Skip to main content
Download PDF

Molecular characteristics of early- and late-onset ovarian cancer: insights from multidimensional evidence

Journal of Ovarian Research volume 18, Article number: 83 (2025) Cite this article

Abstract

Background

Ovarian cancer (OC) is among the most lethal gynecologic malignancies, characterized by poor prognosis. While aging is a well-established risk factor, the underlying mechanisms distinguishing early- and late-onset ovarian cancer remain poorly understood.

Methods

This study analyzed the global burden and age-related trends of ovarian cancer using the GBD database. A cut-off age of 55 years was used to differentiate between early and late onset ovarian cancer, and a Mendelian randomization method was also used to investigate the causal relationship between aging and ovarian cancer. Machine learning was applied to identify tumor-specific age-associated genes, followed by bioinformatics analyses and single-cell sequencing to explore the roles of these genes and immune profile alterations in ovarian cancer. Additionally, models were constructed, and drug sensitivity analyses performed to evaluate their potential as diagnostic markers or therapeutic targets.

Results

Ovarian cancer incidence and mortality exhibit age-related trends, with telomere length positively associated with increased risk (OR = 1.27, 95% CI: 1.01–1.60, P = 3.90 × 10⁻2). Older patients with OC have a worse prognosis. PRKCD and UCP2 were significantly upregulated in ovarian cancer. PRKCD facilitates epithelial-mesenchymal transition (EMT), contributing to ovarian cancer progression, while UCP2 modulates ROS dynamics, influencing chemoresistance. Immune microenvironment analysis revealed differences between high- and low-expression groups, particularly in T cells, macrophages, and other immune cells. Both genes are sensitive to a varity of drugs, including dasatinib, fluvastatin, highlighting their potential as therapeutic targets.

Conclusion

Aging is a significant risk factor for ovarian cancer, with PRKCD and UCP2 closely linked to its onset and progression. These genes show promise as novel biomarkers and therapeutic targets for ovarian cancer.

Introduction

Ovarian cancer (OC) is a significant malignancy among women, with a much higher mortality rate than other cancers, ranking sixth [1]. In 2024, an estimated 19,680 new cases and 12,740 deaths from ovarian cancer are projected in the United States [1]. Similarly, recent statistics indicate that China is expected to report approximately 52,100 new cases and 20,880 deaths [2, 3]. Due to the difficulty of early diagnosis, the 5-year survival rate for ovarian cancer is extremely low at 25–35% [4]. In modern oncology, the primary therapeutic modalities for malignancies encompass surgical resection and molecularly targeted therapies. A profound understanding of the molecular mechanisms underlying carcinogenesis can significantly enhance therapeutic efficacy [5]. With rising global life expectancy, ovarian aging has emerged as a pressing health concern for menopausal women [6]. Ovarian function begins to decline after age 30 and typically ceases around 50, increasing the risk of conditions such as cardiovascular disease, diabetes, and cancer [7, 8]. Thus, identifying novel biomarkers is crucial for improving the diagnosis and treatment of ovarian cancer.

Aging is a key risk factor for many cancers, with incidence rising significantly with age [9]. Natural aging is accompanied by the gradual accumulation of senescent cells, which display heterogeneous phenotypes and can exert both anti-tumor and pro-tumor effects [10]. Recent studies indicate that aging plays a dual role in ovarian cancer. Biomarkers exhibit diverse functional roles in tumor biology. Sodium channel β3 subunit (SCN3B) has emerged as a potential prognostic biomarker for survival prediction in glioma patients, while cyclin-dependent kinase 2 (CDK2) has been identified as a promising diagnostic and prognostic marker in glioblastoma [11, 12]. Other studies have also shown the importance of biomarkers in tumors [13,14,15]. While it is linked to enhanced immune cell infiltration and improved responses to immunotherapy [16], it also contributes to an increased burden of senescent cells, which may exacerbate ascites formation and drive tumorigenesis and metastasis [17, 18]. Research on aging and ovarian cancer risk has largely focused on ovarian aging, with little exploration of the differences in the pathogenesis between early- and late-onset ovarian cancer.

The Global Burden of Disease (GBD) study is a comprehensive epidemiological analysis that quantifies global exposures to major diseases and risk factors using standardized methods, enabling comparisons across populations and time [19]. Mendelian randomization (MR), leveraging single nucleotide polymorphisms (SNPs), assesses causal relationships between exposures and outcomes. By relying on the random distribution of alleles associated with exposure, MR effectively minimizes confounding in observational studies and mitigates bias from reverse causation [20, 21]. Machine learning, broadly defined as the process of developing predictive models or uncovering patterns within data, has become a widely adopted tool in medicine [22, 23]. Traditional machine learning algorithms encompass a variety of approaches, with Random Forests and Support Vector Machines (SVMs) being among the most commonly used. Cancer, as a complex, heterogeneous, and multifaceted disease, generates vast amounts of multimodal data, making clinical oncology a highly promising field for machine learning applications [24].

In the present study, we combined a large-scale observational analysis with Mendelian randomization to investigate the causal relationship between aging and ovarian cancer risk. To explore the mechanisms underlying early- and late-onset ovarian cancer, we integrated data from the GEO and HAGR databases to identify key genes. Functional enrichment, immune infiltration, and single-cell analyses were subsequently employed to uncover the molecular mechanisms driving early and late disease onset. Simultaneously, we developed models to evaluate the predictive capability of the identified key genes for ovarian cancer risk and their potential as novel biomarkers.

Methods

Data sources

The observational data for this study were sourced from the publicly accessible Global Burden of Disease (GBD) Outcomes Database, which provides comprehensive estimates of ovarian cancer incidence, mortality, and disability-adjusted life years (DALYs) across 204 countries and territories from 1990 to 2021. Variations in ovarian cancer indicators were analyzed based on regional categorization by the Socio-Demographic Index (SDI) and age stratification into 20–55 years and > 55 years [25, 26]. 55 years as an age cut-off to distinguish early and late onset ovarian cancer.

RNA-seq data for ovarian cancer were retrieved from the NCBI GEO database (https://www.ncbi.nlm.nih.gov/geo/), including datasets GSE66957 (12 normal and 57 tumor samples) and GSE54388 (6 normal and 16 tumor samples) [27]. Differential gene expression analysis was conducted using the ‘limma’ R package, with genes meeting the criteria of |logFC|> 1 and P < 0.05 identified as differentially expressed.

Single-cell RNA-seq data were obtained from dataset GSE184880, comprising 5 normal and 7 tumor tissue samples [28]. Low-quality cells with mitochondrial gene content exceeding 10% were excluded. Data analysis and processing were performed using the Seurat R package. Cell clustering was achieved by the Uniform Manifold Approximation and Projection (UMAP) method while using ‘FindAllMarkers’ to find key genes for cell class group annotation.

Genome-Wide Association Studies (GWAS) Data Sources and Selection of IVs

GWAS data on aging-related markers and ovarian cancer were sourced from the IEU Open GWAS project (https://gwas.mrcieu.ac.uk/), focusing on populations of European ancestry to reduce bias from population stratification. Based on variations in the age distribution of ovarian cancer incidence, we examined telomere length, frailty index, and facial aging as aging markers to investigate the causal relationship between aging and ovarian cancer risk. To ensure the robustness of the Mendelian randomization (MR) analysis, the instrumental variables (IVs) were selected based on three key assumptions: (1) strong association with the exposure (aging markers), (2) independence from confounding variables, and (3) influence on the outcome (ovarian cancer risk) exclusively through the exposure pathway. SNPs significantly associated with the exposure were identified using a threshold of P < 5 × 10⁻⁸. To address linkage disequilibrium, highly correlated SNPs were excluded using the European reference panel, retaining those with the smallest P values within a 10,000 kb window if R2 < 0.001. Additionally, SNPs with F-statistics > 10 were selected to minimize weak instrument bias [29].

Mendelian randomization

Mendelian randomization (MR), which calculates associations between exposures and outcomes from SNPs, is effective in reducing bias from confounders and reverse causation [30, 31]. In a two-sample MR analysis, inverse variance weighting (IVW) was the primary method used to evaluate the causal relationship between aging and ovarian cancer. To confirm the robustness of IVW results, complementary analyses were conducted using MR-Egger, weighted median, weighted mode, and simple mode methods. Sensitivity analyses were performed to ensure result reliability. Cochrane's Q-test assessed potential heterogeneity, with P < 0.05 indicating significant heterogeneity. Horizontal pleiotropy was evaluated via the MR-Egger intercept, with P < 0.05 suggesting potential pleiotropic bias. Additionally, MR-PRESSO identified and excluded SNPs with pleiotropic effects [32]. Leave-one-out analyses were also conducted to determine if findings were influenced by any single SNP.

Functional enrichment analysis

Gene Ontology (GO) enrichment analysis, encompassing Biological Process (BP), Cellular Component (CC), Molecular Function (MF), and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways, alongside Gene Set Enrichment Analysis (GSEA) using MSigDB gene sets, was performed to uncover the biological functions and potential signaling pathways of key genes [33].

Machine learning

Ovarian cancer-specific senescence-associated genes were identified through LASSO regression, random forests, and SVM-RFE. LASSO performs gene selection by imposing L1 regularization to shrink unimportant features'coefficients to zero. Random Forest assesses gene importance by calculating feature importance scores based on decision tree ensembles. SVM-RFE iteratively removes features with the least contribution to the model's performance, selecting the most relevant genes [34,35,36]. Key genes were determined by intersecting the results using a Venn diagram.

Immune infiltration analysis

Ovarian cancer patient data were analyzed using the CIBERSORT algorithm to estimate the proportions of 22 immune infiltrating cell types, including T and B cells. Correlation analyses between gene expression and immune cell composition were conducted using the TIMER database [37, 38].

Drug sensitivity analysis

Using the GDSC cancer drug sensitivity genomics database (https://www.cancerrxgene.org/), chemotherapy sensitivity for each individual was predicted with the"pRRophetic"R package. Drug sensitivity analysis is performed by linear regression of gene expression data to estimate the contribution of each gene to IC50, while avoiding multicollinearity and improving model stability.

Statistical analysis

The distribution of immune infiltrating cells was evaluated using the Wilcoxon rank-sum test. Survival outcomes were compared with the Kaplan–Meier method and log-rank test. All analyses were conducted using R (v4.3.0), with two-sided statistical tests and P < 0.05 considered significant.

Results

Global burden and age trends in ovarian cancer

Globally, the incidence of ovarian cancer rose significantly from 159.1 thousand cases (95% CI: 145. 7–174.1) in 1990 to 298,900 cases (95% CI: 270.7–324.5) in 2021, reflecting an 87.9% increase. This growth was accompanied by corresponding rises in DALYs and mortality (Supplementary Table 1). The incidence and mortality rates varied notably across age groups and geographic locations, with positive correlations observed between these metrics and the SDI (s Supplementary Table 2). Furthermore, the age-specific analysis revealed a marked upward trend in both incidence and mortality with advancing age (Fig. 1).

Fig. 1
figure 1

Global burden of ovarian cancer and age trends. A line graphs and (B) bar graphs of incidence and mortality of ovarian cancer by age subgroups

Full size image

Causal association between aging markers and ovarian cancer

Following stringent screening criteria, 146, 14, and 40 SNPs were selected as instrumental variables for each marker, respectively, all with F-statistics > 10 (Supplementary Table 3). Mendelian randomization analysis revealed a significant positive association between telomere length and ovarian cancer risk (OR = 1.27, 95% CI: 1.01–1.60, P = 3.90 × 10⁻2), other approaches have verified the existence of this association (Fig. 2A-B). In contrast, no significant associations were observed for frailty index or facial aging with ovarian cancer (P > 0.05) (Fig. 2C-D).

Fig. 2
figure 2

Causal association between aging and ovarian cancer. A Forest plot of causal associations between aging markers and ovarian cancer risk. B Scatterplot of telomere length versus ovarian cancer risk. C Scatterplot of frailty index versus ovarian cancer risk. D Scatterplot of facial aging versus ovarian cancer risk

Full size image

To ensure the reliability of our findings, we conducted several sensitivity analyses. Cochrane's Q-test indicated no heterogeneity between aging markers and ovarian cancer, which suggests that the association of these markers with ovarian cancer is consistent across the different datasets. MR-Egger intercept and MR-PRESSO analyses detected no horizontal pleiotropy, and no significant outliers were identified (Supplementary Table 3). This enhances the robustness of the findings and gives more credibility to the results. Additionally, a"leave-one-out"analysis confirmed that the observed effects were not driven by any single SNP (Supplementary Fig. 2). Overall, the MR analysis established a significant association between aging and the risk of ovarian cancer.

Identification of key genes for early and late onset ovarian cancer

Given the association between aging and ovarian cancer risk, RNA-Seq data from GSE66957 and GSE54388 were analyzed for differential expression using the criteria |logFC|> 1 and P < 0.05 (Fig. 3A-B). A Venn analysis combining the differentially expressed genes from both datasets with the aging-related gene set identified 14 overlapping genes (Fig. 3C). These genes exhibited strong intrinsic associations, with PARP1, JUN1, and AKT1 serving as key regulators. These core factors were interconnected with multiple genes, forming a complex interaction network that modulates various signaling pathways, ultimately influencing cancer progression (Fig. 3D). Enrichment analysis revealed that these genes were associated with key pathways, including apoptosis, base excision repair, and TNF signaling (Fig. 3E). GO analysis indicated enrichment in biological processes such as responses to reactive oxygen species and oxidative stress. In cellular components, the genes were primarily clustered in the nuclear chromosome and protein-DNA complexes, while molecular functions highlighted various transferase activities (Fig. 3F). Multiple machine learning analyses were conducted to identify key genes among the 14 candidates. Lasso regression revealed that PRKCD, UCP2, JUN, and FAS were significant contributors (Fig. 3G, Supplementary Fig. 3 A). Integration of Random Forest and SVM-RFE further refined the selection, pinpointing PRKCD and UCP2 as the genes most closely associated with tumor-specific senescence (Fig. 3H-I, Supplementary Fig. 3B). Multiple complementary methodologies were employed to ensure the robustness and accuracy of these findings (Supplementary Fig. 3 C-D).

Fig. 3
figure 3

Screening of tumor-specific age-related genes. A Volcano plot of differential gene expression from GSE66957. B Volcano plot of differential gene expression from GSE54388. C Venn diagram identifying tumor-specific age-related genes. D PPI network analysis of 14 candidate genes. E–F KEGG and GO enrichment analyses of the 14 genes. (G-I) Machine learning-based selection of key genes. J Expression levels of key genes in tumor and normal tissues, stratified by age

Full size image

Functions of key genes in ovarian cancer

To investigate the roles of PRKCD and UCP2 in ovarian cancer, we first analyzed their expression across cancer stages and found that both were significantly overexpressed in tumor tissues. Notably, their expression also showed significant differences when stratified by age (Fig. 3J). Furthermore, older patients exhibited worse prognoses (Supplementary Fig. 4A). After adjustment, we found that high expression of PRKCD was associated with shorter survival, while UCP2 showed the opposite trend (Fig. 4A-B). To elucidate potential mechanisms, we performed GSEA and observed that high PRKCD expression was enriched for Hallmark pathways such as epithelial-mesenchymal transition (EMT), inflammatory response, and interferon-α response. Intriguingly, these pathways were also enriched in patients with high UCP2 expression (Fig. 4C-D). To evaluate their diagnostic potential, we built a diagnostic model using GSE54388 as a validation set. PRKCD demonstrated an AUC of 0.990, while UCP2 achieved an AUC of 0.969, indicating their strong potential as biomarkers for ovarian cancer diagnosis. These findings highlight the significant roles of PRKCD and UCP2 in ovarian cancer and suggest their promise as novel diagnostic markers.

Fig. 4
figure 4

Molecular characterization of key genes. A Kaplan–Meier curves of PRKCD high and low expression subgroups. B Kaplan–Meier curves of high and low expression groups of UCP2. C Enrichment of GSEA pathways in patients with high and low PRKCD expression using the Hallmark gene set. D Enrichment of GSEA pathways in patients with high and low UCP2 expression using the Hallmark gene set. AUC curve area for ovarian cancer diagnosis using PRKCD (E) and UCP2 (F)

Full size image

Tumor Microenvironmental Characterization of Key Genes

To investigate the relationship between key genes and alterations in the tumor microenvironment, we conducted immunogenomic analyses focusing on specific components of the tumor microenvironment. Using the CIBERSORT algorithm, we calculated the immune cell composition for individual ovarian cancer patients and observed that immune profiles varied significantly among patients (Fig. 5A). Patients were stratified by tumor versus normal tissue, as well as high and low expression levels of PRKCD and UCP2. Notably, B cells and macrophages differed significantly between tumor and normal tissues, while T cells and plasma cells showed distinct changes in the high- and low-expression groups of both PRKCD and UCP2 (Fig. 5B-D). Consistent with these findings, analysis of the TIMER database revealed significant associations between PRKCD and UCP2 expression and immune cells, particularly B cells and macrophages (Supplementary Fig. 5 A-B).

Fig. 5
figure 5

Immunological characterization of key genes. A Components of the immune microenvironment in individual ovarian cancer patients. Based on the CIBERSORT algorithm, the differences in the immune microenvironment between tumor-normal tissues (B), high and low PRKCD expression (C) and high and low UCP2 expression (D). E UMAP plots of all clusters after quality control. (F) Scale diagram of individual somatic cells. G Plot of PRKCD and UCP2 expression versus cell ratio. Expression of PRKCD (H) and UCP2 (I) among the fractions. expression of PRKCD (J) and UCP2 (K) in various immune cells

Full size image

To further validate changes in the tumor microenvironment, we analyzed single-cell RNA-seq datasets from ovarian cancer patients. Cells were filtered based on stringent inclusion criteria (Supplementary Fig. 6 A), and the optimal dimensionality for analysis was determined using PCA (Supplementary Fig. 6B). UMAP dimensionality reduction revealed the distribution and heterogeneity of tumor microenvironment (TME) cells, showing distinct immune cell profiles between tumor and normal tissues. These cells were classified into nine major cell types, including B cells, T cells, and NK cells (Fig. 5E, Supplementary Fig. 6 C-D). Additionally, the proportion of cells varied among individuals (Fig. 5F), and the expression levels of PRKCD and UCP2 were correlated with cell proportions (Fig. 5G). We then examined the expression of PRKCD and UCP2 in the immune microenvironment at the single-cell level. Both genes were highly expressed in tumors, consistent with bulk RNA-seq results (Supplementary Fig. 6E-F). However, their distributions differed: UCP2 exhibited significantly higher expression levels compared to PRKCD (Fig. 5H-K). PRKCD was primarily localized to T cells, monocytes, and NK cells, while UCP2 was broadly expressed across various immune cell types (Fig. 5I-J).

These results highlight the differential immune landscape associated with PRKCD and UCP2 expression, suggesting their roles in modulating the tumor microenvironment in ovarian cancer.

Drug sensitivity of key genes

Given the critical roles of PRKCD and UCP2 in ovarian cancer, we investigated their potential as therapeutic targets. Using the GDSC database and gene expression data from ovarian cancer patients, we estimated drug sensitivity and IC50 values. PRKCD exhibited sensitivity to drugs such as vorinostat and dasatinib, while UCP2 responded to agents like gemcitabine and cytarabine (Fig. 6A-B). Additionally, analysis through the GSCA database revealed that both PRKCD and UCP2 showed the strongest correlation with bleomycin (Fig. 6C). Further exploration using the CTRP database identified a strong association between PRKCD and cytochalasin B, while UCP2 demonstrated a notable correlation with vorinostat (Fig. 6D). These findings highlight significant links between the expression of PRKCD and UCP2 and drug sensitivity, suggesting that these genes may serve as promising therapeutic targets for ovarian cancer (Table 1).

Fig. 6
figure 6

Drug Sensitivity Analysis of Key Genes. A Relationship between PRKCD expression and drug sensitivity. B Relationship between UCP2 expression and drug sensitivity. C Drug-gene correlations derived from the GDSC database. D Drug-gene correlations derived from the CTRP database

Full size image
Table 1 Analysis of the association between aging markers and ovarian cancer
Full size table

Discussion

Ovarian cancer remains the leading cause of death among gynecologic malignancies and is predominantly diagnosed in postmenopausal women [39]. The absence of effective screening programs often results in late-stage detection and a high five-year mortality rate [40]. Although advancements in treatment options and therapeutic targets have been made, the overall cure rate remains disappointingly low [41, 42]. This underscores the urgent need to identify novel diagnostic markers and therapeutic strategies to improve outcomes for ovarian cancer patients.

Aging is a significant risk factor for ovarian cancer [9], and in this study, we observed a positive correlation between telomere length and ovarian cancer risk. Telomeres progressively shorten with age, but in advanced malignancies, the reactivation of telomerase stabilizes telomere length, enabling continuous cell division and granting cancer cells replicative immortality [43, 44]. Interestingly, excessively long telomeres have also been linked to increased cancer susceptibility [45]. These findings align with previous Mendelian randomization studies that reported an association between telomere length and ovarian cancer risk [46]. The development of anti-cancer therapies targeting telomeres further highlights their potential as a therapeutic avenue, suggesting that telomere modulation could represent a novel strategy in ovarian cancer treatment.

In this study, we identified PRKCD and UCP2 as key genes distinguishing early- and late-onset ovarian cancer. PRKCD, a member of the PKC kinase family, plays a critical role in regulating mitochondrial autophagy and is implicated in tumor development and progression through various pathways [47]. For instance, PRKCD mediates the MAPK pathway, contributing to breast cancer progression via phosphorylation [48]. Additionally, TRIM69 has been shown to suppress metastasis and resistance to anoikis in gastric cancer by promoting PRKCD degradation through the ubiquitin–proteasome system [49]. PRKCD, a key regulator of EMT, influences the malignant phenotype of tumor cells, thereby driving cancer progression. PRKCD encodes PKCδ, which enhances its expression, promoting tumor cell invasion and metastasis. Additionally, PRKCD can regulate c-Myc, facilitating EMT in CRC cells and contributing to tumor progression [50]. Meanwhile, PRKCD is predominantly localized in T cells, monocytes, and NK cells, where it plays a critical role in both adaptive and innate immunity by modulating signaling pathways, apoptosis, and inflammatory responses. Although PRKCD has been less extensively studied in ovarian cancer, existing evidence supports its role in promoting ovarian cancer progression, consistent with our findings [51].

Uncoupling protein 2 (UCP2) exhibits a dual role in cancer, with its effects varying depending on the context. In melanoma, UCP2 overexpression has been shown to restore oxidative phosphorylation balance and suppress malignant phenotypes [52]. Conversely, in acute myeloid leukemia (AML), elevated reactive oxygen species (ROS) can induce UCP2 expression, activate AMPK phosphorylation, and upregulate PFKFB3, promoting cancer cell proliferation [53]. UCP2 has been identified as a critical regulator of mitochondrial metabolic processes and the cellular response to oxidative stress. It has been shown that UCP2 modulates mitochondrial ATP synthesis while also influencing the production of ROS, which serve as key second messenger signals within cells [54]. Additionally, research indicates that inhibition of UCP2 within the deacetylase signaling pathway leads to an increase in ROS levels and promotes cell death. Moreover, the downregulation of UCP2 elevates ROS production and suppresses the proliferation of uterine leiomyosarcoma cells, suggesting that UCP2 may serve as a potential therapeutic target for this malignancy [55]. Furthermore, UCP2 expression is upregulated in various cancers and plays a pivotal role in the metabolic reprogramming of tumors [56]. In ovarian cancer, inhibiting or silencing UCP2, a key ROS-protective protein, can increase mitochondrial ROS levels, enhancing cisplatin-induced apoptosis [57]. While UCP2 appears to play a protective role in normal cells, its overexpression in cancer cells may contribute to chemotherapy resistance and improved survival by reducing ROS levels [58]. Our study identified PRKCD and UCP2 as being responsive to several drugs in ovarian cancer. Previous research has shown that PRKCD can modulate chemotherapy response in other tumors, either enhancing sensitivity or promoting resistance through diverse pathways [59, 60]. Similarly, UCP2's role appears context-dependent, varying with tumor type, metabolic state, and disease stage. Targeting UCP2, particularly in combination with conventional chemotherapy or immunotherapy, has the potential to regulate tumor progression and improve treatment efficacy [56]. These findings highlight the significant roles of PRKCD and UCP2 in ovarian cancer, both of which contribute to tumor progression through diverse pathways.

Currently, CA- 125 and HE4 are widely recognized biomarkers for the early detection of ovarian cancer, playing a crucial role in predicting disease progression and treatment response [61, 62]. In this study, PRKCD emerges as a potential prognostic marker, while UCP2 represents a promising therapeutic target. Their potential synergistic effects provide novel insights into personalized treatment strategies and drug resistance mechanisms in ovarian cancer. Future research should further explore their combined prognostic significance to refine prognostic assessment methods and advance targeted therapeutic approaches.

This study has certain limitations. We focused solely on the causal relationship between aging and ovarian cancer risk within a European population, relying on bioconviction analysis to identify differences in the pathogenesis of early- and late-onset ovarian cancer. Also relying on transcriptomic data alone, there are many confounding factors such as technical differences, etc. that cannot be avoided [63, 64]. Therefore, further functional experiments are needed to validate these findings, which we aim to address in future research. Also, we need multicenter studies to avoid bias from a single population.

In conclusion, this study establishes a causal link between aging and ovarian cancer and identifies PRKCD and UCP2 as tumor-specific, age-related genes through various machine learning approaches. Leveraging bioinformatics analysis, we comprehensively evaluated the roles of PRKCD and UCP2 in ovarian cancer, including their impact on immune landscape alterations. These findings offer valuable insights into the distinct mechanisms underlying early- and late-onset ovarian cancer. Furthermore, they highlight PRKCD and UCP2 as potential diagnostic markers and therapeutic targets, paving the way for personalized treatment strategies in ovarian cancer. Future research can leverage CRISPR technology to further elucidate the roles of PRKCD and UCP2 as critical targets in overcoming drug resistance in ovarian cancer [65].

Data availability

No datasets were generated or analysed during the current study.

References

  1. Siegel, R.L., Giaquinto, A.N. & Jemal, A. Cancer statistics, 2024. CA: A Cancer Journal for Clinicians 74, 12–49 (2024).

  2. Chen W, et al. Cancer statistics in China, 2015. CA Cancer J Clin. 2016;66:115–32.

    Article  PubMed  Google Scholar 

  3. Qi J, et al. National and subnational trends in cancer burden in China, 2005–20: an analysis of national mortality surveillance data. The Lancet Public Health. 2023;8:e943–55.

    Article  PubMed  Google Scholar 

  4. Sun H, et al. Aurora-A/SOX8/FOXK1 signaling axis promotes chemoresistance via suppression of cell senescence and induction of glucose metabolism in ovarian cancer organoids and cells. Theranostics. 2020;10:6928–45.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Sonkin D, Thomas A, Teicher BA. Cancer treatments: Past, present, and future. Cancer Genet. 2024;286–287:18–24.

    Article  PubMed  Google Scholar 

  6. Wu M, et al. Spatiotemporal transcriptomic changes of human ovarian aging and the regulatory role of FOXP1. Nature Aging. 2024;4:527–45.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Stankovic S, et al. Genetic links between ovarian ageing, cancer risk and de novo mutation rates. Nature. 2024;633:608–14.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  8. Broekmans FJ, Soules MR, Fauser BC. Ovarian Aging: Mechanisms and Clinical Consequences. Endocr Rev. 2009;30:465–93.

    Article  CAS  PubMed  Google Scholar 

  9. López-Otín C, Pietrocola F, Roiz-Valle D, Galluzzi L, Kroemer G. Meta-hallmarks of aging and cancer. Cell Metab. 2023;35:12–35.

    Article  PubMed  Google Scholar 

  10. Schmitt CA, Wang B, Demaria M. Senescence and cancer — role and therapeutic opportunities. Nat Rev Clin Oncol. 2022;19:619–36.

    Article  PubMed  PubMed Central  Google Scholar 

  11. Liu, H. & Weng, J. A comprehensive bioinformatic analysis of cyclin-dependent kinase 2 (CDK2) in glioma. Gene 822(2022).

  12. Liu, H., Weng, J., Huang, C.L.H. & Jackson, A.P. Is the voltage-gated sodium channel β3 subunit (SCN3B) a biomarker for glioma? Functional & Integrative Genomics 24(2024).

  13. Chhatwal, K.S. & Liu, H.J.b. RAD50 is a potential biomarker for breast cancer diagnosis and prognosis. bioRxiv, 2024.2009. 2007.611821 (2024).

  14. Liu H, Li YJCB. Potential roles of Cornichon Family AMPA Receptor Auxiliary Protein 4 (CNIH4) in head and neck squamous cell carcinoma. Cancer Biomark. 2022;35:439–50.

    Article  PubMed  Google Scholar 

  15. Li Y, Liu HJCB. Clinical powers of Aminoacyl tRNA Synthetase Complex Interacting Multifunctional Protein 1 (AIMP1) for head-neck squamous cell carcinoma. Cancer Biomark. 2022;34:359–74.

    Article  CAS  PubMed  Google Scholar 

  16. Chan JK, et al. Ovarian cancer in younger vs older women: a population-based analysis. Br J Cancer. 2006;95:1314–20.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Hou X, et al. Aging accelerates while multiparity delays tumorigenesis in mouse models of high-grade serous carcinoma. Gynecol Oncol. 2022;165:552–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Uruski, P., et al. Patient-Specific Variables Determine the Extent of Cellular Senescence Biomarkers in Ovarian Tumors In Vivo. Biomedicines 9(2021).

  19. Murray CJL. The Global Burden of Disease Study at 30 years. Nat Med. 2022;28:2019–26.

    Article  CAS  PubMed  Google Scholar 

  20. Sanderson E, et al. Mendelian randomization Nat Rev Methods Primers. 2022;2:6.

    Article  CAS  PubMed  Google Scholar 

  21. Skrivankova, V.W., et al. Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization. Jama 326(2021).

  22. Deo RC. Machine Learning in Medicine. Circulation. 2015;132:1920–30.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Greener JG, Kandathil SM, Moffat L, Jones DT. A guide to machine learning for biologists. Nat Rev Mol Cell Biol. 2021;23:40–55.

    Article  PubMed  Google Scholar 

  24. Swanson K, Wu E, Zhang A, Alizadeh AA, Zou J. From patterns to patients: Advances in clinical machine learning for cancer diagnosis, prognosis, and treatment. Cell. 2023;186:1772–91.

    Article  CAS  PubMed  Google Scholar 

  25. Rudd KE, et al. Global, regional, and national sepsis incidence and mortality, 1990–2017: analysis for the Global Burden of Disease Study. The Lancet. 2020;395:200–11.

    Article  Google Scholar 

  26. Roth GA, et al. Global, regional, and national age-sex-specific mortality for 282 causes of death in 195 countries and territories, 1980–2017: a systematic analysis for the Global Burden of Disease Study 2017. The Lancet. 2018;392:1736–88.

    Article  Google Scholar 

  27. Tsz-Lun Yeung, et al. ELF3 is a negative regulator of epithelial-mesenchymal transition in ovarian cancer cells. Oncotarget, 16951–16963 (2017).

  28. Xu J, et al. Single-Cell RNA Sequencing Reveals the Tissue Architecture in Human High-Grade Serous Ovarian Cancer. Clin Cancer Res. 2022;28:3590–602.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Burgess S, Small DS, Thompson SG. A review of instrumental variable estimators for Mendelian randomization. Stat Methods Med Res. 2015;26:2333–55.

    Article  PubMed  PubMed Central  Google Scholar 

  30. Liu, H., et al. Exploring the mechanism underlying hyperuricemia using comprehensive research on multi-omics. Scientific Reports 13(2023).

  31. Liu H. Association between sleep duration and depression: A Mendelian randomization analysis. J Affect Disord. 2023;335:152–4.

    Article  PubMed  Google Scholar 

  32. Verbanck M, Chen C-Y, Neale B, Do R. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet. 2018;50:693–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Subramanian A, et al. Gene set enrichment analysis a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102:15545–50.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Liu, H. & Tang, T.J.A.J.o.T.R. A bioinformatic study of IGFBPs in glioma regarding their diagnostic, prognostic, and therapeutic prediction value. American Journal of Translational Research 15, 2140 (2023).

  35. Liu, H., Dong, A., Rasteh, A.M., Wang, P. & Weng, J. Identification of the novel exhausted T cell CD8 + markers in breast cancer. Scientific Reports 14(2024).

  36. Liu, H. & Tang, T. MAPK signaling pathway-based glioma subtypes, machine-learning risk model, and key hub proteins identification. Scientific Reports 13(2023).

  37. Li T, et al. TIMER: A Web Server for Comprehensive Analysis of Tumor-Infiltrating Immune Cells. Can Res. 2017;77:e108–10.

    Article  CAS  Google Scholar 

  38. Newman AM, et al. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods. 2015;12:453–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Jayson GC, Kohn EC, Kitchener HC, Ledermann JA. Ovarian cancer. The Lancet. 2014;384:1376–88.

    Article  Google Scholar 

  40. Li Y-K, et al. Validation of ESM1 Related to Ovarian Cancer and the Biological Function and Prognostic Significance. Int J Biol Sci. 2023;19:258–80.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Asante D-B, Calapre L, Ziman M, Meniawy TM, Gray ES. Liquid biopsy in ovarian cancer using circulating tumor DNA and cells: Ready for prime time? Cancer Lett. 2020;468:59–71.

    Article  CAS  PubMed  Google Scholar 

  42. Lee JM, Minasian L, Kohn EC. New strategies in ovarian cancer treatment. Cancer. 2019;125:4623–9.

    Article  CAS  PubMed  Google Scholar 

  43. Shay JW. Role of Telomeres and Telomerase in Aging and Cancer. Cancer Discov. 2016;6:584–93.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. Gao J, Pickett HA. Targeting telomeres: advances in telomere maintenance mechanism-specific cancer therapies. Nat Rev Cancer. 2022;22:515–32.

    Article  CAS  PubMed  Google Scholar 

  45. Schmutz, I., et al. TINF2 is a haploinsufficient tumor suppressor that limits telomere length. eLife 9(2020).

  46. Went, M., et al. Phenome-wide Mendelian randomisation analysis of 378,142 cases reveals risk factors for eight common cancers. Nature Communications 15(2024).

  47. Munson, M.J., et al. GAK and PRKCD are positive regulators of PRKN-independent mitophagy. Nature Communications 12(2021).

  48. Deng, Y., et al. Superbinder based phosphoproteomic landscape revealed PRKCD_pY313 mediates the activation of Src and p38 MAPK to promote TNBC progression. Cell Communication and Signaling 22(2024).

  49. Sun L, et al. TRIM69 suppressed the anoikis resistance and metastasis of gastric cancer through ubiquitin-proteasome-mediated degradation of PRKCD. Oncogene. 2023;42:3619–32.

    Article  CAS  PubMed  Google Scholar 

  50. Jia, H.-t., et al. PKCδ promotes the invasion and migration of colorectal cancer through c-myc/NDRG1 pathway. Frontiers in Oncology 13(2023).

  51. Patel RS, Rupani R, Impreso S, Lui A, Patel NA. Role of alternatively spliced, pro-survival Protein Kinase C delta VIII (PKCδVIII) in ovarian cancer. FASEB BioAdvances. 2021;4:235–53.

    Article  PubMed  PubMed Central  Google Scholar 

  52. Esteves P, et al. Mitochondrial Retrograde Signaling Mediated by UCP2 Inhibits Cancer Cell Proliferation and Tumorigenesis. Can Res. 2014;74:3971–82.

    Article  CAS  Google Scholar 

  53. Robinson AJ, et al. Reactive Oxygen Species Drive Proliferation in Acute Myeloid Leukemia via the Glycolytic Regulator PFKFB3. Can Res. 2020;80:937–49.

    Article  CAS  Google Scholar 

  54. Toda C, Diano S. Mitochondrial UCP2 in the central regulation of metabolism. Best Pract Res Clin Endocrinol Metab. 2014;28:757–64.

    Article  CAS  PubMed  Google Scholar 

  55. Nagao, Y., et al. Novel therapeutic strategies targeting UCP2 in uterine leiomyosarcoma. Pharmacological Research 189(2023).

  56. Luby, A. & Alves-Guerra, M.-C. UCP2 as a Cancer Target through Energy Metabolism and Oxidative Stress Control. International Journal of Molecular Sciences 23(2022).

  57. Kleih, M., et al. Direct impact of cisplatin on mitochondria induces ROS production that dictates cell fate of ovarian cancer cells. Cell Death & Disease 10(2019).

  58. Pons DG, et al. UCP2 inhibition sensitizes breast cancer cells to therapeutic agents by increasing oxidative stress. Free Radical Biol Med. 2015;86:67–77.

    Article  CAS  Google Scholar 

  59. Liu, J., et al. CircRNA ITCH increases bortezomib sensitivity through regulating the miR-615–3p/PRKCD axis in multiple myeloma. Life Sciences 262(2020).

  60. Ke G, et al. MiR-181a confers resistance of cervical cancer to radiation therapy through targeting the pro-apoptotic PRKCD gene. Oncogene. 2012;32:3019–27.

    Article  PubMed  Google Scholar 

  61. Anastasi, E., et al. Recent Insight about HE4 Role in Ovarian Cancer Oncogenesis. International Journal of Molecular Sciences 24(2023).

  62. Tummala, M.K. & McGuire, W.P.J.J.o.C.O. Ovarian cancer, CA-125 addiction, and informed confusion: much ado about less. Vol. 25 3570–3571 (American Society of Clinical Oncology, 2007).

  63. Liu H, Li Y, Karsidag M, Tu T, Wang P. Technical and Biological Biases in Bulk Transcriptomic Data Mining for Cancer Research. J Cancer. 2025;16:34–43.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  64. Liu, H., Guo, Z. & Wang, P.J.G.R. Genetic expression in cancer research: Challenges and complexity. Gene reports, 102042 (2024).

  65. Liu, H. & Wang, P. CRISPR screening and cell line IC50 data reveal novel key genes for trametinib resistance. Clinical and Experimental Medicine 25(2024).

Download references

Acknowledgements

We are grateful to all the people who helped us accomplish this project.

Funding

No funding was received for this topic.

Author information

Author notes

  1. Yanting Shen, Jie Cheng and Qing Ding contributed equally to this work.

Authors and Affiliations

  1. Department of Traditional Chinese Medicine, Shenxin Community Health Service Center, Minhang District, Shanghai, China

    Yanting Shen

  2. Department of General Practice, Shenxin Community Health Service Center, Minhang District, Shanghai, China

    Jie Cheng

  3. Pharmacy Department, Shenxin Community Health Service Center, Minhang District, Shanghai, China

    Qing Ding

  4. Department of Oncology, Seventh People’s Hospital of Shanghai University of Traditional Chinese Medicine, Shanghai, China

    Zhihui Tao

  5. Shanghai University of Traditional Chinese Medicine, Shanghai, China

    Zhihui Tao

Authors
  1. Yanting Shen
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Jie Cheng
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Qing Ding
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Zhihui Tao
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Yanting Shen- original draft, Methodology, Formal analysis, Data curation. Jie Cheng—Validation, Methodology, Data curation. Qing Ding- Data curation, Investigation, Methodology. Zhihui Tao- Writing – review & editing, Supervision, Resources.

Corresponding author

Correspondence to Zhihui Tao.

Ethics declarations

Declarations

Ethical Approval and Consent to participate

Ethics approval and consent to participate do not apply to this manuscript.

Consent for publication

All authors of this manuscript have consented to publication.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Shen, Y., Cheng, J., Ding, Q. et al. Molecular characteristics of early- and late-onset ovarian cancer: insights from multidimensional evidence. J Ovarian Res 18, 83 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13048-025-01664-9

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13048-025-01664-9

Keywords

Journal of Ovarian Research

ISSN: 1757-2215

Contact us