Background
Hepatocellular carcinoma (HCC) is the main type of liver cancer and has poor prognosis and low survival rates [
1‐
3]. The GLOBOCAN database estimates that HCC is the sixth most commonly diagnosed cancer, and the fourth ranked contributor to the cause of cancer-related deaths [
4]. Worldwide, approximately 841,000 new cancer cases and 782,000 deaths were reported to occur because of HCC in 2018 [
4]. For inhibiting tumor-specific biological reactions, there are a variety of medical techniques, such as surgical resection, liver transplantation, and medication [
2]. However, due to the high recurrence rates and metastasis, the prognosis of HCC patients is still poor, and the 5-year disease-free survival rate was 50.2% [
4]. Hence, it is necessary to develop different approaches for the diagnosis and therapy of HCC.
Epigenetic modifications, such as DNA methylation, histone modifications, and non-coding RNAs (ncRNA) are pivotal factors of gene expression regulation without alteration of the DNA sequence [
5]. Among them, DNA methylation is an extensively characterized epigenetic mechanism of gene regulation in mammals. DNA methylation levels of CpG islands (CpGis) around the transcription start sites (promoter CpGis) conversely regulate gene expression [
6] and such alterations are closely involved in various cancers including HCC [
7‐
9]. The tumor suppressor gene
P16 (cyclin-dependent kinase inhibitor 2A) is a well-known target that suffers DNA methylation silencing in human cancers [
10,
11]. Suppression of DNA methylation changes at such specific targets are expected to provide new approaches to cancer treatment [
8,
12].
Several drugs targeting aberrant DNA methylation, such as 5-aza-2′-deoxycytidine (5-aza-dC), have already been utilized for the therapy of refractory or relapsed cancer patients [
12,
13]. 5-aza-dC exerts anticancer activity by effectively reducing DNA methylation levels and restoring the expression of
P16 [
14,
15]. However, since this type of drug globally reduces DNA methylation, severe side effects were observed in clinical use [
16]. To reduce this risk, it is desirable to restore the site-specific DNA methylation levels involved in cancer progression. For the purpose, recent techniques, such as the Crispr-Cas9 system, enable manipulation of DNA methylation/demethylation at specific sites [
17‐
19]. On the other hand, fewer DNA methylation targets are identified in HCC compared to colon or gastric cancers [
20].
The Cancer Genome Atlas (TCGA) database (
https://tcga-data.nci.nih.gov/tcga/) provides valuable information about not only gene expression but also DNA methylation levels in various cancers from patients in multi-stages. Using the TCGA database, the present study aimed to identify DNA methylation changes which regulate the cancer-related gene expressions in HCC tissues to propose candidate targets for treatment.
For the purpose, we searched for cancer-related genes whose expression levels are significantly altered in HCC and which are associated with poor survival rates using data of 371 HCC tissues and 41 non-tumor tissues compiled in the TCGA database. They were further selected by the presence of promoter CpGis and significant changes of CpGi methylation levels in HCC tissues. Among them, we found 5 genes whose expressions are inversely correlated with DNA methylation and related to poor prognosis. Overall, we identified cancer-related genes whose expressions are associated with the DNA methylation of promoter CpGis in HCC tissues. The method described in this study is applicable to other types of cancers to identify candidates of genes that are regulated by DNA methylation.
Discussion
In the present study, using the TCGA database, we searched for the candidates of cancer-related genes whose expressions are regulated by DNA methylation of CpGis and involved in poor prognosis in HCC tissues. We performed GO analysis for the 98 genes which are upregulated and associated with poor prognosis in HCC and are cancer relevant (Fig.
1). The result showed that they were significantly involved in the G2/M transition of the mitotic cell cycle, mitotic nuclear division, mitotic sister chromatid segregation, cell division, and cell proliferation (Fig.
2). These terms were related to cell malignancy, and their abnormalities or incompletions may lead to carcinogenesis [
27‐
29]. In addition, more than half of the genes are located in the nuclei (Fig.
2b) where early events of cell division occur and mainly have a molecular function of protein binding, which suggest that they belong to transcription regulators controlling cell division and proliferation.
Among the 98 genes, we identified five genes (
FANCB,
KIF15,
KIF4A,
ERCC6L, and
UBE2C) whose expressions were inversely correlated with DNA methylation (Figs.
3,
4, and
5). Recently, Sun et al. reported on the correlation analysis between DNA methylation and gene expression in HCC also using the TCGA database [
30]. As their analysis did not adopt inverse correlation between expression and DNA methylation, they assigned different genes from those we identified in the present study. FANCB, one of the Fanconi anemia proteins, is involved in the repair of DNA lesions and its upregulation is suggested to be required for the survival of colon cancer [
31]. FANCB is also reported to be associated with other types of cancers [
32,
33]. KIF family proteins, including KIF4A and KIF15 encode kinesin-related proteins which are molecular motor proteins that travel along microtubule tracks, play multiple roles in intracellular transport and cell division [
34]. Kinesins are reported to have oncogenic functions such as progression and development of cancers [
35]. Knockdown using siRNA and overexpression of
KIF4A resulted in attenuation and promotion of proliferation of HCC cell lines, respectively [
36]. Knockdown of
KIF15 by shRNA suppressed proliferation of HCC cell lines in vitro and in mice [
37]. Recent studies reported that an increased expression of
KIF4A and
KIF15 are potential prognostic factors in prostate cancer [
38] and lung adenocarcinoma [
39], respectively.
ERCC6L encodes a newly discovered DNA helicase that is highly expressed in almost all cancers [
40].
ERCC6L is known to be an oncogenic protein of solid tumors, since the high expression leads to cancer cell proliferation and tumor growth [
40].
ERCC6L knockdown was demonstrated to result in downregulation of PLK1 which serves an important role in the control of the proliferation and cell cycle in cancer cells [
41].
UBE2C encodes a member of the E2 ubiquitin-conjugating enzyme family and is required for the destruction of mitotic cyclin and for cell cycle progression.
UBE2C expression is upregulated in various cancers including the liver [
42] and abnormal expression of
UBE2C promotes cell cycle progression [
43]. The deletion of UBE2C notably reduced the level of phosphorylated aurora kinase A via Wnt/β–catenin and PI3K/Akt and results in inhibition of the cancer growth and metastasis [
44]. In addition, this gene is the target of miRNAs leading to the inhibition of cancer cell growth and survival in vitro and in vivo [
44].
The present study showed promoter hypomethylation of these genes is associated with increased expression in HCC patients. Newly developed technologies using genome editing, such as CRISPR/Cas9, have enabled not only site-specific DNA demethylation but also methylation in vitro and even in vivo [
17‐
19]. Thus, manipulations of the altered methylation sites of these genes might be promising targets of HCC therapy. As described above, the five genes (
FANCB,
KIF15,
KIF4A,
ERCC6L, and
UBE2C) are all involved in fundamental cellular functions and found in many types of cancers. Thus, manipulations of the altered methylation sites of these genes might be promising targets of therapy of HCC and possibly of other types of cancers. The relations between methylation changes of specific gene promoters and cancer etiology are yet to be investigated.
Silencing of the tumor suppressor gene
P16 by DNA methylation is known to lead to development of cancer cells [
45]. The analysis of the TCGA database in the present study, however, indicated upregulation of
P16 in many types of cancers (Fig.
6f).
P16 takes a part as an early gatekeeper against cancer and the silencing begins at preinvasive stages of a variety of cancers [
10,
46]. On the other hand, recent studies reported that cellular senescence, an irreversible cell cycle arrest, becomes rather a promoting factor of cancer exacerbation through acquisition of the senescence-associated secretory phenotype (SASP) and P
16 upregulation associates with cellular senescence [
47‐
49]. Upregulation of
P16 in many types of cancers (Fig.
6f) may reflect the stage when
P16 downregulation is no longer a factor of cancer progression but rather
P16 associates with cellular senescence.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.