Equivalency of Nuclear Transfer-Derived Embryonic Stem Cells to Those Derived from Fertilized Mouse Blastocysts

Authors


Abstract

Therapeutic cloning, whereby nuclear transfer (NT) is used to generate embryonic stem cells (ESCs) from blastocysts, has been demonstrated successfully in mice and cattle. However, if NT-ESCs have abnormalities, such as those associated with the offspring produced by reproductive cloning, their scientific and medical utilities might prove limited. To evaluate the characteristics of NT-ESCs, we established more than 150 NT-ESC lines from adult somatic cells of several mouse strains. Here, we show that these NT-ESCs were able to differentiate into all functional embryonic tissues in vivo. Moreover, they were identical to blastocyst-derived ESCs in terms of their expression of pluripotency markers in the presence of tissue-dependent differentially DNA methylated regions, in DNA microarray profiles, and in high-coverage gene expression profiling. Importantly, the NT procedure did not cause irreversible damage to the nuclei. These similarities of NT-ESCs and ESCs indicate that murine therapeutic cloning by somatic cell NT can provide a reliable model for preclinical stem cell research.

Introduction

The first successful production of embryonic stem cell (ESC)-like lines from somatic cells via nuclear transfer (NT) was reported in a bovine model [1, 2] and was followed by production in mouse models [3, 4]. These ESC-like cell lines are believed to possess the same capacities for unlimited differentiation and self-renewal as those of conventional ESC lines derived from normal embryos produced by fertilization. To distinguish NT embryo-derived ESC lines from those derived from fertilized embryos, the former are referred to as NT-ESC lines [5]. We have previously shown that NT-ESC lines are capable of differentiating into all three germ layers in vitro or into spermatozoa and oocytes in chimeric mice [6]. This was the first demonstration that NT-ESCs have the same potential as ESCs derived from normally fertilized embryos. Moreover, cloned mice can be obtained from NT-ESC lines by a second NT procedure [6, [7]–8]. These NT-ESC lines not only can be used for basic research but may also have applications in tissue repair and transplantation medicine, a concept referred to as “therapeutic cloning” [9]. In theory, if applied in a clinical setting, these cells would carry the nuclear genome of the patient. Therefore, after directed cell differentiation, they could be used for autotransplantation to treat degenerative disorders such as diabetes, osteoarthritis, and Parkinson's disease, without immune rejection. Already, several reports have demonstrated that such NT-ESCs can be used in regenerative medicine to rescue immune-deficient or infertile phenotypes or even to achieve reproductive cloning in mouse models [7, 8, 10].

However, it remains unclear whether these somatically derived NT-ESCs are identical to the ESCs derived from early, normally fertilized embryos. Many NT-ESC lines differentiate into germ cells in chimeric mice [6, 7, 11], the strongest evidence to date that these cells have the same potential as ESCs. By contrast, most, if not all, cloned animals show serious fetal and postnatal abnormalities, perhaps resulting from defects in early genome activation [12], and many exhibit placentomegaly [13, 14]. Although imprinted gene expression patterns appeared normal in apparently “normal” cloned mice [15], they still exhibited epigenetic defects [16, [17]–18], often including obesity [19] and premature death [20]. DNA microarray analysis revealed that 4%–12% of expressed genes in the cloned placentas differed in expression levels from controls [21, 22]. However, those abnormalities seem to indicate epigenetic reprogramming errors, not DNA damage, because they were not transmitted to the next generation [19, 23]. Because NT-ESC lines are established using the same procedures, it is possible that they will also exhibit epigenetic defects.

Interestingly, NT-ESC lines can be established with success rates 10 times higher than reproductive cloning [6, [7]–8, 11, 13, 24]. We have previously reported that the success rate of the procedure depends on the mouse strain: hybrid strains are more tolerant of cloning than inbred strains, such as C57BL/6 and C3H/He, which are in common use in mouse genetic studies but which have never been cloned successfully [25]. However, NT-ESC lines can be established with high rates irrespective of donor mouse strain, tissue type, or sex [11]. Therapeutic cloning is thus at least an order of magnitude more successful than reproductive cloning [8, 26]. Thus, either most NT-ESC lines have been established from NT-embryos with negligible reproductive potential (Fig. 1A) or most NT-embryos die during or after implantation because of abnormal placental development, which does not involve ESCs. If these somatic NT-ESC lines have inherent abnormalities, such as epigenetic defects, there may be potential risks in their clinical use, and the scientific discoveries based on these NT-ESCs might be of limited utility.

Figure Figure 1..

Phenotype of NT-ESCs. (A): Comparisons between the success of cloning and establishment of NT-ESCs from cloned blastocysts. Establishment rates of NT-ESC lines are compared with the success rates for producing cloned mice by somatic cell NT. BDF1 cumulus cells were used as donors. Only 2% of somatic cell-derived cloned blastocysts could develop to term, but 16% could develop to NT-ESC lines, in which 14% (88% of established lines) were established from NT-embryos with little or no reproductive capacities. (B): Sequential establishments of NT-ESC lines by NT using NT-ESC nuclei. For the first series, BCF1 cumulus cell nuclei were used as NT donors to establish S1-NT-ESC lines. S2-NT-ESC lines were then established from S1-NT-ESC nuclei. The rate of establishment of NT-ESC lines did not decrease with later series, even though reproductive cloning fails after six generations. (C): Levels of SSEA-1, SSEA-4, and PDGFR-α expression in NT-ESCs (BDmt-1, BDfc-3) and ESCs, measured by flow cytometry. (a): BDmt-1 NT-ESCs were established from BDF1 male fibroblasts, and PDGFR-α is the characteristic marker for fibroblasts. However, BDmt-1 NT-ESC produced only SSEA-1 and did not produce PDGFR-α. (b): BDfc-3 NT-ESC lines expressed SSEA-1 but not SSEA-4, like E14 ESCs. (D): SKY-FISH staining of NT-ESCs showing trisomy of chromosome 11 (DBA/2 male NT-ESCs) and Y-chromosome deletion (B6D2F1 male NT-ESCs). Although most NT-ESC lines showed predominantly normal karyotypes, four cell lines showed only abnormal karyotype. Abbreviations: NT, nuclear transfer; NT-ESC, nuclear transfer-embryonic stem cell; PDGFR-α, platelet-derived growth factor-receptor-α; SKY-FISH, spectral karyotyping with fluorescent in situ hybridization.

We compared the molecular characteristics of NT-ESCs and ESCs in vitro and in vivo. Our immediate objective is to use these cells to study fundamental biology; ultimately, they may be used clinically if proven safe.

Materials and Methods

Animals

Transgenic mouse lines (B6D2F1 and 129B6 backgrounds) carrying the gene for the green fluorescent protein were used [27]. B6D2F1 oocytes were used as recipients for NT. In experiments to create chimeras, normally fertilized blastocysts of the BALB/c or ICR strains were used as recipients for NT-ESC injections. The surrogate mothers carrying chimeric embryos to term were ICR mice.

All animals (obtained from Japan SLC Inc., Shizuoka, Japan, http://www.jslc.co.jp) were maintained in accordance with the animal experiment handbook at the RIKEN Center for Developmental Biology, Kobe, Japan.

NT and Establishment of NT-ESC Lines

Oocytes were collected from 8- to 10-week-old female BDF1 mice. Cultured tail-tip fibroblasts [13], fresh isolated cumulus cells [24], or testicular Sertoli cells [28] were used as nuclear donors [10, 20, [21], [22]–23]. NT was performed as described [24]. Briefly, a group of oocytes (usually 30) was transferred to a droplet (approximately 10 μl) of HEPES-CZB containing 5 μg/ml cytochalasin B, which had been previously placed under mineral oil in the operation chamber on the microscope stage. After an oocyte was held by an oocyte-holding pipette, its zona pellucida was “drilled” by applying several piezo pulses to the tip of an enucleation pipette [24, 29]. The metaphase II chromosome-spindle complex, distinguished as a translucent spot in the ooplasm, was drawn into the pipette with a small amount of accompanying ooplasm and gently pulled away from the oocyte until a stretched cytoplasmic bridge was pinched off. After the spindles of oocytes in one group were removed (which took approximately 10 minutes), they were transferred into cytochalasin B-free CZB and kept there for up to 2 hours at 37.5°C [24]. The donor cell was drawn in and out of the injection pipette until the cell membrane was broken. In some cases, a few piezo pulses were applied to break the plasma membrane. After a cumulus cell nucleus was drawn deep into the pipette, another cell nucleus was drawn into the same pipette. Within a few minutes, several nuclei were lined up within a single pipette and were injected one by one into an “empty” oocyte at room temperature. After NT, the reconstructed oocytes were activated by 10 mM SrCl2 in Ca2+-free CZB medium in the presence of 5 g/ml cytochalasin B and 1% dimethyl sulfoxide and cultured for 4 days in potassium simplex optimized medium (Specialty Media, Lavallette, NJ, http://www.specialtymedia.com).

When they had developed to the morula or blastocyst stages, embryos were used to establish NT-ESC lines as described [6, 11] with a slight modification: 20% knockout serum replacement and 0.1 mg/ml adrenocorticotropic hormone (American Peptide Company, Sunnyvale, CA, http://www.americanpeptide.com) were added to the ESC medium instead of fetal calf serum [30]. Briefly, morulae/blastocysts were treated with acid Tyrode's solution to remove the zona pellucida and placed in 96 multiwell dishes with mouse embryonic fibroblasts (ICR origin) for more than 10 days. Proliferating outgrowths were dissociated using trypsin digestion and replated on fibroblasts until stable cell lines grew out. We used these NT-ESCs within eight passages for all subsequent experiments and did not perform any clonal selection. The effects of sex or organ of the donor cells on the establishment rate of NT-ES lines has been reported previously [7, 8, 11]. More detailed protocols of mouse NT and NT-ESC establishment were published elsewhere [31, 32].

Sequential Establishment of NT-ESCs

For the sequential establishment of NT-ESCs, the original somatic cell nuclei were collected from 10-week-old BCF1 female cumulus cells. When an NT-ESC line was established from those cloned embryos, it was used as a donor for the NT procedure [6, 33] to establish the next series of NT-ESC lines; this was repeated up to eight times. In all experiments, the original donors were used within five to eight passages. We designated the original NT-ESC lines as S1, the second-series NT-ESC lines as S2, and so on until S9.

Production of Cloned Offspring and Tetraploid Complementation Chimeric Offspring

We used four NT-ESC lines (designated BDmt-2, BDmt-4, B6fc-1, and B6fc-2), which were derived from the nuclei of BDF1 male tail-tip fibroblasts and C57BL/6 cumulus cells, and two male ESC lines derived from BDF1 embryos. They were selected according to the percentage of normal karyotypes, or to coincide with other experiments, and used at between four and eight passages. For cloning, NT from NT-ESCs was as described above. For the production of chimeric offspring, NT-ESCs were introduced into the blastocoels of normal (3.5 days post copulation [dpc]) or tetraploid blastocysts from BALB/c or ICR strain mice using a piezo-actuated microinjection pipette [6]. Tetraploid embryos were produced by the electrofusion of two-cell embryos [34]. Cloned and chimeric embryos were transferred into the uteri of day-2.5 pseudopregnant ICR strain females and examined at 19.5 dpc by Caesarian section.

Immunohistochemistry

We used the appropriate manufacturer's staining procedures throughout. Alkaline phosphatase staining was according to the manufacturer's protocol (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com). Immunohistochemistry was performed using the following antibodies: anti-Oct3/4 (monoclonal 1:100; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, http://www.scbt.com); anti-Nanog (monoclonal 1:200; ReproCELL Inc., Tokyo, http://reprocell.com/en); anti-SSEA-1 and -3, (monoclonal 1:100; Chemicon, Temecula, CA, http://www.chemicon.com); anti-SSEA-4 (monoclonal 1:100; Santa Cruz Biotechnology, Inc.), and anti-platelet-derived growth factor-receptor-α (PDGFR-α) (S1) (both 1:100; provided by S. I. Nishikawa). Alexa Fluor 488-, 350-, or 568-labeled secondary antibodies (Molecular Probes, Eugene, OR, http://probes.invitrogen.com) were used for detection as appropriate.

Flow Cytometric Analysis

NT-ESCs and fertilized ESCs were dispersed in 10 μl of goat serum (Sigma-Aldrich). After a 20-minute incubation on ice, the cells were mixed with staining buffer (phosphate-buffered saline containing 0.2% bovine serum albumin) containing anti-SSEA-1 antibody (0.1 μg/106 cells), anti-SSEA-4 (0.1 μg/106 cells), and PDGFR-α (0.2 μg/106 cells). Subsequently, the cells were washed and resuspended in staining buffer containing Alexa Fluor 488-conjugated anti-mouse IgM (BD Biosciences, Franklin Lakes, NJ, http://www.bdbiosciences.com) or anti-allophycocyanin-conjugated anti-mouse IgG (BD Biosciences). After a 30-minute incubation on ice, the cells were analyzed using a FACSaria Cell Sorter (BD Biosciences).

Karyotype Analysis by Giemsa and Spectral Karyotyping with Fluorescent In Situ Hybridization Painting

Chromosomes from NT-ESCs were stained using Giemsa and spectral karyotyping with fluorescent in situ hybridization (SKY-FISH) chromosome painting techniques (Applied Spectral Imaging Inc., Vista, CA, http://www.spectral-imaging.com) according to the manufacturer's protocols. All NT-ESC and ESC lines were used within eight passages. More than 50 metaphase nuclei (for Giemsa staining) or 15–20 metaphase nuclei (for SKY-FISH staining) were examined for each cell line.

DNA Methylation Analysis

Methylation status at specific loci, detected by restriction landmark genomic scanning, was evaluated using a combination of methylation-sensitive restriction enzyme digestion and quantitative real-time polymerase chain reaction (PCR) [35, [36], [37]–38]. We used 10 NT-ESC lines: BDfc-1, -3, -4, -5, and -11 derived from cumulus cells of BDF1 mice, and B6fc-1 to -5 derived from cumulus cells of C57BL/6 mice. These were the lines first established in our laboratory and used for all subsequent gene expression experiments (DNA arrays and high-coverage gene expression profiling [HiCEP]). For controls, five lines of the same genetic background fertilized ESCs (BDES-1, -2, and -5 cells from BDF1 mice and lines B6ES-1 and B6ES-2 from C57BL/6 mice) were used. Genomic DNA of NT-ESCs and ESCs was digested using PstI and then treated with NotI. Each primer set for PCR was designed to amplify regions that included NotI sites. Ten nanograms of genomic DNA, treated with or without NotI, was analyzed by real-time PCR using the appropriate primers. The amount of undigested DNA both in NotI-treated and untreated genomic DNA was estimated by real-time PCR with SYBR green PCR Master Mix using ABI Prism 7,000 and 7,500 Sequence Detection Systems (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com) according to the manufacturer's protocols. The methylation ratio at each RLGS (Restriction Landmark Genomic Scanning) locus was defined as the proportion of the amount of undigested DNA in the NotI-treated genome to that in the untreated one. The initial amount of DNA in the reaction mix was normalized using control primer sets, which amplify sequences without NotI sites. For all samples, we performed at least three independent PCR amplifications in duplicate. Table 1 is annotated with NotI restriction sites obtained using the UCSC (University of California, Santa Cruz) genome browser (accessed March 2005).

Table Table 1.. Expression of ESCs specific marker and normality of karyotype in various NT-ESC lines
original image

Gene Expression Analysis Using Microarrays

DNA microarray analysis was performed as reported [39]. Briefly, total RNA was extracted from NT-ESCs derived from C57BL/6 female cumulus cell nuclei (B6fc-1 and B6fc-2) and control ESCs (B6ES-1 and B6ES-2, established from C57BL/6 embryos by our laboratory and designated as contESC, and in addition to the CCE and EB5 ESC lines). RNA was isolated using Trizol reagent (Invitrogen Life Technologies, Inc., Carlsbad, CA, http://www.invitrogen.com). Double-stranded cDNA was synthesized using the Superscript Choice System (Invitrogen Life Technologies, Inc.). Biotin-labeled cRNA was synthesized using the BioArray RNA Transcript Labeling Kit (Affymetrix, Inc., Santa Clara, CA, http://www.affymetrix.com), and the biotin-labeled and fragmented cRNA was hybridized to the murine genome U74 version 2 GeneChip array series (Affymetrix, Inc.) according to the manufacturer's instructions. Additional expression data from fertilized ESC lines (E14 and ESVJ ESC) were provided by Derek Symula [40]. Affymetrix, Inc., expression data were analyzed using both the Bioconductor suite of programs (http://www.bioconductor.org) and programs developed in-house (eXintegrator, http://www.cdb.riken.jp/scb/documentation/index.html).

Significance analysis of microarray (SAM) data [41] was performed using the affy and siggenes modules of the Bioconductor suite. The specific commands and outputs of the analysis are available on request as a command and a log file. Relationships between the individual ESC samples were calculated as modified all-against-all Euclidean distances, based on the expression of different sets of probe sets selected for their variability in expression across the ESC lines, using functions of the eXintegrator analysis system. Probe sets with high covariations of individual probe pair profiles (calculated as the analysis of variance score), and with high coefficients of variance across the set of ESC samples used in this study, were selected and used to calculate all-against-all distances for the individual samples. Distances were calculated as Euclidean distances of the mean of individually normalized probe pair profiles (with a variance of 1 and a mean of 0 across the samples) mapped to a sigmoid curve. Representations of the individual samples were then mapped to positions in a two-dimensional (2D) field based on the set of distances using an error minimization algorithm [27].

Gene Expression Analyzed by HiCEP

Total RNA was isolated using the RNeasy mini kit (Qiagen, Hilden, Germany, http://www1.qiagen.com) from NT-ESCs derived from C57BL/6 female cumulus cell nuclei or from DBA/2 male tail cell nuclei (B6fc-1, B6fc-2, DBAmt-1, and DBAmt-2) and ESCs derived from C57BL/6 female embryos or DBA/2 male embryos (B6ES-1, B6ES-2, DBAES-1, and DBAES-5). The HiCEP procedure was performed as described [42]. Briefly, 1 μg of total RNA treated with DNase I was converted to cDNA using the SuperScriptIII First Strand Synthesis system (Invitrogen) with 5′-biotinylated oligo(dT) primers. Double-strand cDNA was prepared, digested with MspI, and trapped by avidin bound to magnetic beads. After the fragments digested by MspI (except for most of the 3′-region bearing oligo(dT)-biotin) were washed off, a synthetic adaptor was ligated, and the trapped templates were digested by MseI. The resulting solution was used as a template for 256 runs of selective PCR. The products were denatured and loaded on an ABI PRISM 3100 electrophoresis system (Applied Biosystems).

Results

Pluripotency and Karyotype Analysis of NT-ESC Lines

Newly established NT-ESCs were analyzed for pluripotency, by immunostaining for alkaline phosphatase, SSEA-1, Oct3/4, and Nanog, and for embryoid body formation. For negative controls among ESCs, we examined SSEA-3, SSEA-4, and the differentiated fibroblast marker PDGFR-α. All NT-ESC lines were positive for ESC-specific markers, and negative for differentiation markers, in a pattern similar to fertilized ESCs (Table 1). We also examined the levels of SSEA-1, SSEA-4, and PDGFR-α by cell sorting, but only SSEA-1 was produced by all NT-ESC lines (Fig. 1C). Thus, once established, all these NT-ESC lines showed phenotypes that were almost identical to fertilized embryo-derived ESCs.

The karyotypes of NT-ESC lines were normal when examined using Giemsa and SKY-FISH (i.e. 42 of 50 NT-ESC lines [87%] had normal karyotypes in 44%–100% of the cells) (means were 64.6% by Giemsa staining and 61.4% by SKY-FISH; Table 1). These data are within the normal range for ESC lines (58%–70% and 40%–60%, respectively; Table 1, control ESCs) [43]. Four of the NT-ESC lines had only 10%–20% normal karyotypes. Another four NT-ESC lines had no normal karyotypes, one line showed trisomy in chromosome 1, two lines showed trisomy in chromosome 11, and one line showed Y-chromosome deletions in all examined cells (Fig. 1D). These aneuploid NT-ESC lines could not be distinguished by morphology or by immunostaining for the ESC markers.

Physical Damage to Nuclei of NT-ESC Lines

We next examined whether donor somatic cell nuclei were damaged by the NT procedure. To do this, we performed sequential NT using nuclei from NT-ESCs and repeated this several times to amplify any possible damage from the procedures. However, the success rates of NT and cell line establishment did not decrease by the ninth series (Fig. 1B and supplemental online Table S2) and no karyotype abnormalities were observed by SKY-FISH staining (supplemental online Table S3). On the other hand, we could not generate cloned mice from somatic cell nuclei beyond six generations [44]. Thus, the process of NT in itself at least did not cause the karyotype abnormalities in the NT-ESC lines, but there may have been damage that can be revealed only by full-term development.

Similarity of NT-ESC Lines by Tissue-Dependent and Differentially Methylated Region Assay

Next, we used DNA methylation assays to examine the epigenetic status of NT-ESCs. We previously reported that, even when newborn cloned mice appear normal, a few tissue-dependent and differentially methylated regions (T-DMRs) are aberrantly methylated [16]. There are many T-DMRs dispersed throughout the genome [45], and DNA methylation patterns are specific for different cell types [35]. We selected 26 loci (Table 1) belonging to five subgroups (D, M, UM, TM, and TUM) for DNA methylation analysis. These include the target loci of DNA methyltransferases 3a and 3b in ESCs (D in Fig. 2A), loci that are hypermethylated (M in Fig. 2B) or hypomethylated (UM in Fig. 2B) in most cell types, and loci that are hypermethylated (TM in Fig. 2C) or hypomethylated (TUM in Fig. 2C) in ESCs [36]. We examined 10 NT-ESC lines derived from BDF1 or C57BL/6 female mice (supplemental online Fig. 1) and compared them with five female ESC lines from the same genetic background. As shown in Figure 2, all 26 loci showed very similar methylation patterns in all NT-ESC and ESC lines examined. Although we did not examine the loci of imprinted genes, each T-DMR pattern is known to be cell type-specific [35]. This suggests that there were few or no differences between NT-ESCs and fertilized ESCs in terms of DNA methylation status.

Figure Figure 2..

DNA methylation profiling between BDF1 background nuclear transfer-embryonic stem cells (NT-ESCs) and normally fertilized ESCs, compared using real-time polymerase chain reaction. (A): DNA methyltransferase (Dnmt) target region. (B): Normally methylated and unmethylated regions in the C57BL/6 strain mouse genome. (C): Tissue-dependent differently methylated regions in the C57BL/6 strain mouse genome. Five NT-ESC lines derived from BDF1 female mice and three ESC lines derived from BDF1 embryos were examined. All 26 loci examined showed a very similar methylation status in all the NT-ESC and ESC lines.

Similarity of NT-ESC Lines by DNA Microarray and HiCEP

We then used two independent methods to examine the gene expression patterns between NT-ESCs and ESCs. Using DNA microarray experiments, two C57BL/6 NT-ESC (B6fc-1 and 2) lines were compared with a wide range of control ESC lines. Differences in gene expression between NT-ESC and normal ESC lines were analyzed using two distinct methods. First, we used the SAM method [41] to determine the numbers of genes that were expressed differentially between NT-ESC and control (cont) ESC lines. SAM uses permutation analysis to estimate the false discovery rate, which in turn is used to assign a delta score for each probe set, indicating the likelihood of differential expression.

The two NT-ESC lines we established were compared against two control ESC lines (NT-ESC vs. contESC) also established in this laboratory as well as against a range of ESCs (NT-ESC vs. total ESC). To determine the number of genes that are expected to be differentially expressed between different ESC lines, we also grouped all ESC lines established by us and compared these with the E14 ESC line (NT-ESC + contESC vs. E14) and the E14 and ESVJ lines considered as a group (NT-ESC + contESC vs. E14 + ESVJ). We also compared the E14 line against the ESVJ line to give an indication of the numbers of genes that were differentially expressed between two established and well-characterized ESC lines.

These comparisons are shown in Figure 3A, where we have plotted the number of probe sets classified as differentially expressed (minus the estimated number of false calls) between the indicated groups of ESCs for delta values between 0.2 and 2. Far more differences were found between different established ESC lines (i.e., between E14 and ESVJ) at all delta values than between NT-ESC and contESC lines, indicating that the NT-ES lines do not differ to a greater degree either from each other or from ES lines, as compared with the differences between ES lines.

Figure Figure 3..

Gene expression profiles by DNA microarray were similar between NT-ES cells and normally fertilized ES cells. (A): DNA microarray analysis. Number of probe sets classified as differentially expressed at different delta values calculated for comparisons between different groups of ES cells. The y-axis indicates the number of probe sets with absolute delta values equal to or higher than the corresponding delta values (x-axis) minus the corresponding estimates for false calls. NT-ES cells derived from C57BL/6 female cumulus cell nuclei (B6fc-1 and B6fc-2) were compared with C57BL/6 strain ES cells prepared in our laboratory (NT-ES cell vs. ES cell), and a range of other mouse ES cells, including CCE, EB5, ESVJ, and the E14 lines (NT-ES cell vs. all ES cell). To specify the normal range of delta values expected between different ES cell lines, we also compared the ES cell and NT-ES cell lines from our laboratory with the E14 line (NT-ES cell + ES cell vs. E14) and the combined dataset of the E14 and ESVJ lines (NT-ES cell and ES cell vs. E14 + ESVJ). Delta values were also calculated for the comparison between an E14 line heterozygous for the cdk4 gene and the wild-type E14 line (E14 vs. XA053). (B): Relationships between the different samples used in this analysis; 264 probe sets with variations between the different ES cell samples were used to calculate distances between the different samples. The resulting relationships were used to map the samples to two-dimensional coordinates using a simple error minimization algorithm. The samples are represented by markers with their sample identifiers indicated (100–102, Abca1+/– ESVJ; 106–108, ESVJ; 103–105 cdk4+/– E14; 109–111, E14; 112–113, B6fc NT-ESC; 114–115, B6ESC; 117, CCE; 118, EB5). Labels added to the plot indicate the different groups of ESCs (NT-ES and contES indicates all ES cells established in our laboratory; contES indicates ES cells established by us from normal embryos; NT-ES indicates ES cell lines established by us from cloned embryos; ESVJ, E14, EB5, and CCE are ES cell lines established elsewhere). The colors of markers indicate the amount of error in the mapping (red high, green low). Lines emanating from markers indicate the force vectors used in the energy minimization algorithm. Abbreviations: ContES, control embryonic stem; ES, embryonic stem; NT-ES, nuclear transfer-embryonic stem; ntES, nuclear transfer-embryonic stem.

Furthermore, an inspection of the expression profiles of the 16 probes sets with the highest delta values (NT-ESC vs. contESC) showed that these genes were all variably expressed across the different ESC lines. There were no specific differences in NT-ESC gene expression patterns compared with an extended set of ESC lines (supplemental online Fig. 2). Thus, we were unable to find any genes that are specifically expressed differently between NT-ESC and ESC lines, with the minor differences probably resulting from subtle variations in sample-to-sample preparation.

We also used the Affymetrix, Inc., array data to determine the relationships between different samples obtained. A set of probes that were variably expressed across the ESC lines was used to calculate distance measurements between the individual samples. These distances were then used to map the individual samples to positions in 2D space in a manner that conserved the intersample distances (Fig. 3B). This analysis is analogous to a classical cluster analysis but is able to better indicate the complete set of relationships between individual samples. The results of this analysis agree with those of the SAM analysis. All ESCs (NT-ESC or contESC) derived in this laboratory are similar to each other but are different to the E14 and ESVJ established lines. The largest difference indicated by the SAM analysis is observed between the E14 and ESVJ lines. We also performed this analysis using different sets of probes, but the results did not differ markedly (data not shown).

We next used HiCEP to examine gene expression patterns (Fig. 4A, 4B). This technique can distinguish less than 20% (1.2-fold) differences in gene expression, with very few false-positive peaks and high coverage. It can detect more than 70% of all transcripts, including noncoding transcripts [42]. We identified approximately 18,000 peaks in these cell lines and following scatter-plot analysis revealed a close similarity in gene expression between ESCs and NT-ESCs. The Pearson coefficients of correlation (r values) were .8501 (B6ES-1 vs. B6ES-2), .9213 (B6fc-1 vs. B6fc-2), .8720 (B6fc-1 vs. B6 ES-1), and .9471 (B6fc-1 vs. B6ES-2) (Fig. 4C). Although genome-wide observation detected some variation in gene expression between ESCs and NT-ESCs, a similar level of variation was also observed between different ESC lines, suggesting that the observed differences are not due to the cloning procedure per se. However, this does not necessarily mean that there are no differences in gene expression between ESCs and NT-ESCs. Subsequently we performed a precise analysis of individual peaks and found six for which the expression levels in both fertilized ESC lines were three times as great as the C57BL/6-derived NT-ESC lines and three peaks in which expression level in the NT-ESCs was three times as great as in the ESC lines (Fig. 4A). We subsequently repeated this experiment, using ESC and NT-ESC lines established from DBA2 strain mice, but could not confirm this difference between the ESCs and NT-ESCs (Fig. 4B). Although one might expect to find variation between cells derived from different strains, given that all fertilized ESC lines are considered as functionally equivalent, those differences should not be functionally relevant. Furthermore, if differences in expression found between NT-ESCs and ESCs are to be considered to have some functional significance, the genes responsible should not be variably expressed across several different ESC lines. This evidence gives credence to the notion that these NT-ESCs are similar to fertilized ESCs.

Figure Figure 4..

Gene expression profiling between B6 or DBA/2 background NT-ES cells and ES cells by HiCEP. (A): Representative HiCEP patterns of NT-ES cell lines derived from C57BL/6 female mice (B6fc-1, B6fc-2) and normally fertilized ES cell lines (B6ES-1, B6ES-2). All cell lines were established independently. The x- and y-axes indicate fragment lengths and intensities of fluorescence (expression rate), respectively. Two independent analyses with the same RNA fraction are shown in blue and red. Primers MspI-AA-3′ and MseI-AA-3′ were used for the selective polymerase chain reaction. (B): Representative HiCEP patterns are shown of NT-ES cell lines derived from tail-tip cells of DBA/2 males (DBAmt-1, DBAmt-2), and ES cell lines derived from male DBA/2 embryos (DBAES-1, DBAES-5). All cell lines were established independently. x- and y-axes are the same as in (A). (C): Scatterplot analyses of B6fc-1 versus B6ES-1, B6fc-1 versus B6ES-2, B6fc-1 versus B6fc-2 and B6ES-1 versus B6ES-2 cells. Pearson's coefficients of correlation are indicated at upper left. The x- and y-axes indicate the intensity of peaks. Blue, red, and yellow lines indicate 1.5-, 2.0-, and 3.0-fold differences, respectively. Abbreviations: ES, embryonic stem; HiCEP, high-coverage gene expression profiling; MEF3T3, a cell line established from mouse embryonic fibroblasts (Clontech, Palo Alto, CA, http://www.clontech.com); NT-ES, nuclear transfer-embryonic stem; SCN, suprachiasmatic nucleus of C57BL/6.

Differentiation into Functional Embryonic Tissue

One interpretation of the data is that these NT-ESC genomes may not have been fully reprogrammed. To test the extent to which the NT-ESC nuclei had the capacity for complete reprogramming, we evaluated the frequency of live-born mice by reproductive cloning and in chimeric embryos generated by tetraploid complementation. The former assay tests reprogramming for generating the entire repertoire of embryonic tissues, while the latter tests for generating embryos alone (i.e. fetal and placental tissues are derived in the former situation, and just fetal tissues in the latter). Previously, we demonstrated that ESC nuclei are slightly better for generating cloned mice than are somatic cell nuclei [13, 24, 33]. If the NT-ESCs are indeed identical to normally fertilized ESCs, the success rate of producing cloned mice from NT-ESC nuclei should again be higher than from somatic cell nuclei. Surprisingly, we could generate 13 cloned mice from BDF1 NT-ESC nuclei by NT (0.9%), but this success rate was lower than cloning from ESC nuclei (4%; Table 2).

Table Table 2.. Success rate of NT-ESC cloned mice and tetraploid chimera mice
original image

The most consistent anomaly observed in cloned mice is abnormal development of the extraembryonic tissue, resulting in large, disorganized placentas [13, 14]. Injecting ESCs with tetraploid embryos produces chimeric offspring, in which most or all of the embryonic tissues are derived from the ESCs, with most of the extraembryonic tissues from the tetraploid cells [34, 46]. If NT-ESCs are able to differentiate normally to all embryonic cell types, they should behave similarly to normal ESCs in such tetraploid chimeric embryos. Indeed, this appears to be the case; we have been able to produce a number of chimeric mice derived from NT-ESCs reaggregated with tetraploid embryos. The rates of generating healthy offspring derived from NT-ESC or ESC lines are similar (Table 2) [6, 7, 29, 33], and all those mice show germline transmission at adulthood. This suggests that, like fertilized ESCs, these NT-ESCs can differentiate into all embryonic tissues.

Discussion

In mice, cloned offspring from somatic cells can be obtained only from freshly isolated [24, 28] or primary cultured [13, 28] somatic cells from young hybrid (F1) individuals. However, we previously established several NT-ESC lines from C57BL/6 and C3H strain mice [11], although we were unable to generate full-term cloned mice from these strains of mice [25]. Thus, NT-ESC lines can be established successfully from those cloned embryos, even if those embryos have no potential for full-term development. However, we have demonstrated that these NT-ESCs were identical to fertilized ESCs in their expression of pluripotency markers and tissue-dependent differentially DNA methylated regions of all 26 examined loci. Karyotype analysis demonstrated that most examined NT-ESC lines had normal karyotypes. We found that a few cell lines had low rates of normal karyotypes. It is impossible to judge whether these NT-ESC lines were heterogeneous from the beginning or whether the abnormalities arose because of culture conditions. Thus, karyotype analysis is a prerequisite for ensuring the normalcy of NT-ESC lines, as with fertilized ESCs. However, similar abnormalities are often found in mouse or human ESCs [43, 47], so they do not seem to be specific for NT-ESC lines. In addition, the sequential NT-ESC establishment data indicate that the process of NT or physical treatment in itself did not cause abnormalities in the NT-ESC lines. Moreover, it is known that XO ESCs, which have lost one sex chromosome, can often form germ lineages in chimeric mice [48]. Therefore, even if a particular NT-ESC line is of poor quality, it may be possible to use it for further applications. In addition, it might be possible to clean up and increase the numbers of normal cells by subcloning from such NT-ESCs.

The DNA microarray profiles and HiCEP experiments demonstrate that all ESCs (NT-ESC or contESC) derived in this laboratory are similar to each other but are different to the E14 and ESVJ established lines. Only insignificant differences in gene expression between the NT-ESC and other ESC lines were found here, with the minor differences probably resulting from subtle variations in sample-to-sample preparations. Smith et al. [49] and Brambrink et al. [50] also demonstrated that bovine and mouse cloned embryonic cells closely resemble naturally fertilized embryos, respectively, using global gene expression profiles. Even if NT-ESCs may inherit some undetectable epigenetic abnormalities, the similarities of NT-ESCs and ESCs in terms of molecular and other characteristics indicate that murine therapeutic cloning can provide a reliable model for preclinical stem cell research.

Finally, we examined the differentiation potential of NT-ESCs by tetraploid complementation. If NT-ESCs are able to differentiate normally to all embryonic cell types, the resulting 4n chimeric mice should have healthy phenotypes; the strongest evidence for normality of NT-ESCs. We found the rates of generating healthy offspring derived from NT-ESC or ESC lines to be similar (Table 2) [7, 51], and all those mice show germline transmission at adulthood. Thus, our experiment clearly demonstrates that, like fertilized ESCs, NT-ESCs can differentiate into all normal embryonic tissues in vivo. To establish an NT-ESC line, a cloned blastocyst must be cultured for more than 1 month in vitro. During this period, survival of the cloned embryo is independent of most developmentally related gene expression and of placental development; it might be able to proceed gradually to obtain the normal pattern of gene expression [12, 15]. Another possibility is that some of the cloned embryos possess only a few normally reprogrammed cells [12], and these few cells are insufficient to sustain a viable term fetus to term. However, viable NT-ESCs can be established even from those few normal cells during 1 month of culture. Alternatively, in reproductive cloning, the cloned embryos must express all the appropriate genes and exhibit relatively normal placental development immediately after implantation. This more stringent set of criteria may account for the higher developmental failure rates of presumably incompletely reprogrammed cloned embryos after implantation. However, recently, we have found that the success rate of cloned mice from several NT-ESC nuclei were different, even those established from the same donor at the same time [8]. This suggests that NT-ESC lines have similar in vitro differentiation potential to ESCs but that the extent of reprogramming differs between cell lines.

Conclusion

When compared with ESCs produced by normal fertilization, the mouse NT-ESCs we produced from cloned blastocysts cultured in vitro exhibited similar phenotypes, normal ranges of karyotypes, and identical DNA methylation patterns and gene expression. Nevertheless, many questions remain about NT-ESCs. Does the specific aneuploid composition or the incidence of aneuploidy correlate with the differences in gene expression observed? Does passage number affect this? Would the developmental potential of NT-ESC lines be increased if the lines were clonally rederived in culture, selecting for normal karyotype? Recently, we have found that the reproductive cloning and NT-ESCs establish procedure can be improved by adding Trichostatin A ([TSA] histone deacetylated inhibitor) [52], but we must further examine the effect of TSA before applying it in clinical practice. At least, our data suggest that the therapeutic cloning approach we describe here appears to be a powerful and reliable method for establishing ESCs, which display biological characteristics nearly identical to normally fertilized ESCs.

Disclosures

The authors indicate no potential conflicts of interest.

Acknowledgements

We thank Dr. J. Cummins, Dr. Y. Tabata, and Dr. G. Schatten for critical and useful comments on the manuscript and the Laboratory for Animal Resources and Genetic Engineering for the housing of mice. This research was supported by a Grant-in-Aid for Creative Scientific Research (13GS0008), Scientific Research in Priority Areas (15080211), Young Scientists A (15681014), and a project for the realization of regenerative medicine (the research field for the technical development of stem cell manipulation) to T.W. and S. N. from the Ministry of Education, Science, Sports, Culture and Technology of Japan. M. M was supported by a grant for the 21st Century COE Program of the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Ancillary