Multiple m6A RNA methylation modulators promote the malignant progression of hepatocellular carcinoma and affect its clinical prognosis

Abstract

Background: Hepatocellular carcinoma (HCC) is the second most common cause of cancer-related death in the world. N⁶-methyladenosine (m⁶A) RNA methylation is dynamically regulated by m⁶A RNA methylation modulators ("writer," "eraser," and "reader" proteins), which are associated with cancer occurrence and development. The purpose of this study was to explore the relationships between m⁶A RNA methylation modulators and HCC.

Methods: First, using data from The Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium (ICGC) databases, we compared the expression levels of 13 major m6A RNA methylation modulators between HCC and normal tissues. Second, we applied consensus clustering to the expression data on the m⁶A RNA methylation modulators to divide the HCC tissues into two subgroups (clusters 1 and 2), and we compared the clusters in terms of overall survival (OS), World Health Organization (WHO) stage, and pathological grade. Third, using least absolute shrinkage and selection operator (LASSO) regression, we constructed a risk signature involving the m⁶A RNA methylation modulators that affected OS in TCGA and ICGC analyses.

Results: We found that the expression levels of 12 major m6A RNA methylation modulators were significantly different between HCC and normal tissues. After dividing the HCC tissues into clusters 1 and 2, we found that cluster 2 had poorer OS, higher WHO stage, and higher pathological grade. Four m⁶A RNA methylation modulators (YTHDF1, YTHDF2, METTL3, and KIAA1429) affecting OS in the TCGA and ICGC analyses were selected to construct a risk signature, which was significantly associated with WHO stage and was also an independent prognostic marker of OS.

Conclusions: In summary, m⁶A RNA methylation modulators are key participants in the malignant progression of HCC and have potential value in prognostication and treatment decisions.

Keywords: Liver hepatocellular carcinoma; Prognosis; RNA methylation; m6A.

Fig. 1

Expression of m6A RNA methylation modulators in hepatocellular carcinoma. a Data from TCGA dataset; b Data from ICGC dataset; Wilcox test was used to determine the differential gene expression between tumor group and normal group. * P < 0.05 and *** P < 0.001

Fig. 2

Differential clinicopathological factors and overall survival of hepatocellular carcinoma in the cluster 1/2 subgroups. a Heatmap and clinicopathological factors of the two clusters, Chi-square test was used for correlation between clinical and cluster, * P < 0.05 and *** P < 0.001. b Kaplan–Meier overall survival (OS) curves for 374 TCGA hepatocellular carcinoma patients. The sample size of cluster 1 and cluster 2 is 257 and 117 respectively. c Principal component analysis of the m6A-related gene expression in the TCGA dataset, hepatocellular carcinoma in the cluster1 subgroup are marked with red

Fig. 3

The process of selecting target genes to construct lasso risk regression model by applying the least absolute shrinkage and selection operator (LASSO) Cox regression algorithm. a Univariate Cox regression analysis of the 13 genes correlated with OS in TCGA dataset; the hazard ratios (HR), 95% confidence intervals .c-d The process of selecting target genes in TCGA dataset. e-h The process of selecting target genes in ICGC dataset

Fig. 4

The differences in OS between the low-and high- risk groups based on the median risk score and ROC curve predict survival in TCGA and ICGC datasets. a-b significant differences in OS between the two categories (c-d) Kaplan–Meier overall survival (OS) curves for patients predicting 3 years survival in the TCGA (c) and ICGC (d) datasets assigned to high- and low-risk groups based on the risk score

Fig. 5

Relationship between the risk score, clinicopathological factors and cluster1/2 subgroups in TCGA dataset. a The heatmap shows the distribution of clinicopathological factors and four genes expression compared between the low- and high- risk groups. ** P < 0.01, *** P < 0.001. (b–d) Distribution of risk scores in the TCGA dataset stratified by WHO grade (b) pathological grade (c) and cluster1/2 subgroups (d). e-f Univariate and multivariate analyses of the association between clinicopathological factors (including the risk score) and overall survival of patients in the TCGA datasets, the hazard ratios (HR), 95% confidence intervals

Fig. 6

Relationship between the risk score and clinicopathological factors in ICGC dataset. a Distribution of risk scores in the TCGA dataset stratified by WHO grade. b-c Univariate and multivariate Cox regression analyses of the association between clinicopathological factors (including the risk score) and overall survival of patients in the ICGC dataset, the hazard ratios (HR), 95% confidence intervals