Department of Urology, China Medical University, The First Hospital of China Medical University, Shenyang, Liaoning, China.
Funct Integr Genomics. 2022 Oct;22(5):797-811. doi: 10.1007/s10142-022-00884-2. Epub 2022 Jul 28.
Although bladder cancer (BLCA) is the 10th most common tumor worldwide, particularly practical markers and prognostic models that might guide therapy are needed. We used a non-negative matrix factorization algorithm to classify PI3K pathway-related genes into molecular subtypes. A weighted gene co-expression network analysis (WGCNA) was generated to identify co-expression modules. Univariate Cox regression, least absolute shrinkage sum selection operator-Cox regression, and multivariate Cox regression were utilized to develop a prognostic score model. Kaplan-Meier analysis and receiver operating characteristics were utilized to measure the model's effectiveness. A nomogram was constructed to improve the predictive ability of the model based on clinical parameters and risk. Decision curve analysis (DCA) was used to evaluate the nomogram. To evaluate the immune microenvironment, an estimate algorithm was used. Drug sensitivity was identified using the R package "pRRophetic." UM-UC-3 cell line was used to measure the effect of CDK6 in Western blotting, proliferation assay, and 5-ethynyl-20-deoxyuridine assay. Based on PI3K pathway-related genes, The Cancer Genome Atlas (TCGA)-BLCA and GSE32894 patients were divided into two subtypes. Twenty-five co-expression modules were established using the WGCNA algorithm. A seven-gene signature (CDK6, EGFR, IGF1, ITGB7, PDGFRA, RPS6, and VWF) demonstrated robustness in TCGA and GSE32894 datasets. Expression levels of CDK6 and risk positively correlated with M2 macrophages and IgG. Cisplatin, gemcitabine, methotrexate, mitomycin C, paclitaxel, and vinblastine are sensitive to different groups based on the expression of CDK6 and risk. Functional experiments suggested that CDK6 promotes the proliferation of UM-UC-3 cells. We constructed a seven-gene prognostic signature as an effective marker to predict the outcomes of BLCA patients and guide individual treatment.
尽管膀胱癌 (BLCA) 是全球第 10 常见的肿瘤，但仍需要特别实用的标志物和预后模型来指导治疗。我们使用非负矩阵分解算法将 PI3K 通路相关基因分类为分子亚型。生成加权基因共表达网络分析 (WGCNA) 以识别共表达模块。单因素 Cox 回归、最小绝对收缩和选择算子-Cox 回归以及多因素 Cox 回归用于开发预后评分模型。Kaplan-Meier 分析和接收者操作特征用于衡量模型的有效性。基于临床参数和风险构建列线图以提高模型的预测能力。决策曲线分析 (DCA) 用于评估列线图。使用估计算法评估免疫微环境。使用 R 包 "pRRophetic" 确定药物敏感性。使用 UM-UC-3 细胞系在 Western blot、增殖测定和 5-乙炔基-20-脱氧尿苷测定中测量 CDK6 的作用。根据 PI3K 通路相关基因，将 TCGA-BLCA 和 GSE32894 患者分为两种亚型。使用 WGCNA 算法建立了 25 个共表达模块。CDK6、EGFR、IGF1、ITGB7、PDGFRA、RPS6 和 VWF 七个基因的特征在 TCGA 和 GSE32894 数据集中具有稳健性。CDK6 的表达水平和风险与 M2 巨噬细胞和 IgG 呈正相关。根据 CDK6 和风险的表达，顺铂、吉西他滨、甲氨蝶呤、丝裂霉素 C、紫杉醇和长春碱对不同组敏感。功能实验表明，CDK6 促进了 UM-UC-3 细胞的增殖。我们构建了一个七基因预后签名作为有效标志物，以预测 BLCA 患者的结局并指导个体化治疗。
