Identification of potential key genes and pathways in hepatitis B virus-associated hepatocellular carcinoma by bioinformatics analyses.

Chronic hepatitis B virus (HBV) is one of the leading causes of hepatocellular carcinoma (HCC). The precise molecular mechanisms by which HBV contributes to HCC development are not fully understood. The key genes and pathways involved in the transformation of nontumor hepatic tissues into HCC tissues in patients with HBV infection are essential to guide the treatment of HBV-associated HCC. Five datasets were collected from the Gene Expression Omnibus database to form a large cohort. Differentially expressed genes (DEGs) were identified between HCC tissues and nontumor hepatic tissues from HBV-infected patients using the 'limma' package. The top 50 upregulated and top 50 downregulated DEGs in HCC vs. nontumor tissues were demonstrated in subsets by heat maps. Based on the DEGs, Gene Ontology functional and Kyoto Encyclopedia of Genes and Genomes pathways enrichment analyses were performed. Several key pathways of the up- and downregulated DEGs were identified and presented by protein-protein interaction (PPI) networks. A total of 1,934 DEGs were identified. The upregulated DEGs were primarily associated with the 'cell cycle'. Among the DEGs enriched in the 'cell cycle' pathway, 6 genes had a log2-fold change >2: SFN, BUB1B, TTK, CCNB1, CDK1 and CDC20. The downregulated DEGs were primarily associated with the metabolic pathways, such as 'carbon metabolism', 'glycine, serine and threonine metabolism', 'tryptophan metabolism', 'retinol metabolism' and 'alanine, aspartate and glutamate metabolism'. The DEGs in the 'cell cycle' and 'metabolic pathways' were presented by the PPI networks respectively. Overall, the present study provides new insights into the specific etiology of HCC and molecular mechanisms for the transformation of nontumor hepatic tissues into HCC tissues in patients with a history of HBV infection and several potential therapeutic targets for targeted therapy in these patients. Copyright: © Zhang et al.

Introduction

Hepatocellular carcinoma (HCC) is a prevalent malignant liver disease (1). In most cases, viral infection contributes to the development, invasion and metastasis of HCC (2), which is a global public health problem. In particular, chronic hepatitis B virus (HBV) is one of the leading causes of HCC with an estimated 400 million individuals currently affected by chronic infection worldwide (3,4). More than 50% of HCC cases arise from chronic HBV infections (5,6). In high-prevalence areas, chronic HBV infection is estimated to account for over 80% of HCC cases (7). Moreover, patients with HBV-associated HCC have notably higher rates of metastasis and recurrence compared with those without HBV infection (8,9). Three-quarters of the world's population live in areas where there are high levels of HBV infection (10). However, the currently available antiviral agents can barely eliminate chronic HBV infection (11). HBV-associated liver diseases cause approximately 1 million deaths per year (3), driving an intensive search for curative treatment approaches (12). Han et al (13) reported that WNT family gene expression is associated with the development of HBV-associated HCC. Tian and Ou's study (14) found that chronic HBV infection could lead to chronic inflammation in the liver, which could cause normal liver cells to transform into cancer cells (15). Although the correlation between chronic HBV infection and HCC development is strong, the precise molecular mechanisms by which HBV contributes to HCC development are not fully understood (16). Therefore, a clearer understanding of the molecular mechanisms of the transformation of nontumor hepatic tissues into HCC tissues in patients with HBV infection is required to guide the treatment of HBV-associated HCC (17).

A large array of data could be analyzed, given the remarkable development of high-throughput technologies for the profiling of genome-wide methylation and expression, such as methylation microarray and MeDip-seq, and RNA-seq, and the datasets publicly available worldwide (18). Potential biomarkers and signaling pathway associated with tumor regression could be identified using bioinformatics methods.

Thus far, there are insufficient bioinformatics studies focusing on the differentially expressed genes (DEGs) between HCC tissues and nontumor tissues from HBV-infected patients based on a large sample size. In the present study, data from more datasets on the same platform were collected in order to increase the sample size. Using a large cohort, DEGs between HCC tissues and nontumor tissues were identified. Furthermore, Gene Ontology (GO) functional enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of the DEGs were performed. In addition, protein-protein interaction (PPI) networks were constructed based on the most enriched pathways. The results of the present study may help to identify key biomarkers for the personalized treatment of patients with HCC and a history of HBV infection, and provide further insights into tumor progression and further studies on HCC.

Materials and methods

Microarray datasets for differential expression analyses

A comprehensive search was conducted for HCC microarray datasets, including tissue samples from HBV-infected patients in the Gene Expression Omnibus (GEO) database of the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/geo/) website. All the data of the selected datasets [GSE17548 (19), GSE55092 (20), GSE62232 (21), GSE84044 (22) and GSE84402 (23)] were produced from the GPL570 platform. Subsequently, the raw intensity files (CEL) of the datasets were downloaded from the GEO database. The robust multiarray average method of the R package ‘affy’ (version 1.60.0; http://bioconductor.org/packages/affy/) was used to process the raw intensity files and generate a large gene expression matrix of all the selected samples from the datasets meeting the criteria (HBV positive liver tissue samples with status information) for differential expression analyses (24). The matrices for each selected dataset were also generated. In addition, we used two independent HCC datasets, The Cancer Genome Atlas (TCGA; http://portal.gdc.cancer.gov/projects/TCGA-LIHC) and GSE76427 (25), including HCC patients without HBV infection to further elucidate the specificity of the expression of target genes in HBV-associated HCC.

Analyses of DEGs

Within the large cohort, the DEGs between HCC tissues and nontumor hepatic tissues were identified using the R package ‘limma’ (version 3.38.3; http://bioconductor.org/packages/limma/), which is based on unpaired t-test (26); with the thresholds of log2-fold change >1 or <-1 and adjusted P-value <0.05. The results of the differential expression analyses were visualized with a volcano plot using the R package ‘ggplot2’ (version 3.1.0; http://bioconductor.org/packages/ggplot2/). The top 50 upregulated and top 50 downregulated DEGs were represented by heatmaps using the MeV software (version 4.9.0; http://mev.tm4.org/) in the selected datasets. The unsupervised hierarchical clustering of the selected genes and samples in the heatmaps was performed using an average linkage method using Pearson's correlation.

Enrichment analysis of GO function and KEGG pathway

GO (http://www.geneontology.org) function and KEGG (https://www.kegg.jp/) pathways enrichment analyses of the upregulated and downregulated DEGs were performed using the WEB-based GEne SeT AnaLysis Toolkit (http://www.webgestalt.org/) via a significance threshold of false discovery rate (FDR) <0.05, in order to understand the critical biological implications of the identified DEGs in HBV-positive HCC tissues.

PPI network analyses

To further understand the direct and indirect associations among the DEGs, PPI networks of the upregulated and downregulated DEGs based on the top pathways of the KEGG pathway enrichment analysis were constructed and visualized using the Search Tool for the Retrieval of Interacting Genes/Proteins (https://string-db.org) database. The aforementioned methods are summarized in Fig. 1.

Figure 1.

The process of identifying key genes and pathways in hepatitis B virus-associated hepatocellular carcinoma. GEO, Gene Expression Omnibus; DEGs, differentially expressed genes; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; PPI, protein-protein interaction.

Results

Selection of microarray datasets for differential expression analysis

From the GEO database of NCBI, five datasets (GSE17548, GSE55092, GSE62232, GSE84044 and GSE84402) that met the study criteria were used for differential expression analyses. Within the five datasets, 321 HBV-positive samples with valid hepatic tissue status were selected to generate the gene expression matrix; 82 of which were tumor tissues, and 239 of which were nontumor tissues (Table I). The clinical characteristics of enrolled subjects can be found in supplementary data (Table SI).

Table I.

Characteristics of the datasets used in this study.

	Samples (n)
Dataset	HCC	Non-tumor	Tissue	Platform
GSE17548	10	11	HBV hepatic tissue	Affymetrix Human Genome U133 Plus 2.0 Array
GSE55092	49	91	HBV hepatic tissue	Affymetrix Human Genome U133 Plus 2.0 Array
GSE62232	10	0	HBV hepatic tissue	Affymetrix Human Genome U133 Plus 2.0 Array
GSE84044	0	124	HBV hepatic tissue	Affymetrix Human Genome U133 Plus 2.0 Array
GSE84402	13	13	HBV hepatic tissue	Affymetrix Human Genome U133 Plus 2.0 Array

HCC, human hepatocellular carcinoma; HBV, hepatitis B virus.

DEGs in HCC tissues compared with nontumor hepatic tissues

The expression values of 42,901 genes among 82 HCC samples were compared with 239 nontumor hepatic samples from five GEO datasets. A total of 1934 DEGs were identified with the thresholds of fold change >1 or <-1 and adjusted P-value <0.05, including 682 upregulated genes and 1,252 downregulated genes. All DEGs are marked as red dots in the volcano plot (Fig. 2). In addition, the top 50 upregulated genes and downregulated genes are listed in Tables II and III, respectively (the top 100 up- and downregulated genes are listed in Tables SII and SIII, respectively). The heat maps of GSE17548, GSE55092 and GSE84402 demonstrate the expression of the top 100 DEGs (50 upregulated and 50 downregulated) in different subsets (GSE17548 and GSE84402 in Fig. 3; GSE55092 in Fig. S1).

Figure 2.

Volcano plot of DEGs A total of 1934 DEGs were identified with the thresholds of log2fold change >1 or <-1 and adjusted P-value <0.05. All DEGs were marked as red dots. DEGs, differentially expressed genes.

Table II.

Top 50 upregulated differentially expressed genes.

Probe ID	Gene	Fold change	Adjusted P-value
206239_s_at	SPINK1	4.567348387	9.90×10−43
211470_s_at	SULT1C2	3.963534914	1.47×10−64
205815_at	REG3A	3.770039412	1.10×10−24
209220_at	GPC3	3.752196862	1.99×10−40
201291_s_at	TOP2A	3.555152759	3.85×10−77
206561_s_at	AKR1B10	3.468810343	1.21×10−24
242881_x_at	DUXAP10	3.301086607	1.01×10−75
204602_at	DKK1	3.048004476	1.88×10−37
203820_s_at	IGF2BP3	3.043781559	1.76×10−56
238021_s_at	CRNDE	2.985078406	4.63×10−52
209921_at	SLC7A11	2.980302339	1.66×10−51
213194_at	ROBO1	2.947155321	1.69×10−71
241418_at	NMRAL1P1	2.911349891	8.91×10−59
223642_at	ZIC2	2.821525207	4.36×10−64
222608_s_at	ANLN	2.787055993	6.53×10−62
209875_s_at	SPP1	2.758625836	3.14×10−23
235004_at	RBM24	2.743050577	1.06×10−57
212551_at	CAP2	2.739651435	9.30×10−91
214612_x_at	MAGEA6	2.698311238	6.80×10−34
202422_s_at	ACSL4	2.663914238	2.38×10−33
219918_s_at	ASPM	2.652302575	4.07×10−52
235763_at	SLC44A5	2.472658451	2.37×10−33
207828_s_at	CENPF	2.472581020	2.98×10−78
214710_s_at	CCNB1	2.403923370	1.64×10−56
225681_at	CTHRC1	2.384346752	1.53×10−25
212531_at	LCN2	2.383664937	3.04×10−31
201890_at	RRM2	2.381259393	3.31×10−38
33323_r_at	SFN	2.365919617	9.97×10−36
218009_s_at	PRC1	2.350952594	8.64×10−54
204162_at	NDC80	2.332439517	1.54×10−54
219787_s_at	ECT2	2.313749179	8.57×10−52
207165_at	HMMR	2.304640216	1.82×10−48
204825_at	MELK	2.298883197	1.43×10−57
203477_at	COL15A1	2.295323078	1.49×10−33
203213_at	CDK1	2.271576100	9.42×10−49
206626_x_at	SSX1	2.253448717	3.68×10−30
207325_x_at	MAGEA1	2.252155461	3.97×10−42
204720_s_at	DNAJC6	2.236082454	2.40×10−52
204105_s_at	NRCAM	2.232778933	1.57×10−38
227892_at	PRKAA2	2.229124796	3.47×10−44
227510_x_at	MALAT1	2.224576694	5.41×10−17
201468_s_at	NQO1	2.209490964	2.70×10−26
205110_s_at	FGF13	2.207535230	1.23×10−32
223381_at	NUF2	2.186987390	5.66×10−56
205476_at	CCL20	2.181812965	1.34×10−20
203755_at	BUB1B	2.178984131	2.76×10−50
218755_at	KIF20A	2.153373571	8.72×10−52
231265_at	COX7B2	2.150792051	5.69×10−41
221558_s_at	LEF1	2.122850025	2.29×10−38
225612_s_at	B3GNT5	2.117997087	7.16×10−31

Table III.

Top 50 downregulated differentially expressed genes.

Probe ID	Gene	Fold change	Adjusted P-value
220491_at	HAMP	−4.983632921	4.89×10−60
205866_at	FCN3	−4.729403108	1.63×10−96
222484_s_at	CXCL14	−4.664513026	2.76×10−105
205984_at	CRHBP	−4.504769969	5.41×10−83
207804_s_at	FCN2	−4.498217224	1.69×10−97
207201_s_at	SLC22A1	−4.496482306	9.93×10−68
220496_at	CLEC1B	−4.489579549	1.13×10−112
217546_at	MT1M	−4.427062929	4.20×10−64
230478_at	OIT3	−4.401727477	1.29×10−106
1559573_at	LINC01093	−4.372165201	7.87×10−89
223699_at	CNDP1	−4.185131569	2.30×10−73
229476_s_at	THRSP	−4.124366630	2.88×10−56
1559065_a_at	CLEC4G	−4.114961753	1.36×10−99
206354_at	SLCO1B3	−4.081557901	3.55×10−63
207102_at	AKR1D1	−4.020716573	6.22×10−69
1564706_s_at	GLS2	−4.018224948	3.74×10−70
207608_x_at	CYP1A2	−4.013090089	7.43×10−78
206727_at	C9	−3.961247256	9.82×10−48
209687_at	CXCL12	−3.885309405	1.16×10−60
219014_at	PLAC8	−3.774128748	3.42×10−72
207995_s_at	CLEC4M	−3.732362194	5.50×10−88
231678_s_at	ADH4	−3.682472263	1.77×10−48
205554_s_at	DNASE1L3	−3.666231016	1.29×10−68
220801_s_at	HAO2	−3.663835833	4.99×10−76
211896_s_at	DCN	−3.625520940	1.19×10−46
220432_s_at	CYP39A1	−3.616388369	2.51×10−64
220116_at	KCNN2	−3.587327181	4.31×10−57
205819_at	MARCO	−3.569710716	5.45×10−76
202992_at	C7	−3.473352727	3.89×10−43
230135_at	HHIP	−3.458281369	1.29×10−91
210328_at	GNMT	−3.412414732	1.25×10−52
213629_x_at	MT1F	−3.412271922	3.79×10−72
214478_at	SPP2	−3.405396399	1.74×10−74
205225_at	ESR1	−3.355285624	2.09×10−74
237350_at	TTC36	−3.334397580	7.78×10−84
214320_x_at	CYP2A6	−3.332114585	3.55×10−54
219954_s_at	GBA3	−3.304920162	1.98×10−59
207262_at	APOF	−3.287687923	1.09×10−72
214621_at	GYS2	−3.272101039	8.61×10−56
206797_at	NAT2	−3.270355860	8.59×10−76
242817_at	PGLYRP2	−3.225627679	2.41×10−53
205498_at	GHR	−3.223775643	1.20×10−66
237390_at	ADRA1A	−3.213027372	8.85×10−54
204704_s_at	ALDOB	−3.208187451	3.84×10−55
209301_at	CA2	−3.203130234	1.38×10−58
206172_at	IL13RA2	−3.185151095	1.11×10−52
206210_s_at	CETP	−3.179352566	1.73×10−66
204428_s_at	LCAT	−3.145019681	1.47×10−76
205363_at	BBOX1	−3.135951206	2.16×10−42
208147_s_at	CYP2C8	−3.115222728	8.68×10−61

Figure 3.

Heatmap of the expression profiles for the top 50 upregulated and downregulated DEGs in different subsets. The expression profiles for the top 50 upregulated and downregulated DEGs in (A) GSE17548 and (B) GSE84402. DEGs, differentially expressed genes. The colours represent the expression level of the genes, and the higher the expression level, the darker the colour: red, upregulated; green, downregulated.

GO functional and KEGG pathway enrichment analyses of the upregulated and downregulated DEGs

In order to understand the biological implications of the identified DEGs in HBV-positive HCC tissues, GO functional and KEGG pathway enrichment analyses of the identified DEGs were performed. The GO terms of the up- and downregulated DEGs are presented in Fig. 4.

Figure 4.

GO terms of the DEGs. Each Biological Process, Cellular Component and Molecular Function category is represented by red, blue and green bars, respectively. The height of the bar represents the number of DEGs observed in the category. The GO terms of (A) upregulated genes and (B) downregulated genes. DEGs, differentially expressed genes; GO, Gene Ontology.

In the GO biological process category, upregulated DEGs were closely associated with the ‘biological regulation’ and ‘metabolic process’ terms, whereas the downregulated DEGs were closely associated with the ‘metabolic process’ and ‘biological regulation’ terms.

In the GO cellular component category, upregulated DEGs were closely associated with the ‘nucleus’ and ‘membrane’ terms, whereas the downregulated DEGs were closely associated with the ‘membrane’ and ‘vesicle’ terms.

In the GO molecular category, upregulated DEGs were closely associated with the ‘protein binding’ and ‘nucleic acid binding’ terms, whereas the downregulated DEGs were closely associated with the ‘protein binding’ and ‘ion binding’ terms.

In addition, the top 10 enriched KEGG pathway terms of the up- and downregulated genes are provided in Tables IV and V, respectively. The upregulated DEGs were primarily associated with the ‘cell cycle’, whereas the downregulated DEGs were primarily associated with the ‘metabolic pathways’.

Table IV.

Top 10 enriched KEGG pathway terms of upregulated differentially expressed genes.

KEGG ID	KEGG pathway	No. ofss genes	P-value
hsa04110	Cell cycle	24	2.51×10−14
hsa05222	Small cell lung cancer	13	5.70×10−07
hsa05200	Pathways in cancer	28	4.02×10−06
hsa04512	ECM-receptor interaction	11	1.42×10−05
hsa04115	p53 signaling pathway	10	1.76×10−05
hsa04151	PI3K-Akt signaling pathwayss	21	4.69×10−04
hsa04510	Focal adhesion	15	4.89×10−04
hsa05146	Amoebiasis	9	1.74×10−03
hsa05213	Endometrial cancer	6	2.97×10−03
hsa05218	Melanoma	7	3.37×10−03

KEGG, Kyoto Encyclopedia of Genes and Genomes; ECM, extracellular matrix.

Table V.

Top 10 enriched KEGG pathway terms of downregulated differentially expressed genes.

KEGG ID	KEGG pathway	No. of genes	P-value
hsa01100	Metabolic pathways	185	<0.01
hsa04610	Complement and coagulation cascades	33	<0.01
hsa05204	Chemical carcinogenesis	27	5.49×10−12
hsa01200	Carbon metabolism	30	2.04×10−10
hsa00260	Glycine, serine and threonine metabolism	17	5.18×10−10
hsa00380	Tryptophan metabolism	17	5.18×10−10
hsa05150	Staphylococcus aureus infection	20	6.30×10−10
hsa00830	Retinol metabolism	21	1.95×10−09
hsa00071	Fatty acid degradation	17	3.05×10−09
hsa00250	Alanine, aspartate and glutamate metabolism	15	4.8×10−09

KEGG, Kyoto Encyclopedia of Genes and Genomes.

Notably, there were six genes with log2-fold change >2 in DEGs enriched in the ‘cell cycle’ pathway: SFN, BUB1B, TTK, CCNB1, CDK1 and CDC20. Differential expression analysis was performed on the aforementioned six genes in non-HBV tissues from two independent HCC datasets (TCGA and GSE76427) on different platforms (TCGA, Illumina RNA Sequencing; GSE76427, Illumina HumanHT-12 V4.0 expression beadchip). None of these genes had a log2-fold change >2 (Table VI), which demonstrates that the high expression of these six DEGs in HBV-associated HCC is more significant compared with non-HBV HCC.

Table VI.

Differential expression of six differentially expressed genes in different datasets.

	Large HBV cohort		GSE76427		TCGA
Gene	Fold change	Adj. P-value	Fold change	Adj. P-value	Fold change	Adj. P-value
CCNB1	2.403923	1.64×10−56	0.587062	1.36×10−11	0.607121	1.92×10−11
SFN	2.365920	9.97×10−36	0.880437	3.36×10−03	0.676444	2.14×10−05
CDK1	2.271576	9.42×10−49	0.802341	5.35×10−11	0.769405	1.93×10−11
BUB1B	2.178984	2.76×10−50	1.121998	5.27×10−12	0.234839	9.11×10−07
TTK	2.066772	2.45×10−47	1.587680	3.34×10−13	0.791737	2.17×10−12
CDC20	2.012448	7.80×10−46	1.073711	8.29×10−09	1.712167	4.21×10−10

HBV, hepatitis B virus; Adj., adjusted; TCGA, The Cancer Genome Atlas.

PPI network analysis of the DEGs

To further understand the biological meaning of the DEGs identified by the top KEGG pathways at the protein level, two PPI networks for the proteins encoded by the DEGs in the top pathways were constructed. The PPI network of the ‘cell cycle’ consisted of 24 nodes and 85 edges, whereas the PPI network of the ‘metabolic pathways’ consisted of 184 nodes and 566 edges (Fig. 5).

Figure 5.

PPI networks of the DEGs. PPI network of DEGs in (A) ‘Cell cycle’ and (B) in ‘Metabolic pathways’. PPI, protein-protein interaction; DEGs, differentially expressed genes.

Discussion

The present study specifically focused on HBV-infected patients, which is different from the previous studies on HCC regardless of etiology (19,27). A total of 682 upregulated DEGs and 1,252 downregulated DEGs were identified in HCC tissues compared with nontumor hepatic tissues in 321 HBV-positive samples. KEGG analyses demonstrated that the upregulated DEGs were enriched in signaling pathways such as the cell cycle, p53 signaling pathway and extracellular matrix-receptor interaction. A previous study showed that HBV infection deregulates the cell cycle pathway (28). Notably, there were 6 genes with a log2-fold change >2 among the DEGs enriched in the ‘cell cycle’ pathway: SFN, BUB1B, TTK, CCNB1, CDK1 and CDC20.

SFN (14-3-3σ) protein is a member of the 14-3-3 superfamily (29). SFN has been found to play a key role in various vital regulatory processes, such as cell cycle regulation and signaling pathways (30). In a previous study, high expression of SFN was detected in HCC tissues but not in adjacent nontumor tissues, which indicated an association between SFN and HCC (31). In another study, SFN exhibited high diagnostic accuracy in the differentiation of HCC from nontumorous hepatocytes (32).

BUB1B (encoding BUBR1) is an important component in the SAC protein family, which has been found to be involved in several forms of human cancer, such as lung cancer and breast cancer (33,34). However, the contradiction of BUB1B expression in cancer cells remains controversial. Low expression of BUB1B is associated with the poor survival of patients with colon adenocarcinomas and lung cancer, however overexpression of BUB1B contributes to the progression and recurrence of gastric cancer and bladder cancer (35). However, several studies showed that the overexpression of BUB1B is associated with worse prognosis in patients with HCC (36,37).

TTK, a dual-specific protein kinase participating in the p53 pathway, has been found to be involved in several cancer types by modulating the mitotic checkpoint (38). A previous study by Miao et al (39), regarding HBV-associated HCC, reported TTK as a promising prognostic marker of HCC. TTK alone can accurately predict the recurrence rate and recurrence time. These findings on TTK drew interest and resulted in further studies on cancer (40–42). The results of the present study supported the conclusion of the study by Miao et al (39).

CCNB1 plays an integral role in regulating the G2/M transition in the cell cycle (43). Several studies have found an elevated expression of CCNB1 in different cancer types, such as breast cancer (44), non-small cell carcinoma (45) and gastric cancer (46). In a previous study, CCNB1 was reported as an independent risk factor of recurrence in patients with HBV-associated HCC following surgery (47). However, it is still unclear how CCNB1 contributes to oncogenesis and tumor progression.

CDK1 is required for the role of CCNB1 in the G2/M transition and mitosis resumption (48). Cheng et al (49) conducted in vitro experiments, which demonstrated that HBV could activate the CCNB1-CDK1 kinase in HCC cells. In other studies, CCNB1 and CDK1 were found to be upregulated in the HCC tissues of HBV-positive patients (50). Moreover, overexpression of these two genes is associated with poor prognosis. CDK1 was considered important as CCNB1, since it could affect both overall survival and recurrence-free survival of HBV-positive patients with HCC (51).

CDC20 functions as a regulatory protein that interacts with several other proteins at multiple points in the cell cycle. Chae et al (52) demonstrated that HBV-infection could attenuate the association between BubR1 and CDC20, thus preventing CDC20 from performing its original function, which provided a novel view on the development of HBV-associated HCC.

The high expression of the six DEGs was more significant in HBV-associated HCC than in non-HBV HCC, and was validated in two independent HCC datasets. In future studies, clinical HCC samples should be collected in order to verify that these genes are affected by HBV infection.

The downregulated DEGs were enriched in signaling pathways such as ‘carbon metabolism’, ‘glycine, serine and threonine metabolism’, ‘tryptophan metabolism’, ‘retinol metabolism’ and ‘alanine, aspartate and glutamate metabolism’. A previous study reported that HBV-infection could induce alterations in metabolic signaling pathways. The consequences may alter normal hepatocyte metabolism, thus contributing to the progression of HBV-associated carcinogenesis (53).

In conclusion, the present study identified several DEGs in HCC tissues compared with nontumor tissues from HBV-infected patients, based on a large cohort. Based on the DEGs, several key pathways were identified. The interactions of the DEGs in the pathways were also presented by PPI networks. Some results were consistent with previous studies (39,50). Furthermore, the present study provides new insights into the specific etiology of HCC and molecular mechanisms for the transformation of nontumor hepatic tissues into HCC tissues, in patients with a history of HBV infection. Importantly, these results may provide several potential therapeutic targets for targeted therapy in these patients, which could aid early diagnosis and treatment of HCC.