Introduction

Humans have relatively low fertility compared to other mammals, with only ∼30% of conceptions resulting in live birth (Palmerola et al., 2022a) Capalbo et al., 2021). A striking feature of early human development is the high frequency of chromosome segregation errors during cleavage stage divisions, leading to aneuploidy – the gain or loss of chromosomes (Allais & FitzHarris, 2022; Currie et al., 2022; Palmerola et al., 2022b; Pauerova et al., 2020; Vanneste et al., 2009). This high incidence of aneuploidy is thought to underlie low human fecundity and many developmental defects (Palmerola et al., 2022a). Both in vivo and in vitro fertilization (IVF) frequently results in embryos that are chromosomally mosaic, containing a mixture of diploid and aneuploid cells. It is estimated that ∼60% of pre-implantation IVF embryos exhibit this form of mosaicism (Capalbo et al., 2021). Despite its prevalence, our understanding of how embryos respond to and cope with aneuploidy during early development remains limited.

The incidence of aneuploidy declines at development progresses (Shahbazi et al., 2020; Yang et al., 2021), but the mechanism underlying this decline remains unclear. Remarkably, mosaic human embryos can develop to term (Capalbo et al., 2021; Greco et al., 2015; Starostik et al., 2020). Specifically, embryos classified as low- or medium-grade mosaics - defined by presence of 20-30% or 30-50% aneuploid cells in the trophectoderm – have implantation and birth rates comparable to fully euploid embryos (Capalbo et al., 2021). These observations raise a fundamental question: how do some mosaic embryos maintain developmental potential despite chromosomal abnormalities? Uncovering the mechanisms that support the viability of mosaic embryos is essential for improving reproductive outcomes and embryo selection strategies.

Mouse models of chromosome mosaicism provide a powerful system to investigate mechanisms that cannot be ethically studied in human embryos. Importantly, mouse and human pre-implantation development are highly similar: both undertake cleavage divisions, compaction, blastocyst cavity formation and hatching, albeit with slightly different timings (Mole et al., 2020; Zhu et al., 2021). During this period, outer cells differentiate into the extra-embryonic trophectoderm (TE), whereas inner cells form the inner cell mass (ICM), which further segregates into the embryonic epiblast (EPI) or extra-embryonic primitive endoderm (PE). The TE will form the placenta, the PE will form the yolk sac and the EPI will form the fetus (Zhu & Zernicka-Goetz, 2020).

Recently, our group developed the first mouse model of chromosome mosaicism by inducing aneuploidy using the small-molecule inhibitor reversine (Bolton et al., 2016; Singla et al., 2020). Reversine is a pan-Aurora kinase inhibitor that also antagonizes the A3 adenosine receptor and inhibits mitotic kinase monopolar spindle 1 (MPS1) (D’Alise et al., 2008; Santaguida et al., 2010). We found that reversine-treated mosaic embryos exhibit widespread aneuploidy and upregulation of p53 (Bolton et al., 2016). Importantly, aneuploid cells in mosaic embryos are progressively eliminated from the EPI, beginning around the time of implantation. Consistent with findings in human embryos (Capalbo et al., 2021), murine mosaic embryos containing at least 50% of euploid cells had a similar developmental potential to fully euploid embryos (Bolton et al., 2016; Singla et al., 2020).

However, because reversine affects p53 and may compromise cellular fitness (D’Alise et al., 2008; Santaguida et al., 2010), we sought to establish a complementary model using a more specific Mps1 inhibitor, AZ3146 (Hewitt et al., 2010). Although both AZ3146 and reversine interfere with the spindle assembly checkpoint, they bind to Mps1 in a different manner (Lan & Cleveland, 2010). In mouse embryos, AZ3146 treatment was shown to double the occurrence of micronuclei, indicative of chromosome segregation defects, but without impairing overall cellular fitness (Vazquez-Diez et al., 2019). In this study, we use both AZ3146- and reversine-treated embryos to dissect the mechanisms governing aneuploid cell elimination and survival during pre-implantation development.

Results

AZ3146 treatment induces chromosome segregation defects in pre-implantation embryos

To generate distinct models of aneuploidy, we treated 4- to 8-cell stage mouse embryos with AZ3146 (20 µM) (Vazquez-Diez et al., 2019), as well as with reversine (0.5 µM) (Bolton et al., 2016) as a positive control, or DMSO (vehicle) as a negative control (Fig. 1A). We evaluated how the different Mps1 inhibitors affect chromosome segregation by detecting nuclei and kinetochores in 8-cell embryos and counted chromosomes in situ (Pauerova et al., 2020). Micronuclei were identified as small DAPI-stained chromosomes that were clearly distinct from the nuclei (Fig. S1A). We examined 32 DMSO-treated control embryos and observed only 22 cells with micronuclei out of a total of 256 cells (Fig. S1B). In contrast, we observed 82 cells with micronuclei in a total of 144 individual cells from 18 reversine-treated embryos, and 182 cells in a total of 304 individual cells from 38 AZ3146-treated embryos (Fig. S1A). After normalization, the average of aneuploidy in DMSO-treated 8-cell embryos was 12.5%, compared to 75% in the reversine-treated embryos and 62.5% in the AZ3146-treated embryos (Fig. S1B). We also detected non-dividing nuclei as rounded circles with distinct DAPI intensities in embryos treated with reversine and AZ3146. Specifically, we observed 28 non-dividing cells (19%) in reversine-treated embryos and 24 non-dividing cells (7.9%) in AZ3146-treated blastomeres, but none in DMSO-treated blastomeres (Fig. S1B). The increased frequencies of micronuclei and non-dividing cells in response to Mps1 inhibition are consistent with elevated aneuploidy and chromosomal instability (Daughtry and Chavez, 2016; Vazquez-Diez et al., 2016), as we showed previously for reversine-treated embryos (Bolton et al., 2016).

Figure 1

Given these distinct responses to reversine and AZ3146, we examined how these treatments impact the lineages in the blastocyst. To this end, we treated embryos with DMSO, reversine or AZ3146 from the 4- to 8-cell stage, and then washed and cultured them until the late blastocyst stage (E4.5). We assessed lineages by immunofluorescence (IF) to detect the TE marker CDX2, the EPI marker NANOG, and the PE marker SOX17. We found that all lineages segregated normally, and that blastocyst morphology was similar in all three conditions (Fig. 1B). We quantified the number of cells in the blastocysts and found that DMSO-treated controls had a median of 93 cells in total, whereas reversine- and AZ3146-treated embryos had a median of only 79 and 82 cells, respectively (n=28 embryos per treatment collected from three independent experiments. (Fig. 1C). DMSO-treated blastocysts contained a median of 15 EPI, 11 PE, and 67 TE cells, whereas reversine-treated blastocysts contained 12 EPI, 8 PE and 59 TE cells, and AZ3146-treated blastocysts had 11 EPI, 7 PE and 64 TE cells (Fig. 1C). These data suggest that reversine treatment compromises the development of all three lineages, as observed previously (Bolton et al., 2016), whereas AZ3146 mainly compromises ICM (EPI and PE) development (Statistical test: Mann–Whitney U-test).

To compare the developmental potential of blastocysts that had been treated with AZ3146 or reversine, we transfer ∼7 embryos per treatment into opposite uterine horns of the same mouse and counted decidua, which reflect successful implantation, as well as viable embryos at E9.5. Although two (8%) reversine-treated blastocysts developed decidua (Fig. S1C), none gave rise to viable E9.5 embryos. In contrast, seven (30%) AZ3146-treated blastocysts developed decidua (Fig. S1C) and five (21%) generated viable E9.5 embryos (Fig. S1D). Thus, AZ3146-treated embryos appear to have a higher developmental potential than reversine-treated embryos.

Considering that cellular fitness and aneuploid stress are related to DNA damage and DNA repair, we first sought to understand if these parameters were affected in our drug treatments. We used IF to quantify the levels of the DNA repair marker PARP1 and of the DNA damage marker, phosphorylated H2A.X (γH2A.X) at the 8-cell and morula stages (Fig. 1D). We observed elevated γH2A.X levels in reversine- and AZ3146-treated 8-cell embryos compared to controls (Fig. 1E) (n=∼15 embryos per treatment, ∼120 cells. Collection was made during three independent experiments. Statistical test: t-test), which eventually returned to normal in AZ3146-treated blastocysts, but not in reversine-treated blastocysts (Fig. S2A).

PARP1 was specifically enriched in the EPI lineage in normal blastocysts, as assessed by IF (n>21 embryos per treatment collected from three independent experiments Statistical test: Mann–Whitney U-test) and re-analysis of published scRNA-seq data (Deng et al., 2014) (Fig. S2B). Intriguingly, PARP1 levels were reduced at the 8-cell and morula stages in reversine-treated embryos compared to controls and to AZ3146-treated embryos (Fig. 1 D-E). Moreover, late morula stage embryos treated with the PARP inhibitor Olaparib (10 µM) (Hou et al., 2022; Prasad et al., 2017) developed into smaller blastocysts with only 78 total cells, reflecting a relatively low number of ICM cells, only 8 EPI and 4 PE cells in DMSO-treated embryos (Fig. S2C-E). Treatment with reversine, but not AZ3146, further reduced the number of cells in the EPI lineage of Olaparib-treated embryos (Fig. S2E) (n=15 embryos per treatment collected from three independent experiments. Statistical test: Mann–Whitney U-test). Overall, these data suggest that PARP1 is required for proper development of the ICM, and that its reduced levels after reversine treatment may negatively impact development.

Reversine and AZ3146 activate distinct stress response pathways in pre-implantation embryos

Chromosome mis-segregation and aneuploidy are associated with different cellular stress pathways (Zhu et al., 2018). For instance, p53 is frequently activated following DNA damage, which can limit proliferation and trigger apoptosis (Abuetabh et al., 2022; Soto et al., 2017; Thompson & Compton, 2010). The p38 mitogen-activated protein kinase (MAPK) can also be activated following DNA damage (Thompson and Compton, 2010). In aneuploid cells, p38 promotes apoptosis by inhibiting the transcription factor Hif-1α (Simoes-Sousa et al., 2018), which otherwise promotes cell survival, proliferation, and metabolic changes in aneuploid cells and in response to hypoxia in different contexts (Hu et al., 2003; Simoes-Sousa et al., 2018).

To investigate how treatment with reversine or AZ3146 affect p53 and Hif1a expression in mouse embryos, we performed RT-qPCR at the morula and blastocyst stages. We normalized to Ppia (Peptidylprolyl Isomerase A) mRNA, which is a stable reference gene in diploid and polyploid embryos (Gu et al., 2014). Reversine-treated embryos displayed a significant increase in p53 transcript levels at the morula (7-fold) and blastocyst (1.3-fold) stages compared with DMSO-treated embryos (Sample: 3 biological replicates and 2 technical replicates per experiment with each replicate having a minimum of 16 embryos. Statistical test: Welch’s t-test) (Fig. 2A), as we showed previously (Singla et al., 2020). In addition, reversine-treated embryos showed reduced expression of Hif1a at the blastocyst stage, which would be consistent with p38 activation. In contrast, AZ3146-treated embryos did not show upregulation of p53 at either the morula or blastocyst stage (Fig. 2A). Moreover, AZ3146-treated embryos showed a transient increase in Hif1a mRNA levels (3-fold) at the morula stage compared to DMSO- and reversine-treated embryos (Fig. 2A), which returned to normal levels at the blastocyst stage. HIF1A protein was present in DMSO- and reversine-treated embryos and elevated at the morula and blastocyst stages in AZ3146-treated embryos (n=∼20embryos per treatment. Statistical test: Mann–Whitney U-test, error bars represent s.e.m) (Fig. 2B). HIF1A appeared to be mostly nuclear in morula, but mostly cytoplasmic in blastocysts, under all three conditions (Fig. 2B). Overall, these data suggest that treatment with reversine, but not AZ3146, induces multiple stress pathways in pre-implantation embryos. These differences may contribute to the increased developmental potential of embryos treated with AZ3146 versus reversine.

Figure 2

HIF1A activity is required for proper blastocyst formation after Mps1 inhibition

It was previously shown that Hif1α^−/− embryos undergo developmental arrest and lethality by E11 (Iyer et al., 1998). To assess the role of HIF1A in the embryo’s response to reversine and AZ3146, we used a pharmacological approach to inhibit its function. Briefly, we tested two different small molecules that have been shown to inhibit HIF1A activity, PX-478 (Lee & Kim, 2011; Zhao et al., 2015) and IDF-11774 (Ban et al., 2017). We treated control zygotes with DMSO, PX-478 (2 µM), or IDF-11774 (20 µM) until the blastocyst stage (Fig. S3A). Unlike PX-478, IDF-11774 treatment did not significantly affect the total number of cells in the TE or the whole blastocyst compared to the control (n=∼12 embryos per treatment collected from three independent experiments. Statistical test: Mann–Whitney U-test) (Fig. S3B). Therefore, we used IDF-11774 to inhibit HIF1A in subsequent experiments.

Next, we treated embryos with DMSO, reversine, or AZ3146 from the 4- to 8-cell stage, washed them, and then inhibited HIF1A with IDF-11774 from the 8-cell to blastocyst stage (Fig. 2C). Inhibition of HIF1A did not abolish cavitation but dramatically lowered the number of TE and especially PE cells in AZ3146-treated embryos compared to DMSO-treated controls (Fig. 2D) and compared to AZ3146-treated embryos not exposed to IDF-11884 (Fig. 1C). Inhibition of HIF1A also lowered the number of cells in reversine-treated embryos compared to controls, but the effect was smaller (n=20 embryos per treatment collected from three independent experiments. Statistical test: Mann–Whitney U-test) (Fig. 2D). Overall, our data suggest that elevated HIF1A activity from the 8-cell stage onwards is particularly important to promote the survival of aneuploid TE and PE cells after AZ3146 treatment.

Taken together, our data suggest that AZ3146 and reversine have distinct effects on the pre-implantation embryo. Compared to reversine-treated embryos, AZ3146-treated embryos appear to have increased developmental potential, lack upregulation of p53, and show transient upregulation of Hif1a. Moreover, AZ3146-treated embryos have a greater dependence on HIF1A activity to form the TE and PE.

Hypoxia exposure attenuates DNA damage and blastomere defects in response to Mps1 inhibition

Pre-implantation development occurs in a hypoxic environment, and many clinics culture human embryos under hypoxic conditions (5%) (Houghton, 2021). It was reported that physiologic oxygen concentration (∼5% oxygen) can improve the yield and quality of mammalian blastocysts (Ciray et al., 2009; Nguyen et al., 2020) and increase the nuclear translocation of HIF1A in mouse blastocysts (Choi et al., 2021) To investigate how hypoxia affects the development of euploid and aneuploid mouse embryos, we repeated our experiments performed under normoxia (20% O₂/5% CO₂; control) (Fig. 1A) in hypoxic (5% O₂/5% CO₂) conditions taking into consideration the standard practices in the IVF clinics (Knudtson et al., 2022) (Fig. 3A). We cultured embryos from the 2-cell stage until the blastocyst stage, treating them with DMSO, reversine, or AZ3146 from the 4-8 cell stage.

Figure 3

First, we examined lineage specification by quantifying the cell numbers in all three lineages at the blastocyst stage. We found that DMSO-treated blastocysts cultured under hypoxia had a median of 74 cells, representing 14 EPI cells, 6 PE cells and 54 TE cells (n=18 embryos per treatment collected from three independent experiments. Statistical test: Mann–Whitney U-test)(Fig. 3B-C). These data suggest that, under our conditions, the cell number in the pre-implantation mouse blastocyst is reduced when cultured in hypoxia compared to normoxia (Fig. 3C, 1C), particularly in the TE and PE. Notably, reversine and AZ3146 treatment did not further reduce cell numbers for embryos cultured under hypoxia, except for in the PE (Fig. 3C).

To investigate how hypoxia exposure affects DNA damage and repair in mouse embryos, we again performed IF for γH2A.X and PARP1 in 8-cell embryos and morula (Fig. 3D-E) (n=25 embryos per treatment collected from three independent experiments. Statistical test: t-test). We found that treated embryos with reversine and AZ3146 under hypoxia generated lower levels of γH2A.X (Fig. 3E) compared with the same treatments under normoxia (Fig. 1E). Whereas PARP1 levels at the 8-cell stage are reduce in reversine-treated embryos (Fig. 3E), similar to normoxic conditions (Fig. 1E). Importantly, at the morula stages, the levels of PARP1 are relatively low compared to morula stage in normoxic conditions (Fig. 1E). Overall, these data suggest that hypoxia exposure lowers the accumulation of DNA damage between the 8-cell to morula stages, and results in a reduction of PARP1 at the morula stage.

Hypoxia increases the proportion of euploid cells in the epiblast of mosaic blastocysts

Despite the high incidence of mosaicism in human pre-implantation embryos, the fate of aneuploid cells in mosaic embryos is incompletely understood. We previously used reversine to generate a mouse model of pre-implantation chromosome mosaicism and found that aneuploid (Reversine-treated) cells are eliminated in the EPI of mosaic embryos via apoptosis, starting from the mature blastocyst stage (Bolton et al., 2016). Whether this response reflects reversine treatment specifically, or aneuploidy more generally, is not known. Moreover, how hypoxia versus normoxia affects the outcomes is also not clear.

To address these questions, we created aggregation chimeras at the 8-cell stage that contained a 1:1 ratio of AZ3146-treated and control blastomeres (DMSO/AZ3146 chimeras), which is expected to reflect low-grade mosaicism, or of AZ3146-treated and reversine-treated blastomeres (reversine/AZ3146 chimeras), which is expected to reflect medium-grade mosaicism (Fig. 4A). We cultured these chimeric embryos in normoxia and hypoxia, and followed the fate of individual blastomeres by using transgenic mouse lines with the membrane markers mTmG (Muzumdar et al., 2007) and E-cadherin (Christodoulou et al., 2019). DMSO/AZ3146 and reversine/AZ3146 blastocysts displayed proper lineage allocation, embryo morphology and cavitation in both normoxia and hypoxia (Fig. 4B-C). In addition, DMSO/AZ3146 and reversine/AZ3146 embryos had comparable total cell numbers in their blastocysts, and blastocysts grown in hypoxia were again smaller than those grown in normoxia (Fig. 4D-E). Intriguingly, despite having fewer total cells, DMSO/AZ3146 blastocysts had more EPI cells when they were cultured in hypoxia compared to normoxia (Fig. 4D-E).

Figure 4

We quantified the proportion of AZ3146 cells in each lineage for each chimera. In DMSO/AZ3146-treated blastocysts, we found that 46.75% of the TE and 42.88% of the EPI originated from AZ3146-treated blastomeres, compared to only 28.57% of the PE (Fig. 4F). In reversine/AZ3146 chimeras, 63% of the TE and 78.4% of the EPI originated from AZ3146-treated cells, compared to only 40% of the PE (Fig. 4F). Thus, under normoxia, DMSO cells appear to outcompete AZ3146-treated cells, which in turn outcompete Reversine-treated cells in the TE and EPI. AZ3146-treated cells appeared to be at a competitive disadvantage in the PE in both contexts. Culturing DMSO/AZ3146 chimeras in hypoxia increased the representation of AZ3146-treated blastomeres in the TE (49.06%) and PE (50%) but, strikingly, reduced their representation in the EPI (33.3%) (n>7 embryos per treatment collected from three independent experiments. Statistical test: Mann–Whitney U-test) (Fig. 4G). These altered frequencies reflect a higher number of DMSO-treated cells in the EPI and a lower number in the PE (Fig. 4E). Hypoxia exposure of reversine/AZ3146 embryos lowered the contribution of AZ3146-treated blastomeres to the TE (55.94%) and EPI (50%) but not the PE (46.43%) (Fig. 4G). We found that there was no correlation between the proportion of AZ3146-treated cells in the TE and the EPI in DMSO/AZ3146 and reversine/AZ3146 aggregation chimeras, under hypoxia or normoxia, and that hypoxia seemed to have a greater impact on the proportion of AZ3146 cells in reversine/AZ3146 aggregation chimeras (Fig. S4). Overall, these data suggest that hypoxia has lineage-specific effects on competitions between aneuploid and euploid cells and increases the contribution of euploid cells to the EPI.

Blastomeres in the 4-cell stage embryo display a lineage bias (Goolam et al., 2016). We considered that mosaicism generated before the 4-cell stage might influence lineage allocation. To test this possibility, we treated zygotes with reversine or AZ3146 during the first cleavage division. Importantly, reversine treatment at the zygote stage seems to strongly affect the morphology of the blastocysts (Fig. S5A). Consistent with this change in morphology, reversine-treatment reduced the number of cells in all lineages, particularly in the TE and PE (Fig. S5B). In contrast, AZ3146 treatment did not affect morphology or cell number in any of the lineages (Fig. S5A-B) (n=∼27 embryos per treatment collected from three independent experiments. Statistical test: Mann–Whitney U-test). To evaluate how early generation of aneuploidies in the embryos affect cell competition, we generated DMSO/AZ3146, Reversine/DMSO, and Reversine/AZ3146 aggregation chimeras at the 2-cell stage and cultured them under normoxia until the blastocyst stage (Fig. S5C). We found that all three 2-cell stage aggregation chimeras developed into blastocysts with proper lineage segregation and the presence of a cavity (Fig. S5D). Interestingly, the reversine-treated cells were extruded from the blastocyst in 54% of the reversine/DMSO chimeras and in 33% of the AZ3146/reversine chimeras, whereas no cell extrusion was observed in the DMSO/AZ3146 chimeras. Quantification of the proportion of AZ3146-treated cells in 2-cell stage derived chimeras (Fig. S5E) showed similar results as 8-cell stage derived chimeras (Fig. 4F-G). In DMSO/AZ3146-treated blastocysts, we found that 49.28% of the TE, 39% of the EPI and 33.33% of the PE originated from AZ3146-treated blastomeres (Fig. S5E). In reversine/AZ3146 chimeras, 68.24% of the TE, 68.83% of the EPI, and 66.67% of the PE originated from AZ3146-treated cells (Fig. S5E). In DMSO/reversine chimeras, 22.27% of the TE, 26.98% of the PE, and 31.7% of the EPI originated from reversine-treated cells. Overall, these results suggest that aneuploidy generated at different stages similarly affect the proportion of aneuploid cells in the blastocysts.

HIF1A inhibition reduces the frequency of aneuploid cells in DMSO/AZ3146 mosaic embryos

Our data suggest that hypoxia affects aneuploid-euploid cell competition in mosaic embryos (Fig. 4). To further assess the effect of HIF1A and, considering most human embryos are culture under hypoxic conditions, we decided to evaluate the effect of inhibiting HIF1A in mosaic embryos cultured in hypoxic conditions. We assembled DMSO/AZ3146 and Reversine/AZ3146 8-cell stage chimeras, treated them with the HIF1A inhibitor IDF-11774 (Fig. 5A, S6A-B), and then assessed lineage allocation in blastocysts. Treatment of DMSO/AZ3146 chimeras with IDF11774 does not affect lineage segregation in the blastocyst but reduces the cavity diameter (Fig. 5B-C). Additionally, diminish the contribution of AZ3146-treated cells to the TE, but DMSO-treated cells seem to compensate for this loss in the PE and EPI (Fig. 5D, Fig. S6C). Treating Reversine/AZ3146 chimeras with IDF11774 compromised blastocyst morphology and cavitation (Fig. 5E-F) and considerable increased the number of PE and EPI cells (n>8 embryos per treatment collected from three independent experiments. Statistical test: Mann–Whitney U-test) (Fig. 5G), without altering the frequency of AZ3146-treated cells (Fig. S6D). Taken together, our data suggest that HIF1A promotes the survival of AZ3146-treated cells in DMSO/AZ3146 chimeras and suggest that inhibiting HIF1A could increase the proportion of karyotypically normal cells in mosaic embryos.

Figure 5

Discussion

Aneuploidy is a frequent outcome of early embryonic divisions (Allais & FitzHarris, 2022; Currie et al., 2022; Palmerola et al., 2022b). Here, we describe complementary strategies for inducing aneuploidy in mouse embryos and reveal how the embryo’s response varies depending on the nature of the insult, clel lineage, and environmental context. Our findings support four key conclusions: 1) the Mps1 inhibitor AZ3146 induces aneuploidy while preserving DNA repair capacity and dampening stress responses, compared to the broader-acting inhibitor reversine; 2) Lineage specific responses to aneuploidy emerge during blastocyst development, consistent with our earlier findings (Bolton et al., 2016; Singla et al., 2020); 3) HIF1A promotes survival of extraembryonic lineages (TE and PE) in AZ3146-treated embryos, whereas PARP1 is particularly required in the EPI of both AZ3146- and reversine-treated embryos; 4) Hypoxia enhances DNA repair and influences the competitive dynamics between aneuploid and euploid cells in mosaic embryos.

Reversine-treated embryos upregulate mRNA levels of senescent markers p53 and p21 (Singla et al., 2020) and fail to give rise to viable embryos (Bolton et al., 2016). Our current results further indicate that reversine downregulates PARP1, consistent with previous findings in glioma cells (Hirakata et al., 2021), thereby exacerbating DNA damage and limiting DNA repair capacity. In contrast, AZ3146-treated embryos maintain PARP1 expression and exhibit increased HIF1A activity, which appears to support the survival of extra-embryonic lineages and improve overall developmental potential.

These findings intersect with a broader conversation about embryo culture conditions. It has been proposed that standard culture practices shift from normoxia (20% oxygen) to more physiological hypoxia (5% oxygen) to better recapitulate the in vivo-like oviductal environment (Alva et al., 2022). Indeed, hypoxic culture conditions have been shown to improve the quality of mammalian blastocysts (Ciray et al., 2009; Nguyen et al., 2020), though the underlying mechanisms remain unclear. Studies in somatic cells suggest that hypoxia enhances DNA repair pathways (Marti et al., 2021; Nakamura et al., 2022; Pietrzak et al., 2018). In line with this, we observed reduced DNA damage in aneuploid embryos cultured under hypoxic conditions, consistent with prior reports of lower γH2A.X levels in mouse blastocysts recovered from the uterus or cultured in low oxygen (Houghton, 2021; Meuter et al., 2014). Furthermore, recent work in cancer cells indicates that hypoxia may promote HIF1A-PARP1 interactions, enhancing DNA repair (Marti et al., 2021; Nakamura et al., 2022; Pietrzak et al., 2018). Together, these findings highlight the interplay between chromosomal stress, lineage context, and the embryonic microenvironment in shaping cell fate during early development. A deeper mechanistic understanding of how DNA repair is regulated at pre-implantation stages will be crucial for elucidating how embryos tolerate or eliminate aneuploid cells — with important implications for improving embryo culture systems and fertility treatments.

Confined placental mosaicism, in which the placenta is aneuploid while the foetus is euploid, is a common phenomenon in human pregnancies (McCoy, 2017). In previous work, we showed that aneuploid TE cells exhibit prolonged cell cycles and signs of senescence, whereas aneuploid ICM cells have an increased frequency of apoptosis in reversine-treated embryos (Singla et al., 2020). These findings suggest that abnormal, yet viable TE cells, may contribute to confined placental mosaicism. In the present study, we found that in AZ3146-treated embryos cultured under normoxic conditions, aneuploid cells contributed predominantly to the TE, followed by the EPI, and were especially depleted from the PE. Under hypoxic conditions, however, we observed a shift: aneuploid cells were more frequently eliminated from the EPI and enriched in the PE. These observations suggest that hypoxia modulates lineage-specific cell competition in mosaic embryos, potentially promoting a higher proportion of euploid cells in the EPI relative to the PE and TE. Moreover, we confirmed that the proportions of aneuploid cells in the TE and EPI do not correlate, consistent with the preferential survival of aneuploid cells in the TE and their selective elimination from the ICM. This is in line with clinical data showing low concordance of aneuploidy between TE and ICM in human mosaic embryos (30–40%) (Dahdouh & Garcia-Velasco, 2021). Lineage-specific elimination of aneuploid cells by apoptosis has been described in an in vitro model of human post-implantation development using reversine (Yang et al., 2021). Our findings mirror this distinction and further suggest that mechanisms of DNA repair—possibly involving PARP1 and HIF1A—may underlie the differential survival of aneuploid cells across lineages. Investigating how these pathways operate during post-implantation development will be crucial for understanding the developmental trajectories of mosaic embryos and the basis of placental mosaicism.

Since low- and medium-grade mosaic human embryos have similar developmental potential to fully euploid embryos (Capalbo et al., 2021; Greco et al., 2015), understanding the molecular pathways that govern the response to aneuploidy holds important translational promise. Given that the extent and proportion of aneuploid cells in mosaic embryos impact embryo viability (Capalbo et al., 2021), strategies that reduce the aneuploid cell burden at the blastocyst stage may enhance the likelihood of a successful pregnancy. In this study, we targeted HIF1A as a potential regulator of aneuploid cell persistence and found that treatment with the HIF1A inhibitor IDF-11774 increased the proportion of euploid cells in mosaic embryos. Taken together, our findings reveal that specific environmental and molecular factors can modulate the composition of mosaic embryos, with potential implications for improving embryo quality and reproductive outcomes.

Materials and methods

Pre-implantation embryo culture

Mice were maintained according to national and international guidelines. Four- to six-week-old B6SJLF1/J female mice were injected with 7.5 IU PMSG followed by 7.5 IU hCG 48 h later, to induce superovulation. The females were then mated with B6CBAF1/J males or were indicated, with E-cadherin GFP or mTmG transgenic males. Pre-implantation mouse embryos were recovered 24 h and 40 h after the hCG injection to obtain zygotes and 2-cell stage embryos, respectively in M2 medium (Sigma, M7167). We incubate the embryos in KSOM until 4-cell stage for the aneuploid treatment, around 53 h post hCG. When recovering zygotes, cumulus cells were removed with 0.3% hyaluronidase (Sigma, H4272) in M2. Embryos were cultured in regular KSOM (Sigma. MR-106) at 37°C under 5% CO₂/air (Normoxia) or with pre-mixed 5% CO₂/ 5% O₂ balance with nitrogen, biologic atmosphere batch (Airgas #Z03NI9022000033) (Hypoxia). Normoxia in our study was defined as the standard atmospheric oxygen levels in culture incubators (Alva et al., 2022), whereas hypoxia (5% oxygen concentration) was the standard level of oxygen in the IVF clinics and can be used to better model physiological oxygen (physoxia), which varies from ∼1.5% to 8% (Alva et al., 2022; Knudtson et al., 2022). B6SJLF1/J females and B6CBAF1/J males were obtained from JAX laboratories. Females were received at 3-4 weeks of age and were maintained in the CALTECH animal facility, where they were housed with 5 same-sex littermates on a 12 h light/12 h dark cycle with food and water ad libitum. The temperature in the facility was controlled and maintained at 21 °C. All experimental procedures involving the use of live animals, or their tissues were performed in accordance with the NIH guidelines and approved by the Institutional Animal Care and Use Committee (IACUC) and the Institutional Biosafety Committee at the California Institute of Technology (Caltech, protocol number 1772). Reversine (Sigma, R3904), AZ3146 (Sigma, SML1427), PX-478 (Selleckchem, S7612), IDF-11774 (Selleckchem, S8771) and olaparib (Selleckchem, S1060) were dissolved in DMSO (Sigma, D2650) before use to specific concentration. They were respectively used at following final concentrations: 0.5 µM, 20 µM, 2 µM, 20 µM and 10 µM. Control embryos were incubated in the equivalent DMSO concentration. Drugs were dissolved in regular KSOM to the concentration of use. DMSO concentration in media should never pass 0.4%.

Immunofluorescence

Embryos were fixed in 4% PFA (Thermo Scientific, AA47340) for 20 min at room temperature (RT). Following by three washes with 0.1% Tween-20 (Sigma, P1379) dissolved in PBS (PBST). The embryos were then permeabilized with 0.3%, Triton X-100 (Sigma, X100) in PBS. Washes were then performed three times in PBST before embryos were transferred to blocking solution (3% BSA in PBS) for at least 3h at RT. Incubation with primary antibodies was performed in blocking solution overnight at 4°C. Next day, washes were performed three times in PBST before incubation with Alexa Fluor secondary antibodies (Thermo Fisher Scientific, 1:500) in blocking solution for 2 h. Washes were performed three times before incubation with DAPI (Thermo Fisher Scientific) for 5 min. Washes after DAPI were perform two times in PBST before the final incubation in M2. Embryos were then mounted in M2 micro-drops on 35 mm glass bottom dish (MATTEK, P35G-1.5-14-C). Confocal imaging was carried out using Leica SP8, 40x objective, 1 µM Z-step. Image files were viewed and analyzed using ImageJ and IMARIS 9.9 software.

Primary antibodies used: mouse anti-CDX2 (Biogenex, 1:500), rabbit anti-NANOG (1:500; Abcam), goat anti-SOX17 (1:300; R&D Systems), rabbit anti-HIF1A (1:300, Novusbio), mouse anti-PARP1 (1:500, Proteintech) and rabbit anti-Phospho-Histone H2AX (1:500, R&D Systems).

In situ chromosome counting

For determination of aneuploidy, treated 8-cell stage embryos were synchronized in metaphase by 10h treatment with 0.03 µg/mL colcemid (Cayman, 15364) diluted in KSOM. Following by 1h treatment with 10 µM Mg132 (Selleckchem, S2619) in KSOM. Finally, 1 h treatment with 5 µM dimethylenastron (MedChemExpress, HY-19944) and 10 µM Mg132 in KSOM. Synchronized embryos were then fix in 2% PFA for 20 min, permeabilized with PBST for 15 min and blocked 3h at RT before incubation with human anti-centromere protein antibody (1:300, Antibodiesinc 15-234) overnight at 4°C. Next day, three washes with PBST were performed before 2 h incubation with goat anti-Human secondary antibody, Alexa Fluor 647 (1:400, Invitrogen A-21445) in blocking solution. Washes were performed three times before 20 min incubation with DAPI (1:500) and Alexa Fluor 488 Phalloidin (1:300, Invitrogen A12379). Washes then were performed three times before clearing overnight with AF1 plus (Citifluor, AF1/DAPI-15). Embryos were then mounted in AF1 plus micro-drops on 35 mm glass bottom. Confocal imaging was carried out using Leica SP8, 60x objective, 0.5 µM Z-step. Image files were viewed and analyzed using ImageJ and IMARIS 9.9 software.

Embryo transfers and post implantation recovery and biopsy

Embryo transfer were performed as described previously (Bermejo-Alvarez et al., 2014). CD1 females and vasectomizes CD1 males were obtained from Charles River. Reversine and AZ3146-treated blastocyst were transfer into 2.5-days pseudopregnant CD1 females. Around 16 AZ3146-treated embryos were transfer in the right uterine horn, whereas around 16 reversine-treated embryos were transfer in the left uterine horn of the same female as control. To evaluate implantation and embryo survival potency, uterine horns were recovered 5 days after surgery. Deciduas were dissected and post implantation embryos at stage E9.5 were recovered into PBST on ice. Embryos were fixed in 4% PFA at RT for 1 h, followed by washes through PBST before imaging. Images were taken using an Olympus LS stereo microscope with a 10x objective. Uterine transfers were performed in accordance with the NIH guidelines and approved by the Institutional Animal Care and Use Committee (IACUC) and the Institutional Biosafety Committee at the California Institute of Technology (CALTECH).

qRT-PCR

Around 14 to 17 morulas and blastocysts were collected for quantitative reverse transcriptase polymerase chain reaction (qRT-PCR). Total RNA from morulas was obtained using the NucleoSpin RNA Plus XS kit (Takara, 740990.10). Whereas total RNA from blastocyst was extracted using the Arcturus PicoPure RNA Isolation Kit (Thermo Fisher, KIT0204). qRT-PCR was performed using the Power SYBR Green RNA-to-CT 1-Step Kit (Applied Biosystems, 4389986) in a StepOne Plus Real-time PCR machine (Applied Biosystems). The following program was used: 30 min 48 °C (reverse-transcription) followed by 10 min 95 °C followed by 45 cycles of 15 s 95 °C (denaturing) and 1 min 60 °C (annealing and extension). The ddCT method was used to determine relative levels of mRNA expression, with Ppia as an endogenous control (Gu et al., 2014). Primers were obtained from IDT.

Generation of chimeric embryos

Chimeric embryos were generated following a previous publish protocol (Eakin & Hadjantonakis, 2006). Briefly, wild-type, E-cadherin GFP or mTmG 8-cell stage embryos were transfer to micro-drops of M2 after aneuploid treatment. Zona pellucida was removed by treatment with acidic Tyrode’s solution (Sigma, T1788). The embryos were then incubated in Ca2+/Mg2+-free M2 (made in house) for 5 min and then disaggregated into individual blastomeres by gentle mouth pipetting. Mosaic chimeras were form by aggregation of four and four (1:1) of each treatment. Low mosaicism DMSO/AZ3146-treated blastomeres (1:1). Whereas medium mosaicism Reversine/AZ3146-treated blastomeres (1:1). Culture of the chimeras were performed under normoxic and hypoxia condition in KSOM for 48 h to reach blastocyst stage.

Statistical analysis

The statistical tests used are indicated in the corresponding figure legends. Calculations were carried out in Microsoft Excel and data analysis and visualization in Prism 9 software. All graphs show mean values, error bars: s.e.m.

Supplementary figures

Figure S.1

Figure S.2

Figure S.3

Figure S.4

Figure S.5

Figure S.6

Acknowledgements

This work was supported by M.Z.G’ National Institutes of Health R01 (R01HD101489) grant. E.S.V. is supported by a Pew Latin America fellowship. We thank the Life Science Editors and the Life Science Editors Foundation for invaluable comments and suggestions on the manuscript, and Ariane Helou for copy editing. Data analysis was performed in the Biological Imaging Facility, with the support of the Caltech Beckman Institute and the Arnold and Mabel Beckman Foundation.

Additional information

Author contributions

E.S.V designed and conducted the experiments; analyzed and interpreted the data. E.S.V and M.Z.G conceived the project. M.E.B and M.Z.G supervised the study.

Funding

National Institutes of Health (R01HD101489)