Characterization of Exosome-Related Gene Risk Model to Evaluate the Tumor Immune Microenvironment and Predict Prognosis in Triple-Negative Breast Cancer

Pengjun Qiu, Qiao‐Nan Guo, Qingzhi Yao, Jianpeng Chen, Jianqing LinFrontiers in Immunology 2021 · DOI

Abstract

The results of our study suggest that an exosome-related risk model was related to the prognosis and ratio of immune cell infiltration in patients with Triple-Negative Breast Cancer (TNBC). This discovery may make significant contributions to improving immunotherapy strategies for TNBC by providing a robust tool for patient stratification.

Introduction

Triple-Negative Breast Cancer (TNBC) remains the most aggressive and challenging subtype of breast cancer to treat. Defined by the absence of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2), TNBC lacks the targeted therapies available to other subtypes. Consequently, patients often face higher recurrence rates and poorer overall survival.

Recent research has highlighted the critical role of exosomes—small extracellular vesicles—in cancer progression. Exosomes function as intercellular messengers, carrying proteins, lipids, and nucleic acids that can "reprogram" the tumor immune microenvironment (TIME). While the influence of exosomes is recognized, translating this biological process into a clinical prognostic tool has been difficult. This study aims to bridge this gap by developing a gene-based risk model that quantifies exosomal influence to predict survival and characterize the immune landscape in TNBC.

Methods

The researchers utilized a multi-step bioinformatics pipeline to derive a prognostic signature from genomic data.

Gene Selection and Model Construction

The study began by identifying 121 exosome-related genes from the ExoBCD database. These were cross-referenced with expression data from the Cancer Genome Atlas (TCGA) TNBC cohort ($n=123$) to identify genes relevant to this specific cancer subtype.

Venn diagram showing the intersection of exosome-related genes and TCGA expression data.

To refine this list into a predictive model, the authors employed Univariate Cox regression followed by LASSO (Least Absolute Shrinkage and Selection Operator) regression. LASSO is a sophisticated statistical method that prevents model "overfitting" by penalizing less significant variables, effectively narrowing the gene set to the most impactful predictors.

LASSO regression analysis (A: coefficient paths, B: cross-validation deviance) used to identify the optimal prognostic genes.

The final model identified a three-gene signature consisting of IVL, CXCL13, and AP2S1.

Immune and Drug Sensitivity Analysis

To understand the biological implications of the risk score, the authors used CIBERSORT and ESTIMATE algorithms to deconvolve bulk RNA data into estimates of immune cell infiltration. Furthermore, they correlated the risk score with drug sensitivity data from the CellMiner database, covering 163 FDA-approved drugs.

Results

Prognostic Accuracy

The three-gene risk model demonstrated high predictive power. In the TCGA training cohort, the Area Under the Curve (AUC) for 3-year survival was 0.869. These results were successfully validated using an independent GEO dataset (GSE58812, $n=105$), where low-risk patients consistently showed significantly better survival outcomes ($p < 0.01$).

The "Hot" vs. "Cold" Tumor Landscape

A key finding was the distinct immune profile of the risk groups. The low-risk group exhibited characteristics of "hot" tumors, featuring:

  • Significantly higher infiltration of $CD8^+$ T cells and M1 macrophages.
  • Higher overall Immune Scores and ESTIMATE Scores.
  • Increased expression of immune checkpoints, including PD-L1 (CD274) and CTLA4.

While higher checkpoint expression might seem counterintuitive for a "low-risk" group, it indicates an active immune response that the tumor is attempting to suppress. This suggests that low-risk patients may be the most suitable candidates for immune checkpoint inhibitor therapies.

Therapeutic Correlations

The risk score also correlated with sensitivity to specific chemotherapeutic agents. For instance, the model suggested varying sensitivity to drugs like Sunitinib and Pralatrexate, providing a potential framework for personalized treatment selection.

Discussion

This study establishes a novel, exosome-related three-gene signature that serves as a robust prognostic indicator for TNBC. The inclusion of CXCL13 is particularly noteworthy; its negative coefficient in the model aligns with its known role in recruiting B and T cells to form tertiary lymphoid structures, which are associated with better clinical outcomes.

Despite the strong statistical correlations, the study has limitations. As a retrospective analysis of public databases, it requires prospective clinical validation. Additionally, while the model identifies IVL, CXCL13, and AP2S1 as key markers, the specific molecular mechanisms by which these exosomal genes modulate the TIME remain to be explored through functional "wet lab" experiments.

In conclusion, this exosome-related risk model provides a valuable tool for predicting TNBC prognosis and understanding the underlying immune microenvironment, potentially guiding more effective, personalized immunotherapy strategies.

Summary updated

Abstract

The results of our study suggest that exosome-related risk model was related to the prognosis and ratio of immune cell infiltration in patients with TNBC. This discovery may make contributions to improve immunotherapy for TNBC.
BibTeX
@inproceedings{qiu2021characterization,
  author = {Pengjun Qiu and Qiao‐Nan Guo and Qingzhi Yao and Jianpeng Chen and Jianqing Lin},
  title = {{Characterization of Exosome-Related Gene Risk Model to Evaluate the Tumor Immune Microenvironment and Predict Prognosis in Triple-Negative Breast Cancer}},
  booktitle = {Frontiers in Immunology},
  year = {2021},
  doi = {10.3389/fimmu.2021.736030},
}
Figure