All animal experiments were performed in accordance with protocols approved by the Animal Care and Use Committee of Children’s Hospital of Fudan University (NO. 2021-191). The Zmym2 mutant strain (080214049-HLA) was established on the FVB/N background, and maintained on 12/12-h light/dark cycles. In the Zmym2 PB allele, the PB insertion was mapped in the 15th intron of Zmym2 (Chr:14.57557590, Ensembl ID: ENSMUSG00000021945) (Figure 1). FVB/N strain, Hoxb7-mVnus fluorescent mice were used. Hoxb7-EGFP plasmid was a generous gift from Professor Frank Constatini (Columbia University, New York City, NY) (Wang et al., 2018). When the mice were 6–8 weeks old, they were housed in the same cage with a female to male ratio of 3:1. The female mice were observed for pregnancy at 8:00 a.m. The earliest time when a vaginal plug was observed was counted as E0.5D. After the vaginal plug was observed, the female mice were randomly divided into two groups.

The toes of neonatal mice at postnatal 7 days (P7D) or the yolk sac of E8.5D to E13.5D embryos were dissected from intercrossed Zmym2 PB/+ female mice and used to extract genomic DNA, followed by genotyping polymerase chain reaction (PCR). Offspring with the transposon inserted into the Zmym2 gene were genotyped using the primers P1 (5′-CTGAGATGTCCTAAATGCACAGCG-3′), P2 (5′-TCCAAAAGAGCTGGCATACTAAAGG-3′), P3 (5′-AGCACATTGTACTGTGTTCAGAGAG-3′). Zmym2 PB/PB and Zmym2 PB/+ were amplified by P1 and P3, which produced a 600 bp fragment. Wild type and Zmym2 PB/+ were amplified by P2 and P3, which yielded an 864 bp fragment (Supplementary Figure S1). PCR conditions were as follows: initial denaturation at 93°C for 90 s; 40 cycles of 93°C for 30 s, 57°C for 30 s, and 65°C for 2 min; and a final extension at 65°C for 10 min. Genomic DNA extracted from mouse toes was used as template.

Phenotypes of Zmym2 PB/+ and Zmym2 +/+ mice were assessed. The animals were anesthetized with CO₂.The location, morphology, and number of kidneys, ureters, and bladders of mice in each group were observed under a fluorescence stereoscope. The pregnant mice in difficult groups from E8.5D to E13.5D were anesthetized by CO₂, the dissection was performed layer by layer, and the embryos were removed and observed. The kidney primordia were separated under a fluorescence stereomicroscope to observe the number and location of ND and UB. The images were recorded under a microscope (Leica, Germany).

The RNeasy Mini kit (QIAGEN, Germany) was used to extract the total RNA of E9.5D embryos from the Zmym2 PB/PB and Zmym2 +/+ mice. The sequencing data was filtered with SOAP nuke (v1.5.2), afterward clean reads were obtained and stored in FASTQ format. The clean reads were mapped to the reference genome using HISAT2 (v2.0.4), Bowtie2 (v2.2.5) was applied to align the clean reads to the reference coding gene set, then expression level of gene was calculated by RSEM (v1.2.12). Using Fragments Per Kilobase of exon model per Million mapped fragments (FPKM) algorithm for the expression of standardized, which counts by total exon fragments/[mapped reads (millions) × exon length (kb)], reflects the gene expression level. Essentially, differentially expressed genes (DEGs) analysis was performed using the DESeq2 (v1.4.5) (Abdi, 2007) with fold change ≥2 and false discovery rate (FDR) ≤ 0.001. To take insight to the change of phenotype, GO1 enrichment analysis of annotated different expressed gene was performed by Phyper2 based on Hypergeometric test. The significant levels of terms and pathways were corrected by Q-value (=FDR) with a rigorous threshold (Q-value ≤0.05) by Bonferroni (Tan et al., 2021).

The E12.5D kidney primordia were isolated and fixed in formaldehyde overnight. After washing with 0.3% Triton X-100/PBS, tissues were blocked by 5% donkey serum (Jackson) and placed on a shaker at 4°C overnight. The tissues were incubated with anti-Caspase 3 antibody (CST, #9662, 1:400), then placed in a shaker at 4°C overnight. The tissues were washed with PBS and incubated with Cy5-donkey anti-rabbit antibody (Jackson, 1:1,000). DAPI (1:1,000) was used to stain the nuclei. Six kidney primordia from each group were used, and the samples were mounted with 50% glycerol/PBS. Immunofluorescence done in the sections of kidney tissues was as previously described in the kidney primordia. Anti-Zmym2 antibody (NOVUS, #NB100-56490, 1:400) and Anti-Tbx18 antibody (4A Biotech, # ABIN522536, 1:400) were used as primary antibodies. Images were taken under a confocal microscope (Leica) and analyze date using ImageJ software.

To evaluate anxiety, an open field test was conducted. Mice, approximately 8 weeks old, were acclimatized in the experimental room for over 1 h before the test began. The open field consisted of a 40 cm × 40 cm area enclosed by 40 cm high wooden walls painted white. Testing occurred between 13:00 and 17:00, with the arena cleaned between each mouse’s trial. Each mouse was placed in the center of the square, and its behavior was recorded on videotape for 15 min. The duration spent in both the peripheral zone and the center, as well as the distance traveled and instances of rearing in either area, were analyzed using a computerized technique with Etho Vision 3.0 (Noldus, Wageningen, Netherlands) (Walsh et al., 2017).

For the aggressive behavior test, the resident–intruder paradigm Koolhaas et al. (2013) was employed. Each male mouse was isolated for 1 h prior to testing and was evaluated in his home cage (acting as the resident) against a group-housed male intruder (four mice per cage). The behavior was recorded on video for up to 15 min. Key metrics included the latency to the first biting attack and the number of biting attacks during the subsequent 5 min following the first attack. If no biting behavior was observed for 10 min, the test was terminated, and the latency was recorded as 10 min. Testing was conducted between 12:00 and 16:00 and repeated three times for each male at 3-day intervals, ensuring that each male encountered a novel intruder during each trial. To assess locomotor activity, the duration of walking and the number of rearings were recorded for the 5 min following the first attack in the initial trial. If no biting behavior occurred, the latter 5 min of the 10-min period recorded was used to evaluate locomotor activity.

Blood glucose concentrations of 0-, 7-, 14-, 21-, 30-, 60-, 90- and180-day-old mice were measured using a glucometer, typically taken in either a random or fasted state. Oral glucose tolerance tests (OGTT) were performed in 90-day-old Zmym2 PB/+ mutant mice and Zmym2 +/+ littermate control mice as previously described (Kennard et al., 2021). Serum insulin concentrations in a random or fasted state and 30, 60, and 120 min after oral glucose application were quantified by ELISA (Abcam, #ab285341).

Numerical data are presented as the mean ± standard deviation. Statistical analysis was performed using GraphPad Prism version 10.0 and SPSS 20.0 statistical software. Two-sided unpaired t tests were used for between-group data, and χ2 tests were used for assessing count data. The significance level was set at p < 0.05, p > 0.05 representing no statistically significant distinction, all n = 6 at least in each group. In all figures, a single asterisk indicates p < 0.05, while double, triple, and quadruple asterisks indicate p < 0.005, p < 0.0005, and p < 0.0001, respectively.

The PB transposon was inserted into the 15th intron of Zmym2 (Figure 1A). In E9.5D embryos, quantitative PCR analysis showed that Zmym2 mRNA levels were reduced by nearly 100% in homozygous mutants (Zmym2 PB/PB) compared to wild-type littermates (Zmym2 +/+) and by approximately 50% in heterozygous mutants (Zmym2 PB/+) (Figure 1B). Western blot analysis demonstrated that PB transposon insertion resulted in significant truncation of the Zmym2 protein. In the whole embryos tissues at E9.5D, a truncated protein band (95 kDa) was detected in both Zmym2 PB/+ and Zmym2 PB/PB lanes. Native Zmym2 protein (155 kDa) expressed in Zmym2 PB/+ and Zmym2 +/+ groups, and the protein level of native Zmym2 in Zmym2 PB/+ group was reduced to approximately 50% of that in the Zmym2 +/+ group. Notably, the native Zmym2 band is absent in the Zmym2 PB/PB group (Figures 1C, D).

Furthermore, quantitative PCR analysis and immunofluorescence were performed to assess Zmym2 expression levels and distribution in the kidneys of mutant mice. Zmym2 mRNA expression in Zmym2 PB/+ kidneys was approximately 50% of that in Zmym2 +/+ kidneys (Figure 1E). In kidney tissue sections, Zmym2 was predominantly localized in the nuclei of tubular cells, with weak expression detected in glomerular cells. Fluorescence quantification results were consistent with the quantitative PCR results, confirming that Zmym2 expression was lower in Zmym2 PB/+ kidneys than that in Zmym2 +/+ kidneys (Figures 1F, G).

Zmym2 PB/+ mice were mated and their neonatal offspring were genotyped by PCR amplification on postnatal 7 days. No Zmym2 PB/PB mice were identified across more than three consecutive generations (Supplementary Figure S1A), indicating that Zmym2 PB/PB embryos are homozygous lethal. This finding is consistent with a previous study and the Mouse Genome Informatics database (MGI:5706016) (Graham-Paquin et al., 2023). To investigate the role of Zmym2 in embryonic development, embryos were collected at different stages. Zmym2 PB/PB embryos were present at mendelian ratios until E9.5D (Figure 2A). However, by E10.5D, embryonic resorption was observed, and no Zmym2 PB/PB embryos were recovered, confirming embryonic lethality (Figure 2B). At E9.5D, most Zmym2 PB/PB embryos exhibited gross phenotypic abnormalities, and a statistically significant reduction in length was noted (Figures 2C, D). Zmym2 PB/PB embryos seemly appeared reduction in length at E8.5D, but there was no statistical difference (Supplementary Figure S1C).

In contrast, Zmym2 PB/+ mice are viable and fertile. We monitored the general condition (weight, fur, activity and appetite) and survival rates among different experimental groups over a period of 12 weeks (Supplementary Figures S1D, E). The weight of Zmym2 PB/+ mice was normal, and their fur, dietary habits, and activity levels were basically the same as those of Zmym2 +/+ mice. However, a higher proportion of the Zmym2 PB/+ than Zmym2 +/+ mice died throughout the observation period (P60D-P180D).

A comprehensive analysis of the urinary system was performed in 60 newborn Zmym2 PB/+ mice. Among these, 28 displayed CAKUT phenotypes: 13 (21.7%) had unilateral duplex kidneys (DK), 9 (15.0%) had unilateral hydronephrosis (HN) and 6 (10.0%) had unilateral RA. No significant differences in phenotype distribution were observed between sexes or between left and right kidneys (Figures 3A, C). VUR testing was conducted in all mice (Figure 3B), revealing that unilateral VUR was present in 13 (46.4%) of the 28 CAKUT mice: 4 cases from DK group, 6 cases from HN group, and 3 cases from RA group (Figure 3C). No obvious CAKUT phenotypes were detected in Zmym2 +/+ mice. Histopathological analysis of the newborn Zmym2 PB/+ mice was performed. The structure and morphology of the glomeruli and kidney tubules appeared normal in mice with DK. However, in mice with HN, the numbers of glomeruli and tubules were reduced.

To determine whether the Zmym2 mutation affects kidney function, blood urea nitrogen (BUN) and serum creatinine (Scr) levels were measured. Both BUN and Scr levels remained stable over a 180-day period in Zmym2 +/+ and Zmym2 PB/+ mice, with no statistically significant differences between the two groups (Supplementary Figure S2A). Furthermore, H&E staining revealed no structural abnormalities in the glomeruli of either Zmym2 PB/+ or Zmym2 +/+ mice (Supplementary Figure S2B).

The morphology of the testis, seminal vesicles, and caudal epididymis was examined. Most male adult Zmym2 PB/+ mice (77.78%, 7/9) exhibited seminal vesicle atrophy (Figure 4A), with a 32% reduction in seminal vesicle weight compared to Zmym2 +/+ controls (Figure 4B). Additionally, some Zmym2 PB/+ mice displayed unilateral cryptorchidism (22.22%, 2/9), bilateral cryptorchidism with hydrocele (11.11%, 1/9) or intratesticular cysts (11.11%, 1/9). The morphology of mature epididymal sperm in Zmym2 PB/+ mice was normal, with sperm heads exhibiting typical morphology.

To assess the impact of the Zmym2 mutation on sperm quantity and quality, sperm counts were compared between Zmym2 PB/+ and Zmym2 +/+ mice (Table 1). No significant difference in total sperm number in the epididymis. However, the average path velocity (VAP), straight-line velocity (vSL) and curvilinear velocity (vCL) were significantly reduced in Zmym2 PB/+ mice (p < 0.05). The percentage of progressive mobile sperm (level a and b) was slightly lower in the Zmym2 PB/+ group (p = 0.14), whereas the percentage of sperm with slow velocity (level c) was significantly higher (p < 0.05). These findings indicate that sperm motility is impaired in Zmym2 PB/+ mice due to the Zmym2 mutation.

Despite the observed morphological abnormalities in the testis, seminal vesicles, and epididymis, as well as changes in sperm motility Zmym2 heterozygosity did not appear to affect male fertility. Fertility was assessed by continuously mating Zmym2 +/+ and Zmym2 PB/+ males with 8-week-old wild-type females at a 1:2 male-to-female ratio. To control for genetic background differences, females from different strains (S129 and C57BL/6) were used. No significant differences were observed between the two groups in terms of the ability to induce vaginal plugs or achieve pregnancy in wild-type females. Furthermore, the mean litter size was nearly identical in both groups (Supplementary Table S1), indicating that the Zmym2 mutation does not affect vaginal plug formation or pregnancy rate.

The craniofacial morphology, gross brain structure, and histopathology (H&E staining) of brain tissue were examined in Zmym2 PB/+ and Zmym2 +/+ mice. No craniofacial abnormalities were detected in either genotype (Figures 5A–C). Behavioral phenotypes were evaluated in Zmym2 PB/+ mice. Anxiety levels were assessed using the open-field test. Male Zmym2 PB/+ mice were significantly more active than male Zmym2 +/+ controls during the first 15 min of the test and spent less time in the center of the chamber (Figures 5D–G) suggesting increased anxiety-like behavior. However, no significant difference in anxiety levels was observed between female Zmym2 PB/+ and Zmym2 +/+ mice. Aggressive behavior, was assessed using the resident-intruder paradigm. The latency to the first biting attack did not differ between male Zmym2 PB/+ and Zmym2 +/+ mice (Figure 5H). The frequency of biting attacks was similar between genotypes during the first and second trials. However, in the third trial, the frequency of biting attacks was significantly higher in Zmym2 PB/+ mice than in Zmym2 +/+ mice (Figure 5I) indicating that Zmym2 PB/+ mice exhibited increased aggression following repeated encounters with intruders. Locomotor activity was evaluated by recording walking duration and the number of rearings during the first trial. No significant differences were observed between for these measures Zmym2 PB/+ and Zmym2 +/+ mice for either measure (Figures 5J, K).

ZMYM2 plays a critical role in embryogenesis, potentially by regulating gene expression and cellular functions essential for development (Graham-Paquin et al., 2023). To further investigate the role of Zmym2 in embryonic development and understand the lethality of Zmym2 PB/PB embryos, mRNA was were extracted from the E9.5D embryos of Zmym2 +/+ and Zmym2 PB/PB mice for RNA-sequencing (RNA-seq). The analysis identified 342 DEGs between wild-type and homozygous mutant mice, including 221 upregulated genes and 121 downregulated genes. These DEGs were categorized and enriched for functions related to DNA-bind transcription factor activity, nervous system development and so on. Notably, the most significantly downregulated DEGs associated with Zmym2 mutation were linked to maturity-onset diabetes of the young (MODY). RNA-seq profiles between Zmym2 +/+ and Zmym2 PB/+ embryos highlight the consistent results that MODY associated genes are the most significantly downregulated DEGs (Figures 6A, B).

Subsequent glucose metabolism tests in Zmym2 PB/+ mice revealed progressive disturbances in glucose homeostasis until 90 days of age, and then stabilizing (Figure 6C). At 90 days, both random and fasting blood glucose levels were significantly elevated in Zmym2 PB/+ mice and remained stable throughout the study. During the OGTT at 90 days, blood glucose concentrations in Zmym2 PB/+ mice were significantly higher than those in Zmym2 +/+ mice at all measured time points. The area under the glucose curve (AUC) in 90-day-old Zmym2 PB/+ mice was significantly greater than that in Zmym2 +/+ mice (Figures 6D, E). Additionally, random serum insulin levels in Zmym2 PB/+ mice were significantly lower starting from P90D (Figure 6F). During the OGTT performed with 90-day-old mice, serum insulin levels in Zmym2 PB/+ mice were significantly elevated at 30 min post-glucose administration (Figure 6G). However, the AUC for insulin in Zmym2 PB/+ mice was significantly lower than that in Zmym2 +/+ mice. Between 60 and 90 min after glucose administration, AUC insulin levels in Zmym2 PB/+ mice were comparable to those in Zmym2+/+ mice (Figure 6H). The pancreatic morphology and histological appearance of Zmym2 PB/+ mice were unchanged compared to those of Zmym2 +/+ mice (Supplementary Figure S3).

Duplex kidney was the most frequently observed urinary system abnormality in Zmym2 PB/+ mice, occurring in 13 of 28 cases (46.4%). Urinary system malformations originate early in embryonic development, and duplicated UB formation is associated with duplex kidney in mice. To examine the effect of Zmym2 mutation on UB budding, kidney primordia were isolated at E12.5D from Zmym2 PB/+ and Zmym2 +/+ mice. Duplicated UBs were observed in nine of 36 (25.0%) Zmym2 PB/+ samples, whereas no aberrant UB formation occurred in Zmym2 +/+ mice (Figure 7A).

Defects in apoptosis can contribute to various CAKUT phenotypes. Apoptosis in the UB is crucial for normal UB growth and subsequent branching morphogenesis. A wave of apoptosis in the ND that is likely necessary for separating the ureters from the ND (Mendelsohn et al., 2009). The number of apoptotic cells in the ND and UB was compared, revealing a significant increase in caspase3 positive apoptotic cells in these regions in Zmym2 PB/+ mutants (Figures 7B–D), which may contribute to the development of VUR and other CAKUT phenotypes. Previous proteomic analyses identified ZMYM2 as endogenous binding partner of TBX18 in 293 and A549 cells. Tbx18 is co-expressed with Zmym2 in the mesenchymal compartment of the developing ureter in mice, and mutations in TBX18 and in ZMYM2 have recently been linked to CAKUT, supporting the functional relevance of TBX18 and ZMYM2 interaction in ureter development (Ludtke et al., 2022). To investigate the effect of Zmym2 mutation on Tbx18 expression in kidney, immunofluorescence staining and western blot were performed to assess co-localization and expression changes in the kidney tissue. Zmym2 and Tbx18 were primarily co-localized in kidney interstitial cells, with particularly high expression in the nephrogenic zone(Figure 8A). Zmym2 PB/+ kidneys exhibited discontinuous nephrogenic zones, which was consistent with the formation of multiple UBs. In comparing to Zmym2 +/+ group, Tbx18 expression in Zmym2 PB/+ kidney tissue was decreased approximately 50% (Figures 8B, C).