基于RNA结合蛋白构建宫颈癌预后模型并实验验证
Construction and validation of cervical cancer prognosis model based on RNA-binding proteins
张哲 1吴杰 2伍婷婷 3胡成彪 4宋韶晖 1马文龙 1董训虎4
作者信息
- 1. 解放军69330部队卫生分队,新疆哈密 839000
- 2. 解放军92730部队医院,海南 三亚 572000
- 3. 陆军军医大学西南医院妇产科,重庆 400038
- 4. 解放军73091部队卫生分队,江苏镇江 212424
- 折叠
摘要
目的 构建宫颈癌(CC)中核心RNA结合蛋白(RBPs)相关预后模型并初步筛选预后标志物.方法 比较基因型-组织表达数据库(GTEx)和癌症基因组图谱(TCGA)中1 542个RBPs在CC中的表达水平,筛选得到差异表达RBPs[DERBPs,|Log2(FC)|＞1且FDR＜0.05];将TCGA队列按照2:1随机分为训练队列和测试队列,在训练队列中,通过单因素、最小绝对收缩选择算子和多元逐步Cox回归构建CC预后模型,并在测试队列中进行验证;最后,建立列线图并基于模型筛选核心基因体外功能实验以评价和验证模型效能.结果 从348个DERBPs中筛选出9个核心RBPs并构建模型.生存分析发现,高风险组CC患者的总生存率显著低于低风险组(P＜0.001),ROC曲线(AUC＞0.8)提示该模型具有中等预测能力,该模型可作为CC患者独立的预后因子(HR=1.152,95％CI:1.106～1.200,P＜0.001);以上结果均在测试队列中得到验证.在TCGA队列中构建基于9个RBPs的列线图,校准曲线显示其具有较好的准确性.最后,体外实验证实沉默CC细胞RBM38表达后,细胞增殖、迁移及侵袭能力显著增强(P＜0.05),与模型预测结果一致.结论 本研究成功构建由9个RBPs组成的CC预后模型,可作为CC独立的预后标志物并具有中等预测能力,该结果可能为CC患者提供潜在的治疗靶点和预后标志物.
Abstract
Objective To construct a core RNA-binding proteins(RBPs)related prognosis model in cervical cancer(CC)and preliminarily screen the prognostic biomarkers.Methods The expression levels of 1 542 RBPs in CC were compared in Genotype-Tissue Expression and The Cancer Genome Atlas(TCGA)to screen the differential expressed RBPs(DERBPs,log2FC＞1 and FDR＜0.05).The TCGA cohort was randomly divided into training cohort and testing cohort according to the ratio of 2:1.The CC prognosis model was constructed by univariate Cox regression,least absolute shrinkage and selection operator and stepwise multivariate Cox regression in the training cohort,and verified in the testing cohort.Finally,a nomogram was established and hub gene was screened based on the model in vitro to evaluate and verify the efficacy of the model.Results Nine core RBPs were screened out from 348 DERBPs and the model was constructed.Survival analysis showed that the overall survival rate of CC patients in the high-risk group was significantly lower than that in the low-risk group(P＜0.001),and ROC curves(AUC＞0.8)indicated that the model had moderate predictive ability,which could be used as an independent prognostic factor for CC patients(HR=1.152,95％CI:1.106-1.200,P＜0.001).All the above results were verified in the testing cohort.A nomogram based on 9 RBPs was constructed in the TCGA cohort,and the calibration curves showed that it had good accuracy.Finally,in vitro experiments confirmed that after silencing the expression of RBM38 in CC cells,the cell proliferation,migration and invasion ability were significantly enhanced(P＜0.05),which was consistent with the predicted results.Conclusion In this study,a CC prognostic model composed of 9 RBPs is successfully constructed,which can be used as an independent prognostic biomarker of CC and has moderate predictive ability.The results may provide potential therapeutic targets and prognostic biomarkers for CC patients.
关键词
宫颈癌/生物信息学/癌症基因组图谱/预后模型Key words
cervical cancer/bioinformatics/The Cancer Genome Atlas/prognostic model引用本文复制引用
基金项目
国家自然科学基金青年科学基金(81803279)
出版年
2024
国家科技期刊平台
NSTL
万方数据