TL;DR
This study develops a novel four-mRNA prognostic signature for Cervical Squamous Cell Carcinoma and Endocervical Adenocarcinoma (CESC) by integrating a competing endogenous RNA (ceRNA) network. This signature demonstrates promising prognostic ability, offering a new perspective on CESC pathogenesis and potential for personalized medicine.
What Problem Does This Paper Address?
Cervical cancer, specifically CESC, remains a significant global health challenge with a substantial mortality rate, particularly for recurrent or metastatic disease. Current prognostic indicators, like the TNM staging system, are based on anatomical information and fail to capture the complex biological heterogeneity of CESC. This limitation hinders precise, individualized treatment strategies. The authors aim to address this by identifying novel biomarkers that can more accurately predict patient outcomes and provide insights into the underlying molecular mechanisms of CESC. The competing endogenous RNA (ceRNA) hypothesis, which describes how non-coding RNAs (lncRNAs and miRNAs) can regulate mRNA expression through competitive binding, offers a promising framework for understanding these complex regulatory networks.
Methods
The researchers employed a multi-step bioinformatics approach:
- Data Acquisition: RNA-sequencing (RNA-seq) data and clinical information for CESC patients were downloaded from The Cancer Genome Atlas (TCGA).
- Differential Expression Analysis: The
edgeRR package was used to identify differentially expressed mRNAs, lncRNAs, and miRNAs between CESC tumor samples and adjacent normal tissues. A stringent FDR-adjusted P-value < 0.05 and an absolute log2 fold change |log2FC| > 2 were used as criteria for statistical significance. - Survival Analysis: Univariate Cox regression analysis was performed using the
survivalR package to identify RNAs significantly associated with overall survival (OS). RNAs with a P-value < 0.05 were considered prognostic. - ceRNA Network Construction: A ceRNA network was built using the
multiMiRR package and the miRcode database to predict miRNA-mRNA and lncRNA-miRNA interactions. Cytoscape software was used for visualization. - Prognostic Signature Development: A subset of prognostic mRNAs (P < 0.001 in univariate analysis) were subjected to multivariate Cox regression analysis to identify key genes for a prognostic model.
- Functional Enrichment Analysis: Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis and Gene Ontology (GO) analyses were conducted using the
clusterProfilerandggplot2R packages to understand the biological functions of the identified prognostic RNAs. - Model Evaluation: A risk score (RS) model was developed based on the expression levels and regression coefficients of the selected prognostic mRNAs. Time-dependent Receiver Operating Characteristic (ROC) analysis, using the
pROCR package, was employed to evaluate the model's prognostic accuracy, with the Area Under the Curve (AUC) serving as a key metric.
Datasets
- The Cancer Genome Atlas (TCGA): This public database was the primary source for RNA-seq data and clinical information of CESC patients. The study utilized data from 240 CESC tumor samples and 3 adjacent non-tumor samples.
- Human Protein Atlas: This database was used to obtain information on protein expression levels of the key identified mRNAs through immunohistochemistry (IHC) and other immunoassay techniques.
The study did not introduce new benchmark datasets.
Key Insights and Ideas
- ceRNA Network as a Prognostic Framework: The study highlights the power of integrating ceRNA networks with prognostic signature development. By considering the complex interplay between mRNAs, lncRNAs, and miRNAs, the researchers could identify more robust prognostic markers.
- Beyond Single Genes: The paper emphasizes the limitations of relying on single gene biomarkers for cancer prognosis. The development of a multi-mRNA signature underscores the need to capture the collective impact of multiple molecular players.
- TGF-β and Cell Adhesion in CESC Prognosis: The enrichment analysis revealing the TGF-β signaling pathway and cell adhesion molecules as key pathways associated with prognostic RNAs provides crucial insights into the molecular mechanisms driving CESC progression and metastasis.
- Potential for Personalized Medicine: The developed four-mRNA signature demonstrates a significant ability to stratify patients into high-risk and low-risk groups, suggesting its potential utility in guiding personalized treatment decisions and improving patient outcomes.
Key Results
- Differentially Expressed RNAs: The study identified a substantial number of differentially expressed RNAs: 1398 downregulated and 1495 upregulated mRNAs, 5 downregulated and 15 upregulated miRNAs, and 493 downregulated and 658 upregulated lncRNAs.
- Survival-Associated RNAs: A total of 298 mRNAs, 8 miRNAs, and 129 lncRNAs were found to be significantly associated with overall survival in CESC patients.
- Four-mRNA Prognostic Signature: A prognostic signature was developed based on four key mRNAs: OPN3, DAAM2, HENMT1, and CAVIN3.
- Prognostic Model Performance: The developed prognostic signature achieved an AUC of 0.726 in ROC analysis, indicating a moderate to good ability to predict patient survival.
- Risk Stratification: Patients in the high-risk group, as defined by the signature, exhibited significantly worse overall survival (P < 0.0001).
- Clinical Parameter Validation: In multivariate Cox regression analysis, the signature's risk score remained a statistically significant independent prognostic factor (HR = 6.35, P = 0.01), even after adjusting for clinical parameters like T, N, and M stages.
Limitations and Future Work
- Experimental Validation: The authors acknowledge that the exact effects of the identified four mRNAs (OPN3, DAAM2, HENMT1, and CAVIN3) on CESC tumorigenesis have not been fully elucidated and require experimental verification.
- Independent Cohort Validation: The study relied on data from a single online database (TCGA). The authors suggest that an independent cohort is necessary to validate these findings and confirm the prognostic value of the signature.
- Further Mechanistic Studies: While the study identified enriched pathways, further experimental investigations are needed to fully elucidate the biological functions and mechanisms of the identified genes and their roles in CESC pathogenesis.
- Protein Expression Discrepancies: While transcript levels of OPN3 and HENMT1 were upregulated in CESC, their protein expression was higher in CESC samples. Conversely, DAAM2 and CAVIN3 transcript levels were decreased, but CAVIN3 protein expression was lower in CESC. These discrepancies highlight the complexity of gene regulation and the need for protein-level validation.
Summary updated