摘要
背景同源重组缺陷（HRD）与乳腺癌的发生、发展和预后有重要关系。然而，原发性乳腺癌的 HRD 阳性率仅为 24%。确定与 HRD 相关的标志物对于开发新型乳腺癌治疗方法至关重要。癌基因 ERCC6L 在乳腺癌中的作用尚不明确，其与放疗的相互作用也有待探索，因此有必要进行进一步研究以明确其作用。 方法 我们利用 WGCNA，利用 TCGA 乳腺癌的公开 HRD 评分和遗传数据，结合临床特征，找出与 HRD 评分相关的基因。随后，我们采用多种机器学习方法筛选相关基因。通过随机森林和 stepCox 方法筛选出最终的四个基因，并在 TCGA、GSE96058 和 METABRIC 数据集中验证了它们的性能。接下来，我们使用ssGSEA、GSVA、CIBERSORT、ESTIMATE和单细胞分析等方法评估了肿瘤免疫微环境。最后，我们通过免疫荧光、流式细胞术、平板克隆和 Western blot 验证了 ERCC6L 的下调增加了 DNA 损伤并提高了辐射敏感性。 结果 通过机器学习建立了一个名为 HRAS 的预后模型，该模型由四个基因（ERCC6L、UBE2T、TPX2 和 SLC7A5）组成。在独立数据集中，该指标对乳腺癌患者的预后以及免疫疗法和放射疗法的疗效具有出色的预测性能。HRAS评分高的乳腺癌患者TMB和干性更高，免疫检查点表达增加，免疫细胞浸润减少，在接受免疫治疗和放射治疗时预后较差。实验验证表明，敲除 ERCC6L 会明显增加 DNA 损伤、增强细胞凋亡并诱导细胞周期停滞，从而使细胞对放疗敏感。 结论 HRD 相关特征对多个数据集的预后以及乳腺癌患者的免疫疗法和放疗疗效具有很强的预测能力。此外，综合指标反映了免疫微环境的特征，可作为预测乳腺癌患者预后和临床治疗效果的新型标记物。我们的实验首次阐明了ERCC6L在增强辐射诱导的DNA损伤中的作用，为旨在使癌细胞对放疗敏感的策略提供了一个新的靶点。
Background
Homologous recombination deficiency (HRD) exhibits significant associations with the occurrence, progression, and prognosis of breast cancer. However, the primary breast cancer HRD positivity rate is merely 24%. The identification of markers associated with HRD is crucial for the development of novel therapeutic approaches for breast cancer. The role of the oncogene ERCC6L in breast cancer remains unclear, and its interaction with radiotherapy has yet to be explored, necessitating further investigation for clarification.
Methods
We employed WGCNA to identify genes associated with the HRD score, utilizing public HRD score and genetic data from TCGA breast cancer, with their clinical characteristics. Subsequently, we employed various machine learning methods to filter relevant genes. The final four genes were obtained through random forest and stepCox, and their performance was validated in TCGA, GSE96058, and METABRIC datasets. Next, we assessed the tumor immune microenvironment using methods such as ssGSEA, GSVA, CIBERSORT, ESTIMATE, and single-cell analysis. Finally, we validated the downregulation of ERCC6L, increasing DNA damage and enhancing radiation sensitivity, through immune fluorescence, flow cytometry, plate cloning, and western blot.
Results
A prognostic model named HRAS was established through machine learning, consisting of four genes (ERCC6L, UBE2T, TPX2, and SLC7A5). The indicator exhibited excellent predictive performance on the prognosis and the efficacy of immunotherapy and radiotherapy of breast cancer patients in independent datasets. Breast cancer patients with high HRAS scores showed higher TMB and stemness, increased expression of immune checkpoints, reduced immune cell infiltration, and poorer prognosis in the context of immunotherapy and radiotherapy. Experimental validation demonstrated that knockdown of ERCC6L markedly elevated DNA damage, enhanced apoptosis, and induced cell cycle arrest in response to radiation therapy, thereby sensitizing cells to radiation.
Conclusion
The HRD-related signatures displayed strong predictive capabilities for the prognosis in multiple datasets and the efficacy of immunotherapy and radiotherapy of breast cancer patients. Moreover, the composite indicator reflected the immune microenvironment characteristics and could be novel markers for predicting the prognosis and clinical treatment outcomes in breast cancer patients. Our experiments first elucidated the role of ERCC6L in enhancing radiation-induced DNA damage, presenting a novel target for strategies aimed at sensitizing cancer cells to radiotherapy.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
