CDH4/UBA1/RBMX axis promotes polycystic ovary syndrome progression through YAP1 activation.

Recent studies have demonstrated that any dysregulation in the Hippo signaling pathway, particularly in its key effector YAP1, plays a crucial role in the pathogenesis of PCOS and the associated insulin resistance. To further investigate this molecular mechanism, comprehensive screening of the PCOS differential gene expression dataset GSE168404 and PMID29344314 and the known regulatory components of the Hippo pathway was conducted. This integrated analysis revealed CDH4 to be a promising candidate gene potentially involved in the regulation of Hippo signaling (Fig.  1 A). The subsequent protein–protein interaction network analysis revealed that CDH4 interacted with ten core associated proteins, forming a distinct molecular cluster (Fig.  1 B). To elucidate the roles of the ten core CDH4-interacting proteins identified in our analysis, we performed a functional pathway enrichment analysis using the Wiki Pathways database ( https://www.wikipathways.org/ ). This further confirmed the significant association between CDH4 and the components of the Hippo signaling pathway (WP4541, https://www.wikipathways.org/pathways/WP4541.html , and WP4540, https://www.wikipathways.org/pathways/WP4540.html ), supporting the potential regulatory role of CDH4 in this pathway (Fig.  1 C). Subsequently, on the basis of the data from GEO ( GSE168404 ) (Fig.  1 D) and PMID29344314 (Fig.  1 E), CDH4 was subsequently found to be upregulated in patients with PCOS. Next, in vitro experiments were conducted using different concentrations of DHEA. When the concentration of DHEA was 20 μM or greater, the mRNA levels of Cdh4 in the DHEA culture for 1 day was greater than that in the control group (dimethyl sulfoxide, DMSO) (Fig.  1 F), and the Cdh4 mRNA levels in the DHEA culture for 2 days was greater than that in the control group (Fig.  1 G). Similarly, when KGN (human granulosa cell tumor) cells were cultured with DHEA at concentrations of 20 μM or higher, the CDH4 protein levels were also higher than those in the control group (Fig.  1 H). These results revealed that CDH4 expression is upregulated in PCOS. Fig. 1 CDH4 is associated with the Hippo pathway and is highly expressed in the GCs of patients with PCOS. A Venn diagram of gene sets and Hippo-related gene sets from GSE168404 and PMID29344314. B A protein–protein interaction (PPI) plot showing the interaction between CDH4 and ten core-related proteins. C Enrichment analysis in Wiki Pathways for the ten core CDH4-binding proteins. D The mRNA levels of Cdh4 in GSE168404 . The GSE168404 data are derived from human ovarian granulosa cells, N  = 5. E The mRNA levels of Cdh4 in PMID29344314. The PMID29344314 data are derived from human ovarian granulosa cells, N  = 3. F The mRNA levels of Cdh4 cultured with DHEA for 1 day in KGN, N  = 4. G The mRNA levels of Cdh4 cultured with DHEA for 2 days in KGN, N  = 4. H The protein levels and quantitative analysis of CDH4 cultured with DHEA for 2 days in KGN CDH4 is associated with the Hippo pathway and is highly expressed in the GCs of patients with PCOS. A Venn diagram of gene sets and Hippo-related gene sets from GSE168404 and PMID29344314. B A protein–protein interaction (PPI) plot showing the interaction between CDH4 and ten core-related proteins. C Enrichment analysis in Wiki Pathways for the ten core CDH4-binding proteins. D The mRNA levels of Cdh4 in GSE168404 . The GSE168404 data are derived from human ovarian granulosa cells, N  = 5. E The mRNA levels of Cdh4 in PMID29344314. The PMID29344314 data are derived from human ovarian granulosa cells, N  = 3. F The mRNA levels of Cdh4 cultured with DHEA for 1 day in KGN, N  = 4. G The mRNA levels of Cdh4 cultured with DHEA for 2 days in KGN, N  = 4. H The protein levels and quantitative analysis of CDH4 cultured with DHEA for 2 days in KGN To investigate the role of CDH4 in PCOS pathogenesis, a DHEA-induced PCOS mouse model was established through continuous subcutaneous injection of DHEA for 21 days, while control mice received an equivalent volume of oil (Fig.  2 A). Wild-type (WT) and CDH4-knockout (CDH4-KO) mice were divided into four groups: WT + oil, KO + oil, WT + DHEA (PCOS), and KO + DHEA (KO-CDH4-PCOS). A comparative analysis revealed distinct phenotypic differences among the experimental groups: While body weight measurements revealed no significant differences between the KO + DHEA (KO-CDH4-PCOS) and WT + DHEA (PCOS) groups (Fig.  2 B), the KO + DHEA group presented significantly lower serum testosterone levels than the WT + DHEA model group did (Fig.  2 C). Furthermore, the inflammatory factor analysis revealed elevated IL-6 and TNF-α concentrations in WT + DHEA mice compared with those in both the WT + oil and KO + DHEA groups (Fig.  2 D, E). No significant difference in body weight, serum testosterone levels, IL-6, and TNF-α was observed between the WT + oil and KO + oil groups. Compared with the sustained estrus of WT + DHEA mice, vaginal cytology showed that KO + DHEA mice partially restored the estrous cycle in some mice, with the reappearance of proestrus, estrus, and diestrus (Supplementary Fig. S1A). A histopathological examination of ovarian morphology revealed that CDH4 deficiency by reducing antral and cystic follicles, which promoting the restoration of normal follicular architecture (Fig.  2 F and Supplementary Fig. S1B). Molecular analysis of ovarian tissues revealed significant alterations in the CDH4 and YAP1 expression profiles across the experimental groups. Quantitative PCR analysis demonstrated that Cdh4 mRNA and Yap1 mRNA expression were markedly upregulated in WT + DHEA mice compared with that in the WT + oil animals (Fig.  2 H, I). Consistent with the observed transcriptional changes, western blotting revealed elevated protein levels of both CDH4 and YAP1 in the ovaries of the WT + DHEA group relative to those of the WT + oil group (Fig.  2 J and Supplementary Fig. S3C). The results of immunohistochemistry showed that the expression of CDH4 in the ovaries of the WT + DHEA group was elevated (Fig.  2 G). Furthermore, Yap1 expression analysis revealed that WT + DHEA mice presented significantly higher mRNA levels than KO + DHEA groups (Fig.  2 I), with the corresponding protein expression patterns showing similar trends (Fig.  2 J and Supplementary Fig. S3C). Apoptosis-related marker analysis revealed that the WT + DHEA group ovarian tissues presented reduced levels of cleaved Caspase-3 (cl-Caspase-3), elevated Bcl-2 and PCNA expression compared with those in both control and KO + DHEA groups (Fig.  2 J, Supplementary Figs. S1C and S3C, D). Fig. 2 Role of CDH4 in a mouse model of PCOS induced by DHEA. A The PCOS mouse model was established via 21-day continuous subcutaneous injection of DHEA (60 mg/kg; control mice received the oil vehicle alone). B Body weight of mice in different groups, N  = 6. C Serum testosterone content in different groups of mice, N  = 6. D Serum IL-6 content in different groups of mice, N  = 6. E Serum TNF-α content in different groups of mice, N  = 6. F HE-stained sections of mouse ovary in different groups. Scale, 500 μM. G Immunohistochemical section of CDH4 in mouse ovary. Scale, 100 μM. H mRNA levels of Cdh4 in the ovaries of mice from the WT + oil and WT + DHEA groups determined by qPCR and expressed as −ΔCT, N  = 6. I Yap1 mRNA levels expressed as −ΔCT. Data are shown for different experimental groups, N  = 6. J Protein levels of YAP1, Bcl-2, and Caspase-3 in ovary of mice in different groups Role of CDH4 in a mouse model of PCOS induced by DHEA. A The PCOS mouse model was established via 21-day continuous subcutaneous injection of DHEA (60 mg/kg; control mice received the oil vehicle alone). B Body weight of mice in different groups, N  = 6. C Serum testosterone content in different groups of mice, N  = 6. D Serum IL-6 content in different groups of mice, N  = 6. E Serum TNF-α content in different groups of mice, N  = 6. F HE-stained sections of mouse ovary in different groups. Scale, 500 μM. G Immunohistochemical section of CDH4 in mouse ovary. Scale, 100 μM. H mRNA levels of Cdh4 in the ovaries of mice from the WT + oil and WT + DHEA groups determined by qPCR and expressed as −ΔCT, N  = 6. I Yap1 mRNA levels expressed as −ΔCT. Data are shown for different experimental groups, N  = 6. J Protein levels of YAP1, Bcl-2, and Caspase-3 in ovary of mice in different groups To investigate the metabolic characteristics of PCOS, a PCOS mouse model of insulin resistance was established through continuous subcutaneous administration of letrozole along with a high-fat (HF) diet for 30 days, while the control group received an equal volume of oil vehicle and was fed a standard diet (Fig.  3 A). WT and KO-CDH4 mice were divided into six groups: WT + oil, KO + oil, WT + HF, KO + HF, WT + HF + letrozole (PCOS), and KO + HF + letrozole (KO-CDH4-PCOS). A comparative analysis of body weight results revealed that a high-fat (HF) diet increased mouse weight, and this effect was more pronounced when combined with letrozole. Specifically, mice in the WT + HF + letrozole model exhibited a significant weight gain compared with the WT + oil, WT + HF, and KO + HF + letrozole groups (Fig.  3 B). At baseline, no significant differences were observed between the WT + oil and KO + oil groups across multiple parameters, including follicular development, serum testosterone, IL-6, TNF-α levels, oral glucose tolerance test (OGTT) and the corresponding area under the curve (AUC), and insulin levels (Fig.  3 B–I). Biochemical analyses revealed elevated serum levels of testosterone, IL-6, and TNF-α in the WT + HF + letrozole group compared with the control and KO + HF + letrozole groups (Fig.  3 C–E). Notably, while serum testosterone showed no significant change between the WT + HF and WT + oil groups, the WT + HF group displayed higher levels of IL-6, TNF-α, and insulin, along with impaired glucose tolerance (OGTT/AUC), compared with the WT + oil group (Fig.  3 B–H). Furthermore, the IL-6 and insulin levels were significantly higher in the WT + HF group than in the KO + HF group (Fig.  3 B–H). Metabolic assessment through OGTT and the corresponding AUC analysis indicated pronounced insulin resistance in WT + HF + letrozole group, which was partially ameliorated in the mice from the KO + HF + letrozole group (Fig.  3 F, G). Consistently, fasting serum insulin levels were significantly greater in the WT + HF + letrozole group than in both the control and KO + HF + letrozole groups (Fig.  3 H). Fig. 3 Knockout of CDH4 can alleviate metabolic dysfunction and inflammatory responses in a PCOS mouse model induced by letrozole. A The PCOS mouse model was established by 30 days of continuous subcutaneous injection of letrozole (50 μg/day) combined with a high-fat diet. Control groups were set as follows: control (oil injection + regular diet); high-fat diet control. B Body weight of mice in different groups, N  = 6. C Serum testosterone content in different groups of mice, N  = 6. D Serum IL-6 content in different groups of mice, N  = 6. E Serum TNF-α content in different groups of mice, N  = 6. F Blood glucose levels at each time point in the OGTT experiment in different groups of mice, N  = 6. G The total area under the blood glucose curve in different groups of mice during the OGTT experiment, N  = 6. H Serum insulin levels in mice, N  = 6. I HE-stained section of mouse ovary. Scale, 500 μM. J mRNA levels of Cdh4 in mouse ovaries measured by qPCR and expressed as −ΔCT, N  = 6. K mRNA levels of Yap1 in mouse ovaries expressed as −ΔCT, N  = 6. L Protein levels of CDH4, YAP1, Bcl-2, and Caspase-3 in ovary of mice in different groups Knockout of CDH4 can alleviate metabolic dysfunction and inflammatory responses in a PCOS mouse model induced by letrozole. A The PCOS mouse model was established by 30 days of continuous subcutaneous injection of letrozole (50 μg/day) combined with a high-fat diet. Control groups were set as follows: control (oil injection + regular diet); high-fat diet control. B Body weight of mice in different groups, N  = 6. C Serum testosterone content in different groups of mice, N  = 6. D Serum IL-6 content in different groups of mice, N  = 6. E Serum TNF-α content in different groups of mice, N  = 6. F Blood glucose levels at each time point in the OGTT experiment in different groups of mice, N  = 6. G The total area under the blood glucose curve in different groups of mice during the OGTT experiment, N  = 6. H Serum insulin levels in mice, N  = 6. I HE-stained section of mouse ovary. Scale, 500 μM. J mRNA levels of Cdh4 in mouse ovaries measured by qPCR and expressed as −ΔCT, N  = 6. K mRNA levels of Yap1 in mouse ovaries expressed as −ΔCT, N  = 6. L Protein levels of CDH4, YAP1, Bcl-2, and Caspase-3 in ovary of mice in different groups Analysis of the estrous cycle revealed that CDH4 knockout partially restored normal cyclicity, characterized by the reappearance of proestrus, estrus, and diestrus stages. This finding contrasts with the persistent disruption of the estrous cycle observed in the letrozole-induced PCOS mouse model (Supplementary Fig. S2A). The histopathological examination revealed that, compared with the WT + HF + letrozole model mice, KO + HF + letrozole model mice presented reduced antral and cystic follicles and improved ovarian morphology (Fig.  3 I and Supplementary Fig. S2B). Molecular analysis revealed upregulation of both CDH4 and YAP1 at the mRNA and protein levels in PCOS ovarian tissues relative to those in controls (Fig.  3 J–L and Supplementary Fig. S3E). The results of immunohistochemistry showed that the expression of CDH4 in the ovaries of the PCOS group was higher than that in the control group (Supplementary Fig. S3A). Both transcriptional and translational expression levels of YAP1 were significantly lower in the KO + HF + letrozole mice than in the WT + HF + letrozole model group (Fig.  3 K, L and Supplementary Fig. S3E). Apoptosis-related marker analysis revealed decreased cleaved Caspase-3 levels (cl-Caspase-3), increased Bcl-2 and PCNA expression in PCOS ovaries compared with those in the control and KO + HF + letrozole groups (Fig.  3 L, Supplementary Fig. S3B, E, F). The above findings collectively suggested that CDH4 plays a crucial role in PCOS pathogenesis and may serve as a potential therapeutic target for this condition, as its suppression alleviates key PCOS symptoms, including metabolic dysfunction, reproductive abnormalities, and ovarian morphological alterations. The functional role of CDH4 was further investigated in vitro using lentiviral vectors encoding CDH4-specific shRNAs and thus establishing stable KGN cell lines with CDH4 knockdown (Fig.  4 A, B and Supplementary Fig. S4A). The subsequent colony formation assays revealed that the colony number and area of CDH4-knockdown cells were significantly lower and smaller than those of the negative control group cells (Fig.  4 C and Supplementary Fig. S4D), suggesting impaired cell proliferation upon CDH4 suppression. This finding was further corroborated by the results of the CCK-8 assays (Fig.  4 D). Fig. 4 Functional implications of CDH4 downregulation in the KGN cells stimulated with DHEA. A Cdh4 mRNA expression after knockdown, N  = 4. B CDH4 protein expression after knockdown. C Colony formation assays for indicated cells. D CCK-8 assay for indicated cells, N  = 4. E Protein expression levels of different molecules in KGN cells treated with 20 μM DHEA for 2 days. F Representative Mito-Tracker images (and quantification) of cells treated with 20 μM DHEA for 2 days. Scale, 100 μM, N  = 3. G Representative TUNEL images (and quantification) of cells treated with 20 μM DHEA for 2 days. Scale, 100 μM N  = 3 Functional implications of CDH4 downregulation in the KGN cells stimulated with DHEA. A Cdh4 mRNA expression after knockdown, N  = 4. B CDH4 protein expression after knockdown. C Colony formation assays for indicated cells. D CCK-8 assay for indicated cells, N  = 4. E Protein expression levels of different molecules in KGN cells treated with 20 μM DHEA for 2 days. F Representative Mito-Tracker images (and quantification) of cells treated with 20 μM DHEA for 2 days. Scale, 100 μM, N  = 3. G Representative TUNEL images (and quantification) of cells treated with 20 μM DHEA for 2 days. Scale, 100 μM N  = 3 For in vitro PCOS modeling, KGN cells were treated with 20 μM DHEA. After CDH4 knockdown, a decrease in YAP1 expression, an increase in the cleaved Caspase-3 (cl-Caspase-3) levels, and a reduction in Bcl-2 expression were observed, indicating enhanced apoptosis. Additionally, the expression of the inflammatory factors IL-6 and TNF-α was significantly reduced, suggesting that CDH4 knockdown attenuated inflammation and proliferation in PCOS-like KGN cells (Fig.  4 E and Supplementary Fig. S4B, C). The mitochondrial membrane potential, assessed through Mito-tracker staining, was markedly decreased in CDH4-knockdown cells (Fig.  4 F), further supporting the role of CDH4 in maintaining mitochondrial function. Consistent with these findings, TUNEL staining analysis demonstrated a significant increase in apoptosis in CDH4-knockdown KGN cells (Fig.  4 G). In summary, the in vitro experiments demonstrated that CDH4 knockdown inhibits inflammation and abnormal proliferation in PCOS-model KGN cells, highlighting the potential role of CDH4 in regulating the cellular processes associated with PCOS pathogenesis. Previous studies have demonstrated a close association between CDH4 and the Hippo signaling pathway (Fig.  1 A–C). On the basis of this evidence, it was hypothesized that CDH4 might regulate YAP1, the core effector of the Hippo pathway. The experimental results confirmed that knockdown of CDH4 led to significant downregulation of YAP1 at mRNA and protein levels (Fig.  5 A, B and Supplementary Fig. S4E), whereas overexpression of CDH4 resulted in upregulation of YAP1 expression (Fig.  5 C, D and Supplementary Fig. S4F). Since YAP1 activity is primarily governed by its nuclear localization, which in turn is regulated by phosphorylation, we proceeded to evaluate YAP1 protein levels. This indicated that CDH4 knockdown coordinately decreased both total and phosphorylated YAP1 protein levels (Fig.  5 B and Supplementary Fig. S4E), whereas CDH4 overexpression enhanced them (Fig.  5 D and Supplementary Fig. S4F). Subcellular fractionation and western blot analysis verified that these alterations occurred in both the nuclear and cytoplasm (Supplementary Fig. S4H, I). Collectively, these data demonstrate that CDH4 regulates Yap1 at the transcriptional level, consequently governing the abundance of YAP1 protein that is available for phosphorylation and nuclear translocation. Immunofluorescence analysis further supported these findings, revealing reduced YAP1 fluorescence intensity upon CDH4 knockdown (Fig.  5 E) and increased YAP1 fluorescence intensity following CDH4 overexpression (Fig.  5 F). Fig. 5 CDH4-mediated regulation of YAP1 and its functional role in KGN cells stimulated by DHEA. A The Yap1 mRNA level after knocking down CDH4, N  = 4. B The YAP1 and p-YAP1 protein level after knocking down CDH4. C The YAP1 mRNA level after overexpressing CDH4, N  = 4. D The YAP1 and p-YAP1 protein level after overexpressing CDH4. E Fluorescence intensity of YAP1 after knocking down CDH4. Scale, 50 μM, N  = 3. F Fluorescence intensity of YAP1 after overexpressing CDH4. Scale, 50 μM, N  = 3. G Mito-Tracker staining in KGN cells under different treatments for 2 days. Scale, 100 μM, N  = 3. H Protein levels of IL-6 and TNF-α in KGN cells after 2-day culture under different treatments CDH4-mediated regulation of YAP1 and its functional role in KGN cells stimulated by DHEA. A The Yap1 mRNA level after knocking down CDH4, N  = 4. B The YAP1 and p-YAP1 protein level after knocking down CDH4. C The YAP1 mRNA level after overexpressing CDH4, N  = 4. D The YAP1 and p-YAP1 protein level after overexpressing CDH4. E Fluorescence intensity of YAP1 after knocking down CDH4. Scale, 50 μM, N  = 3. F Fluorescence intensity of YAP1 after overexpressing CDH4. Scale, 50 μM, N  = 3. G Mito-Tracker staining in KGN cells under different treatments for 2 days. Scale, 100 μM, N  = 3. H Protein levels of IL-6 and TNF-α in KGN cells after 2-day culture under different treatments To investigate the functional role of YAP1 in PCOS, cells were treated with 1 μM Verteporfin (VP), a YAP1 inhibitor, in combination with DHEA. Compared with the DHEA group, the Verteporfin + DHEA group presented a significant reduction in the mitochondrial membrane potential (Fig.  5 G). Additionally, the levels of the IL-6 and TNF-α were markedly lower in the Verteporfin + DHEA group than in the DHEA group (Fig.  5 H and Supplementary Fig. S4G). In summary, CDH4 was indicated to positively regulate YAP1 expression and inhibition of YAP1 attenuates inflammation and abnormal proliferation in PCOS-like cellular models. These findings reveal a link between CDH4, YAP1, and PCOS pathology. To elucidate the molecular mechanism through which CDH4 regulates YAP1, a core effector of the Hippo pathway, co-immunoprecipitation (IP) and mass spectrometry (MS) experiments were performed. Given that CDH4 is a cadherin and unlikely to directly modulate the mRNA levels of YAP1 , this study focused on identifying the potential interacting partners of CDH4 that could mediate this regulation. The specific binding peptides associated with CDH4 were sequenced and ranked by abundance, which revealed the RNA motif protein RBMXL1 among the top ten candidate molecules (Fig.  6 A). Fig. 6 Mechanism through which CDH4 regulates YAP1. A The mass spectrum of CDH4 specifically binds the first ten molecules of the protein. B Co-IP of CDH4 and RBMX. C Fluorescence co-localization of CDH4 and RBMX with scale bar of 20 μM. D The protein level of RBMX after knocking down CDH4. E RBMX protein expression after knockdown. F Rbmx mRNA expression after knockdown, N  = 4. G The mRNA of Yap1 after knocking down RBMX, N  = 4. H Yap1 mRNA stability assay. Cells were transfected with siRBMX or siNC, followed by treatment with actinomycin D (5 µg/mL). Yap1 mRNA levels were quantified by qRT-PCR at the indicated times (0, 1, 2, 4, 6, and 8 h) post-treatment. We normalized the mRNA level at the 0-h time point to 1 for each group (siNC and siRBMX), using it as the baseline, N  = 3. I YAP1 protein levels after RBMX knockdown. J Fluorescence intensity of YAP1 after knocking down RBMX. Scale, 50 μM N  = 3 Mechanism through which CDH4 regulates YAP1. A The mass spectrum of CDH4 specifically binds the first ten molecules of the protein. B Co-IP of CDH4 and RBMX. C Fluorescence co-localization of CDH4 and RBMX with scale bar of 20 μM. D The protein level of RBMX after knocking down CDH4. E RBMX protein expression after knockdown. F Rbmx mRNA expression after knockdown, N  = 4. G The mRNA of Yap1 after knocking down RBMX, N  = 4. H Yap1 mRNA stability assay. Cells were transfected with siRBMX or siNC, followed by treatment with actinomycin D (5 µg/mL). Yap1 mRNA levels were quantified by qRT-PCR at the indicated times (0, 1, 2, 4, 6, and 8 h) post-treatment. We normalized the mRNA level at the 0-h time point to 1 for each group (siNC and siRBMX), using it as the baseline, N  = 3. I YAP1 protein levels after RBMX knockdown. J Fluorescence intensity of YAP1 after knocking down RBMX. Scale, 50 μM N  = 3 RBMXL1 is a retrogene of the RNA, binding motif protein X-linked (RBMX) located on the X chromosome. RBMX and its retrogene RBMXL1 share 95% protein homology, rendering them indistinguishable when using commercial RBMX antibodies and western blotting. The proteins encoded by RBMX and RBMXL1 are collectively referred to as RBMX. Accordingly, it was hypothesized that CDH4 interacts with RBMX and, thereby, potentially influences the mRNA and protein levels of YAP1 [ 56 ]. Co-immunoprecipitation experiments confirmed the physical interaction between CDH4 and RBMX (Fig.  6 B), supporting the hypothesis that this interaction may play a role in regulating YAP1 expression. Immunofluorescence co-localization experiments demonstrated that RBMX binds to CDH4, and the fluorescence intensity of RBMX increased upon CDH4 downregulation (Fig.  6 C). Consistent with this observation, the western blot analysis revealed increased RBMX protein levels upon CDH4 knockdown (Fig.  6 D and Supplementary Fig. S5A). To further validate the role of RBMX, small interfering RNAs (siRNAs) were used to reduce RBMX expression, which resulted in significant decreases in both mRNA and protein levels of RBMX (Fig.  6 E, F and Supplementary Fig. S5B). Supplementary tables (NIHMS1703553) from the published literature (PMID34458856) indicate that RBMX binds to the intron region of YAP1 and thereby negatively regulates its mRNA level [ 56 ]. This finding was corroborated by the results of the qPCR experiments in this study, which revealed that the Yap1 mRNA level was increased following RBMX knockdown (Fig.  6 G). To investigate the mechanism underlying this regulation, RNA stability assays were performed using RBMX-knockdown cells. Actinomycin D was added to both negative control (NC) and siRBMX groups at different time points to inhibit RNA synthesis. The results demonstrated that YAP1 mRNA degradation was decelerated in the siRBMX group compared with the NC group, further confirming the regulatory role of RBMX in Yap1 mRNA stability (Fig.  6 H). Western blot analysis and immunofluorescence assays further supported these findings, demonstrating that both YAP1 protein levels and fluorescence intensity were significantly increased upon RBMX downregulation (Fig.  6 I, J and Supplementary Fig. S5C). The above results indicated that CDH4 negatively regulates the RBMX protein levels, whereas RBMX, in turn, negatively regulates both mRNA and protein levels of YAP1. Thus, it was understood that CDH4 modulates YAP1 expression through a direct interaction with RBMX. Given that RBMX expression was downregulated in ovarian tissues from a PCOS mouse model compared with controls, we sought to investigate its functional role in KGN cells (Supplementary Figs. S1C and S3B, C, E). To this end, a series of functional assays were conducted following RBMX knockdown. The colony formation assay revealed that both the number and the area of colonies significantly increased after RBMX knockdown (Fig.  7 A and Supplementary Fig. S5D, E). Consistent with this finding, the results of CCK-8 assays demonstrated enhanced cell proliferation upon RBMX knockdown (Fig.  7 B), consistent with the proliferative effects mediated by YAP1 upregulation. Fig. 7 RBMX-dependent regulation of YAP1 by CDH4. A Colony formation assays for indicated cells. B CCK-8 assay for indicated cells, N  = 4. C Protein expression levels of different molecules under 20 μM DHEA culture for 2 days. D , E Mito-Tracker staining in NC and RBMX-knockdown cells treated with 20 μM DHEA for 2 days. Scale, 100 μM, N  = 3. F CCK-8 assay for functional recovery experiments, N  = 4. G Protein expression levels of different molecules in functional recovery experiments under 20 μM DHEA culture for 2 days. H , I Representative images and quantification of Mito-Tracker staining for different groups under 20 μM DHEA for 2 days. Scale, 100 μM, N  = 3 RBMX-dependent regulation of YAP1 by CDH4. A Colony formation assays for indicated cells. B CCK-8 assay for indicated cells, N  = 4. C Protein expression levels of different molecules under 20 μM DHEA culture for 2 days. D , E Mito-Tracker staining in NC and RBMX-knockdown cells treated with 20 μM DHEA for 2 days. Scale, 100 μM, N  = 3. F CCK-8 assay for functional recovery experiments, N  = 4. G Protein expression levels of different molecules in functional recovery experiments under 20 μM DHEA culture for 2 days. H , I Representative images and quantification of Mito-Tracker staining for different groups under 20 μM DHEA for 2 days. Scale, 100 μM, N  = 3 In KGN cells cultured with 20 μM DHEA, RBMX knockdown led to an increase in YAP1 protein levels, elevated expression of the antiapoptotic protein Bcl-2, and reduced levels of cleaved Caspase-3 (cl-Caspase-3), all of which indicated suppression of apoptosis. Additionally, the expressions of IL-6 and TNF-α were significantly increased (Fig.  7 C and Supplementary Fig. S5F, G). The mitochondrial membrane potential, assessed using Mito-Tracker fluorescence staining, was markedly increased following RBMX knockdown (Fig.  7 D, E). TUNEL assays further confirmed a reduction in apoptosis in RBMX-knockdown compared with control cells (Supplementary Fig. S5H). In summary, these results demonstrated that RBMX knockdown promotes cell proliferation, enhances the mitochondrial membrane potential, reduces apoptosis, and enhances inflammation in KGN cells, likely through upregulation of YAP1. On the basis of the above experimental findings, it was proposed that CDH4 modulates the development of PCOS through the regulation of the function of YAP1 dependent on RBMX. To investigate the functional relationship between CDH4 and RBMX in PCOS pathogenesis, a series of functional recovery experiments were conducted using four distinct cell line groups, cell lines with negative control, cell lines with shCDH4, cell lines with siRBMX, and cell lines with shCDH4 and siRBMX. Western blotting results revealed a different regulatory relationship between CDH4 and RBMX in modulating YAP1 expression. Specifically, CDH4 knockdown resulted in decreased YAP1 expression, whereas RBMX knockdown led to increased YAP1 levels. Notably, RBMX knockdown partially rescued the reduction in YAP1 expression induced by CDH4 knockdown (Fig.  7 G and Supplementary Fig. S5I). Similarly, on the basis of the expression levels of Bcl-2 and cl-Caspase-3, it was demonstrated that CDH4 knockdown suppressed KGN cell proliferation, whereas RBMX knockdown promoted cell proliferation and partially reversed the antiproliferative effects of CDH4 knockdown, as evidenced by the expression levels of Bcl-2 and cl-Caspase-3 (Fig.  7 G and Supplementary Fig. S5I, J). Inflammatory response assessment revealed that CDH4 knockdown reduced IL-6 and TNF-α levels, whereas RBMX knockdown increased the levels of these inflammatory markers and partially restored the inflammation that was suppressed upon CDH4 knockdown (Fig.  7 G and Supplementary Fig. S5I). CCK-8 assays confirmed that RBMX knockdown could partially counteract the inhibition of cell proliferation induced by CDH4 knockdown (Fig.  7 F). Mitochondrial function analysis indicated that RBMX knockdown partially restored the reduction in the mitochondrial membrane potential caused by CDH4 knockdown (Fig.  7 H, I). Apoptosis evaluation through TUNEL assay demonstrated that RBMX knockdown partially reversed the apoptotic effects induced by CDH4 knockdown (Supplementary Fig. S6A). These findings demonstrated that RBMX knockdown effectively counteracted the physiological consequences of CDH4 suppression, suggesting that CDH4 primarily affects YAP1 expression and PCOS pathophysiology through RBMX-mediated mechanisms. To elucidate the mechanism through which CDH4 regulates RBMX expression, the mRNA levels of Rbmx in CDH4-knockdown cells were determined. The results revealed that CDH4 knockdown did not significantly alter the mRNA levels of Rbmx (Fig.  8 A), suggesting posttranscriptional regulation. To investigate how CDH4 affects RBMX protein stability, cells were treated with cycloheximide (CHX) to inhibit protein synthesis, and RBMX degradation was monitored. Notably, RBMX protein degradation was significantly slower in CDH4-knockdown cells than in control cells (Fig.  8 B and Supplementary Fig. S6B), indicating that CDH4 depletion enhances RBMX protein stability. Fig. 8 CDH4/UBA1/RBMX complex mediates the ubiquitination and proteasomal degradation of RBMX in KGN cells. A The mRNA level of Rbmx after knocking down CDH4, N  = 4. B The protein level of RBMX after CDH4 was knocked down after cycloheximide was added at different times. C RBMX protein levels in different groups with MG132 and chloroquine. D Ubiquitination status of RBMX upon CDH4 knockdown assessed by IP. E Co-IP experiments of UBA1 with RBMX and CDH4, respectively. F IP experiment with RBMX antibody pull-down after knocking down CDH4. G The mRNA expression of Uba1 after overexpressing, N  = 4. H The protein expression of UBA1 after overexpressing. I The protein expression levels of UBA1 and RBMX in overexpressed UBA1 and added MG132 groups. J The UBA1 and RBMX protein levels following CHX chase in UBA1-overexpressing cells. K The protein expression level of UBA1 after knocking down CDH4 CDH4/UBA1/RBMX complex mediates the ubiquitination and proteasomal degradation of RBMX in KGN cells. A The mRNA level of Rbmx after knocking down CDH4, N  = 4. B The protein level of RBMX after CDH4 was knocked down after cycloheximide was added at different times. C RBMX protein levels in different groups with MG132 and chloroquine. D Ubiquitination status of RBMX upon CDH4 knockdown assessed by IP. E Co-IP experiments of UBA1 with RBMX and CDH4, respectively. F IP experiment with RBMX antibody pull-down after knocking down CDH4. G The mRNA expression of Uba1 after overexpressing, N  = 4. H The protein expression of UBA1 after overexpressing. I The protein expression levels of UBA1 and RBMX in overexpressed UBA1 and added MG132 groups. J The UBA1 and RBMX protein levels following CHX chase in UBA1-overexpressing cells. K The protein expression level of UBA1 after knocking down CDH4 To further explore the degradation pathway of RBMX mediated by CDH4, CDH4-knockdown KGN cells were treated either with the proteasome inhibitor MG132 or with the lysosome inhibitor chloroquine. Western blotting results demonstrated that MG132 treatment restored RBMX expression in CDH4-knockdown cells to levels comparable to those observed in negative control cells (Fig.  8 C), suggesting that CDH4 primarily regulates RBMX through the ubiquitin–proteasome pathway. This finding underscores the role of CDH4 in modulating RBMX ubiquitination. Consistent with these findings, results of endogenous ubiquitination assays revealed that CDH4 knockdown reduced RBMX ubiquitination levels (Fig.  8 D and Supplementary Fig. S6C), further supporting the involvement of ubiquitin-mediated degradation. To identify the molecular link between CDH4 and RBMX ubiquitination, CDH4-associated proteins were identified using mass spectrometry, which revealed ubiquitin-like modifier activating enzyme 1 (UBA1), a key ubiquitin-activating enzyme, as a prominent interactor. The mass spectrum results of CDH4 revealed that UBA1 was the site associated with the greatest degree of ubiquitination-related enzyme binding to CDH4. In addition, CDH4, RBMX, and UBA1 were bound to each other in the Co-IP experiments (Fig.  8 E). In IP experiments in CDH4 knockdown cells, it was found that UBA1 pulled down by RBMX was reduced when CDH4 was knocked down (Fig.  8 F and Supplementary Fig. S6D). Further studies on UBA1 revealed that the level of UBA1 was not significantly affected when CDH4 was knocked down (Fig.  8 K and Supplementary Fig. S6G). To further investigate the functional role of UBA1, UBA1 was overexpressed in KGN cells, and increased UBA1 mRNA and protein levels were confirmed (Fig.  8 G, H and Supplementary Fig. S6E). After adding CHX to the cells, western blot analysis was conducted, which revealed that the overexpression of UBA1 accelerated the degradation of RBMX (Fig.  8 J and Supplementary Fig. S6F). Conversely, UBA1 knockdown slowed RBMX degradation (Supplementary Fig. S6H–J). Experimental results showed that, when UBA1 was overexpressed, the level of RBMX protein decreased, and MG132 could block the degradation of RBMX (Fig.  8 I), indicating that UBA1 promotes RBMX degradation via the ubiquitin–proteasome pathway. Collectively, these findings suggested that CDH4, RBMX, and UBA1 formed a regulatory complex in KGN cells, wherein CDH4 modulates RBMX stability through UBA1-mediated ubiquitination and subsequent proteasomal degradation. This mechanism likely contributes to the regulation of YAP1 and its downstream effects in PCOS pathogenesis.

CDH4-KO C57BL/6J mice (cat no. T028771) were procured from GemPharmatech Co., Ltd. (Nanjing, China). This study utilized CRISPR/Cas9 to knock out the mouse Cdh4 gene. Targeting exon 3 of the major transcript Cdh4-201 (ENSMUST00000000314.12) eliminated a 227-bp coding sequence, which is expected to disrupt CDH4 protein function. Wild-type (WT) C57BL/6J mice were procured from Beijing HFK Bioscience Co., Ltd. (Beijing, China). To establish a PCOS model, we administered 4-week-old female mice daily subcutaneous injections of DHEA (60 mg/kg) for 21 days [ 9 , 29 , 30 ]. Control mice received equivalent sesame oil injections. Alternatively, a PCOS-like model was established using letrozole combined with a high-fat diet, for which 5-week-old female mice received subcutaneous injections of 50 μg of letrozole daily for 30 days [ 31 – 36 ]. The control group received subcutaneous injections of the sesame oil vehicle (equivalent volume to that of the experimental groups) daily and was maintained on a standard diet. All animal experiment procedures were reviewed and approved by the Animal Welfare and Ethics Review Committee of Beijing Obstetrics and Gynecology Hospital, Capital Medical University (approval no. BOGH21-2404-1). The animal experiments were then conducted in strict compliance with National Institutes of Health (NIH) guidelines for the care and use of laboratory animals. The human granulosa cell line KGN (cat. no. CTCC-003-0105) used in this study was obtained from the Meisen Chinese Tissue Culture Collection (Meisen CTCC). Cellular propagation was conducted at physiological temperature (37 °C) in a humidified atmosphere containing 5% carbon dioxide. Nutrient support was provided through high-glucose Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum and a 1% concentration of penicillin–streptomycin antibiotic mixture [ 37 , 38 ]. Short hairpin RNA sequences targeting human Cdh4 were designed, synthesized, and cloned into the hU6 promoter-driven multiple cloning site of the GV112 lentiviral vector. A nontargeting scrambled shRNA sequence was cloned into the same site to generate the control virus. Polybrene (Genechem) and Enhanced Infection Solution (Genechem) were used to transfect the KGN cell line with a lentivirus expressing low levels of Cdh4 and the corresponding control vector [ 39 ]. Three days post-transfection, we assessed the transfection efficiency by quantitative reverse-transcription polymerase chain reaction (qRT-PCR) and western blot. The plasmid was transfected (2 µg of the CDH4 plasmid introduced into about 3 × 10 5 cells) using Lipofectamine 3000 (Invitrogen, CA, USA). Additionally, 5 µg of the siRNA targeting Rbmx was transfected into about 3 × 10 5 cells using the same transfection reagent. Protein lysates and total RNA were collected 72 h post-transfection to assess the effectiveness of the transfection. The CDH4 coding sequence was cloned into the pCDH-CMV-MCS-EF1-Puro lentiviral expression vector. The full-length coding sequence of human UBA1 was cloned into the GV350 mammalian expression vector to generate a plasmid for expressing C-terminal 3 × FLAG-tagged UBA1 protein. The empty GV350 vector served as the control. The sequences of all siRNAs employed in our investigation are provided in Table  1 . Table 1 siRNA and shRNA target sequences Gene Sequence (5′–3′) siNC Guangzhou RiboBio Co., Ltd. siN0000001-1–5 siRBMX CGGATATGGTGGAAGTCGA siUBA1 CAAGGCTGTTACCCTACAT shCDH4 CGACCTGTACATCTACGTCAT siRNA and shRNA target sequences Tissue samples preserved in 4% paraformaldehyde and embedded in paraffin were sectioned for immunohistochemical analysis. Primary antibody treatment involved overnight exposure (12 h, 4 °C) to mouse-derived anti-CDH4 immunoglobulin. After incubation with the primary antibody, three phosphate-buffered saline (PBS) washes were performed to remove unbound antibodies. For signal detection, tissue sections were treated with 50 µL of secondary antibody and then maintained at room temperature (25 °C) for 120 min to facilitate antibody binding [ 40 ]. Nucleic acid extraction was performed on both tissue samples and cultured cells using a total RNA isolation kit (Magen; Guangzhou, CHN). For subsequent molecular analysis, 1 μg of the purified extracted RNA was reverse transcribed by using ABScript III RT Master Mix (ABclonal, China) to generate complementary DNA (cDNA) templates [ 41 ]. Quantitative polymerase chain reaction was then conducted using a Roche thermocycler (Switzerland) and Universal SYBR Green Fast qPCR reagents (ABclonal, China) [ 42 ], with reduced glyceraldehyde 3-phosphate dehydrogenase (GAPDH) used as the endogenous control for data normalization. Using −ΔCt = Ct GAPDH −Ct Gene , the 2 −ΔΔCt method was applied to determine target gene expression. All PCR primer sequences are provided in Table  2 . Table 2 Primer sequences for qRT-PCR Gene Species Sequences (5′–3'′) CDH4 Human Forward TCCGGTCCGACAAAGACAATG CDH4 Human Reverse CATGGGCCTTGTGACGTACAT YAP1 Human Forward GCAGTTGGGAGCTGTTTCTC YAP1 Human Reverse GCCATGTTGTTGTCTGATCG GAPDH Human Forward GAGTCAACGGATTTGGTCGT GAPDH Human Reverse GACAAGCTTCCCGTTCTCAG RBMX Human Forward GCACCACCACCACGAGATTA RBMX Human Reverse TCAATCAGCACTCCACGACC UBA1 Human Forward GGAACCGGCATTGATGTCCA UBA1 Human Reverse GGTCCAGTGTAGGCAGTGAC CDH4 Mouse Forward CTCTCAGCCGCCAAATGACA CDH4 Mouse Reverse AGGGTCCACGGGAGTTCTC YAP1 Mouse Forward TGAGATCCCTGATGATGTACCAC YAP1 Mouse Reverse TGTTGTTGTCTGATCGTTGTGAT GAPDH Mouse Forward AAGAGGGATGCTGCCCTTAC GAPDH Mouse Reverse TACGGCCAAATCCGTTCACA Primer sequences for qRT-PCR Protein extraction from cellular and tissue samples was performed using radioimmunoprecipitation assay (RIPA) lysis buffer (Beyotime, China) supplemented with protease inhibitors and phenylmethylsulfonyl fluoride (PMSF) [ 43 ]. For electrophoretic separation, 30 µg of the protein lysate was loaded onto sodium dodecyl sulfate (SDS)–polyacrylamide gels, and after electrophoretic migration, separated protein bands were transferred onto polyvinylidene fluoride (PVDF) membranes through electroblotting. Membrane blocking was then achieved through incubation with 5% skim milk solution at ambient temperature. Overnight incubation with primary antibody at 4 °C was conducted next using target-specific immunoglobulins. After extensive washing with PBS with Tween 20 (PBST) to eliminate the non-specifically bound proteins, the membranes were probed with species-specific secondary antibodies, either rabbit or mouse monoclonal antibodies (horseradish peroxidase (HRP)-conjugated goat anti-rabbit immunoglobulin G (IgG) (H + L) and HRP-conjugated goat anti-mouse IgG (H + L), purchased from ABclonal and diluted 1:5000) [ 44 ]. The following antibodies were employed: CDH4 (1:500; Santa Cruz, sc-398306), YAP1 (1:500; Santa Cruz, sc-376830), Phospho-YAP1-S127 (1:500; ABclonal, AP0489), RBMX (1:2000; ABclonal, A21963), UBA1 (1:2000; ABclonal, A9254), Ubiquitin (1:2000; Proteintech, 10201-2-AP), Bcl-2 (1:2000; ABclonal, A19693), Caspase-3 (1:2000; ABclonal, A2156), IL-6 (1:1000; Proteintech, 66146-1-Ig), TNF-α (1:1000; Proteintech, 60291-1-Ig), and GAPDH (1:50,000, Proteintech, 60004-1-Ig). For endogenous immunoprecipitation (IP) assays, KGN cells were lysed with cell lysis buffer (Beyotime). The lysate was then incubated with magnetic beads at 4 °C for 1–2 h to remove the non-specifically bound proteins, after which the beads were discarded. The lysate was subsequently incubated overnight at 4 °C with IgG or the target-specific antibody, along with new magnetic beads [ 45 , 46 ]. Following incubation, the lysate was removed, and the magnetic beads were washed three times to eliminate any unbound proteins. The bound proteins were eluted next by adding the loading buffer and heating at 100 °C for 10 min. The resulting samples were then loaded onto an SDS–polyacrylamide gel electrophoresis (PAGE) gel for further analysis [ 47 ]. KGN cells (5 × 10 7 ) were lysed and pretreated with 40 μL A/G magnetic beads (MCE, HY-K0202) for 1–2 h at 4 °C to remove nonspecific combination. The lysate was then incubated overnight at 4 °C with anti-CDH4 antibody or control IgG, together with 50 μL new A/G magnetic beads. After washing, bound proteins were eluted, and separated by SDS-PAGE, and gel bands were excised. Following destaining, reduction with dithiothreitol (DTT), alkylation with iodoacetamide (IAA), and in-gel tryptic digestion overnight at 37 °C, the extracted peptides were desalted and subjected to liquid chromatography (LC)–MS/MS analysis. Peptides were separated on an Acclaim PepMap RSLC C18 column using an Easy-nLC system at a flow rate of 300 nL/min. Eluted peptides were analyzed by a Q Exactive Plus mass spectrometer in positive ion mode. Raw data were searched against the UniProt human database using Proteome Discoverer 2.2 with carbamidomethylation of cysteine as a fixed modification and oxidation of methionine as a variable modification. Precursor and fragment mass tolerances were set to 10 ppm and 0.02 Da, respectively. Peptide and protein identifications were filtered at a false discovery rate (FDR) of ≤ 1%, and proteins with at least one unique peptide were considered confident. CDH4-specific interactors were defined by subtracting proteins identified in the IgG control. The MS protein sequencing and data analysis mentioned above were provided by GeneChem, Shanghai, China. For the assessment of cellular proliferation, 2 × 10 3 cells per well were seeded into the wells of 96-well culture plates. Cellular metabolic activity was quantified at 24-h intervals over 4 days using the CCK-8 assay. Cell viability was evaluated at 0, 1, 2, 3, and 4 days after treatment, at 450 nm [ 48 ]. Cells were seeded in six‑well plates at 2000 cells/well in complete medium and cultured for 7 days. Colonies were fixed, stained with Crystal Violet, and photographed. Colony number and size were quantified using ImageJ. Data are presented as mean ± standard deviation (SD) from three independent experiments and analyzed by Student’s t ‑test. To assess protein stability and degradation pathways, cells were treated with specific inhibitors. Cycloheximide (50 μM) was used to inhibit protein synthesis. To block proteasomal and autophagic degradation, cells were treated with MG132 (20 μM) for 12 h to inhibit proteasomal degradation, and with chloroquine (50 μM) for 12 h to inhibit autophagic–lysosomal degradation. Cycloheximide, MG132, and chloroquine were purchased from AbMole Bioscience. The cells were uniformly seeded onto climbing slides, and we performed fixation with 4% paraformaldehyde (10 min). Then cells were permeabilized with 0.5% saponin for 5 min and subsequently washed three times with phosphate-buffered saline (PBS) [ 49 ]. To prevent nonspecific binding, the cells were blocked with the immunostaining blocking solution for 1 h and then washed three times with the PBS [ 50 , 51 ]. Primary antibodies targeting specific molecules were then applied, and the samples were incubated overnight at 4 °C. After incubation, the primary antibodies were removed, followed by three washes. A species-specific fluorescent secondary antibody was added, followed by 1–2 h incubation at room temperature and another three washes. Finally, cells were mounted using 4′,6-diamidino-2-phenylindole (DAPI) anti-fade medium and examined by fluorescence microscopy. Cells of each treatment group were inoculated at the same density in 96-well plates. All groups were stained and imaged simultaneously. Images were obtained under the same settings on the fluorescence microscope. For each experimental group, images were randomly captured from at least three independent biological repeat wells ( N  ≥ 3). All acquired images were analyzed in a blinded manner using ImageJ software, where the mean fluorescence intensity per field of view was quantified as an individual data point. Final plotted values and statistical comparisons (Student’s t ‑test or one‑way analysis of variance (ANOVA)) are based on the mean fluorescence intensity calculated from the pooled data of at least three biological replicates ( N  ≥ 3). The research utilized multiple data sources, including the transcriptomic profiles from the GSE168404 dataset acquired through the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus platform, the reported PMID 29344314 collected through the PubMed database [ 52 ], and the Hippo signaling pathway information obtained from the Gene Set Enrichment Analysis database. The GSE168404 dataset includes mRNA-seq profiles of ovarian GCs from five patients with PCOS and five matched controls. In addition, data from the study associated with PMID 29344314 also originate from human ovarian granulosa cells, with samples obtained from three nonobese patients with PCOS and three comparable control women. For the identification of significant gene expression changes, stringent selection criteria were applied, requiring p   1 [ 53 ]. The estrous cycle phase was assessed by analyzing the dominant cell populations in the vaginal smear samples. In the last 10 days of modeling, the vaginal epithelial cells shed from the mice were collected by lavaging the vaginal cavity with physiological saline. These samples were then subjected to Wright stain solution and observed under a microscope [ 29 ]. The mice were subjected to 16 h of fasting, which started at 5 p.m. and ended at 9 a.m. the following day. Next, the fasted mice were administered a glucose solution by oral gavage. The solution concentration was 200 mg/mL in normal saline; the dosage administered was 2 g/kg. We then measured blood glucose levels at six time points following administration: 0 min and then 30, 60, 90, 120, and 180 min post-administration [ 54 , 55 ]. Samples were obtained from the tail vein and analyzed using a Sinocare blood glucose monitoring system. Statistical analyses were performed using Microsoft Excel 2020 and GraphPad Prism version 9.5 software packages. The experimental data derived from in vitro investigations were collected from three independent replicates and are expressed as the mean with standard deviation (SD). Comparisons between two groups were performed using Student’s t -test, and linear relationships were assessed on the basis of the Pearson correlation coefficient. For pairwise comparisons among three or more groups, one-way ANOVA was conducted, followed by Tukey’s post hoc correction. A probability threshold of p  < 0.05 was established to determine statistical significance [ 53 ].

This study revealed that CDH4 is a key driver of the pathogenesis of PCOS through the regulation of the UBA1–RBMX–YAP1 axis. It was demonstrated that CDH4 regulates Yap1 mRNA by facilitating the UBA1-mediated ubiquitination of RBMX, which subsequently promotes GC proliferation and inflammation. These findings provide the first evidence linking the cadherin family of proteins to the posttranscriptional regulation in PCOS. Accordingly, it was proposed that the mechanistic cascade involved three critical steps: First, CDH4 interacts physically with UBA1 to form an E1 ubiquitin-activating complex. Second, this complex targets RBMX for proteasomal degradation, as evidenced by the results of the MG132 rescue experiments. Third, RBMX negative feedback regulates the stability of Yap1 mRNA. This multistep regulation explains the proliferative as well as the inflammatory phenotypes observed in PCOS. PCOS is a complex endocrine disorder with heterogeneous clinical manifestations, and no definitive cure is available currently. Metformin combined with lifestyle modification remains the first-line therapy at present, although optimization of individualized treatment strategies is warranted. Insulin resistance (IR) and obesity are the key pathophysiological features of PCOS. Future therapeutic breakthroughs may focus on targeting energy metabolism pathways (e.g., AMPK activators to improve the ovarian microenvironment) and addressing hyperandrogenemia and chronic inflammation through targeted therapies. It is further proposed that YAP1, a key effector of the Hippo signaling pathway, could be utilized as a novel therapeutic target to mitigate both inflammation and hyperandrogenemia in PCOS. In this study, CDH4 was identified as an upregulated gene in PCOS, with its elevated expression levels correlated to PCOS-related dysfunction. While CDH4 has been studied extensively in the context of cancer metastasis, its potential involvement in endocrine pathologies remains largely uncharacterized. Notably, the role of CDH4 in cellular heterodifferentiation processes, particularly in PCOS, remains to be investigated to date. The discovery of the involvement of RBMX ubiquitination in this study extends the prior reports, evidencing the protein-binding capacity of CDH4. This discrepancy may reflect the tissue-specific ubiquitin ligase recruitment patterns. As with all research, this study also has certain limitations. While the KGN cell line serves as a valuable in vitro model for studying granulosa cell biology, it cannot fully recapitulate the complex ovarian microenvironment, including the paracrine signaling, that may influence CDH4 function in vivo. Also, regarding the mechanistic exploration of the Hippo/YAP pathway, the KGN cell line may not fully represent the functional state of native granulosa cells within the complex ovarian milieu of PCOS. Consequently, CDH4 may be a key factor in the pathogenesis of PCOS acting via the UBA1–RBMX–YAP1 pathway, but further investigation is warranted. Therefore, to further advance these findings, future studies should use conditional knockout (KO) mice with CDH4 to verify the effects in vivo to comprehensively evaluate the role of CDH4 in ovarian follicle development and the systemic PCOS phenotype in a physiologically relevant context. Second, clinical studies have to be conducted to evaluate the potential diagnostic utility of CDH4 by determining its expression levels in relation to in vitro fertilization (IVF) treatment outcomes across well-characterized cohorts of patients with PCOS. Such investigations could establish CDH4 as a clinically relevant biomarker. Collectively, these studies will position CDH4 as a pivotal molecular link connecting cell adhesion processes with transcriptional regulation in PCOS pathogenesis. The involvement of CDH4 in both structural and signaling pathways in this study underscores its potential as a multifunctional therapeutic target in this complex endocrine disorder. This study elucidated a novel molecular pathway involved in PCOS pathogenesis, revealing the critical role of the CDH4–RBMX–YAP1 regulatory axis in PCOS. Through comprehensive in vitro and in vivo studies, we found that CDH4 promotes granulosa cell dysfunction by facilitating the UBA1-mediated ubiquitination of RBMX, thereby increasing YAP1 stability and activity. These findings not only advance the understanding of PCOS pathophysiology at the molecular level but also position CDH4 as a promising therapeutic target for PCOS. The CDH4/UBA1/RBMX axis offers promising translational opportunities for PCOS diagnosis and therapy. From a diagnostic perspective, circulating soluble CDH4 could serve as a novel biomarker for disease stratification and monitoring. Therapeutically, several targeted approaches emerge: tissue-specific UBA1 inhibitors delivered via ovarian-targeted nanoparticles could selectively prevent RBMX degradation, thereby restoring its regulatory role in granulosa cell function; YAP1-focused interventions mitigate inflammation and hyperandrogenemia through Hippo pathway modulation. The therapeutic potential of targeting UBA1 and YAP1 has been demonstrated in diverse preclinical murine models. For instance, the UBA1 inhibitor TAK-243 has been shown to reduce leukemic burden and target leukemia stem cells in vivo without evidence of toxicity, and mutations conferring resistance to TAK-243 were mapped to the UBA1 adenylation domain, confirming its on-target activity in malignant cells and supporting its clinical investigation [ 57 ]. Similarly, the YAP1 antagonist verteporfin (VP) inhibits keratinocyte proliferation and inflammatory cytokine production in a dose-dependent manner. In psoriasis models, VP alleviates disease by impeding epidermal hyperplasia and systemic inflammation [ 58 ]. Moreover, in endometriosis, VP synergizes with progestins to cause lesion regression and improve endometrial decidualization, with progestin treatment itself found to reduce YAP1 expression [ 59 ]. These targeted agents could be strategically integrated with established metabolic therapies to form personalized regimens. However, successful translation will require challenges in tissue-selective drug delivery and long-term safety validation to be overcome. Overall, the discovery of this previously unrecognized regulatory mechanism provides a conceptual framework for developing targeted interventions that would address both reproductive and metabolic manifestations of PCOS. Future studies focusing on the translational potential of these findings could facilitate the development of novel treatment strategies for this complex endocrine disorder.

This study underscores the critical role of YAP1, the core effector of the Hippo signaling pathway, in the pathogenesis of PCOS. CDH4 regulates YAP1 through an interaction with RBMX. CDH4 modulates RBMX stability by binding to UBA1, thereby promoting RBMX ubiquitination and proteasomal degradation. Therefore, CDH4 may be a key factor in the pathogenesis of PCOS acting via the UBA1–RBMX–YAP1 pathway, but further investigation is warranted.

Polycystic ovary syndrome (PCOS) is a multifaceted endocrine disorder with a prevalence of 5–18% among the female population [ 1 , 2 ]. PCOS has important clinical implications and can lead to health conditions associated with lipid metabolism, such as obesity, metabolic syndrome, insulin resistance, and type 2 diabetes [ 3 – 6 ]. Furthermore, PCOS manifests as a persistent low-grade inflammatory state, as evidenced by the chronically elevated circulating levels of inflammatory factors, including tumor necrosis factor-alpha (TNF-α), C-reactive protein (CRP), and interleukins (IL-6, IL-8, and IL-18) [ 7 ]. PCOS has three typical clinical features, namely elevated androgen levels, impaired ovulatory function, and multiple ovarian cysts, and is the main cause of ovulation infertility [ 8 – 10 ]. There are multiple cystic follicles in an ovary. The density of the small preantral follicles is increased and these abnormalities observed in anovulatory PCOS are further defined as abnormal granulosa cell proliferation [ 11 ]. Abnormal proliferation of granular cells (GCs) is an important cause of PCOS [ 12 ]. However, the mechanism underlying the development of these abnormal follicles remains unclear [ 13 ]. The pathological condition manifests as inadequate follicular maturation signals, which refers to the dysregulation of key signaling pathways governing follicular development. This dysregulation impedes the normal maturation process, ultimately resulting in anovulation. Consequently, this condition is characterized by disturbances in both endocrine and intraovarian signaling, and frequently co-occurs with metabolic disorders, particularly impaired insulin sensitivity and obesity [ 8 , 14 ]. Structurally, the ovarian follicle is composed of multiple layers, featuring an external theca cell (TC) population that is responsible for androgen synthesis and an inner GC compartment of greater thickness. External theca cells (TCs) are less likely to initiate follicle rupture and ovulation as they are separated from oocytes by multiple layers of GCs and may not receive the oocyte maturation signals, which are a prerequisite for ovulation. In contrast, granulosa cells maintain direct communication with the oocyte through follicular fluid-mediated signaling pathways and are, therefore, considered to be the primary mediators of follicular rupture initiation [ 14 ]. Consequently, GCs have a high rate of apoptosis during the major follicular selection and preovulation maturation stages of folliculation, whereas PCOS GCs display reduced apoptosis. The apoptosis rate of GCs in patients with PCOS is significantly lower than that in women with normal ovulation cycles. The lower apoptosis rate in the PCOS group was correlated to a decreased level of caspase-3 and an increased level of Bcl-2, an antiapoptotic survival factor [ 15 , 16 ]. Biomechanically, the mechanosensitive Hippo signaling pathway is significantly influenced by structural alterations in the polycystic ovaries [ 17 ]. Suppression of Hippo pathway induces sustained YAP1 activation, inducing two distinct pathological consequences: stromal tissue enlargement and theca cell hyperproliferation. These cellular changes subsequently drive excessive androgen synthesis by hyperplastic glandular cells and simultaneously impair follicular development, leading to the accumulation of numerous arrested immature follicles [ 18 ]. Most studies have focused on the role of YAP1, which functions downstream in the Hippo pathway. Forced overexpression of YAP1 in studies led to cell reprogramming and loss of particle identity. As a PCOS susceptibility gene, YAP1 is implicated in Hippo pathway dysregulation. Furthermore, genome-wide association studies (GWASs) demonstrated that the single-nucleotide polymorphisms (SNPs) rs11225161 and rs11225138 from the YAP1 gene are associated with PCOS, such as impaired glucose tolerance and elevated luteinizing hormone (LH) [ 19 – 22 ]. Studies employing a range of models, from in vivo mouse systems to in vitro cultures, have suggested that dysregulation of the Hippo pathway leads to overactivation of YAP1 in granulosa cells, ultimately impairing tissue homeostasis and fertility. YAP1 is linked to both the ovarian PCOS phenotype and metabolic disorders [ 18 , 23 ]. Hyperandrogen leads to overactivation of YAP1, which in turn contributes to a reduced ovulation rate [ 24 ]. Elevated YAP1 results in a decreased expression of the mRNA transcripts of Wnt4, FSHR, RHGR, and CYP19A1, which has been previously identified as essential for GC function and differentiation [ 25 ]. A dysregulation of the Hippo pathway may lead to elevated YAP1 expression and enlarged ovaries and ovarian cysts, which are characteristic of PCOS [ 14 , 26 ]. By increasing interleukin 6 (IL-6) expression, YAP1 promotes proinflammatory responses [ 27 , 28 ]. Cadherin 4 (CDH4, R-Cadherin) is a classical cadherin from the cadherin superfamily that participates in calcium-dependent cell adhesion and regulates cell proliferation, cell arrangement, and tissue development. It may, therefore, be involved in the sorting of heterogeneous cell types. However, its mechanism is still unclear in PCOS. This study confirmed high CDH4 expression in PCOS through both in vitro and in vivo approaches. The downregulation of CDH4 significantly inhibited the degree of PCOS and the abnormal proliferation of GCs via YAP1. The in vivo experiments demonstrated that a reduction in CDH4 alleviated inflammation and abnormal proliferation in dehydroepiandrosterone (DHEA)-induced PCOS mice. Moreover, in a letrozole-induced insulin-resistant PCOS mouse model, decreased CDH4 levels were observed to be associated with reduced body weight, improved insulin resistance, and attenuated inflammatory responses. Therefore, we speculate that CDH4 may be involved in function of PCOS through the regulation of RBMX-mediated YAP1 in granulosa cells. CDH4 regulates YAP1 through RBMX, and partial reversal of this process can inhibit PCOS progression. These findings provide the rationale for using CDH4 as a potential therapeutic target for PCOS.

Supplementary material 1. Supplementary material 2. Supplementary material 1. Supplementary material 2.