Human embryo development occurs through a process that encompasses reprogramming, sequential cleavage divisions and mitotic chromosome segregation and embryonic genome activation. Chromosomal abnormalities may arise during germ cell and/or pre-implantation embryo development, and are a major cause of spontaneous miscarriage or birth defects. Nonetheless, model systems suitable for the study of human germ cell and embryo development have been limited until recently. We suggest that human embryonic stem cells may provide a valuable human cell-based model for genetic studies of human pre-implantation pluripotent cells. Here, we review the current literature on diagnosing chromosomal abnormalities in the pre-implantation embryo, and the importance of provisions from the human oocyte in establishing and maintaining the human embryonic genome during the first 3 days post‐conception. We focus on transcriptional analysis of human oocytes and embryos and the unique cell cycle and checkpoint requirements in the early embryo. Taken together, data suggest that the unique programs of the early human embryo, including management of aneuploid cells, may paradoxically promote embryo development but contribute to the high rate of spontaneous miscarriages in human pregnancies.
HUMAN EMBRYO DEVELOPMENT
Human embryo development begins at fertilization. Fertilization triggers the completion of meiosis II in the oocyte and subsequent fusion of the haploid male and female pronuclei in the zygote. The maternal and paternal chromosomes are then replicated and subsequently a series of cleavage divisions ensue until approximately the fourth day of development when the embryo compacts to form the morula and then begins the first distinct cellular differentiations with the formation of the trophectoderm and inner cell mass of the blastocyst. In vivo, these events occur in the fallopian tubes and uterus and consequently are difficult to observe. In vitro observation, however, has been facilitated by the use of assisted reproductive technologies such as in vitro fertilization (IVF) in which an egg is fertilized by sperm in vitro.
In vitro fertilization was first successfully used in England in 1978 to treat infertility and was initially developed as a way of bypassing common somatic reproductive health problems such as endometriosis or tubal blockage. Since its introduction, the use of IVF has increased dramatically so that as many as 1% of all births are now being conceived in vitro. As a consequence of IVF and the culturing of human embryos in the laboratory, we have gained insights into many of the morphological events that take place during normal development of the human embryo. In addition, since the process of IVF generates, on average, more embryos that can be transferred for reproductive purposes, some couples may donate embryos to research. Indeed, the donation of embryos for research directly led to the generation of human embryonic stem cells (hESCs) (1). hESCs constitute an important model for understanding the earliest events in human embryo development, in combination with studies of human oocytes and embryos donated for research.
THE IMPORTANCE OF GENOMIC INTEGRITY IN GAMETES AND BLASTOMERES
All cells in the human body are derived from the two haploid gametes and subsequent cleaving diploid blastomeres. While chromosomal abnormalities and genetic lesions that arise in the gametes are inherited by all daughter cells, chromosomal abnormalities or genetic lesions acquired in one or a few blastomeres during cleavage are inherited in a mosaic pattern with a portion of cells being of normal ploidy (euploid) whereas others may be chromosomally abnormal (aneuploid). It is predicted that the earlier in cleavage a genetic lesion occurs, the greater the number of daughter cells that will inherit the genetic defect. Therefore, maintenance of genomic integrity during gametogenesis, fertilization and cleavage is essential for normal human embryogenesis and fetal disorders.
Some of the most significant outcomes of chromosomal and genetic lesions inherited from the gametes or blastomeres are spontaneous miscarriages and congenital abnormalities. Spontaneous miscarriage may occur in ∼10% of all pregnancies, with an increased risk associated with increasing maternal age. Birth defects are observed in 3% of live born infants, with a significant proportion (20%) attributed to chromosomal aberrations and/or gene mutations (2). Aneuploidies that can be compatible with viable pregnancies include those associated with chromosomes 13, 18, 21, X and Y. Therefore, understanding the underlying causes of aneuploidy in the oocyte and early cleaving human embryo provides essential information for understanding the causes of these reproductive diseases.
METHODS FOR DETECTING CHROMOSOMAL ABNORMALITIES IN PRE-IMPLANTATION HUMAN EMBRYOS
Diagnosing chromosomal abnormalities during IVF, prior to transfer of embryos for reproductive purposes, is termed pre-implantation genetic diagnosis (PGD). PGD is accomplished by removing a single blastomere during the human embryo cleavage. This procedure is used for genetic diagnosis of common and rare genetic diseases with a simple inheritance pattern and for diagnosis of chromosomal aneuploidies associated with advanced maternal age (though the success of this later technology has been debated extensively). Currently, PGD is used to diagnose more than 200 diseases, including cystic fibrosis, Tay Sach’s, hemophilias, Huntington’s disease, sickle cell anemia and Fragile X syndrome and can be used for human leukocyte antigen (HLA)-matching (3–5). PGD is highly effective for diagnosing genetic defects originating in the gametes; however, in cases where the pre-implantation embryo is mosaic, the genetic lesion may not be detected in the single blastomere that is biopsied and thus, alternate strategies to identify embryos with genetic lesions without biopsy is highly desirable.
Recently, human embryo morphology was evaluated as a surrogate marker for chromosomal abnormalities (6,7). These studies suggested that there may be a correlation between the distribution and number of nucleoli in the male and female pronuclei, together with the position of the second polar body relative to the first, and the chromosomal status of the zygote (6,7). With regard to cleavage-stage embryos, a recent study comparing morphology with the results of PGD found that arrested cleavage-stage embryos exhibited a high frequency of chromosomal abnormalities (8). This study also suggested that abnormal rates of cleavage (either too fast or too slow) might predict chromosomal lesions in embryos. Although preliminary, these studies suggest that morphology could be a useful indicator of aneuploidy in some embryos and under some conditions. Nonetheless, more researches with a larger number of participants are required before alternative techniques will be verified and integrated into clinical practice.
ROLE OF THE OOCYTE IN REGULATING GENOMIC INTEGRITY OF THE PRE-IMPLANTATION EMBRYO
Advanced maternal age is associated with chromosomal defects that affect one or two specific chromosomes such as chromosomes 18, 21 and X (9). In contrast, random or complex chromosomal abnormalities in which three or more chromosomes are affected are hypothesized to occur through post-zygotic mechanisms rather than genomic errors originating in the haploid gametes. In humans, the oocyte plays a central role in providing RNAs and proteins that maintain genomic integrity of the zygote and cleavage-stage embryos until Day 3 when embryonic genome activation occurs (10). Nonetheless, we know little about the mechanisms in the oocyte that provide resources to ensure genomic integrity in pre-implantation cells after fertilization.
Several studies have now assessed genome-wide transcriptional profiles of human oocytes and embryos (11–14). In one study, which evaluated oocyte maturation, primary oocytes were profiled and compared with oocytes in metaphase I and metaphase II, as well as embryos with Day 3 of development (11). This study revealed that oocyte maturation and the first few days of embryo development were characterized by a significant decrease in transcript levels, suggesting that decay of RNAs associated with gamete identity is integral to embryo development (11). Furthermore, this study revealed that developmental arrest and activation of the embryonic genome are uncoupled events.
Another study sought to compare transcripts in human and mouse metaphase II oocytes with those in mouse and human ESCs (12). In this study, the authors identified 5331 transcripts that were highly expressed in the human oocyte relative to somatic cells. Intersection of these 5331 transcripts enriched in human metaphase II oocytes with the 1626 transcripts enriched in hESCs identified 388 common genes that are hypothesized to play an important role in oocyte reprogramming and early events associated with human pre-implantation embryo development. To further examine these transcripts, we performed gene ontology analysis of these 388 common genes (12), using the DAVID conversion tool in order to identify highly represented functional groups in common between hESCs and oocytes. This analysis revealed that common oocyte and hESC genes are enriched in gene ontologies associated with mitosis, cell cycle, cell division and DNA repair. Therefore, genes involved in regulating passage through mitosis are highly enriched in oocyte and hESCs. Analysis of these various oocyte data sets was also performed by Sudheer and Adjaye (14). In this study, 5862 transcripts in human blastocysts were compared with 4038 combined oocyte expressed genes from the previous studies. This analysis revealed 894 genes in common between the human oocyte and the human blastocyst, further suggesting that genes indicative of mitosis as well as transcriptional and translational control are enriched in the oocyte and hESCs (14).
In summary, early human embryo development in the 5 days following fertilization follows a precise series of events that culminate in the formation of a blastocyst. The first 2 days of human embryo formation are controlled by factors expressed solely by the oocyte, demonstrating the central importance of genes expressed by the oocyte in regulating transcription and translation by the newly formed embryo. Furthermore, the abundance of genes associated with mitosis and DNA repair in both oocytes and ESCs suggests that this molecular program is required for all stages of pre-implantation embryo development and not just for the first 2 days of human embryo development (12).
USING HESCS AS A MODEL FOR EVALUATING GENOMIC INTEGRITY IN PRE-IMPLANTATION EMBRYOS
As discussed above, donated embryos from IVF clinics have provided significant albeit descriptive insight into the morphology and gene expression patterns in oocytes and pre-implantation human embryos. In contrast, hESCs provide a genetically malleable cell-based tool for examining human pluripotent cell biology and lineage differentiation (15). For example, recently, hESCs were derived from embryos that had been diagnosed with Fragile X following PGD (16). The use of these cell lines revealed that the critical gene FMR1 is normally expressed in pluripotent cells, with abnormalities in gene regulation occurring upon lineage differentiation (16). Importantly, these experiments demonstrate the utility of hESCs as a model for understanding the molecular basis for inherited genetic mutations and the consequence on embryonic linage formation.
hESCs can be used to evaluate human pluripotent cell biology and early lineage differentiation into the three embryonic layers as well as the germline as described previously (Fig. 1; (17–19)). As discussed above, hESCs are cultured as self-renewing colonies on mouse (MEFs) and/or human embryonic fibroblasts (HEFs) in the presence of fibroblast growth factor 2 (FGF2) (Fig. 1A–C). To induce differentiation, colonies are removed from the feeder layer and cultured in suspension as embryoid bodies (EBs) in differentiation media containing fetal bovine serum (FBS) and no FGF2 as described. During the course of differentiation, samples are taken at various time points and processed for further analysis by mRNA and protein analysis, morphology and functional characteristics. Some genes such as NANOG and OCT4 characteristically decrease over the differentiation time course as pluripotent cell populations decline. In contrast, other markers of embryonic lineage differentiation including Neural cell adhesion molecule 1 (NCAM1), an ectodermal marker, Kinase insert domain receptor (KDR), a mesodermal marker, Alfa fetoprotein (AFP), an endodermal marker, and VASA, a germline marker, increase with differentiation. Using this simple system, it is possible that germ cell-expressed genes and genes identified in pluripotent cells, which are hypothesized to play a role in pre-implantation human embryo biology and the regulation of genomic integrity, can be genetically evaluated.
In mammalian somatic cells, genomic integrity is maintained by a coordinated effort between the macromolecular machinery of the mitotic apparatus, cell cycle checkpoint controls, DNA repair and cell cycle arrest. From the end of mitosis (M), through to the end of Gap1 (G1), this translates into a series of phosphorylation events involving the cdks, CyclinD and CyclinE that ultimately phosphorylate Rb, liberating E2F and enabling passage into S-phase. DNA synthesis occurs in the S-phase of the cell cycle. In somatic cells, cell cycle checkpoints in G1 include the activity of p21, p27, p15, p16, p18 and p19, which inhibit the activity of the cyclin-dependent kinases. By inhibiting kinase activity, the checkpoint remains ‘on’, preventing DNA replication in S-phase. In somatic cells, these checks and balances maintain genomic integrity by providing time for DNA repair, exit from the cell cycle or apoptosis. Before the availability of ESCs as a model for pluripotent cell biology, it was presumed that these control mechanisms were a common feature to all cells; however, recent evidence suggests that the cell cycle in ESCs cultured under self-renewing conditions is remarkably different.
RELAXED CELL CYCLE CHECKPOINTS AT G1-S PHASE IN ESCS COMPARED WITH SOMATIC CELLS
One difference between somatic cells and ESCs is found in the timing of the phases of cell cycle, and lack of phasic expression of many cell cycle regulators. The cell cycle of ESCs is characterized by a very short G1. As a result, a high proportion of cells are observed in S-phase (20). In somatic cells, G1 is predicted to take 6‐12 h, with transition from G1 to S-phase committing cells to the proliferative cycle and mitosis. Therefore, monitoring genomic quality in G1 prior to this commitment is an evolutionarily conserved method for preserving genomic integrity in daughter cells particularly in response to DNA damage. The reason for a short G1 in ESCs is not entirely clear; however, it has been hypothesized that the shortened G1 is essential for rapid proliferation and pluripotency (21). Alternatively, a shortened G1 could result in reduced integrity of cell cycle checkpoints, and therefore propagation of genetic mutations.
In somatic cells, p53 is a critical sensor for DNA damage, inducing p21-, p15- and p27-mediated inhibition of Cdk2/Cyclin E and Cdk4/Cyclin D. This prevents the normal phosphorylation of Rb and blocks the G1 to S transition. In hESCs, p27 is expressed at or below the limits of detection, which may provide one mechanism for driving cells into S-phase at a faster rate (therefore, shortening G1). Skp2 is a ubiquitin ligase that targets p27 for degradation by the proteosome. Skp2 expression in undifferentiated hESCs is particularly high in G1, where p27 is barely detectable (22). Therefore, the ratios of p27/Skp2 most likely translates into the non-phasic expression of Cyclin E, which together enables a G1 to S transition and presumably a relaxed G1/S checkpoint (23). Furthermore, the Rb protein, which is normally hypophosphorylated in somatic cells and is subsequently phosphorylated by Cdk2/4CyclinE/D to pass the G1 checkpoint, is constitutively phosphorylated in mouse ESCs. Taken together, these unique cell cycle features of ESCs appear to all contribute toward relaxed G1–S checkpoints compared with somatic cells, and therefore a faster turnaround of ESCs in mitosis.
RELAXED MITOTIC CHECKPOINTS IN ESCS COMPARED WITH SOMATIC CELLS: A REASON FOR ANEUPLOIDY?
The function of this mitotic checkpoint in pluripotent cells is not clear. However, the analysis for genes enriched in oocytes and hESCs demonstrates that a large number of mitotic spindle assembly checkpoint (SAC) proteins are enriched in metaphase II oocytes and hESCs. This suggests that high levels of these proteins may be necessary in order to safeguard against aneuploidy in cleavage-stage pre-implantation embryos, given that the downstream G1–S checkpoints may not be triggered. It will be important to evaluate whether oocytes from women who have higher risk of conceiving children with aneuploidy (for example, women older than 40) have reduced levels of any of these cell cycle or mitotic proteins in order to address this hypothesis.
Critically, one study examining the SAC during mitosis in mouse ESCs suggests that the SAC is also relaxed relative to somatic cells, and a higher proportion of ESCs can be found with polyploidy and aneuploidy. In somatic cells, aneuploidies are normally cleared from the cell cycle by apoptosis (24). The reason why pluripotent cells have these relaxed checkpoints seems paradoxical, given that the maintenance of genomic integrity would seem paramount to enabling high embryo quality. Thus, given that these checkpoints are inefficient, is there an alternate method for clearing genetically abnormal cells from the human embryo?
DIFFERENTIATION AS A MEANS FOR REMOVING ANEUPLOID CELLS FROM THE CELL CYCLE
Recently, an alternate checkpoint in ESCs was hypothesized that depends upon ESC differentiation (24). In somatic cells, drug-induced mitotic spindle breakdown and SAC activation result in exit from the cell cycle and apoptosis or senescence. In contrast, mouse and human ESCs have a functional SAC, but the checkpoint fails to prevent re-replication of aneuploid cells. This results in the chromosomally abnormal ESCs remaining in the cell cycle and resisting apoptosis (24). Interestingly, upon ESC differentiation to EBs, the polyploidy/aneuploid ESCs initiate caspase-dependent apoptosis of the chromosomally abnormal cells (24), therefore suggesting that differentiation is a mechanism that selects for diploid ESCs. Furthermore, upon ESC differentiation, p27 expression also appears to be globally restored in G1 (22) and presumably a longer G1 phase is induced in the differentiating cells which is more reflective of differentiation. To further support a molecular program in which the detection of DNA damage in ESCs is coupled with differentiation, the relationship between p53 and the pluripotent transcription factor NANOG was evaluated (25). In somatic cells, the activity of p53 in response to DNA damage results in cell cycle arrest or apoptosis. In contrast, ESCs do not immediately undergo apoptosis in response to DNA damage. Instead, it was determined that p53 bound the NANOG promoter, suppressing NANOG transcription and signaling differentiation (25,26). This study did not evaluate whether genetically damaged cells were subsequently removed from the cell cycle following suppression of NANOG. However, we would predict that, with differentiation and establishment of cell cycle checkpoints, damaged cells would exit the cell cycle.
Human embryo development requires the regulation of genomic stability and genomic quality via mechanisms that appear to be fundamentally different in somatic cells and embryonic cells/ESCs. Evidence suggests that embryonic cells and ESCs may have relaxed checkpoints that contribute to the accumulation of pluripotent cells with chromosomal aneuploidies (Fig. 2 showing chromosomal duplication). Interestingly, oocytes and hESCs are enriched in genes required for mitosis and cell division. Therefore, the expression of high levels of mitotic and cell cycle genes may be one mechanism for limiting chromosomal aneuploidies in pluripotent cells by ensuring that critical mitotic proteins required for sister chromatid separation are abundant. The model shown in Figure 2 suggests that, upon differentiation, the normal cell cycle checkpoints are re-established, and aneuploid cells may be targeted for removal from the cell cycle by well-described mechanisms and/or may fail to proliferate. This method for managing aneuploidy could result in a number of outcomes including a karyotypically normal embryo (46 chromosomes), spontaneous miscarriage of an aneuploid embryo (for example, duplications of chromosome 16) or a birth defect if the genetic abnormality is not repaired but the cell is still capable of differentiation (for example, trisomy of chromosome 21 or loss of an X chromosome). Therefore, the high degree of spontaneous miscarriages and birth defects in humans might be traced back to the unique relaxed regulation of DNA checkpoints in human pluripotent cells of the pre-implantation embryo. Many questions remain; however, understanding these checkpoints and the function of the oocyte-expressed transcripts in regulating pluripotent cell biology may significantly increase our understanding of the underlying causes of these disorders of human reproduction.
G.A. is supported by a training grant from the NIH/NICHHD UCLA Women’s Reproductive Health Research Center (5 K12 HD001281). A.T.C. is supported by a grant from the Lance Armstrong Foundation and STOP Cancer.
The authors would like to thank Sudip Mandel, Anne Lindgren, John Vincent and Mary Clark and for critical reading of this manuscript.
Conflict of Interest statement. None declared.