For the preparation of fertilized eggs, we used C57BL/6N male mice (8 weeks old) and female mice (4 weeks old) purchased from Japan SLC (Hamamatsu, Japan). All animals were maintained under specific pathogen-free conditions in individually ventilated cages (TECNIPLAST S. p.A., Buguggiate, Italy) on a 12-h light/12-h dark cycle at 22 °C ± 2 °C and 40%–60% relative humidity. They were provided with Labo MR Stock food (Nosan Corporation, Yokohama, Japan) ad libitum. All animal protocols were reviewed and approved by the Institutional Animal Care and Use Committee of Kanazawa Medical University and conducted in accordance with the university’s guidelines.

For all embryo culture experiments, one-cell stage C57BL/6N embryos were cryopreserved and subsequently thawed for use. The cryopreservation and thawing protocols have been described in our previous publication (Darwish et al., 2019). Briefly, to collect oocytes, 4-week-old C57BL/6N female mice were intraperitoneally injected with HyperOva (KYUDO CO., Ltd., Tosu, Japan) to induce ultra-superovulation. Forty-eight hours later, 7.5 IU of human chorionic gonadotropin (Sigma-Aldrich, St. Louis, MO, United States) was injected, and oocytes were harvested from the oviducts 16 h thereafter. Fertilization was performed in HTF medium (ARK Resources, Kumamoto, Japan) designed for in vitro fertilization, by incubating oocytes with sperm. Four hours after fertilization, excess sperm were removed, and the embryos were incubated in a 200 µL HTF drop supplemented with 20% fetal bovine serum for 10 min. Subsequently, the embryos were cryopreserved in liquid nitrogen using a freezing solution containing 1 M DMSO (ARK Resources) and DAP213 solution (ARK Resources). Thawing of the frozen embryos was rapidly carried out using 900 µL of 0.25 M sucrose solution, followed by two washes in KSOM medium (ARK Resources) before being used in subsequent experiments. To assess the effects of ultra-superovulation and freeze–thaw procedures, in vivo fertilized zygotes obtained after natural mating served as controls.

We utilized two standardized inhibitor libraries, SCADS Inhibitor Kit II ver. 2.0 and SCADS Inhibitor Kit III ver. 1.6, which systematically compiles a broad range of chemical inhibitors commonly used in chemical biology research. A complete list of the inhibitors included in these kits is provided in Supplementary Table S1. The two SCADS inhibitor kits were obtained from the Molecular Profiling Committee of the Advanced Animal Model Support platform, Japan.

Each kit was supplied in a 96-well plate format, with 5 μL (10 nmol) of each inhibitor pre-dispensed per well. To prepare 100 μM stock solutions, 95 μL of 50% methanol was added to each well. These stock solutions were then diluted with KSOM medium, which was used for the subsequent embryo culture experiments.

For each embryo culture experiment, 20 thawed one-cell stage embryos were used per treatment group. Each inhibitor was added to KSOM medium at a final concentration of 1 μM, and embryos were cultured in this medium. In total, 96 experimental groups were prepared, including 95 inhibitor-treated groups and one control group. Each group was independently replicated three times, resulting in 288 embryo culture experiments.

The developmental rate (%) was calculated by dividing the number of developed embryos by the total number of embryos used in the experiment at the time of observation and then multiplying by 100. Mathematically, this is expressed as: Developmetal rate  % = N d e v e l o p e d N t o t a l × 100 where N _developed is the number of developed embryos, and N _total is the total number of embryos used in the experiment.

To analyze expression levels of Cxcr2 and Ctsd genes across human and mouse embryonic development stages, we re-analyzed the GSE44183 (Xue et al., 2013) single-cell RNA-seq dataset. FASTQ files were downloaded from the Sequence Read Archive using SRA Toolkit and quantified using Salmon v1.10.0 (Patro et al., 2017) NCBIBioinformatics Answers with bias correction options (--gcBias, --seqBias, --validateMappings) against a combined human (GRCh38) and mouse (GRCm39) transcriptome reference from Ensembl release 114. Transcript-level abundances were imported and summarized to gene-level using tximport v1.30.0 (Soneson et al., 2015) GSE44183 (GEO), with transcript-to-gene mapping obtained via biomaRt. Differential expression analysis between species and developmental stages was performed using DESeq2 v1.42.0 (Love et al., 2014) Salmon: Accurate, versatile and ultrafast quantification from RNA-Seq data using lightweight-alignment | Request PDF with default parameters. Expression levels were visualized as log2(TPM+1) values and statistical significance was assessed at an adjusted p-value threshold of 0.05.

Immunofluorescence staining of mouse embryos was based on the method of Sauvegarde et al. (Sauvegarde et al., 2016). For CTSD detection, an anti-CTSD rabbit polyclonal IgG antibody (1:200, 55021-1-AP, Proteintech group Inc., IL, United States), an anti-CXCR2 (1:200, 19538-1-AP, Proteintech) for CXCR2 detection, and as a negative control, a rabbit IgG isotype control antibody (98136-1-RR, Proteintech) was used. Anti-rabbit IgG recombinant VHH CoraLite® Plus 488 (1:500, srb2GCL488-1, Proteintech) was used as the secondary antibody. After the secondary antibody reaction, embryos were washed and then embryos were mounted on a 35-mm glass-bottom dish (3960–035, AGC TECHNO GLASS Co. Ltd., Shizuoka, Japan) and examined using an All-in-one fluorescence microscope (BZ-X1000, KEYENCE, Osaka, Japan).

Genome editing experiments were performed as previously described (Darwish et al., 2019; Nishizono et al., 2020). Briefly, CRISPR-Cas9 ribonucleoprotein (RNP) complexes were prepared by combining two guide RNAs (each consisting of 100 ng/μL crRNA and 200 ng/μL tracrRNA) with 100 ng/μL Cas9 protein in Nuclease-Free Duplex buffer (Integrated DNA Technologies, Inc., Coralville, IA, United States). Approximately 100 one-cell stage embryos were placed in the solution, and electroporation was conducted using a NEPA21 electroporator (Nepa Gene Co., Ltd., Ichikawa, Japan) to introduce the RNP complexes into the embryos. All reagents, including Cas9 protein, RNA oligonucleotides, and buffer, were purchased from Integrated DNA Technologies.

The sequences of the crRNAs used and the PCR primers for genotyping knockout alleles are listed in Supplementary Table S2. Because electroporation itself can cause physical damage to embryos (Miyagoshi et al., 2025), control embryos were electroporated in buffer with Alt-R™ Cas9 Negative Control crRNA (1072544, Integrated DNA Technologies), tracrRNA and Cas9 protein. Genome editing experiments for each target gene were independently repeated seven times.

All numerical data are presented as mean ± standard deviation (SD). Statistical analyses for each experiment were performed using JMP® Pro 18.0.2 (JMP Statistical Discovery LLC, Cary, NC, United States). Heatmap visualization and clustering analysis of the inhibitor library screening data were conducted using Python 3 with the Seaborn library.

Prior to conducting the inhibitor library screening, we first evaluated the effects of the solvent used in the library on mouse fertilized eggs. The inhibitor library used in this study was prepared in 50% methanol, resulting in a stock solution concentration of 47.5%. To assess methanol tolerance, we tested its impact on embryo development at various final concentrations (Figure 1A). Under the standard usage condition, which yields a final methanol concentration of 4.75%, embryo development rates at 96 h after culture initiation were severely impaired, with a mean of 2.60% ± 3.43% (Figure 1B), indicating near-complete developmental arrest. This data suggests that mouse embryos are more sensitive to methanol toxicity compared to other cell types.

In contrast, methanol concentrations of 0.475% and 0.0475% (corresponding to 1/10 and 1/100 of the standard usage, respectively) did not show statistically significant differences in developmental rates compared to embryos cultured without methanol: 0.475% group: p = 0.1481; 0.0475% group: p = 0.1439, Tukey-Kramer test. Furthermore, no significant difference was observed between the 0.475% and 0.0475% methanol groups (p = 1.0000). The corresponding inhibitor concentrations at these dilutions were 1 μM (0.475% methanol) and 100 nM (0.0475% methanol), respectively. Based on these findings, we conducted all subsequent screening experiments using 1 μM inhibitors in 0.475% methanol diluted in KSOM medium. Furthermore, to evaluate whether ultra-superovulation and freeze–thaw procedures affect inhibitor efficacy, we compared in vivo fertilized embryos (in vivo embryos) with IVF-derived embryos obtained after ultra-superovulation and freeze–thawing (Cryo IVF embryos) under DMSO (1A) treatment (Figure 1C, n = 6). This analysis demonstrated a significant pre-stress in the Cryo IVF embryos (p < 0.05). Accordingly, subsequent analyses should be interpreted in the context of baseline pre-stress in the ultra-superovulation and freeze–thaw group.

Using the optimized inhibitor and solvent concentrations established in preliminary experiments, we screened for compounds that affect early embryonic development. To minimize genetic variability from parental origin, we employed a strategy that combined ultra-superovulation and in vitro fertilization (IVF) to generate large cohorts of fertilized eggs from the same parents. These one-cell stage embryos were cryopreserved, allowing us to conduct high-throughput screening using embryos with a uniform genetic background (Figure 2A).

Through this screening, we identified 16 compounds that significantly impaired embryo development, including five known regulators (Figure 2B). Hierarchical clustering analysis revealed that the inhibitors could be categorized into five groups based on their effects on developmental progression: Group I–inhibitors that caused severe toxicity at the one-cell stage; Group II–those that blocked development from the one-cell to two-cell stage; Group V–those that interfered with the transition from the 4-cell to morula stage; Group IV–those that impaired blastocyst formation; and Group III–those with minimal or no effect on development. Among the 95 inhibitors tested, the majority (68.42%) fell into Group III, followed by Group V (15.79%). In contrast, Group II (3.16%) and Group IV (2.11%) were relatively rare.

Of the five known factors identified, four were involved in energy metabolism: two ATPases and two mitochondrial complex inhibitors. Notably, inhibitors targeting F1-ATPase or mitochondrial complex I/III exerted strong deleterious effects at the one-cell stage, while V-ATPase inhibition specifically disrupted morula formation (Figure 2C). The remaining known factor, cathepsin B, markedly inhibited the transition from the one-cell to two-cell stage. In addition to these, we discovered 11 previously uncharacterized compounds that significantly impaired development at distinct stages, including inhibitors of p53 activator (PRIMA-1), cathepsin D, CXCR2, and potassium channels (Table 1).

Among them, cathepsin D inhibition interfered with progression from the 4-cell to morula stage, while CXCR2 inhibition affected the one-cell stage. RNA-seq analysis showed that Cathepsin D (Ctsd) transcripts were detectable from the oocyte stage through the morula stage, whereas Cxcr2 transcripts, although near the detection limit, were present from the two-cell stage through the morula stage (Figure 3A). In human embryos, similar analyses indicated that CTSD was detected but with a distinct temporal profile, while CXCR2 transcripts were not detected above background (Figure 3B). Protein-level assessment by immunofluorescence (IF) corroborated these trends. CTSD signal was observed from the oocyte stage through the blastocyst stage, broadly consistent with the transcript data (Figure 3C). CXCR2 IF signal was very faint but reproducibly detectable from the two-cell stage through the morula stage, and it was not detected in blastocysts (Figure 3D). Based on these findings, we selected cathepsin D and CXCR2 for further validation using gene knockout experiments.

To verify that the developmental arrest observed with specific inhibitors was attributable to inhibition of their intended targets, we performed loss-of-function experiments using the CRISPR-Cas9 system. Cas9 protein and guide RNAs (crRNA and tracrRNA) targeting either Ctsd (Figure 4) or Cxcr2 (Figure 5) were delivered into one-cell-stage embryos by electroporation.

For Ctsd, because the gene is long and contains multiple exons and splice variants, we designed guides to the start codon in exon 1, which is conserved across all isoforms (Figure 4B). After electroporation, zygotes were collected for PCR and Sanger sequencing, which confirmed genomic deletions at the predicted cut sites (Figure 4C, genome editing efficiency = 100%). Developmental outcomes for these embryos versus those electroporated with a negative-control gRNA are shown in Figure 4D. Ctsd knockout embryos exhibited a decline in developmental rate from the 4-cell stage, with a significant reduction at the blastocyst stage (n = 7, p < 0.05). This phenotype was slightly delayed relative to the morula-stage arrest seen with pharmacologic inhibition (Figure 2B).

For Cxcr2, because the entire coding sequence resides within exon 3, we targeted this exon to achieve functional disruption (Figure 5B). Sanger sequencing indicated that double-strand breaks occurred within the region targeted by crRNA1 (of the two crRNAs tested), resulting in a partial deletion predicted to introduce a premature stop codon (Figure 5C, genome editing efficiency = 75%). Cxcr2 knockout embryos showed a significant reduction in blastocyst formation (n = 7, p < 0.05).

Although the onset of developmental failure in both knockout models was slightly later than with small-molecule inhibition, these results collectively support essential roles for Ctsd and Cxcr2 during early embryonic development.