Identification of prognostic cancer-associated fibroblast markers in luminal breast cancer using weighted gene co-expression network analysis

Abstract

Background: Cancer-associated fibroblasts (CAFs) play a pivotal role in cancer progression and are known to mediate endocrine and chemotherapy resistance through paracrine signaling. Additionally, they directly influence the expression and growth dependence of ER in Luminal breast cancer (LBC). This study aims to investigate stromal CAF-related factors and develop a CAF-related classifier to predict the prognosis and therapeutic outcomes in LBC.

Methods: The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases were utilized to obtain mRNA expression and clinical information from 694 and 101 LBC samples, respectively. CAF infiltrations were determined by estimating the proportion of immune and cancer cells (EPIC) method, while stromal scores were calculated using the Estimation of STromal and Immune cells in MAlignant Tumors using Expression data (ESTIMATE) algorithm. Weighted gene co-expression network analysis (WGCNA) was used to identify stromal CAF-related genes. A CAF risk signature was developed through univariate and least absolute shrinkage and selection operator method (LASSO) Cox regression model. The Spearman test was used to evaluate the correlation between CAF risk score, CAF markers, and CAF infiltrations estimated through EPIC, xCell, microenvironment cell populations-counter (MCP-counter), and Tumor Immune Dysfunction and Exclusion (TIDE) algorithms. The TIDE algorithm was further utilized to assess the response to immunotherapy. Additionally, Gene set enrichment analysis (GSEA) was applied to elucidate the molecular mechanisms underlying the findings.

Results: We constructed a 5-gene prognostic model consisting of RIN2, THBS1, IL1R1, RAB31, and COL11A1 for CAF. Using the median CAF risk score as the cutoff, we classified LBC patients into high- and low-CAF-risk groups and found that those in the high-risk group had a significantly worse prognosis. Spearman correlation analyses demonstrated a strong positive correlation between the CAF risk score and stromal and CAF infiltrations, with the five model genes showing positive correlations with CAF markers. In addition, the TIDE analysis revealed that high-CAF-risk patients were less likely to respond to immunotherapy. Gene set enrichment analysis (GSEA) identified significant enrichment of ECM receptor interaction, regulation of actin cytoskeleton, epithelial-mesenchymal transition (EMT), and TGF-β signaling pathway gene sets in the high-CAF-risk group patients.

Conclusion: The five-gene prognostic CAF signature presented in this study was not only reliable for predicting prognosis in LBC patients, but it was also effective in estimating clinical immunotherapy response. These findings have significant clinical implications, as the signature may guide tailored anti-CAF therapy in combination with immunotherapy for LBC patients.

Keywords: anti-CAF therapeutic approach; cancer-associated fibroblasts (CAFs); luminal breast cancer (LBC); prognostic CAF markers; weighted gene co-expression network analysis (WGCNA).