Sage Journals: Discover world-class research

Abstract

Hepatocellular carcinoma (HCC) is a cancer with relatively high mortality, yet little attention has been devoted for related prognostic biomarkers. This study analyzed differential expression of m5C RNA methyltransferase-related genes in normal samples and tumors samples in TCGA-LIHC using Wilcoxon test. K-means consensus clustering analysis was implemented to subdivide samples. Independent prognostic factors were screened by univariate and multivariate Cox regression analyses. KEGG pathway enrichment analysis was performed on the screened independent prognostic factor using GSEA tools. qPCR was conducted to test mRNA expression of key m5C RNA methyltransferase-related genes in tissues and cells. There were 7 m5C RNA methyltransferase-related genes (NOP2, NSUN4, etc.) differentially expressed in HCC tumor tissues. HCC samples were classified into 3 subgroups through clustering analysis according to the expression mode of m5C RNA methyltransferase-related genes. It was also discovered that patients in different subgroups presented significant differences in survival rate and distribution of grade. Additionally, NOP2, NSUN4 and NSUN5 expression notable varied in different grades. Through regression analyses combined with various clinical pathological factors, it was displayed that NSUN4 could work as an independent prognostic factor. KEGG analysis showed that NSUN4 mainly enriched in signaling pathways involved in ADHERENS JUNCTION, RNA DEGRADATION, MTOR SIGNALING PATHWAY, COMPLEMENT and COAGULATION CASCADES. As examined by qPCR, NSUN4 was conspicuously upregulated in HCC patient’s tissues and cells. Altogether, our study preliminarily developed a novel biomarker that could be independently used in prognosis of HCC, which may provide a new direction for the study of related molecular mechanism or treatment regimen.

Keywords

m5C RNA methyltransferase NSUN4 HCC prognosis biomarker

1. Introduction

Hepatocellular carcinoma (HCC) is a cancer with relatively high mortality, strong invasiveness and frequent recurrence. Identified clinical risk factors for HCC are liver cirrhosis, alcohol abuse, steatohepatitis, obesity, diabetes, intake of the aflatoxin B1, etc. [1]. A study unveiled that HCC morbidity is increasing in patients with chronic liver disease and liver cirrhosis [2]. Due to the high incidence of hepatitis B, more than half of HCC patients in the world are in China [3]. Metastatic HCC may achieve distant metastasis directly or indirectly via the circulatory system, arriving at lungs or lymph nodes in the chest through the lymph system [4]. Presently, clinical treatment for HCC includes surgery, chemotherapy, radiotherapy and immunotherapy, but the prognosis of these remains dissatisfied with the 5-year survival rate lower than 50% [5]. Differential expression of some molecules can be used as a biomarker for the prognosis of HCC. For instance, a study pointed out that the expression level of Enolase-1 enzyme can be an independent prognostic factor of HCC [6]. Despite this, few studies about biomarkers related to the prognosis of HCC are available. Finding new biomarkers can provide new entry points and thoughts for the study of HCC-related molecular mechanism.

An increasing number of evidences indicate that post-transcriptional modification of RNA involves in the progression of various cancers in different degrees. Highly expressed NSUN4, for example, is considered to be relevant to the processes of methylation and demethylation and HCC patient’s survival [7]. NSUN4 is a 5-methylcytosine (m5C) methyltransferase responsible for mitochondria rRNA methylation, and it can participate in the assembly of ribosomes in mitochondria and the translation of related molecules by forming a stable dimer complex with MTERF4 [8]. In addition, other members of the M5C methyltransferase family also play an important role in RNA post-transcriptional modification. For example, NSUN1 and NSUN5 modify cytoplasmic ribosomal RNA [9, 10]. Cytoplasmic transport RNA is methylated by NSUN2, NSUN6, and DNMT2 [11, 12, 13]. A study demonstrated that NSUN4 is essential in embryonic development, and knocking out NSUN4 can kill embryos and can sometimes induce myocarditis in the myocardium of mice [14]. Besides, a relevant report uncovered by bioinformatics analysis that m5C distribution and expression are significantly different in HCC tissues and para-carcinoma tissues, and the peak value of m5C methylation in HCC tissues is much higher than that in para-carcinoma tissues [15]. These suggest that m5C methyltransferase family may be involved in the progression of HCC. However, there are rare studies or reports about the relationship between NSUN4 and HCC.

In recent years, the mining of cancer prognostic signals and establishment of prognostic models have been the focus of research. A relevant study explored long non-coding RNAs (lncRNAs) in glioblastoma via constructing a competing endogenous RNA (ceRNA) network mediated by cancer-related lncRNAs [16]. Moreover, a study dug immune-related genes that can guide prognosis using immunological database and ImmPort database, and a prognostic model constructed by 30 immune-related genes was built which can predict patient’s overall survival (OS) [17]. Another study applied multivariate Cox regression analysis to indicate that the 5 studied lncRNAs are all independent clinical pathological factors of HCC [18]. Since previous studies disclosed that highly expressed NSUN4 is significantly correlated with the survival of HCC patients, this study focused on m5C methyltransferase-related genes and mainly analyzed expression differences of these genes. Consensus clustering analysis was performed to classify HCC subgroups. The survival status of patients and the expression mode of m5C methyltransferase-related genes in different subgroups were investigated. Univariate and multivariate Cox regression analyses were adopted to screen genes related to HCC survival and prognosis. The independence of these genes in prognosis was identified. Kyoto encyclopedia of genes and genomes (KEGG) pathway enrichment analysis was undertaken on the independent prognostic factor using gene set enrichment analysis (GSEA) tools.

2. Materials and methods

2.1 Acquisition of HCC clinical information and gene expression data for gene expression analysis

Patient’s clinical information (Table S1) and gene expression profile (FPKM format) in TCGA-LIHC were downloaded from TCGA database (https://www.cancergenome.nih.gov/). A total of 50 normal samples and 374 HCC samples were obtained. Previous studies discovered that human m5C RNA methyltransferase plays an important role in biological processes like protein transcriptional regulation, RNA processing or stress response, while abnormal m5C methyltransferase is potentially associated with varying diseases [12]. Here, 7 m5C RNA methyltransferase-related genes were screened (NOP2, NSUN2, NSUN3, NSUN4, NSUN5, NSUN6, NSUN7) and their expression data were extracted. Afterwards, Wilcoxon test was applied to analyze the expression differences of the above 7 genes in normal tissues and HCC tissues samples, and different clinical pathological stages and grades.

2.2 Consensus clustering analysis

To determine whether HCC could be subdivided by the expression mode of m5C RNA methyltransferase-related genes in HCC, “ConsensusClusterPlus” package was applied to undertake k-means consensus clustering analysis on samples [19]. The resampling method was adopted, with each sampled 80% and repeated 1,000 times. The cluster number K value was set from 2 to 10 in turn to evaluate the stability of the results and the optimal K value was chosen to subdivide the samples. To explore the difference in the survival rate in different subgroups divided by consensus clustering analysis, “survival” package was used to carry out Kaplan-Meier survival analysis on samples in each group [20]. The test method was log-rank test ( $p<$ 0.05 was considered statistically significant).

2.3 Screening of prognosis-related genes and construction of risk-model by univariate and multivariate Cox regression analyses

To screen prognosis-related m5C RNA methyltrans-ferase-related genes, “survival” package was first applied to perform univariate Cox regression analysis on 7 m5C RNA methyltransferase-related genes ( $p<$ 0.05 was considered statistically significant). Next, “survival” package was used to undertake multivariate Cox regression analysis on the screened prognosis-related genes. A risk model was constructed using the final identified optimal signature genes that relate to prognosis. The risk model-based score was formulated as:

$\displaystyle\text{Risk score }=\sum_{i=1}^{n}(\text{Coef}_{i}\ast\text{Exp}_{% i})$

In the formula, $\text{Coef}_{i}$ is the risk coefficient and $\text{Exp}_{i}$ is the expression of genes.

HCC samples were scored according to the risk model. Samples were divided into high- and low-risk groups by the median risk score to continue the following survival analysis. “survivalROC” package was used to draw receiver operating characteristic (ROC) curves to calculate corresponding area under the curve (AUC) values for 1-year OS and 3-year OS [21]. The accuracy of the prognosis predicted by the model was evaluated by the AUC value.

To explore the independence of the signature genes in prognostic assessment, whether the signature genes were related to prognosis was firstly determined by univariate Cox regression analysis ( $p<$ 0.05 represents statistically significant). Afterwards, multivariate Cox regression analysis was used to determine the independence of the signature genes in prognosis. Survival analysis of HCC patients was undertaken on a prognosis-related gene-based model using TIMER website (https://cistrome.shinyapps.io/timer/) [22].

2.4 KEGG pathway enrichment analysis on GSEA

Signature genes that had independent prognostic ability were obtained via univariate and multivariate Cox regression analyses. The threshold value was defined by the median expression of the feature genes, and samples were divided into high- and low-expression groups. KEGG pathway enrichment analysis was performed on the feature genes using GSEA tools to predict potential signaling pathways [23].

2.5 Clinical samples

Clinical samples used in this investigation were provided by HCC patients treated in Tangshan Gongren Hospital. Recruited patients all met the subsequential conditions: i. Patients received surgery for cancer therapy purpose; ii. The associated clinical information and medical histories of patients were accessible; iii. Patients didn’t receive preoperative treatments (radiotherapy and chemotherapy, etc.); iv. Patients didn’t have other tumors except for HCC. Clinical samples included 18 HCC patient’s tumor tissues and corresponding adjacent normal tissues. Fresh samples were preserved at $-$ 80 ${}^{\circ}$ C until qPCR detection. This investigation was approved by Ethics Committee of Tangshan Gongren Hospital and patient’s written informed consent was obtained.

2.6 Cell culture

LO2, HepG2, Hep3B, SMMC-7721, MHCC97L cells were procured from BeNa Culture Collection and were routinely grown in Dulbecco’s Modified Eagle Medium (Gibico, BRL, USA) containing 10% fetal bovine serum (FBS).

2.7 qPCR detection

Total RNA isolation was performed using Trizol (Invitrogen, USA). RNA concentration was calculated using Nanodrop2000 (Thermo fish, USA). TransScript First-Strand (TransGen, China) was employed for RNA reverse transcription. Green qPCR SuperMix and 7500 sequence detection system were used for real-time quantitative amplification. NSUN4 primers were: forward: AGCGTTGCTTCAGACTGGCTGT; reverse: CCTGATCTCTTCAGGCACATAGC; GAPDH primers were: forward: GTCTCCTCTGACTTCAACAGCG; reverse: ACCACCCTGTTGCTGTAGCCAA. Calculation method for relative expression data was 2(̂-delta delta CT).

3. Results

3.1 Expression of m5C RNA methyltransferase-related genes in HCC samples

Wilcoxon test was used to analyze the differences in the expression of 7 m5C RNA methyltransferase-related genes (NOP2, NSUN2, NSUN3, NSUN4, NSUN5, NSUN6, NSUN7) in 50 normal samples and 374 HCC samples. The results showed that these 7 m5C RNA methyltransferase-related genes had highly significant differential expression in normal samples and HCC samples (Fig. 1A and B). Besides, the expression differences in the genes in different stages and grades of tumor were analyzed. It was displayed that NSUN4 showed different expression in varying stages and grades (Fig. 1C and D), and the expression level of NSUN4 increased as HCC progressed (Fig. 1E and F).

Figure 1.

Expression of m5C RNA methyltransferase-related genes in samples. A: Heatmap of the expression of m5C RNA methyltransferase-related genes (NOP2, NSUN2, NSUN3, NSUN4, NSUN5, NSUN6, NSUN7) in normal samples and HCC samples; B: Violin plot of the expression of m5C RNA methyltransferase-related genes in normal samples and HCC samples, with blue plot representing the normal group and red plot representing the tumor group; C-D: Heatmaps of the expression of m5C RNA methyltransferase-related genes in different tumor stages and grades; E-F: Boxplots of the expression of NSUN4 in different tumor stages and grades; * $p<$ 0.05; ** $p<$ 0.01; *** $p<$ 0.001; $p<$ 0.05 represents that the difference is statistically significant.

3.2 HCC subgroup analysis based on k-means consensus clustering analysis

In order to explore whether HCC could be subdivided by the expression mode of m5C RNA methyltransfer-ase-related genes, unsupervised consensus clustering analysis was performed on the profile of these genes in 374 HCC samples in TCGA-LIHC dataset. The results showed that when k went from 2 to 3, the area difference under the CDF curve was the largest (Fig. 2A and B). Meanwhile, when $k=$ 3, it can be seen from the expression calorimeter that the expression patterns of the samples in the three clusters were different (Fig. 2C). It was exposed that the subgroup result was the most stable when $K=$ 3. Thereafter, Kaplan-Meier method was applied to analyze the OS in each subgroup, and it was revealed that the 5-year OS was relatively high in cluster 1 while it was the lowest in cluster 3 (Fig. 2D). Combined with clinical information (Table S2), it was uncovered that patients in different subgroups had significant differences in the distribution of grades. Besides, the expression of other 6 genes except NSUN7 had statistical significance in three subtypes (Fig. 2E). Moreover, it was further illustrated that the expression of NOP2, NSUN4 and NSUN5 was conspicuously different in varying grades (Fig. 2F).

Figure 2.

HCC subgroup analysis based on k-means consensus clustering analysis. A: Curves of cumulative distribution function (CDF) of consensus clustering as cluster number ranging from 2 to 10; B: Curve of CDF delta area as cluster number ranging from 2 to 10; C: Heatmap of sample consistency as cluster number K $=$ 3; D: Analysis of OS in each subgroup based on k-means consensus clustering analysis; E: Heatmap of the expression of m5C RNA methyltransferase-related genes in each subgroup and clinical pathological differences in different subgroups; F: Boxplots of the expression of m5C RNA methyltransferase-related genes in varying clinical pathological grades; * $p<$ 0.05; ** $p<$ 0.01; *** $p<$ 0.001; $p<$ 0.05 represents that the difference is statistically significant.

3.3 Screening of prognosis-related genes, construction and assessment of risk model

Following univariate Cox regression analysis, 4 m5C RNA methyltransferase-related genes of survival significance were screened: NOP2, NSUN2, NSUN4, NSUN5 (Fig. 3A). Afterwards, multivariate Cox regression analysis was carried out, and 2 high-risk factors, NOP2 and NSUN4 (HR ${}_{\textit{NOP2}}=$ 1.4, HR ${}_{\textit{NSUN4}}=$ 2.2), related to the prognosis, were screened. Then, a risk model was constructed (Fig. 3B). Next, samples were grouped according to the model-based risk score into high- and low- risk groups, and OS of the 2 groups were analyzed. It was unveiled that the 5-year survival rate in the low-risk group was higher than that in the high-risk group (Fig. 3C). To further validate the accuracy of the risk model, the AUC value of 1-year OS and 3-year OS was calculated according to ROC curves. It was demonstrated that AUC ${}_{\textit{1year}}=$ 0.72 and AUC ${}_{\textit{3years}}=$ 0.62 (Fig. 3D), suggesting that the risk model was relatively accurate. Later, the differences in expression of signature genes NOP2 and NSUN4 and in pathological factors in the high- and low-risk groups were explored. The results disclosed that the expression of NOP2 and NSUN4 elevated as the risk score increased. Significant differences in pathological grades and stages were found in the high- and low-risk groups (Fig. 3E). Moreover, the risk score increased with HCC progression, although stage 4 showing no upward trend of the risk score due to its few samples (Fig. 3F).

Figure 3.

Screening of prognosis-related genes, construction and assessment of risk model. A: Screening of prognosis-related genes based on univariate Cox regression analysis; B: Screening of signature genes based on multivariate Cox regression analysis; C: Survival curve of high- and low-risk groups (Risk value of each sample was calculated based on prognostic prediction model. The median of risk value was used as the node to divide the samples into high and low risk groups); D: The predictive performance of the prognostic prediction model was evaluated by drawing the ROC curve based on the survival time of patients; E: Heat map of NSUN4 and NOP2 expression in high and low risk groups based on common clinical features; F: Boxplots of the risk score in different clinical stages and grades; * $p<$ 0.05; ** $p<$ 0.01; *** $p<$ 0.001; $p<$ 0.05 represents that the difference is statistically significant.

3.4 Screening of independent prognostic factors and analysis of signaling pathways

To investigate whether the screened prognosis-related genes (NOP2, NSUN4) could work as an independent prognostic factor to assess the prognosis of HCC, firstly, univariate and multivariate Cox regression analyses were performed on the 2 genes with clinical pathological factors. Results of univariate Cox regression analysis displayed that NOP2 and NSUN4 were notably associated with prognosis in addition to clinical stage and T stage (Fig. 4A). Multivariate Cox regression analysis was undertaken on NOP2 and NSUN4 combined with clinical information, and it was illuminated that only NSUN4 could be an independent factor to predict the prognosis of HCC (Fig. 4B). Afterwards, TIMER website was used to analyze survival status of HCC patients with high or low NOP2 and NSUN4 expressions. Similarly, the prognosis of HCC patients with high NSUN4 expression was poor (Fig. S1). Later, samples were divided into NSUN4 highly and lowly expressed groups with the median NSUN4 expression as the threshold. GSEA was employed for KEGG pathway enrichment analysis (Fig. 4C). The results manifested that NSUN4 mainly enriched in signaling pathways like ADHERENS JUNCTION, RNA DEGRADATION, MTOR SIGNALING PATHWAY, COMPLEMENT and COAGULATION CASCADES.

Figure 4.

Screening of independent prognostic factors and signaling pathway analysis. A: Forest map of univariate Cox regression analysis on the expression of NOP2 and NSUN4 combined with clinical information; B: Forest map of multivariate Cox regression analysis on the expression of NOP2 and NSUN4 combined with clinical information; C: GSEA enrichment analysis of high- and low-expression groups of NSUN4 based on KEGG database. FDR $<$ 0.25 is considered that the difference is statistically significant.

3.5 Validation of expression of NSUN4 in HCC tissues and cells

Patient’s tumor tissues and corresponding normal tissues were gathered for validation of bioinformatics prediction (Table S3). NSUN4 expression was also measured in normal liver cell line (LO2) and HCC cell lines (Hep3B, Huh-7, HepG2, SMCC7721, MHCC97L) Then, NSUN4 expression was found to be conspicuously upregulated in tumor tissues and HCC cells (Fig. 5A and B).

Figure 5.

Expression of NSUN4 in tissues and cells. A: qPCR detection of NSU4N expression in HCC tumor tissues and normal tissues; B: qPCR detection of NSU4N expression in LO2, Hep3B, Huh-7, HepG2, SMCC-7721, MHCC97L cells. * $p<$ 0.05.

4. Discussion

RNA epigenetic modification is an important biological process. In recent years, more and more evidences indicate that the aberrant expression of RNA epigenetic modification-related molecules is closely related to the progression of cancers. Relevant studies unmasked that the abnormal expression of NSUN family relates to a variety of diseases including cancers. For instance, NSUN2 mutation can result in autosomal recessive intelligence disorder [24], and knocking out NSUN2 inhibits the migration of glioma cells [25]. Furthermore, studies found that NSUN4 and MTERF4 participate in mitochondrial ribosomes together to modulate the expression of mitochondrial DNA, which is relevant to the regulation of mammalian oxidative phosphorylation [26]. However, abnormal oxidative phosphorylation is closely associated with the progression of cancers [27, 28, 29]. These suggest that m5C RNA methyltransferase-related genes may be involved in progression of cancers. This study mainly discussed the expression of m5C RNA methyltransferase-related genes in HCC and corresponding effect on prognostic prediction. It was disclosed that the expression of m5C RNA methyltransferase-related genes in normal samples and HCC samples was different. The expression levels of genes like NOP2, NSUN2 and NSUN4 in HCC were higher than that in normal tissues. A relevant study represented that high NOP2 expression can stimulate the migratory ability of prostate cancer. [30]. A study revealed that stable NOP2 helps to promote the proliferative ability of HCC [31]. However, direct evidence about the effect of NSUN4 in HCC is rare. Even so, a study clarified that the expression level of NSUN4 elevates with HCC progression [7], which is in good consistent with our research.

Using k-means consensus clustering analysis, HCC was classified into 3 subgroups according to the expression mode of m5C RNA methyltransferase-related genes. Previously, a study divided HCC into 4 subgroups according to the expression mode of hypoxia-related genes [32]. After subdividing HCC, this study analyzed OS of samples in each subgroup and explored whether the survival status of samples in different subgroups is different. Besides, samples in each subgroup were further subdivided, and it was discovered that the expression of m5C RNA methyltransferase-related genes was different in varying clinical grades. The results may provide new references and evidences for the clinical grading of HCC in the future.

A related study classified cancer samples into high- and low-risk groups by constructing a predictive model based on DNA methyltransferase-related gene signature, and then assessed the predictive accuracy of the model via AUC of ROC curve [33]. In this study, firstly, univariate and multivariate Cox regression analyses were applied to construct a risk model of HCC which contained 2 high-risk factors related to prognosis (NOP2, NSUN4). Afterwards, the accuracy of the model was evaluated through AUC of ROC curve. AUC values for 1-year OS and 3-year OS were 0.72 and 0.62, respectively, indicating that the model was relatively accurate in predicting the survival of HCC patients. Most studies adopted ROC curves to evaluate the accuracy of models in predicting survival [34, 35]. However, in addition to ROC curve, this study analyzed the difference in the risk score based on the model in different clinical stages and grades. According to the results, the risk score elevated with HCC progression, which further testified the accuracy of the model.

Combining with clinical information of samples, univariate and multivariate Cox regression analyses revealed that only NSUN4 could be an independent prognostic factor for HCC. A study established a 3-miRNA based predictive model, and testified these 3-miRNA signatures as independent predictors of vascular invasion in HCC patients by multivariate Cox regression analysis [36]. In the context, NSUN4 is potential to be a novel factor to guide the prognostic management of HCC patients. Besides, it was uncovered by KEGG pathway enrichment analysis that NSUN4 was mainly enriched in signaling pathways including ADHERENS JUNCTION, RNA DEGRADATION, MTOR SIGNALING PATHWAY, COMPLEMENT and COAGULATION CASCADES. Aberrant ADHERENS JUNCTION directly affects the invasive and migratory abilities of cancer cells. For example, prostate cancer can release exosomes under hypoxic conditions to regulate the expression of related molecules in the pathways involved in ADHERENS JUNCTION, thereby enhancing the invasive ability of the cancer [37]. Nevertheless, no study can support that NSUN4 can modulate the expression of ADHERENS JUNCTION-related molecules at present, which may provide a new idea for future research directions. So far, biological functions of NSUN4 in cells have not been fully dug. NSUN2 is the most well-known member of NSUN family so far, and its effect has been clarified in various cancers. For instance, NSUN2 shortage can cause skin cancer progression [38]. Another study about the expression of NSUN family in mouse embryos demonstrated that NSUN family may have overlaps in biological functions, indicating that NSUN4 may play an unknown biological function in cancers.

Viewed in toto, this study mainly analyzed the expression of m5C methyltransferase-related genes in HCC. HCC was firstly divided into 3 subgroups by k-means consensus clustering analysis. A risk model was then established based on m5C methyltransferase-related genes by univariate and multivariate Cox regression analyses. The accuracy of the model was also evaluated. Combining patient’s clinical information, it was found that NSUN4 could be an independent prognostic factor of HCC. However, there are some shortages in this study. For instance, this study didn’t use HCC datasets in other databases in validating the accuracy of the risk model, thus the accuracy of the risk model should be further examined.

Ethical approval

Ethical approval is not applicable for this article. This article does not contain any studies with human or animal subjects.

Competing interest

The authors declare no conflicts of interest.

Availability of data and materials

The data used to support the findings of this study are included within the article. The data and materials in the current study are available from the corresponding author on reasonable request.

Funding statement

Not applicable.

Authors’ contributions

Conception: Dr. Mingxin Cui and Dr. Fengzhi Qu.

Interpretation and Analysis of Data: Dr. Libing Wang and Dr. Xiaogang Liu.

Preparation of the manuscript: Dr. Fengzhi Qu and Dr. Zhaoyuan Tang.

Revision for Important Intellectual Content: Dr. Mingxin Cui and Dr. Daming Cheng.

Supervision: Dr. Jingkun Yu.

Supplementary data

The supplementary files are available to download from http://dx.doi.org/10.3233/CBM-210154.

sj-pdf-1-cbm-10.3233_CBM-210154.pdf - Supplemental material

Supplemental material, sj-pdf-1-cbm-10.3233_CBM-210154.pdf

sj-xlsx-2-cbm-10.3233_CBM-210154.xlsx - Supplemental material

Supplemental material, sj-xlsx-2-cbm-10.3233_CBM-210154.xlsx

sj-xlsx-3-cbm-10.3233_CBM-210154.xlsx - Supplemental material

Supplemental material, sj-xlsx-3-cbm-10.3233_CBM-210154.xlsx

sj-xlsx-1-cbm-10.3233_CBM-210154.xlsx - Supplemental material

Supplemental material, sj-xlsx-1-cbm-10.3233_CBM-210154.xlsx

Footnotes

Acknowledgments

Not applicable.

References

Zhao

Dou

Cao

Wang

Gao

Zeng

Liu

Cui

Teng

Zhang

and Zhao

, Uncovering the potential differentially expressed miRNAs as diagnostic biomarkers for hepatocellular carcinoma based on machine learning in The Cancer Genome Atlas database, Oncol Rep 43 (2020), 1771–1784.

Santopaolo

Lenci

Milana

Manzia

T.M.

and Baiocchi

, Liver transplantation for hepatocellular carcinoma: Where do we stand, World J Gastroenterol 25 (2019), 2591–2602.

Zhou

P.C.

Sun

L.Q.

Shao

L.Z.

and Fan

X.G.

, Establishment of a pattern recognition metabolomics model for the diagnosis of hepatocellular carcinoma, World J Gastroenterol 26 (2020), 4607–4623.

Alemayehu

and Hailu

, Hepatocellular carcinoma with vertebral metastasis in a child with hepatitis B virus infection: a case report, J Med Case Rep 13 (2019), 158.

Ding

Liu

Zou

Cheng

and Su

, Long non-coding RNA TPTEP1 inhibits hepatocellular carcinoma progression by suppressing STAT3 phosphorylation, J Exp Clin Cancer Res 38 (2019), 189.

Zhu

Chen

Ren

and Cui

, Enolase-1 serves as a biomarker of diagnosis and prognosis in hepatocellular carcinoma patients, Cancer Manag Res 10 (2018), 5735–5745.

Zhang

Zheng

and Guo

, Role of m(5)C-related regulatory genes in the diagnosis and prognosis of hepatocellular carcinoma, Am J Transl Res 12 (2020), 912–922.

Yakubovskaya

Guja

K.E.

Mejia

Castano

Hambardjieva

Choi

W.S.

and Garcia-Diaz

, Structure of the essential MTERF4:NSUN4 protein complex reveals how an MTERF protein collaborates to facilitate rRNA modification, Structure 20 (2012), 1940–7.

Schosserer

Minois

Angerer

T.B.

Amring

Dellago

Harreither

Calle-Perez

Pircher

Gerstl

M.P.

Pfeifenberger

Brandl

Sonntagbauer

Kriegner

Linder

Weinhausel

Mohr

Steiger

Mattanovich

Rinnerthaler

Karl

Sharma

Entian

K.D.

Kos

Breitenbach

Wilson

I.B.

Polacek

Grillari-Voglauer

Breitenbach-Koller

and Grillari

, Methylation of ribosomal RNA by NSUN5 is a conserved mechanism modulating organismal lifespan, Nat Commun 6 (2015), 6158.

10.

Bourgeois

Ney

Gaspar

Aigueperse

Schaefer

Kellner

Helm

and Motorin

, Eukaryotic rRNA Modification by Yeast 5-Methylcytosine-Methyltransferases and Human Proliferation-Associated Antigen p120, PLoS One 10 (2015), e0133321.

11.

Blanco

Dietmann

Flores

J.V.

Hussain

Kutter

Humphreys

Lukk

Lombard

Treps

Popis

Kellner

Holter

S.M.

Garrett

Wurst

Becker

Klopstock

Fuchs

Gailus-Durner

Hrabe de Angelis

Karadottir

R.T.

Helm

Ule

Gleeson

J.G.

Odom

D.T.

and Frye

, Aberrant methylation of tRNAs links cellular stress to neuro-developmental disorders, EMBO J 33 (2014), 2020–39.

12.

Liu

R.J.

Long

and Wang

E.D.

, Structural basis for substrate binding and catalytic mechanism of a human RNA:m5C methyltransferase NSun6, Nucleic Acids Res 45 (2017), 6684–6697.

13.

Goll

M.G.

Kirpekar

Maggert

K.A.

Yoder

J.A.

Hsieh

C.L.

Zhang

Golic

K.G.

Jacobsen

S.E.

and Bestor

T.H.

, Methylation of tRNAAsp by the DNA methyltransferase homolog Dnmt2, Science 311 (2006), 395–8.

14.

Metodiev

M.D.

Spahr

Loguercio Polosa

Meharg

Becker

Altmueller

Habermann

Larsson

N.G.

and Ruzzenente

, NSUN4 is a dual function mitochondrial protein required for both methylation of 12S rRNA and coordination of mitoribosomal assembly, PLoS Genet 10 (2014), e1004110.

15.

Zhang

Zheng

and Guo

, Overview of distinct 5-methylcytosine profiles of messenger RNA in human hepatocellular carcinoma and paired adjacent non-tumor tissues, J Transl Med 18 (2020), 245.

16.

Gong

Hong

Wang

Zhang

and Chen

, An eight-mRNA signature outperforms the lncRNA-based signature in predicting prognosis of patients with glioblastoma, BMC Med Genet 21 (2020), 56.

17.

Song

Shang

Yang

Zhang

Chen

and Wu

, Identification of an immune signature predicting prognosis risk of patients in lung adenocarcinoma, J Transl Med 17 (2019), 70.

18.

Zhao

Q.J.

Zhang

and Liu

F.F.

, Identification of a five-long non-coding RNA signature to improve the prognosis prediction for patients with hepatocellular carcinoma, World J Gastroenterol 24 (2018), 3426–3439.

19.

Wilkerson

M.D.

and Hayes

D.N.

, ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking, Bioinformatics 26 (2010), 1572–3.

20.

Ørnulf Borgan. Modeling Survival Data: Extending the Cox Model. Terry M. Therneau and Patricia M. Grambsch, Springer – Verlag, New York, 2000. No. of pages: xiii

+

350. Price: $69.95. ISBN 0-387-98784-3. Statistics in Medicine 20(13) (2001) 2053–2054.

21.

Kamarudin

A.N.

Cox

and Kolamunnage-Dona

, Time-dependent ROC curve analysis in medical research: current methods and applications, BMC Med Res Methodol 17 (2017), 53.

22.

Severson

Pignon

J.C.

Zhao

Novak

Jiang

Shen

Aster

J.C.

Rodig

Signoretti

Liu

J.S.

and Liu

X.S.

, Comprehensive analyses of tumor immunity: implications for cancer immunotherapy, Genome Biol 17 (2016), 174.

23.

Subramanian

Tamayo

Mootha

V.K.

Mukherjee

Ebert

B.L.

Gillette

M.A.

Paulovich

Pomeroy

S.L.

Golub

T.R.

Lander

E.S.

and Mesirov

J.P.

, Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles, Proc Natl Acad Sci USA 102 (2005), 15545–50.

24.

Abbasi-Moheb

Mertel

Gonsior

Nouri-Vahid

Kahrizi

Cirak

Wieczorek

Motazacker

M.M.

Esmaeeli-Nieh

Cremer

Weissmann

Tzschach

Garshasbi

Abedini

S.S.

Najmabadi

Ropers

H.H.

Sigrist

S.J.

and Kuss

A.W.

, Mutations in NSUN2 cause autosomal-recessive intellectual disability, Am J Hum Genet 90 (2012), 847–55.

25.

Zhang

and Zhang

, NSun2 promotes cell migration through methylating autotaxin mRNA, JBiol Chem (2020).

26.

Camara

Asin-Cayuela

Park

C.B.

Metodiev

M.D.

Shi

Ruzzenente

Kukat

Habermann

Wibom

Hultenby

Franz

Erdjument-Bromage

Tempst

Hallberg

B.M.

Gustafsson

C.M.

and Larsson

N.G.

, MTERF4 regulates translation by targeting the methyltransferase NSUN4 to the mammalian mitochondrial ribosome, Cell Metab 13 (2011), 527–39.

27.

Tan

J.L.

Yeo

J.Z.

Yong

K.J.

Bassal

M.A.

G.H.

Lee

M.Y.

Leong

C.Y.

Tan

H.K.

C.S.

Liu

B.H.

Chan

T.H.

Tan

Z.H.

Chan

Y.S.

Wang

Lim

Z.H.

Toh

T.B.

Hooi

Low

K.N.

Kong

N.R.

Stein

A.J.

Thangavelu

M.T.

Suzuki

Periyasamy

Asara

J.M.

Dan

Y.Y.

Bonney

G.K.

Chow

E.K.

G.D.

H.H.

Kanagasundaram

S.B.

Tam

W.L.

Tenen

D.G.

and Chai

, New high-throughput screening identifies compounds that reduce viability specifically in liver cancer cells that express high levels of SALL4 by inhibiting oxidative phosphorylation, Gastroenterology 157 (2019), 1615–1629 e17.

28.

Rodriguez-Enriquez

Pacheco-Velazquez

S.C.

Marin-Hernandez

Gallardo-Perez

J.C.

Robledo-Cadena

D.X.

Hernandez-Resendiz

Garcia-Garcia

J.D.

Belmont-Diaz

Lopez-Marure

Hernandez-Esquivel

Sanchez-Thomas

and Moreno-Sanchez

, Resveratrol inhibits cancer cell proliferation by impairing oxidative phosphorylation and inducing oxidative stress, Toxicol Appl Pharmacol 370 (2019), 65–77.

29.

Boukalova

Hubackova

Milosevic

Ezrova

Neuzil

and Rohlena

, Dihydroorotate dehydrogenase in oxidative phosphorylation and cancer, Biochim Biophys Acta Mol Basis Dis 1866 (2020), 165759.

30.

Sun

Yao

Zheng

Sun

Wang

Liu

and Zhu

, Long Noncoding RNA PVT1 Promotes Prostate Cancer Metastasis by Increasing NOP2 Expression via Targeting Tumor Suppressor MicroRNAs, Onco Targets Ther 13 (2020), 6755–6765.

31.

Wang

Yuan

J.H.

Wang

S.B.

Yang

Yuan

S.X.

Yang

Zhou

W.P.

W.L.

and Sun

S.H.

, Oncofetal long noncoding RNA PVT1 promotes proliferation and stem cell-like property of hepatocellular carcinoma cells by stabilizing NOP2, Hepatology 60 (2014), 1278–90.

32.

Zhang

Tang

Gao

Kong

and Qin

, A hypoxia-related signature for clinically predicting diagnosis, prognosis and immune microenvironment of hepatocellular carcinoma patients, J Transl Med 18 (2020), 342.

33.

G.X.

Ding

Z.Y.

Wang

Y.W.

Liu

T.T.

Chen

W.X.

J.J.

W.Q.

Zhu

and Zhang

B.X.

, Integrative analysis of DNA methylation and gene expression identify a six epigenetic driver signature for predicting prognosis in hepatocellular carcinoma, J Cell Physiol 234 (2019), 11942–11950.

34.

Zhou

Cai

Xie

Gao

Huang

Bai

and Tang

, A Novel Ten-Gene Signature Predicting Prognosis in Hepatocellular Carcinoma, Front Cell Dev Biol 8 (2020), 629.

35.

Zhan

X.H.

Jiao

J.W.

Zhang

H.F.

C.Q.

Zhao

J.M.

Liao

L.D.

J.Y.

B.L.

Z.Y.

Wang

S.H.

Z.P.

Shen

J.H.

Zou

H.Y.

Neufeld

L.Y.

and Li

E.M.

, A three-gene signature from protein-protein interaction network of LOXL2- and actin-related proteins for esophageal squamous cell carcinoma prognosis, Cancer Med 6 (2017), 1707–1719.

36.

Han

Zheng

Wang

Sun

and Zhang

, A novel microRNA signature predicts vascular invasion in hepatocellular carcinoma, J Cell Physiol 234 (2019), 20859–20868.

37.

Ramteke

Ting

Agarwal

Mateen

Somasagara

Hussain

Graner

Frederick

Agarwal

and Deep

, Exosomes secreted under hypoxia enhance invasiveness and stemness of prostate cancer cells by targeting adherens junction molecules, Mol Carcinog 54 (2015), 554–65.

38.

Chellamuthu

and Gray

S.G.

, The RNA Methyltransferase NSUN2 and Its Potential Roles in Cancer, Cells 9 (2020).

m5C RNA methyltransferase-related gene NSUN4 stimulates malignant progression of hepatocellular carcinoma and can be a prognostic marker

Abstract

Keywords

1. Introduction

2. Materials and methods

2.1 Acquisition of HCC clinical information and gene expression data for gene expression analysis

2.2 Consensus clustering analysis

2.3 Screening of prognosis-related genes and construction of risk-model by univariate and multivariate Cox regression analyses

2.4 KEGG pathway enrichment analysis on GSEA

2.5 Clinical samples

2.6 Cell culture

2.7 qPCR detection

3. Results

3.1 Expression of m5C RNA methyltransferase-related genes in HCC samples

Ethical approval

Competing interest

Availability of data and materials

Funding statement

Authors’ contributions

Supplementary data

sj-pdf-1-cbm-10.3233_CBM-210154.pdf - Supplemental material

sj-xlsx-2-cbm-10.3233_CBM-210154.xlsx - Supplemental material

sj-xlsx-3-cbm-10.3233_CBM-210154.xlsx - Supplemental material

sj-xlsx-1-cbm-10.3233_CBM-210154.xlsx - Supplemental material

Footnotes

Acknowledgments

References