Crucial Role of C-Myc in the Generation of Induced Pluripotent Stem Cells

Abstract

c-Myc transduction has been considered previously to be nonessential for induced pluripotent stem cell (iPSC) generation. In this study, we investigated the effects of c-Myc transduction on the generation of iPSCs from an inbred mouse strain using a genome integration-free vector to exclude the effects of the genetic background and the genomic integration of exogenous genes. Our findings reveal a clear difference between iPSCs generated using the four defined factors including c-Myc (4F-iPSCs) and those produced without c-Myc (3F-iPSCs). Molecular and cellular analyses did not reveal any differences between 3F-iPSCs and 4F-iPSCs, as reported previously. However, a chimeric mice formation test indicated clear differences, whereby few highly chimeric mice and no germline transmission was observed using 3F-iPSCs. Similar differences were also observed in the mouse line that has been widely used in iPSC studies. Furthermore, the defect in 3F-iPSCs was considerably improved by trichostatin A, a histone deacetyl transferase inhibitor, indicating that c-Myc plays a crucial role in iPSC generation through the control of histone acetylation. Indeed, low levels of histone acetylation were observed in 3F-iPSCs. Our results shed new light on iPSC generation mechanisms and strongly recommend c-Myc transduction for preparing high-quality iPSCs. STEM CELLS 2011; 29:1362–1370

INTRODUCTION

Induced pluripotent stem cells (iPSCs) were first demonstrated via the ectopic expression of four defined factor genes, Oct3/4, Sox2, Klf4, and c-Myc, in mouse embryonic fibroblasts (MEFs) using a retrovirus gene transduction system (hereafter, we refer to these cells as R-4F-iPSCs) [1, 2]. However, it has been subsequently reported that c-Myc is not essential for the generation of iPSCs and that such cells generated by Oct3/4, Sox2, and Klf4 only (hereafter, referred to as R-3F-iPSCs) are indistinguishable from their 4F counterparts [3, 4]. In addition, negative effects of c-Myc introduction into iPSCs have also been demonstrated. For example, various cancers arise in chimeric mice developed from 4F-iPSCs, in which the silencing of exogenous genes usually occurs and the reactivation of the c-Myc oncogene is evident [3]. These observations led us to generate iPSCs in the absence of c-Myc transduction.

However, the importance of c-Myc transduction for iPSC generation has been suggested previously by genome-wide chromatin immunoprecipitation experiments [5], and the involvement of c-Myc in the pluripotency-maintenance machinery in embryonic stem (ES) cells has also been reported [6–11]. To date, substantial numbers of genes including noncoding transcripts have been identified as c-Myc-regulated genes [9, 12]. Hence, a new regulatory system that differs from the classical pathways has been proposed for specific genes [13]. The involvement of histone acetylation in such a c-Myc gene expression system, particularly the histone acetylation transferase (HAT) enzyme complex, has also been suggested [14]. We therefore hypothesized that there are crucial roles of c-Myc in iPSC generation other than its function in enhancing the incidence of these cells. In this study, we focused on the role of c-Myc in iPSC generation using newly developed genome integration-free 3F-iPSCs and 4F-iPSCs that we established from an inbred mouse strain using the 2A plasmid vector (hereafter, we refer to these cells as 2A-3F-iPSCs and 2A-4F-iPSCs) [15].

MATERIALS AND METHODS

iPSC Generation and Validation

The following mouse strains were used in this study: C57BL/6J (Japan SLC, Hamamatsu, Japan), Nanog-GFP tg (STOCK Tg(Nanog-GFP, puro)1Yam, RIKEN BRC), and C57BL/6-Tg(CAG-EGFP) (Japan SLC) [16]. MEFs were prepared from embryos of embryonic day 13.5 of each strain and used within three cell passages. The generation of iPSCs and subsequent real-time polymerase chain reaction verification were performed as described previously [17]. For immunocytochemical staining, anti-Nanog (1:50, ReproCELL, Yokohama, Japan), anti-Oct3/4 h-134 (1:100, Santa Cruz Biotechnology, Santa Cruz, CA), and anti Sox-2 Y-17 (1:200, Santa Cruz) antibodies were used [18].

Generation of Chimeric Mice and Germline Transmission Assays

Chimeric mice were produced via the aggregation method using ICR eight-cell embryos as described previously [19]. Briefly, after an overnight in vitro incubation to generate blastocysts, transplantations were performed. Chimerism was assessed by coat color in 2-week-old mice. Chimeric mice were mated with BALB/c mice to investigate germline transmission in 9-week-old animals. For 2i treatments, iPSCs were incubated in medium supplemented with 3 μM CHIR99021 (Wako, Osaka, Japan) and 1 μM PD0325901 (Wako) for 4 or 8 days. For trichostatin A (TSA) or 5-azacytidine treatments, iPSCs were incubated in medium supplemented with 20 nM TSA for 24 hours twice at a 24-hour interval, or with 1 μM 5-azacytidine for 24 hours, and then used in subsequent aggregation experiments [18, 20].

In Vitro Analysis of Aggregated Embryos

After 24, 48, and 72 hours incubations in vitro, aggregated embryos were analyzed by fluorescence microscopy. Time-lapse recordings were performed as reported previously using a device for observing embryos at around the blastocyst stage (Astec, Fukuoka, Japan) [21]. Recordings were performed from the aggregation step over a 96-hour period at 10 minute intervals.

Immunoblotting

Whole-cell extracts were prepared by sonication in sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer. Lysates of 6,000–7,000 cells were loaded per lane. Proteins were separated on 15% SDS-PAGE gels. Antiacetylated lysine Ac-K-103 (1:500, Cell Signaling Technology, Danvers, MA) and anti-Histone H4 ab10158 (1:600, Abcam, Cambridge, U.K.) antibodies were used.

RESULTS

Clear Differences Exist Between 4F-iPSCs and 3F-iPSCs in Terms of Developmental Ability

We attempted to carefully compare 4F-generated and 3F-generated iPSCs by using newly established cells generated from the inbred mouse strain, C57BL/6J, using a nonviral vector [17]. We anticipated that a common genetic background would allow us to better investigate the role of the introduced c-Myc in iPSCs, that is, with minimal external influences such as the integration of the exogenous genes into the genome through the use of a viral vector, which could potentially induce mutations. In terms of the genetic background, the use of an inbred strain was expected to significantly reduce the very large number of polymorphisms, including single nucleotide polymorphisms that arise in other mouse lines.

We examined 2A-3F-iPSC and 2A-4F-iPSC clones, which were verified previously to be genome integration-free [17]. Few differences were observed between these two types of iPSCs in terms of phenotypic, molecular biological, or cytochemical analyses. All iPSCs were found to be strikingly similar to ES cells (Fig. 1A). With regards to the expression of transcripts for the stem cell marker genes Oct3/4, Sox2, Nanog, and Fgf4, we did not detect any differences between the two types of iPSCs (Fig. 1B). No differences were observed following staining for Nanog, Oct3/4 and Sox2 proteins (Fig. 1C). Furthermore, all of the iPSC clones examined efficiently generated teratomas, including three germ layers in nude mice, thus confirming their pluripotency (Supporting Information Fig. S1). We also could not distinguish these iPSCs from ES cells [17]. Hence, our initial results were strictly consistent with previously reported findings [3, 4].

Figure 1. Comparative analysis of 2A-4F-induced pluripotent stem cells (iPSCs) and 2A-3F-iPSCs using cellular and molecular analyses. (A): Phase-contrast micrograph of iPSC colonies. Scale bar = 500 μm. (B): Analysis of stem cell marker genes by real-time polymerase chain reaction. The data were normalized using the expression of the Gapdh gene. The relative expression levels compared with each transcript in the B6ES cells are presented. SDs are also shown (n = 3). (C): Immunocytochemical analysis of Oct3/4, Sox2, and Nanog expression. Scale bar = 100 μm. Abbreviation: DAPI, 4′,6-diamino-2-phenylindole.

However, a marked difference between 3F-iPSCs and 4F-iPSCs was evident, when we performed aggregation experiments to produce chimeric mice. We used three of the 4F-iPSC clones and five of the 3F-iPSC clones for aggregation testing with eight-cell embryos. We unexpectedly observed differences in not only the numbers of embryos produced but also significantly the numbers of highly chimeric mice (Fig. 2A). All of the 4F-iPSC lines used, 4F-1, 4F-33, and 4F-EGFPtg-1, efficiently developed chimeric mice, most of which exhibited 100% chimerism. In contrast, none of the 3F-iPSC clones developed more than a few chimeric mice and few highly chimeric animals were produced in each case (Fig. 2B). Furthermore, while germline transmission was frequently observed in the mice derived from the 4F-iPSC clones, none was evident for any of the 3F-iPSC clones examined (Fig. 2B and Supporting Information Fig. S2A, S2B).

Figure 2. Comparative analysis of 2A-4F-induced pluripotent stem cells (iPSCs) and 2A-3F-iPSCs using a chimeric mouse generation test. (A): Numbers of developed blastocysts, embryos, and highly chimeric mice. SDs are also shown. (B): Chimerism of the mice developed from 2A-4F-iPSCs and 2A-3F-iPSC clones. Coat colors were assessed at 2 weeks after birth. Aggregation tests were conducted three times on clones 2A-3F-1 and 2A-3F-3. *, Could not be mated as these mice had died by 9 weeks after birth. (C): Survival curve of the chimeric mice. The survival of mice with black eyes, considered to be an indicator of high chimerism [22], was investigated. Day 0 is the day of birth. Abbreviations: iPSCs, induced pluripotent stem cells; n, germline transmission negative; p, germline transmission positive.

Interestingly, although a significant number of black-eyed mice, which are thought to grow into highly chimeric mice [22], was observed for 3F-iPSCs just after birth, few of these animals survived for longer than 2 weeks, during which we assessed their chimerism by coat color (Fig. 2C and Supporting Information Fig. S3). Our results strongly suggest that almost all of the other highly chimeric mice generated using 3F-iPSCs also died soon after birth in contrast to those produced using 4F-iPSCs. We further found that although the 3F-EGFPtg-4 and 3F-EGFPtg-15 clones had the ability to develop highly chimeric mice, that is, 100% chimerism, it was impossible to progress these mice to mating with females because they died within a short period after birth. Interestingly, when we assessed the germ cells in the highly chimeric mice produced using the 2A-3F-EGFPtg-4 clone, in which the green fluorescent protein (GFP) gene is controlled by a ubiquitous promoter, at 9 days old, strong GFP signals were detected in the spermatogonial cells (gonocytes). This strongly suggests an ability of 2A-3F-iPSCs to develop into germ cells (Supporting Information Fig. S2C–S2E). Hence, a distinct difference was observed between 4F-iPSCs and 3F-iPSCs in terms of their developmental potential.

We observed positive effects of c-Myc transduction upon iPSC generation, most notably on pluripotency, using an integration-free vector from C57BL/6J mice [15, 17]. We then conducted similar analyses of iPSCs generated from this inbred strain using another transduction system involving retroviral vectors that has been widely used for this purpose (two of the 4F-iPSC and five of the 3F-iPSC clones). Similar to iPSCs generated by 2A plasmid transfection, although the 4F-iPSCs exhibited an ability to develop fully and achieve germline transmission, few 3F-iPSCs exhibited this developmental ability, that is, 3F-iPSCs showed low levels of chimerism and inefficient germline transmission (Supporting Information Fig. S4A). In addition, we also examined another mouse strain, Nanog-GFP tg, which has been widely used in previous iPSC studies, and observed similar effects of c-Myc transduction on iPSC generation (Supporting Information Fig. S4B). Furthermore, defects in germline transmission have also been quite recently observed in the Nanog-GFP tg mouse line [23], indicating that our current observations are not limited to the C57BL/6J strain.

Defects in Blastocysts Generated by the Aggregation of iPSCs with Eight-Cell Embryos

During our testing for chimeric mice generation, we observed abnormal embryos that developed from 3F-iPSCs at the expected delivery date via a caesarean operation (Supporting Information Fig. S5). Furthermore, a small but distinct difference was evident in the implantation frequency of blastocysts developed from 3F-iPSCs and 4F-iPSCs, indicating that this difference develops before the implantation step (Fig. 3A). Thus, we carefully investigated in vitro blastocyst formation through the aggregation of iPSCs with eight-cell embryos. For this purpose, we used three GFP-expressing iPSC clones, 2A-4F-EGFPtg-1, 2A-3F-EGFPtg-4, and 2A-3F-EGFPtg-15, in which EGFP genes are governed by a ubiquitous promoter. We performed 30 aggregations for the 2A-4F clone and 29 and 30 aggregations for the 2A-3F-EGFPtg-4 and 2A-3F-EGFPtg-15 clones, respectively (Fig. 3B and Supporting Information Fig. S6A–S6C). The incorporated frequency was measured at 83.3%, 13.8%, and 83.3%, respectively. A clear difference was detected in the GFP signal area and intensity between the blastocysts generated from 3F and those from 4F-iPSCs (Fig. 3C). In most blastocysts generated from aggregations using 4F-iPSCs, the entire inner cell mass (ICM) region was GFP-positive, while in most blastocysts derived from clone 2A-3F-EGFPtg-15, the GFP signals did not cover the entire ICM region and the intensity of these signals was much lower. Furthermore, in the case of the 2A-3F-EGFPtg-4 clone, few blastocysts exhibited GFP signals and then only in small regions (Fig. 3B, 3C and Supporting Information Fig. S6A–S6C). Abnormal phenotypes were also observed in such cases. We then performed the same investigation for four more iPSC clones, 2A-3F-EGFPtg-5, 2A-3F-EGFPtg-7, 2A-3F-EGFPtg-14, and 2A-3F-EGFPtg-20 and obtained similar results (Supporting Information Fig. S6D).

Figure 3. Induced pluripotent stem cell (iPSC) potential for blastocyst formation. (A): Implantation frequency of blastocysts developed from iPSCs. The number of implantation sites is shown. A total of 19–20 blastocysts were transplanted for each of the five recipients (n = 5). (B): Blastocysts aggregated with GFP-labeled iPSCs and eight-cell embryos. Observations were performed at 48 hours postaggregation. Scale bar = 100 μm. (C): Intensity and area of GFP-positive signals in blastocysts. **, Statistical analysis was performed using Welch's t test (p < .001). (D): GFP-positive period in the aggregated blastocysts. *, Statistical analysis was performed using Welch's t test (p < .01). Abbreviations: BF, bright field; GFP, green fluorescent protein.

Because we detected a clear difference between the blastocysts generated from 3F-iPSCs and 4F-iPSCs at 48 and 72 hours postaggregation, we next performed time-lapse analysis of these aggregations over 96 hours at 10 minute intervals (Supporting Information Movies S1–S3) to better understand the process from aggregation to implantation. We found that 2A-4F-EGFPtg-1 iPSCs were efficiently incorporated into the recipient embryos and then spread over the entire ICM in most aggregations. In contrast, 2A-3F-EGFPtg-iPSCs seemed to be impaired in terms of invasion of the embryos and subsequent survival in the blastocysts. Even when invasion did occur for 2A-3F-EGFPtg-iPSCs, GFP-positive cells did not cover the entire ICM region in many cases, and GFP-positive areas within the ICM did not fuse with each other in several blastocysts (Supporting Information Fig. S6 and Movies S2, S3). Moreover, GFP signals in 3F-iPSC-derived blastocysts were lost more quickly when compared with their 4F-iPSC counterparts (Fig. 3D and Supporting Information Movies S2, S3).

Improvement of 3F-iPSCs by TSA

Having observed that 2A-3F-iPSCs are severely defective in their developmental ability, we initially hypothesized that this aberration was due to an epigenetic control defect and investigated the DNA methylation status of the promoter regions of the stem cell marker genes, Oct3/4, Nanog [17], Dppa2, Rex1, Stella, Utf1, and Esg1 [24]. However, the promoters of these genes were demethylated to a similar extent in 3F-iPSCs and 4F-iPSCs (Supporting Information Fig. S7A) [17]. We also examined the expression of imprinting genes, as a close relationship between iPSC quality, and these genes were recently suggested [25, 26]. Nevertheless, no significant reduction in the expression of imprinting genes between 3F-iPSCs, 4F-iPSCs, or control ES cell lines was observed except for the 2A-3F-4-iPSC clone in which the expression of the Gtl2, Rtl1, Rian, and Mirg genes was suppressed (Supporting Information Fig. S7B). We further investigated the ploidy of each cell because it is well known that the number of chromosomes is a good indicator of ES cell quality and the ability to achieve germline transmission. However, no differences were observed between 3F-iPSCs and 4F-iPSCs (84.1% ± 14.7% and 81.2% ± 5.3% 40 chromosome cells among 2A-4F-iPSCs and 2A-3F-iPSCs, respectively; Supporting Information Fig. S8).

Finally, we examined the effects of 5-azacytidine, TSA, and “2i medium” on the developmental potential of 3F-iPSCs [18, 20]. We cultured 2A-3F-iPSCs in these media and then performed aggregation experiments. iPSCs died after 24 hours in the 5-azacytidine cultures (Fig. 4A). Importantly, a significant improvement in these cells was observed following TSA treatment. Highly chimeric mice were successfully born, black-eyed mice did not die when compared with untreated 3F-iPSCs, and germline transmission was also confirmed (Fig. 4B, 4C). Furthermore, western blotting analysis demonstrated a low level of histone acetylation in 3F-iPSCs (Fig. 4D). Although little improvement in developmental ability was observed for the 2i medium treatment and the rapid death of embryos ensued, improvements were observed in the implantation frequency and chimeric mice generation using these cells, particularly after 8 days in culture, albeit with a lower chimerism (Figs. 4B, 4C).

Figure 4. Effects of 2i, 5-azacytidine, and trichostatin A (TSA) on the developmental potential of 3F-induced pluripotent stem cells (iPSCs). (A): Phase-contrast micrograph of iPSC colonies. Scale bar = 500 μm. (B): Effects of 2i medium and TSA on 3F-iPSCs. The results of chimeric mice generation tests are shown. (C): Survival curve of the chimeric mice. (D): Western blotting analysis. The graph shows relative acetylated lysine band-intensity normalized to the amount of histone H4 band-intensity. Abbreviations: 5-Aza, 5-azacytidine; TSA, trichostatin A.

DISCUSSION

We reveal in this study that c-Myc transduction is required for preparing high-quality iPSCs that have full developmental ability from MEFs. Thus far, different observations that were previously considered controversial have been reported for c-Myc function in pluripotent stem cells. For example, it was reported previously that c-Myc transduction was dispensable for iPSC generation [3, 4] but, on the other hand, crucial roles of c-Myc in ES cells were demonstrated from transcriptome analysis and studies of c-Myc knockout mice [7, 8, 11]. However, our current results show that these contrasting findings have an underlying explanation.

It was not until we performed an aggregation test that any difference between 3F-iPSCs and 4F-iPSCs was observed. Hence, it required a developmental engineering approach to reveal that these cells are in fact distinct. We additionally reveal from our present data that TSA, an HDAC inhibitor, can compensate for the absence of c-Myc transduction during iPSC generation, strongly suggesting an involvement of this oncogene in the gene expression regulation mechanism via histone acetylation [27]. Indeed, western blotting further showed lower levels of histone H4 acetylation in 3F-iPSCs compared with 4F-iPSCs. Our present findings are thus consistent with the previous finding that c-Myc functions with the HAT complex in ES cells [28]. A large number of genes including micro-RNAs may therefore be governed by c-Myc through histone acetylation modifications during iPSC generation [9]. We next addressed whether events that occur downstream of histone acetylation modifications such as gene expression and DNA methylation profiles differed between 3F-iPSCs and 4F-iPSCs. The question of whether such differences occur at specific gene loci or nonspecific loci was also an important consideration. Gene expression profiling of 3F-iPSCs and 4F-iPSCs was undertaken as described previously [29, 30] and several differentially expressed genes were detectable. These included Gtl2 and Rtl1, both imprinting genes [25, 26] (Supporting Information Fig. S9). However, few gene expression changes were observed in other 3F-iPSC clones (Supporting Information Fig. S7B). Although these data suggest that abnormal expression occurs at nonspecific gene loci in 3F-iPSCs, further and careful investigation of the entire transcriptome of these stem cells will be required to verify this.

It is noteworthy that genome integration-free 4F-iPSCs generated using a plasmid vector possess full developmental potential, equivalent to 4F-iPSCs generated with retrovirus vectors. This indicated that approximately 1 week of forced expression from the point of gene introduction is sufficient for the generation of high-quality iPSCs and that no constitutive expression of c-Myc is necessary. Indeed, in our present experiments, 2A plasmids containing Sox2, Klf4, and Oct3/4 genes and the c-Myc plasmid were mixed, then transfected four times at 2 day intervals, resulting in about 8 days of forced c-Myc expression. iPSC clones subsequently appeared at approximately day 16. Hence, an abundant level of c-Myc at a specific time is crucial for iPSC generation and causes a drastic change in the subsequent developmental potential of these stem cells.

It is also of great interest that similar to the 2A plasmid-generated iPSCs, positive effects of c-Myc transduction were also observed for iPSCs generated by retrovirus gene transduction (R-iPSCs). However, R-iPSCs seemed to be marginally but definitively different from 2A-iPSCs. Indeed, we successfully obtained one iPSC line from an R-3F transduction (clones 5–7) that was indistinguishable from 4F-iPSCs and normal ES cells, even in chimeric mice generation experiments. Highly chimeric mice were efficiently obtained from this R-3F-iPSC clone and germline transmission was also confirmed (Supporting Information Fig. S4A) [17]. Similar results was also found in an additional mouse strain, whereby one out of the four 3F cell lines, 3F-9, exhibited a highly pluripotent state equivalent to 4F and ES cells, and also clones 5–7 (Supporting Information Fig. S4B). Moreover, unlike 2A-iPSCs, chromosome analysis could be used to discriminate R-3F-iPSCs and R-4F-iPSCs as the ratio of the cells possessing a normal number of chromosomes in 4F-iPSCs was clearly higher than that in 3F-iPSCs (Supporting Information Fig. S8). In addition, the extent of the abnormalities among R-iPSCs was higher than that among 2A-iPSCs. As expected, the iPSC clone which exhibited the highest level of normalcy by chromosome analysis among the 3F-iPSCs generated was the 5–7 clones (86%), the only R-3F-iPSC clone to show similar properties to 4F-iPSCs and normal ES cells. The differences between R-3F-iPSCs and 2A-3F-iPSCs may be due to differing levels of Oct3/4, Sox2, and Klf4 transcripts. In addition, the retrovirus system may induce intracellular responses more readily than the plasmid system.

Our current aggregation test results indicate differences between 3F-iPSCs and 4F-iPSCs in terms of their invasion capability into blastocysts. However, an evaluation of whether a close relationship exists between the ratio of the remaining GFP-positive cells in the aggregated blastocysts and the chimerism and germline transmission capability of the mice developed from the embryos investigated in vitro will be needed to demonstrate the usefulness of our in vitro assay.

Recently, considerable variations among iPSC clones have been described, which is an extremely critical issue when considering any future medical use of human iPSCs. Because no in vivo evaluation is possible in humans, generation of iPSCs using c-Myc and/or TSA would be quite useful and minimize variations among iPSCs. In addition, our current results also require careful evaluation using in vivo assays such as the aggregation test for other iPSCs that have been established through various procedures, particularly through the use of small-molecule compounds.

CONCLUSION

It has been postulated previously that c-Myc transfection is not essential for iPSC generation and in fact has negative effects, including oncogenic transformation. Our current results from a chimeric mice generation test involving a total of 22 iPSC lines consisting of seven 4F-iPSC lines and 15 3F-iPSC lines, all of which are genome integration-free and had been generated using an inbred mouse strain C57BL/6, reveal, however, that c-Myc has crucial positive effects upon iPSC generation, particularly in terms of achieving mature pluripotent stem cells. Furthermore, we demonstrate that TSA, an HDAC inhibitor, improves the defects in 3F-iPSCs. Hence, our results indicate the important involvement of c-Myc in histone acetylation regulation during iPSC generation. Based on our current findings, we strongly recommend the use of c-Myc and/or TSA for preparing high-quality iPSCs.

Acknowledgements

We thank Shinya Yamanaka (University of Kyoto, Kyoto, Japan) for providing the 2A plasmid and retrovirus vectors through Addgene and Nanog-GFP tg mice through RIKEN BioResource Center; Toshio Kitamura (University of Tokyo, Tokyo, Japan) for providing PlatE cells and the pMXs retrovirus vector; Andras Nagy for providing R1 ES cells; and Akira Nifuji and Hisashi Ideno for helpful discussions and Hana Yoshida, Kinuko Nishikawa, Akemi Ishibashi, Yuko Noda, Ryoko Watanabe, Koji Douzono, and Matsuko Nakamura for technical assistance. This work was supported partially by a research grant from Precursory Research for Embryonic Science and Technology (PRESTO), Japan Science and Technology Agency.

DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

The authors indicate no potential conflicts of interest.