中文摘要

第一部分 铁死亡相关mRNA在膀胱癌中的潜在预后价值研究

目的

铁死亡是不同于传统凋亡和坏死的一种新型细胞死亡方式，其在多种恶性肿瘤发生机制中发挥着重要的调控作用。然而，铁死亡这一新型机制与膀胱癌的潜在关系至今仍未明确。本研究拟初步探索铁死亡相关mRNA是否在膀胱癌中发挥潜在预后价值，并就研究中发现的重要靶点做进一步体外验证，以期探索靶向诱导铁死亡可否为继化疗、免疫治疗后的膀胱癌新型治疗方式提供新的方向。

方法

本研究首先借助当前高通量二代测序技术的优势，对膀胱癌大样本转录组测序数据进行了系统性分析和外部体系验证。根据膀胱肿瘤组织和正常膀胱组织间显著差异表达的铁死亡相关基因和单因素回归分析中筛选出的预后基因，利用最小绝对收缩和选择运算符（least absolute shrinkage and selection operator cox regression, LASSO）回归方法构建出铁死亡相关mRNA最佳预后模型。研究进一步通过Kaplan-Meier分析、单因素cox回归分析、多因素cox回归分析等方法验证铁死亡预后模型的生存预后价值。通过基因功能注释富集分析（Gene ontology, GO）、京都基因和基因组百科全书（Kyoto Encyclopedia of Genes and Genomes, KEGG）富集分析、单样本基因集富集分析（Single-sample Gene Set Enrichment Analysis, ssGSEA）探索铁死亡相关mRNA在膀胱癌中所涉及的潜在生物学基础机制。研究进一步筛选出关键铁死亡相关mRNA靶点，培养并传代正常膀胱永生化上皮细胞（SV-HUC-1）和常见膀胱癌细胞系（HT-1376, BIU-87, T24），进行实时荧光定量聚合酶链式反应（Quantitative Real-time Polymerase Chain Reaction, RT-qPCR）和免疫印迹实验（Western Blot, WB），以验证关键铁死亡相关mRNA在膀胱癌细胞和正常上皮细胞之间的表达差异。研究进一步借助CCK细胞增殖实验、克隆形成实验、细胞划痕愈合实验等方法验证了其潜在细胞学功能，并通过透射电子显微镜识别验证膀胱癌细胞线粒体铁死亡形态。同时收集北京医院膀胱癌临床样本进行免疫组织化学染色，以验证关键铁死亡相关基因在膀胱癌临床组织中的表达情况及其与膀胱癌临床分期的潜在关联。最后，研究运用ESTIMATE肿瘤纯度分析和CIBERSORT免疫浸润分析等方法进一步探索了目标靶点与膀胱癌免疫微环境的潜在关系。

结果

本研究首先识别出33种铁死亡相关mRNA在膀胱癌组织和正常膀胱组织间呈显著差异表达。（p < 0.05）单因素cox回归分析发现仅8种铁死亡相关mRNA的表达水平在膀胱癌样本中具备显著预后价值。（p < 0.05）借助交叉预后基因，LASSO回归分析成功构建出仅包含四种铁死亡相关mRNA（FADS2, HMGCR, FANCD2, ALOX5）的最佳模型。利用各铁死亡相关mRNA表达水平和回归系数将所有膀胱癌样本进行危险评分量化。Kaplan–Meier生存分析发现，无论在分析组队列还是外部验证组队列，较高的铁死亡相关mRNA危险评分均预示着显著较差的生存预后。（p < 0.05）多因素cox回归分析表明，无论在分析组还是外部验证组，铁死亡相关mRNA危险评分均是膀胱癌总体生存期（Overall Survival, OS）的独立预后因素。（p < 0.05）研究亦表明铁死亡相关mRNA危险评分与膀胱癌分子亚型分类、免疫细胞浸润水平呈现显著相关性。qPCR和WB分析表明，模型中关键铁死亡相关mRNA靶点-FADS2，在多种膀胱癌细胞系中的表达水平显著高于正常膀胱细胞。（p < 0.05）CCK8细胞增殖实验、克隆形成实验、划痕愈合实验等结果表明，靶向敲低FADS2表达后，膀胱癌细胞的增殖和迁移能力均受到显著影响。膀胱癌细胞进一步经过固定、脱水、包埋、切片、染色处理，借助透射电子显微镜优势，可观察到敲低组膀胱癌细胞的线粒体体积明显缩小，线粒体膜密度显著增高。此外，收集我院临床膀胱样本组织行免疫组化实验，结果表明，FADS2在膀胱癌组织中较正常膀胱组织显著高表达，且其表达量与膀胱癌临床分期呈现显著相关性。（p <0.01）本研究进一步借助FADS2单基因表达量和多种膀胱癌临床元素成功构建出膀胱癌临床预测模型-列线表，借助该列线表模型可高效、准确地预测出膀胱癌患者1年、3年和5年总体生存概率。最后，ESTIMATE和CIBERSORT等结果表明，FADS2可能通过肿瘤免疫微环境相关机制影响膀胱癌的发生和发展。

结论

铁死亡作为一种新型细胞死亡方式，其相关mRNA在膀胱肿瘤中显示出重要的潜在预后价值。靶向诱导铁死亡可为膀胱癌未来治疗手段提供新的思路和研究方向。

第二部分 铁死亡相关lncRNA在膀胱癌中的潜在预后价值研究

目的

铁死亡相关lncRNA在膀胱癌中的作用目前仍未知，本研究旨在第一部分基础上深入探索铁死亡相关lncRNA在膀胱癌中的潜在预后价值。

方法

本研究首先从癌症基因组图谱（the cancer genome atlas, TCGA）数据库中系统性下载整理所有膀胱癌样本的转录组测序数据以及临床信息数据。借助第一部分研究中整理出来的铁死亡相关mRNA数据矩阵，借助皮尔逊相关分析（|R| > 0.5, p < 0.001）识别出膀胱癌铁死亡相关lncRNA矩阵。将所有膀胱癌样本随机分成分析队列和验证队列，借助LASSO回归方法在分析队列中尝试建立铁死亡相关lncRNA最佳预后模型，并在验证队列中对其预后价值进行验证。本研究针对模型中回归系数值最高的关键靶点-AC006160.1，进行了多种细胞系体外实验验证，包括SV-HUC-1细胞、 BIU-87细胞、HT-1376细胞、 T24细胞、 RT4细胞、RT-112细胞、5637细胞和UMUC3细胞。本研究进一步构建出过表达AC006160.1的质粒- pcDNA3.1-AC006160.1，并将其成功转染入膀胱癌细胞系T24细胞和BIU-87细胞，利用CCK8细胞增殖实验、细胞克隆形成实验、Transwell细胞迁移实验、细胞划痕愈合实验等进一步验证其在膀胱癌中的生物学功能。此外，本研究借助药物敏感性分析法继续探索了AC006160.1高、低表达组对不同抗肿瘤药物的潜在治疗反应。

结果

借助具有显著预后价值的铁死亡相关lncRNA行一致性聚类分析发现，在膀胱癌样本中存在两种显著不同的铁死亡相关肿瘤亚型。本研究运用LASSO回归法在分析队列中成功构建出铁死亡相关lncRNA最佳预后模型，其在分析队列和验证队列中，均具备出色的临床生存预测价值。根据铁死亡相关lncRNA表达水平可有效预测膀胱癌样本的总体生存期、肿瘤临床分期和肿瘤病理分级等多项重要指标。本研究分析结果亦表明，铁死亡相关lncRNA模型亦可用以预测膀胱癌免疫治疗临床疗效从而在临床免疫治疗中发挥潜在指导价值。此外，实验验证结果表明，相较于正常尿路上皮细胞SV-HUC-1，AC006160.1在膀胱癌BIU-87细胞、T24细胞、RT4细胞、RT-112细胞、5637细胞中表达量均明显下调。（p < 0.01）细胞功能实验表明，在膀胱癌细胞中过表达AC006160.1可显著抑制膀胱癌细胞的增殖能力和细胞迁移能力。（p < 0.05）药物敏感性分析和体外实验验证表明，AC006160.1高表达患者可从二甲双胍或甲氨蝶呤药物治疗中显著获益。

结论

铁死亡相关lncRNA对膀胱癌多组学特征均具备理想的预后价值。AC006160.1被发现为膀胱癌细胞生长增殖的负性抑制因子。铁死亡诱导可为膀胱癌未来治疗提供新的方向。

Potential prognostic roles of Ferroptosis-related mRNAs and lncRNAs in bladder cancer

Abstract

Part I: Potential prognostic roles of Ferroptosis-related mRNAs in bladder cancer

Objective

Ferroptosis is a new cell death mode which is different from traditional apoptosis and necrosis. It is reported to play important regulatory roles in the pathogenesis of various malignant cancers. However, the potential relationships between ferroptosis and bladder cancer still keep unclear. This study aims to preliminarily explore whether Ferroptosis-related mRNAs play a potential prognostic role in bladder cancer, and further verify the vital target molecule in vitro. This study aims to identify whether the ferroptosis induction can provide a new direction for the new treatment of bladder cancer after chemotherapy and immunotherapy.

Methods

This study firstly utilized the advantages of the high-throughput second-generation sequencing technology to systematically analyze and verify the mRNA transcriptome matrix of bladder cancer. Depending on differentially expressed ferroptosis-related mRNAs (DEGs) between tumor and normal tissues and the prognostic ferroptosis-related mRNAs, the optimal ferroptosis-related mRNA prognosis model was constructed by utilizing the least absolute shrinkage and selection operator cox regression (LASSO) analysis. The Kaplan-Meier analysis, univariate cox regression analysis and multivariate cox regression analysis were conducted to further verify the prognostic roles of optimal risk model. The gene ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis, and single sample Gene Set Enrichment Analysis (ssGSEA) were performed to explore potential biological and immunological mechanisms involved in ferroptosis-related mRNAs during bladder cancer samples. The study further screened out targeted ferroptosis-related mRNA molecule and validate it by experiments in vitro. Quantitative real-time polymerase chain reaction (RT-qPCR) and western blot (WB) experiments were performed to verify the expressions of target ferroptosis-related genes between typical bladder cancer cell lines (HT-1376, BIU-87, T24) and normal bladder immortalized epithelial cells (SV-HUC-1). Short hairpin RNA (shRNA) was constructed to knock down the expressions of target gene in bladder cancer cells. Its potential cell function was further verified by CCK cell proliferation assay, clone formation assay and the wounding healing assay. At the same time, the mitochondrial morphology of bladder cancer cells was identified by transmission electron microscopy. Furthermore, this study collected clinical bladder cancer samples from Beijing hospital for immunohistochemical staining to verify the detailed expressions of target ferroptosis-related gene in this study. The correlation between target Ferroptosis-related genes and the clinical stages was also further explored in this study. The research further explored the potential relationships between the target mRNA and the immune microenvironment of bladder cancer by using the ESTIMATE analysis and CIBERSORT immune infiltration analysis.

Results

In this study, totally 33 ferroptosis-related mRNAs were differentially expressed between bladder cancer tissues and normal tissues. (p < 0.05) Only eight ferroptosis-related mRNAs were identified to pose significant prognostic-roles for overall survivals of bladder cancer samples. (p < 0.05) The intersection of prognostic ferroptosis-related genes and differentially expressed genes identified 4 prognostic targets, including ALOX5, FANCD2, HMGCR and FADS2. The least absolute shrinkage and selection operator cox regression successfully built a 4-gene signature: risk score value = e^sum (each gene’s normalized expression * each gene’s coefficient). Kaplan – Meier survival analysis found that a high Ferroptosis-related mRNA risk predicted a significantly poor survival prognosis both in the derivation cohort and the external validation cohort. (p < 0.05) Multivariate Cox regression analysis in the derivation cohort identified age (p < 0.001), grade (p = 0.129) and risk score (p = 0.016) as independent prognostic predictors for overall survivals. Multivariate Cox regression analysis in the validation cohort also identified age (p = 0.002), stage (p < 0.001) and risk score (p = 0.006) as independent prognostic predictors for overall survivals. Further analysis showed that the ferroptosis-related mRNA risks were significantly correlated with the molecular subtypes of bladder cancers and also the enrichment levels of various immune cells and immune activities. The vitro experiments further demonstrated that the key ferroptosis-related gene in the target model, FADS2, showed significantly higher expressions in cancer cell lines than normal bladder cells in this study. (p < 0.05) CCK8 cell proliferation assay, clone formation assay and the wounding healing assay showed that the proliferation and migration activity of bladder cancer cells were significantly affected after targeted knockdown of FADS2. Bladder cancer cells were further fixed, dehydrated, embedded, sectioned, and stained. With the help of transmission electron microscopy, it was observed that mitochondria were significantly reduced and the mitochondrial membrane density was significantly increased in knockdown cancer cells. In addition, immunohistochemical staining revealed that FADS2 was significantly overexpressed in bladder cancer tissues compared with normal bladder tissues, and its expression level was significantly correlated with the clinical stages of bladder cancer. (p < 0.01). Furthermore, with the help of FADS2 expressions and a variety of clinical elements, this study successfully constructed a clinical prediction nomogram for bladder cancer. Finally, the results of ESTIMATE and CIBERSORT analysis suggested that FADS2 might play a regulatory role in bladder cancer through the tumor immune microenvironment mechanisms.

Conclusion

Ferroptosis-related mRNAs posed excellent predictive-abilities for overall survivals of bladder cancers. Ferroptosis induction could provide new ideas and research directions for the future treatment of bladder cancers.

Part II: Potential prognostic roles of Ferroptosis-related lncRNAs in bladder cancer

Objectives

The role of ferroptosis-related long non-coding RNAs (lncRNAs) in bladder cancer remains elusive. This study aims to examine the prognostic roles of ferroptosis-related lncRNAs in bladder cancer.

Methods

The transcriptomic matrix and clinical information of patients with bladder cancer were obtained from The Cancer Genome Atlas (TCGA) database. A ferroptosis-related lncRNA signature was developed via least absolute shrinkage and selection operator (LASSO) analysis using data from the training cohort, and the signature was further validated using data from the validation cohort. The role of AC006160.1, the most significant lncRNA in the risk signature, was examined in various cell lines including SV-HUC-1, BIU-87, HT-1376, T24, RT4, RT-112, 5637 and UMUC3. The pcDNA3.1-AC006160.1 plasmid was constructed and transfected into the bladder cancer cell lines T24 and BIU-87. In addition, cell proliferation, colony formation, transwell and wound healing assays were performed to examine the biological function of AC006160.1 in T24 and BIU-87 cell lines. Furthermore, this study continued to explore the potential therapeutic responses of high- or low- AC006160.1 group to different therapeutic drugs by utilizing drug sensitivity analysis.

Results

Two clusters were identified through consensus clustering based on prognostic ferroptosis-related lncRNAs. A 5-lncRNA risk signature was successfully constructed using data from the training cohort and validated using data from the validation cohort. The risk signature had excellent ability to predict survival outcomes, clinical stages, pathological grades, expression of immune checkpoints and immunotherapeutic responses in bladder cancer samples. Furthermore, AC006160.1 expression was found to be lower in the cancer cell lines BIU-87, T24, RT4, RT-112 and 5637 than in the normal control cell line SV-HUC-1. Cell proliferation, colony formation, transwell migration and wound healing assays validated that overexpression of AC006160.1 significantly inhibited the proliferation and invasion abilities of both T24 and BIU-87 cells. Drug sensitivity analysis revealed that patients with high expression of AC006160.1 were sensitive to metformin and methotrexate, and the results were further validated via in vitro drug experiments.

Conclusion

Ferroptosis-related lncRNAs play a vital role in predicting the multiomic characteristics of bladder cancer. The lncRNA AC006160.1 serves as a protective factor for the development of bladder cancer.