Department of Gastrointestinal Surgery, Tianjin First Central Hospital, No.24 Fukang Road, Nankai District, Tianjin, 300190, People's Republic of China.
Appl Biochem Biotechnol. 2024 Jul;196(7):4156-4165. doi: 10.1007/s12010-023-04751-z. Epub 2023 Oct 31.
Aberrant expression of gene is driven by its promoter methylation and is the key molecular basis of carcinogenic processes. This study aimed at identifying a risk signature of methylation-driven (MD) genes and evaluating its prognostic value for colon cancer (CC) patients. The expression profiles of methylation and mRNA in CC samples were obtained from the TCGA database, and the MethylMix algorithm was used to identify MD genes. The relationships between their expression levels and overall survival (OS) of CC patients were analyzed, and a prognostic signature of MD genes was established. The risk score of gene signature was calculated, and the median was used to divide all patients into high (H) and low (L) risk groups. The prognostic value of gene signature was tested by the TCGA cohort and an independent validation cohort (GSE17538 dataset). In total, 69 MD genes were identified, and 7 were associated with OS of CC patients. Ultimately, 4 (TWIST1, LDOC1, EPHX3, and STC2) were screened out to establish a risk signature. The H-risk patients (>0.923) had a worse OS than L-risk patients (≤0.923) in both the TCGA (5-year cumulative survival: 52.9% vs 72.0%, P=0.005) and GSE17538 cohort (49.4% vs 69.3%, P=0.004). The AUC values of MD genes signature for the prediction of 3- and 5-year OS were 0.648 and 0.643 in the TCGA dataset and 0.634 and 0.624 in the GSE17538 dataset, respectively. The risk signature of four MD genes was identified as an independent predictor of OS for CC patients (HR for TCGA dataset: 2.071, 95% CI=1.196-3.586, P=0.009; HR for GSE17538 dataset: 2.021, 95% CI=1.290-3.166, P=0.002). The risk signature of four MD genes might be a useful prognostic tool and help doctors improve the clinical management of CC patients.
基因的异常表达是由其启动子甲基化驱动的，是致癌过程的关键分子基础。本研究旨在鉴定一种由甲基化驱动（MD）基因驱动的风险特征，并评估其对结肠癌（CC）患者的预后价值。从 TCGA 数据库中获取 CC 样本的甲基化和 mRNA 表达谱，使用 MethylMix 算法鉴定 MD 基因。分析其表达水平与 CC 患者总生存期（OS）的关系，建立 MD 基因的预后特征。计算基因特征的风险评分，中位数用于将所有患者分为高（H）和低（L）风险组。通过 TCGA 队列和独立验证队列（GSE17538 数据集）验证基因特征的预后价值。总共鉴定出 69 个 MD 基因，其中 7 个与 CC 患者的 OS 相关。最终，筛选出 4 个（TWIST1、LDOC1、EPHX3 和 STC2）基因建立风险特征。在 TCGA（5 年累积生存率：52.9%对 72.0%，P=0.005）和 GSE17538 队列（49.4%对 69.3%，P=0.004）中，H 风险患者（>0.923）的 OS 均差于 L 风险患者（≤0.923）。MD 基因特征对 TCGA 数据集 3 年和 5 年 OS 预测的 AUC 值分别为 0.648 和 0.643，在 GSE17538 数据集分别为 0.634 和 0.624。在 TCGA 数据集（HR：2.071，95%CI=1.196-3.586，P=0.009）和 GSE17538 数据集（HR：2.021，95%CI=1.290-3.166，P=0.002）中，4 个 MD 基因的风险特征均被鉴定为 OS 的独立预测因子。由四个 MD 基因组成的风险特征可能是一种有用的预后工具，有助于医生改善 CC 患者的临床管理。
