Zheng Jing, Zhang Yi-Wen, Pan Zong-Fu
Department of Pharmacy,Zhejiang Medical & Health Group Hangzhou Hospital,Hangzhou 310022,China.
Department of Pharmacy,Zhejiang Provincial People's Hospital,People's Hospital of Hangzhou Medical College,Hangzhou 310014,China.
Zhongguo Yi Xue Ke Xue Yuan Xue Bao. 2021 Oct;43(5):685-695. doi: 10.3881/j.issn.1000-503X.13271.
Objective To study the stemness characteristics of uterine corpus endometrial carcinoma(UCEC)and its potential regulatory mechanism.Methods Transcriptome sequencing data of UCEC was obtained from The Cancer Genome Atlas.Gene expression profile was normalized by edgeR package in R3.5.1.A one-class logistic regression machine learning algorithm was employed to calculated the mRNA stemness index(mRNAsi)of each UCEC sample.Then,the prognostic significance of mRNAsi and candidate genes was evaluated by survminer and survival packages.The high-frequency sub-pathways mining approach(HiFreSP)was used to identify the prognosis-related sub-pathways enriched with differentially expressed genes(DEGs).Subsequently,a gene co-expression network was constructed using WGCNA package,and the key gene modules were analyzed.The clusterProfiler package was adopted to the function annotation of the modules highly correlated with mRNAsi.Finally,the Human Protein Atlas(HPA)was retrieved for immunohistochemical validation.Results The mRNAsi of UCEC samples was significantly higher than that of normal tissues(=25.095,<0.001),and the lower degree of differentiation corresponded to higher mRNAsi in tumor tissues.The mRNAsi of UCEC increased gradually with tumor staging.The prognostic analysis showed that high mRNAsi was correlated with short overall survival in patients with UCEC(=6.864,=0.0088).There were 570 DEGs between the high-and low-mRNAsi groups.By using the HiFreSP algorithm,we identified that the oocyte meiosis sub-pathway(Oocyte meiosis_1)and cell cycle sub-pathway(Cell cycle_3)had significant prognostic significance.These pathways contained 11 DEGs(MAD2L1,CAMK2A,PTTG1,PLK1,CCNE1,CCNE2,ESPL1,CDC20,CCNB1,CCNB2,and SMC1B),which were significantly associated with the prognosis of UCEC patients.Gene co-expression network showed that mRNAsi,as well as MAD2L1,CAMK2A,and PTTG1,was associated with three gene modules.The immunohistochemical analysis demonstrated that MAD2L1 and PTTG1 showed up-regulated expression while CAMK2A showed down-regulated expression in UCEC,which was consistent with the results of transcriptome sequencing.Conclusions On the basis of machine learning,this study characterizes the stemness characteristics of UCEC.We identify the key sub-pathways related to prognosis and demonstrate that MAD2L1,CAMK2A,PTTG1 are closely related to the stemness of UCEC,which provides insight into the regulatory mechanism of cancer stemness and reveals the potential therapeutic targets of UCEC.
目的 研究子宫内膜癌（UCEC）的干性特征及其潜在调控机制。方法 从癌症基因组图谱获取UCEC的转录组测序数据。基因表达谱通过R3.5.1中的edgeR软件包进行标准化。采用单类逻辑回归机器学习算法计算每个UCEC样本的mRNA干性指数（mRNAsi）。然后，通过survminer和survival软件包评估mRNAsi及候选基因的预后意义。使用高频子通路挖掘方法（HiFreSP）识别富含差异表达基因（DEG）的预后相关子通路。随后，使用WGCNA软件包构建基因共表达网络，并分析关键基因模块。采用clusterProfiler软件包对与mRNAsi高度相关的模块进行功能注释。最后，检索人类蛋白质图谱（HPA）进行免疫组化验证。结果 UCEC样本的mRNAsi显著高于正常组织（ = 25.095， < 0.001），肿瘤组织中分化程度越低，mRNAsi越高。UCEC的mRNAsi随肿瘤分期逐渐升高。预后分析表明，高mRNAsi与UCEC患者的总生存期缩短相关（ = 6.864， = 0.0088）。高、低mRNAsi组之间有570个DEG。通过HiFreSP算法，我们确定卵母细胞减数分裂子通路（Oocyte meiosis_1）和细胞周期子通路（Cell cycle_3）具有显著的预后意义。这些通路包含11个DEG（MAD2L1、CAMK2A、PTTG1、PLK1、CCNE1、CCNE2、ESPL1、CDC20、CCNB1、CCNB2和SMC1B），它们与UCEC患者的预后显著相关。基因共表达网络显示，mRNAsi以及MAD2L1、CAMK2A和PTTG1与三个基因模块相关。免疫组化分析表明，UCEC中MAD2L1和PTTG1表达上调，而CAMK2A表达下调，这与转录组测序结果一致。结论 本研究基于机器学习对UCEC的干性特征进行了表征。我们识别了与预后相关的关键子通路，并证明MAD2L1、CAMK2A、PTTG1与UCEC的干性密切相关，这为癌症干性的调控机制提供了见解，并揭示了UCEC的潜在治疗靶点。
