(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Single-cell multiomics analyses of spindle-transferred human embryos suggest a mostly normal embryonic development [1]

['Shuyue Qi', 'Biomedical Pioneering Innovation Center', 'School Of Life Sciences', 'Peking University', 'Beijing', 'Wei Wang', 'Department Of Obstetrics', 'Gynecology', 'The Seventh Medical Center Of Chinese Pla General Hospital', 'Chinese Pla General Hospital']

Date: 2022-08 

Mitochondrial DNA (mtDNA) mutations are often associated with incurable diseases and lead to detectable pathogenic variants in 1 out of 200 babies. Uncoupling of the inheritance of mtDNA and the nuclear genome by spindle transfer (ST) can potentially prevent the transmission of mtDNA mutations from mother to offspring. However, no well-established studies have critically assessed the safety of this technique. Here, using single-cell triple omics sequencing method, we systematically analyzed the genome (copy number variation), DNA methylome, and transcriptome of ST and control blastocysts. The results showed that, compared to that in control embryos, the percentage of aneuploid cells in ST embryos did not significantly change. The epiblast, primitive endoderm, and trophectoderm (TE) of ST blastocysts presented RNA expression profiles that were comparable to those of control blastocysts. However, the DNA demethylation process in TE cells of ST blastocysts was slightly slower than that in the control blastocysts. Collectively, our results suggest that ST seems generally safe for embryonic development, with a relatively minor delay in the DNA demethylation process at the blastocyst stage.

Funding: This work was supported by grants from the National Key R&D Program of China (2018YFA0107601) (to FT) and the National Key R&D Program of China (2018YFC1003003) (to WS) ( http://www.most.gov.cn/index.html ). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Data Availability: All the raw sequence data reported in this paper have been deposited in the Genome Sequence Archive50 in National Genomics Data Center (National Genomics Data Center Members and Partners, 2020), Beijing Institute of Genomics (China National Center for Bioinformation), Chinese Academy of Sciences, under accession number HRA001110 that are publicly accessible at http://bigd.big.ac.cn/gsa-human . All the code associated with this research has been deposited on GitHub ( https://github.com/sherryxuePKU/MRT_scTrio-seq ).

Copyright: © 2022 Qi et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Additionally, replacement of oocyte mutant mtDNA and normal development of offspring by ST was shown to be effective in nonhuman primates [ 12 , 18 , 19 ]. The results from several groups using human oocytes suggest that the development to the blastocyst stage and derivation of embryonic stem cells (ESCs) of ST embryos are indistinguishable from those of control embryos in vitro, with negligible mtDNA carryover (<1%) [ 18 , 20 ]. In addition, mitochondrial respiratory chain enzyme activities and oxygen consumption rates of isolated ESCs and differentiated cells of ST embryos were shown to be comparable to those of control [ 20 – 22 ]. These results suggest that isolating maternal metaphase II spindles for nuclear transfer can be accomplished with minimal mtDNA carryover and has potential clinical applications [ 23 ]. In April 2016, Zhang and colleagues used the ST procedure to deliver a live boy in Mexico whose mother’s mitochondria carried the mutation causing Leigh syndrome [ 24 ]. The boy is still under long-term observation, but to date, no systematic omics study has been implemented on the safety of the ST procedure in humans.

To prevent the transmission of mitochondrial disease through the germline, clinicians have developed several different strategies for mitochondrial replacement therapy, including pronuclear transfer [ 9 , 10 ], spindle transfer (ST) [ 11 , 12 ], and polar body transfer [ 13 – 15 ]. The core principle of the abovementioned methods involves uncoupling the inheritance of mtDNA from that of the nuclear genome through the combination of nuclear DNA from oocytes or zygotes of a woman with mutated mtDNA with the cytoplasm of oocytes or zygotes from a healthy donor to obtain reconstructed embryos with both a normal nuclear genome and normal mtDNAs; these techniques have been developed over the past few decades. Currently, mitochondrial replacement therapy, a type of germline gene therapy, seems to be the only promising way to block the inheritance of mtDNA mutation-related diseases [ 16 , 17 ].

Mitochondria are indispensable organelles of most eukaryotic cells and play an essential role in numerous metabolic processes such as oxidative phosphorylation, apoptosis, and calcium homeostasis. Mitochondria are also the only organelles in humans with their own genetic material, which includes 37 genes with 13 encoding mitochondrial proteins and 24 generating noncoding RNAs. Mutations in mitochondrial DNA (mtDNA) may cause mitochondrial dysfunction and lead to debilitating or devastating diseases that can affect virtually any organ at any age [ 1 , 2 ]. Mitochondrial diseases are among the most common inherited metabolic diseases [ 3 , 4 ], with 1 in 5,000 being the minimum prevalence rate for mtDNA mutations in adults [ 5 ]. Moreover, mitochondrial genome is invariably maternal inheritance [ 6 ], and at least 1 in 200 offspring of female carriers carry an mtDNA mutation that potentially causes disease [ 7 ]. Prenatal and preimplantation genetic diagnosis are available reproductive options for preventing the intergenerational transmission of mtDNA mutation-related diseases. However, these alternatives are only appropriate for women who can produce oocytes and embryos with low enough percentage of mutant mtDNAs [ 8 ].

Results

First, we collected 2,207 individual cells from 46 blastocysts (S1 Table) from the 2 experimental groups. After stringent filtration, 1,397 cells (768 cells from 23 ST blastocysts and 629 cells from 22 intracytoplasmic sperm injection (ICSI) control blastocysts) were retained for subsequent analyses (Fig 1A and 1B). On average, in each individual cell, we detected 7,683 expressed genes and 190,494 copies of mRNAs (Fig 1C).

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Single-cell Trio-seq2 sequencing of spindle-transferred human embryos. (A) Schematic of single cell collection, transcriptome sequencing, and DNA methylome sequencing and analysis. (B) Summary of the number of cells and embryos before and after quality control for the ICSI and ST groups. (C) Average number of UMIs and genes and percentage of MT transcripts in 1,397 cells passing quality control. (D) Ternary plots of lineage scores. (E) A total of 1,397 cells retained after quality control projected onto a t-SNE map and colored according to group (left) and lineage (right). The numerical data are listed in S1 Data. EPI, epiblast; ICSI, intracytoplasmic sperm injection; MT, mitochondrial; PBAT-seq, post-bisulfite adaptor tagging sequencing; PE, primitive endoderm; QC, quality control; scMethylome-seq, single-cell DNA methylome sequencing; scRNA-seq, single-cell RNA-seq; ST, spindle transfer; STRT-seq, single-cell tagged reverse transcription sequencing; TE, trophectoderm; t-SNE, t-distributed stochastic neighbor embedding; UMI, unique molecular identifier. https://doi.org/10.1371/journal.pbio.3001741.g001

Unsupervised t-distributed stochastic neighbor embedding (t-SNE) analysis revealed that these cells could be partitioned into 3 main clusters. According to the canonical markers and lineage scores, these clusters were determined to be epiblast (EPI), primitive endoderm (PE), and trophectoderm (TE). Lineage-specific marker genes were verified in our data, including SOX2, NANOG, and POU5F1 (also known as OCT4) for EPI; SOX17 and FOXA2 for PE; and GATA2, GATA3, and CDX2 for TE (Figs 1D and 1E and 2). For all 3 lineages, the cells from ST blastocysts and from ICSI blastocysts were tightly clustered in the t-SNE plot, indicating that the differentiation of these 3 lineages is generally normal in ST embryos (Fig 1E). Hierarchical clustering analysis based on the expression of lineage-specific genes also showed that the developmental potential of ST blastocysts was comparable for all 3 lineages of cells (Fig 2C).

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Lineage identification of scRNA-seq data. (A, B) The lineage identification results were verified by (A) the expression levels of high-confidence lineage markers and (B) the expression levels and percentages of lineage marker sets. (C) Expression patterns of lineage signature genes. Hierarchical clustering of 1,397 cells with Euclidean distance according to the Ward.D2 method. And lineage-specific genes were identified and intersected with known markers of the corresponding lineages. The numerical data are listed in S1 Data. EPI, epiblast; ICSI, intracytoplasmic sperm injection; PE, primitive endoderm; ST, spindle transfer; TE, trophectoderm. https://doi.org/10.1371/journal.pbio.3001741.g002

Next, to explore the differences between the ST and control blastocysts in more detail, we analyzed the differentially expressed genes (DEGs). The patterns of gene expression were nearly identical for all of the TE, EPI, and PE lineages between the ST and control embryos. There were 24, 11, and 0 DEGs in the TE, EPI, and PE lineages, respectively (Figs 3A and S1A and S2 Table). Regression analysis showed that the ST and control blastocysts were highly similar between corresponding lineages, with high correlation coefficients (R2 values) for TE (0.990), EPI (0.987), and PE (0.942) (Fig 3A). Calculated variance (SD) of expression level for EPI, PE, and TE lineages in the ST and control ICSI embryos was similar in all 3 lineages and even lower in PE (S1C Fig). And random down-sampling analysis suggested R2 values decreased as the sample size decreased. And when the sample size was 10, R2 values for EPI (0.948) and TE (0.952) were close to that of PE (0.942) (Figs 3A and S1B). These suggested that the small sample size rather than significant variation within the group was the cause of lower R2 value for PE.

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Comparative analysis of ST and ICSI control embryos. (A) Analysis of DEGs and linear regression analysis of ST and ICSI embryos. The ST-to-ICSI ratio of the mean expression level (ln(TPM/10+1)) for each commonly expressed gene was fitted to y~x. The correlation coefficient of each fitting and the number of up-regulated (red) or down-regulated (blue) DEGs are listed for each lineage. (B) ST and ICSI embryonic cells projected onto a t-SNE map of AUC scores from SCENIC analysis and colored according to lineage (left) and group (right). (C) Hierarchical clustering of AUC scores from SCENIC analysis. Cells of the same lineage (red, EPI; green, PE; blue, TE) clustered together. EPI and PE cells belonged to the same branch and converged to the root with TE branches. ST (yellow) and ICSI (purple) clustered together regardless of cell lineage. The numerical data are listed in S2 Data. DEG, differentially expressed gene; EPI, epiblast; ICSI, intracytoplasmic sperm injection; PE, primitive endoderm; SCENIC, single-cell regulatory network inference and clustering; ST, spindle transfer; TE, trophectoderm; t-SNE, t-distributed stochastic neighbor embedding. https://doi.org/10.1371/journal.pbio.3001741.g003

Several sex chromosome-linked genes, such as XIST, VCX2, VCX3B, EIF1AY, RPS4Y1, and DDX3Y, were identified among the DEGs between the ST and control blastocysts. Presumably, this is because the percentages of male and female embryos were different between the ST and control groups. In the ST group, there were 6 male and 17 female embryos, whereas in the control group, there were 14 male and 8 female embryos (Figs 1B and S2A). Supervised hierarchical clustering plots showed that the unbalanced sex ratio of the embryos resulted in sex-linked genes seemingly being expressed group specifically in the TE and EPI lineages (S2B Fig). Comparison of the expression levels of identified sex-linked DEGs in the embryos of the same gender between ICSI control group and ST group showed that although these genes tended to be expressed differentially between ST and ICSI embryos, almost all of them were not statistically significant with Bonferroni correction (S2C Fig). Significant difference between ST and ICSI groups was only seen for XIST in TE cells of male embryos with Log 2 (fold change) less than 1 (0.91) (S2C Fig). Simulation within TE also carried out and when the number of female embryos in the ST group was 3, these 2 groups had the closest sex ratio and the lowest ratio (6.3%) of sex-linked DEGs (S3 Table). These confirmed our hypothesis that imbalanced sex ratio was indeed the main confounding factor for the DEG analysis of the sex-linked genes.

Notably, single-cell regulatory network inference and clustering (SCENIC) analysis based on regulons also divided the cells into 3 distinct lineages, and cells of the ST blastocysts still clustered together with the same lineage of cells of ICSI control blastocysts in both the t-SNE map and the hierarchical clustering plot (Fig 3B and 3C). Specifically, the ST and control blastocysts were highly similar, if not identical, in terms of transcriptional regulatory networks.

DNA methylation plays critical roles in the development of mammalian embryos. Using the scTrio-seq2 method, we subsequently selected 268 individual cells from 32 embryos for DNA methylome sequencing, and 217 cells from 29 embryos (132 cells from ST blastocysts and 85 cells from ICSI control blastocysts) passed quality control (S3A Fig). The t-SNE analysis and DEG analysis revealed the same conclusion that, in general, there was no difference in the transcriptomes of the same lineage of cells between the ST and ICSI control groups for cells whose DNA methylomes were also sequenced and analyzed (Figs 4A and S3E and S2 Table).

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. DNA methylome analysis of selected cells. (A) Two hundred and two cells remained after quality control and were projected onto a t-SNE map of their transcriptome data. The 3 lineages were clearly separated, while the ST and control groups clustered together well. (B, C) Overall DNA methylation levels for each lineage of cells. The p-values in the boxplot (B) were calculated by the Wilcoxon sum test (ns, p > 0.05; **, p < 0.01). The DNA methylation levels of the embryonic cells distinguished by group (column) and lineage (row) are listed in Table (C). (D) DMR analysis. Hypomethylated regions (genomic regions with DNA methylation levels low in ST cells but high in ICSI cells) and hypermethylated regions (regions with DNA methylation levels low in ICSI cells but high in ST cells) were annotated to DNA elements. The DMR frequency shows the enrichment of DMRs within DNA elements. For example, a hypermethylated DMR frequency in CGIs indicates the ratio of the number of hypermethylated DMRs located in CGIs to the total number of CGIs. (E) Ridge plot of the DNA methylation pseudotime trajectory. Cells of an earlier stage (8-cell stage, morula, blastocyst; Ref, reference) were mixed with cells in this research, and PCA was performed. The value of PC1 was interpreted as pseudotime for each cell. (F) DNA methylation levels of gene bodies and CGIs and their flanking 15-kb regions. The methylation level of each cell in the corresponding group is indicated by the transparent area around the line. The solid lines indicate the mean methylation levels of ST (purple) and ICSI (yellow) groups. The numerical data are listed in S2 Data. CGI, CpG island; DMR, differentially methylated region; EPI, epiblast; ICSI, intracytoplasmic sperm injection; PE, primitive endoderm; ST, spindle transfer; TE, trophectoderm; t-SNE, t-distributed stochastic neighbor embedding. https://doi.org/10.1371/journal.pbio.3001741.g004

We found that the global DNA methylation levels of the cells of ST blastocysts were slightly higher than those of control embryos (Figs 4B and 4C and S3C). This indicates that global DNA demethylation in the ST embryos may be slightly delayed compared with that in the control embryos.

We further investigated the DNA methylation level in each lineage. Specifically, the DNA methylation levels of EPI and PE of ST blastocysts were comparable to those of control blastocysts, with no significant differences (27.6% versus 28.3% for EPI; 26.2% versus 25.6% for PE) (Fig 4B and 4C). Additionally, the profiles of the DNA methylation levels of the gene body and CpG islands (CGIs) were similar between these two groups (Figs 4F and S5 and S6). However, the DNA methylation level of TE cells from ST blastocysts (28.5%) was 2.4% higher than that of TE cells from control blastocysts (26.1%) (Fig 4B and 4C). We also analyzed the DNA methylation levels of different genomic elements between ST and control blastocysts. In general, there were no DNA methylation level differences in the genomic regions we analyzed in EPI or PE cells between the ST and control groups. On the other hand, for the TE lineage, the majority of genomic elements we analyzed presents methylation levels that were higher in the ST embryos than in the control embryos (S3F Fig).

Differentially methylated region (DMR) analysis revealed that 152,258 300-bp genomic tiles were hypermethylated in TE cells of the ST group compared with those of the control group. On the other hand, 121,255 300-bp genomic tiles were hypomethylated in TE cells from the ST group compared with those from the ICSI control group. Both hypermethylated and hypomethylated regions of ST blastocysts were enriched in repetitive sequences (Fig 4D). And there is no obvious clustering in the genome-wide spatial distribution of these DMRs (S4B Fig).

Since human embryos undergo a drastic global DNA demethylation process during preimplantation development, it is reasonable to speculate that the DNA demethylation process is slightly delayed in ST blastocysts, resulting in higher residual DNA methylation levels. To test this hypothesis, a pseudotime trajectory was constructed using our previously published DNA methylome data of human preimplantation embryos [27]. The DNA methylation patterns of ST blastocysts were closer to those of the earlier preimplantation developmental stage, while the DNA methylation patterns of ICSI blastocysts were very similar to those of the reference blastocysts (Fig 4E). Therefore, we conclude that, compared with ICSI control embryos, ST embryos showed a slightly delayed global demethylation process at the blastocyst stage. Moreover, we profiled the DNA methylation levels of each individual embryos or cells using sequencing data of TE and the results suggested that the DNA methylation levels of ST embryos or cells had an overall increasing trend (S4A Fig), and the delay in DNA demethylation process was not caused by abnormally high DNA methylation level of a small fraction of the embryos or a small fraction of the cells.

Next, we used the transcriptome data and previously published method [28] to infer the CNVs for each individual cell and identify aneuploid cells (Fig 5A). The reliability of the deduction was confirmed by comparison with the preimplantation genetic testing results for several of the embryos. These findings were further verified by comparison with the single-cell genome sequencing data of the same individual cell by scTrio-seq2 (S7A and S7B Fig). In general, the percentages of aneuploid cells in ST blastocysts were comparable to those in ICSI blastocysts (22.7% versus 17.0%) (Fig 5B and 5C). This suggests that ST manipulation does not increase the percentage of aneuploid cells in blastocysts.

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. CNV analysis. (A) Global CNV patterns of the 1,397 cells from 45 embryos. (B, C) The CNV frequency for embryos was calculated based on the percentage of aneuploid cells in each embryo and tested by the Wilcoxon sum test (B). The numbers of euploid or aneuploid cells in the ST or ICSI control groups are listed in Table (C). The numerical data are listed in S2 Data. CNV, copy number variation; EPI, epiblast; ICSI, intracytoplasmic sperm injection; PE, primitive endoderm; ST, spindle transfer; TE, trophectoderm. https://doi.org/10.1371/journal.pbio.3001741.g005

Additionally, after the aneuploid cells were removed, the transcriptome landscapes of the euploid cells of these 3 lineages were still nearly identical between the ST and control groups, with fewer than 30 DEGs identified (S7C Fig and S2 Table), further verifying that ST manipulation in general did not affect the gene expression patterns of any of the 3 lineages of blastocyst cells.

Finally, the relationship between gene expression levels and CNVs was investigated, which suggested a positive correlation. Because the size and location of CNV varied on each chromosome, gene expression levels fluctuated between 1.04 to 1.29 when the copy number increased and between 0.92 to 0.54 when the copy number decreased (S7D Fig).

To further integrate transcriptome and DNA methylome, multiomics factor analysis (MOFA) [29] was applied to cells that had both omics data. As an unsupervised dimensionality reduction method, MOFA inferred 5 hidden factors capturing the principal sources of biological variations in multiomics data (S8A Fig). The factor 1 and factor 2 (or factor 1 and factor 5) (ranked by variance explained) together captured the characteristics of 3 cell lineages best, which could divide the cells into 3 lineages (EPI, PE, and TE), and cells from the 2 groups were still clustered together (S8C Fig). And pairwise combination of the top 5 factors could not separate the cells of ST and ICSI blastocysts (S8C Fig). Uniform manifold approximation and projection (UMAP) analysis based on these factors also divided the cells into 3 cell lineages, and cells of ST and ICSI embryos were still clustered together (S8B Fig).

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001741

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/