Li Yong, Chen Chunmei, Li Weiwen, Shao Mingtao, Dong Yan, Zhang Qunchen
Department of Breast, Jiangmen Central Hospital Jiangmen, Guangdong, P. R. China.
Department of General Surgery, The First Affiliated Hospital of Jinan University Guangzhou, Guangdong, P. R. China.
Am J Clin Exp Immunol. 2025 Apr 25;14(2):45-67. doi: 10.62347/XPCM9169. eCollection 2025.
This study aimed to utilize single-cell RNA sequencing (scRNA-seq) to elucidate the autophagic landscape in breast cancer and to develop a prognostic model for breast cancer patients based on traditional high-throughput RNA sequencing (bulk RNA-seq).
We analyzed scRNA-seq data from the GSE75688 dataset to explore the expression patterns of autophagy-related genes (ARGs) across distinct cellular clusters. ARGs were retrieved from the GeneCards database, and bulk RNA-seq data were obtained from The Cancer Genome Atlas (TCGA). Cox proportional hazards regression was employed to construct a prognostic risk model based on ARGs. Patients were subsequently stratified into high-risk and low-risk groups according to their risk scores. For external validation, we used gene expression data from the GSE20685 and GSE48390 datasets. Receiver operating characteristic (ROC) curve analysis was performed to evaluate the performance of the 3-gene signature.
Using the FindClusters function in Seurat, all cells were grouped into four distinct clusters, highlighting the intratumoral heterogeneity within the samples. Significant differences in autophagy scores were observed among the clusters. Fifteen differentially expressed autophagy-related genes were identified, and a prognostic signature consisting of three autophagy-related genes - FEZ1, STX11, and ADAMTSL1 - was developed. Based on this model, patients were classified into high- and low-risk groups, with a statistically significant difference in survival between the two groups (log-rank test, = 0.0011). The model demonstrated robust predictive performance with an AUC of 0.761 in the external validation dataset. A nomogram incorporating the 3-gene signature and clinical factors showed strong prognostic discrimination.
This study uncovered significant variation in autophagy levels among different breast cancer cell clusters. Furthermore, we established a novel 3-gene autophagy-related prognostic model that effectively stratifies patient risk and provides a potential tool for personalized prognosis in breast cancer.
本研究旨在利用单细胞RNA测序（scRNA-seq）阐明乳腺癌中的自噬格局，并基于传统的高通量RNA测序（批量RNA-seq）为乳腺癌患者开发一种预后模型。
我们分析了来自GSE75688数据集的scRNA-seq数据，以探索自噬相关基因（ARG）在不同细胞簇中的表达模式。从GeneCards数据库中检索ARG，并从癌症基因组图谱（TCGA）获得批量RNA-seq数据。采用Cox比例风险回归基于ARG构建预后风险模型。随后根据患者的风险评分将其分为高风险和低风险组。为了进行外部验证，我们使用了来自GSE20685和GSE48390数据集的基因表达数据。进行了受试者工作特征（ROC）曲线分析以评估三基因特征的性能。
使用Seurat中的FindClusters函数，将所有细胞分为四个不同的簇，突出了样本中的肿瘤内异质性。在各簇之间观察到自噬评分存在显著差异。鉴定出15个差异表达的自噬相关基因，并开发了一种由三个自噬相关基因——FEZ1、STX11和ADAMTSL1组成的预后特征。基于该模型，将患者分为高风险和低风险组，两组之间的生存率存在统计学显著差异（对数秩检验，=0.0011）。该模型在外部验证数据集中显示出强大的预测性能，AUC为0.761。包含三基因特征和临床因素的列线图显示出强大的预后判别能力。
本研究揭示了不同乳腺癌细胞簇之间自噬水平的显著差异。此外，我们建立了一种新的三基因自噬相关预后模型，该模型有效地对患者风险进行分层，并为乳腺癌的个性化预后提供了一种潜在工具。
