Identification of Molecular Subtypes and Prognostic Features for Triple-Negative Breast Cancer Based on Golgi Apparatus-Related Gene Signature.

Abstract

Objectives: Triple-negative breast cancer (TNBC) presents a major treatment challenge due to its aggressive behavior. The dysfunction of the Golgi apparatus (GA) contributes to the development of various cancers. This study aimed to utilize GA-related genes (GARGs) to forecast the prognosis and immune profile of TNBC. Methods: The data were downloaded from The Cancer Genome Atlas (TCGA) database, including 175 TNBC and 99 healthy samples. The differentially expressed GARGs (DEGARGs) were analyzed using the TCGA biolinks package. The patients with TNBC were classified into two clusters utilizing the ConsensusClusterPlus package according to prognosis-related DEGARGs, followed by comparing the differences in prognosis and immune infiltration between the two clusters. Next, LASSO and stepwise Cox regression were applied to establish a GARGs signature to forecast the TNBC prognosis. The association of the GARGs signature with immune infiltrates and drug sensitivity was further explored. Results: In total, 430 DEGARGs were identified between TNBC and healthy samples, among which 20 were related to TNBC prognosis. Two GARG-related molecular clusters associated with different survival times and immune heterogeneity were identified. A risk model for TNBC was established based on six GARGs, and the high-risk (HR) group exhibited a poor prognosis. The HR group demonstrated a distinctly high M2 macrophage infiltration and low M1 macrophage infiltration, which contributed to an immunosuppressive tumor microenvironment and thus led to poor prognosis of the HR group. Immune dysfunction scores and programmed cell death ligand 1 (PD-L1) expression were substantially elevated in the HR group. The HR group showed increased sensitivity to anticancer drugs, such as cisplatin. Conclusion: Our findings suggest that GARGs are involved in the pathogenesis of TNBC and provide new insights into prognostic prediction. The identified clusters and GARGs signatures have the potential to guide individualized therapy.