Genome Organization and Stability in Mammalian Pre-Implantation Development

Abstract

A largely stable genome is required for normal development, even as genetic change is an integral aspect of reproduction, genetic adaptation and evolution. Recent studies highlight a critical window of mammalian development with intrinsic DNA replication stress and genome instability in the first cell divisions after fertilization. Patterns of DNA replication and genome stability are established very early in mammals, alongside patterns of nuclear organization, and before the emergence of gene expression patterns, and prior to cell specification and germline formation. The study of DNA replication and genome stability in the mammalian embryo provides a unique cellular system due to the resetting of the epigenome to a totipotent state, and the de novo establishment of the patterns of nuclear organization, gene expression, DNA methylation, histone modifications and DNA replication. Understanding DNA replication and genome stability in the early mammalian embryo is relevant for studies of both normal and disease-causing genetic variation, and to uncover basic principles of genome regulation.

Introduction

DNA replication passes genetic information from cell to cell within the soma, and from generation to generation within the germ line. Genome integrity is not only consequential to the function of the gene products, but also to the decision of whether a cell exits the cell cycle, proliferates, or undergoes cell death. Genome duplication intrinsically carries the potential for instability, which refers to an increased risk of genetic changes. The impact of genome instability is influenced by the timing of genetic changes. Mammalian preimplantation embryo development can be generally divided into several stages: fertilization, ~4 cleavage divisions which reduce a very large cell to smaller size, compaction which begins cell type specification due to different signals between inner and outer cells, and blastocyst formation (Rossant 2018). At the cleavage stage, all cells, which are termed blastomeres, are totipotent, highlighting the profound effect when genetic variation is being introduced at this stage of development. Blastocysts consist of epiblast, which differentiate to all embryonic tissues including germline cells, and trophoblast, which becomes extraembryonic tissues. If genetic change occurs at the 1-cell stage, or as early as the first few cleavages, the consequences can be significant.

Despite the high risks of adverse outcomes in reproduction, the genome at the cleavage stage of some mammalian embryos is surprisingly unstable. Genome instability is, at least in part, conferred by DNA replication stress during the first cell divisions (Palmerola, Amrane et al. 2022). DNA replication stress, the slow or impeded progression of DNA replication forks, is an intrinsic characteristic of mammalian embryos. Though there are mechanisms to recover stalled forks at intrinsic or extrinsic obstacles, such as at G4 structures, transcription-replication conflicts, DNA crosslinks, spontaneous or induced DNA damage, they do not ensure that all stress events are inconsequential (Saxena and Zou 2022). Either due to a high load of stress, or due to inefficient repair, replication stress can result in incomplete replication of the genome, gaps and DNA breaks that can alter the genome profoundly, and impact development. In human embryos, the most visible consequences of DNA replication stress are chromosome breakages and aneuploidies (Palmerola, Amrane et al. 2022).

The exact cause of impaired DNA replication fork progression in mammalian embryos remains unknown, but the consequence is relatively clear: replication stress can be detrimental for development. At the chromosomal level, replication stress causes DNA damage and aneuploidy in embryos, which can result in an arrested cell cycle and subsequent cell death. This is evidenced by the high level of aneuploidy, attrition of blastomeres, and low developmental efficiency in humans and bovines (Palmerola, Amrane et al. 2022, Xu, Wang et al. 2024). Additionally, replication stress has been shown to be a source of DNA damage in stem cells, limiting stem cell function and promoting differentiation (Alvarez, Díaz et al. 2015, Sui, Xin et al. 2021, Do, Hsu et al. 2024).

Understanding DNA replication and genome stability during early mammalian development offers insights into both normal and abnormal embryo development, as well as the origins of genetic variation in the germline. In addition to aneuploidy, mechanisms that mitigate replication stress can be mutagenic: forks stalled at DNA damage sites can be rescued by translesion synthesis that can induce point mutations in human cells (Anand, Chiou et al. 2023). Recombinogenic events at stalled forks can induce copy number changes in eukaryotes (Larsen, Liberti et al. 2017). Genome instability and DNA replication have been extensively studied in mammalian somatic cells and lower vertebrate embryos, including zebrafish and Xenopus laevis and of course in lower eukaryotes, particularly in yeast. Though many basic principles apply to all eukaryotes, our knowledge of DNA replication at the earliest stages of mammalian development remains limited. This is largely due to the limited availability of cells. Recent advances in single-cell technologies have made the 1-cell and cleavage stage mammalian embryo accessible to the study of DNA replication.

In somatic cells, DNA replication follows a temporal pattern of progression correlating with gene expression, nuclear organization, and epigenetic modifications (Marchal, Sima et al. 2019). Early replication is associated with transcriptionally active regions in Drosophila, mouse, and human somatic cells (Schübeler, Scalzo et al. 2002, Woodfine, Fiegler et al. 2004, Hiratani, Ryba et al. 2008). This correlation suggests that DNA replication is co-regulated or determined by principles regulating transcription. Early replicating regions are enriched with lineage specific genes (Rivera-Mulia, Buckley et al. 2015). For example, in hESCs, genes involved in maintaining pluripotency, such as DPPA2 and DPPA4 are early replicating, and these genes are late replicating in other cell types. A telling study in yeast induced global DNA early replication by overexpressing replication factors, leading to disrupted transcription activity of more than a quarter of all genes in yeast. Normal gene expression could be restored by deleting replication origins to delay replication timing after the global early replication induction (Santos, Johnson et al. 2022). These studies indicate that replication timing can regulate transcription. Underlying both patterns of DNA replication and gene expression are common principles of genome regulation, including histone modifications and nuclear organization (Rhind and Gilbert 2013). The genome can be divided into two compartments, referred to as the A and B compartments. These compartments are identified through genome contact maps, which reveal stronger interaction within each compartment. The A compartment is referred to be the one enriched with open chromatin, while the B compartment generally contains more closed chromatin (Dekker, Marti-Renom et al. 2013). Both active transcription and early replication correlate with high chromatin accessibility and internal nuclear localization to the active (A) compartment (Corces, Granja et al. 2018, Vouzas and Gilbert 2021, Harris, Gu et al. 2023). Replication timing is thought to play a role in maintaining the epigenetic state. Disruption of the replication timing profile using RIF1 knockout results in global depletion of many histone markers such as H3K9me3 H3K27ac and H3K4me3 in human embryonic stem cells (Klein, Zhao et al. 2021). This result indicates that replication timing may serve to direct and/or maintain epigenetic modifications. However, the causal relationships between replication, transcription, histone modifications, DNA methylation, and chromatin organizations are difficult to dissect, in part because most studies involved cell types with already established patterns. In the early embryo, different dimensions of genome organization and regulation are newly established, thereby providing a unique model system.

The patterns of DNA replication are relevant to genome stability: Late replicating regions are prone to incomplete replication, chromosome breakage and to copy number change (Wilson, Arlt et al. 2015, Ji, Liao et al. 2020, Macheret, Bhowmick et al. 2020). Late replicating regions also show a higher rate of point mutations (Juan, Rico et al. 2013, Yaacov, Vardi et al. 2021). Studies on somatic cells do necessarily apply to the early mammalian embryo as there are numerous differences in all aspects of genome regulation, as epigenetic patterns are reset and the genome prepared to give rise to all different cell types. Surprisingly, patterns of genome stability are apparent at the earliest stages: the recent mapping of fragile sites in the early human embryo revealed a pattern of chromosome breakage as early as the first cell division, and prior to the major wave of embryonic transcription (Palmerola, Amrane et al. 2022). This points to patterns of DNA replication in early mammalian development. Here we will review our current knowledge on genome organization and epigenetic remodeling in the context of DNA replication and genome stability. We aim to discuss recent findings on DNA replication patterns and genome stability in the early mammalian embryo.

Genome re-organization in the mammalian embryo

In the first cell cycle after fertilization, the paternal and maternal genomes are prepared for the development into all cell types of an adult organism. The programming of the genome for these tremendous changes involves the new establishment of nuclear architecture, DNA methylation, histone modifications, gene expression, and DNA replication. In somatic cells, these different layers of genome regulation have already been established, and thus their study and experimental manipulation occurs in the context of preexisting patterns. The oocyte erases epigenetic patterns to establish a totipotent state, whether the genome is from an egg, a sperm, or from a differentiated somatic cell. The study of genome regulation in the early embryo thus provides insight into basic principles, including the temporal order and hierarchy of genome organization establishment.

Demethylation and re-establishment of the methylome of the genome is an important step of developmental reprogramming. Upon fertilization, both oocyte and sperm undergo demethylation to remove the parental epigenetic modifications and reprogram back to the totipotent state (Saadeh and Schulz 2014). The level of methylation remains low and starts re-methylation at the blastocyst stage. In humans and bovines, the genome is gradually demethylated, reaching the lowest methylation level at the blastocyst stage in the inner cell mass and the trophectoderm (Jiang, Lin et al. 2018, Zhu, Guo et al. 2018). In mice, the paternal genome is rapidly demethylated within the first cell cycle upon fertilization before and during the S phase, while the maternal genome retains methylation at the 1-cell stage (Fig. 1A) (Santos, Peat et al. 2013). Demethylation consists of active and passive pathways. Active demethylation removes 5-methylcytosine (5mC) by ten-eleven translocation (TET) (Wu and Zhang 2017). The passive pathway involves progressive loss of 5mC through DNA replication. The maternal genome in mice is gradually demethylated as cells replicate, following a trajectory similar to bovine embryos (Wang, Zhang et al. 2014).

Figure 1. Genome organization and replication timing pattern in mammalian embryos.

A) Schematic for genome methylation in mouse maternal and paternal genomes and in bovine embryos at indicated developmental stages. The timing of the onset of embryonic transcription is shown as solid bars for mouse and bovine embryos (Barnes and First 1991, Zeng and Schultz 2005). Mouse methylation data adopted from Guo et al (Guo, Li et al. 2017) and Mayer et al (Mayer, Niveleau et al. 2000). Bovine methylation data from Jiang et al (Jiang, Lin et al. 2018). B) Immunocytochemistry in mouse zygotes stained for H3K9me3. Maternal and paternal pronuclei are indicated. With permission from (Burton, Brochard et al. 2020). C) Schematic of mouse maternal and paternal genome A/B compartment strength at indicated developmental stages. Data from Du et al (Du, Zheng et al. 2017). The onset of embryonic transcription is shown as solid bars. D) Hi-C interaction maps in mouse preimplantation embryos and mouse embryonic stem cells (Du, Zheng et al. 2017). E) Replication timing profile at chromosome 3 for parthenogenetic and fertilized mouse zygotes. Lamina associated domains and A/B compartment from 1-cell stage embryos (Du, Zheng et al. 2017, Borsos, Perricone et al. 2019). Origin density is taken from mouse embryonic stem cells (Cayrou, Ballester et al. 2015).

The dynamics of histone modification patterns shows a similar asymmetric trajectory. The repressive histone mark H3K9me3 is lowest in zygotes and increases through preimplantation development to the blastocyst stage (Liu, Wang et al. 2016, Wang, Liu et al. 2018). The paternal genome in particular starts with very low levels of repressive histone marks H3K9me3, likely because histones are assembled de novo on the paternal genome after the exchange of sperm protamines. Extensive asymmetry persists for the entire first cell cycle (Burton, Brochard et al. 2020). (Fig. 1B). Likewise, there is extensive asymmetry in the levels of methylation between maternal and paternal genome in H3K27me3 and in H3K4me3, in particular at the zygote stage (Wang, Liu et al. 2018). Unlike in differentiated cells, histone modification marks at the cleavage stage are not repressive and compatible with gene expression (Burton, Brochard et al. 2020). These dynamic changes in histone marks and the asymmetry between maternal and paternal genomes can provide insight regarding the relevance of these marks in establishing DNA replication patterns during development.

The nuclear architecture is also established de novo in embryos, starting from the 1-cell stage. The Lamina-associated domains (LADs) are heterochromatic regions of the genome interacting with the nuclear lamina and nucleolus. LADs contribute to the 3-D organization of the chromatins inside the nucleus by anchoring chromatin to the nuclear membrane. LADs are associated with low transcription activity in various cell types such as human fibroblast and mouse embryonic stem cells (Guelen, Pagie et al. 2008, Peric-Hupkes, Meuleman et al. 2010). In addition, LADs also have the feature of low gene density, late replication and are enriched with the more closely packed inactive chromatin domains, namely the B compartment, in mammalian cells (van Steensel and Belmont 2017). In mouse embryos, LADs were detected on maternal and paternal genomes as early as the 1-cell stage, and their formation is dependent on H3K4me3 (Borsos, Perricone et al. 2019). Based on chromosome contact studies, segregation of A and B compartments were also detected in mouse embryos already at the 1-cell stage (Du, Zheng et al. 2017, Gassler, Brandão et al. 2017, Ke, Xu et al. 2017). However, patterns of nuclear architecture were gradually more defined as development progressed to the blastocyst stage (Du, Zheng et al. 2017, Ke, Xu et al. 2017). (Fig. 1C-D). In human embryos, essentially no A/B compartment distinction was found at the 2-cell stage and was only apparent at the 8-cell stage (Chen, Ke et al. 2019). Whether these differences reflect biological species-specific differences or technical challenges due to a scarcity of material is not currently known. Interestingly, in mice, the maternal genome showed less pronounced LADs and lower compartmentalization than the paternal genome in early S-phase (~6-9h post-fertilization) of 1-cell embryos, but became more similar to the paternal nucleus by the end of the first cell cycle (~15h post-fertilization) (Flyamer, Gassler et al. 2017, Gassler, Brandão et al. 2017, Borsos, Perricone et al. 2019). Another group similarly found compartmentalization on the paternal genome, and essentially none on the maternal genome in zygotes, and also found a greater compartmentalization after S-phase than during early S-phase (Du, Zheng et al. 2017). These congruent findings suggest that embryonic nuclear architecture is established during and after the first S-phase and solidifies further in the subsequent cell cycles.

All these extensive changes in genome regulation and the asymmetries between maternal and paternal genomes may impact the progression of DNA replication and impact genome stability. The reverse may also be true. DNA replication progression may affect nuclear architecture, DNA methylation, histone modifications and transcriptional patterns. Their temporal relation to each other can help determine causality.

DNA replication is patterned from the first cell cycle in mouse embryos on paternal and maternal genomes

In Drosophila melanogaster and Xenopus laevis embryos, DNA replication progresses with no apparent patterns before embryonic genome activation (EGA) (Hyrien and Méchali 1993, Sasaki, Sawado et al. 1999). In zebrafish, a pattern begins to emerge just prior to EGA, at 2.8h post-fertilization, when the embryo consists of several hundred cells (Siefert, Georgescu et al. 2017). However, this pattern is far less structured than in cells after the midblastula transition, beginning at the 10th cell cycle. No studies have examined DNA replication timing at the first cell division nor the ones immediately following it in these organisms. A unique hallmark of the initial cleavage divisions in lower vertebrates and invertebrates is the rapid expansion of the number of nuclei. Both zebrafish and Xenopus omit G1 and G2 phases before EGA, and cell cycles are less than 25 minutes (Siefert, Clowdus et al. 2015). The genome size of zebrafish and Xenopus is 1.4 GBp and 3.1 GBp respectively (Kimmel, Ballard et al. 1995, Hellsten, Harland et al. 2010). With the DNA polymerase progressing at ~1-2kbp/min, the short cell cycles of lower vertebrate embryos do not allow for replicons of several hundred kb as observed in differentiated cells. To enable such short cell cycles, replication origins are very closely spaced, only 10-30 kilobases apart (Lemaitre, Danis et al. 2005). Cell cycle regulation in mammalian embryos differs from lower vertebrates: the duration of the S-phase and of the cell cycle overall are similar to differentiated cells; the S-phase lasts approximately 6-7 hours, and the duration of the first cell cycle ranges from 15-24 hours, with an extended G2 phase of 6-10 hours. The normal cell cycle duration and a long G2 phase might allow for replicons with long traveling replication forks of dozens or even hundreds of kilobases.

Four recent studies now show that DNA replication patterns in mammalian embryos are established no later than the 4-cell stage (Halliwell, Martin-Gonzalez et al. 2024, Nakatani, Schauer et al. 2024, Takahashi, Kyogoku et al. 2024, Xu, Wang et al. 2024). As in differentiated cells, these patterns included distinct late and early replicating regions. Nakatani and colleagues compared late replication with histone modifications and gene expression and observed an increasing correlation of early replication with transcription after the 4-cell stage (Nakatani, Schauer et al. 2024). Early replication peaks correlated with H3K36me3, a mark indicative of active transcription. H3K9me3 were high in later replicating regions and H3K27me3 were lower at replication timing peaks, consistent with their generally more repressive nature. The authors concluded that replication timing profiles might be established through heterochromatin maturation from the 2-cell stage onward and after embryonic genome activation. These studies affirm an anticipated role of histone modifications in modulating DNA replication patterns. However, the relationship of histone marks to DNA replication timing may be context and cell type dependent. H3K9me3 can be associated with either late or early replication in embryonic stem cells, depending on other histone modifications (Ding, Edwards et al. 2021). In early embryos, the repressive histone marks are not repressive for transcription (Burton, Brochard et al. 2020) and may not be repressive for DNA replication either. DNA replication patterns correlated with LADs, but surprisingly, and unlike the establishment of LADs, did not require methylation of H3K4me3 (Borsos, Perricone et al. 2019, Nakatani, Schauer et al. 2024). The role of histone in shaping DNA replication patterns is therefore an important subject of further studies.

Takahashi and colleagues observed a distinct replication timing profile starting at the 4-cell stage (Takahashi, Kyogoku et al. 2024), and no replication pattern was observed at the 1-cell or the 2-cell stage. They found a strong correlation between late replication and the B compartment and concluded that replication timing pattern correlate with nuclear compartmentalization. Their data suggest that replication timing is established downstream of nuclear organization. An A/B compartment is present in mouse embryo at the 1-cell stage, as are LADs. However, the authors reported cytological patterns of DNA replication progression at the 1-cell stage consistent with this compartmentalization, contrasting with the conclusion that replication timing in mouse embryo is established abruptly at the 4-cell stage.

Halliwell et al. found that a replication timing program is established as early as the 2-cell stage (Halliwell, Martin-Gonzalez et al. 2024). Similar to Nakatani and Takahashi et al, no replication patterns were detected at the 1-cell stage. Due to the asymmetry of epigenetic modifications between the maternal and paternal genomes, the authors examined differences in replication timing patterns between the two parental genomes in mouse 2 cell stage embryos. Both parental genomes exhibit high levels of association between late replicating regions with LADs. Regions with oocyte deposited H3K27me3 were early replicating at the 2-cell stage in the maternal but not in the paternal genome. This appears to differ from Nakatani et al that drew a correlation between H3K27me3 and late replication.

Another study found that DNA replication timing patterns are well defined already at the zygote stage (Xu, Wang et al. 2024) (Fig. 1E). Late replicating regions correlated strongly with LADs and with the B compartment, while early replicating regions correlated with the A compartment. Replication patterns were detectable at the 1-cell stage on both maternal and paternal genomes. The time point used for analysis of DNA replication timing (~7-9h post-fertilization or activation) corresponds to the mid-pronuclear stage, a time point used in nuclear structure analysis when extensive asymmetries are observed. Embryos consisting of only a maternal genome, so-called parthenotes, also showed a temporal progression of DNA replication resembling that of fertilized embryos (Fig. 1E). This is all the more surprising as the maternal genome at the time of sampling shows barely any detectable nuclear organization, without segregation of the A and B compartments (Du, Zheng et al. 2017) (Fig. 1C, D). The emergence of DNA replication timing patterns therefore precedes or at least is concordant with the segregation into A and B compartments, and with LAD association. At the one-cell stage, DNA replication timing does not reflect the major asymmetries in both DNA methylation and histone modifications between maternal and paternal genomes. Though this is surprising, other studies have shown that late replication can persist without the heterochromatin marker H3K9me3 in cancer cells, indicating that epigenetic modifications alone do not necessarily affect replication timing (Spracklin, Abdennur et al. 2023). Furthermore, mutations in DNA methylases in cultured HAP1 cells showed no significant changes in DNA replication timing profile (Caballero, Ge et al. 2022). However, loss of DNA methylation has been shown to reduce replication timing precision and increase heterogeneity within a population of cells, in particular at the boundaries of early/late replication transition regions (Du, Smith et al. 2021). These, and the study by Xu et al are consistent with the notion that DNA replication timing is a remarkably resilient and basic feature of genome regulation.

The four studies on mouse embryos differed in how early replication patterns were observed, ranging from the 1-cell to the 4-cell stage. The differences between the four studies may be related to the timing of collection of the samples, or due to other technical differences. Capturing replication timing profiles at the 1-cell stage is challenging because results are impacted by the timing of cell cycle progression relative to fertilization, while collection later in development can rely on sufficient intrinsic asynchrony across a larger population of cells. Despite the discrepancy in the timing at which replication patterns were first observed, a consensus emerged on the correlation of late replication with nuclear organization in all four studies. Starting from the first cell cycle, late replicating regions localize to the lamina and the nucleolus (Ferreira and Carmo-Fonseca 1997, Nakatani, Schauer et al. 2024, Takahashi, Kyogoku et al. 2024). Whether only some zygotes show a temporal pattern of replication progression is possible in principle, and remains to be further investigated.

The underlying molecular mechanisms for the establishment of DNA replication patterns in the embryo remain elusive. Although epigenetic modifications do not appear to be the primary determinant, some studies point to epigenetic modifications ability to modulate the replication timing program. Upon a global DNA methylation loss, only about 3% of the genome was found to shift replication timing either from early to late or vice versa, despite extensive changes in epigenetic markers including H3K27me3, H3K4me3 and H3K9me3 (Du, Smith et al. 2021). This suggests that histone modification may not be fundamental for the establishment of the replication pattern. In addition, it has been shown that knock out of RIF1, which is known to regulate replication timing pattern, results in change in histone modification H3K27ac and H3K4me3 markers in human cells (Cornacchia, Dileep et al. 2012), indicating replication timing may act upstream of epigenetic patterns, or in a positive feedback loop. Consistent with this finding, replication timing patterns facilitate maintenance of epigenetic states (Klein, Zhao et al. 2021). Other determinants of replication timing may be transcription factors. In studies on SNPs altering OCT4 binding sites, higher binding affinity shifts timing to earlier replication (Ding, Edwards et al. 2021). Interestingly however, replication timing of embryos resembles embryonic stem cells despite great differences in transcription factor expression (Guo, Li et al. 2017, Takahashi, Kyogoku et al. 2024, Xu, Wang et al. 2024). For instance, Oct4, Sox2, Nanog, are missing in 2-cell embryos, but expressed in ES cells. Overall, although both epigenetic modifications and transcription factors affect DNA replication timing patterns, individually, they are not likely the key determinant for the establishment of a replication timing profile.

Cell cycle progression in mammalian embryos resembles that of embryonic stem cells and includes a lack of a true G1 phase where cells can pause, and a very long and extendable G2 phase. This differs from the rapid cell cycles of lower vertebrates, as well as from differentiated somatic cells which can arrest in G1. The long G2 phase in mammals can accommodate a DNA replication timing program, with late replicating regions completing in G2 phase. Together with segregation into A/B and LADs, DNA replication timing appears to be one of the first and most basic layers of genome organization during mammalian development, preceding transcription and the segregation of cell lineages. The early emergence of a structured cell cycle and of DNA replication patterns is one of the distinguishing features of mammals as opposed to lower vertebrates, including fish and amphibians. Whether reptilians and birds show DNA replication patterns like in mammals or are more analogous to fish and amphibians is not currently known. In avian embryos, each cell cycle is approximately 20 minutes which is comparable to cell cycles of zebrafish and Xenopus (Zwaan and Pearce 1970, Gelens, Huang et al. 2015, Adar-Levor, Nachmias et al. 2021). Thus, it is reasonable to speculate that the presence of replication timing patterns from the first cell cycle is a mammalian specific feature.

Patterns of genome instability are established from the first cell cycle in mammals in correlation with DNA replication timing patterns

Human as well as bovine embryos show frequent spontaneous chromosome breakages and aneuploidy acquired post-fertilization (Destouni, Zamani Esteki et al. 2016, Cavazza, Takeda et al. 2021, Palmerola, Amrane et al. 2022, Xu, Wang et al. 2024). These breakages result from intrinsic replication stress and occur independent of transcription (Palmerola, Amrane et al. 2022). Mapping of spontaneous chromosomal breakages in human and bovine embryos revealed that the segmental chromosomal breakages follow a pattern: most break sites localize to genomic regions with low gene density or contain genes encoding very long transcripts. Others yet localize to neuronal gene clusters, such as olfactory gene clusters (Palmerola, Amrane et al. 2022, Xu, Wang et al. 2024). Because of this shared feature of frequent spontaneous aneuploidy and chromosome breakage, bovine embryos offer a suitable model to study mammalian genome instability. Functional annotation of genes located at spontaneous break sites in human and bovine embryos reveals genes involved in neuronal development and function. For example, a recurring break site shared by bovine and human embryos is located at the gene Dipeptidyl peptidase 10 (DPP10). Mutation in this gene results in a higher risk of Alzheimer’s and other neurodegenerative diseases (Chen, Gai et al. 2014, Xu, Wang et al. 2024). Other overlapping genes include FHOD3 and KCNMA1, which both are in common fragile sites and are crucial for neurodevelopment (Tabarki, AlMajhad et al. 2016, Sulistomo, Nemoto et al. 2019). The overlap of embryo break sites with common fragile sites and with sites of mitotic DNA synthesis in somatic cells (Ji, Liao et al. 2020, Macheret, Bhowmick et al. 2020) is notable, as it points to conserved principles in the patterns of genome stability and replication starting from the first cell cycle and throughout development and cell differentiation.

Similar to mice, DNA replication timing patterns in bovine embryos are established prior to embryonic genome activation and show late replication in gene-poor regions and neuronal gene clusters (Xu, Wang et al. 2024). Spontaneous chromosome breakage was enriched for late replicating regions. Thus, there is a pattern of genome stability connected to DNA replication patterns in bovine embryos. DNA replication timing patterns are not yet available in human embryos, but it is plausible that these principles also apply to human embryos. The spontaneous aneuploidy rate in bovine embryos derived from in vitro matured oocytes ranges from 40% to 85%, which is comparable to human (Destouni, Zamani Esteki et al. 2016, Hornak, Kubicek et al. 2016, Tšuiko, Catteeuw et al. 2017, Palmerola, Amrane et al. 2022).

Unlike human embryos, mouse embryos do not commonly show spontaneous chromosomal breakages. The spontaneous aneuploidy rate in mouse embryos is less than 1% during the first two cell divisions, compared to over 50% in humans (Bolton, Graham et al. 2016, Palmerola, Amrane et al. 2022). The robustness of mouse embryos offers a suitable model to understand species differences and potentially uncover mechanisms that protect genome integrity more effectively in mice. The low spontaneous aneuploidy rate in mice also provides a model for experimental manipulation. Low concentrations of aphidicolin have been used in somatic cells to slow DNA replication progression and induce chromosomal breakages at sites of unreplicated DNA (Glover, Berger et al. 1984). This kind of treatment exposes genomic regions of low origin density, and with long-traveling replication forks that are limiting to DNA replication completion (Hosseini, Horton et al. 2013). When mouse embryos are exposed to low levels of aphidicolin to slow DNA replication during S-phase, chromosome breakage is common (Xu, Wang et al. 2024). Chromosomal breakages identified after such treatments show association with low gene density, late DNA replication timing, as well as LAD and B compartment association. This pattern of differential genome stability is established already during the first cell cycle prior to embryonic genome activation. Breakages of the same pattern can also be induced in human embryos in the first cell cycle, indicating a conservation of basic principles and conservation of patterns of genome stability in different mammalian species (Palmerola, Amrane et al. 2022).

Though spontaneous aneuploidies in the mouse embryo are rare, sites of chromosomal damage and repair are common as early as the first G2 phase. EdU-pulse staining shows late DNA replication located around the nuclear lamina or nucleolus (Fig. 2A). Sites of DNA repair show a concordant cytological pattern (Fig. 2B): the majority of the foci for the DNA double-strand break marker γH2AX are associated with either nuclear lamina or nucleolus (Fig. 2C). In somatic cells, very late replicating regions are prone to chromosomal breakages due to the scarcity of replication origin at those sites, which also tends to be gene deserts (Cadoret, Meisch et al. 2008). In mouse embryos, the genomic location of chromosomal break sites reveals a correlation with lamina association, B compartment, low origin density as measured in ESCs, and with low gene density (Xu, Wang et al. 2024). Remarkably, break sites identified from maternal/paternal-only mouse embryo exhibit identical genomic features of late replication, lamina and B compartment association (Fig. 2D). Thus, the pattern of DNA replication in the early embryo predisposes the same regions to fragility on both paternal and maternal genomes. This concordance is independent of EGA and occurs despite major asymmetries in DNA methylation and histone modifications.

Figure 2. Chromosomal fragile sites are patterned and correlate with replication timing.

A) Time course of DNA replication through EdU incorporation for mouse zygotes. Maternal(M) and paternal (P) pronuclei are indicated. Figure adapted from Nakanati et al (Nakatani, Schauer et al. 2024). B) Immunostaining for γH2AX and LaminB of fertilized zygotes at late G2 phase. Arrows indicate foci localized on the nucleolus and arrowheads indicate foci localized on the nuclear lamina. Figure adapted from Xu et al (Xu, Wang et al. 2024). C) Schematic of the spatial location of late and early replicating regions. Yellow indicates early replication and blue indicates late replication. Percentages refer to the location of γH2AX foci. D) Replication timing profile, gene density, lamina associated domains (LAD), and A/B compartment (A/B) in fertilized zygotes. Vertical bars are annotated chromosomal break sites that are caused by DNA synthesis and repair in the G2 phase from maternal (red) and paternal (maroon) genomes.

The association between late replication with genome instability motivates the study of replication timing patterns in mammalian embryos. It has been shown in yeast that late replicating regions correlates with higher mutation rate, and deletion of replication origin leads to increased mutation (Lang and Murray 2011). In somatic cells, common fragile sites are hotspots for mutations (Durkin and Glover 2007, Nesta, Tafur et al. 2021). This increased level of genome instability in late replicating regions does not necessarily mean higher incidence of DNA damage or breakage, but could also be due to inefficient repair, a lack of converging replication forks, or reduced time for repair prior to mitotic entry. Mammalian zygotes have few dormant origins, and thus regions of a relative paucity of origins are at greater risk of incomplete replication and repair. Origin poor regions may also show greater dependence on inaccurate mechanisms of replication fork recovery, contributing to higher mutagenesis. This is particularly relevant for the early embryo due to the high level mutation rate in the first few cleavage divisions (Manders, van Boxtel et al. 2021). The early patterns of replication and genome stability may have profound implications for genome evolution in mammals.

Perspective

The early embryo incurs spontaneous DNA replication stress, characterized by slow replication fork progression, replication fork stalling and foci indicative of DSB repair (Palmerola, Amrane et al. 2022). In some, but not all mammalian species, replication stress results in frequent chromosomal breakage and aneuploidy. Mice are resistant to spontaneous chromosome breakage and aneuploidy, but when induced, the pattern of breakage is concordant with the spontaneous breakages in human and bovine embryos (Palmerola, Amrane et al. 2022, Xu, Wang et al. 2024). The early establishment of a DNA replication pattern in the mammalian embryo confers a pattern of genome stability from the first cell cycle. Chromosomal aneuploidies can result in the loss of the cell or the embryo and thus provide a reproductive disadvantage. Why this property is conserved across the mammalian species is an enigma. In accordance with the Aristotelian principle that “nature does nothing without use”, there may also be a function of genetic change in reproduction, albeit at the cost of frequent embryo loss and the risk of genetic disease.

DNA replication stress may lead to genetic changes evidenced by chromosomal aneuploidies, as well as more subtle changes in the genome including point mutations. Somatic mosaicism provides a means to infer the timing of mutation acquisition by identifying variant allele frequency (VAF). For a variant to have high VAF, it needs to be present in multiple cell populations and tissues and is thus inferred to occur early during development. One study identified 8 variants that were present in all 15 sequenced organs from a fetus of 21 weeks gestational age. This and other somatic mosaicism studies suggest that in early embryos, cell divisions frequently incur point mutations with an average of 2-5 mutations per cell per cell division, compared with ~1 mutation per cell division later in development (Bae, Tomasini et al. 2018, Park, Mali et al. 2021, Spencer Chapman, Ranzoni et al. 2021). The proportion of somatic mosaicism is linked to the severity of symptoms. For example, a recent case study of asymmetric congenital myopathy shows patients with milder symptoms when the level of mosaicism is lower, as compared to the more severe case in which the mutation is constitutive (Lehtokari, Sagath et al. 2024). Additionally, since the onset of certain diseases, such as neurodegenerative diseases, are tissue-specific, an early incidence of mutation that leads to somatic mosaicism may have a high risk of being disease-causing (Iourov, Vorsanova et al. 2009, Sherman, Rodin et al. 2021). Thus, the study of DNA replication and genome stability and DNA replication in the cleavage stage embryo is relevant to genetic disease.

Despite the recent advances, there is still a significant gap in our knowledge regarding the molecular mechanisms of the replication patterns. A comprehensive longitudinal analysis on differences and how correlation with other genomic features, such as epigenetic modification and 3-D genome organizations, may change overtime has not yet been performed. Additionally, studies to demonstrate causal relationships with those genomic features have not been conducted. Another pressing question is the cause and the consequences of replication stress. Several mechanisms have been proposed but remain untested in embryos, such as DNA demethylation, nucleoside depletion and active replication origin exhaustion. Furthermore, the consequence of replication stress in embryos has not yet been studied at the structural variant and point mutations level. And lastly, genome stability and resulted genetic changes in other mammals such as bat, cetaceans and other non-human primates can provide a more comprehensive view of the role of embryo replication stress in reproduction and evolution in mammals.

Highlights.

• DNA replication is patterned in the early mammalian embryo

• Genome stability and replication timing are linked to nuclear organization

• Patterns of replication predispose neuronal regions to breakage from the first cell cycle