Following fertilization in mice and humans, the first two blastomeres are not equivalent, but one produces more epiblast than the other (imbalance); therefore, they do not feature equal totipotency. Research into the causes has overlooked that the epiblast imbalance is preceded by a fertilization imbalance, since in nature, the spermatozoon fertilizes the oocyte preferentially in the animal hemisphere near the animal-vegetal midline (equator). We conceived a hypothesis that the two imbalances are linked to each other, and broke it down into testable predictions. If the two imbalances were interdependent, then changing the site of sperm entry into the oocyte should change the extent of the epiblast imbalance. Thus, we evened out the fertilization imbalance, using ICSI to fertilize mouse oocytes also in the vegetal hemisphere and the equator. Resultant embryos were split at the 2-cell stage, and the twin blastocysts originating from the sister blastomeres were analyzed. Against the similarity in mRNA levels of epiblast genes, twin blastocysts differed in epiblast function, as measured by NANOG protein expression and derivation of embryonic stem cells, and the epiblast imbalance was greater after oocyte fertilization at the equator. There is no simple way to explain the positional effect other than through differences in the molecular composition of the ooplasm, which, moreover, should also be apportioned variably at the first zygotic division. We tested these predictions by measuring the orientation of the first zygotic division regarding the ICSI site, and the composition of bisected oocytes' hemispheres using half-cell proteomics. Since we found that the hemispheres have different compositions depending on the bisection axis, and the angle of the first division is variable, we propose that the variable partition of non-homogeneous ooplasm sets the stage for the epiblast imbalance. These results revive the role of the oocyte's molecular architecture on embryogenesis in a mammalian species hitherto considered mostly regulative in development.

Dear Editors, We have read with great interest the recently published White Paper by Arias et al., ‘Human stem cell-based embryo models: innovation, ethics, and policy’ (Arias et al., 2026). We would like to commend the authors for addressing their considerations now to a reproductive sciences community forum, rather than to the stem cell science community. Ever since we first took a stand on this topic in 2022 (Boiani and Duncan, 2022), we had noted that it seemed unbalanced to us that guidelines on embryo models were coming from a stem cell society (the International Society for Stem Cell research (ISSCR)) (Lovell-Badge et al., 2021; Clark et al., 2025) rather than a reproductive society like the European Society for Human Reproduction and Embryology (ESHRE). Now this asymmetry has finally been rebalanced in the White Paper. On the other hand, the White Paper is authored by a group of scientists many of whom are directly involved in stem cell-based embryo models (SCBEMs) research, hold patents related to these technologies, or have commercial interests in the field (as the conflicts of interest section makes clear (Arias et al., 2026)). While recognizing the merit of Arias et al.’s contribution, we believe some aspects lack sufficient detachment. We raise three points that require further consideration from our perspective as outsider scientists without vested interest in the technology’s success. In essence, we are proposing a different perspective on what SCBEMs are and what they are good for.

Trim-Away is an antibody-based method for the degradation of endogenous cellular proteins without prior genetic manipulation. From the effect that protein degradation has on the phenotype, one can infer the function of the protein-coding gene. Among the strengths of Trim-Away, the authors highlight the possibility to target embryonic proteins that have not yet been produced or are not yet functionally required at the time of antibody delivery. Conversely, a limitation of Trim-Away is the possibility that while the target protein is being degraded, de novo translation may replenish it, whereby the two processes offset each other, and the phenotype is inconspicuous. Appropriate controls, as described, are therefore essential for the correct interpretation of Trim-Away results. With these considerations in mind, the authors provide guidance on how to set up a Trim-Away experiment in fertilized mouse oocytes, using microinjection as the delivery method and the epithelial CADHERIN protein as an example.

The zona pellucida (ZP) is the quintessential extracellular structure of mammalian oocytes. Contrary to long-standing view that the synthesis of ZP proteins is specific to oocytes and muted in embryos, we report here that the major zona pellucida protein ZP2 is re-synthesized and functionally required during mouse embryo development. The orthogonal methods of mass spectrometry and monoclonal immunofluorescence revealed an increase of ZP2 abundance at the 8-cell / morula stage, which did not occur when zygotes were microinjected with translation-blocking oligonucleotides (morpholinos). To shed light on the functional significance of embryonic ZP2, we performed protein knockdown using immunodepletion (by ‘Trim-Away’) while at the same time preventing replenishment (by translation-blocking morpholino). ZP2 knockdown resulted in morula stage retardation and formation of defective blastocysts, whose cell lineages trophectoderm and primitive endoderm were smaller and less able to support post-implantation development. The transcriptional correlates of these morphological alterations had a gene ontology (biological process) signature that included cell lineage-relevant terms (‘endoderm development’, ‘gastrulation’), while the proteomic correlates had a gene ontology signature related to protein synthesis. Taken together, these results call into question the traditional model that ZP proteins function solely in the extracellular space and accompany embryogenesis as passive bystanders: on the contrary, ZP proteins also participate actively in the intracellular processes of early embryogenesis.

What does a secreted oocyte-specific protein, ZP2, do inside mouse embryos during cleavage? Contrary to prevalent notion, ZP2 protein synthesis is not silenced after ovulation, but goes on in preimplantation embryos to support the morula-blastocyst transition A tenet of reproductive biology is that the expression of oocyte-specific genes must remain silent in embryos. Among such genes, those of the ZP encode glycoproteins that are secreted until end of oogenesis and tasked with extracellular roles during folliculogenesis and fertilization. No further synthesis of ZP proteins is thought to occur after ovulation. However, already since the mouse gene knockout was reported in 2001 it was noticed that ZP2 null oocytes failed to develop to term even when fertilized correctly (PMID 11245577). Likewise, rabbit oocytes in which ZP4 was mutated at pronuclear stage featured defective embryogenesis (PMID 31635692). This is a treatment vs control study conducted in wild-type mice. For profiling of ZP2 protein expression,zygotes were produced in vivo and cultured 4 days in the presence of natural vs isotopic amino acids, and compared to embryos grown in vivo. For analysis of ZP2 function, zygotic knock-down of ZP2 was compared with that of a protein not present in wild-type embryos (green fluorescent protein, GFP), followed by transfer of both groups to uterus. B6C3F1 female mice were stimulated with gonadotropins and mated to CD1 males to produce zygotes. These were labeled continuously in vitro or in vivo using isotopic Arginine + Lysine, and sampled at consecutive stages for monoclonal immunofluorescence and mass spectrometry. To achieve knockdown of ZP2 during embryogenesis, zygotes were microinjected with Trim-Away reagents (PMID 29153837) and/or morpholino oligonucleotide anti-ZP2, followed by scoring of cleavage rates, blastocyst quality (germ layers) and transcriptome and proteome compositions. The most abundant protein of the mouse zona pellucida, ZP2, not only was still present inside embryos up to blastocyst, but also became significantly more abundant in the 4-8 cell stage, as determined by direct immunofluorescence with monoclonal antibody (+25%, p < 0.05, Wilcoxon test). Mass spectrometry after isotopic labeling of protein synthesis during embryo culture confirmed that ZP2 incorporated the label, meaning: de novo protein synthesis of an oocyte-specific gene product in embryos. Zygotic knockdown via morpholino oligonucleotides and Trim-Away resulted in 50% reduction of ZP2 abundance at the 8-cell stage (p < 0.05, Wilcoxon test), accompanied by growth retardation at the 8-cell and morula stage. Embryos with depletion of ZP2 formed blastocysts with significantly smaller trophectoderm and primitive endoderm compartments. These effects were not observed in controls subjected to mock knockdown (GFP). Consistent with these morphometric data, transcriptome and proteome analyses returned ‘embryonic lethality’ and ‘embryonic growth retardation’ among the over-represented gene ontology terms of ZP2 (but not GFP) knockdown embryos. Thus, intra-embryonic ZP2 appears necessary to keep up the pace of cleavage progression and secure blastocyst fitness. This discovery can explain the historical conundrum of why ZP2 null oocytes formed blastocysts and yet these could not develop to birth (PMID 11245577). Although there is sufficient evidence for expanding on the tasks of ZP proteins in mouse development, human oocytes contain ZP4 in addition to ZP1-3. While the novel tasks of ZP2 in mice hinge on de novo protein synthesis, it is unclear if this occurs also in humans. Our study provides novel insight into the pathogenic mechanism of ZP gene mutations, adding a novel intracellular route to the longstanding extracellular roles. If ZP proteins function inside human oocytes and embryos, then using ICSI to overcome problems of thick or thin zona pellucida remains ineffective. No

Across metazoan species, the centromere-specific histone variant CENP-A is essential for accurate chromosome segregation, yet its regulation during the mammalian parental-to-zygote transition is poorly understood. To address this, we generated a CENP-A-mScarlet mouse model that revealed sex-specific dynamics: mature sperm retain 10% of the CENP-A levels present in MII oocytes. However, this difference is resolved in zygotes prior to the first mitosis, using maternally inherited cytoplasmic CENP-A. Notably, the increase in CENP-A at paternal centromeres is independent of sensing CENP-A asymmetry or the presence of maternal chromosomes. Instead, CENP-A equalization relies on the asymmetric recruitment of maternal CENP-C to paternal centromeres. Depletion of maternal CENP-A decreases total CENP-A in both pronuclei without disrupting equalization. In contrast, reducing maternal CENP-C or disruption of its dimerization function impairs CENP-A equalization and chromosome segregation. Therefore, maternal CENP-C acts as a key epigenetic regulator that resets centromeric symmetry at fertilization to preserve genome integrity.

Recent advances in embryology have shown that the sister blastomeres of two-cell mouse and human embryos differ reciprocally in potency. An open question is whether the blastomeres became different as opposed to originating as different. Here we wanted to test two relevant but conflicting models: one proposing that each blastomere contains both animal and vegetal materials in balanced proportions because the plane of first cleavage runs close to the animal-vegetal axis of the fertilized oocyte (meridional cleavage); and the other model proposing that each blastomere contains variable proportions of animal and vegetal materials because the plane of the first cleavage can vary - up to an equatorial orientation - depending on the topology of fertilization. Therefore, we imposed the fertilization site in three distinct regions of mouse oocytes (animal pole, vegetal pole, equator) via ICSI. After the first zygotic cleavage, the sister blastomeres were dissociated and subjected to single-cell transcriptome analysis, keeping track of the original pair associations. Non-supervised hierarchical clustering revealed that the frequency of correct pair matches varied with the fertilization site (vegetal pole > animal pole > equator), thereby, challenging the first model of balanced partitioning. However, the inter-blastomere differences had similar signatures of gene ontology across the three groups, thereby, also challenging the competing model of variable partitioning. These conflicting observations could be reconciled if animal and vegetal materials were partitioned at the first cleavage: an event considered improbable and possibly deleterious in mammals. We tested this occurrence by keeping the fertilized oocytes immobilized from the time of ICSI until the first cleavage. Image analysis revealed that cleavage took place preferentially along the short (i.e. equatorial) diameter of the oocyte, thereby partitioning the animal and vegetal materials into the two-cell blastomeres. Our results point to a simple mechanism by which the two sister blastomeres start out as different, rather than becoming different.

Abstract Study question Is de novo protein synthesis during the conversion of oocytes into embryos coupled with aminoacid availability more than with gene transcription? Summary answer The oocyte-to-embryo transition is modulated post-transcriptionally by exogenous aminoacids, including translational reactivation of otherwise transcriptionally silent maternal genes. What is known already Countless studies documented how genes’ mRNAs are regulated up or down during the oocyte-to-embryo transition, with maternal genes being switched off and embryonic genes being switched on. This switching off and on has been poorly explored at the protein level, because it required to administer radioactive or non-canonical aminoacids to embryos in vitro or in the genital tract of the mother, in animal models. However, radioactive aminoacids could not be applied for too long as they harm cells, while non-canonical aminoacids are recognized by cells as unnatural and are metabolized differently. A promising solution is provided by non-radioactive isotopic aminoacids. Study design, size, duration In vivo-fertilized mouse oocytes (zygotes) were cultured to blastocysts in KSOM medium, which was adapted for live-cell proteomic labeling by replacing 2 of the 12 Eagle’s essential aminoacids (EAAs) with isotopic but non-radioactive counterparts (Arginine and Lysine, 13C 15N). The 7 non-essential Eagle’s aminoacids (NEAAs) were left untouched. Labeled blastocysts were sampled at 96 hours of culture in vitro. Controls were cultured in parallel under non-labeling conditions in conventional KSOM. Participants/materials, setting, methods B6C3F1 females were stimulated with eCG+hCG and mated to CD1 males. Zygotes were labeled in vitro by culture in KSOM containing polyvinylpyrrolidone in place of albumin, isotopic Arginine and Lysine in place of non-isotopic counterparts, with or without NEAAs and with or without the remaining 10 EAAs, at 37 °C under 5% CO2. In vitro-labeled blastocysts were transplanted to uterus to ascertain viability, or compared to non-labeled blastocysts by cell counting, mass spectrometry and RNA sequencing. Main results and the role of chance A total of 5426 mouse blastocysts were generated. Isotopically labeled embryos formed blastocysts with total cell numbers and full-term ability similar to unlabeled controls (blastocyst rate ≥68%, p ≥ 0.11; cell count ≥58, p ≥ 0.62; birth rate ≥25%, p ≥ 0.33; Wilcoxon test). When isotopic Arginine and Lysine were present as the sole aminoacids, the blastocyst proteome consisted of 54% unlabeled proteins, 30% semi-labeled proteins and 16% completely labeled proteins. The balance changed to 12%-51%-37% and 6%-56%-38% when NEAAs and NEAAs+EAAs, respectively, were added in culture on top of isotopic Arginine and Lysine (p = 1.9E-09; chi test). This rebalancing did not mirror in the transcriptomes, which were conserved in the three groups (p > 0.08, Wilcoxon test). The proteins of 42 maternal-effect genes (PMID 35047854), which are downregulated during the oocyte-to-embryo transition, were unlabeled when isotopic Arginine and Lysine were present as the sole aminoacids, but became increasingly more labeled (49%, 66%) when NEAAs and NEAAs+EAAs were provided on top of isotopic Arginine and Lysine. Collectively, this information supports that the oocyte-to-embryo transition is not a cell-autonomous process, but features a dynamic post-translational response to the composition of culture medium, including an altered pace of degradation and de novo synthesis of maternal proteins. Limitations, reasons for caution This is an animal study. Mature and fertilized mouse oocytes were collected from oviduct, which is not how human embryos are produced in medically assisted reproduction. The role of autophagy as internal source of aminoacids was not investigated. The oocyte-to-embryo transition takes 2-3 days longer in humans than in mice. Wider implications of the findings As seen in mice, so also in humans the oocyte-to-embryo transition could be modulated by aminoacids, considering that human media are mouse embryo-tested and the requirement of aminoacids for development is conserved across species. To recognize that culture media actively change the embryos, also in medically assisted reproduction, is overdue. Trial registration number not applicable

In oocyte biology, the zona pellucida has long been known to operate three extracellular functions downstream of the secretory pathway, namely, encasing the oocytes in ovarian follicles, mediating sperm-oocyte interaction, and preventing premature embryo contact with oviductal epithelium. The present study uncovers a fourth function that is fundamentally distinct from the other three, being critical for embryonic cell survival in mice. Intriguingly, the three proteins of the mouse zona pellucida (ZP1, ZP2, ZP3) were found abundantly present also inside the embryo 4 days after fertilization, as shown by mass spectrometry, immunoblotting, and immunofluorescence. Contrary to current understanding of the roles of ZP proteins, ZP3 was associated more with the cytoskeleton than with secretory vesicles in the subcortical region of metaphase II oocytes and zygotes, and was excluded from regions of cell-cell contact in cleavage-stage embryos. Trim-away-mediated knockdown of ZP3 in fertilized oocytes hampered the first zygotic cleavage, while ZP3 overexpression supported blastocyst formation. Transcriptome analysis of ZP3-knockdown embryos pointed at defects of cytoplasmic translation in the context of embryonic genome activation. This conclusion was supported by reduced protein synthesis in the ZP3-knockdown and by the lack of cleavage arrest when Trim-away was postponed from the one-cell to the late two-cell stage. These data place constraints on the notion that zona proteins only operate in the extracellular space, revealing also a role during the oocyte-to-embryo transition. Ultimately, these data recruit ZP3 into the family of maternal factors that contribute to developmental competence of mouse oocytes.

Abstract Study question Would new embryo properties emerge if fertilization was imposed on oocytes in regions, such as the animal pole, where it does not take place naturally? Summary answer Oocyte fertilization at the animal pole, vegetal pole or equator results in 2-cell embryos with distinguishable transcriptomes and functional peculiarities. What is known already There is fertilization bias in embryos used in basic and clinical research, because: 1) sperm–oocyte fusion hardly occurs at the surface above the meiotic spindle – the animal pole; 2) the region opposite the spindle - vegetal pole - is poorly accessible to sperm due to tiny perivitelline space. These regions are also avoided during intracytoplasmic sperm injection (ICSI) for fear of damaging the spindle or losing the sperm nucleus into the 2nd polar body. Pole materials are held non-essential for mouse development, but this conclusion relies on the 2nd polar body, which is an unreliable topological marker. Study design, size, duration Oocytes were rotated using a micromanipulator fitted with Nomarski optics, using the meiotic spindle as a landmark. Between 9:30 and 10:30 am single sperm heads were microinjected (ICSI) at the animal or vegetal pole (treatments) vs. the naturally prevalent equatorial region (control). This way two fertilization topologies and possibly also two classes of zygotes were created that are otherwise seldom or not at all represented in natural fertilization, in vitro insemination or conventional ICSI. Participants/materials, setting, methods Metaphase II oocytes were collected from 8-week-old B6C3F1 mice stimulated with 10 I.U. eCG+hCG. Sperm heads from a single batch of cryopreserved CD1 semen were deposited via ICSI in the cortex at the animal pole, vegetal pole or half-way between poles i.e. equatorially. Zygotes were cultured in KSOM(aa) and analyzed (triplicate or more) for: cleavage rates, transcriptomes at the 2-cell stage (RNAseq), blastocyst germ layers (immunostaining for trophectoderm, primitive endoderm, epiblast), and postimplantation development. Main results and the role of chance Although full development was supported irrespective of ICSI site, embryos clustered by site, as revealed by single-cell RNAseq of 21, 21 and 13 two-cell embryos whose oocytes were fertilized at the animal pole, vegetal pole or equator, respectively. When examining the sister blastomeres together, 462 genes of the shared transcriptome were differently expressed between ICSI sites, with the equatorial class contributing most to the difference (adj.p<0.05, Wilcoxon test). This was true also when examining the sister blastomeres separately: interblastomere differences of the equatorial class exceeded those of the pole classes (72% vs. 14% of differently expressed genes). Ontology analysis of the differently expressed genes using Enrichr pointed at the endomembrane system – an acquaintance of oocyte polarity studies (PMID 10545249; PMID 29746690). Follow-up of 2-cell embryos to blastocysts revealed that sister blastomere contribution to each germ layer was less balanced in the equatorial class, as measured by linear correlation of cell numbers (e.g. equatorial R2=0.00 vs. polar R2>0.23 for the epiblast). Summing up, it is difficult to reconcile these data with a mainstream view that fertilization at the animal pole is harmful. Rather they support that the topology of fertilization defines functional classes of 2-cell embryos with distinguishable transcriptomes. Limitations, reasons for caution This is an animal study. Mouse ICSI uses mercury-loaded piezo-driven needles, human ICSI does not. The higher consistency of using the one and same batch of cryopreserved spermatozoa was traded off against lower developmental rates. Single-cell resolution posed a limit on RNAseq depth. Results need confirmation in other mouse strains. Wider implications of the findings The higher blastomere similarity observed after ICSI at the vegetal pole compared to ICSI at the equator is not consistent with the prevalent model of first zygotic cleavage that is driven by the topology of the two apposing pronuclei. Trial registration number not applicable

A long-standing question in mammalian embryology is whether regional differences of oocyte composition matter for the properties of blastomeres receiving those regions after fertilization. A hitherto untested hypothesis is that allocation depends on the orientation of 1st cleavage. However, the orientation is influenced by the site of sperm entry, which can be almost anywhere on the membrane of oocytes when these are inseminated. This variability undermines consistency and reproducibility of studies. Therefore, we harnessed the intracytoplasmic sperm injection to impose the site of fertilization in three specific ooplasmic regions (animal pole, vegetal pole, equator) in mice. Notwithstanding this categorical distinction, after 1st cleavage, the sister blastomeres differed from each other nearly the same way, as measured by gene expression and twin blastocysts formation following 2-cell embryo splitting. We reasoned that either the oocyte territories did not matter, or their effect was obscured by other factors. To shed light on these possibilities, we immobilized the oocytes on the micromanipulation stage during sperm injection and for 24 h thereafter. Imaging revealed that the orientation of 1st cleavage, instead of varying with the fertilization site, followed the shorter diameter of the unfertilized oocyte. This led in most cases to the segregation of animal and vegetal hemispheres into the sister blastomeres of 2-cell embryos. Since one blastomere received more of the animal materials and the other blastomere more of the vegetal materials, this offers a rationale to explain the distinct properties of monozygotic twins derived from 2-cell embryos in mice.### Competing Interest StatementThe authors have declared no competing interest.

Journal Article A reproductive science perspective: deliberations on the stem cell guidelines update Get access Michele Boiani, Michele Boiani Editor-in-Chief, MHR Correspondence address. E-mail: mboiani@mpi-muenster.mpg.de https://orcid.org/0000-0003-2765-2781 Search for other works by this author on: Oxford Academic PubMed Google Scholar Francesca E Duncan, Francesca E Duncan Deputy Editor, MHR https://orcid.org/0000-0002-3756-9394 Search for other works by this author on: Oxford Academic PubMed Google Scholar MHR-ISSCR guidelines working group MHR-ISSCR guidelines working group Search for other works by this author on: Oxford Academic PubMed Google Scholar Molecular Human Reproduction, Volume 28, Issue 4, April 2022, gaac008, https://doi.org/10.1093/molehr/gaac008 Published: 21 March 2022 Article history Published: 21 March 2022 Corrected and typeset: 05 April 2022

Abstract Study question Are the intracellular deposits of zona pellucida (ZP) proteins relevant to embryonic development? Summary answer Contrary to longstanding views, ZPs are not only oocyte-specific proteins with extracellular tasks, but also have intracellular tasks that are essential for embryo survival. What is known already Mutation studies always defined the ZP functions as extracellular, namely: to encase oocytes in ovarian follicles, to ensure species-specific sperm binding, and to dampen shear stress on the embryo surface. Therefore, ZP mutations cause primary infertility due to empty follicles, polyspermic fertilization or harmful contact between embryo and oviductal epithelium. However, when these limitations were obviated in mice by monospermic fertilization in vitro and blastocyst transfer to uterus, the concepti of ZP2-null and ZP3-null oocytes were still unviable (PMID 11245577). This suggests that the tasks of ZPs don’t end in the extracellular space as previously assumed. Study design, size, duration After monospermic fertilization, wild-type mouse oocytes were depleted of intracellular ZP proteins using the ‘Trim-away’ method (PMID 29153837). ZP1-3 were targeted by prevalidated antibodies (producing a dominant band in Western blot) and committed for proteasomal degradation by the ubiquitine ligase Trim21. Antibodies were either folded or denatured (heat-inactivated), thereby allowing for efficient or abolished depletion, respectively (negative control). The two groups were examined for development and efficiency of ‘Trim-away’ at degrading the ZPs. Participants/materials, setting, methods Mature B6C3F1 female mice were stimulated with gonadotropins (eCG, hCG) and mated to CD1 stud males to collect fertilized superovulated oocytes. These were microinjected with a cocktail of mCherry-Trim21 mRNA and anti-ZP antibody, followed by culture in KSOM(aa) medium for 96 hours. Successful depletion of ZP proteins was measured by mCherry fluorescence, ZP Western blotting and mass spectrometry. Developmental rates were compared between the groups of the folded vs. denatured antibodies. Main results and the role of chance Fertilized oocytes depleted of ZP3 did not cleave and remained arrested at 1-cell stage (0 blastocysts / 300 oocytes; 7 replicates). Oocytes depleted of ZP2 cleaved once or twice, and remained arrested at 2- or 4-cell stage (0 blastocysts/ 100 oocytes; 3 replicates). These failures stood in contrast to the blastocyst progression of oocytes depleted of ZP1 (42 blastocysts / 100 oocytes; 3 replicates) and negative controls injected with either of the denatured antibodies (>50% blastocysts) (p < 0.01; chi test). Molecular efficacy of ‘Trim-away’ was manifest from the decline of mCherry fluorescence, and from the reduction of the ZP signal intensity in Western blot and mass spectrometry; whereas the negative controls remained bright-fluorescent and retained the dominant band in Western blot. Inspection of publicly available datasets from two-hybrid and affinity capture screens reveals that ZP3 has more than 20 interaction partners in cell lines. Collectively, this information supports that ZP proteins are required for developmental competence not only outside but also inside of oocytes, where the ZPs might interact with more proteins than just ZP1, ZP2 and ZP3. Our study provides novel insight into the pathogenic mechanism of ZP gene mutations, adding a novel intracellular route to the longstanding extracellular roles. Limitations, reasons for caution Although there is sufficient evidence for expanding on the tasks of ZPs in mice, human oocytes contain ZP4 in addition to ZP1-3. While the novel tasks in mice rely on stable protein deposits that persist from oocyte to blastocyst, it is not known if the human counterparts persist that long. Wider implications of the findings ZPs may qualify as ‘moonlighting proteins’ that enrich - but at the same time also complicate - the determination of primary infertility phenotype from genotype. Trial registration number not applicable

Growth factors became attractive candidates for medium supplementation to further improve the quality of embryo culture and to mimic in vivo nutrition. Granulocyte macrophage colony-stimulating factor (GM-CSF) is a cytokine influencing the maternal-fetal interface and supporting placental development in mouse and human. It is expressed in epithelial cells of the endometrium under the regulation of estrogens. The factor is already in clinical use and a large clinical trial showed that, if supplemented to an embryo culture medium, it leads to increased survival of embryos, especially in women with previous miscarriages. Animal and cell culture studies on isolated trophectoderm cells support an effect mainly on cellular expansion. Aim of this study was to investigate, if the supplementation of GM-CSF either in a human ART medium or in a mouse optimized medium, leads to a change in cell number and cell lineages in the early pre-implantation mouse embryo. Our data shows that mouse GM-CSF increased total cell numbers with increasing concentrations. This increase of cell number has not been found in embryos cultured in ART media with or without human GM-CSF (hGM-CSF) or in a mouse medium supplemented with different concentrations of hGM-CSF. The changes were caused by a marked difference in TE and primitive endoderm cell numbers but not due to a change in epiblast cell numbers. Additionally, results show an ectopic expression of NANOG among trophectoderm cells in both, human ART media (with and without GM-CSF) and at increasing concentrations in the mouse and the human GM-CSF supplemented media. In conclusion, we could show that GM-CSF has an effect on cell identity in mice, which might probably also occur in the human. Therefore, we would like to rare awareness that the use of supplements without proper research could bare risks for the embryo itself and probably also in the post-implantation phase.