Zhang Peng, Li Cuicui, Li Fen, Wu Jiezhong, Hu Kunpeng, Huang He
Department of Thyroid and Breast Surgery, the Third Affiliated Hospital of Sun Yat-sen University, Guangzhou, Guangdong Province 510630, China.
Department of Nephrology, the Fifth Affiliated Hospital of Guangzhou Medical University, 621 Gangwan Road, Huangpu District, Guangzhou, Guangdong Province 510730, China.
Transl Oncol. 2025 Sep;59:102444. doi: 10.1016/j.tranon.2025.102444. Epub 2025 Jul 9.
Breast cancer remains one of the most prevalent malignancies globally, with metabolic reprogramming contributing significantly to tumor progression, immune evasion, and treatment resistance. Understanding the metabolic heterogeneity and its interaction with the tumor microenvironment is crucial for improving prognostic predictions and therapeutic strategies.
We integrated single-cell RNA sequencing (scRNA-seq), bulk RNA sequencing, and clinical data to characterize metabolic patterns in breast cancer. Immunoregulatory genes were obtained from the TISIDB database and analyzed by weighted gene co-expression network analysis (WGCNA) to identify key metabolic-related modules and hub genes. A metabolic risk signature was constructed using machine learning algorithms. Immune cell infiltration and immune checkpoint profiles were assessed to explore tumor microenvironment differences. Drug sensitivity prediction was performed via the OncoPredict tool. Functional assays investigated the oncogenic role of PDCD1 in breast cancer cell lines.
We identified distinct breast cancer epithelial subpopulations with highly activated glycolysis and associated metabolic pathways. Two patient clusters showed significant prognostic differences; the cluster with elevated glycolytic activity exhibited increased immune suppression, higher M2 macrophage infiltration, and poorer survival outcomes. The metabolic risk signature demonstrated robust prognostic power across multiple cohorts. High-risk patients displayed increased immune suppressive markers and reduced chemotherapy sensitivity. PDCD1 knockdown experiments confirmed its role in promoting proliferation, migration, and invasion of breast cancer cells.
Our study reveals metabolic heterogeneity linked to glycolytic reprogramming and immune modulation in breast cancer. The established metabolic signature offers a powerful prognostic tool and identifies potential therapeutic targets such as PDCD1. These findings contribute to precision oncology by guiding tailored treatment strategies based on metabolic and immune profiles.
乳腺癌仍然是全球最常见的恶性肿瘤之一，代谢重编程在肿瘤进展、免疫逃逸和治疗耐药性方面发挥着重要作用。了解代谢异质性及其与肿瘤微环境的相互作用对于改善预后预测和治疗策略至关重要。
我们整合了单细胞RNA测序（scRNA-seq）、批量RNA测序和临床数据，以表征乳腺癌中的代谢模式。从TISIDB数据库中获取免疫调节基因，并通过加权基因共表达网络分析（WGCNA）进行分析，以识别关键的代谢相关模块和枢纽基因。使用机器学习算法构建代谢风险特征。评估免疫细胞浸润和免疫检查点图谱，以探索肿瘤微环境差异。通过OncoPredict工具进行药物敏感性预测。功能实验研究了PDCD1在乳腺癌细胞系中的致癌作用。
我们鉴定出具有高度激活糖酵解和相关代谢途径的不同乳腺癌上皮亚群。两个患者簇显示出显著的预后差异；糖酵解活性升高的簇表现出免疫抑制增加、M2巨噬细胞浸润增加和较差的生存结果。代谢风险特征在多个队列中显示出强大的预后能力。高危患者显示免疫抑制标志物增加，化疗敏感性降低。PDCD1敲低实验证实了其在促进乳腺癌细胞增殖、迁移和侵袭中的作用。
我们的研究揭示了与乳腺癌中糖酵解重编程和免疫调节相关的代谢异质性。建立的代谢特征提供了一个强大的预后工具，并识别出潜在的治疗靶点，如PDCD1。这些发现通过基于代谢和免疫特征指导定制治疗策略，为精准肿瘤学做出了贡献。
