Defining three ferroptosis-based molecular subtypes and developing a prognostic risk model for high-grade serous ovarian cancer

Xiang Sun, Wenbin He, Baohua Lin, Weiming Huang, Danping YeAging 2024 · DOI

TL;DR

This study identifies three distinct molecular subtypes of high-grade serous ovarian cancer (OV) based on ferroptosis-related genes, revealing significant differences in their tumor microenvironment and prognosis. They also developed and validated a novel 8-gene prognostic risk model that accurately predicts patient outcomes and potential responses to therapy, offering new avenues for targeted treatment strategies.

What Problem Does This Paper Address?

Ovarian cancer (OV), particularly high-grade serous OV (HGSOC), remains a significant cause of cancer-related mortality, often diagnosed at late stages with limited effective treatment options. While ferroptosis, a regulated cell death pathway, shows promise as a biomarker, its role in the OV tumor microenvironment (TME) and its clinical predictive value are not fully understood. Existing prognostic models for OV often don't specifically leverage ferroptosis-related mechanisms, leaving a gap in personalized treatment strategies. This paper aims to address this by classifying OV based on ferroptosis signatures and developing a predictive model to guide clinical decision-making.

Methods

The researchers employed a multi-pronged approach:

  • Consensus Clustering: They used consensus clustering analysis on RNA-Seq data from the TCGA-OV cohort to classify HGSOC patients into distinct molecular subtypes based on the expression of ferroptosis-related genes.
  • Prognostic Model Development: A LASSO Cox regression model was used to identify key ferroptosis-related genes that predict prognosis. This was followed by stepwise multivariate regression to build a parsimonious 8-gene risk model.
  • Bioinformatic Analysis: Extensive analyses were conducted, including:
  • Univariate Cox Regression: To identify initial prognostic ferroptosis-related genes.
  • Principal Component Analysis (PCA): To visualize the separation of subtypes.
  • Kaplan-Meier (KM) Curves: To assess survival differences between subtypes and risk groups.
  • Gene Ontology (GO) and KEGG Pathway Enrichment Analysis: To understand the biological functions associated with different subtypes and risk groups.
  • ESTIMATE Algorithm and CIBERSORT: To analyze immune cell infiltration and TME characteristics.
  • Gene Set Enrichment Analysis (GSEA): To identify enriched pathways in high- and low-risk groups.
  • Receiver Operating Characteristic (ROC) Curves: To evaluate the predictive performance of the risk model.
  • In Vitro Experiments: The study included cell culture experiments to investigate the functional role of a specific gene (PDP1) in OV cell proliferation, invasion, and migration.

Datasets

  • TCGA-OV Cohort: Used as the primary training dataset for subtype classification and risk model development. This dataset comprised 361 primary high-grade serous OV samples.
  • GSE70769: An independent validation cohort of 85 OV samples.
  • ICGC-AU Cohort: Another independent validation cohort of 93 OV samples.

Key Insights and Ideas

  • Ferroptosis Defines Distinct OV Subtypes: The study successfully identified three distinct molecular subtypes (C1, C2, C3) based on ferroptosis gene expression, revealing that these subtypes are not merely statistical groupings but reflect significant biological differences.
  • Subtype C1: The "Bad Actor" Subtype: Subtype C1 emerged as the most aggressive, characterized by mesenchymal features, high infiltration of M2 macrophages, lower tumor purity, and enrichment of EMT-related pathways. This subtype is strongly associated with a poor prognosis.
  • A Robust Prognostic Predictor: The developed 8-gene risk model demonstrated strong predictive power across multiple cohorts, suggesting its potential for clinical utility in stratifying patients and guiding treatment decisions.
  • EMT as a Key Driver of Poor Prognosis: The research strongly links epithelial-mesenchymal transition (EMT) pathways to high-risk patients, highlighting its critical role in OV progression and resistance.
  • Immune Microenvironment Differences: Significant variations in immune cell infiltration and TME composition were observed between subtypes and risk groups, providing insights into immune evasion mechanisms and potential immunotherapeutic targets.
  • PDP1's Role in OV Progression: The experimental validation showed that PDP1 knockdown significantly reduced proliferation, invasion, and migration of OV cells, suggesting PDP1 as a potential therapeutic target.

Key Results

  • Three Ferroptosis Subtypes Identified: Consensus clustering identified three distinct subtypes (C1, C2, C3). Subtype C3 showed the most favorable prognosis, while C1 had the poorest.
  • 8-Gene Risk Model Developed: A risk model comprising PDP1, FCGBP, EPHA4, GAS1, SLC7A11, BLOC1S1, SPOCK2, and CXCL9 was constructed.
  • Strong Predictive Performance: The risk model achieved high AUC values (e.g., 0.74 for 5-year survival in TCGA) and demonstrated robustness in validation cohorts (GSE70769 and ICGC-AU).
  • High-Risk Association with EMT: GSEA analysis revealed enrichment of EMT-related pathways (e.g., Wnt, TGF-beta) in high-risk patients.
  • Immune Cell Infiltration Differences: High-risk patients showed increased Macrophages_M2 and enrichment of angiogenesis and CAF pathways, while low-risk patients had higher ImmuneScore and more T cells (follicular Tregs, CD8, CD4 memory activated).
  • Chemotherapy Sensitivity: Low-risk patients showed higher sensitivity to Cisplatin, Gemcitabine, and Gefitinib.
  • PDP1 Knockdown Effects: In vitro experiments showed that si-PDP1 treatment significantly reduced proliferation (p < 0.001), migration, and invasion of OV cells.

Limitations and Future Work

The authors acknowledge several limitations:

  • Retrospective Data: The study relied on retrospective data from public databases. A prospective study with larger sample sizes is needed for future validation.
  • Mechanism Validation: The underlying mechanisms of the identified ferroptosis-related genes in HGSOC require further validation through in vitro and in vivo experimental studies.
  • Clinical Application: The long-term clinical applicability and reliability of the developed risk model need to be verified through clinical trials.

Future work could focus on translating these findings into clinical practice by further validating the risk model in diverse patient populations and exploring the therapeutic potential of targeting ferroptosis pathways and specific genes like PDP1 in HGSOC.

Summary updated

Abstract

Our results could improve the understanding of ferroptosis in OV, providing promising insights for the clinical targeted therapy for the cancer.
BibTeX
@inproceedings{sun2024defining,
  author = {Xiang Sun and Wenbin He and Baohua Lin and Weiming Huang and Danping Ye},
  title = {{Defining three ferroptosis-based molecular subtypes and developing a prognostic risk model for high-grade serous ovarian cancer}},
  booktitle = {Aging},
  year = {2024},
  doi = {10.18632/aging.205857},
}
Figure