Exploration of the mechanism of 5-Methylcytosine promoting the progression of hepatocellular carcinoma

Abstract

5-Methylcytosine (m5C) is a ubiquitous RNA modification that is closely related to various cellular functions. However, no studies have comprehensively demonstrated the role of m5C in hepatocellular carcinoma (HCC) progression. In this study, six pairs of HCC and adjacent tissue samples were subjected to methylated RNA immunoprecipitation sequencing to identify precise m5C loci. Non-negative matrix factorization (NMF) was used to identify HCC subtypes in TCGA-LIHC cohort. Immune, metabolic, and tumor-related pathways in HCC subtypes with differences in methylation status were analyzed and a prognostic model based on m5C-related genes was constructed. Finally, using RIP and molecular interaction analysis, we demonstrated that YBX1 binds to TPM3 in an m5C dependent manner and regulates HCC progression. Widespread m5C sites were identified and found to be differentially distributed in HCC compared with adjacent tissues. Metabolic processes were inhibited in hypermethylated HCC, whereas immune checkpoint and multiple classical tumor pathways were significantly upregulated. More importantly, we have identified an m5C dependent regulatory axis. The m5C reader YBX1 binds to TPM3 in an M5C dependent manner and promotes the progression of hepatocellular carcinoma. These results provide new evidence for further understanding the comprehensive role of m5C in HCC and the regulatory mechanism of m5C.

Keywords: 5-methylcytosine; HCC; Immunity; Metabolism; Tumor-related pathways; YBX1.

Fig. 1

a Flowchart of the study. b Schematic of methylated RNA immunoprecipitation sequencing (meRIP-Seq).

Fig. 2

m5C sites are differentially distributed in hepatocellular carcinoma (HCC) and adjacent tissues. a Significantly more m5C sites were found in HCC than in adjacent tissues. b Principal component analysis (PCA) of HCC and adjacent tissues by methylomic data. c PCA of HCC and adjacent tissues by transcriptomic data. d Most significant motifs for mRNA in HCC and adjacent tissues. e Most significant motifs for lncRNA in HCC and adjacent tissues. f Most significant motifs for circRNA in HCC and adjacent tissues. g Volcano map of differentially methylated genes in mRNA. h Volcano map of differentially methylated genes in lncRNA. i Volcano map of differentially methylated genes in circRNA.

Fig. 3

Functional enrichment analysis of methylated genes. a-b Gene set enrichment analysis (GSEA) were performed on the methylation genes and differentially methylated genes of HCC. c Gene ontology (GO) biological process terms associated with genes with significant differences in methylation and transcriptomics between HCC and adjacent tissues. d The majority of pathways with correlations >0.8 based on methylation and transcriptomic single-sample GSEA (ssGSEA) scores were negatively correlated. e Among genes with differences in both methylation and transcriptomics, most genes with upregulated transcript levels also showed upregulated m5C levels in HCC. f Protein–protein interaction network based on m5C-related genes. Genes in red circles are high-connectivity modules defined by MCODE.

Fig. 4

Identification of m5C-related hepatocellular carcinoma (HCC) subtypes. a Non-negative matrix factorization clustering heatmap when K = 4. b The changing trends in cophenetic, dispersion, and other indicators when K = 2–6. c Kaplan–Meier analysis between Clusters 2 and 3. d Differences between m5C regulators between Clusters 2 and 3. Almost all regulators were significantly upregulated in Cluster 2. We defined Cluster 2 as the hypermethylated group and Cluster 3 as the hypomethylated group. e Volcano plot of differentially expressed genes between Cluster 2 and 3.

Fig. 5

Differences in immune and classical tumor pathways between the hypermethylated and hypomethylated groups. a Differences in immune infiltration scores between groups based on single-sample gene set enrichment analysis (ssGSEA). b Differences in immune checkpoint gene expression between groups. c Metabolism, transcriptional regulation, and multiple canonical tumor pathways differ significantly between the hypermethylated and hypomethylated groups based on ssGSEA.

Fig. 6

Construction of m5C-related prognostic model. a Trajectories of lambda changes in LASSO for prognosis-related differentially expressed genes between hypermethylated and hypomethylated groups. b Trajectories of confidence intervals for lambda versus independent variables. We identified 15 signature prognostic models. c Significant prognostic differences between high-risk and low-risk groups. d Receiver operating characteristic curve analysis based on the RiskScore we identified. e Violin plot for RiskScore and Stage_T. f Violin plot for RiskScore and Stage. g Univariate Cox analysis of RiskScore and clinical characteristics. h Multivariate Cox analysis of RiskScore and clinical characteristics. i Nomogram construction of RiskScore and clinical characteristics.

Fig. 7

The expression and diagnostic value of YBX1 in HCC. a-c Differential expression of YBX1 between HCC and normal tissues in TCGA-LIHC, GSE14520 and 6 pairs of tissues used for MeRIP Seq. d Differences in YBX1 protein expression between 10 pairs of liver cancer and adjacent tissues in our center; e In situ hybridization immunohistochemistry results of paraffin sections of HCC and adjacent tissues; f-g The difference in IHC total staining area and average staining area between HCC and adjacent tissues; h-i KM analysis of YBX1 in TCGA-LIHC queue and GSE14520 queue; J-K: Differences in YBX1 expression between different stages in the TCGA-LIHC cohort.

Fig. 8

YBX1 binds to TPM3 in an m5C dependent manner. a RNA levels after stable silencing of YBX1 in HepG2 cell lines; b Protein levels after stable silencing of YBX1 in HepG2 cell lines; c CCK8 cell viability experiment between sh1 and sh2 cell lines and control group cell lines; d Clone formation experiment between sh1 and sh2 cell lines and control group cell lines. e The m5C peaks and corresponding m5C consistency motifs of MEX3A, TPM3, and TMCO3 sequences in 6 pairs of liver cancer and adjacent tissues; f-h Expression levels of MEX3A, TPM3, and TMCO3 in control group cells and YBX1 knockout cell lines. i The affinity between TPM3 RNA molecules without m5C modification and YBX1 protein; j Affinity between TPM3 RNA molecules modified with m5C and YBX1 protein.