Cholangiocarcinoma (CCA) is a rare disease, but it is amongst the most lethal cancers with a median survival under 1 year. Variations in DNA methylation and gene expression have been extensively studied in other cancers for their role in pathogenesis and disease prognosis, but these studies are very limited in CCA. This study focusses on the identification of DNA methylation and gene expression prognostic biomarkers using multi-omics data of CCA tumors from The Cancer Genome Atlas (TCGA).

We have conducted a genome-wide analysis of differential DNA methylation and gene/miRNA expression using data from 36 CCA tumor and 9 normal samples from TCGA. The impact of DNA methylation in promoters and long-range distal enhancers on the regulation and expression of CCA-associated genes was examined using linear regression. Next, we conducted network analyses on genes which are regulated by DNA methylation as well as by miRNA. Finally, we performed Kaplan–Meier and Cox proportional hazards regression analyses in order to identify the role of selected methylation sites and specific genes and miRNAs in patient survival. We also performed real-time quantitative PCR (qPCR) to confirm the change in gene expression in CCA patients’ tumor and adjacent normal samples.

Altered DNA methylation was observed on 12,259 CpGs across all chromosomes, of which 78% were hypermethylated. We observed a strong negative relationship between promoter hypermethylation and corresponding gene expression in 92% of the CpGs. Differential expression analyses revealed altered expression patterns in 3,305 genes and 101 miRNAs. Finally, we identified 17 differentially methylated promoter CpGs, 72 differentially expressed genes, and two miRNAs that are likely associated with patient survival. Pathway analysis suggested that cell division, bile secretion, amino acid metabolism, PPAR signaling, hippo signaling were highly affected by gene expression and DNA methylation alterations. The qPCR analysis further confirmed that MDK, HNF1B, PACS1, and GLUD1 are differentially expressed in CCA.

Based on the survival analysis, we conclude that DEPDC1, FUT4, MDK, PACS1, PIWIL4 genes, miR-22, miR-551b microRNAs, and cg27362525 and cg26597242 CpGs can strongly support their use as prognostic markers of CCA.

胆管癌 (CCA) 是一种罕见疾病，但它是最致命的癌症之一，中位生存期不到 1 年。DNA甲基化和基因表达的变化已在其他癌症中广泛研究，因为它们在发病机制和疾病预后中的作用，但这些研究在CCA中非常有限。本研究的重点是使用来自癌症基因组图谱 (TCGA) 的 CCA 肿瘤的多组学数据鉴定 DNA 甲基化和基因表达预后生物标志物。

我们使用来自 36 个 CCA 肿瘤和 9 个来自 TCGA 的正常样本的数据对差异 DNA 甲基化和基因/miRNA 表达进行了全基因组分析。使用线性回归检查启动子和远程远端增强子中 DNA 甲基化对 CCA 相关基因的调节和表达的影响。接下来，我们对受 DNA 甲基化和 miRNA 调控的基因进行了网络分析。最后，我们进行了 Kaplan-Meier 和 Cox 比例风险回归分析，以确定选定的甲基化位点和特定基因和 miRNA 在患者生存中的作用。我们还进行了实时定量 PCR (qPCR) 以确认 CCA 患者肿瘤和相邻正常样本中基因表达的变化。

在所有染色体的 12,259 个 CpG 上观察到改变的 DNA 甲基化，其中 78% 是高甲基化的。我们在 92% 的 CpG 中观察到启动子高甲基化与相应基因表达之间存在很强的负相关关系。差异表达分析揭示了 3,305 个基因和 101 个 miRNA 的表达模式改变。最后，我们确定了 17 个差异甲基化启动子 CpG、72 个差异表达基因和两个可能与患者生存相关的 miRNA。通路分析表明，细胞分裂、胆汁分泌、氨基酸代谢、PPAR 信号、河马信号受到基因表达和 DNA 甲基化改变的高度影响。qPCR 分析进一步证实 MDK、HNF1B、PACS1 和 GLUD1 在 CCA 中差异表达。

基于生存分析，我们得出结论，DEPDC1、FUT4、MDK、PACS1、PIWIL4 基因、miR-22、miR-551b microRNA 以及 cg27362525 和 cg26597242 CpG 可以强烈支持它们用作 CCA 的预后标志物。