Identification of immune-related gene signature for predicting prognosis in uterine corpus endometrial carcinoma

Siyuan Song, Haoqing Gu, Jingzhan Li, Peipei Yang, Xiafei Qi, Jiatong Liu, Jiayu Zhou, Li Ye, Peng ShuScientific 2023 · DOI

TL;DR

This study identifies a novel immune-related gene signature that can predict prognosis in uterine corpus endometrial carcinoma (UCEC). This signature not only helps stratify patients into distinct immune microenvironment subtypes but also forms the basis for a prognostic model that could guide personalized immunotherapy strategies.

What Problem Does This Paper Address?

Uterine Corpus Endometrial Carcinoma (UCEC) is a significant gynecological malignancy, and while early detection is key, metastatic UCEC poses a serious threat. Current treatments often involve surgery or hormone therapy, which can be costly and lead to recurrence. Emerging research highlights the crucial role of the tumor immune microenvironment (TIME) in cancer development and patient prognosis. However, there's a need for better tools to understand and leverage this immune context for personalized treatment, particularly for immunotherapy. This paper aims to address this by developing an immune-related gene signature to predict UCEC prognosis and identify patients who might benefit most from immunotherapy.

Methods

The researchers employed a multi-faceted bioinformatics approach. They started by using consensus clustering to categorize UCEC samples into distinct immune subtypes, hypothesizing these clusters represent different immune states (inflammation, rejection, desert). Immune correlation algorithms were then used to analyze the tumor immune microenvironment (TIME) within these clusters. Gene Set Enrichment Analysis (GSEA) was performed to understand the biological functions associated with these clusters. A prognostic risk model was developed by integrating identified genes with clinical features, visualized using a Nomogram. Finally, in vitro experimental validation was conducted to confirm the prognostic value of key genes. Specific tools and packages mentioned include ConsensusClusterPlus, CIBERSORT, Limma, DAVID, STRING, Cytoscape, Maftools, GEPIA, TIMER, and WGCNA.

Datasets

  • Training Cohort: TCGA (The Cancer Genome Atlas) database for UCEC data and clinical information.
  • Validation Dataset: GEO database (GSE17025) for mRNA expression data and clinical features.

Key Insights and Ideas

  • Three Distinct Immune Subtypes: UCEC patients can be reliably classified into three distinct immune microenvironment subtypes (C1: inflammation, C2: rejection, C3: desert), each with different prognostic implications.
  • Cluster C1 as Immunotherapy Candidate: Cluster C1, characterized by a high immune status and expression of immune checkpoint genes, appears to be the most promising subtype for immunotherapy.
  • MAPK and Immune Checkpoint Pathways are Key: The identified hub genes are significantly enriched in pathways like MAPK signaling and PD-1/PD-L1 immune checkpoint, underscoring their role in UCEC's immune landscape.
  • Novel Prognostic Genes Identified: The study identified STAC and SRD5A1 as key prognostic genes, with STAC showing lower expression in tumors and correlating with poorer survival, while SRD5A1 expression was higher in tumors.
  • Prognostic Model Reflects TIME: The developed prognostic risk model not only predicts patient outcomes but also effectively reflects the underlying tumor immune microenvironment, suggesting a link between prognosis and immune status.

Key Results

  • The study successfully classified UCEC patients into three immune clusters (C1, C2, C3) with distinct survival outcomes, with C1 showing the best prognosis.
  • The prognostic risk model, built on STAC and SRD5A1, demonstrated a strong predictive ability, with a high AUC for predicting overall survival (OS) at 1, 3, and 5 years.
  • The high-risk group identified by the model exhibited a higher ESTIMATE score, Immune score, and Stromal score, along with increased infiltration of certain immune cells (e.g., CD8+ T cells, Tregs), suggesting a more inflamed but potentially less responsive TIME.
  • Experimental validation confirmed that STAC expression was significantly lower in UCEC cell lines compared to normal cells, while SRD5A1 was significantly higher, supporting their roles as prognostic markers.

Limitations and Future Work

  • In Vivo Validation Needed: While in vitro experiments were performed, the authors acknowledge the need for further in vivo experimental verification to fully validate the prognostic risk model.
  • Mechanism Elucidation: While key pathways and genes were identified, a deeper understanding of the precise mechanisms by which STAC and SRD5A1 influence UCEC progression and immune response is warranted.
  • Clinical Translation: The study provides a foundation for personalized immunotherapy, but further clinical trials are necessary to translate these findings into routine clinical practice. The authors suggest that the established prognostic risk signature could offer new possibilities for predicting immunotherapy efficacy and guiding personalized treatment for UCEC patients.

Summary updated

Abstract

The objective of this study is to develop a gene signature related to the immune system that can be used to create personalized immunotherapy for Uterine Corpus Endometrial Carcinoma (UCEC). To classify the UCEC samples into different immune clusters, we utilized consensus clustering analysis. Additionally, immune correlation algorithms were employed to investigate the tumor immune microenvironment (TIME) in diverse clusters. To explore the biological function, we conducted GSEA analysis. Next, we developed a Nomogram by integrating a prognostic model with clinical features. Finally, we performed experimental validation in vitro to verify our prognostic risk model. In our study, we classified UCEC patients into three clusters using consensus clustering. We hypothesized that cluster C1 represents the immune inflammation type, cluster C2 represents the immune rejection type, and cluster C3 represents the immune desert type. The hub genes identified in the training cohort were primarily enriched in the MAPK signaling pathway, as well as the PD-L1 expression and PD-1 checkpoint pathway in cancer, all of which are immune-related pathways. Cluster C1 may be a more suitable for immunotherapy. The prognostic risk model showed a strong predictive ability. Our constructed risk model demonstrated a high level of accuracy in predicting the prognosis of UCEC, while also effectively reflecting the state of TIME.
BibTeX
@inproceedings{song2023identification,
  author = {Siyuan Song and Haoqing Gu and Jingzhan Li and Peipei Yang and Xiafei Qi and Jiatong Liu and Jiayu Zhou and Li Ye and Peng Shu},
  title = {{Identification of immune-related gene signature for predicting prognosis in uterine corpus endometrial carcinoma}},
  booktitle = {Scientific},
  year = {2023},
  doi = {10.1038/s41598-023-35655-x},
}
Figure