysis of all DEGs indicated that compared to the developed group, the Wnt signaling pathway, stem cell pluripotency regulation and phosphatidylinositol signal pathway in the blastocyst formation failed group were all affected (Fig. 3E). Finally, GO terms related to negative regulation of meiosis cell cycle, stem cell division and negative regulation of the Wnt signaling pathway were selected for in-depth analysis in terms of gene expression (Fig. 3F). The expression levels of genes related to the selected GO terms were significantly altered in the blastocyst formation failed group, among which 21 maternal genes were upregulated and four zygotic genes were downregulated.

Correlation between differentially expressed genes and blastocyst quality

Our study has found that the expression levels of DEGs related to developmental regulation may be one of the important factors for the subsequent development process of human eight-cell embryos. To investigate the relation between DEGs and embryo developmental outcomes, the P-values were multi-corrected for each DEG and 19 of them (Padj < 0.2) were selected for correlation analysis. We found the expression level of low-density lipoprotein receptor-related protein 6 (LRP6) and zinc finger protein 557 (ZNF557) to be highly correlated with the embryo quality, with coefficients of −0.72 and 0.74, respectively (Fig. 4A). In addition, compared with the developed group, the average expression level of LRP6 was significantly upregulated in the blastocyst formation failed group (Fig. 4B, P = 0.001) at the eight-cell stage. The average expression level of ZNF557, however, was significantly downregulated in the blastocyst formation failed group (Fig. 4B, P = 0.003). According to the DevOmics database, both LRP6 and ZNF557 showed high expression levels during early embryonic development (Fig. 4C and D). However, unlike ZNF557, LRP6 expression was lowered throughout the ZGA phase, suggesting that the expression level of LPR6 and ZNF557 may affect the embryonic development process.

Next, the protein interactions of LRP6 and ZNF557 were analyzed using STRING, and the results suggested that almost all proteins that directly interact with LRP6 were also involved in regulation of the Wnt signaling pathway (Fig. 4E), and there was a strong interaction between TRIM28 and ZNF557 (Fig. 4F). These data suggest that these genes may play an important role in human embryonic development from the eight-cell stage to the blastocyst stage.

Discussion

Mammalian preimplantation embryonic development involves a series of complex processes, and the successful application of ART technology allows us to observe the developmental process of human preimplantation embryos (Niakan et al., 2012; Wamaitha and Niakan, 2018). Compared with mouse embryos, human embryos are more prone to fail to form blastocysts after the eight-cell stage (Schwarzer et al., 2012), and tracking the development of an embryo to understand its ultimate fate would offer some insights to help reveal the causes of abnormal embryonic development.

During ART, the fertilized embryos undergo multiple cleavages in vitro, and finally form transferable blastocysts. Although the cleavage biopsy will inevitably cause mild trauma to the embryo, studies have shown that would not significantly affect the clinical outcomes (Cimadomo et al., 2016), and our results were consistent with this point. Previous studies have analyzed the developmental potential of oocytes through polar bodies; however, the mechanisms of embryo arrest are unclear owing to the joint regulation of parental and zygotic sources in embryos after fertilization (Keefe et al., 2015). Therefore, tracking the development process after the eight-cell stage is important in order to study the mechanisms of embryonic development. After biopsy, the remaining cells in the embryo were found to have the potential to continue developing, suggesting that our biopsy had minor effects on the subsequent progression of development. In addition, we found that compared to embryos that developed to blastocyst, the gene expression levels of embryos with abnormal blastocyst formation had already altered at the eight-cell stage.

During the human preimplantation embryo development process, the chromosomal segmental abnormalities involving loss or gain were often found to occur, especially during the first 3 days (Babariya et al., 2017). Abnormal chromosome segregation may cause embryo aneuploidy, which would, in turn, lead to failed embryo implantation, miscarriage, and a low IVF success rate (Tšuiko et al., 2019). In this study, out of the 12 analyzed embryos that formed a blastocyst, three showed chromosomal deletions and arrested, as could be expected. The remaining nine arrested embryos showed chromosomal normalcy, and therefore, their arrest might be attributed to transcriptional factors. Indeed, deep GO analysis found significant differences between developed and arrested embryos in genes regulating cell growth, differentiation and division. These differences suggested that chromosomal aneuploidy and altered expression levels of embryonic development-regulating genes may cause developmental abnormalities after the eight-cell stage.

With the gradual degradation of most maternal mRNAs, the embryonic genome is transcriptionally activated, including a large number of key activator genes that regulate the MZT process (Li et al., 2013; Yu et al., 2016). Mouse embryos begin to enter ZGA from the two-cell stage (Schultz, 1993), while ZGA of human embryos would occur after the four eight-cell stage (Niakan et al., 2012). In this study, we observed incomplete degradation of maternal transcription and incomplete activation of zygotic genes in the blastocyst formation failed group, and the MZT process was also disturbed. These factors might correlate with the abnormal embryonic development process after the eight-cell stage. Maternal degradation requires enhanced expression of zygotic genes to clear most maternal mRNAs in time to increase the developmental potential of preimplantation embryos (Sha et al., 2020a,b). Therefore, the timely degradation of most maternal factors and the correct progression of ZGA is of great significance to subsequent embryonic development.

Interestingly, the expression levels of the maternal gene LRP6 and the zygotic gene ZNF557 are closely related to the quality of embryonic development. Although the regulatory effects of ZNF557 on embryonic development are still unclear, studies have reported that the lack of TRIM28, an interactive protein of ZNF557, will lead to epigenetic variation or even death during preimplantation embryonic development (Messerschmidt et al., 2012). In addition, Trim28 can maintain the inherent differentiation potential of embryonic stem cells and participate in the expression of embryonic stem cell pluripotency molecular markers (Hu et al., 2009). Therefore, TRIM28 plays an irreplaceable role in maintaining the stability of totipotent stem cells and epigenetic reprogramming.

By further analysis, we found that classical signaling pathways, such as Wnt and stem cell pluripotency regulation, differed in the blastocyst formation failed group and normally developing embryos. LRP6, as a key regulatory gene in the Wnt signaling pathway, showed a significant negative correlation between its expression level and embryonic development. Although a high expression level of LRP6 had been observed during the early embryonic development process, a rapid decline in LRP6 expression in human, mouse and bovine embryos would occur at the ZGA stage (Yan et al., 2013; Deng et al., 2014; Tribulo et al., 2017). This suggests that as one of the genes highly expressed after fertilization, incomplete degradation of LRP6 during the MZT may alter the subsequent embryonic development potential. Furthermore, the Wnt-β-catenin signaling pathway regulates embryonic development and stem cell proliferation through LRP5 and LRP6 (González-Sancho et al., 2004; Wang et al., 2012). Within the pathway, LRP6 is involved in degradation of β-catenin (Swiatek et al., 2006), and the accumulation of β-catenin further affects early embryonic development (Kofron et al., 2007).

Conclusion

Through scRNA-seq of single blastomeres from human eight-cell embryos, the developmental potential of different embryos can be evaluated by examining their full transcriptome level without significantly affecting embryonic development outcomes. Therefore, gene expression and changes in the transcriptome related to embryonic development may be an important cause of abnormal blastocyst formation. Our findings indicated that the use of single-cell biopsy technology to obtain valuable embryonic information can help to evaluate the potential of preimplantation embryo development. Furthermore, our study offers some novel insights into improving the success rate of ART and the quality of preimplantation embryos. It may be possible to select embryos with normal gene expression levels during the ZGA period, and optimize the in vitro culture conditions to increase the rate of blastocyst formation during the early embryonic development process.

Data availability

All data relevant to the study are included in the article or uploaded as supplementary information. The raw sequencing data generated in this paper have been deposited in the Genome Sequence Archive in the BIG Data Center (PRJCA006930), Chinese Academy of Sciences under accession codes HRA001473 that are publicly accessible at https://ngdc.cncb.ac.cn/gsa-human/.

Acknowledgments

The authors wish to thank Hai-Quan Wang from the Medical College of Nanjing University for his generous technical assistance. In addition, the authors would like to thank Yi-Ru Zhu and Jia-Xi Cheng from the Reproductive Medical Center of Peking University Third Hospital for their kind help in suggestions on polishing this manuscript. The authors also thank all the volunteers and their families who have participated in this study.

Authors’ roles

Q.-L.H., P.Y. and L.-Y.Y. conformed and designed the experiments. L.Y. and L.-Y.Y. carried out embryo manipulation and molecular biology experiments. Q.-L.H., P.Y., Z.-Q.Y., Y.-D.C., W.C. and S.-M.K. conducted bioinformatics analysis. Q.-L.H. wrote the manuscript with P.Y., L.-Y.Y. and F.-C.T. J.Q. gave guidance and suggestions. The manuscript was reviewed by all authors.

Funding

This project was funded by the National Key Research and Development Program of China (2017YFA0103801, 2019YFA0801400 and 2018YFC1004000), the National Natural Science Foundation of China (82101743) and the China Postdoctoral Science Foundation (2021M700285).

Conflict of interest

The authors declare that they have no conflict or financial interest.

References

Author notes