Multiomics analyses reveal key circadian rhythm genes implicated in Premature ovarian insufficiency

Multiomics analyses reveal key circadian rhythm genes implicated in Premature ovarian insufficiency | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Multiomics analyses reveal key circadian rhythm genes implicated in Premature ovarian insufficiency yong tan, Jin Ren, Jie chen This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6850329/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Premature ovarian insufficiency (POI), defined as the loss of ovarian function before the age of 40 years, severely disrupts reproductive and endocrine health. While circadian rhythm disruption has been implicated in reproductive aging, its mechanistic contribution to POI remains largely unexplored in humans. Methods We performed an integrative multiomics analysis combining single-nucleus RNA sequencing (snRNA-seq), bulk RNA-seq, genome-wide association studies (GWAS), and expression quantitative trait locus (eQTL) data. Co-expression networks were constructed via weighted gene coexpression network analysis (WGCNA), while cell-cell communication and trajectory analyses were conducted via CellChat and Monocle. Regulatory networks were inferred via SCENIC, and causality was assessed via summary-data-based Mendelian randomization (SMR). Candidate hub genes were prioritized through machine learning and validated via in vitro assays assessing rhythmicity and gene expression. Results snRNA-seq identified a granulosa cell subpopulation (GC1) with the highest circadian rhythm score, suggesting a pivotal role in regulating the ovarian clock. WGCNA and SCENIC analyses revealed age-associated downregulation of the core circadian regulators CLOCK and ARNTL, accompanied by disruptions in lipid metabolism and stress response pathways. SMR analysis revealed 120 circadian-related genes associated with POI risk, 30 of which were enriched in GC1-specific modules. CLOCK, CRY1, APOE, and GSTA1 emerged as key regulators on the basis of machine learning prioritization. Functional assays confirmed impaired rhythmicity and altered gene expression in KGN cells and senescent mouse granulosa cells. CLOCK knockdown increased P16 and P21 expression, underscoring its role in preserving granulosa cell homeostasis. Conclusions Our findings implicate circadian rhythm disruption as a hallmark and potential driver of ovarian aging. CLOCK, BMAL1, CRY1, APOE, and GSTA1 may serve as early biomarkers and therapeutic targets for POI. Biological sciences/Computational biology and bioinformatics Biological sciences/Systems biology Premature ovarian insufficiency Circadian rhythm Granulosa cells Multiomics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Premature ovarian insufficiency (POI), characterized by ovarian dysfunction before the age of 40, affects approximately 1–3% of reproductive-aged women globally, with an increasing incidence in younger populations[ 1 ]. Clinically, POI presents with elevated gonadotropin levels, diminished estrogen fluctuations, and menstrual irregularities or amenorrhea, often causing infertility, hormonal imbalances, and early menopause. The pathophysiology of POI remains incompletely understood, although chromosomal anomalies, autoimmune disorders, and environmental factors have been identified as contributing factors[ 2 ]. Current management largely relies on hormone replacement therapy (HRT) to alleviate symptoms, yet this approach fails to address the underlying ovarian dysfunction. The absence of reliable early biomarkers and effective long-term therapies underscores the need for deeper exploration into POI mechanisms[ 3 ]. Emerging studies suggest that circadian rhythms may significantly influence ovarian function and reproductive health. Circadian rhythms orchestrate physiological processes such as sleep‒wake cycles, hormone secretion, and metabolic regulation[ 4 ]. Clinical evidence indicates that circadian disruptions, due to nocturnal light exposure, irregular sleep patterns, or shift work, accelerate ovarian aging, thereby contributing to POI. Circadian genes have been shown to regulate hormonal balance, follicle development, and oocyte maturation, which are crucial for reproductive longevity[ 4 , 5 ]. Animal studies provide further support, demonstrating that mutations in Per1/Per2 result in premature ovarian reserve depletion[ 6 ], whereas Bmal1 deficiency induces accelerated reproductive aging and reduced fertility[ 7 , 8 ]. These findings may suggest that circadian rhythm genes play a significant role in POI detection and treatment. However, existing evidence originates primarily from animal models, providing limited insights into human pathogenesis. To bridge this gap, we conducted an integrated multiomics analysis combining single-nuclear RNA sequencing (snRNA-seq), bulk RNA sequencing (RNA-seq), genome-wide association studies (GWAS), and expression quantitative trait loci (eQTL) analyses, leveraging machine learning algorithms to identify hub circadian rhythm genes in the ovarian tissues of women with POI. Our findings establish a novel foundation for biomarker discovery, offering critical insights for early POI diagnosis and future mechanistic research. Materials and methods 1.1 Study design and data sources In this study, we accessed the Gene Expression Omnibus (GEO) database and analysed the GSE202601 snRNA-seq dataset, which includes ovarian samples from four reproductively healthy women and four patients with diminished reproductive function. Additionally, we compiled all publicly available bulk RNA sequencing data of ovarian granulosa cells (GCs) in POI, including data from GSE201276, GSE243720, and GSE232306, which together encompass ovarian GCs from 12 POI patients and 11 control subjects. We curated circadian rhythm-related genes from the GeneCards, MSigDB, and gene set variation analysis (GSVA) databases. After deduplication and filtering, a total of 2,556 circadian rhythm-related genes were included in the analysis. Summary-level data from GWAS and eQTL studies were used. Case control GWAS data for primary ovarian failure were derived from the FinnGen project R12, adhering to the International Classification of Diseases, 10th Revisio. The whole blood and ovary tissue eQTL database was derived from the GTEx database (Table 1 ). Table 1 Basic data information Dataset Database Platform Sample Species GSE202601 GEO GPL21697 4 cases of young women ovaries and 4 reproductively aged women ovaries Homo spaines GSE201276 GEO GPL20795 granulosa cells (GCs) in 6 POI patients and 5 control patients Homo spaines GSE243720 GEO GPL23227 GCs in 3 POI patients and 3 control patients Homo spaines GSE232306 GEO GPL23227 3 cases of GCs by cyclophosphamide and normal Homo spaines Circadian rhythm-related gene Genecard Genecard - Homo spaines Primary ovarian failure Finnen Finnen R12 655 cases of POF and 267780 cases of NOR Homo spaines eQTL eQTLGen consortium expression QTL (eQTL) data 31 684 individuals Homo spaines eQTL GTEx database expression QTL (eQTL) data 984 individuals Homo spaines 1.2 snRNA-seq dataset analysis In this snRNA-seq dataset, we processed the raw data into Seurat objects via the Seurat R 4.5.1 package. To ensure data quality, we filtered cells on the basis of the following criteria: detection of gene numbers ranging from 200 to 7000, mitochondrial gene expression levels less than 20%, and red blood cell gene expression levels less than 5%[ 9 ]. To reduce data dimensionality, we conducted principal component analysis (PCA) on genes with relatively high expression levels and selected the top 20 principal components. We subsequently utilized the “FindCluster” function for clustering and set the resolution parameter to 0.8[ 10 ]. We used UMAP to visualize the results. We used a manual typical marker gene annotation method to annotate individual cells to ensure the accuracy of the results. We also reannotated the cellular subpopulations of ovarian granulosa cells with reference to previous literature descriptions so that the granulosa cells could be classified into more detailed categories. 1.3 Differentially expressed genes (DEGs) and pathway enrichment analysis To identify DEGs for different cell clusters, we employed Seurat's “FindAllMarkers” function to compare cells within specific clusters with all other cells. To explore the potential molecular mechanisms of POI, we performed Gene Ontology (GO) and Kyoto Encyclopedia of Genes Genomes (KEGG) analyses on the differentially expressed genes. Gene set enrichment analysis (GSEA) was also applied to explore significant functional terms between both groups in each subtype[ 11 ]. 1.4 Construction of weighted gene coexpression networks to identify circadian rhythm-related genes Weighted gene coexpression network analysis (WGCNA) was conducted to identify modules associated with circadian rhythm[ 12 ]. After data preprocessing, outlier removal, and selection of an optimal soft-threshold power (scale-free topology, R²>0.85), genes were clustered via hierarchical clustering on the basis of topological overlap matrix (TOM) dissimilarity. Module eigengenes (MEs) were calculated, and the two modules most correlated with circadian traits were chosen. Hub genes exhibiting high module membership (MM) and gene significance were identified as potential biomarkers or therapeutic targets. 1.5 Cell communication and pseudotime analysis CellChat (v2.0) was used to evaluate cell‒cell communication by analysing ligand‒receptor interactions via the default CellChatDB[ 13 , 14 ]. Ligand‒receptor pairs that are overexpressed in specific cell subsets were identified, allowing inference of cell type-specific communication patterns and detection of enhanced signalling pathways. For pseudotime trajectory analysis, Monocle was utilized to reconstruct cellular developmental trajectories. The cells were classified into distinct developmental states on the basis of single-cell expression profiles, and trajectories were mapped according to gene expression dynamics[ 15 ]. 1.6 Single-cell regulatory network inference and clustering (SCENIC) To identify master transcriptional regulators, single-cell regulatory network inference and clustering (SCENIC) was performed via the pySCENIC workflow. Coexpression patterns were analysed to predict transcription factors (TFs) and their target genes. CisTarget was used to identify enriched TF-binding motifs and direct regulatory targets, while AUCell was used to quantify regulon activities across single cells. 1.7 Screening Hub Genes via Machine Learning Candidate genes identified via snRNA-seq were validated via bulk RNA-seq data[ 16 ]. Four machine learning algorithms—random forest (RF), logistic regression (LR), multilayer perceptron (MLP), and k-nearest neighbors (KNN)—were employed to screen for hub genes. Genes consistently identified by all four algorithms were defined as optimal features. RF, configured with 500 trees and 10-fold cross-validation, effectively captures feature interactions through ensemble decision-tree modelling[ 16 ]. LR provided interpretable coefficients for feature importance in binary classifications. KNN, which is optimal for small-sample datasets with distinct local structures, predicts outcomes on the basis of proximity in feature space, optimizing 'k' via 10-fold cross-validation[ 17 ]. An MLP, an artificial neural network, models complex nonlinear patterns through multiple hidden layers, effectively managing high-dimensional data. 1.8 Summary-data-based Mendelian randomization (SMR) SMR analysis was performed to evaluate the causal relationships between circadian rhythm-related gene expression and POI[ 18 ]. Single nucleotide polymorphisms (SNPs) are instrumental variables, gene expression levels are exposures, and POI is the outcome. Candidate signals were defined on the basis of the following criteria: (1) SMR false discovery rate (FDR) < 0.05, (2) genome-wide suggestive significance (P 0.05. 1.9 Isolation of primary granulosa cells Three-week-old female C57BL/6J mice (Qinglongshan Technology Co., Ltd., Jiangsu, China) received intraperitoneal injections of pregnant mare serum gonadotropin (PMSG, 10 IU). After 48 hours, the mice were euthanized via cervical dislocation. The ovaries were punctured to release the follicular contents, followed by enzymatic digestion with hyaluronidase (1 mg/mL, 30 s). Primary granulosa cells were isolated by filtration through a 200-mesh cell strainer. 1.10 Cell culture and model establishment Primary granulosa cells and KGN cells were cultured in DMEM/F-12 medium (Gibco, USA) supplemented with 10% fetal bovine serum (FBS; Gibco, USA) and 1% penicillin‒streptomycin. After 72 h, primary granulosa cells were treated with 20 µmol/L 4-hydroxycyclophosphamide (4-HC) for 12 h to induce premature cellular senescence. Protein and gene expression levels were measured via Western blot (WB) and quantitative PCR (qPCR), respectively. 1.11 Cell Rhythm Synchronization Training KGN cells were inoculated in 24-well plates at a density of 1*10 4 /well, and when the cell density reached 70–80%, the cells were treated with 100 µm dexamethasone buffer (Sigma, Merck) and then replaced with new culture medium after 2 h, which was considered ZT0. Then, the cellular proteins and RNAs were extracted every 4 h, and the mRNAs and proteins of the genes were plotted at different time points. The mRNA and protein expression of each gene at different time points were plotted, and the results are shown as a cosine function. 1.12 Western blot and quantitative PCR Protein expression was assessed by WB. Proteins extracted via RIPA buffer were separated via SDS‒polyacrylamide gel electrophoresis (SDS‒PAGE, ACE Tris‒acetate Precast gels), transferred onto polyvinylidene fluoride (PVDF) membranes (Millipore, Germany), and blocked with 5% nonfat milk for 2 h. Membranes were incubated overnight at 4°C with primary antibodies and subsequently incubated with secondary antibodies at 37°C for 1 h. Bands were visualized via hypersensitive–enhanced chemiluminescence (ECL, Beyotime, China), with GAPDH serving as a loading control. Total RNA from cells was extracted via a total RNA extraction kit (Vazyme, China) according to the manufacturer’s instructions and then reverse transcribed into complementary DNA (cDNA). Quantitative PCR was performed via ChamQ SYBR qPCR Master Mix (Vazyme, China), and relative gene expression was quantified via the 2^ –ΔΔCt method and normalized to that of GAPDH. The primer sequences are listed in Supplementary Meterials. 1.13 Statistical analysis Statistical analysis was conducted using Graphpad Prism 9.5.0. The relationship between variates was evaluated through Pearson correlation analysis. Group comparisons were performed via student test. R software 3.5.1 were used to visualize the results. Statistical significance was defined as p < 0.05. Results 2.1 Identification of DEGs via snRNA-seq Following stringent filtering and normalization of the snRNA-seq dataset, 42,475 high-quality cells were retained for downstream analysis. Principal component analysis (PCA) confirmed a relatively uniform distribution across groups, whereas UMAP classified cells into 24 distinct clusters. Manual annotation based on canonical marker genes identified ten major cell types: theca and stroma (T&S), smooth muscle, endothelial, epithelial, and granulosa (GC) cells, monocytes, B cells, T cells, T lymphocytes, and oocytes (Fig. 1 . A). The GCs were further subclustered via marker gene expression, yielding two distinct subtypes: GC1 and GC2 (Fig. 1 . B). Differentially expressed genes (DEGs) across cell types were identified via “FindAllMarkers” with selection criteria of adjusted p value 0.5, resulting in 3,219 DEGs across 11 cell types (Fig. 1 . C). Gene Ontology (GO) analysis revealed enrichment in functions such as protein folding, heat shock protein binding, ribosomal structure, and cytoplasmic translation. KEGG pathway analysis revealed significant alterations in circadian rhythm, circadian entrainment, and cellular senescence (Fig. 1 D). Gene set enrichment analysis (GSEA) of the granulosa cell gene sets further highlighted biological processes related to ribosomal function, mitochondrial activity, and lipid metabolism (Fig. 1 . E). 2.2 Screening Circadian Rhythm-Associated Genes Single-sample gene set enrichment analysis (ssGSEA) was used to calculate the circadian rhythm correlation score (CRCS) for each cell. The average CRCS values across cell types revealed marked heterogeneity in circadian gene expression following ovarian aging (Fig. 2 A). The cells were subsequently stratified into high-CRCS and low-CRCS groups on the basis of circadian gene enrichment scores. As shown in Fig. 2 B, high-CRCS cells were predominantly localized in the GC1 subpopulation, whereas low-CRCS scores were enriched in smooth muscle and epithelial cells. These findings indicate that GC1 cells exhibit the strongest association with circadian rhythm activity. Given this observation, we performed weighted gene coexpression network analysis (WGCNA) on the GC1 subset to identify rhythm-associated gene modules. The soft-thresholding power was set to 6 (Fig. 2 C), and two modules (GC1-M1 and GC1-M2) were identified. Both modules had module eigengene-based connectivity (KME) values greater than 0.6, indicating strong internal coherence. The top-ranked genes from each module were selected for further analysis. 2.3 Analysis of cell crosstalk and proposed time series analysis of core GC genes To explore the crosstalk between GCs and other cells in the ovarian microenvironment, we used CellChatDB to assimilate information on cellular interactions on the basis of a database of ligand‒receptor signals. Compared with those in the young group, GCs in the aging group exhibited a significantly reduced capacity for intercellular communication. The downregulation of signalling pathways such as the VEGF, PDGF, and FGF pathways suggests impaired proliferative and reparative functions of GCs. Suppressing the expression level of VISFATIN, a protein secreted by adipose tissue, may disrupt lipid metabolism and steroid hormone synthesis and disrupt the ovarian microenvironment. In contrast, the elevated expression of signalling molecules such as CXCL, PGN, WNT, and GPF may promote proinflammatory signalling (Fig. 3 A-C). These molecular alterations may serve as key biomarkers for the early diagnosis of POI and represent potential targets for therapeutic intervention. Monocle analysis further revealed changes in the developmental trajectory of ovarian GCs with aging, revealing that the developmental trajectory of cells in the older group significantly deviated from that of cells in the younger group, which may be related to the decrease in the differentiation potential of the cells or the intensification of cellular aging. We clustered the differentially expressed genes by monocle according to the expression trend on the proposed temporal trajectory to discover different dynamic regulation patterns. The expression trends of key genes on the proposed chronological trajectories were plotted (Fig. 3 D-F). 2.4 SCENIC analysis reveals basal motifs in granulosa cells We used SCENIC to identify a total of 280 transcription factors and motifs in different cellular subpopulations and screened 18 transcription factors whose transcriptional activities differed significantly between young and senescent GCs, of which 8 were significantly upregulated and 10 were significantly downregulated (Fig. 4 A). The rhythm-related transcription factors CLOCK and ARNTL presented significantly greater transcriptional activity in the young GC group than in the aging GC group, and their gene expression also decreased with age. The downregulation of circadian rhythm-related gene activity may indicate that granulosa cell desynchronization occurs with aging. In contrast, the transcriptional activities of the JUN and CEPBP matrices were progressively upregulated, and their expression was significantly upregulated with age (Fig. 4 B-D). The major signalling pathways involved in these transcription factors, including the WNT pathway and the NF-κB pathway, highlight the involvement of proinflammatory factors, oxidative stress and other unfavourable factors in the aging process of GCs. The downregulation of circadian rhythm-related gene activity may indicate that granulosa cell desynchronization occurs with aging. 2.5 Summary-data-based Mendelian methods We performed summary-data-based Mendelian randomization (SMR) analyses by integrating eQTL data from both whole blood and ovarian tissue with genome-wide association study (GWAS) data for POI and identified a total of 120 circadian rhythm-related genes significantly associated with POI risk, all of which presented P(HEIDI) values greater than 0.05 (Fig. 5 A). To evaluate whether these SMR-identified genes align with transcriptomic coexpression patterns, we compared them with hub genes derived from WGCNA modules. Notably, 30 of these genes overlapped with the GC1-specific circadian rhythm module (Fig. 5 B). This overlap suggests that a substantial subset of rhythm-related genes under genetic regulation also demonstrate strong coexpression within aging-associated transcriptomic modules. Functional enrichment analysis of the 30 overlapping genes revealed significant associations with lipid metabolism, DNA repair, and circadian rhythm synchronization, further supporting their potential mechanistic involvement in the process of ovarian aging (Fig. 5 C). 2.6 Validation at the bulk RNA level and results of machine learning algorithms We screened a total of three sets of human GC RNA sequencing data for POI and NOR from the GEO database, all of which were high-throughput sequencing data. To ensure sample quality, we merged and cleaned the data. Batch effects between different datasets were removed via the R package ‘sva’, and datasets were normalized via the R package ‘preprocessCore’. PCA confirmed that the batch effect of the three datasets was low and that the data distribution was stable (Fig. 5 D). We verified the expression levels of 30 overlapping genes at the Bulk-RNA level, and 10 genes, such as CLOCK, ARNTL, and APOE, were identified as differentially expressed (Fig. 5 E). We independently assessed gene importance via four ML models: GLM, RF, KNN, and NNET. The residual boxplot results indicate that the KNN model has the best data fit, with the smallest residuals. In contrast, the GLM has higher median residuals and RMSE, and its interquartile range (IQR) is notably wider, suggesting that it has the largest residual errors and slightly poorer fit than the other three models. Additionally, the area under the ROC curve (AUC) for all four models was greater than 0.8, indicating that the predictive performance of all the models was satisfactory (Fig. 5 F). Following Z score normalization of the gene importance scores across the four models, a soft voting ensemble learning method was employed for gene selection. As a result, the core genes were ultimately determined: CLOCK, CRY1, APOE and GSTA1. 2.7 WB and qPCR confirmed the expression of circadian genes in granulosa cells We used WB and qPCR to confirm the differences in the expression of overlapping genes in primary GCs, as shown in Fig. 6 A-B. Compared with that in the control group, the expression of GSTA1 was upregulated in the model group, and the expression of APOE was downregulated. The expression of the transcription factors BMAI1, CLOCK, and CRY1 was downregulated. Under physiological conditions, the expression levels of the CLOCK, CRY1, APOE and GSTA1 genes exhibit marked circadian rhythmicity over a 24-hour cycle, with peak expression of CLOCK and APOE occurring at Zeitgeber Time 8 (ZT8), CRY1 peaking at ZT12, and GSTA1 peaking at ZT20. In contrast, the model group presented attenuated rhythm amplitude and a phase delay, indicating a disruption in circadian rhythm synchronization. 2.8 CLOCK −/ GCs show an aging phototype We employed siRNA to knock down the CLOCK gene in primary mouse cells and subsequently evaluated the cellular senescence phenotype. Western blot analysis revealed that, compared with negative control cells, CLOCK-deficient cells presented significant upregulation of P16 and P21 protein expression, indicating that CLOCK gene knockdown may contribute to the induction of GC senescence. Discussion Our integrated multiomics analysis provides compelling evidence that circadian clock disruption in GCs plays a central role in the pathogenesis of POI. GCs—key somatic supporters of oocyte development—undergo profound circadian dysregulation with ovarian aging, contributing to their functional decline. This elevates circadian misregulation from a peripheral observation to a mechanistic hallmark of premature ovarian aging. Using snRNA-seq, we first identified a GC subtype, GC1, characterized by a high enrichment of circadian gene expression, suggesting its role as an intrinsic follicular "timekeeper". Through WGCNA, we identified circadian-related gene modules within the GC1 subtype that converge on critical pathways, including lipid metabolism, protein folding, and DNA repair—processes essential for the proper development of follicles. The consequences of circadian disruption were evident in aged GCs, which presented significantly diminished intercellular communication. Downregulated signalling pathways such as the VEGF, PDGF, and FGF pathways suggest impaired vascular support, proliferation, and survival. Moreover, increased expression of proinflammatory and stress-related signalling molecules, including CXCL, PGN, and WNT, implies a shift toward an inflammatory microenvironment[ 19 ]. Circadian regulators, such as CLOCK/BMAL1, orchestrate various physiological rhythms[ 20 , 21 ]; diminished activity of these factors may disrupt follicular nutrient metabolism and activate inflammatory signalling pathways. SCENIC analysis further demonstrated that the transcriptional activity of the core circadian genes CLOCK/BMAL1 declines, whereas the stress response factors JUN and CEBPB exhibit increased activity. This reciprocal shift—from circadian regulation to stress-dominated transcription—reflects a transition in the regulatory landscape of aging GCs, reinforcing the role of circadian erosion in driving ovarian dysfunction. A key question is whether circadian rhythm alterations in POI are merely correlative or truly causative. Our SMR analysis revealed a causal relationship and identified 120 circadian-related genes whose expression is influenced by genetic variants associated with POI. Notably, 30 of these genes overlap with circadian transcriptional modules in GC1, integrating both genetic and transcriptomic evidence. These genes are enriched in pathways related to mitochondrial function, metabolism, and DNA repair—critical processes that govern oocyte quality and follicle longevity—thereby establishing a direct mechanistic link between circadian disruption and POI pathogenesis. Compared with those of the controls, the protein and mRNA levels of CLOCK, BMAL1, CRY1, and APOE, alongside markedly elevated GSTA1, were significantly lower in the model group. Circadian rhythm assays further revealed disrupted intrinsic rhythmicity of CLOCK, CRY1, APOE and GSTA1 in the model group, characterized by delayed phases and decreased amplitudes. Following CLOCK knockdown in primary mouse granulosa cells, the P16 and P21 proteins were notably upregulated, leading to cell cycle arrest and progression to irreversible senescence. The concordance between multiomics predictions and in vitro results not only strengthens confidence in our findings but also highlights a broad circadian network of coregulated genes. Together, our data suggest that GCs harbor local circadian clocks critical for maintaining ovarian function. In addition to governing hormone secretion and ovulation timing, these intrinsic clocks orchestrate cell cycle progression and intercellular communication. Disruption of this timing system induces desynchronization within the follicle, contributing to accelerated ovarian aging. GSTA1, a glutathione S-transferase involved in antioxidant defense and apoptosis regulation, detoxifies lipid peroxides and harmful metabolites to alleviate oxidative stress[ 22 ]. Its hepatic expression has been shown to follow circadian regulation by CLOCK/BMAL1, potentially determining peak detoxification timing[ 23 ]. In this study, GSTA1 was significantly upregulated in the CTX-induced granulosa cell injury model, suggesting a compensatory antioxidant response. These findings imply that circadian genes may modulate oxidative stress in granulosa cells via the regulation of GSTA1[ 24 ]. APOE, expressed in ovarian granulosa cells, facilitates cholesterol uptake and transport, contributing to steroid hormone synthesis[ 25 ]. Several clinical studies have reported that women carrying the APOEε4 allele tend to exhibit increased reproductive potential[ 26 ]. Here, APOE expression was decreased in aging GCs, accompanied by dampened circadian rhythmicity, suggesting that APOE may be transcriptionally regulated by the CLOCK/BMAL1 axis and influence hormone biosynthesis through circadian mechanisms. Clinically, our findings facilitate early diagnosis and intervention. Circadian genes, including CLOCK, BMAL1, CRY1, GSTA1, and APOE, have potential as early biomarkers for a diminished ovarian reserve; monitoring their expression in granulosa cells could identify women at risk of POI, especially those exposed to environmental or iatrogenic factors. Lifestyle modifications aligned with circadian rhythms or chronotherapy might offer noninvasive strategies to protect ovarian function. Additionally, pharmacological modulation of circadian regulators represents a promising therapeutic avenue, pending further validation. However, several limitations should be noted. The modest sample size of our single-cell dataset limits the exploration of population-level heterogeneity. Although the incidence of POI is increasing, clinical sample acquisition remains challenging, constraining the reliability and generalizability of our findings. Furthermore, the specificity of the SMR results is restricted by the available GWAS data. Finally, since in vitro models cannot fully replicate ovarian physiology, further in vivo validation is essential. Conclusions Through multiomics integration, we demonstrated that disruption of circadian rhythms in the ovary contributes to the accelerated development of POI and identified CLOCK, BMAL1, CRY1, GSTA1, and APOE as potential POI-associated risk genes. These findings provide a foundation for innovative diagnostic and therapeutic approaches to safeguard reproductive longevity and improve women’s health. Abbreviations Premature ovarian insufficiency POI single-nuclear RNA sequencing snRNA-seq bulk RNA sequencing RNA-seq genome-wide association studies GWAS expression quantitative trait loci eQTL summary-data-based Mendelian randomization SMR granulosa cells GCs Declarations Data availability: FinnGen (https://www.finngen.finn/en); GEO (https://www.ncbi.nlm.nih.gov/geo/); CGDB (http://cgdb.biocuckoo.org)/ Author Contributions: Conceptualization, RJ; methodology, RJ and CJ; data curation, CJ; writing-original draft preparation, RJ and TY; writing-review & editing, TY. All authors reviewed the manuscript. Funding statement: This study was supported by the National Natural Science Foundation of China (No. 82330125). The funding source did not influence the study design, analyses, or interpretation of the data. Disclosure Statement: None. Data sharing statement: All the data in this paper are publicly available. References Genes linked to premature ovarian insufficiency show no pathogenicity in the general population. Nat Med 2023;29:1617–8. Bompoula MS, Valsamakis G, Neofytou S, et al. Demographic, clinical and hormonal characteristics of patients with premature ovarian insufficiency and those of early menopause: data from two tertiary premature ovarian insufficiency centers in Greece. Gynecological Endocrinology 2020;36:693–7. The Lancet. Time for a balanced conversation about menopause. The Lancet 2024;403:877. https://doi.org/10.1016/S0140-6736(24)00462-8. Li Y, Pei T, Zhu H, et al. Melatonin Alleviates Circadian Rhythm Disruption-Induced Enhanced Luteinizing Hormone Pulse Frequency and Ovarian Dysfunction. J Pineal Res 2025;77. Lane JM, Qian J, Mignot E, et al. Genetics of circadian rhythms and sleep in human health and disease. Nat Rev Genet 2023;24:4–20. Zheng Y, Liu C, Li Y, et al. Loss-of-function mutations with circadian rhythm regulator Per1/Per2 lead to premature ovarian insufficiency. Biol Reprod 2019;100:1066–72. Zhang A, Li S, Huang L, et al. Bmal1 regulates female reproduction in mice via the hypothalamic–pituitary–ovarian axis. FASEB Journal 2024;38. Kawamura T, Dai Y, Ono M, et al. BMAL1 positively correlates with genes regulating steroidogenesis in human luteinized granulosa cells. Reproduction 2024;167. Wu J, Lv Y, Hao P, et al. Immunological profile of lactylation-related genes in Crohn’s disease: a comprehensive analysis based on bulk and single-cell RNA sequencing data. J Transl Med 2024;22. Du C, Wang C, Liu Z, et al. Machine learning algorithms integrate bulk and single-cell RNA data to unveil oxidative stress following intracerebral hemorrhage. Int Immunopharmacol 2024;137. Le DC, Ngo MHT, Kuo YC, et al. Secretome from estrogen-responding human placenta-derived mesenchymal stem cells rescues ovarian function and circadian rhythm in mice with cyclophosphamide-induced primary ovarian insufficiency. J Biomed Sci 2024;31:95. An Q, Zheng N, Ji Y, et al. Exploration the interaction of cadmium and copper toxic effects in pakchoi (Brassica chinensis L) roots through combinatorial transcriptomic and weighted gene coexpression network analysis. J Environ Manage 2024;359. https://doi.org/10.1016/J.JENVMAN.2024.120956. Su J, Song Y, Zhu Z, et al. Cell–cell communication: new insights and clinical implications. Signal Transduct Target Ther 2024;9. https://doi.org/10.1038/S41392-024-01888-Z. Clyde D. Single cell–cell communication. Nat Rev Genet 2023;24:488. Hymel LA, Anderson SE, Turner TC, et al. Identifying dysregulated immune cell subsets following volumetric muscle loss with pseudotime trajectories. Commun Biol 2023;6. Asnicar F, Thomas AM, Passerini A, et al. Machine learning for microbiologists. Nat Rev Microbiol 2024;22:191–205. https://doi.org/10.1038/S41579-023-00984-1. Samet H. K-nearest neighbor finding using MaxNearestDist. IEEE Trans Pattern Anal Mach Intell 2008;30:243–52. https://doi.org/10.1109/TPAMI.2007.1182. Lin P-W, Lin Z-R, Wang W‒W, et al. Identification of immune-inflammation targets for intracranial aneurysms: A multiomics and epigenome-wide study integrating summary-data-based mendelian randomization, single-cell-type expression analysis, and DNA methylation regulation. International Journal of Surgery 2024;111. Zhou C, Gao Y, Ding P, et al. The role of CXCL family members in different diseases. Cell Death Discov 2023;9:212. Kulkarni SS, Singh O, Zigman JM. The intersection between ghrelin, metabolism and circadian rhythms. Nat Rev Endocrinol 2024;20:228–38. McAlpine CS, Swirski FK. Circadian influence on metabolism and inflammation in atherosclerosis. Circ Res 2016;119:131–41. Ma S, Xie F, Wen X, et al. GSTA1/CTNNB1 axis facilitates sorafenib resistance by suppressing ferroptosis in hepatocellular carcinoma. Pharmacol Res 2024;210. https://doi.org/10.1016/J.PHRS.2024.107490. Xu YQ, Zhang D, Jin T, et al. Diurnal Variation of Hepatic Antioxidant Gene Expression in Mice. PLoS One 2012;7. Dai C, Sharma G, Liu G, et al. Therapeutic detoxification of quercetin for aflatoxin B1-related toxicity: Roles of oxidative stress, inflammation, and metabolic enzymes. Environmental Pollution 2024;345. https://doi.org/10.1016/J.ENVPOL.2024.123474. Oriá RB, de Almeida JZ, Moreira CN, et al. Apolipoprotein E Effects on Mammalian Ovarian Steroidogenesis and Human Fertility. Trends in Endocrinology and Metabolism 2020;31:872–83. Lu H, Jing Y, Zhang C, et al. Aging hallmarks of the primate ovary revealed by spatiotemporal transcriptomics. Protein Cell 2024;15:364–84. Additional Declarations There is NO Competing Interest. Supplementary Files 00Graphicabstract.jpg SupplementaryMaterials1.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6850329","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":473223831,"identity":"463c1ebb-f3ff-42c9-8b7e-087818cf26c4","order_by":0,"name":"yong tan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABAElEQVRIiWNgGAWjYBACfmbmA4YfDCR47NsbGAwgYgn4tUi2tyUUS1TYyBjwHCBSi8GZMwYfeM6k2RhIwFUS0MJwI8Fwg2TbYR5zybcHinlqtjHws+cYMPzcgVsH44yEZINCoBbL2XkJxjzHbjNI9rwxYOw9g1sLs0TCMQOQLQy3cwyMeRtuMxjcyDFgZmzDrYVNIrH9By9Iy80zEC32hLTw8BxmMAB6n8fgBg/UFgkCWiTY2xiMgYHMI9mTY2A459htHokzzwoO9uLRYn+Y/wMoKu352c+YGbypuS3H35688cFPPFpQ/AWKSh4Q6wBxGoCh94BYlaNgFIyCUTCyAABseVAGehev+wAAAABJRU5ErkJggg==","orcid":"","institution":"Department of Reproductive Medicine, Affiliated Hospital of Nanjing University of Chinese Medicine, Jiangsu Province Hospital of Chinese Medicine","correspondingAuthor":true,"prefix":"","firstName":"yong","middleName":"","lastName":"tan","suffix":""},{"id":473223832,"identity":"cea47681-81e9-4379-8676-bd298dae3a3c","order_by":1,"name":"Jin Ren","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Jin","middleName":"","lastName":"Ren","suffix":""},{"id":473223833,"identity":"ee1c5251-af4f-4c90-9836-dd0236b0e3c0","order_by":2,"name":"Jie chen","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Jie","middleName":"","lastName":"chen","suffix":""}],"badges":[],"createdAt":"2025-06-09 03:40:29","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6850329/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6850329/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":85051855,"identity":"a98193bb-9b91-42a8-aa5b-e2139aadba2a","added_by":"auto","created_at":"2025-06-20 11:51:54","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2689099,"visible":true,"origin":"","legend":"\u003cp\u003eA. UMAP plot showing SnRNA-seq cell clustering;\u003c/p\u003e\n\u003cp\u003eB. Clustering of granulosa cells ;\u003c/p\u003e\n\u003cp\u003eC. Differentially expressed genes identified at the single-cell level;\u003c/p\u003e\n\u003cp\u003eD. GO and KEGG enrichment analysis of differentially expressed genes;\u003c/p\u003e\n\u003cp\u003eE. GSEA analysis of SnRNA-seq cell subpopulations.\u003c/p\u003e","description":"","filename":"Fig100.png","url":"https://assets-eu.researchsquare.com/files/rs-6850329/v1/b6081ae5073654a0bd6f40cd.png"},{"id":85051860,"identity":"1ec36d42-5079-4616-93f3-70883baa7957","added_by":"auto","created_at":"2025-06-20 11:51:54","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1547361,"visible":true,"origin":"","legend":"\u003cp\u003eA. UMAP plot illustrating cell distribution profiles between high and low CRCS scoring groups;\u003cbr\u003e\nB. ssGSEA analysis indicates that GCs1 exhibits the highest average CRCS scores;\u003cbr\u003e\nC. Visualization of two gene modules in GCs1, ranked by KME values;\u003cbr\u003e\nD. Representative gene expression maps of the two modules;\u003cbr\u003e\nE. Bubble plot showing the correlation of the GCs1 module with all cell populations.\u003c/p\u003e","description":"","filename":"Fig200.png","url":"https://assets-eu.researchsquare.com/files/rs-6850329/v1/3957d635e13a953bc44239d0.png"},{"id":85051858,"identity":"b0e04bc3-0f06-4ee1-82e2-4adfc315bd99","added_by":"auto","created_at":"2025-06-20 11:51:54","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":2500562,"visible":true,"origin":"","legend":"\u003cp\u003eA. Cell-cell communication network and input/output signaling pathway patterns in the young group;\u003cbr\u003e\nB. Cell-cell communication network and input/output signaling pathway patterns in the senescent group;\u003cbr\u003e\nC. Expression profiles of signaling pathway-related genes across cellular subpopulations;\u003cbr\u003e\nD. Inferred pseudotime trajectory analysis of granulosa cells;\u003cbr\u003e\nE. Hierarchical clustering of differentially expressed genes identified by Monocle;\u003cbr\u003e\nF. Temporal expression dynamics of core genes in granulosa cells.\u003c/p\u003e","description":"","filename":"Fig300.png","url":"https://assets-eu.researchsquare.com/files/rs-6850329/v1/dbeb8dae0a08f575ac19cabd.png"},{"id":85051861,"identity":"3e338425-d47f-4c99-a9d2-5478f362b00c","added_by":"auto","created_at":"2025-06-20 11:51:54","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":2602112,"visible":true,"origin":"","legend":"\u003cp\u003eA. Heatmap displaying binomial test values of transcription factors;\u003cbr\u003e\nB. Volcano plot showing differentially expressed transcription factors between granulosa cell groups;\u003cbr\u003e\nC. Heatmap of transcription factor expression in granulosa cells.\u003c/p\u003e","description":"","filename":"Fig4001.png","url":"https://assets-eu.researchsquare.com/files/rs-6850329/v1/cc12486241a48eea737b7fc9.png"},{"id":85051864,"identity":"3c24f5fa-46e7-4510-bc78-347877eee82b","added_by":"auto","created_at":"2025-06-20 11:51:55","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1918171,"visible":true,"origin":"","legend":"\u003cp\u003eA: Manhattan plot of eQTL and GWAS;\u003c/p\u003e\n\u003cp\u003eB: Venn diagram of WGCNA and SMR genes;\u003c/p\u003e\n\u003cp\u003eC: GO pathway plot of SMR genes;\u003c/p\u003e\n\u003cp\u003eD: PCA plot of the three groups from bulk RNA-seq;\u003c/p\u003e\n\u003cp\u003eE: Expression differences of two gene groups in bulk RNA-seq;\u003c/p\u003e\n\u003cp\u003eF: Machine learning model validation plot, AUC curve, and gene importance plot.\u003c/p\u003e","description":"","filename":"Fig500.png","url":"https://assets-eu.researchsquare.com/files/rs-6850329/v1/6426f6305692deb295f58e24.png"},{"id":85051866,"identity":"ae83fddd-8883-450b-9726-0ed92e5b7a15","added_by":"auto","created_at":"2025-06-20 11:51:55","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1934386,"visible":true,"origin":"","legend":"\u003cp\u003eA. Protein expression levels of core genes in the two groups;\u003c/p\u003e\n\u003cp\u003eB. mRNA expression levels of core genes in the two groups;\u003c/p\u003e\n\u003cp\u003eC. Rhythmic expression patterns of individual genes modeled using a cosine function;\u003c/p\u003e\n\u003cp\u003eD. Changes in P16 and P21 protein expression following CLOCK\u003csup\u003e -/- \u003c/sup\u003ein GCs.\u0026nbsp;\u003c/p\u003e","description":"","filename":"Fig6.png","url":"https://assets-eu.researchsquare.com/files/rs-6850329/v1/2a144937bc933f6b03fb3ca7.png"},{"id":86354806,"identity":"4d81b642-ced3-4abb-9a4e-1dcbbc12fb68","added_by":"auto","created_at":"2025-07-09 16:49:13","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":12503525,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6850329/v1/f4b216e0-fb9b-448b-b382-5155320ef14f.pdf"},{"id":85052447,"identity":"c2f623f9-28af-40a7-ad15-1e5cf218fc48","added_by":"auto","created_at":"2025-06-20 11:59:55","extension":"jpg","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":185339,"visible":true,"origin":"","legend":"","description":"","filename":"00Graphicabstract.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6850329/v1/69d63bf6be0ddbdb498142dc.jpg"},{"id":85051862,"identity":"47eefda0-a82e-470e-b9d8-b8b62d2ec8a9","added_by":"auto","created_at":"2025-06-20 11:51:54","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":14779,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterials1.docx","url":"https://assets-eu.researchsquare.com/files/rs-6850329/v1/dab56e63a106256f99abb135.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Multiomics analyses reveal key circadian rhythm genes implicated in Premature ovarian insufficiency","fulltext":[{"header":"Introduction","content":"\u003cp\u003ePremature ovarian insufficiency (POI), characterized by ovarian dysfunction before the age of 40, affects approximately 1\u0026ndash;3% of reproductive-aged women globally, with an increasing incidence in younger populations[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Clinically, POI presents with elevated gonadotropin levels, diminished estrogen fluctuations, and menstrual irregularities or amenorrhea, often causing infertility, hormonal imbalances, and early menopause. The pathophysiology of POI remains incompletely understood, although chromosomal anomalies, autoimmune disorders, and environmental factors have been identified as contributing factors[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Current management largely relies on hormone replacement therapy (HRT) to alleviate symptoms, yet this approach fails to address the underlying ovarian dysfunction. The absence of reliable early biomarkers and effective long-term therapies underscores the need for deeper exploration into POI mechanisms[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eEmerging studies suggest that circadian rhythms may significantly influence ovarian function and reproductive health. Circadian rhythms orchestrate physiological processes such as sleep‒wake cycles, hormone secretion, and metabolic regulation[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Clinical evidence indicates that circadian disruptions, due to nocturnal light exposure, irregular sleep patterns, or shift work, accelerate ovarian aging, thereby contributing to POI. Circadian genes have been shown to regulate hormonal balance, follicle development, and oocyte maturation, which are crucial for reproductive longevity[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Animal studies provide further support, demonstrating that mutations in Per1/Per2 result in premature ovarian reserve depletion[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e], whereas Bmal1 deficiency induces accelerated reproductive aging and reduced fertility[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. These findings may suggest that circadian rhythm genes play a significant role in POI detection and treatment.\u003c/p\u003e \u003cp\u003eHowever, existing evidence originates primarily from animal models, providing limited insights into human pathogenesis. To bridge this gap, we conducted an integrated multiomics analysis combining single-nuclear RNA sequencing (snRNA-seq), bulk RNA sequencing (RNA-seq), genome-wide association studies (GWAS), and expression quantitative trait loci (eQTL) analyses, leveraging machine learning algorithms to identify hub circadian rhythm genes in the ovarian tissues of women with POI. Our findings establish a novel foundation for biomarker discovery, offering critical insights for early POI diagnosis and future mechanistic research.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cp\u003e1.1 Study design and data sources\u003c/p\u003e \u003cp\u003eIn this study, we accessed the Gene Expression Omnibus (GEO) database and analysed the GSE202601 snRNA-seq dataset, which includes ovarian samples from four reproductively healthy women and four patients with diminished reproductive function. Additionally, we compiled all publicly available bulk RNA sequencing data of ovarian granulosa cells (GCs) in POI, including data from GSE201276, GSE243720, and GSE232306, which together encompass ovarian GCs from 12 POI patients and 11 control subjects. We curated circadian rhythm-related genes from the GeneCards, MSigDB, and gene set variation analysis (GSVA) databases. After deduplication and filtering, a total of 2,556 circadian rhythm-related genes were included in the analysis. Summary-level data from GWAS and eQTL studies were used. Case control GWAS data for primary ovarian failure were derived from the FinnGen project R12, adhering to the International Classification of Diseases, 10th Revisio. The whole blood and ovary tissue eQTL database was derived from the GTEx database (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBasic data information\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDataset\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDatabase\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePlatform\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSample\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSpecies\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE202601\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGEO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGPL21697\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4 cases of young women ovaries and 4 reproductively aged women ovaries\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHomo spaines\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE201276\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGEO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGPL20795\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003egranulosa cells (GCs) in 6 POI patients and 5 control patients\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHomo spaines\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE243720\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGEO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGPL23227\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGCs in 3 POI patients and 3 control patients\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHomo spaines\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGSE232306\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGEO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGPL23227\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 cases of GCs by cyclophosphamide and normal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHomo spaines\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCircadian rhythm-related gene\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGenecard\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGenecard\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHomo spaines\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePrimary ovarian failure\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFinnen\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFinnen R12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e655 cases of POF and 267780 cases of NOR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHomo spaines\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eeQTL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eeQTLGen consortium\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eexpression QTL (eQTL) data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e31 684 individuals\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHomo spaines\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eeQTL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGTEx database\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eexpression QTL (eQTL) data\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e984 individuals\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHomo spaines\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e1.2 snRNA-seq dataset analysis\u003c/p\u003e \u003cp\u003eIn this snRNA-seq dataset, we processed the raw data into Seurat objects via the Seurat R 4.5.1 package. To ensure data quality, we filtered cells on the basis of the following criteria: detection of gene numbers ranging from 200 to 7000, mitochondrial gene expression levels less than 20%, and red blood cell gene expression levels less than 5%[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. To reduce data dimensionality, we conducted principal component analysis (PCA) on genes with relatively high expression levels and selected the top 20 principal components. We subsequently utilized the \u0026ldquo;FindCluster\u0026rdquo; function for clustering and set the resolution parameter to 0.8[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. We used UMAP to visualize the results. We used a manual typical marker gene annotation method to annotate individual cells to ensure the accuracy of the results. We also reannotated the cellular subpopulations of ovarian granulosa cells with reference to previous literature descriptions so that the granulosa cells could be classified into more detailed categories.\u003c/p\u003e \u003cp\u003e1.3 Differentially expressed genes (DEGs) and pathway enrichment analysis\u003c/p\u003e \u003cp\u003eTo identify DEGs for different cell clusters, we employed Seurat's \u0026ldquo;FindAllMarkers\u0026rdquo; function to compare cells within specific clusters with all other cells. To explore the potential molecular mechanisms of POI, we performed Gene Ontology (GO) and Kyoto Encyclopedia of Genes Genomes (KEGG) analyses on the differentially expressed genes. Gene set enrichment analysis (GSEA) was also applied to explore significant functional terms between both groups in each subtype[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e].\u003c/p\u003e \u003cp\u003e1.4 Construction of weighted gene coexpression networks to identify circadian rhythm-related genes\u003c/p\u003e \u003cp\u003eWeighted gene coexpression network analysis (WGCNA) was conducted to identify modules associated with circadian rhythm[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. After data preprocessing, outlier removal, and selection of an optimal soft-threshold power (scale-free topology, R\u0026sup2;\u0026gt;0.85), genes were clustered via hierarchical clustering on the basis of topological overlap matrix (TOM) dissimilarity. Module eigengenes (MEs) were calculated, and the two modules most correlated with circadian traits were chosen. Hub genes exhibiting high module membership (MM) and gene significance were identified as potential biomarkers or therapeutic targets.\u003c/p\u003e \u003cp\u003e1.5 Cell communication and pseudotime analysis\u003c/p\u003e \u003cp\u003eCellChat (v2.0) was used to evaluate cell‒cell communication by analysing ligand‒receptor interactions via the default CellChatDB[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Ligand‒receptor pairs that are overexpressed in specific cell subsets were identified, allowing inference of cell type-specific communication patterns and detection of enhanced signalling pathways. For pseudotime trajectory analysis, Monocle was utilized to reconstruct cellular developmental trajectories. The cells were classified into distinct developmental states on the basis of single-cell expression profiles, and trajectories were mapped according to gene expression dynamics[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e].\u003c/p\u003e \u003cp\u003e1.6 Single-cell regulatory network inference and clustering (SCENIC)\u003c/p\u003e \u003cp\u003eTo identify master transcriptional regulators, single-cell regulatory network inference and clustering (SCENIC) was performed via the pySCENIC workflow. Coexpression patterns were analysed to predict transcription factors (TFs) and their target genes. CisTarget was used to identify enriched TF-binding motifs and direct regulatory targets, while AUCell was used to quantify regulon activities across single cells.\u003c/p\u003e \u003cp\u003e1.7 Screening Hub Genes via Machine Learning\u003c/p\u003e \u003cp\u003eCandidate genes identified via snRNA-seq were validated via bulk RNA-seq data[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Four machine learning algorithms\u0026mdash;random forest (RF), logistic regression (LR), multilayer perceptron (MLP), and k-nearest neighbors (KNN)\u0026mdash;were employed to screen for hub genes. Genes consistently identified by all four algorithms were defined as optimal features. RF, configured with 500 trees and 10-fold cross-validation, effectively captures feature interactions through ensemble decision-tree modelling[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. LR provided interpretable coefficients for feature importance in binary classifications. KNN, which is optimal for small-sample datasets with distinct local structures, predicts outcomes on the basis of proximity in feature space, optimizing 'k' via 10-fold cross-validation[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. An MLP, an artificial neural network, models complex nonlinear patterns through multiple hidden layers, effectively managing high-dimensional data.\u003c/p\u003e \u003cp\u003e1.8 Summary-data-based Mendelian randomization (SMR)\u003c/p\u003e \u003cp\u003eSMR analysis was performed to evaluate the causal relationships between circadian rhythm-related gene expression and POI[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Single nucleotide polymorphisms (SNPs) are instrumental variables, gene expression levels are exposures, and POI is the outcome. Candidate signals were defined on the basis of the following criteria: (1) SMR false discovery rate (FDR)\u0026thinsp;\u0026lt;\u0026thinsp;0.05, (2) genome-wide suggestive significance (P\u0026thinsp;\u0026lt;\u0026thinsp;1 \u0026times; 10^\u003csup\u003e\u0026minus;5\u003c/sup\u003e) in eQTL and GWAS analyses, and (3) heterogeneity in dependent instrument (HEIDI) test results with P\u0026thinsp;\u0026gt;\u0026thinsp;0.05.\u003c/p\u003e \u003cp\u003e1.9 Isolation of primary granulosa cells\u003c/p\u003e \u003cp\u003eThree-week-old female C57BL/6J mice (Qinglongshan Technology Co., Ltd., Jiangsu, China) received intraperitoneal injections of pregnant mare serum gonadotropin (PMSG, 10 IU). After 48 hours, the mice were euthanized via cervical dislocation. The ovaries were punctured to release the follicular contents, followed by enzymatic digestion with hyaluronidase (1 mg/mL, 30 s). Primary granulosa cells were isolated by filtration through a 200-mesh cell strainer.\u003c/p\u003e \u003cp\u003e1.10 Cell culture and model establishment\u003c/p\u003e \u003cp\u003ePrimary granulosa cells and KGN cells were cultured in DMEM/F-12 medium (Gibco, USA) supplemented with 10% fetal bovine serum (FBS; Gibco, USA) and 1% penicillin‒streptomycin. After 72 h, primary granulosa cells were treated with 20 \u0026micro;mol/L 4-hydroxycyclophosphamide (4-HC) for 12 h to induce premature cellular senescence. Protein and gene expression levels were measured via Western blot (WB) and quantitative PCR (qPCR), respectively.\u003c/p\u003e \u003cp\u003e1.11 Cell Rhythm Synchronization Training\u003c/p\u003e \u003cp\u003eKGN cells were inoculated in 24-well plates at a density of 1*10\u003csup\u003e4\u003c/sup\u003e/well, and when the cell density reached 70\u0026ndash;80%, the cells were treated with 100 \u0026micro;m dexamethasone buffer (Sigma, Merck) and then replaced with new culture medium after 2 h, which was considered ZT0. Then, the cellular proteins and RNAs were extracted every 4 h, and the mRNAs and proteins of the genes were plotted at different time points. The mRNA and protein expression of each gene at different time points were plotted, and the results are shown as a cosine function.\u003c/p\u003e \u003cp\u003e1.12 Western blot and quantitative PCR\u003c/p\u003e \u003cp\u003eProtein expression was assessed by WB. Proteins extracted via RIPA buffer were separated via SDS‒polyacrylamide gel electrophoresis (SDS‒PAGE, ACE Tris‒acetate Precast gels), transferred onto polyvinylidene fluoride (PVDF) membranes (Millipore, Germany), and blocked with 5% nonfat milk for 2 h. Membranes were incubated overnight at 4\u0026deg;C with primary antibodies and subsequently incubated with secondary antibodies at 37\u0026deg;C for 1 h. Bands were visualized via hypersensitive\u0026ndash;enhanced chemiluminescence (ECL, Beyotime, China), with GAPDH serving as a loading control.\u003c/p\u003e \u003cp\u003eTotal RNA from cells was extracted via a total RNA extraction kit (Vazyme, China) according to the manufacturer\u0026rsquo;s instructions and then reverse transcribed into complementary DNA (cDNA). Quantitative PCR was performed via ChamQ SYBR qPCR Master Mix (Vazyme, China), and relative gene expression was quantified via the 2^\u003csup\u003e\u0026ndash;ΔΔCt\u003c/sup\u003e method and normalized to that of GAPDH. The primer sequences are listed in Supplementary Meterials.\u003c/p\u003e \u003cp\u003e1.13 Statistical analysis\u003c/p\u003e \u003cp\u003eStatistical analysis was conducted using Graphpad Prism 9.5.0. The relationship between variates was evaluated through Pearson correlation analysis. Group comparisons were performed via student test. R software 3.5.1 were used to visualize the results. Statistical significance was defined as p\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e2.1 Identification of DEGs via snRNA-seq\u003c/p\u003e \u003cp\u003eFollowing stringent filtering and normalization of the snRNA-seq dataset, 42,475 high-quality cells were retained for downstream analysis. Principal component analysis (PCA) confirmed a relatively uniform distribution across groups, whereas UMAP classified cells into 24 distinct clusters. Manual annotation based on canonical marker genes identified ten major cell types: theca and stroma (T\u0026amp;S), smooth muscle, endothelial, epithelial, and granulosa (GC) cells, monocytes, B cells, T cells, T lymphocytes, and oocytes (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. A). The GCs were further subclustered via marker gene expression, yielding two distinct subtypes: GC1 and GC2 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. B). Differentially expressed genes (DEGs) across cell types were identified via \u0026ldquo;FindAllMarkers\u0026rdquo; with selection criteria of adjusted \u003cem\u003ep value\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05 and average log2-fold change\u0026thinsp;\u0026gt;\u0026thinsp;0.5, resulting in 3,219 DEGs across 11 cell types (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. C). Gene Ontology (GO) analysis revealed enrichment in functions such as protein folding, heat shock protein binding, ribosomal structure, and cytoplasmic translation. KEGG pathway analysis revealed significant alterations in circadian rhythm, circadian entrainment, and cellular senescence (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD). Gene set enrichment analysis (GSEA) of the granulosa cell gene sets further highlighted biological processes related to ribosomal function, mitochondrial activity, and lipid metabolism (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. E).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e2.2 Screening Circadian Rhythm-Associated Genes\u003c/p\u003e \u003cp\u003eSingle-sample gene set enrichment analysis (ssGSEA) was used to calculate the circadian rhythm correlation score (CRCS) for each cell. The average CRCS values across cell types revealed marked heterogeneity in circadian gene expression following ovarian aging (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). The cells were subsequently stratified into high-CRCS and low-CRCS groups on the basis of circadian gene enrichment scores. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB, high-CRCS cells were predominantly localized in the GC1 subpopulation, whereas low-CRCS scores were enriched in smooth muscle and epithelial cells. These findings indicate that GC1 cells exhibit the strongest association with circadian rhythm activity.\u003c/p\u003e \u003cp\u003eGiven this observation, we performed weighted gene coexpression network analysis (WGCNA) on the GC1 subset to identify rhythm-associated gene modules. The soft-thresholding power was set to 6 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC), and two modules (GC1-M1 and GC1-M2) were identified. Both modules had module eigengene-based connectivity (KME) values greater than 0.6, indicating strong internal coherence. The top-ranked genes from each module were selected for further analysis.\u003c/p\u003e \u003cp\u003e2.3 Analysis of cell crosstalk and proposed time series analysis of core GC genes\u003c/p\u003e \u003cp\u003eTo explore the crosstalk between GCs and other cells in the ovarian microenvironment, we used CellChatDB to assimilate information on cellular interactions on the basis of a database of ligand‒receptor signals. Compared with those in the young group, GCs in the aging group exhibited a significantly reduced capacity for intercellular communication. The downregulation of signalling pathways such as the VEGF, PDGF, and FGF pathways suggests impaired proliferative and reparative functions of GCs. Suppressing the expression level of VISFATIN, a protein secreted by adipose tissue, may disrupt lipid metabolism and steroid hormone synthesis and disrupt the ovarian microenvironment. In contrast, the elevated expression of signalling molecules such as CXCL, PGN, WNT, and GPF may promote proinflammatory signalling (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA-C). These molecular alterations may serve as key biomarkers for the early diagnosis of POI and represent potential targets for therapeutic intervention. Monocle analysis further revealed changes in the developmental trajectory of ovarian GCs with aging, revealing that the developmental trajectory of cells in the older group significantly deviated from that of cells in the younger group, which may be related to the decrease in the differentiation potential of the cells or the intensification of cellular aging. We clustered the differentially expressed genes by monocle according to the expression trend on the proposed temporal trajectory to discover different dynamic regulation patterns. The expression trends of key genes on the proposed chronological trajectories were plotted (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD-F).\u003c/p\u003e \u003cp\u003e2.4 SCENIC analysis reveals basal motifs in granulosa cells\u003c/p\u003e \u003cp\u003eWe used SCENIC to identify a total of 280 transcription factors and motifs in different cellular subpopulations and screened 18 transcription factors whose transcriptional activities differed significantly between young and senescent GCs, of which 8 were significantly upregulated and 10 were significantly downregulated (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). The rhythm-related transcription factors CLOCK and ARNTL presented significantly greater transcriptional activity in the young GC group than in the aging GC group, and their gene expression also decreased with age. The downregulation of circadian rhythm-related gene activity may indicate that granulosa cell desynchronization occurs with aging. In contrast, the transcriptional activities of the JUN and CEPBP matrices were progressively upregulated, and their expression was significantly upregulated with age (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB-D). The major signalling pathways involved in these transcription factors, including the WNT pathway and the NF-κB pathway, highlight the involvement of proinflammatory factors, oxidative stress and other unfavourable factors in the aging process of GCs. The downregulation of circadian rhythm-related gene activity may indicate that granulosa cell desynchronization occurs with aging.\u003c/p\u003e \u003cp\u003e2.5 Summary-data-based Mendelian methods\u003c/p\u003e \u003cp\u003eWe performed summary-data-based Mendelian randomization (SMR) analyses by integrating eQTL data from both whole blood and ovarian tissue with genome-wide association study (GWAS) data for POI and identified a total of 120 circadian rhythm-related genes significantly associated with POI risk, all of which presented P(HEIDI) values greater than 0.05 (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). To evaluate whether these SMR-identified genes align with transcriptomic coexpression patterns, we compared them with hub genes derived from WGCNA modules. Notably, 30 of these genes overlapped with the GC1-specific circadian rhythm module (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). This overlap suggests that a substantial subset of rhythm-related genes under genetic regulation also demonstrate strong coexpression within aging-associated transcriptomic modules. Functional enrichment analysis of the 30 overlapping genes revealed significant associations with lipid metabolism, DNA repair, and circadian rhythm synchronization, further supporting their potential mechanistic involvement in the process of ovarian aging (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e2.6 Validation at the bulk RNA level and results of machine learning algorithms\u003c/p\u003e \u003cp\u003eWe screened a total of three sets of human GC RNA sequencing data for POI and NOR from the GEO database, all of which were high-throughput sequencing data. To ensure sample quality, we merged and cleaned the data. Batch effects between different datasets were removed via the R package \u0026lsquo;sva\u0026rsquo;, and datasets were normalized via the R package \u0026lsquo;preprocessCore\u0026rsquo;. PCA confirmed that the batch effect of the three datasets was low and that the data distribution was stable (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD). We verified the expression levels of 30 overlapping genes at the Bulk-RNA level, and 10 genes, such as CLOCK, ARNTL, and APOE, were identified as differentially expressed (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE). We independently assessed gene importance via four ML models: GLM, RF, KNN, and NNET. The residual boxplot results indicate that the KNN model has the best data fit, with the smallest residuals. In contrast, the GLM has higher median residuals and RMSE, and its interquartile range (IQR) is notably wider, suggesting that it has the largest residual errors and slightly poorer fit than the other three models. Additionally, the area under the ROC curve (AUC) for all four models was greater than 0.8, indicating that the predictive performance of all the models was satisfactory (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eF). Following Z score normalization of the gene importance scores across the four models, a soft voting ensemble learning method was employed for gene selection. As a result, the core genes were ultimately determined: CLOCK, CRY1, APOE and GSTA1.\u003c/p\u003e \u003cp\u003e2.7 WB and qPCR confirmed the expression of circadian genes in granulosa cells\u003c/p\u003e \u003cp\u003eWe used WB and qPCR to confirm the differences in the expression of overlapping genes in primary GCs, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA-B. Compared with that in the control group, the expression of GSTA1 was upregulated in the model group, and the expression of APOE was downregulated. The expression of the transcription factors BMAI1, CLOCK, and CRY1 was downregulated. Under physiological conditions, the expression levels of the CLOCK, CRY1, APOE and GSTA1 genes exhibit marked circadian rhythmicity over a 24-hour cycle, with peak expression of CLOCK and APOE occurring at Zeitgeber Time 8 (ZT8), CRY1 peaking at ZT12, and GSTA1 peaking at ZT20. In contrast, the model group presented attenuated rhythm amplitude and a phase delay, indicating a disruption in circadian rhythm synchronization.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e2.8 CLOCK\u003csup\u003e\u0026minus;/\u003c/sup\u003eGCs show an aging phototype\u003c/p\u003e \u003cp\u003eWe employed siRNA to knock down the CLOCK gene in primary mouse cells and subsequently evaluated the cellular senescence phenotype. Western blot analysis revealed that, compared with negative control cells, CLOCK-deficient cells presented significant upregulation of P16 and P21 protein expression, indicating that CLOCK gene knockdown may contribute to the induction of GC senescence.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eOur integrated multiomics analysis provides compelling evidence that circadian clock disruption in GCs plays a central role in the pathogenesis of POI. GCs\u0026mdash;key somatic supporters of oocyte development\u0026mdash;undergo profound circadian dysregulation with ovarian aging, contributing to their functional decline. This elevates circadian misregulation from a peripheral observation to a mechanistic hallmark of premature ovarian aging.\u003c/p\u003e \u003cp\u003eUsing snRNA-seq, we first identified a GC subtype, GC1, characterized by a high enrichment of circadian gene expression, suggesting its role as an intrinsic follicular \"timekeeper\". Through WGCNA, we identified circadian-related gene modules within the GC1 subtype that converge on critical pathways, including lipid metabolism, protein folding, and DNA repair\u0026mdash;processes essential for the proper development of follicles.\u003c/p\u003e \u003cp\u003eThe consequences of circadian disruption were evident in aged GCs, which presented significantly diminished intercellular communication. Downregulated signalling pathways such as the VEGF, PDGF, and FGF pathways suggest impaired vascular support, proliferation, and survival. Moreover, increased expression of proinflammatory and stress-related signalling molecules, including CXCL, PGN, and WNT, implies a shift toward an inflammatory microenvironment[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Circadian regulators, such as CLOCK/BMAL1, orchestrate various physiological rhythms[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]; diminished activity of these factors may disrupt follicular nutrient metabolism and activate inflammatory signalling pathways. SCENIC analysis further demonstrated that the transcriptional activity of the core circadian genes CLOCK/BMAL1 declines, whereas the stress response factors JUN and CEBPB exhibit increased activity. This reciprocal shift\u0026mdash;from circadian regulation to stress-dominated transcription\u0026mdash;reflects a transition in the regulatory landscape of aging GCs, reinforcing the role of circadian erosion in driving ovarian dysfunction.\u003c/p\u003e \u003cp\u003eA key question is whether circadian rhythm alterations in POI are merely correlative or truly causative. Our SMR analysis revealed a causal relationship and identified 120 circadian-related genes whose expression is influenced by genetic variants associated with POI. Notably, 30 of these genes overlap with circadian transcriptional modules in GC1, integrating both genetic and transcriptomic evidence. These genes are enriched in pathways related to mitochondrial function, metabolism, and DNA repair\u0026mdash;critical processes that govern oocyte quality and follicle longevity\u0026mdash;thereby establishing a direct mechanistic link between circadian disruption and POI pathogenesis.\u003c/p\u003e \u003cp\u003eCompared with those of the controls, the protein and mRNA levels of CLOCK, BMAL1, CRY1, and APOE, alongside markedly elevated GSTA1, were significantly lower in the model group. Circadian rhythm assays further revealed disrupted intrinsic rhythmicity of CLOCK, CRY1, APOE and GSTA1 in the model group, characterized by delayed phases and decreased amplitudes. Following CLOCK knockdown in primary mouse granulosa cells, the P16 and P21 proteins were notably upregulated, leading to cell cycle arrest and progression to irreversible senescence. The concordance between multiomics predictions and in vitro results not only strengthens confidence in our findings but also highlights a broad circadian network of coregulated genes. Together, our data suggest that GCs harbor local circadian clocks critical for maintaining ovarian function. In addition to governing hormone secretion and ovulation timing, these intrinsic clocks orchestrate cell cycle progression and intercellular communication. Disruption of this timing system induces desynchronization within the follicle, contributing to accelerated ovarian aging.\u003c/p\u003e \u003cp\u003eGSTA1, a glutathione S-transferase involved in antioxidant defense and apoptosis regulation, detoxifies lipid peroxides and harmful metabolites to alleviate oxidative stress[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Its hepatic expression has been shown to follow circadian regulation by CLOCK/BMAL1, potentially determining peak detoxification timing[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. In this study, GSTA1 was significantly upregulated in the CTX-induced granulosa cell injury model, suggesting a compensatory antioxidant response. These findings imply that circadian genes may modulate oxidative stress in granulosa cells via the regulation of GSTA1[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAPOE, expressed in ovarian granulosa cells, facilitates cholesterol uptake and transport, contributing to steroid hormone synthesis[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Several clinical studies have reported that women carrying the APOEε4 allele tend to exhibit increased reproductive potential[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. Here, APOE expression was decreased in aging GCs, accompanied by dampened circadian rhythmicity, suggesting that APOE may be transcriptionally regulated by the CLOCK/BMAL1 axis and influence hormone biosynthesis through circadian mechanisms.\u003c/p\u003e \u003cp\u003eClinically, our findings facilitate early diagnosis and intervention. Circadian genes, including CLOCK, BMAL1, CRY1, GSTA1, and APOE, have potential as early biomarkers for a diminished ovarian reserve; monitoring their expression in granulosa cells could identify women at risk of POI, especially those exposed to environmental or iatrogenic factors. Lifestyle modifications aligned with circadian rhythms or chronotherapy might offer noninvasive strategies to protect ovarian function. Additionally, pharmacological modulation of circadian regulators represents a promising therapeutic avenue, pending further validation.\u003c/p\u003e \u003cp\u003eHowever, several limitations should be noted. The modest sample size of our single-cell dataset limits the exploration of population-level heterogeneity. Although the incidence of POI is increasing, clinical sample acquisition remains challenging, constraining the reliability and generalizability of our findings. Furthermore, the specificity of the SMR results is restricted by the available GWAS data. Finally, since in vitro models cannot fully replicate ovarian physiology, further in vivo validation is essential.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThrough multiomics integration, we demonstrated that disruption of circadian rhythms in the ovary contributes to the accelerated development of POI and identified CLOCK, BMAL1, CRY1, GSTA1, and APOE as potential POI-associated risk genes. These findings provide a foundation for innovative diagnostic and therapeutic approaches to safeguard reproductive longevity and improve women\u0026rsquo;s health.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003ePremature ovarian insufficiency \u0026nbsp; POI\u003c/p\u003e\n\u003cp\u003esingle-nuclear RNA sequencing \u0026nbsp; snRNA-seq\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;bulk RNA sequencing \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;RNA-seq\u003c/p\u003e\n\u003cp\u003egenome-wide association studies \u0026nbsp;GWAS\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;expression quantitative trait loci \u0026nbsp;eQTL\u003c/p\u003e\n\u003cp\u003esummary-data-based Mendelian\u0026nbsp;randomization \u0026nbsp;\u0026nbsp;SMR\u003c/p\u003e\n\u003cp\u003egranulosa cells \u0026nbsp; GCs\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFinnGen (https://www.finngen.finn/en);\u003c/p\u003e\n\u003cp\u003eGEO (https://www.ncbi.nlm.nih.gov/geo/);\u003c/p\u003e\n\u003cp\u003eCGDB (http://cgdb.biocuckoo.org)/\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions:\u0026nbsp;\u003c/strong\u003eConceptualization, RJ;\u0026nbsp;methodology, RJ and CJ;\u0026nbsp;data curation, CJ;\u0026nbsp;writing-original draft preparation, RJ and TY;\u0026nbsp;writing-review \u0026amp; editing, TY. All authors reviewed the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding statement:\u0026nbsp;\u003c/strong\u003eThis study was supported by the National Natural Science Foundation of China (No. 82330125). The funding source did not influence the study design, analyses,\u0026nbsp;or\u0026nbsp;interpretation of\u0026nbsp;the\u0026nbsp;data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDisclosure Statement:\u003c/strong\u003e None.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData sharing statement:\u0026nbsp;\u003c/strong\u003eAll the data in this paper are publicly available.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eGenes linked to premature ovarian insufficiency show no pathogenicity in the general population. Nat Med 2023;29:1617\u0026ndash;8.\u003c/li\u003e\n\u003cli\u003eBompoula MS, Valsamakis G, Neofytou S, et al. Demographic, clinical and hormonal characteristics of patients with premature ovarian insufficiency and those of early menopause: data from two tertiary premature ovarian insufficiency centers in Greece. Gynecological Endocrinology 2020;36:693\u0026ndash;7.\u003c/li\u003e\n\u003cli\u003eThe Lancet. Time for a balanced conversation about menopause. The Lancet 2024;403:877. https://doi.org/10.1016/S0140-6736(24)00462-8.\u003c/li\u003e\n\u003cli\u003eLi Y, Pei T, Zhu H, et al. Melatonin Alleviates Circadian Rhythm Disruption-Induced Enhanced Luteinizing Hormone Pulse Frequency and Ovarian Dysfunction. J Pineal Res 2025;77.\u003c/li\u003e\n\u003cli\u003eLane JM, Qian J, Mignot E, et al. Genetics of circadian rhythms and sleep in human health and disease. Nat Rev Genet 2023;24:4\u0026ndash;20.\u003c/li\u003e\n\u003cli\u003eZheng Y, Liu C, Li Y, et al. Loss-of-function mutations with circadian rhythm regulator Per1/Per2 lead to premature ovarian insufficiency. Biol Reprod 2019;100:1066\u0026ndash;72.\u003c/li\u003e\n\u003cli\u003eZhang A, Li S, Huang L, et al. Bmal1 regulates female reproduction in mice via the hypothalamic\u0026ndash;pituitary\u0026ndash;ovarian axis. FASEB Journal 2024;38.\u003c/li\u003e\n\u003cli\u003eKawamura T, Dai Y, Ono M, et al. BMAL1 positively correlates with genes regulating steroidogenesis in human luteinized granulosa cells. Reproduction 2024;167.\u003c/li\u003e\n\u003cli\u003eWu J, Lv Y, Hao P, et al. Immunological profile of lactylation-related genes in Crohn\u0026rsquo;s disease: a comprehensive analysis based on bulk and single-cell RNA sequencing data. J Transl Med 2024;22.\u003c/li\u003e\n\u003cli\u003eDu C, Wang C, Liu Z, et al. Machine learning algorithms integrate bulk and single-cell RNA data to unveil oxidative stress following intracerebral hemorrhage. Int Immunopharmacol 2024;137.\u003c/li\u003e\n\u003cli\u003eLe DC, Ngo MHT, Kuo YC, et al. Secretome from estrogen-responding human placenta-derived mesenchymal stem cells rescues ovarian function and circadian rhythm in mice with cyclophosphamide-induced primary ovarian insufficiency. J Biomed Sci 2024;31:95.\u003c/li\u003e\n\u003cli\u003eAn Q, Zheng N, Ji Y, et al. Exploration the interaction of cadmium and copper toxic effects in pakchoi (Brassica chinensis L) roots through combinatorial transcriptomic and weighted gene coexpression network analysis. J Environ Manage 2024;359. https://doi.org/10.1016/J.JENVMAN.2024.120956.\u003c/li\u003e\n\u003cli\u003eSu J, Song Y, Zhu Z, et al. Cell\u0026ndash;cell communication: new insights and clinical implications. Signal Transduct Target Ther 2024;9. https://doi.org/10.1038/S41392-024-01888-Z.\u003c/li\u003e\n\u003cli\u003eClyde D. Single cell\u0026ndash;cell communication. Nat Rev Genet 2023;24:488.\u003c/li\u003e\n\u003cli\u003eHymel LA, Anderson SE, Turner TC, et al. Identifying dysregulated immune cell subsets following volumetric muscle loss with pseudotime trajectories. Commun Biol 2023;6.\u003c/li\u003e\n\u003cli\u003eAsnicar F, Thomas AM, Passerini A, et al. Machine learning for microbiologists. Nat Rev Microbiol 2024;22:191\u0026ndash;205. https://doi.org/10.1038/S41579-023-00984-1.\u003c/li\u003e\n\u003cli\u003eSamet H. K-nearest neighbor finding using MaxNearestDist. IEEE Trans Pattern Anal Mach Intell 2008;30:243\u0026ndash;52. https://doi.org/10.1109/TPAMI.2007.1182.\u003c/li\u003e\n\u003cli\u003eLin P-W, Lin Z-R, Wang W‒W, et al. Identification of immune-inflammation targets for intracranial aneurysms: A multiomics and epigenome-wide study integrating summary-data-based mendelian randomization, single-cell-type expression analysis, and DNA methylation regulation. International Journal of Surgery 2024;111.\u003c/li\u003e\n\u003cli\u003eZhou C, Gao Y, Ding P, et al. The role of CXCL family members in different diseases. Cell Death Discov 2023;9:212.\u003c/li\u003e\n\u003cli\u003eKulkarni SS, Singh O, Zigman JM. The intersection between ghrelin, metabolism and circadian rhythms. Nat Rev Endocrinol 2024;20:228\u0026ndash;38.\u003c/li\u003e\n\u003cli\u003eMcAlpine CS, Swirski FK. Circadian influence on metabolism and inflammation in atherosclerosis. Circ Res 2016;119:131\u0026ndash;41.\u003c/li\u003e\n\u003cli\u003eMa S, Xie F, Wen X, et al. GSTA1/CTNNB1 axis facilitates sorafenib resistance by suppressing ferroptosis in hepatocellular carcinoma. Pharmacol Res 2024;210. https://doi.org/10.1016/J.PHRS.2024.107490.\u003c/li\u003e\n\u003cli\u003eXu YQ, Zhang D, Jin T, et al. Diurnal Variation of Hepatic Antioxidant Gene Expression in Mice. PLoS One 2012;7.\u003c/li\u003e\n\u003cli\u003eDai C, Sharma G, Liu G, et al. Therapeutic detoxification of quercetin for aflatoxin B1-related toxicity: Roles of oxidative stress, inflammation, and metabolic enzymes. Environmental Pollution 2024;345. https://doi.org/10.1016/J.ENVPOL.2024.123474.\u003c/li\u003e\n\u003cli\u003eOri\u0026aacute; RB, de Almeida JZ, Moreira CN, et al. Apolipoprotein E Effects on Mammalian Ovarian Steroidogenesis and Human Fertility. Trends in Endocrinology and Metabolism 2020;31:872\u0026ndash;83.\u003c/li\u003e\n\u003cli\u003eLu H, Jing Y, Zhang C, et al. Aging hallmarks of the primate ovary revealed by spatiotemporal transcriptomics. Protein Cell 2024;15:364\u0026ndash;84.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Premature ovarian insufficiency, Circadian rhythm, Granulosa cells, Multiomics","lastPublishedDoi":"10.21203/rs.3.rs-6850329/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6850329/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cb\u003eBackground\u003c/b\u003e\u003c/p\u003e \u003cp\u003ePremature ovarian insufficiency (POI), defined as the loss of ovarian function before the age of 40 years, severely disrupts reproductive and endocrine health. While circadian rhythm disruption has been implicated in reproductive aging, its mechanistic contribution to POI remains largely unexplored in humans.\u003c/p\u003e\u003cp\u003e\u003cb\u003eMethods\u003c/b\u003e\u003c/p\u003e \u003cp\u003eWe performed an integrative multiomics analysis combining single-nucleus RNA sequencing (snRNA-seq), bulk RNA-seq, genome-wide association studies (GWAS), and expression quantitative trait locus (eQTL) data. Co-expression networks were constructed via weighted gene coexpression network analysis (WGCNA), while cell-cell communication and trajectory analyses were conducted via CellChat and Monocle. Regulatory networks were inferred via SCENIC, and causality was assessed via summary-data-based Mendelian randomization (SMR). Candidate hub genes were prioritized through machine learning and validated via in vitro assays assessing rhythmicity and gene expression.\u003c/p\u003e\u003cp\u003e\u003cb\u003eResults\u003c/b\u003e\u003c/p\u003e \u003cp\u003esnRNA-seq identified a granulosa cell subpopulation (GC1) with the highest circadian rhythm score, suggesting a pivotal role in regulating the ovarian clock. WGCNA and SCENIC analyses revealed age-associated downregulation of the core circadian regulators CLOCK and ARNTL, accompanied by disruptions in lipid metabolism and stress response pathways. SMR analysis revealed 120 circadian-related genes associated with POI risk, 30 of which were enriched in GC1-specific modules. CLOCK, CRY1, APOE, and GSTA1 emerged as key regulators on the basis of machine learning prioritization. Functional assays confirmed impaired rhythmicity and altered gene expression in KGN cells and senescent mouse granulosa cells. CLOCK knockdown increased P16 and P21 expression, underscoring its role in preserving granulosa cell homeostasis.\u003c/p\u003e\u003cp\u003e\u003cb\u003eConclusions\u003c/b\u003e\u003c/p\u003e \u003cp\u003eOur findings implicate circadian rhythm disruption as a hallmark and potential driver of ovarian aging. CLOCK, BMAL1, CRY1, APOE, and GSTA1 may serve as early biomarkers and therapeutic targets for POI.\u003c/p\u003e","manuscriptTitle":"Multiomics analyses reveal key circadian rhythm genes implicated in Premature ovarian insufficiency","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-20 11:51:50","doi":"10.21203/rs.3.rs-6850329/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"8a09e61a-92ac-4ddb-b905-70a665ecd3ea","owner":[],"postedDate":"June 20th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":50257727,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":50257728,"name":"Biological sciences/Systems biology"}],"tags":[],"updatedAt":"2025-07-09T16:40:59+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-20 11:51:50","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6850329","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6850329","identity":"rs-6850329","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}