基于GEO数据库构建APP、APBA和APBB家族的胃癌预后评估模型  

The Establishment of Prognostic Model of APP,APBA and APBB Families for Gastric Cancer Based on GEO Database

作　　者：李文诗 俞思琦 王梦欧 吴歆姝 包博文 车晓芳[1] 郑春雷 LI WenShi;YU Siqi;WANG Meng′ou;WU Xinshu;BAO Bowen;CHE Xiaofang;ZHENG Chunlei(Department of Oncology,the First Hospital of China Medical University,Shenyang 110001 China;Department of Oncology Shanghai Electric Power Hospital,Shanghai 200050 China)

机构地区：[1]中国医科大学附属第一医院肿瘤内科,辽宁沈阳110001 [2]上海电力医院肿瘤科,上海200050

出　　处：《锦州医科大学学报》2024年第5期48-54,共7页Journal of Jinzhou Medical University

基　　金：国家自然科学基金面上项目,项目编号:32170791;国中康健集团科技项目,项目编号:GZKJ-KJXX-QTHT-20230425;上海市长宁区重点专科建设项目,项目编号:20232010。

摘　　要：目的基于淀粉样β前体蛋白(amyloid beta precursor protein,APP)家族、淀粉样β前体蛋白结合蛋白A(amyloid beta precursor protein binding family A,APBA)家族和淀粉样β前体蛋白结合蛋白B(amyloid beta precursor protein binding family B,APBB)家族构建胃癌预后评估模型。方法从基因表达综合(gene expression omnibus,GEO)数据库下载GSE62254胃癌数据集作为训练集,GSE15459作为验证集。利用Cox回归分析筛选APP家族、APBA家族和APBB家族中胃癌预后的独立危险因素;分别建立基于三家族独立预后因素的风险评分1(risk score 1,RS1)、RS1联合病理学参数的RS2、传统TNM分期的RS3;卡方检验分析RS1与胃癌患者临床病理特征的关系;利用单细胞在线分析网站,分析纳入RS1模型的基因在不同细胞亚群中的表达情况;利用CIBERSORT分析RS1对不同免疫细胞浸润的影响;利用基因集富集分析(gene set enrichment analysis,GSEA)进行通路富集分析。结果APLP2、APBB1、APBB2是胃癌患者预后的独立危险因素(P<0.05),基于三者的风险评分RS1高组患者生存期明显短于RS1低组患者。联合临床病理学参数的Cox回归分析显示,N分期、M分期、Lauren分型和RS1是胃癌患者预后的独立危险因素(P<0.05)。基于此构建的RS2(AUC=0.767)比仅基于T分期、N分期、M分期构建的RS3(AUC=0.719)预测准确率提高了4.8%。RS1和肿瘤T分期呈正相关(P<0.05),RS1高组CD4静息细胞浸润较高,激活细胞浸润较低,M2巨噬细胞浸润较高。GSEA通路分析显示,高RS1组患者富集于MAPK、MTOR和WNT等通路。结论本研究成功构建了基于APP、APBA和APBB家族的胃癌预后评估模型,该模型能够较准确地判断胃癌患者预后。Objective To establish prognostic model for gastric cancer based on the amyloid beta precursor protein(APP)family,amyloid beta precursor protein binding protein A(APBA)family and amyloid beta precursor protein binding protein B(APBB)family.Methods GSE62254 gastric cancer dataset was downloaded from GEO database as the training set and GSE15459 as the validation set.Cox regression analysis was used to screen the independent prognostic risk factors in APP family,APBA family and APBB family;the risk score 1(RS1)based on the independent prognostic factors of the three families,RS2 based on RS1 combined with pathological parameters,and RS3 based on the traditional TNM staging were established.Chi-square test was used to analyze the relationship between RS1 and clinicopathological features of gastric cancer patients;the expression of RS1 genes in different cell subsets was analyzed using the single cell online analysis website;the effect of RS1 on the infiltration of different immune cells was analyzed by CIBERSORT;Gene Set Enrichment Analysis(GSEA)was used for pathway enrichment analysis.Results By Cox regression analysis,APLP2,APBB1 and APBB2 were selected as the independent risk factors of gastric cancer(P<0.05).The overall survival(OS)time of patients with high-RS1 based on APLP2,APBB1 and APBB2 was significantly shorter than that of patients with low-RS1.The further Cox regression analysis including RS1 and clinical pathological parameters showed that N stage,M stage,Lauren type and RS1 were independent risk factors for prognosis of gastric cancer(P<0.05).The prediction accuracy of RS2(AUC=0.767)based on these parameters was 4.8%higher than that of RS3(AUC=0.719)based on T,N and M stages.RS1 was positively correlated with T stage of tumor(P<0.05).CD4 resting cell infiltration was higher,active cell infiltration was lower and M2 macrophage infiltration was higher in high RS1 group.GSEA pathway analysis showed that patients with high RS1 were enriched in MAPK,MTOR and WNT pathways.Conclusion This study has successfu

关 键 词：胃癌 APP APBA家族 APBB家族 风险评分 

分 类 号：R735.2[医药卫生—肿瘤]