Key Genes in Endometrial Hyperplasia & Cancer: Unlocking Prognostic Insights (2025)

Imagine a world where a woman's risk of developing a deadly cancer could be predicted years in advance, giving her a fighting chance. That world may be closer than you think. Endometrial cancer (EC), a frequent gynecological malignancy, is preceded by a precancerous condition called endometrial hyperplasia (EH), which is at least three times more common. The stakes are high: in China alone, a 2015 analysis revealed approximately 69,000 endometrial cancer cases, resulting in 16,000 deaths.

Both EH and EC manifest with irregular vaginal bleeding, particularly during perimenopause, often causing significant anxiety and disruption in women's lives. Common risk factors for both conditions include obesity, diabetes mellitus, prolonged estrogen or tamoxifen use without progesterone, and hereditary factors like Lynch syndrome. But here's where it gets controversial... While we know these risk factors, the precise genetic pathways that link them and drive the progression from EH to EC remain murky. This is a critical gap in our understanding, hindering the development of effective prevention and treatment strategies.

The World Health Organization (WHO) classifies EH into two types: endometrial hyperplasia without atypia and atypical endometrial hyperplasia (AEH). AEH carries a concerning 8% yearly risk of progressing to endometrial cancer. Diagnosis relies on endometrial histopathological examination, typically obtained through diagnostic curettage or hysteroscopic biopsy.

Growing evidence highlights a strong connection between type 2 diabetes and an increased risk of various cancers, including pancreatic, breast, endometrial, colorectal, prostate, and liver cancers. The culprits are believed to be hyperglycemia, hyperinsulinemia, insulin resistance, chronic inflammation, and obesity. Obesity, hypertension, and diabetes are even collectively known as the "triad of endometrial cancer". Epidemiological studies show that diabetic patients face a 2.12-fold higher risk of endometrial cancer compared to their non-diabetic counterparts. This risk escalates with body weight: overweight patients (BMI ≥ 25 kg/m²) have a 2.45-fold higher risk, while those with both obesity and hypertension face a staggering 3.5-fold increased risk.

Previous genomic studies utilizing TCGA and GEO datasets have shed light on molecular subtypes and prognostic signatures in EC. However, the shared genetic underpinnings connecting EH, diabetes, and EC progression remain largely unexplored. And this is the part most people miss... Understanding these connections is crucial for developing targeted prevention strategies and personalized treatments. Progesterone receptor (PGR) and vimentin (VIM) have been implicated in endometrial carcinogenesis. PGR helps maintain endometrial homeostasis, while VIM facilitates epithelial-mesenchymal transition (EMT), a process where cells gain migratory and invasive properties. Their specific roles in diabetes-associated EC warrant further investigation.

Imagine being able to identify women at the highest risk of developing endometrial cancer, years before the disease manifests. That's the promise of this research. This study aimed to identify genes commonly associated with both EH and type 2 diabetes, both significant risk factors for EC, to potentially prevent EC occurrence. Researchers identified overlapping genes in EH and EC samples from the GSE106191 dataset and type 2 diabetes-related genes from the GeneCards database. Lasso-Cox regression analysis was then used to pinpoint key genes. The study further investigated the expression and prognostic implications of these key genes within the TCGA-UCEC dataset. Finally, a nomogram was created to visualize a predictive model based on these hub genes and EC risk.

Materials and Methods

Data Collection and Workflow: Gene expression profiles for endometrial hyperplasia and endometrial cancer were obtained from the GSE106191 dataset in the GEO database (https://www.ncbi.nlm.nih.gov/geo/). Type 2 diabetes-related genes were retrieved from the GeneCards database (https://www.genecards.org/). Clinical and transcriptome data for endometrioid carcinoma (UCEC) patients were collected from the TCGA-UCEC project (https://portal.gdc.cancer.gov/). The overall workflow is illustrated in Figure 1.

Identifying Differentially Expressed Genes (DEGs) and Intersection Genes: Differential expression analysis between endometrial hyperplasia and carcinoma samples from the GSE106191 dataset was performed using the "limma" package in R, with significance thresholds set at FDR < 0.01 and |log₂FC| ≥ 1.5. The "Venn" package was then used to identify common genes between the DEGs and the type 2 diabetes-related gene set. Essentially, this step aimed to find genes whose activity levels were significantly different between normal, hyperplastic, and cancerous endometrial tissues, and which were also linked to type 2 diabetes.

Lasso-Cox Regression: Lasso-Cox regression analysis was performed using the "glmnet" package in R to identify genes significantly associated with overall survival in the TCGA-UCEC cohort. A risk score was calculated for each patient based on the expression levels of selected genes and their regression coefficients: risk score = Σ (gene expression × coefficient). Patients were stratified into high- and low-risk groups based on their risk scores. Survival differences between groups were analyzed using Kaplan-Meier curves ("survminer" package), and predictive accuracy was evaluated by time-dependent ROC analysis ("timeROC" package). Principal component analysis (PCA) and t-SNE were performed for dimensionality reduction and visualization. Both univariate and multivariate Cox regression analyses were conducted to identify independent prognostic factors, and a nomogram was constructed using the "rms" package. This sophisticated statistical technique helps identify the most important genes influencing survival, while minimizing the risk of overfitting the data.

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway Analysis: Gene ontology (GO) analysis involves functional enrichment analysis, encompassing biological process (BP), molecular function (MF), and cellular component (CC). KEGG analysis delves deeper into advanced functions and mechanisms within the biological system at the molecular level. The DAVID online tool (http://david.ncifcrf.gov/) was used to perform both analyses, with enrichment results downloaded and visualized using R software. A significance threshold of P <0.05 and FDR <0.25 was applied. These analyses help to understand what biological processes, cellular components, and molecular functions are affected by the identified genes.

Protein-Protein Interaction (PPI) Network: The PPI network for the intersecting genes was constructed using the STRING database (http://string-db.org/). Cytoscape v. 3.6.1 (http://www.cytoscape.org/) was employed for network analysis and visualization. Using Cytoscape’s built-in Network Analyzer tool, the degree value of each gene was calculated, with hub genes identified based on betweenness centrality (BC) values. This step identifies the key players (hub genes) in the network of interacting proteins.

Analysis of mRNA Expression Profile of Hub Gene Using UALCAN: UALCAN (http://ualcan.path.uab.edu (http://ualcan.path.uab.edu/) ) is an online resource designed for analyzing relative mRNA expression patterns of potential genes (TCGA and MET500 transcriptome sequencing) and their relationship with various tumor subtypes. UALCAN was utilized to investigate the mRNA expression profile of hub genes identified within TCGA-UCEC tissue samples and assess their potential association with clinical characteristics.

Tissue Collection and Immunohistochemistry (IHC): Paraffin-embedded tissues from normal endometrium (NE), endometrial hyperplasia (EH), EH with diabetes (EHD), endometrial carcinoma (EC), and EC with diabetes (ECD) (n = 5 per group) were subjected to antigen retrieval and immunostaining using antibodies against PGR (Abcam, ab184337, 1:200) and VIM (Proteintech, 60330-1-Ig, 1:500). Detection was performed with DAB chromogenic development. IHC is a technique used to visualize the location of specific proteins within tissue samples.

Statistical Analysis: Student’s t-test (R function t-test) was performed to determine statistically significant differences between the two groups. A P-value < 0.05 was considered significant. The ggplot package was used to generate the corresponding graphs for data visualization.

Results

Differential Gene Expression in Endometrial Hyperplasia, Carcinogenesis, and Type 2 Diabetes: DEGs in EH and EC were identified using the GSE106191 dataset from the GEO database. The analysis revealed 52 upregulated genes and 225 downregulated genes. A Venn diagram demonstrated 162 intersection genes between DEGs and type 2 diabetes-related genes.

Construction and Verification of the Stability of the Prognostic Model: Lasso-Cox regression analysis with the TCGA-UCEC dataset identified 162 genes that potentially influence patient prognosis based on the best value of log λ. A risk score was calculated based on the expression levels of 16 key genes (MMP9, TRPC6, VIM, PGR, NOG, CDH3, PXDN, PEG3, ASPM, KCTD15, NLGN1, KCNK6, RAPGEF4, GNLY, ID3, and PRDM6). Patients were then divided into high- and low-risk groups based on the median risk score. The overall survival (OS) and prognosis were significantly better in the low-risk group than in the high-risk group (P = 2.51e-7, HR(High group) = 3.475, 95% CI(2.16–5.58)). The AUC for the ROC curves indicated good performance in predicting 1-, 3-, and 5-year survival rates of 0.740, 0.749, and 0.783, respectively. This means the model has good accuracy in predicting patient survival based on the expression of these 16 genes.

Gene Ontology and Kyoto Encyclopedia of Genes and Genomes Pathway Analyses of Key Genes: The GO enrichment and KEGG pathway analyses revealed that the 16 key genes are involved in cell differentiation, cellular developmental process, regulation of multicellular organismal process, phagocytic vesicle, ion channel complex, transmembrane transporter complex, transporter complex, and dimerization activity. Enriched KEGG pathways included the TGF-beta signaling pathway, leukocyte transendothelial migration, estrogen signaling pathway, and cell adhesion molecules (CAMs).

PPI Network Identification and Prognostic Analysis of Hub Genes in EC: The protein interaction network identified PGR, VIM, MMP9, CDH3, and PXDN as hub genes. Prognostic analysis revealed that PGR and VIM exhibited significant associations with survival outcomes, while MMP9, CDH3, and PXDN showed no differential impact. Therefore, based on the PPI network and prognostic results, PGR and VIM were identified as the key hub genes.

Expressions of PGR and VIM with Clinical Characteristics and Immunohistochemistry of EC: PGR and VIM expression was assessed across TCGA-UCEC and GSE106191 datasets. Both genes were significantly downregulated in EC compared to normal tissues. Notably, while EH tissues showed expression levels comparable to normal endometrium, both PGR and VIM were markedly reduced in EC. Analysis using the UALCAN database explored correlations between PGR/VIM expression and key EC clinical characteristics (age, histological subtype, stage, TP53 mutation status, weight). Interestingly, when comparing tumor samples specifically, expression of both markers was significantly higher in EC with diabetes compared to EC without diabetes. Immunohistochemical staining confirmed these expression patterns in normal, EH, and EC tissues for both PGR and VIM.

Discussion

Endometrial hyperplasia and type 2 diabetes mellitus are well-established risk factors for endometrial cancer, a disease whose incidence has been rising globally. Epidemiological studies underscore this link, showing diabetic patients have a 2.12-fold increased risk of EC, which is further compounded by obesity and hypertension. The risk of progression from EH to EC is significantly higher in atypical cases (25–33%) than in non-atypical cases (1–3%), and diabetes further elevates the risk of development and lymph node metastasis. Although antidiabetic therapies show potential benefits in managing EC and EH, the precise molecular mechanisms connecting T2DM and EC malignant progression remain poorly defined. This gap highlights the necessity of identifying common genetic drivers to improve prognostic prediction and personalized treatment.

In this study, researchers integrated transcriptomic profiles of EH and EC (GSE106191) with T2DM-associated genes from GeneCards. Through a comprehensive bioinformatics approach—including Venn analysis, Lasso-Cox regression, and nomogram construction—they identified VIM and PGR as central hub genes influencing EC prognosis.

These findings on VIM align with its well-characterized role in promoting cell migration, invasion, and metastasis across various cancers. For instance, in colorectal cancer, VIM expression correlates with malignancy grade and metastatic potential and has even been proposed as a diagnostic biomarker in stool samples. Similarly, in breast cancer, VIM positivity aids in distinguishing malignant from benign tumors. Most notably, Martinez-Garcia et al (2022) recently identified VIM as a promising protein biomarker in cervical fluid for detecting early-stage EC, showing high specificity (81%) and sensitivity (78%). This study confirms that VIM is underexpressed in EC tissues and its expression level is a significant predictor of patient survival.

The role of PGR in endometrial homeostasis and carcinogenesis is equally critical. Long-term unopposed estrogen stimulation is a known driver of EH and EC, whereas progesterone-based therapy is the standard treatment for non-atypical hyperplasia, effectively reducing the risk of progression to malignancy. The receptor’s two main isoforms, PRA and PRB, mediate distinct transcriptional responses influencing cell proliferation and differentiation. Clinically, progesterone is a recommended endocrine therapy for recurrent EC per NCCN guidelines, particularly for patients with positive hormone receptor status. These results reinforce the crucial protective role of PGR, demonstrating its significant downregulation in EC and strong correlation with favorable clinical outcomes.

Notably, both VIM and PGR were significantly downregulated in EC compared to normal tissues. Their expression levels were correlated with key clinicopathological variables, including age, histological subtype, disease stage, TP53 mutation status, and body weight. Most importantly, it was established that their expression levels are independent predictive factors for overall survival. A prognostic model integrating these two molecular markers with clinical staging demonstrated improved predictive accuracy, highlighting their translational potential for risk stratification.

It is important to acknowledge that the IHC validation in this study was conducted on a relatively small sample size. While the results are consistent with bioinformatic predictions and show clear trends, future studies with larger, independent cohorts are warranted to confirm the generalizability of these protein-level findings.

Looking forward, the differential expression of PGR and VIM offers promising translational implications. The incorporation of these biomarkers into existing diagnostic algorithms could improve risk stratification for patients with EH and diabetes, potentially enabling earlier detection of malignant transformation. Furthermore, given the established role of progesterone therapy in EC management, the status of PGR expression could serve as a predictive biomarker for selecting patients who are most likely to benefit from endocrine-based interventions. From a therapeutic perspective, the association of VIM with aggressive tumor phenotypes positions it as a potential target for drugs aimed at inhibiting epithelial-mesenchymal transition and metastasis. Future studies validating these genes in larger prospective cohorts and developing standardized assays for clinical use will be essential steps toward translating these findings into practice.

In conclusion, VIM and PGR have been identified as key genes at the intersection of EH, diabetes, and EC. Their downregulation in tumor tissues and association with aggressive clinical features underscore their roles in EC progression. These findings provide novel insights into the molecular mechanisms linking metabolic disease and endometrial carcinogenesis and offer a promising biomarker signature for prognostic stratification in EC patients, which could guide more personalized treatment decisions in the future.

Conclusion

EH and DM are high-risk factors for EC. Prevention and treatment management of EC are equally important. VIM and PGR were identified as DEGs of EH, type 2 diabetes, and EC. These genes exhibited low expression in EC and played a role in influencing EC prognosis. When combined with clinical characteristics (stage), these genes contributed to an effective prognostic model for EC.

Data Sharing Statement

All data in this study can be obtained by GSE106191, TCGA-UCEC, and UALCAN databases.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Funding

This work was supported by Zhuzhou Central Hospital Commission (grant number: 202322) and Social Investment Project of Zhuzhou City in 2024 (No.: 2024-130).

Disclosure

All authors declare no competing interest in this work.

References