Author contributions: L.L.: conception and design, financial support, final approval of manuscript; Z.C.: collection and assembly of data, data analysis and interpretation, manuscript writing; J.H.: collection and assembly of data, data analysis and interpretation; Z.L., C.L., B.L., C.W., S.C., T.A.: collection and assembly of data; L.Z.: provision of study material; M.G.C.: collection and assembly of data, data analysis and interpretation; D.L.K., X.Y., data analysis and interpretation, financial support, final approval of manuscript.
First published online in STEM CELLS EXPRESS June 19, 2009.
Mammalian parthenogenetic embryos are not viable and die because of defects in placental development and genomic imprinting. Parthenogenetic ESCs (pESCs) derived from parthenogenetic embryos might advance regenerative medicine by avoiding immuno-rejection. However, previous reports suggest that pESCs may fail to differentiate and contribute to some organs in chimeras, including muscle and pancreas, and it remains unclear whether pESCs themselves can form all tissue types in the body. We found that derivation of pESCs is more efficient than of ESCs derived from fertilized embryos, in association with reduced mitogen-activated protein kinase signaling in parthenogenetic embryos and their inner cell mass outgrowth. Furthermore, in vitro culture modifies the expression of imprinted genes in pESCs, and these cells, being functionally indistinguishable from fertilized embryo-derived ESCs, can contribute to all organs in chimeras. Even more surprisingly, our study shows that live parthenote pups were produced from pESCs through tetraploid embryo complementation, which contributes to placenta development. This is the first demonstration that pESCs are capable of full-term development and can differentiate into all cell types and functional organs in the body. STEM CELLS 2009;27:2136–2145
Parthenogenetically derived mammalian embryos, with no paternal genome, are not viable and undergo spontaneous resorption at midgestation, largely from defective placental growth [1, 2] attributed to a lack of bi-parental imprinting [1, 3, 4]. Parthenogenetic mouse pups have been produced from reconstructed parthenogenetic embryos after genetic modification of the imprinted loci Igf2 and H19 . Parthenogenetic ESCs (pESCs) derived from parthenogenetic embryos are capable of contributing to chimeras , although it remains unclear whether pESCs themselves can form all tissue types in the body, and the current hypothesis is that parthenogenetic embryos cannot develop to term without genetic modification . Previously, pESCs derived from oocytes activated with 6-7% ethanol, which only triggers a single rise in Ca2+, showed limited developmental potential in mice [6–8]. In contrast, multiple Ca2+ elevations induced by strontium (Sr2+) treatment fully activate oocytes, similar to fertilization [9, 10], and significantly enhance parthenogenetic embryo development [11, 12]. Derivation of pESCs from fully activated oocytes by Sr2+ may increase the pluripotency of pESCs. Sr2+-activated cytoplasts were used for somatic cell nuclear transfer to successfully produce cloned mice  and, when using pESCs for nuclear transfer to regenerate pESCs (NT-enhanced pESC or NT-pESC), contribution of pESC in the chimeras was considerably improved . We generated new pESC lines, which are shown indistinguishable from ESCs derived from fertilized embryos (fESCs), and these pESCs can contribute to all organs in chimeras. Surprisingly, through tetraploid embryos complementation (TEC) , these hybrid pESCs can develop to full term without gene modification, showing that pESCs are capable of term development and can differentiate into all cell types and functional organs in the body. Our report provides scientific support for the possibility of female patient-specific pESC-based therapeutic applications, because demonstration of pESC pluripotency is critical to the feasibility of pESC therapy in regenerative medicine [16, 17]. pESCs are a potential source of histocompatible tissues for transplantation , which will not induce the immuno-rejection that may be generated by bi-parental genes in fESCs.
MATERIALS AND METHODS
Production of Parthenogenetic and Fertilized Embryos
B6C3F1 female mice at 6-10 weeks of age were superovulated with 5 IU pregnant mare serum gonadotrophin (Calbiochem, San Diego, http://www.emdbiosciences.com), followed by 5 IU human chorionic gonadotropin (hCG) 46-48 hours later. Oocytes enclosed in cumulus masses were collected from oviduct ampullae 14 hours after hCG injection. Cumulus cells were removed by pipetting, after a brief incubation in 0.03% hyaluronidase prepared in potassium simplex optimized medium (KSOMAA) containing 14 mM HEPES and 4 mM sodium bicarbonate (HKSOM). Oocytes were reliably activated by a 4-hour treatment with 10 mM SrCl2 and cytochalasin D prepared in Ca2+-free KSOM [19, 20]. Activated oocytes with two pronuclei were defined as diploid parthenotes and cultured in 50-μl droplets of pre-equilibrated KSOM, covered with mineral oil at 37°C in a humidified atmosphere of 6% CO2 in air for 4-5 days. To obtain fertilized F1 embryos, C57Bl/6 females were superovulated and mated individually with C3H/He males of proven fertility. Oviducts of successfully mated females were flushed 3.5 days after mating with HKSOM using a 30-gauge needle, and blastocysts were obtained to derive normal fESC lines.
Establishment and In Vitro Culture of ESC Lines
Blastocysts were transferred to mitomycin C-treated murine embryonic fibroblast feeder (MEF) cell layers and cultured for 4-5 days in stem cell medium: 80% Knockout Dulbecco's modified Eagle's medium (Gibco, Grand Island, NY, http://www.invitrogen.com) supplemented with 20% fetal bovine serum (FBS), 50 units/ml penicillin, 50 μg/ml streptomycin, 1 mM glutamine, 1% nonessential amino acid stock, 0.1 mM β-mercaptoethanol, and 1,000 units/ml ESGRO leukemia inhibitory factor (LIF; Chemicon International, Temecula, CA, http://www.chemicon.com). Inner cell mass (ICM)-derived outgrowths were mechanically dissociated into clumps with 32-gauge needles and replated on mitomycin C-treated feeder layers in fresh medium, designated as P1. Mechanical dissociation was used for passaging until typical ES-like cells appeared on the feeder layer (P1–P3). Afterward, propagation of colonies was accomplished by exposure to 0.25% trypsin-EDTA. Fertilized B6C3F1 blastocysts were transferred to feeder cell layers and cultured in ES-medium supplemented with 50 μM mitogen-activated protein kinase (MAPK) inhibitor PD98059 (Cell Signaling Technology, Beverly, MA, http://www.chemicon.com), which was replaced by standard ES medium when typical ES-like cells appeared. Otherwise, the method used to derive B6C3F1 fESCs was basically the same as for pESCs.
Recipient blastocysts (3.5 day post coitum [d.p.c.]) were collected from uteri of superovulated females. The collected blastocysts were incubated in KSOMAA at 37°C in a humidified atmosphere of 6% CO2 in air. Approximately 10-20 pESCs or fESCs were injected into the blastocoels of blastocysts obtained from ICR mice using a Piezo micromanipulator (PrimeTech, Ibaraki, Japan, http://www.primetech.org). The blastocysts were transferred into the uteri of 2.5 d.p.c. pseudo pregnant ICR mice 1-4 hours after the ESC injection. Chimeras were initially identified by coat color. The contribution of pESC in chimeras was determined by standard DNA microsatellite analysis, as described below.
Production of Tetraploid Embryos for pESC/fESC Complementation
Tetraploid embryos were produced by electrofusion of blastomeres at the two-cell stage. Two-cell embryos were induced to fuse by two pulses of 1,200 V/cm DC for 50 μs, after an AC pulse of 750 V/cm for 5 s, generated from an Eppendorf Multiporator. About 98% efficiency of fusion was routinely achieved 1 hour later. Fused embryos were cultured in KSOMAA to form blastocysts. pESCs or fESCs were injected into tetraploid blastocysts as done for diploid blastocysts. Injected blastocysts were transferred into the uteri of 2.5 d.p.c. pseudopregnant Kunming (KM) mice.
Metaphase chromosomes were prepared by exposing cultured cells to nocodazole (0.4 μg/ml) for 2 hours or colcemid (0.1 μg/ml) for 1 hour, followed by hypotonic treatment with 75 mM KCl solution, fixation with methanol:glacial acetic acid (3:1) and spreading onto slides. The spreads were treated with 0.005% trypsin, stained with Giemsa and 20-30 separate metaphase spreads were examined for each culture.
DNA Microsatellite Polymorphism Analysis
Polymerase chain reaction (PCR)-based haplotype analysis was used to determine homo- or heterozygosity at specific loci in established pESC and fESC lines, as well as in tissues from chimeric mice and pups produced by TEC. Microsatellite markers D14mit5, D16mit4, and D18mit17 were used to confirm homozygosity in hybrid pESC lines C2 and C3. D1Mit46, D6Mit9, D14Mit5, and D16Mit4 were used to assess pESC contribution in pESC chimeras. pESC contribution in tetraploid embryo complemented pup T1 was measured using D8Mit4, D12Mit136, D16Mit4, and DXMit1. Briefly, genomic DNA was extracted from tissues collected from chimeras, recipient tissues, and donor ESCs. By screening >36 microsatellite markers from Mouse Genome Informatics Website (The Jackson Laboratory, Bar Harbor, ME, http://www.informatics.jax.org), some that are located on different chromosomes were identified to be polymorphic in our experiments, and therefore used for genotyping the chimeras. Microsatellite analysis was performed by PCR amplification with primers designed from conserved sequences flanking each marker and subsequent electrophoresis using 15% polyacrylamide gels. The gels were silver-stained and scanned, and the genotype was determined. Ratios of pESC to fESC contribution in chimeras were estimated by quantification of the density of the bands after subtraction of the background noise using Bio-Rad Quantity One software (Bio-Rad, Hercules, CA, http://www.bio-rad.com).
Teratoma Formation Assay
Approximately 2 × 106 ESCs were injected subcutaneously into 4-week-old immunodeficient nude mice to test for teratoma formation. Four weeks after injection, the mice were sacrificed and the resulting teratomas were excised, fixed in 4% paraformaldehyde, embedded in paraffin, sectioned, and stained with hematoxylin-eosin for histological examination.
Alkaline Phosphatase Staining and Immunofluorescence Microscopy
Alkaline phosphatase staining was performed using the Vector blue kit from Vector Laboratories (DAKO, Glostrup, Denmark, http://www.dako.com). For immunocytochemistry, cells were cultured in tissue culture grade 4-well plates for 2 days. The media were removed, the cells washed twice in phosphate-buffered saline (PBS) and fixed in freshly prepared 4% paraformaldehyde in PBS for 15 minutes at 4°C. Cells were rinsed with PBS and blocking buffer (3% goat serum in PBS) and permeabilized in 0.1% Triton X-100 in blocking buffer, washed three times, and left in blocking solution for 1 hour at room temperature. Cells were incubated overnight at 4°C with mouse monoclonal antibody OCT3/4, ERK2, pERK (Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com) or SSEA-1 (DSHB, The University of Iowa), diluted 1:100 in blocking solution, washed, incubated for 1 hour with Alexa Flur 568 (or 488)-conjugated goat anti-mouse IgG or IgM (Molecular Probes), diluted 1:100 in blocking solution, washed, and counterstained with 0.2 μg/ml 4′,6-diamidino-2-phenylindole or Hoechst 33342 in Vectashield mounting medium. Primary antibodies were omitted from control samples. Fluorescent images were taken under a Leica microscope using appropriate filters (Leica, Germany, http://www.leica-microsystems.com).
Cultured cells were washed twice with ice-cold PBS and lysed on ice for 30 minutes in radioimmunoprecipitation buffer containing protease inhibitors. The lysates were centrifuged at 20,000g for 30 minutes at 4°C, and the supernatant was collected. The protein content was determined by Bradford protein assay. The lysates (20 ng protein) were separated by SDS-PAGE (10%) and transferred to polyvinyldene fluoride membrane (Millipore, Bilerica, MA, http://www.millipore.com), followed by incubation in PBS containing 0.05% Tween-20 and 2% block powder (Amersham Bioscience, Piscataway, NJ, http://www.gelifesciences.com) overnight at 4°C. The blots were incubated with mouse monoclonal antibody OCT3/4, pERK1, pERK2 (1:200), rabbit polyclonal TERT (1:200; Santa Cruz Biotechnology), or mouse monoclonal antibody β-actin (1:200) for 2 hours at room temperature followed by incubation with horseradish peroxidase-conjugated secondary antibodies (1:10,000 goat-anti-mouse, goat-anti-rabbit; Pierce, Rockford, IL, http://www.piercenet.com) for 45 minutes. The bands were detected by chemiluminescence signals using ECL Advance Western Blotting Detection Kit (Amersham) and scanned to produce digital images.
Total RNA was isolated with TRIzol (Invitrogen, Shanghai, CHN, http://www.bioasia.cn/) and treated with DNase I (Invitrogen). After quantitative measurement of the total RNA extracted from each sample using a spectrophotometer (Eppendorf, Westbury, NY, http://www.eppendorf.com), the samples (1 μg RNA) were subjected to cDNA synthesis using ReverTra Ace Kit (Toyobo, Osaka, Japan, http://www.toyobo.co.jp/e/). Total RNA and PCR products were separated on a 2% agarose gel, stained with ethidium bromide, visualized, and photographed on a UV transluminator.
For real-time quantitative PCR, primers were designed using GeneTool software (supporting information Table S4). Real-time PCR was performed using F1P9 as controls, and each sample was analyzed in triplicate with β-actin as the internal control. The final PCR reaction volume of 20 μl contained 10 μl SYBR Green PCR Master Mix (Real-time PCR Master Mix; Toyobo), 1 μl cDNA template, 1 μl primer mixture, and 8 μl water. Thermal cycling was carried out with a 5-minute denaturation step at 94°C, followed by 40 three-step cycles: 30 seconds at 94°C, 30 seconds at 60°C, and 30 seconds at 72°C. Amplification data were collected by the ABI PRISM 7900 and analyzed by the Sequence Detection System 2.0 software (Applied BioSystems, Foster City, CA, http://www.appliedbiosystems.com).
pESCs Have Characteristics Similar to fESCs
In our first series of experiments, we generated 17 pESC lines at various passages from B6C3F1 parthenogenetic blastocysts (supporting information Fig. 1A; supporting information Table S1), using methods described in our previous study [19, 20]. We systematically characterized two pESC lines (C2 and C3) and tested their pluripotency by teratoma formation (supporting information Fig. 1B) and chimera production (supporting information Fig. 1C) and by generation of pESC pups by TEC (supporting information Fig. 1D).
The two hybrid F1 pESC lines (C2 and C3) proliferated in culture for more than 40 passages and maintained morphology typical of ESCs (Fig. 1A–1C). pESCs formed round colonies with well-defined boundaries, a pattern similar to fESCs (Fig. 1E, 1F). These cells were comparatively small, typically had large, clear, shiny nuclei containing one or more prominent nucleoli, and were tightly packed within a multilayered, primary colony (Fig. 1B). These pESC lines maintained normal karyotypes at passages 10 and 50, and G-banding results showed that pESCs at early passages exhibited the expected XX karyotype (Fig. 1D). Furthermore, these cell lines were positive for alkaline phosphatase (AP), SSEA-1, and Oct4 at early (p9) and late passages (p30), comparable to fESCs (Fig. 1G). Oct4 and TERT protein were present in early and late passages of pESCs like fESCs, but absent from MEFs served as negative controls (Fig. 1H). Real-time PCR analysis also showed that mRNA levels of the stem cell markers Oct4 and Nanog  did not differ between pESCs and fESCs (data not shown). These data show that expression patterns of stem cell markers were comparable between pESCs and fESCs.
Our B6C3F1 hybrid pESCs were derived from parthenogenetically activated MII oocytes by Sr2+, supplemented with cytochalasin D, which causes retention of the second polar body. Genotype analysis with markers for D14mit5, D16mit4, and D18mit17 showed homozygosity in both C2 and C3 pESC lines. Cell line C3 is homozygous at D16mit4 and D18mit17 for alleles found in C57Bl/6 mice and at D14mit5 for a C3H/He allele (Fig. 1I). Cell line C2 is homozygous at D16mit4 for an allele in C3H/He mice, homozygous at D18mit17 for a C57Bl/6 allele, but heterozygous at D14mit5, which is not surprising as shown in a recent report that human pESC lines also exhibit heterozygous genotypes by genome-wide single nucleotide polymorphism (SNP) analysis .
Reduced MAPK Signaling Is Associated with Efficient Derivation of pESCs from Parthenogenetic Blastocysts
Maintaining appropriate ICM outgrowth of mouse blastocysts placed on the MEF feeder in ES medium is the first critical step for successful derivation of ESC lines. ICM outgrowth of parthenogenetic blastocysts grew slower than did fertilized blastocysts. ICM outgrowth formed and appeared large from fertilized blastocysts than from parthenogenetic blastocysts (supporting information Fig. 2). However, it seemed that the growth rate of ICM did not associate with isolation efficiency of ESCs. Initially, no fESC lines were derived from fertilized B6C3F1 mouse blastocysts in the ES medium supplemented with FBS, in contrast to an efficiency of 10% for the derivation of pESC lines of the same mouse strain (supporting information Table S1). Notably, fESC lines were generated with efficiency of up to 29%, when MAPK inhibitor PD98059 was added into the medium during ICM outgrowth. Therefore, we speculated that MAPK signaling might be reduced in parthenogenetic blastocysts compared with fertilized blastocysts. Indeed, expression of pERK protein was significantly lower in parthenogenetic than fertilized embryos (Fig. 2A). ICM outgrowths of parthenogenetic blastocysts exhibited noticeably reduced fluorescence intensity of ERK2 and pERK, indicative of decreased MAPK signaling, compared to those of fertilized blastocysts (Fig. 2B). Inhibition of MAPK signaling by PD98059 further reduced expression of ERK2 and pERK and size of ICM outgrowth from both fertilized and parthenogenetic embryos (Fig. 2B; supporting information Fig. 2). These results show that MAPK pathway activation was reduced in parthenogenetic blastocysts and MAPK signaling remained low during ICM outgrowth for derivation of pESCs. Reduced MAPK signaling in parthenogenetic embryos and ICM outgrowth is associated with increased efficiency in the derivation of pESC lines. Moreover, inhibition of MAPK by PD98059 in ICM outgrowth of fertilized blastocysts also increased efficiency in derivation of fESC lines. Thus, reduced MAPK pathway found in parthenogenetic embryos also plays critical role in the efficient derivation of pESC lines, as in the derivation of fESCs.
Differentiation Capacities In Vivo Are Indistinguishable Between pESCs and fESCs
pESCs have been shown to retain an extensive differentiation capability in vitro [16, 23]. To assess the in vivo differentiation potential of pESC lines, we injected C3 pESC cells at early, middle, and late passages into the hypodermic cavity of immunodeficient nude mice. After injection of C3 cells at early passages [5–10] under the skin of nude mice (n = 4), all nude mice produced teratomas, weighing 0.8-1.9 g, comparable to teratomas from fESCs (Fig. 3A, 3B). By contrast, injection of MEF into nude mice (n = 5) did not produce teratomas (Fig. 3C). Four weeks after injection, teratomas were sectioned and analyzed by histology, showing mature tissues and low mitotic frequencies, indicating their benign nature. Furthermore, the teratomas contained derivatives of all three germ layers, including cartilage, neurons, and skin (ectoderm), intestinal epithelia (endoderm), and muscle (mesoderm; Fig. 3D–3I). Injection of pESCs at passages 22 and 30 also led to teratoma formation in nude mice (n = 2).
To further assess the pluripotency of pESC in vivo, we injected C3 pESC cells carrying the agouti coat color marker into the blastocoels of albino ICR diploid blastocysts and transferred them into the uteri of pseudopregnant surrogate mice. We generated 18 chimeras from these pESCs, including 10 chimeras from early passages (passages 6-9) and 8 from a late passage (passage 30; Table 1). These chimeras developed well postnatally, with no growth retardation. The contributions of pESCs in these chimeras varied from as little as 5% to more than 70%, as judged by coat color (Fig. 3J, 3K). We showed participation of pESCs in various tissues of chimeras at ages of 3 weeks to 9 months, estimated by density measurement of bands resolved by microsatellite gel electrophoresis (Table 2; Fig. 3L–3N). pESC generally did not exhibit tissue-specific distribution, except for that pESCs contributed to pancreas at a higher ratio than to other tissues. Distribution of pESCs in the chimeric tissues correlated with the extent of coat color. Remarkably, in our study, the contribution of pESCs to skeletal muscle (up to 31% vs. 34% for fESCs) and gonads (up to 39% vs. 40% for fESCs) did not differ from other tissues (Table 2; Fig. 3L, 3M). This is in contrast to a previous report showing limited pESC contributions to pancreas, muscle, and gonads in pESC chimeras . Repressed expression of Igf2 was associated with defects in contributions by pESCs to muscle . Moreover, the distribution and extent of pESC contribution to chimeras produced in our study were markedly greater than those previously reported for chimeras produced with either parthenogenetic embryos or pESCs [6, 14, 24]. Notably, a high contribution of pESCs was found in gonads of chimeras, like chimeras produced from fESC. Although some chimeras produced from pESC C3 lines were bitten to death by surrogate mothers and the survived pups failed to produce agouti pups after limited breeding, other pESC lines did show a high capacity through germline in the resultant chimeras after extensive breeding, as reported in details elsewhere (unpublished data). Our results also showed that the age of chimeras did not affect the proportions of tissues contributed by pESCs. However, pESCs at later passages (P30) resulted in lower frequencies of chimera production (Table 1) and decreased contributions to chimeras compared with early passage pESCs (Table 2; Fig. 3N), implying that, just as with fESC lines, in vitro culture conditions and long-term passaging may alter the differentiation potential of pESCs in vivo.
Table 1. Production of pESC chimeras identified by coat colora
Table 2. Contribution by ratio of pESCs in the chimera tissues analyzed by microsatellites
pESCs Can Produce Term Pups by Tetraploid Embryo Complementation
To determine whether pESCs have the capacity to differentiate into all tissue types in the body, we took advantage of the TEC method, which was considered the most stringent test of pluripotency and genomic stability of ESCs and has been used previously to test term development of fESCs, ntESCs, and induced pluripotent stem cells [25–27]. In a series of experiments, 10-20 agouti pESCs were injected into albino tetraploid blastocysts (also see research plan in supporting information Fig. 1). Surprisingly, one B6C3F1 pESC-derived pup (designated as T1) was obtained by natural delivery after transfer of 420 tetraploid embryos injected with C3 pESCs into 26 pseudo-pregnant recipients (Fig. 4A). The newborn T1 was alive, appeared anatomically normal, and had normal birth weight, similar to other newborns (Fig. 4B), although T1 died shortly after birth. We dissected uteri of nine recipient mice that had received 161 C3 pESC-tetraploid embryos and found 23 implantation sites (14.3%). No pups were produced from the transfer of 148 C2 pESC-tetraploid embryos, although implantation did occur. In our B6C3F1 fESCs by TEC as controls, only one fES (F1)-derived pup was obtained by cesarean from the transfer of 247 tetraploid embryos injected with fESCs into 15 pseudo-pregnant recipients. However, when using 129B6F1 hybrid fESCs for tetraploid complementation, we obtained five pups at a much higher efficiency (14%, n = 37; Table 3). The differential efficiency in obtaining pups from fESCs is likely caused by strain differences of donor ESCs and recipient embryos, as shown previously in the literature [25, 28]. Together, these data suggest that our pESC lines are comparable to fESCs for B6C3F1 strain in terms of TEC efficiency.
Table 3. Production of pESC and fESC pups by tetraploid blastocyst complementationa
Microsatellite analysis at D12mit136 and Dxmit1 showed that the alleles of various tissues from the pup T1 were identical to those of C3 pESC lines but completely different from the tetraploid blastocysts KM17 and KM18 (Fig. 4C). The identity of the fESCs and pESC-derived pups confirmed by DNA microsatellite profiling is shown in supporting information Tables S2 and S3, respectively. We indicated by microsatellite profiling that the tetraploid embryos made no contribution to the tissues tested from the derived pup (Fig. 4C–4E). The origin of T1 from pESCs was further confirmed by SNP analysis (supporting information Fig. 3), although we could not exclude the possibility of contribution of tetraploid embryos in the pES fetuses, because of limitations of the assays uses. The pattern of C3P7 pESC differs from that of B6C3F1, suggesting that pESCs had undergone some loss of heterozygosity in culture. Together, this is the first demonstration that pESCs can support full-term development, resulting in a pESC-derived newborn, similar to those reported previously for fESCs .
To confirm our surprising observation of term development from the newly established pESC line in this study, we independently performed experiments using a pESC line constitutively expressing enhanced green fluorescent protein (EGFP) complemented by TEC in the laboratory of X.Y. Thus far, three live pups with GFP expression have been produced from pESCs, albeit also with low efficiency (<1%), but these pups also died shortly after birth. Preliminary data by microsatellite analysis also showed that these pups were derived from the pESCs. Further analysis of the tissues and organs of the pups is expected to be reported shortly.
Epigenetic Reprogramming of Imprinted Genes During Derivation of pESCs by In Vitro Culture
Genomic imprinting plays critical roles for embryonic and fetal development and may have been changed during isolation and in vitro culture of pESCs. Our group (data not shown) and others have found that parthenogenetic hybrid embryos complemented with tetraploid embryos die abruptly around 13-14 days of fetal development . Abnormal imprinting likely contributes to the limited development of parthenogenetic embryos and the restricted potential of pESCs [1, 3, 4, 6, 8, 30, 31]. To understand expression and methylation of imprinted genes in our newly derived pESCs reported in this study, we initially analyzed expression of imprinted genes in blastocysts using quantitative real-time PCR. Not unexpectedly, maternally expressed imprinted genes H19 and Igf2r exhibited remarkably higher expression (p < .01) in parthenogenetic blastocysts, whereas paternally expressed imprinted gene Snrpn was expressed at lower levels than in fertilized blastocysts (p < .01). Expression TSSC3 did not differ between parthenogenetic and fertilized blastocysts (Fig. 4F). To understand whether expression of imprinted genes in the resultant ESCs differ from their progenitor blastocysts, we compared expression of the imprinted genes in pESCs with fESCs. Notably, expression of Igf2r, Tssc3, and Snrpn did not differ (p > .05) between pESCs and fESCs (Fig. 4G). H19 showed sixfold increase in parthenogenetic blastocysts, but only about twofold increase in the derived pESCs, suggesting reprogramming of H19 to certain degrees. Consistent with expression data, pESC lines showed a hypomethylated status of imprinted genes (Igf2/H19; Fig. 4H), which is most consistent with the parental origin of the hypomethylated oocytes that they were derived from. Further experiments showed that pESCs and pESC fetuses complemented by TEC exhibited balanced methylation of Snrpn, Peg1, and U2af1-rs1, and global methylation increased in parthenogenetic embryos but decreased in pESCs . These data suggest that culture in vitro modifies the expression of imprinted genes in pESC after isolation from parthenogenetic blastocysts. Our data are consistent with previous observations that in vitro culture itself can alter the expression of imprinted genes in both fESCs and pESCs [33, 34]. Appropriate epigenetic alterations may facilitate parthenogenetic embryonic development from pESCs.
In this study, we successfully established hybrid pESCs that are indistinguishable from fESCs. These pESCs express molecular markers typical of fESCs in vitro, differentiate fully in vivo, and exhibit unrestricted potential to generate various tissues in vivo, as evidenced by teratoma formation and chimera production, as well as by the surprising births of parthenote pups from these pESCs by TEC. By microsatellite and SNP analysis, the pups were in fact derived from pESCs, and no detectable contribution, if any, could be found from tetraploid host blastocyst cells to the tissues or organs of the pups. Any contribution from tetraploid host blastocysts would have to be below the detection limits of the assays.
Based on current knowledge of the role of and requirement for proper genetic imprinting in development, it is not entirely clear how or why pESC are capable of full-term development. What is clear is that these in vitro systems offer only a limited and partial view of the action of imprinted gene regulation and function in development, and these results open the door to future studies examining additional imprinting modifications that may occur at various stages of development complemented by TEC. It is also not clear why term development for pESC-TEC has such a low success rate, but in future studies, we plan to examine the effects of different genetic backgrounds on efficiency, as well as early embryonic contribution and support by tetraploid cells, followed by selection against them in embryonic tissues as development proceeds. Efficiency in producing ES pups by TEC generally is very low and depends on the genetic background of donor ESCs and recipient tetraploid embryos, as well as passage numbers of ESCs [35, 36]. We show that pESCs could directly produce pups by TEC albeit at very low frequency, like fESCs of the B6C3F1 strain. This could result from modification of genomic imprinting, particularly activation of paternally expressed imprinted genes by in vitro culture, such that expression of most imprinted genes did not differ between pESCs and fESCs . However, pESC pups died within 1 day, but fESC pups survived to adults, and this could be because of relative high homozygosity of pESCs in contrast to heterozyosity of fESC. Indeed, ES pups produced from inbred mouse ESCs often died shortly after birth because of respiratory deficiency, whereas ES pups produced from hybrid mouse ESCs survived and grew normally . Genetic manipulation of ICR of H19/Igf2 resulted in normal expression of the genes, leading to live birth of parthenogenetic pups [5, 30]; thus, incomplete epigenetic changes in the ICR of H19/Igf2 found in pESCs may also negatively influence survival of pESC pups.
Mammalian parthenogenetic embryos are not viable for term development and die from defects in genomic imprinting [1, 2]. pESCs are isolated from the ICM of parthenogenetic blastocysts through an in vitro derivation and culture process and thus may escape factors that regulate their development in vivo, including imprinting. Genomic imprinting thus may have less influence on those pESCs derived from parthenogenetic embryos than the parthenogenetic embryos themselves. Szabo and Mann  showed that H19 is not expressed but Igf2r is expressed in normally fertilized preimplantation mouse embryos using RT-PCR, as did Hamatani et al.  using gene expression microarrays. Interestingly, our study showed that expression levels of Snrpn, Igf2r, and Tssc3 in C3 pESCs were comparable to fESCs controls and that H19 also show significantly reduced expression; these expression patterns differed from their progenitor blastocysts. This observation supports the notion that expression of imprinted genes in embryonic cells can be changed by in vitro culture [33, 34]. These epigenetic changes of pESCs in culture may partially explain why those pESCs by TEC can lead to full-term development. This novel finding is well supported by the report that parthenogenetic and bi-maternal mouse pups can be successfully produced from reconstructed parthenogenetic embryos after genetic modification of the key imprinted loci particularly Igf2 and H19 .
We further found that derivation of the pESC line is more efficient than that of fESCs. Favorable signaling involved in parthenogenesis could explain the efficient derivation of pESCs. Naturally decreased MAPK signaling in parthenogenetic embryos and ICM outgrowth is associated with derivation of pESCs at higher efficiency than that of fESCs from B6C3F1 mouse strain. This is consistent with the notion that inhibition of MAPK signaling promotes derivation of fESC lines from hybrid embryos [40, 41]. In contrast, the Grb2/Mek pathway represses Nanog, a key transcription factor that maintains self-renewal and pluripotency in mESCs [21, 42, 43]. Indeed, inhibition of MAPK pathway by PD98059 also increased derivation of fESCs from fertilized B6C3F1 mouse embryos. It is also important for the LIF-STAT3 pathway to maintain the self-renewal and pluripotency of mouse ESCs even though the MAPK pathway activated by LIF-STAT3 can induce ESC differentiation. Decreased MAPK signaling in parthenogenetic embryos reduces differentiation of ICM outgrowth, likely contributing to high efficient generation of pESCs.
Overall, newly established pESCs from activation of oocytes by strontium reported here are indistinguishable to fESCs for tissue/organ contribution. Reliable and efficient derivation of pluripotent pESCs is a critical step toward the feasibility of female patient-specific ESC therapy in regenerative medicine [16, 17]. Our findings that parthenogenetic embryo-derived ESCs can differentiate fully in vivo, exhibit unrestricted potential to generate various tissues in vivo, and lead to full-term development via tetraploid embryo complementation open the door for more epigenetic studies in stem cell biology and may show promising evidence for future therapeutic applications of an alternative stem cell resource for regenerative medicine, without involving fertilized embryos or somatic cell nuclear transfer cloning. More work is needed to understand the molecular dynamics and mechanisms underlying epigenetic changes of pESCs in vitro and to further explore the therapeutic potential of pESCs for basic developmental biology research and for regenerative medicine.
We thank Jingping Yun for histological examination and technical support and Peng Xiong for providing a cell line. This work was supported in part by National Natural Science Foundation, Science and Technology Division of Guangdong Province, and China Ministry of Science and Technology (2009CB941000) to L.L. and by USDA-ARS Contracts AG 58-1265-2-018 and 58-1265-2-020 to X.Y.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.