The formation of the meiotic spindle–chromosomal complex from G0/G1 somatic chromosomes
We postulated that the somatic chromosomes with pseudo-meiotic spindles originating from the diploid G0/G1 nucleus (2n) of the somatic cell are composed of two single chromatids (2n/2c). These chromatids can be segregated by fertilization and haploid somatic pronucleus (1n/1c) could be produced in the zygote (Fig. 1a).
a Schematic of diploid somatic nucleus haploidization by mature ooplasm. SCNT oocyte with pseudo-meiotic spindle derived from diploid G0/G1 somatic nucleus (2n) could be composed of two single chromatids (2n/2c), and chromatids could be segregated (1n) at fertilization. Half of the segregated chromatids (1n/1c) could be extruded to the pseudo-polar body (PPB), whereas the rest (1n/1c) remained in a zygote with a sperm nucleus. b Spindle reformation in SCNT oocytes. The spindle was cleared 2 h after SCNT. Scale bars, 20 μm. c Spindle-chromosomal complexes in intact MII and SCNT oocytes. Metaphase-like spindle-chromosomal complex was observed at 2 h (blue, DNA; green, α-tubulin). Scale bars, 25 µm. d Spindle–chromosome complexes in intact MI, MII, and SCNT oocyte. SCNT spindles showed a similar chromosomal arrangement to MI oocyte (blue, DNA; green, α-tubulin; red, kinetochore). Scale bars, 5 µm. e PPB extrusion depending on resting time after SCNT. 2 h resting after SCNT resulted in significantly higher PPB extrusion than that in other groups. mean ± s.e.m. (*P < 0.05, by ANOVA with Tukey analysis) among the groups. n means the number of 2PN/1PPB embryos/the number of fertilized embryos. Three technical replications for each group. f Spindle and nuclear changes after fertilization in intact MII and SCNT oocytes. SCNT oocytes showed PPB extrusion (red arrows) and 2 PN formation (white arrows), Scale bars, 20 μm.
We imaged SCNT oocytes generated from somatic cells under a noninvasive polarized microscope. The spindles were not observed within 30 min after SCNT. The newly formed spindles first became visible 1 h 30 min after SCNT and were clearly organized 2 h after it (Fig. 1b and Supplementary Movie 1). In addition, α-tubulin, a protein required for chromosome segregation during cell division, was stained at 1, 1.5, 2, and 3 h after SCNT. The microtubules were not detected until 1 h and exhibited a prometaphase-like arrangement at 1.5 h (Fig. 1c). The comparable spindle–chromosomal complex to that in intact MII oocytes was observed at 2 h. The anaphase-like microtubule arrangement was detected at 3 h. Moreover, the SCNT oocytes, as well as control MI and MII oocytes, were fixed and labeled with α-tubulin, a protein of microtubules, and kinetochore, a multiprotein complex that assembles on centromeric DNA and constitutes the main attachment interface between chromosomes and microtubules8. As expected, the kinetochores in the control MI oocytes appeared as punctate, parallel signals (red) at the equatorial region of the spindle (Fig. 1d). Chromosomes were stretched by centromere–kinetochore pairs toward opposite spindle poles. By contrast, MII spindles carried one row of centromeres with microtubules connected on both sides of each centromere. The chromosomal arrangement pattern in some SCNT spindles was similar to those observed in MI oocytes.
We next examined the ability of the SCNT spindles to segregate chromosomes into the pseudo polar body (PPB) and SH zygote via NT-IVF. The PPB extrusion rate was evaluated in SH zygotes, which were rested for 30 min as well as 1, 1.5, 2, and 3 h after SCNT. The SCNT oocytes that were rested for 2 h exhibited a significantly higher PPB extrusion rate than that of other groups (Fig. 1e, Supplementary Table 1, and Supplementary Data 1).
Spindles in the SH zygote disappeared gradually, a PPB was extruded, and two pronuclei (PN) were formed (Fig. 1f). Staining confirmed that the PPB and both PN contained DNA. The timing of these events and the morphology of the SH zygote were indistinguishable from those of the control.
Fasudil, retinoic acid, and RS-1 promote the segregation of homologous chromosomes
The SCNT technique has been performed as described previously9. Briefly, a hemagglutinating virus of Japan envelope was applied to fuse the donor somatic cells with the enucleated oocytes. After resting time for 30 min to 1 h, the reconstructed oocytes were activated with strontium and HDAC inhibitors such as Trichostatin A or Scriptaid. Based on this conventional method, we modified the protocol for NT-IVF, which was extended resting time for 2 h. Additionally, the caffeine was treated before and during SCNT micromanipulation to prevent premature oocyte activation and to prompt spindle reformation in SCNT oocytes10,11. The normal SH zygote morphology was two PN and one PPB (2PN/1PPB), while irregular SH zygotes were 2PN/0PPB, 1PN/1PPB, 3PN/0PPB, and 1PN/0PPB (Fig. 2a).
a Various morphologies of somatic haploid (SH) zygotes. A normal SH zygote (2PN/1PPB), which was similar to the intact control zygote (2PN/PB), was observed. Other morphologies of SH zygotes were also produced. Blue arrows: PN; red arrows: the second PB and PPB in the intact control zygotes and SH zygotes, respectively. Scale bar, 20 μm. b Improved 2PN/1PPB formation with Fasudil (Fa) treatment during IVF (*P < 0.05, by Independent-group t test). Sc, scriptaid. c Enhanced 2PN/1PPB formation with retinoic acid (RA) treatment during the resting time (*P < 0.05, by Independent-group t test). d, Increased spindle formation in SCNT oocytes with RA treatment. The spindle formation rate was significantly increased with RA treatment compared with the RA-free condition (*P < 0.05, by Independent-group t test). However, once the spindle was reformed, the rates of 2PN/1PPB became comparable. e Improved 2PN/1PPB formation with RS-1 treatment during IVF (*P < 0.05, by Independent-group t test). f Advanced schematic protocol for somatic haploidization. Caffeine (250 μg/ml) was supplemented before and during SCNT. RA (300 ng/ml) was treated for 30 min after SCNT. After RA treatment, the SCNT oocytes were rested for 1.5 h before IVF. Sc (80 ng/ml), Fa (3 μg/ml), and RS-1 (4 μg/ml) were added to the medium during IVF and the overnight culture. g SCNT oocytes with reconstructed spindles and the development of SH embryos. The SCNT oocytes with reconstructed spindles were fertilized, and the SH zygotes showed normal development up to the blastocyst. Scale bar, 20 μm. h Improved development of preimplantation embryos with the advanced protocol (aNT-IVF). aNT-IVF significantly increased the rates of 2PN/1PPB and blastocysts compared with NT-IVF. (*P < 0.05, by ANOVA with Tukey analysis) among the groups. n in the 2PN/1PPB graphs and the blastocyst graphs of b, c, and e; the number of 2PN/1PPB embryos/the number of fertilized embryos and the number of blastocysts/the number of 2PN/1PPB embryos, respectively. Six–10 technical replications for b, c, e, and h, and four technical replications for d. mean ± s.e.m.
To improve the somatic haploidy, fasudil (ROCK, rho-associated protein kinase, inhibitor), retinoic acid (RA), and RAD51-Stimulatory Compound 1 (RS-1) were treated in SCNT oocytes or SH embryos. The first, the fasudil was treated during IVF. Since ROCK supports spindle assembly in mature oocytes12, fasudil might assist the spindle decomposition during fertilization and enhance the PPB extrusion through the regulation of microtubule polarity13. The rate of 2PN or 2PN/1PPB formation was calculated based on the number of fertilized embryos. Approximately 75% of the SH zygotes formed 2PN (Supplementary Fig. 1a). The yield of SH zygotes with the proper morphology, 2PN/1PPB, in the fasudil-treated group (n = 58/184, 32%) was significantly higher than that of such SH zygotes in the untreated group (n = 24/109, 22%; P < 0.05), whereas blastocyst rates were similar between the groups (Fig. 2b, Supplementary Table 1, and Supplementary Data 1).
RA initiates the entrance of the prophase of meiosis I during oogenesis14,15. Because we proposed that the premature chromosomes from the G0/G1 somatic cell could be similar to the prophase of meiosis I of the oocyte, we tested several incubation times in SCNT oocytes for 30 min, 1 h, and 2 h (Supplementary Fig. 1b and Supplementary Table 1). Treatment with RA for 30 min produced the highest number of 2PN/1PPB SH zygotes compared with the other groups (52%, 46/89 for 30 min group vs. both 24%, 20/83 for 1 h and 17/70 for 2 h groups, respectively, P < 0.05). The 2PN/1PPB rate in the RA-treated group was significantly higher than that in the untreated group (51%, n = 104/202 vs. 32%, n = 58/184; Fig. 2c, Supplementary Table 1 and Supplementary Data 1). We then observed the spindle formation in SCNT oocytes using a noninvasive imaging system to examine the effect of RA (Fig. 2d). In the RA-treated group, 59% (n = 65/110) SCNT oocytes developed spindles, which was significantly higher than the rate in the untreated group (44%, n = 46/104; P < 0.05). By contrast, once the novel spindle was formed, the rates of 2PN/1PPB were comparable between the treatment groups (92%, 60/65 in RA vs. 89%, 41/46 in RA free; 2PN/1PPB zygotes/spindle reconstructed oocytes; Fig. 2d and Supplementary Data 1).
Finally, we evaluated the effect of RS-1, which enhances the expression of the homologous recombinases, such as Rad51 and Dmc116. The RS-1 could support the alignment of homologous chromosomes of the somatic nucleus in SCNT oocytes thus increasing PPB extrusion. To determine the optimum treatment timing, we applied RS-1 during the resting time after SCNT, during IVF, and overnight after IVF (Supplementary Fig. 1c and Supplementary Table 1), resulting that the treatment during IVF resulted in a significantly higher rate of 2PN/1PPB than that in other groups. 2PN/1PPB rate with RS-1 treatment during IVF was significantly higher compared to the non-treatment (67%, n = 148/221 vs. 51%, n = 104/202, P < 0.05; Fig. 2e, Supplementary Table 1, and Supplementary Data 1). However, the rates of blastocysts were comparable. Based on the results of these treatments, we established an advanced (a)NT-IVF protocol; RA was treated for 30 min after SCNT, rested SCNT oocytes for 2 h before IVF, and scriptaid, fasudil, and RS-1 were treated during IVF and the overnight culture (Fig. 2f).
These treatments could promote the spindle reconstruction and formation of normal SH zygotes. The SH zygotes had a normal morphological development up to the blastocyst stage (Fig. 2g and Supplementary Movie 1). The incorporation of these treatments significantly increased 2PN/1PPB (67%, n = 148/ 221, vs. 17%, n = 31/184; 2PN/1PPB/fertilized) and blastocyst (50%, n = 27 /54 vs. 29%, n = 4/14; blastocysts/morula) formation rates compared with non-treated (P < 0.05; Fig. 2h and Supplementary Table 1). When the blastocyst rate was calculated from MII oocytes, 12% oocytes (n = 27/229) could develop blastocysts with the aNT-IVF protocol, while only 2% (n = 4/198) with the non-treated NT-IVF one (Supplementary Table 2). The blastocyst rates were 91% (n = 138/152; blastocysts/morula) and 69% (n = 11/16) in intact IVF and regular SCNT, respectively. As we expected, intact IVF showed a significantly higher blastocyst rate than that of aNT-IVF (Fig. 2h, Supplementary Table 1, and Supplementary Data 1). However, the rate of blastocysts tended to be higher in regular SCNT (69%) compared to aNT-IVF (50%).
Reciprocal segregation of somatic homologous chromosomes
Although morphological examinations of SCNT oocytes and SH zygotes are informative and conclusive, evaluations of somatic chromosomal haploidization can only be determined by genetic analysis of chromosomal content. We initially conducted whole-exome sequencing (WES) analyses of FVB/N (FVB) and C57BL/6 N (B6) mouse strains and determined homozygous single-nucleotide polymorphisms (SNPs) that differentiate each strain (Supplementary Fig. 2a and Supplementary Data 2). A catalog representing unique high-confidence SNPs could distinguish FVB from B6 chromosomes. We designed primers amplifying one region (1100 to 1500 bp) with the highest SNP frequency in each of 19 autosomal chromosomes and sequenced those regions using the MiSeq platform for FVB or B6 genotyping (Supplementary Fig. 2b, c).
To determine the segregation pattern of somatic chromosomes, we generated SCNT oocytes by transferring FVB/B6 (a cross between female FVB and male B6) or B6/FVB (the opposite combination: a cross between female B6 and male FVB) somatic cells. The SCNT oocytes were fertilized with B6 sperm (Fig. 3a). A total of 15 SH zygotes, 10 from FVB/B6 and 5 from B6/FVB, with normal 2PN/1PPB appearance, were cultured to the 2-cell stage. PPB and blastomeres were separated to perform whole-genome amplification (WGA) for the assessment of the segregation of the 19 autosomes using MiSeq (Supplementary Data 3, sheet 1). The X chromosome was analyzed by Sanger sequencing (Supplementary Fig. 3a). The segregation of somatic FVB (red) and B6 (blue) homologous chromosomes from the somatic cells were observed in most PPBs and SH embryos (Fig. 3b). Initially, we hypothesized that FVB SNPs could only be detected in either PPB or SH embryo because the somatic donor was heterozygous. However, the exome sequencing showed the recombined homozygous SNPs in heterozygous somatic donors (Supplementary Fig. 3b and Supplementary Data 2), which could make to detect the FVB SNPs in both PPB and SH embryo. Therefore, if FVB SNPs were detected in SH embryos and their corresponding PPB, we also considered the proper segregation of somatic chromosomes. We first checked the zygosity of chromosomes in each PPB, resulting in that 10–20 chromosomes were homozygous, which could be haploidy segregated from somatic genomes (Fig. 3c). Among them, either FVB or B6 chromosomes were identified randomly in each chromosome. Next, the number of properly segregated chromosomes into PPB and SH embryos was analyzed. In total, 9–20 homologous chromosomes were properly segregated between SH embryos and their corresponding PPBs (Fig. 3d). Some chromosomes in PPBs showed heterozygosity (74%) or were not amplified (26%), suggesting that these homologous chromosomes were not separated and extruded to PPBs or remained in embryos (Fig. 3e and Supplementary Data 1). In the SH embryos, 66% and 68% of haploid chromosomes in the FVB/B6 and B6/FVB combinations were originated from the FVB strain respectively, suggesting that the somatic genome remaining in SH embryos after haploidization was more species-specific rather than maternally or paternally biased (Fig. 3f and Supplementary Data 1). We also analyzed the segregation for each chromosome in 15 SH embryos. Chromosome 1 was segregated properly in all 15 embryos, whereas the other 19 chromosomes were separated in 8–14 embryos (Supplementary Fig. 3c).
a Schematic procedure to determine the segregation pattern of somatic chromosomes. FVB/B6 or B6/FVB fibroblasts were transferred into enucleated oocytes and fertilized with B6 sperm. The 2PN/1PPB 2-cell embryos were separated into SH embryos and PPB, and artificial whole-genome amplification (WGA) was performed for genotyping. b Chromosome genotypes in PPBs and SH embryos using MiSeq. Proper segregation patterns of somatic homologous chromosomes were displayed in most PPBs and SH embryos. Red bars, FVB; blue, B6; light blue, heterozygosity with FVB and B6; white, the absence of PCR amplicons; S above the last bars, sex chromosome. c Number of homozygous chromosomes in PPBs. 10–20 chromosomes showed haploid, which could be haploidy segregated from somatic genomes. The genotype between FVB and B6 was random in each chromosome. d Number of properly segregated homologous chromosomes. Properly segregated chromosomes were analyzed by considering the homozygosity of donor cells as revealed by exome sequencing. Nine to twenty homologous chromosomes were segregated reciprocally between SH embryos and corresponding PPBs. e Location of nondisjunction chromosomes. 74% of nonseparated homologous chromosomes were located in PPBs. mean ± s.d. f Origin of somatic haploid in SH embryos. In all, 66–68% of haploid chromosomes were originated from FVB. Mean ± s.d.
In addition, SCNT oocytes transferred with FVB/B6 somatic cells were chemically activated without sperm. Ten PPBs were isolated and analyzed for homozygosity in all chromosomes (Supplementary Fig. 3d and Supplementary Data 3, sheet 2). Five PPBs were detected FVB or B6 homozygous genotype (GT) in all chromosomes, suggesting that chemical activation could also promote the segregation of somatic homologous chromosomes.
Next, we selected four pairs of 2-cell embryos and their corresponding PPBs, two from FVB/B6 (SH embryos 3 and 9) and two from B6/FVB (SH embryos 11 and 14) somatic cells, and performed WES (Supplementary Data 2). In the FVB/B6 combination, the copy numbers of chromosomes, analyzed by exome data, showed that 13 (in SH embryo 3) and 11 (in SH embryo 9) homologous chromosomes were properly segregated (Fig. 4a, b). The remaining chromosomes were nondisjunct and either extracted to PPBs (green boxes in Supplementary Fig. 4a, b) or remained in SH embryos (blue boxes in Supplementary Fig. 4a, b). In the properly segregated homologous chromosomes, the homologous chromosomes were separated as completely FVB or B6 between the PPB and the embryo in the whole-exome area (Fig. 4c, d and Supplementary Fig. 4a, b). Of those, 6 and 7 chromosomes, respectively, were of the FVB genome in embryos.
a, b Copy number variation (CNV) profiles of SH embryos 3 and 9 and their corresponding PPBs generated by FVB/B6 somatic donor. In total, 13 and 11 homologous chromosomes were properly segregated in SH embryos 3 and 9, respectively. The other chromosomes were nondisjunct and extracted to PPBs (red asterisk) or remained in SH embryos (blue asterisk). Relative CNV was interpreted by comparison with the control IVF embryo, second PB, and C57BL/6 mouse tissue as a control. c, d Chromosome map of SH-embryo 3 and 9 and their corresponding PPB. Thirteen (in SH-embryo 3) and 11 (in SH-embryo 9) chromosomes were segregated reciprocally. e, f CNV profiles of SH embryos 11 and 14 and their corresponding PPBs generated by B6/FVB somatic donor. Copy numbers of chromosomes displayed proper segregation of somatic homologous chromosomes in all 20 chromosomes of both embryos. g, h Chromosome map of SH-embryo 11 and 14 and their corresponding PPB. All chromosomes were segregated reciprocally in both embryos.
By contrast, in both embryos from the B6/FVB combination, the copy numbers of chromosomes were properly segregated in all 20 chromosomes (Fig. 4e, f). However, unlike in the FVB/B6 combination, both the PPB and the embryo contained the FVB genome (Fig. 4g, h and Supplementary Fig. 5a, b). Much of the shared FVB genome between the PPB and the embryo showed homozygosity in somatic donor cells (Supplementary Fig. 5c, d), indicating that the shared FVB SNPs were derived from homozygous SNPs of somatic cells and the homozygous FVB SNPs were segregated into the PPB and the embryo. However, some shared SNPs between the PPB and embryo were not carried in somatic cells as homozygous status, because (1) every single somatic cell transferred into an ooplast could harbor different SNPs, and (2) the pooled somatic cell cannot represent all homozygous SNPs of every single cell.
These results suggest that two single homologous chromosomes (2n/2c) in the SCNT oocytes were segregated randomly after fertilization, producing truly haploid cells from somatic cells in zygotes.
Contribution of somatic genomes in all chromosomes of SH embryos
To determine the contribution of somatic chromosomes to SH embryos, we used adult fibroblasts derived from female homozygous FVB strains to generate SCNT oocytes. SCNT oocytes were fertilized with B6 sperm, and embryos at the 2-cell or blastocyst stage were examined for all 20 chromosome GTs (Fig. 5a).
a Schematic illustrating the contribution of somatic chromosomes to SH embryos. The adult fibroblasts derived from homozygous FVB mice were used for somatic cell nuclear transfer. SCNT oocytes were fertilized with B6 sperm. Artificial whole-genome amplification was performed on 2-cell embryos and blastocysts and analyzed all 20 chromosome genotypes. b The sex ratio of the SH embryos. Of all the embryos, 22 (55%) were male and 17 (45%) were female. c The number of heterozygous chromosomes. Among the blastocysts showing heterozygosity in all 20 chromosomes, WES was performed for BL6, 10, and 14 (red font). M and F indicate male and female, respectively. d The frequency of somatic detection in 20 chromosomes. Three 2-cell embryos (n = 3/19, 16%) and 6 blastocysts (n = 6/19, 32%) harbored the somatic origin in all 20 chromosomes. e Chromosome map of the male SH blastocysts. FVB genomes were confirmed by WES in blastocysts 6, 10, and 14 (somatic origin) in all 20 chromosomes. f CNV profile of the SH blastocysts by exome data. Euploidy was shown in blastocysts 6, 10, and 14 in whole chromosomes.
A total of 19 2-cell stage embryos and 19 blastocysts were screened in chromosome 2 and sex chromosomes (Supplementary Fig. 6a, b). The FVB (somatic origin) and B6 (sperm origin) genomes were detected in nine 2-cell embryos (47%) and 11 blastocysts (58%) in chromosome 2 (Supplementary Fig. 6c). Sex determination revealed that 55% (n = 21/38) of the SH embryos were male, indicating that these embryos inherited their X chromosome from the somatic cell genome (Fig. 5b and Supplementary Fig. 6d).
The 20 FVB-positive embryos in chromosome 2 were genotyped for all autosomes using MiSeq (Supplementary Data 3, sheet 3 and 4), and the X chromosome of the female embryos was analyzed by Sanger sequencing (Supplementary Fig. 7a). The results showed that three 2-cell embryos (n = 3/19, 16%) and six blastocysts (n = 6/19, 32%) were FVB/B6 heterozygous across all 20 chromosomes, suggesting the presence of SH embryos in all 20 chromosomes with somatic haploidy (Fig. 5c, d and Supplementary Fig. 7b, c). All three 2-cell embryos and five blastocysts were male. The remaining 11 embryos included B6 or FVB homozygosity in a few chromosomes, which could indicate partial haploidization in those embryos or amplification errors (Supplementary Data 3, sheets 3 and 4). The increase in SH in blastocysts could be due to the arrest of many aneuploid SH 2-cell embryos before reaching the blastocyst stage.
To confirm the somatic origin and euploidy in whole chromosomes, we performed WES for selected male blastocysts (BL6, BL10, BL12, and BL14; Supplementary Data 2). The FVB genomes were detected in the whole chromosomes in 3 blastocysts (BL6, BL10, and BL14), but not in chromosomes 1, 8, 9, 18, and 19 of blastocyst 12 (Fig. 5e and Supplementary Fig. 7d). The normal euploidy copy number of all chromosomes was detected in blastocysts 6, 10, and 14, suggesting complete somatic haploidy in all chromosomes (Fig. 5f). As expected, in blastocyst 12, chromosomes 1, 8, 9, 18, and 19 were haploid with only the sperm genome, whereas chromosome 6 was triploid (Supplementary Fig. 7e).
Global gene expression in SH-ESCs
Because the analysis of embryos had limitations due to WGA or technical errors, ESCs from SH blastocysts (SH-ESCs) were established using B6/FVB somatic cells to validate the results found in the embryos. We generated several SH-ESC lines, and the efficiency of ESC derivation was 7% (4 ESCs/55 NT-IVF BL), significantly lower than that of ESC from IVF embryos (75%, 13 ESCs/23 IVF BL) (Supplementary Fig. 8a). These SH-ESCs were demonstrated a typical morphology and a normal diploid karyotype (Fig. 6a–d and Supplementary Figs. 8b and 13). The single male cell line, with the biopsied sample of the same blastocyst, was further investigated (Supplementary Fig. 8c). The results showed that 3 or 4 chromosomes were heterozygous in the SH embryo derivatives in regions analyzed using MiSeq (Fig. 6e). The remaining chromosomes harbored the B6 GT. The typical 2n copy number for all 20 chromosomes was confirmed in biopsied blastocysts and SH-ESCs (Fig. 6f). The GTs by exome sequencing revealed that the FVB genome was similarly displayed between the blastocysts and SH-ESCs (Fig. 6g and Supplementary Data 2). Chromosomes 2, 5, and 11 originated mainly from the FVB, whereas chromosome 13 did not harbor any FVB SNPs in the SH-ESCs.
a The morphology of the SH-ESCs. SH-ESCs (SH-ES4) derived from SH blastocysts showed the typical morphology of mouse ESCs. Scale bar, 100 μm. b Three germ layer formations of the SH-ESCs by teratoma assay. Scale bars, 100 μm. c Diploid configuration of the SH-ESCs by cell cycle analysis. The histogram refers to the cell cycle profile of the SH-ESCs resulting in a 2n nuclear configuration. d Representative image of the diploid SH-ESCs chromosome spread of all 40 chromosomes. Scale bars, 10 μm. e Genotype of SH embryo derivatives by MiSeq. Outgrowth refers to an inner cell mass of outgrowth in the SH blastocyst-plated dish. Blue bars: B6 genotype; light blue: heterozygous status with FVB and B6. Three or four chromosomes were heterozygous in SH embryo derivatives and the remaining chromosomes harbored the B6 genotype. f Copy number variation profile of the SH-blastocysts and SH-ES4 with exome data. Diploid copy numbers were displayed in all 20 chromosomes. Relative CNV was interpreted by comparison with the controls, in vitro fertilization embryo, second polar body, and C57BL/6 mouse tissue. g Chromosome map of SH-blastocysts and SH-ES4. The distribution of FVB SNPs was similar between the blastocysts and the SH-ES4.
We examined the global gene expression patterns of SH-ESCs using RNA-seq in comparison with intact ESCs. A total of 22,014 genes were expressed in at least one of the intact ESCs or SH-ESCs. Only 200 genes were determined with a significant P value (P < 0.05), and the clustering of these 200 genes resulted in a separation between the intact ESC and SH-ESC lines (Fig. 7a). Among them, seven genes with significant adjusted p-values were considered to be differentially expressed (false discovery rate, FDR < 0.05) in SH-ESCs (Fig. 7b and Supplementary Fig. 8d). Six genes (GM29100, Lef1os1, 8030451A03Rik, Sncg, Gm41724, and Gm21992) were expressed at a significantly higher level, whereas one gene (Antxrl) was expressed at a significantly lower level in SH-ESCs (Fig. 7c). The phenotype or detailed in vivo function of all these genes, except Sncg, has not been reported yet (Supplementary Fig. 8e). The Sncg is known to be related to neurodegenerative disease based on its overexpression17, which could interrupt the full development of SH embryos in vivo.
a Heat map displaying 200 genes with significant difference (P < 0.05) between intact ESCs and SH-ESCs. The clustering of gene expression with a significant P value resulted in a separation between intact ESC and SH-ESC lines. b Venn diagram showing the number of upregulated or downregulated genes in SH-ESCs compared to intact ESCs. c Heat map displaying seven differentially expressed genes between intact ESCs and SH-ESCs (false discovery rate, FDR < 0.05). Six genes were upregulated, whereas, one gene was downregulated in SH-ESCs.
Because the regulation of imprinting genes is important for embryonic and fetal growth or reprogramming to a pluripotent state, a total of 105 gene expressions were investigated in SH-ESCs18. We focused on several imprinting genes, such as H19, Igf2r, and Grb10 as paternally imprinted genes, and Igf2 and Snrpn as maternally imprinted genes (Supplementary Fig. 8f). The results suggested that these genes showed no significant difference in SH-ESCs compared with intact ESCs (P > 0.05). The remaining imprinting genes (100 genes) also displayed no significant difference between SH-ESCs and intact ESCs. Based on these results, we concluded that the gene expressions of SH-ESCs were comparable to those of intact ESCs.
SH-embryos are able to produce live offspring
To evaluate the full-term development of SH embryos, we generated SH blastocysts using somatic cells from various strains, including FVB fetal and adult fibroblasts, B6/FVB fetal fibroblasts, as well as B6D2F1/Crl (BDF1, a cross between C57BL/6NCrl female x DBA/2NCrl male) and FVB cumulus cells (Fig. 8a and Supplementary Fig. 9a). The rates of 2PN/1PPB were comparable among donor strains, in approximately 65%, but were lower than the rates in the unmanipulated IVF controls (94%). BDF1 donors yielded higher blastocyst rates among donor strains (30% vs. 11–18%).
a In vitro development depends on the somatic cell and sperm. The rates of 2PN/1PPB were comparable among donor strains, type, and sperm strain, whereas BDF1 donor cells showed higher blastocyst rates than other donor strains. *,#P < 0.05; **P < 0.01 between the two groups by Fisher’s exact test. b In vivo production from SH embryos with different strain combinations of somatic cell and sperm. n indicates the number of in vivo production/number of transplanted blastocysts. c Diploidy of the SH-implanted embryos. SH-implanted embryos had a diploid genome in all chromosomes by array comparative genomic hybridization analysis, and the results for the SH-implanted embryos were analyzed by comparison with DBA/2 mouse tissues as a control. d CNV profile of SH-implanted embryo 1 by exome sequencing. CNV using exome sequencing data showed diploidy in SH-implanted embryo 1. Relative CNV was interpreted by comparison with the controls, in vitro fertilization-embryo, second PB, and C57BL/6 mouse tissue. e Chromosome map of SH-implanted embryo 1. FVB unique indicates specific FVB SNPs of somatic cells against those of DBA/2. All 20 chromosomes contained FVB-unique SNPs. f SH mice on day 1 (left) and 15 weeks (middle) after birth and the first generation (F1) of SH mice (right). g The body weight of SH mice and F1 of SH mice after birth. The body weights of the SH mice were significantly lower than those of SH-F1 and intact intracytoplasmic sperm injection control pups. Mean ± s.d. n means the number of mice for each group. *P < 0.05, by ANOVA with Tukey analysis.
Only SH blastocysts with a good morphology were transferred into recipients (Supplementary Fig. 9b). The quality of SH blastocysts was poor compared with that in the unmanipulated IVF controls (33% vs. 96%; Supplementary Fig. 9c). Only 8.6% 2PN/1PPB zygotes were developed to blastocysts evaluated as being of good quality (n = 18/232), compared with 65% (n = 132/206) in the controls (P < 0.05; Supplementary Fig. 9d). In addition, 9 SH blastocysts underwent trophectoderm biopsy and were subjected to WGA to determine embryo ploidy with copy number variation (CNV) analysis using Illumina. Of these, three were normal euploid (33%, n = 3/9), of which two were female and one was male (Supplementary Fig. 9e, f).
A total of 95 blastocysts were generated using FVB somatic cells and B6 sperm were transferred into nine recipients, resulting in no pregnancies (Fig. 8b and Supplementary Fig. 9g). The somatic cells that originated from the B6/FVB hybrid mouse also failed to result in pregnancy (118 blastocysts to 10 recipients). The B6/FVB hybrid somatic cells with sperm from BDF1 hybrid mice improved the pregnancy rate. A total of 121 blastocysts were transferred into 35 recipients, resulting in 3 pregnancies. The implanted embryos from two recipients were lost, but the uterus from the third recipient was collected on day 7 after embryo transfer. The three implanted spots (embryos) were observed in the collected uterus, and their DNA was extracted (Supplementary Fig. 9h). Array comparative genomic hybridization was performed to detect euploidy in the three SH embryos. All three embryos had a diploid genome, and genomic alterations were not detected (Fig. 8c). One of these embryos (SH embryo 1) was selected to analyze chromosomal CNV using WES data, with the results confirming the diploidy (Fig. 8d). The FVB origin in SH-implanted embryo 1 was analyzed using WES data (Supplementary Data 2). We conducted WES of DBA/2 and determined FVB-specific SNPs compared with DBA/2. All 20 chromosomes contained FVB-specific SNPs (Fig. 8e). FVB genetic bias occurred in the SH-implanted embryos, similar to the case of SH-ESCs.
Finally, 81 SH blastocysts combined with BDF1 somatic cells and BDF1 sperm were transferred into 27 recipients, resulting in one pregnancy and the delivery of 3 female pups (Fig. 8f). The average bodyweight of the pups was 1.1 g, which was significantly lower than that of the intracytoplasmic sperm injection (ICSI) controls (1.66 ± 0.17 g) (Fig. 8g, Supplementary Fig. 9i and Supplementary Data 1). All three pups survived and grew into adulthood (Fig. 8f). Furthermore, all SH mice were mated with BDF1 males and produced three healthy first-generation litters, which had birth weights similar to those of ICSI controls (Fig. 8g, Supplementary Fig. 9j, and Supplementary Data 1). All first-generation litters of the SH mice survived and grew into adulthood (Fig. 8f).
