Derivation of Euploid Human Embryonic Stem Cells from Aneuploid Embryos

Authors


Abstract

Human embryonic stem cells (HESCs) are pluripotent cells derived from the inner cell mass of preimplantation embryos. In this study, to isolate new lines of HESCs, we used blastocyst-stage embryos diagnosed as aneuploid in preimplantation genetic screening (PGS). During in vitro fertilization treatments, PGS is widely applied to identify chromosomal aneuploidies, especially in cases of advanced maternal age. Embryos that are detected as carrying aneuploidies are destined to be discarded unless donated for research. From 74 fresh PGS-defined aneuploid embryos, we derived seven HESC lines. Most of the embryos were left to hatch spontaneously through the hole created for blastomere biopsy and further treated by immunosurgery. The seven HESC lines exhibited morphology and markers typical of HESCs and the capacity for long-term proliferation. The derived HESC lines manifested pluripotent differentiation potential both in vivo and in vitro. Surprisingly, karyotype analysis of the HESC lines that were derived from these aneuploid embryos showed that the cell lines carry a normal euploid karyotype. We show that the euploidy was not achieved through chromosome duplication. Alternatively, we suggest that the euploid HESC lines originated from mosaic embryos consisting of aneuploid and euploid cells, and in vitro selection occurred to favor euploid cells. We assume that aneuploid HESC lines could be isolated mostly from embryos that are uniform for the aneuploidy. These results led us to conclude that the aneuploid mosaic embryos that are destined to be discarded can serve as an alternative source for normal euploid HESC lines.

Disclosure of potential conflicts of interest is found at the end of this article.

Introduction

Author contributions: N.L.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing; K.N.: administrative support, collection and/or assembly of data, data analysis and interpretation; T.G.-L.: collection and/or assembly of data, data analysis and interpretation; N. Buehler: administrative support, provision of study material or patients; D.H.: provision of study material or patients, final approval of manuscript; N. Benvenisty: conception and design, data analysis and interpretation, final approval of manuscript.

Human embryonic stem cells (HESCs) are derived from the inner cell mass (ICM) of blastocyst-stage embryos [1, 2]. HESCs have indefinite proliferation potential, and they were shown to differentiate into cells from the three embryonic germ layers [3]. These cells may serve as a source of cells for therapy and for the study of human embryogenesis. Furthermore, HESC lines with different mutations could also be used for disease modeling and for drug testing.

HESCs are isolated from surplus embryos of in vitro fertilization (IVF) treatments [1, 2]. One common source of surplus embryos is couples undergoing IVF treatments who have finished their family plan and donated their frozen blastocysts for research. Another source is embryos that were identified by preimplantation genetic diagnosis (PGD) or preimplantation genetic screening (PGS) as carriers of genetic disorders or having chromosomal aberrations.

Embryonic chromosomal aneuploidy is considered to be the main reason for first-trimester abortions. Thus, PGS is used in cases of recurrent implantation failure and recurrent spontaneous abortions. Almost 50% of IVF preimplantation embryos are aneuploid, and the rate of aneuploidy is higher in embryos of older patients [4]. The biopsy of embryos going through PGS is commonly performed in eight-cell-stage, 3-day-old embryos. A hole is drilled in the zona pellucida (ZP), and one or two of the blastomeres are removed for genetic analysis. The blastomeres are analyzed using a single-cell fluorescence in situ hybridization (FISH), which allows the analysis of chromosomes in interphase cells. The first report of FISH analysis of blastomeres was for the purpose of stating the embryo sex using probes of the sex chromosomes [5]. Today FISH analysis is applied in PGS of IVF embryos using multicolor florescence probes for 5–12 chromosomes [6]. By the time the embryos reach their blastocyst stage (5-day-old embryos), the abnormal embryos are identified. Normal embryos are either returned to the uterus or frozen for future pregnancies, and the abnormal embryos are either discarded or used for research purposes.

The widespread use of PGD and PGS in IVF clinics has created a new source of surplus blastocyst-stage embryos that are available for the establishment of HESC lines. Recently, HESC lines were derived from embryos carrying various disorders, among them, Duchenne muscular dystrophy, Fanconi anemia, and fragile-X syndrome [7, [8]–9]. More recently, Peura et al. used PGS embryos with chromosomal aneuploidy for establishing HESC lines [10].

In this study we aimed to derive aneuploid HESC lines for studying different aspects of abnormal embryogenesis. We used embryos for which transfer into the uterus had been avoided because of aneuploidies detected by PGS. We have derived seven HESC lines from these monosomic and trisomic embryos. These HESC lines exhibited markers of undifferentiated HESCs and had a morphology typical of HESCs, capacity for long-term proliferation, and pluripotent differentiation potential in vivo and in vitro.

Surprisingly, karyotype analysis of these HESC lines showed that they carry a normal euploid karyotype. In this article we report our derivation methodology of HESCs from freshly biopsied blastocyst-stage embryos. In addition, we explore the reasons leading to the development of euploid HESC lines from aneuploid embryos.

Materials and Methods

Derivation of HESCs from Aneuploid IVF Embryos After PGS

Isolation of HESCs from blastocyst-stage embryos was performed after approval by the Stem Cell Research Oversight Committee at Cedars-Sinai Medical Center, according to protocol 9647, “Establishment of Human Embryonic Stem Cells from Genetically Abnormal and Spare IVF-Derived Embryos.” Blastocyst-stage embryos were donated according to informed consent under the approved protocol. Blastocysts after PGS analysis were allowed to hatch spontaneously. Blastocysts without the ZP were treated by immunosurgery (IS) to remove the trophectoderm (TE) layer from the ICM [9]. Using a stripper, a naked blastocyst was transferred to a drop of anti-human serum (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) diluted 1:5 with G2V3 blastocyst medium supplemented with human serum albumin (Vitrolife, Englewood, CO, http://www.vitrolife.com) and incubated for 30 minutes at 37°C in a 5% CO2 incubator. Then, the blastocyst was washed three times with the blastocyst medium, and the blastocyst was transferred to a drop of guinea pig complement (Sigma-Aldrich) diluted 1:5 with blastocyst medium for 30 minutes at 37°C in a 5% CO2 incubator. The blastocyst was again washed three times with the blastocyst medium, and blebbing of the TE was apparent. After the TE blebbing, if the TE was not completely removed from the ICM, it was removed by passing the blastocyst through a narrow pipette. The ICM was placed on mouse embryonic fibroblasts (MEF) as a feeder layer with HESC medium [3] supplemented with 30 ng/ml of basic fibroblast growth factor (bFGF). The medium was changed every 2 days, and after 5–12 days, when the colony was approximately 300 μm in diameter, it was cut (using a filter-isolated mouth pipette) into two or three parts and transferred to new plates. The new colonies were transferred every 5–9 days, using the mouth pipette, for the first few weeks. At passages 4–8, as the cell line was more established, we used either trypsin/EDTA or collagenase type IV for transferring the cells [3]. Once the cell line was established, we lowered the bFGF concentration to 5 ng/ml. The established cell lines were named sequentially, according to the date of derivation, Cedars Sinai Embryonic Stem 1 (CSES1) to CSES7.

In Vitro and In Vivo Differentiation of HESCs

Embryoid bodies (EBs) were created as described [3]. EBs were generated by aggregation of the HESCs for different time points: early EBs (2–4 days), mid EBs (7–10 days), and late EBs (20–30 days).

For teratoma formation, 5 × 106 HESCs were injected under the kidney capsule of 6–8-week-old Nude mice [11]. Each cell line was injected into two mice. One month later, the mice were euthanized and the teratomas were removed. Part of the tissue was fixed by 10% buffered formalin phosphate (Sigma-Aldrich), and RNA was extracted from the rest of the tissue as detailed below. The fixed tissues were collected from the mice and embedded in paraffin blocks, and 4-μm sections were stained with hematoxylin and eosin. Care of animals was in accordance with institutional guidelines, as approved by the Cedars-Sinai Medical Center Institutional Animal Care and Use Committee, according to protocol 2182.

Immunostaining and Alkaline Phosphatase Staining

HESCs were washed three times with phosphate-buffered saline (PBS) and fixed onto the plate with 10% buffered formalin phosphate (Sigma-Aldrich). As first antibodies we used mouse IgG anti-human OCT-3/4, mouse IgM anti-human TRA-1–60 (Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com), and rabbit anti-NANOG (Abcam, Cambridge, MA, http://www.abcam.com). As secondary antibodies we used goat anti-mouse IgG PE, goat anti-mouse IgM fluorescein isothiocyanate (FITC), and donkey anti-rabbit FITC (Santa Cruz Biotechnology). The first antibody was incubated with the cells overnight at 4°C. The next morning, the cells were washed twice with PBS and were incubated with the second antibody for an hour at room temperature. The cells were washed, and nucleus staining was performed by incubation with 1 μg/ml bis-benzimide, Hoechst 33258 (Sigma-Aldrich), for 10 minutes. The HESCs grown in a 12-well plate were fixed and labeled with the alkaline phosphatase kit 86R-1KT (Sigma-Aldrich).

Fluorescence-Activated Cell Sorting Analysis

To analyze the cells by fluorescence-activated cell sorting (FACS), the HESCs were trypsinized and resuspended in PBS. The first antibody, either mouse IgG anti-human SSEA4 or mouse IgM anti-human TRA-1–60 (Santa Cruz Biotechnology), was added, and the cells were incubated on ice for 20 minutes. Then the cells were washed; a secondary antibody, either goat anti-mouse IgG FITC or goat anti-mouse IgM FITC (Santa Cruz Biotechnology), was added; and the cells were incubated on ice for 20 minutes in the dark. The cells were washed and immediately analyzed by FACS in a FACSCalibur system (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com). Data analysis was performed by CellQuest software (Becton Dickinson). Forward and side scatter plots were used to exclude dead cells and debris from the histogram analysis plots.

Reverse Transcription-Polymerase Chain Reaction Analysis

Total RNA was extracted using RNeasy Mini (Qiagen, Valencia, CA, http://www1.qiagen.com), and 1 μg of RNA was reverse transcribed by random hexamer priming (Promega, Madison, WI, http://www.promega.com). Complementary DNA samples were subjected to polymerase chain reaction (PCR) amplification with human-specific primers from different exons. All reverse transcription-PCR experiments were performed under nonsaturating conditions. PCR conditions included a first step of 3 minutes at 94°C, a second step of 35 cycles of 30 seconds at 94°C, a 30-second annealing step at 60°C–62°C and 45 seconds at 72°C, and a final step of 5 minutes at 72°C. Primer sequences and the sizes of the final products are described in Table 1. PCR products were assessed by gel electrophoresis on 1% agarose ethidium bromide-stained gels and verified by direct sequencing.

Table Table 1.. Primer sequences and PCR conditions
original image

Gene Expression by Low-Density Arrays

RNA from undifferentiated HESCs and from 10-day-old EBs was used for gene expression studies using the TaqMan Human Stem Cell Pluripotency Array (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com). Gene expression is reported as Δ cycle threshold (Ct) value. For each individual gene the number of amplification cycles for the fluorescent reporter signal to reach a Ct value was estimated and then normalized by subtracting the Ct value obtained for the same sample for a positive control transcript (β actin), to give the ΔCt value. To show the change of expression upon differentiation of the HESCs into EBs, the ΔCt value for a specific gene in EBs was subtracted from the ΔCt values for the same gene in ES cells.

Karyotype Analysis

A 10-cm plate of HESCs in log growth phase was used from each cell line. The night before the procedure, the cells were fed with 10 ml of fresh HESC medium. KaryoMAX colecemid (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) was added directly to the plate of cells during the following morning to give a final concentration of 100 ng/ml, and the cells were incubated for 30 minutes at 37°C in a 5% CO2 incubator. Then the cells were trypsinized, treated with hypotonic solution, and fixed. Metaphases were spread on microscope slides, and by using standard G banding technique, the chromosomes were classified according to the International System for Human Cytogenetic Nomenclature. At least 20 metaphases were analyzed per cell line.

Single-Nucleotide Polymorphism Analysis

The HESC lines CSES1, 3, 5, and 6 were analyzed using the Genechip Human Mapping 500K Array Set (Affymetrix, Santa Clara, CA, http://www.affymetrix.com). The data analysis was based on that of Kim et al. [12]. For each cell line, we analyzed four chromosomes, 14, 16, 18, and 21. Each chromosome was divided to blocks of 1,000 single-nucleotide polymorphisms (SNPs), and in each block the percentage of heterozygous SNPs was calculated. Kim et al. tested the error rate and found that an average of 2.3% of each block consisted of errors [12]. Thus, they assigned homozygosity to any block of 1,000 SNPs where the heterozygous SNP frequency was at or below 5%. We used the same parameters and analyzed blocks of 1,000 SNPs ordered by their chromosomal position along the whole chromosomes detailed above.

Results

In this study we focused on the derivation of HESCs from PGS-defined aneuploid embryos. Typical derivation of HESCs uses frozen blastocyst-stage normal embryos. Thus, we initially processed frozen PGS-defined aneuploid embryos for the derivation of HESCs. Out of 40 frozen blastocysts, 43% of the embryos attached to the feeder layer, but none grew to give a HESC line. Thus, we decided to focus our efforts on deriving new HESCs from fresh embryos. In our analysis we used 6-day-old blastocysts. Blastocysts that were completely inside the ZP were left to spontaneously hatch overnight. Embryos that did not completely hatch were either left to spontaneously hatch or assisted mechanically to release the embryo from the ZP. Since a hole was created in the ZP during the blastomere biopsy for PGS, we decided not to use enzymatic digestion of the ZP to avoid leakage of the enzymes into the embryos.

From 74 blastocysts, we derived seven HESC lines. Approximately two-thirds of the blastocysts were of high quality (grade A, 16%; grade B, 50%). Of the embryos, 51% were already fully hatched upon their arrival at our derivation laboratory, 38% were partially hatched, and 11% were completely inside the ZP. Of the nonhatching blastocysts, 28% required manual assistance for hatching. In most cases (93%), the TE was removed by IS, and a minority required mechanical dissection. In only 55% of the embryos that went thorough the IS did we observe clear blebbing of the TE cells. Of the 74 blastocysts, 53 attached to the feeder layer of MEF, and for 7 blastocysts, HESC lines were derived. Supplemental online Table 1 summarizes the details of derivation of the seven HESC lines.

To define the new cell lines as HESCs, we had to show that the cells express markers of undifferentiated cells, have a self-renewal capacity, and have the pluripotent ability to differentiate into derivatives of the three embryonic germ layers. In Figure 1, we show a comprehensive examination of pluripotency to the first HESC line that we derived, named CSES1. The HESC colonies are positively stained for alkaline phosphatase, TRA-1–60, and NANOG. FACS analysis of CSES1 cells grown on MEF with antibody against TRA-1–60 shows that 77% of the cells are positive for TRA-1–60 and 70% are positive for SSEA4. In Figure 2 we show characterization of the other six HESC lines that we derived, named CSES2–CSES7. The HESCs show typical morphology of HESC colonies and are positive for alkaline phosphatase and OCT4. In colonies of HESCs that were expanded by using collagenase, the center of the colonies is sometimes negative to OCT4 because of cell death of the original clump of cells that started the colonies. All seven cell lines were grown to passage 30, and they kept more than 80% of SSEA4 and TRA-1–60-positive cells (data not shown). To show that the cells are pluripotent, we tested their differentiation potential in vitro and in vivo. In Figure 3A, we demonstrate the differentiation potential of our different cell lines upon in vitro differentiation into EBs. The graph shows quantitatively that 10-day-old EBs express the differentiation markers for the trophectoderm, ectoderm, mesoderm, and endoderm (Fig. 3A). We also show that the CSES1 cell line exhibits markers for differentiated cells representing the three embryonic germ layers, ectoderm, mesoderm, and endoderm, similarly to the established cell line H9 (supplemental online Fig. 1). To analyze the cell lines' differentiation in vivo, the cells were injected into immunocompromised mice. All seven cell lines formed teratomas, as shown in Figure 3B.

Figure Figure 1..

The Cedars Sinai Embryonic Stem 1 (CSES1) cell line exhibits markers of undifferentiated human embryonic stem cells (HESCs). (A): Typical morphology of defined HESC colonies of small undifferentiated cells on mouse embryonic fibroblast (MEF) feeder layer (passage [P] 5). (B): The HESC colonies (P2) were positively stained for alkaline phosphatase (red). (C): HESC colony (P6) positively stained for TRA-1–60 (green). Hoechst-stained nuclei of HESCs and MEFs are shown in blue. (D): Fluorescence-activated cell sorting (FACS) analysis of CSES1 (P19) grown on MEF with antibody against TRA-1–60 (77% were positive). (E): FACS analysis of CSES1 (P19) grown on MEF with antibody against SSEA4 (70% were positive). (F): Hoechst staining of HESCs and MEF nuclei (blue). (G): HESC colony positively stained for Nanog (green). (H): HESC colony positively stained for OCT4 (red).

Figure Figure 2..

The CSES2–CSES7 cell lines exhibit markers of undifferentiated human embryonic stem cells (HESCs). (A): Typical morphology of HESC colonies on mouse embryonic fibroblast (MEF) feeder layer (passage [P] 4–P8). (B): HESC colonies (P5–P10) positively stained for alkaline phosphatase (red). (C): HESC colonies (P7–P14) positively stained for OCT4 (red). Hoechst-stained nuclei of HESCs and MEF are shown in blue. Abbreviation: CSES, Cedars Sinai Embryonic Stem.

Figure Figure 3..

The Cedars Sinai Embryonic Stem (CSES) cell lines exhibit markers for differentiated cells representing the three embryonic germ layers (ectoderm, mesoderm, and endoderm) upon in vitro differentiation into EBs and in vivo differentiation into teratomas. (A): A graph showing quantitatively that 10-day-old EBs express the differentiation markers for the trophectoderm, ectoderm, mesoderm, and endoderm. The bars represent the average of the expression values of Δ cycle threshold (Ct) of embryonic stem cells − ΔCt of 10-day-old EBs for the representative genes. (B–H): Shown are 4-μm sections of teratomas labeled by hematoxylin and eosin. (B): CSES1, neural rosettes. (C): CSES2, gut-like structures. (D): CSES3, gut-like structures. (E): CSES4, cartilage. (F): CSES5, connective tissue. (G): CSES6, adipose cells. (H): CSES7, adipose and muscle cells. Abbreviations: AFP, alpha-fetoprotein; Brachyury-T, brachyury homolog (mouse); CDH5, Cadherin 5, type 2, VE-cadherin; CDX2, caudal type homeobox 2; EB, embryoid body; GCM1, glial cells missing homolog 1; IPF1, pancreatic and duodenal homeobox 1; ISL1, ISL LIM homeobox 1; KRT1, Keratin 1; MYF5, myogenic factor 5; MYOD1, myogenic differentiation 1; NEUROD1; neurogenic differentiation 1; PAX6, paired box 6; PECAM1, platelet/endothelial cell adhesion molecule; SERPINA1, serpin peptidase inhibitor, clade A, member 1.

To further characterize the cell lines, we tested their karyotype. Surprisingly, the analysis showed that the seven cell lines carry a normal euploid karyotype (Fig. 4). Both HESC lines derived from monosomic embryos (CSES1, 2, 3, 4, 7) and those isolated from trisomic embryos (CSES5, 6) showed a normal karyotype.

Figure Figure 4..

Karyotype analysis of Cedars Sinai Embryonic Stem 1 (CSES1)–CSES7 cell lines. (A): CSES1, analyzed in passage 12, 46XX. (B): CSES2, analyzed in passage 12, 46XX. (C): CSES3, analyzed in passage 5, 46XX. (D): CSES4, analyzed in passage 33, 46XY. (E): CSES5, analyzed in passage 16, 46XX. (F): CSES6, analyzed in passage 10, 46XX. (G): CSES7, analyzed in passage 9, 46XX.

One explanation how monosomic embryos can give rise to euploid HESC lines is a duplication of the monosomic chromosome. In this self-correction mechanism the cells will end up with diploid number of chromosomes but in uniparental disomy. To test this hypothesis we analyzed CSES1, 2, 3, 4, and 7 for heterozygosity in those chromosomes that showed monosomy by PGS. This analysis was performed by examining polymorphism of multiple microsatellites along these chromosomes. Supplemental online Table 2 shows that in all of these cell lines there is clear heterozygosity in the analyzed chromosomes. To verify the results, we tested CSES1 and CSES3 using the Genechip Human Mapping 500K Array Set. This analysis showed that many heterogeneous SNPs are distributed along the chromosomes that were initially reported as monosomies, similarly to all other chromosomes (Fig. 5). These results suggest that a mechanism of duplication of the monosmic chromosome was not the one that led to the diploid number of chromosomes.

Figure Figure 5..

The distribution of heterozygous SNPs along four representative autosomes in CSES1, CSES3, CSES5, and CSES6 cell lines. Using the Genechip Human Mapping 500K Array Set, genotyping approximately 500,000 SNPs, we showed that many heterogeneous SNPs are distributed along the chromosomes that were initially reported as carrying monosomies (CSES1, CSES3) or trisomy of chromosome 18 (CSES5, CSES6) similar to other chromosomes. Each bar represents the percentage of the heterozygous SNPs in a block of 1,000 SNPs ordered by the physical chromosomal position. The SNP signature showed heterozygosity for all the autosomes, ruling out the option that the chromosomes originated from duplication. The average number of heterozygous SNPs in these four chromosomes in the four cell lines is 23.3% (ranging between 19.3% and 25.3%). Abbreviations: CSES, Cedars Sinai Embryonic Stem; SNP, single-nucleotide polymorphism.

In a similar fashion, to explain the fact that two embryos that were reported as trisomic for chromosome 18 gave rise to euploid HESC lines, we tested whether a cellular self-correction was the mechanism that led the cells to lose an extra third chromosome, ending up as diploid. In cases where the trisomy resulted from chromosome duplication, if it later lost the other chromosome in becoming diploid, we can again have uniparental disomy. We therefore tested CSES5 and CSES6 by the SNP microarray analysis and showed that similar to all other chromosomes, many heterogeneous SNPs are distributed along chromosome 18 (Fig. 5). Therefore, we have no support for the hypothesis that the euploidy was achieved by chromosome loss.

A second explanation that may lead to euploid HESCs from embryos analyzed as aneuploid is that the embryos were mosaic for aneuploid and euploid cells. There is a possibility that during the derivation of HESC lines, euploid cells had a growth advantage over the aneuploid cells and a selection favoring the euploid cells occurred in culture. To test this hypothesis we did FISH analysis to see whether mosaicsm is detectable in the early passage of HESCs. One hundred seventy-seven cells from CSES1 at passages 3 and 10 were analyzed for chromosomes 13 and 21. We found that in passage 3, 15% of the cells showed 21 monosomy, whereas in passage 10, less than 4% showed this anomaly. Chromosome 13, which served as internal control, showed a negligible number of anomalies in both passages. We conclude that in the early passages of HESCs we still have a mosaic of euploid normal cells and cells abnormal to chromosome 21, and in higher passages the euploid cells took over the culture.

Discussion

Typical derivation protocols uses spare frozen 5–6-day-old blastocyst-stage embryos [1, 2]. Usually at that stage the embryos are contained within the ZP, and enzymatic or acidic digestion is used to release the embryo from the ZP [1, 2]. The embryos that we used for the derivation of HESCs arrived at the derivation laboratory only on day 6, and they were all biopsied for PGS analysis on day 3. Some were fully hatched, some were partially hatched, and several were completely inside the ZP. In several reports, derivation of HESCs from biopsied embryos used pronase digestion to remove the ZP [7, 13, 14]. We did not use pronase and allowed the blastocysts to spontaneously hatch from the ZP. We derived seven new HESC lines from these PGS-biopsied embryos and showed that the cells express markers of undifferentiated cells, have a self-renewal capacity, and have the pluripotent ability to differentiate into derivatives of the three embryonic germ layers (Figs. 1, Figure 2.3; supplemental online Fig. 1).

The embryos that we used for the derivation of new HESC lines were identified as having chromosomal aberrations by PGS and were destined to be discarded. We aimed to derive aneuploid HESC lines from these embryos to study different aspects of abnormal embryogenesis. Surprisingly, the seven HESC lines were found to carry a normal euploid karyotype (Fig. 4; supplemental online Table 1). These results led us to conclude that these aneuploid embryos that are destined to be discarded can serve as an alternative source for normal euploid HESC lines. Of the couples having aneuploid blastocysts, 60% agreed to donate them for the derivation of HESCs, a percentage that is much higher than the percentage of couples agreeing to donate euploid spare embryos. This new source of embryos, together with arrested and poor-quality embryos, expands the pool of embryos that can be used for derivation of euploid HESC lines [15, 16].

We speculate that there may be three reasons for the derivation of euploid HESC lines from embryos diagnosed by PGS as aneuploid: (a) misdiagnosis of the blastomere as a result of technical factors, (b) cellular self-correction to a diploid number of chromosomes, and (c) embryonic self-correction in mosaic embryos consisting of aneuploid and euploid cells. (a) The technical factors that may lead to misdiagnosis can occur during the preparation of the blastomere and in the course of FISH analysis [17, [18]–19]. FISH artifacts may lead to a 7%–15% rate of misdiagnosis [18]. Thus, we cannot rule out misdiagnosis as the reason for the discrepancies, but it is not reasonable to assume that all seven cell lines were misdiagnosed. In addition, the appearance of chromosomal aneuploidies in early passages of CSES1 cells, shown by FISH, also argues against misdiagnosis. (b) In cellular self-correction, the diploid cells might be a result of duplication of the original haploid chromosome in the embryos carrying monosomies or a result of losing the extra chromosome in cases of trisomic embryos. To test these mechanisms, we checked whether there is heterozygosity in several loci along the chromosome. In the case of a haploid chromosome that was duplicated, we would expect homozygosity along the chromosome. In the cases of trisomic chromosomes that turned out to be diploid, we might see homozygosity along the chromosome, telling us that indeed there were three copies of the specific chromosome, and one was lost. Our analysis of few microsatellites and global SNP analysis demonstrated that heterozygosity was shown in all chromosomes, ruling out the self-correction mechanism. (c) We assume that these aneuploid embryos originated from embryos that were mosaic for aneuploid and euploid cells. Mosaic embryos have been reported at all stages of embryonic development [20]. Almost half of the chromosomally abnormal cleavage-stage embryos are mosaic [21, [22]–23]. Mosaicism is suggested to originate and persist mainly because of the low expression of certain cell cycle checkpoint genes at early developmental stages. The embryonic genome is activated at day 3 of development [24]. Therefore, early embryonic development relies on maternal transcripts. The maternal transcripts might lead to aneuploidy until the embryo genome is activated [25]. A mosaic embryo can develop to a morphologically normal blastocyst or even to a live birth if at least two-thirds are normal, overwhelming the abnormal cells [26]. Coulam et al. published an analysis of biopsy of two blastomeres from the eight-cell-stage embryos and found that in only 25.5% of the embryos is there concordance between the two blastomeres [27]. Previous report showed that 3-day-old embryos have approximately 50% mosaicism [28]. In addition, FISH results from embryos that were biopsied twice, once at day 3 and again at day 5 or 6, showed a confirmation rate of the first analysis in approximately 60% of the embryos [26, 28]. It has been suggested that mosaicism in early embryos represents a capability of early embryonic development to “self-correct” [26, 29]. Munné et al. have demonstrated by FISH analysis of day 12 cultures that chromosome self-normalization occurs in a significant proportion of chromosomally abnormal embryos [29]. According to our FISH results for early and late passages of HESCs, we assume that during the growth of the HESCs, the aneuploid cells had a disadvantage compared with the euploid cells, and thus the euploid cells divided much more efficiently and took over the culture. Other mechanisms that may cause the mosaic embryos to give rise to euploid HESC lines are as follows: (a) there may be a contribution of the aneuploid cells to the trophectoderm and therefore elimination from the ICM; (b) aneuploid cells may have a disadvantage compared with euploid cells within the ICM and therefore can be eliminated before the initiation of the derivation process; and (c) clonal growth of HESCs from a single euploid blastomere may occur.

We used 34 embryos that were reported as carrying a blastomere with abnormal aneuploid monosomy of one of the autosomes or several chromosomes. Out of these 34 embryos, we derived five HESC lines. Karyotype analysis of HESC lines CSES1 and CSES2 in passage 12 showed that both cell lines have a normal female karyotype, 46XX. Karyotype analysis of HESC lines CSES3 and CSES7 in earlier passages, passages 5 and 9, respectively, also showed a normal female karyotype, 46XX. CSES4 showed a normal male euploid karyotype, 46XY (Fig. 4). Looking back at the literature, we could not find any reports of a HESC line that is aneuploid with monosomy in autosomes. No live birth of embryos carrying monosomies in autosomes is known [30]. The only live birth carrying monosomy is the absence of a second normal sex chromosome, which is diagnostic of Turner syndrome. It is hypothesized that the 45X conceptuses that survive to live birth are frequently mosaic for a cell line with normal chromosomes [30]. Assuming that the monosomy detection is not due to misdiagnosis of PGS, we decided to test whether a process of cellular self-correction was responsible for that phenomenon. For each chromosome that was reported as haploid and turned out to be diploid, we choose four or more different microsatellites along the chromosome and tested their status using microsatellites analysis (supplemental online Table 2). For each chromosome we found that at least two microsatellites show clear heterozygosity. To verify the results, we sent samples of DNA from CSES1 and CSES3 for SNP microarray. In Figure 6 we show the distribution of heterozygous SNPs along four representative autosomes of CSES1 and CSES3. The SNP signature showed a similar percentage of heterozygosity along all the chromosomes. To test the third option that the euploid karyotype was due to mechanisms of embryo self-correction of a mosaic embryo, we did FISH analysis of an early passage of CSES1 and found a significant number of cells carrying abnormal numbers of chromosome 21. Thus, we conclude that the embryos that were reported as carrying a monosomy were mosaic and had euploid cells, and these euploid normal cells took over the culture.

Figure Figure 6..

Our model for the derivation of euploid human embryonic stem cells (HESCs) from abnormal embryos. According to our results, we hypothesize that the euploid HESCs that we derived from embryos that were detected as carrying chromosomal abnormalities originated from mosaic blastocyst-stage embryos. In vivo, these mosaic blastocysts will develop either to a normal embryo or to an abnormal mosaic embryo. Depending on the specific abnormality and percentage of mosaicism, the mosaic embryos will either spontaneously abort or develop to sick newborns. Abbreviation: ES, embryonic stem.

From the embryos carrying a blastomere with trisomy of one or more chromosomes (n = 24), we derived two HESC lines. Karyotype analysis for these two cell lines showed that they carry a 46XX karyotype (Fig. 4). In the past, several laboratories reported the derivation of HESC lines carrying a trisomy. Heins et al. derived HESC lines from surplus embryos and found that five were euploid but one was triploid and another one had a trisomy of chromosome 13 [31]. Peura et al. derived three aneuploid HESC lines with the karyotype of 47XX+16, Mosaic of 46XX, i(13)(q10), and 46XX and 47XY,+5 [10]. It was also reported that during long-term culture of HESCs normal euploid HESC lines tend to become aneuploid and gain chromosomes or parts of them. HESCs were reported to have a recurrent gain of chromosomes 17 and 12 [32, 33]. In vivo, autosomal trisomy syndromes are known to affect live births. Autosomal trisomy was identified in 20% of the spontaneous aborted pregnancies [30]. Trisomy for every chromosome except chromosome 1 has been reported. The reports on HESCs carrying trisomies, together with the fact that live births of embryos carrying trisomies are sometimes viable, led us to conclude that cells carrying an extra chromosome are viable and capable of proliferating and differentiating. We assume that creating trisomic HESC lines from trisomic embryos is feasible depending on whether the extra chromosome gives the cells a growth advantage. In our case the two embryos that led to euploid 46XX HESC lines originated from embryos carrying an extra copy of chromosome 18. We assume that the extra copy of chromosome 18 did not give the cells a growth advantage. Thus, either the cells went through cellular self-correction and depleted the extra copy or, in the case of mosaic embryos, this outcome most likely reflects a positive growth selection pressure during the stem cell derivation and culture, favoring a subset of cells with a normal chromosome complement.

Summary

The fact that embryos that had an aneuploid blastomere carrying either monosomies or trisomies gave rise to euploid HESC lines raises the question of whether they could have developed to euploid normal babies. As summarized in our model, we assume that the embryo development is different from the development of HESCs (Fig. 6). The derivation and expansion of HESC lines to passage 5, the earliest possible point at which to analyze for karyotype, demands that the ICM cells will go through many cell divisions. The growth conditions in vitro in culture are not optimal, and thus we assume that the aneuploid cells divide slowly, whereas the euploid normal cells have a self-renewal advantage, and they proliferate faster and take over the culture. This phenomenon is similar to adaptation that was reported in the cases of long-term culture of HESCs [32, 34]. In the case of long-term HESC cultures, the opposite result was seen: karyotypically normal euploid HESCs became aneuploid. The aneuploid cells were gaining genes and chromosomes that provided the cells with a strong advantage for maintenance in vitro. In the case of embryogenesis, as opposed to culture of HESCs, the life span of the ICM is very short, and they go through only a few cell divisions and then differentiate. We assume that in vivo, these mosaic blastocysts will develop either to a normal embryo or to an abnormal mosaic embryo.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.

Acknowledgements

We thank Dr. Rachel Eiges for assisting us with the initial derivation attempts of HESC lines, Dr. Dalit Ben-Yosef and Tzvia Frumkin for the FISH analysis, and Lauren Rimoin and Jonathan Bolotin for excellent technical assistance.

Ancillary