HCC has been reported to be the second most common cause of cancer-related death worldwide . HBV and HCV infections have resulted in a high rate of HCC in China. Because of the current lack of effective interventions, high metastasis rate, and high mortality rate, it is crucial to develop a deeper understanding of the molecular mechanism of HCC development. A growing body of evidence suggests that liver cancer is a multistep process associated with complex interactions between genetics, epigenetics, and transcriptional changes .
About 100 post-transcriptional chemical modifications can occur in biological RNAs  and m6A, which was discovered in the 1970s, is one of the most abundant endochemical modifications in eukaryotic mRNAs . The biological significance of m6A RNA methylation has been increasingly recognized; it has important and diverse biological functions in mammals, including sex determination, tissue development, DNA damage response, circadian rhythm, and tumorigenesis . In this study, we demonstrated that the expression of m6A RNA methylation modulators, which relates to the field of epigenetics, is also closely related to HCC prognosis. Based on the expression of m6A RNA methylation modulators, two subgroups of HCC tissues were identified by consensus clustering. These clusters had different prognoses and clinicopathological characteristics. In addition, we used four selected m6A RNA methylation modulators to derive a prognostic risk score, which was used to classify the HCC patients into high- and low-risk groups based on the median risk score.
Several studies have pointed out that the occurrence of liver cancer is related to the abnormal expression of m6A RNA methylation modulators [20, 25, 27-29, 35-37]. Zhao et al.  reported that increased YTHDF1 is related to poor prognosis of liver cancer patients, and YTHDF1 plays an important role in regulating liver cancer cell metabolism and cell cycle progression. Cheng and colleagues  reported that KIAA1429 facilitated HCC migration and invasion by inhibiting ID2 via increasing m6A modification of ID2 mRNA. Furthermore, Tanabe et al.  reported that YTHDC2 plays an important role in the growth of liver cancer cells.
The functions of METTL14, METTL3, and YTHDF2 in HCC are controversial. Ma et al.  demonstrated that METTL14 positively regulates the primary miRNA 126 (miR126) in an m6A-dependent manner by interacting with microprocessor complex subunit DiGeorge syndrome critical region 8 (DGCR8), and Ma et al. concluded that METTL14 can inhibit liver cancer metastasis. Ma et al.  also reported that METTL14 and m6A levels were decreased in HCC tissues compared with normal tissues or tissues adjacent to HCC tissues, while METTL3 and WTAP levels were basically unchanged, In contrast, Chen et al.  reported that METTL14 levels were slightly higher in liver cancer tissues than in normal tissues, and METTL3 levels were considerably higher. Based on this, Chen and colleagues concluded that both METTL14 and METTL3 play carcinogenic roles in HCC, and they are necessary for HCC growth and metastasis. Zhong et al.  reported that YTHDF2 may play an anti-tumor role in HCC because its overexpression inhibited cell proliferation and growth and promoted the apoptosis of HCC cells. In contrast, Yang et al.  and Chen et al.  found that YTHDF2 played a pro-cancer role in HCC. These studies indicate that abnormal expression of m6A RNA methylation modulators is closely related to HCC occurrence and development.
In this study, we comprehensively analyzed the expression of the 13 most common m6A RNA methylation modulators in HCC tissues. There was no difference in the expression of METTL14 between HCC and normal tissues from either database, which is consistent with the results reported by Zhou et al. . Additionally, we analyzed the CNV data of these 13 genes using the TCGA database. The HCC tissues were significantly different from the normal tissues in terms of CNV. The expression of all 13 genes except HNRNPC was related to CNV. Only three genes (YTHDF1, YTHDC2, and KIAA1429) mainly exhibited increased copy numbers in the HCC tissues compared to the normal tissues, while the other 10 genes mainly exhibited reduced copy numbers. Furthermore, we analyzed data on SNP mutations in the 13 genes in HCC tissues and found very low mutation rates. These data indicate that abnormal expression of m6A RNA methylation modulators in HCC tissues is not entirely caused by genetic changes (CNV or SNP mutations).
Whether the expression of m6A RNA methylation modulators can be used to predict cancer prognosis is an important research subject . In this study, we used a risk signature constructed with four m6A RNA methylation modulators (YTHDF1, YTHDF2, KIAA1429, and METTL3) to predict OS among HCC patients. In the TCGA database, patients with high risk scores were more likely to have a higher WHO stage and higher pathological grade and be in HCC subtype cluster 2. Also, in the ICGC database, high risk scores were associated with higher WHO stage. It should be noted that the risk score was independently associated with OS among HCC patients in both the TCGA and ICGC analyses. However, unlike in the TCGA analysis, sex was also an independent prognostic factor for HCC in the ICGC analysis, which may have been due to ethnic differences between the two datasets.
The METTL3 RNA methyltransferase is a “writer” protein responsible for m6A modification and is involved in mRNA biogenesis, decay, and translation. METTL3 may play a carcinogenic role in lung cancer , bladder cancer [23, 40], gastric cancer [41, osteosarcoma , cutaneous squamous cell carcinoma , and acute myeloid leukemia (AML) . Li et al.  reported that METTL3 promoted colorectal cancer progression through an m6A-IGF2BP2-dependent mechanism, while Deng et al.  reported that METTL3 inhibited the proliferation and migration of colorectal cancer cells through the p38/ERK pathway. Additionally, Cui and colleagues  reported that METTL3 downregulation significantly promoted the growth, self-renewal, and tumorigenesis of human glioblastoma stem cells, while METTL3 overexpression inhibited the growth and self-renewal of these cells. However, Visvanathan et al.  reported that METTL3 transcription was increased in glioblastoma, while METTL3 silencing inhibited tumor growth and prolonged the survival of mice. These results suggest that METTL3 may play varied roles in different types of cancer, and the study of METTL3 in colorectal cancer and glioblastoma remains controversial.
As a component of the m6A “writer” complex, KIAA1429 is reported to be an RNA-binding protein involved in m6A modification and RNA splicing and processing. At present, its role as an m6A “writer” in tumorigenesis and its mechanism have not been fully reported. However, Cheng et al.  reported that KIAA1429 inhibits ID2 by increasing the m6A modification of ID2 mRNA, thus promoting HCC migration and invasion. Additionally, Qian and colleagues  reported that KIAA1429 can regulate CDK1 in breast cancer in an m6A-independent manner, and act as a carcinogenic factor. These studies suggest that KIAA1429 promotes tumorigenesis and development.
YTH N6-methyladenosine RNA binding protein 1 (YTHDF1) is a member of the YTH domain family, which includes YTHDF1, 2, and 3 and YTHDC1 and 2. As a “reader” of m6A-modified mRNA, cytoplasmic YTHDF1 interacts with binding sites on m6A-modified mRNA to promote the initiation of translation. However, the link between YTHDF1 and cancer is largely unknown. Han et al. reported that m6A RNA modification, involving YTHDF1, modulates the anti-tumor immune response. Zhao et al.  and Zhou et al.  reported that YTHDF1 is highly expressed in liver cancer and is significantly associated with poor prognosis. Furthermore, Nishizawa et al.  and Bai et al.  reported that YTHDF1 is highly expressed in colorectal cancer and plays an important role in carcinogenesis.
The main role of YTHDF2 is to regulate the degradation of m6A-modified mRNAs . However, the relationship between YTHDF2 and cancer is largely unknown. Yang et al.  reported that miR-145 regulates m6A by targeting the 3'-untranslated region of YTHDF2 in HCC cells, and YTHDF2 expression is closely related to the malignant degree of HCC. Thereafter, Chen and colleagues reported that YTHDF2 was highly expressed in pancreatic cancer, which promoted the proliferation and inhibited the migration and invasion of pancreatic cancer cells. Furthermore, many studies have focused on the relationship between YTHDF2 and AML, indicating that YTHDF2 is increased in the broad spectrum of human AML tissues. Targeting YTHDF2 to inhibit its expression can enlarge hematopoietic stem cells, enhance bone marrow reconstruction, and selectively impair AML [53-55], suggesting that it may be useful in the treatment of hematological malignant tumors.