Cui Jie, Wang Liping, Zhong Waisheng, Chen Zhen, Chen Jie, Yang Hong, Liu Genglong
Department of Head and Neck Surgery, Affiliated Cancer Hospital & Institute of Guangzhou Medical University, Guangzhou, 510095 Guangdong P. R. China.
Department of Otorhinolaryngology Head and Neck Surgery, The First Affiliated Hospital of Hainan Medical University, Haikou, 570102 Hainan P. R. China.
Cancer Cell Int. 2020 Sep 29;20:472. doi: 10.1186/s12935-020-01567-3. eCollection 2020.
Recurrence remains a major obstacle to long-term survival of laryngeal squamous cell carcinoma (LSCC). We conducted a genome-wide integrated analysis of methylation and the transcriptome to establish methylation-driven genes prognostic signature (MDGPS) to precisely predict recurrence probability and optimize therapeutic strategies for LSCC.
LSCC DNA methylation datasets and RNA sequencing (RNA-seq) dataset were acquired from the Cancer Genome Atlas (TCGA). MethylMix was applied to detect DNA methylation-driven genes (MDGs). By univariate and multivariate Cox regression analyses, five genes of DNA MDGs was developed a recurrence-free survival (RFS)-related MDGPS. The predictive accuracy and clinical value of the MDGPS were evaluated by receiver operating characteristic (ROC) and decision curve analysis (DCA), and compared with TNM stage system. Additionally, prognostic value of MDGPS was validated by external Gene Expression Omnibus (GEO) database. According to 5 MDGs, the candidate small molecules for LSCC were screen out by the CMap database. To strengthen the bioinformatics analysis results, 30 pairs of clinical samples were evaluated by digoxigenin-labeled chromogenic in situ hybridization (CISH).
A total of 88 DNA MDGs were identified, and five RFS-related MDGs (LINC01354, CCDC8, PHYHD1, MAGEB2 and ZNF732) were chosen to construct a MDGPS. The MDGPS can effectively divide patients into high-risk and low-risk group, with the area under curve (AUC) of 0.738 (5-year RFS) and AUC of 0.74 (3-year RFS). Stratification analysis affirmed that the MDGPS was still a significant statistical prognostic model in subsets of patients with different clinical variables. Multivariate Cox regression analysis indicated the efficacy of MDGPS appears independent of other clinicopathological characteristics. In terms of predictive capacity and clinical usefulness, the MDGPS was superior to traditional TNM stage. Additionally, the MDGPS was confirmed in external LSCC cohorts from GEO. CMap matched the 9 most significant small molecules as promising therapeutic drugs to reverse the LSCC gene expression. Finally, CISH analysis in 30 LSCC tissues and paired adjacent normal tissues revealed that MAGEB2 has significantly higher expression of LSCC compared to adjacent non-neoplastic tissues; LINC01354, CCDC8, PHYHD1, and ZNF732 have significantly lower expression of LSCC compared to adjacent non-neoplastic tissues, which were in line with bioinformatics analysis results.
A MDGPS, with five DNA MDGs, was identified and validated in LSCC patients by combining transcriptome and methylation datasets analysis. Compared TNM stage alone, it generates more accurate estimations of the recurrence prediction and maybe offer novel research directions and prospects for individualized treatment of patients with LSCC.
复发仍然是喉鳞状细胞癌（LSCC）长期生存的主要障碍。我们进行了全基因组甲基化和转录组综合分析，以建立甲基化驱动基因预后特征（MDGPS），从而精确预测复发概率并优化LSCC的治疗策略。
从癌症基因组图谱（TCGA）获取LSCC DNA甲基化数据集和RNA测序（RNA-seq）数据集。应用MethylMix检测DNA甲基化驱动基因（MDGs）。通过单因素和多因素Cox回归分析，从DNA MDGs的五个基因构建了无复发生存（RFS）相关的MDGPS。通过受试者工作特征（ROC）和决策曲线分析（DCA）评估MDGPS的预测准确性和临床价值，并与TNM分期系统进行比较。此外，通过外部基因表达综合数据库（GEO）验证MDGPS的预后价值。根据5个MDGs，通过CMap数据库筛选出LSCC的候选小分子。为加强生物信息学分析结果，采用地高辛标记的显色原位杂交（CISH）对30对临床样本进行评估。
共鉴定出88个DNA MDGs，选择5个与RFS相关的MDGs（LINC01354、CCDC8、PHYHD1、MAGEB2和ZNF732）构建MDGPS。MDGPS可有效将患者分为高风险和低风险组，5年RFS的曲线下面积（AUC）为0.738，3年RFS的AUC为0.74。分层分析证实，MDGPS在不同临床变量的患者亚组中仍是一个具有显著统计学意义的预后模型。多因素Cox回归分析表明，MDGPS的疗效似乎独立于其他临床病理特征。在预测能力和临床实用性方面，MDGPS优于传统的TNM分期。此外，MDGPS在来自GEO的外部LSCC队列中得到证实。CMap匹配了9种最显著的小分子作为有前景的治疗药物，可逆转LSCC基因表达。最后，对30例LSCC组织和配对的相邻正常组织进行CISH分析，结果显示与相邻非肿瘤组织相比，MAGEB2在LSCC中的表达显著更高；与相邻非肿瘤组织相比，LINC01354、CCDC8、PHYHD1和ZNF732在LSCC中的表达显著更低，这与生物信息学分析结果一致。
通过整合转录组和甲基化数据集分析，在LSCC患者中鉴定并验证了一个包含5个DNA MDGs的MDGPS。与单独的TNM分期相比，它能更准确地预测复发，可能为LSCC患者的个体化治疗提供新的研究方向和前景。
