卵巢癌乳酸化修饰亚型的鉴定及预后评分模型构建与免疫疗效预测  

Identification of lactylation-related subtypes,construction of prognostic scoring model and immunotherapy prediction in ovarian cancer

作　　者：林子丹 周琛斐 黄舒婷 巢锦瑜 何善阳 LIN Zidan;ZHOU Chenfei;HUANG Shuting;CHAO Jinyu;HE Shanyang(Guangdong Cardiovascular Institute,Guangzhou 510080,China;Department of Gynecology,Guangdong Provincial People’s Hospital,Guangzhou 510080,China;Guangdong Academy of Medical Sciences,Guangzhou 510080,China;Department of Gynecology,the Fifth Affiliated Hospital of Southern Medical University,Guangzhou 510920,China)

机构地区：[1]广东省心血管病研究所,广东广州510080 [2]广东省人民医院妇科,广东广州510080 [3]广东省医学科学院,广东广州510080 [4]南方医科大学第五附属医院妇科,广东广州510920

出　　处：《新医学》2025年第1期3-19,共17页Journal of New Medicine

基　　金：国家自然科学基金(82272850);广东省自然科学基金(2024A1515010596);国家卫生健康委医药卫生科技发展研究中心基金(WKZX2023CX130001)。

摘　　要：目的 在卵巢癌中鉴定与乳酸化修饰相关的分子分型,构建预后评分模型,并预测免疫治疗效果。方法基于癌症基因组图谱-卵巢癌(TCGA-OV)数据集及基因表达综合数据库(GEO)的GSE63885、GSE26193数据集,对乳酸化修饰相关基因进行预后和生存分析。采用无监督聚类方法将卵巢癌患者分为4种乳酸化修饰分型(LRGCluster),并对其差异基因进行基因本体论(GO)和京都基因与基因组百科全书(KEGG)富集分析。通过单因素Cox回归分析(P <0.000 1),筛选出预后相关基因(PRGs),并通过实时荧光定量聚合酶链反应(RT-qPCR)验证其在正常卵巢组织、早期及晚期卵巢癌组织中的表达情况。基于这些基因,进一步通过二次无监督聚类将患者分为2种基因分型(geneCluster),并构建预后评分系统LactyScore。同时,依据LactyScore分层,对样本进行免疫细胞浸润分析、免疫治疗预测及药物敏感性分析。结果 鉴定出4种LRGCluster和2种geneCluster,并筛选出5个与卵巢癌预后密切相关的差异表达基因:胶原蛋白ⅩⅥ型α1链(COL16A1)、家族转录抑制因子SPEN、含AT钩DNA结合基序1(AHDC1)、亮氨酸拉链蛋白1(LUZP1)和基质细胞衍生因子2样1(SDF2L1)。RT-qPCR结果显示,SPEN、COL16A1、AHDC1、LUZP1可能为预后的危险因素,而SDF2L1可能为预后的保护因素。基于这5个基因构建的LactyScore预后评分系统显示,高LactyScore组患者的生存率高于低LactyScore组。高LactyScore组患者具有较高的免疫逃逸潜力和较低的免疫治疗应答率。结论 COL16A1、SPEN、AHDC1、LUZP1和SDF2L1为与卵巢癌预后密切相关的乳酸化修饰基因亚型,这些基因具有作为卵巢癌生物标志物的潜力。基于这5个基因构建的LactyScore预后评分系统可有效预测卵巢癌患者的预后,并为不同患者分层提供免疫治疗效果预测依据。Objective To identify lactylation-related molecular subtypes,construct a prognostic model,and predict immunotherapy efficacy in ovarian cancer(OC).Methods The prognostic significance of lactylation-related genes(LRGs)was analyzed using data from TCGA-OV,GSE63885,and GSE26193 datasets.Unsupervised clustering identified four distinct lactylation-related clusters(LRGClusters).Differential expression and Gene Ontology(GO)and Kyoto Encyclopedia of Genes and Genomes(KEGG)enrichment analyses were performed for these clusters.Using univariate Cox regression analysis(P<0.0001),five prognostic-related genes(PRGs)were identified.The expression levels of these PRGs in normal ovarian tissues,as well as early and advanced-stage ovarian cancer tissues were validated via RT-qPCR.Based on the five PRGs,a second round of unsupervised clustering was conducted to identify two gene clusters(geneClusters),and a prognostic scoring system,termed LactyScore,was developed.Immune cell infiltration,immunotherapy response,and drug sensitivity analyses were then performed based on LactyScore stratification.Results Four LRGClusters and two geneClusters were identified.Five differentially expressed genes including COL16A1,SPEN,AHDC1,LUZP1,and SDF2L1 were significantly associated with prognosis of OC patients.RT-qPCR indicated that SPEN,COL16A1,AHDC1,and LUZP1 were the potential risk factors for poor prognosis,whereas SDF2L1 might serve as a protective factor.Based on these PRGs,the LactyScore prognostic scoring system was established.Survival analysis revealed that patients in the high LactyScore group exhibited significantly better overall survival compared to those in the low LactyScore group.Moreover,patients with a high LactyScore showed increased immune evasion potential and lower response rates to immunotherapy.Conclusions Five prognostic genes including COL16A1,SPEN,AHDC1,LUZP1,and SDF2L1 are associated with OC and these genes demonstrate their potential as biomarkers for OC.Furthermore,the development of the robust LactyScore system off

关 键 词：卵巢癌 乳酸化修饰 预后评分模型 免疫治疗 

分 类 号：R737.31[医药卫生—肿瘤]