中文摘要: |
︿
<p>
目的:旨在构建内质网应激基因相关风险模型预测宫颈癌患者的生存和分析不同风险组之间免疫原 性和免疫治疗的差异,为宫颈癌预后判断和精准治疗提供临床参考价值。 方法: 自 TCGA、GEO、 GTEx 数据库获取宫颈癌和正常宫颈的测序数据和相应的临床信息,Gene Card 数据库下载内质网应激相关基因,应用 R 软件行差异化、交集分析,确定内质网应激 (Endoplasmic reticulum stress, ERS)差异基因。采用 Cox 回归分析构建预后风险模型,根据模型 计算风险评分,绘制风险曲线、Kaplan-Meier(KM)生存曲线,应用受试者工作特征(receiver operating characteristic, ROC)曲线评估模型的效能,应用 Wilcoxon 秩和检验和卡方检验对高低 风险组临床特征分析,利用 GSE52903 数据集验证模型的适用性。采用单样本富集分析 (ssGSEA)确定与风险评分相关的 KEGG 通路,自肿瘤免疫组图谱(The Cancer Immunome Atlas,TCIA)获得 TCGA-CESE 免疫表型评分,分析免疫原性的差异,应用 TIDE 评分和 SubMap 算法预测宫颈癌患者免疫治疗的敏感性。应用 GSE52903 和 GSE7410 数据集验证 6 个 预后生物标志物的表达水平。用免疫组化法验证预后生物标志物在宫颈癌和癌旁组织的表达。 结果: 1. 本研究筛选出 42 个内质网应激相关共表达差异基因,单因素 Cox 分析得到 10 个 (DES,JUN,PLOD2,SLC2A1,TFRC,IL1B,SPP1,CXCL8,GJB2,DSG2)与预后相关的 ERS 相关基 因,多因素 Cox 分析确定 6 个 ERS 相关基因(DES,PLOD2, SPP1,CXCL8,GJB2,DSG2)构建最优 预后模型。K-M 生存曲线示高风险组生存率更低(P<0.05),ROC 曲线分析示 1 年 AUC 为 0.757,3 年 AUC 为 0.759,5 年 AUC 为 0.778。 2. 在 GSE52903 验证集验证风险模型,K-M 生存曲线分析示高风险组生存率更低(P=0.0039), ROC 曲线分析示 1 年 AUC 为 0.816,3 年 AUC 为 0.657,5 年 AUC 为 0.634。 3. 临床特征相关分析示临床病理 T 分期、FIGO-stage 在高低风险组差异显著(P <0.05),单因素 Cox 分析示 Pathologic-T,FIGO-stage 和 Risk score 为独立预后因素(P <0.05),多因素 Cox 分析示 Risk score 为显著的独立预后因素(P <0.05)。 4. ssGSEA 分析得到与风险评分正相关的 11KEGG 通路 (r>0.35),TNF 通路,IL-17 通路,HIF-1 通路,内质网中蛋白质加工通路、细胞衰老通路与免疫原性相关。 5. 低风险组 IPS-CTLA4、IPS-PD1-PD-L1-PD-L2、IPS- PD1-PD-L1-PD-L2-CTLA4 显著高于高风 险组(p<0.05),低风险组有更强的免疫原性表型。TIDE 评分分析免疫检测点抑制剂治疗敏感 性,高风险组(49/137, 35.8%)较低风险组(38/136, 27.9%)敏感(P =0.029),SubMap 预测免疫治疗 的响应性,低分险组对 PD-1 疗法敏感(Nominal P<0.05 ),高风险组对 CTLA4 疗法敏感 (Nominal P<0.05 ) III 6. 验证 TCGA-CESE、GSE52903 和 GSE7410 数据集中 6 个预后风险因素标志物的表达,得到 DES、SPP1、PLOD2 在宫颈癌和正常组织中表达有差异,趋势一致。 7. 免疫组化染色法验证 ERS 相关基因 PLOD2、SPP1、DSG2、GJB2、DES、CLCX8 在宫颈癌组 织和癌旁组织的表达存在显著差异。PLOD2、SPP1、DSG2、GJB2、DES、CXCL8 在宫颈癌 组织中阳性表达(P<0.05),DES 在宫颈癌组织中阴性表达(P<0.001)。 结论: 1. 本研究通过生物信息学方法构建了基于内质网应激相关基因预测宫颈癌预后的风险模型,发现 6 个与预后相关的基因,PLOD2、SPP1、DSG2、GJB2、CXCL8 为危险因素, DES 为保护性 因素。 2. 风险模型经其他数据集进行验证后,证实模型有效,能预测宫颈癌的预后,预测效能优于其他 临床特征,风险评分为宫颈癌预后的独立预后因素。 3. ERS 预后特征能预测宫颈癌患者的免疫原性和免疫治疗,风险基因也需要更多的功能分析来探 索其可能的免疫治疗机制和临床价值。</p>
﹀
|
外文摘要: |
︿
<p>
Objective: To construct an endoplasmic reticulum stress-related signature risk model to predict the survival of cervical cancer patients and analyze the differences in immunogenicity and immunotherapy between different risk groups to provide clinical reference value for prognosis determination and precise treatment of cervical cancer. Methods: RNA-Seq and corresponding clinical information for cervical cancer and normal cervix were obtained from TCGA, GEO, and GTEx databases, and endoplasmic reticulum stress-related genes were downloaded from Gene Card database, and R software was applied to differential and intersection analysis to identify ERS differential genes. The prognostic risk model was constructed using Cox regression analysis, risk scores were calculated based on the model, Kaplan-Meier (KM) survival curves, and receiver operating characteristic (ROC) curves were plotted to assess the efficacy of the model, and Wilcoxon test and chi-square test were applied to the clinical characteristics of the high and low risk groups were analyzed, and the applicability of the model was validated using the GSE5290 dataset. Single sample Gene Set Enrichment Analysis (ssGSEA) was used to identify KEGG pathways associated with risk scores, TCGA-CESE immunophenoscores (IPS) were obtained from The Cancer Immunome Atlas (TCIA), differences in immunogenicity were analyzed, and the TIDE score and SubMap algorithm were applied to predict the sensitivity of immunotherapy in cervical cancer patients. The expression levels of six biomarkers were validated using the GSE52903 and GSE7410 datasets. The expression of prognostic biomarkers in cervical cancer and normal cervical tissue was verified by Immunohistochemical staining. Results: 1. In this study, 42 endoplasmic reticulum stress-related co-expression differential genes were screened, and 10 (DES, JUN, PLOD2, SLC2A1, TFRC, IL1B, SPP1, CXCL8, GJB2, DSG2) prognosis-related ERS genes were obtained by single-factor Cox analysis, and 6 ERS genes were identified by multi-factor Cox (DES, PLOD2, SPP1, CXCL8, GJB2, DSG2) to construct the optimal prognostic model. The K-M survival curves analysis showed lower survival in the high-risk group (p<0.05), and the ROC curve validated the model, and 1-year AUC was 0.757, and 3-year AUC was 0.759, and 5-year AUC was 0.778. 2. The risk model was verified in the GSE52903 validation set. K-M survival curve analysis showed that the high-risk group had a lower survival (P=0.0039). ROC curve analysis showed that the1-year AUC V was 0.816, 3-year AUC was 0.657, and 5-year AUC was 0.634. 3. Correlation analysis of clinical characteristics showed significant differences in clinicopathologic T-stage and FIGO-stage in high and low risk groups (P<0.05), Univariate analysis showed that Pathologic-T, FIGO-stage and Risk score were independent prognostic factors (p<0.05), and multi-variate Cox analysis showed that Risk score was a significant independent prognostic (p<0.05), and multivariate Cox analysis showed that Risk score was a significant independent prognostic factor (p<0.05). 4. The ssGSEA analysis yielded positive correlations with risk scores for 11KEGG pathway (r > 0.35), TNF pathway, IL-17 signaling pathway, HIF-1 pathway, protein processing in the endoplasmic reticulum pathway, and cellular senescence pathway associated with immunogenicity. IPS-CTLA4, IPS-PD1-PD-L1-PD-L2, and IPS- PD1-PD-L1-PD-L2-CTLA4 were significantly higher in the low-risk group than in the high-risk group (P<0.05), and the low-risk group had a stronger immunogenic phenotype. The TIDE analysis of immune checkpoint inhibition treatment sensitivity was more sensitive (P=0.029) in the high-risk group (49/137, 35.8%) than in the low-risk group (38/136, 27.9%). 5. SubMap predicted responsiveness to immunotherapy, with the low-risk group sensitive to PD-1 therapy (Nominal P <0.05) and the high-risk group sensitive to CTLA4 therapy (Nominal P<0.05). 6. Validation of the expression of six prognostic risk factor markers in the TCGA-CESE, GSE52903, and GSE7410 datasets yielded differential expression of DES, SPP1, and PLOD2 in cervical cancer and normal tissues, with consistent trends. 7. Immunohistochemical staining confirmed that there were significant differences in the expression of ERS-related genes PLOD2, SPP1, DSG2, GJB2, CXCL8, and DES in cervical cancer tissues and adjacent cancerous tissues. PLOD2, SPP1, DSG2, GJB2, DES, and CXCL8 were positive expression in cancer tissues (P < 0.05), and DES was negative expression in cancer tissues (P < 0.001). Conclusion: 1. In this study, a prognostic risk model of cervical cancer was established based on endoplasmic reticulum stress-related genes by bioinformatics methods. Six prognostic genes were found, including PLOD2, SPP1, DSG2, GJB2, CXCL8 as risk factors and DES as protective factors. 2. The risk model was validated by other data sets and confirmed that the model was valid and could predict the prognosis of cervical cancer with better predictive efficacy than other clinical characteristics, and the risk score was an independent prognostic factor for the prognosis of cervical cancer. 3. ERS prognostic features predict immunogenicity and immunotherapy in patients with cervical cancer, and more functional analysis of risk genes is needed to explore their possible immunotherapeutic mechanisms and clinical value.</p>
﹀
|