Characteristic mutations in hepatitis C virus core gene related to the occurrence of hepatocellular carcinoma

Abstract

Chronic hepatitis C virus (HCV) infection often results in hepatocellular carcinoma (HCC). Previous studies have shown that there might be some characteristic mutations in the core region of HCV related to HCC. Thus, we downloaded and analyzed HCV genotype 1b core gene sequences from HCV databases online to identify them. Based on the information of the sequences, 63 from patients with HCC and 188 from non‐HCC were enrolled into our analysis. Then, the nucleotides at each position were compared by χ²‐test between the two groups, and 24 polymorphisms were found to be associated with HCC. Further analysis of these 24 polymorphisms by logistic regression indicated that eight were significantly related to the increased HCC risk: A028C, G209A, C219U/A, U264C, A271C/U, C378U/A, G435A/C, and G481A. Moreover, U303C/A was associated with the decreased HCC risk. These mutations could bring about four amino acid substitutions: K10Q, R70Q, M91L, and G161S. In conclusion, eight characteristic mutations in the HCV‐1b core gene related to the occurrence of HCC were identified. The structural and functional alterations of core protein due to these mutations and the relationship with the occurrence of HCC need to be further studied. (Cancer Sci 2009; 100: 2465–2468)

Hepatitis C virus (HCV) is a major cause of chronic hepatitis worldwide and a major risk factor for hepatocellular carcinoma (HCC). More than 70% of patients with HCC in Japan are infected with HCV, and HCC has become the major cause of death in patients chronically infected with HCV.^{( 1} , ² , ^{3 )} In Asia, especially in Japan, HCV‐1b is the most prevalent genotype and is more likely to develop into HCC than other genotypes.^{( 4} , ^{5 )} Although there are many published data about the correlation between HCV infection and HCC development, details of hepatocarcinogenesis by HCV remain unclear. HCV core protein is implicated in hepatocarcinogenesis for its ability to modulate cellular gene transcription and protein expression, intracellular signal transduction, cell proliferation and apoptosis.^{( 6} , ⁷ , ⁸ , ^{9 )}

Being an RNA virus, the HCV genome exhibits a considerable degree of sequence variation. Studies have proved that certain kinds of mutations may lead to functional changes of the virus, such as the resistance to antiviral treatments and the association with HCC.^{( 10} , ^{11 )} Furthermore, there are more mutations in the core region in patients with HCC than that in those without HCC. Moreover, the rate of nucleotide substitutions in the core gene is significantly greater for isolates from HCC patients compared to those from individuals with chronic hepatitis.^{( 12 )} It was also found that there was a significantly higher variability within the core region of tumor tissue isolates than that of isolates from non‐tumor tissue. Mutant sequence diversity ranged from silent mutations, as well as amino acid substitutions, to appearance of in‐frame stop codons and deletions leading to frame‐shifts. In contrast, the variability of the NS5 region sequences between isolates from tumor and non‐tumor tissue was not significantly different.^{( 13 )} Recently, a report indicated that amino acid substitutions in the HCV core region are important predictor of hepatocarcinogenesis.^{( 11 )} Even in patients without HCC, the substitution of amino acid 70 in the hepatitis C virus core region of genotype 1b is an important predictor of elevated alpha‐fetoprotein.^{( 14 )}

Based on the data above, we suppose that there might be some characteristic mutations in the HCV‐1b core gene related to the occurrence of HCC. Because we can find thousands of HCV sequences from databases online, we have tried to identify these mutations by analyzing such sequences in this study.

Materials and Methods

Collection of HCV core gene sequences  HCV core gene sequences were downloaded from the following databases: European HCV database (euHCVdb), HCV Databases from Los Alamos National Laboratory, HCV Database (HCVdb) from the Viral Bioinformatics Resource Center (VBRC), and National Center for Biotechnology Information (NCBI).

Inclusion criteria  All the sequences enrolled into our analysis needed to comply with the following inclusion criteria: (i) the diagnosis of patients could be confirmed; (ii) the genotype of HCV could be confirmed and samples were isolated from sera of patients infected with HCV‐1b; and (iii) the core gene sequence was of full length.

Exclusion criteria  Sequences were excluded if: (i) the diagnosis of patients or the genotype of HCV was unknown; (ii) samples came from liver, other tissues, ascites, or pooled sera; (iii) patients were co‐infected with hepatitis B virus or human immunodeficiency virus; (iv) patients had ever been treated with interferon (IFN); (v) patients had ever undergone orthotopic liver transplantation (OLT); (vi) samples came from cell culture, animal experiments, chimeric DNA, or fusion protein; (vii) sequences came from repeated samples; and (viii) existence of deletion mutations or insertion mutations.

Statistical analysis  According to the diagnosis of the cases, sequences enrolled were divided into two groups: group HCC and group non‐HCC. The first base of the start codon (AUG) was designated position 1. All the 573 nucleotides of the core region were compared respectively between group HCC and non‐HCC. All the 573 nucleotides of the core region were compared respectively between group HCC and non‐HCC by Pearson χ²‐test to find out significantly different mutations (when there were cells that the expected count was less than 5 or 1, we chose different χ²‐test methods – Continuity correction or Fisher's exact test). Then, the significant bases were further analyzed by logistic regression to identify the mutations significantly related to the occurrence of HCC. A P‐value of <0.05 was considered statistically significant.

Results and Discussion

Enrollment of sequences  We downloaded sequences of HCV core region from HCV databases online, mainly from euHCVdb. A total of 2841 sequences of HCV‐RNA with full‐length core region were downloaded. Then, we checked the information provided by the databases and the published papers so that we could get the details about these sequences such as the diagnoses of patients, genotypes of HCV, the origin of samples, and so on. Finally, we picked out the sequences that complied with the inclusion and exclusion criteria and enrolled them into our analysis. As a result, from 2841 sequences, we found 1336 sequences of genotype 1b. Of these, 547 sequences were from identified samples and the others lacked information that identified the types of the samples. Among the 547 sequences with definite origin, 292 were from sera and the others (228 from livers, 20 from cultured cells, seven from plasmids) were ruled out. Of the 292 sequences derived from sera, seven were from animal experiments, 24 were from repeated cases at different time points, and four and six were from cases that had undergone OLT and IFN treatments, respectively. Finally, 251 sequences complied with the inclusion and exclusion criteria and were enrolled into our analysis. Sixty‐three were from patients with HCC and 188 were from HCV carriers or patients with acute or chronic hepatitis without HCC. For the 251 sequences, 73.7% (185/251) were directly sequenced and 26.3% (66/251) represented one clone of available clones. Thus, it is possible that about a quarter of our data could contain mutations due to PCR error. However, the sequences were collected from 42 different studies and the statistical analysis may have overcome this problem.

Identification of the wild‐type nucleotides  For the reason that there are frequent mutations of the HCV sequence, it is impossible to know the real wild‐type nucleotides. It might be reasonable to identify the wild‐type nucleotide based on the consensus of a large number of sequences. In our analysis, we defined the consensus of the 1336 HCV‐1b sequences as the wild type, except the position of nucleotide (nt) 209. Because 209G was treated as wild type in almost all the published papers and it was also the consensus type of 2841 all genotypes sequences, we still defined “G” as the wild type even if “A” was the consensus.

Significantly different mutations between group HCC and non‐HCC  By performing the χ²‐test or Fisher’s exact test, we compared all the 573 nucleotides of the core gene between group HCC and non‐HCC. A total of 24 nucleotide mutations were found to be associated with occurrence of HCC. Of the 24 nucleotide mutations, 21 increased the risk of HCC while three decreased the risk of HCC (Table 1).

Table 1.

 Significant nucleotide mutations found by χ²‐tests

Gene position	Types of Nucleotide			Percentage of mutant				Amino acid			P‐values
	Wild		Mutant	Non‐HCC		HCC		Wild type		Mutant
nt 009	G	→	A/U	20.2	→	68.3		Thr		Thr	<0.001*
nt 028	A	→	C	1.1	→	7.9		Lys	→	Gln	0.015†
nt 039	U	→	G/A	0	→	4.8		Arg		Arg	0.015‡
nt 048	C	→	G/U	0	→	4.8		Asn	→	Lys/Asn	0.015‡
nt 066	C	→	U	14.4	→	30.2		Val		Val	0.005*
nt 078	C	→	U	28.7	→	3.2	↓	Gly		Gly	<0.001*
nt 116	G	→	C	0	→	7.9		Arg	→	Pro	0.001†
nt 209	G	→	A	28.7	→	52.4		Arg	→	Gln	0.001*
nt 219	C	→	U/A	7.4	→	23.8		Gly		Gly	<0.001*
nt 264	U	→	C	12.2	→	36.5		Asn		Asn	<0.001*
nt 270	C	→	U	8	→	33.3		Gly		Gly	<0.001*
nt 271	A	→	C/U	21.8	→	65.1		Met	→	Leu	<0.001*
nt 303	U	→	C/A	64.9	→	47.6	↓	Arg		Arg	0.015*
nt 309	U	→	C	22.3	→	39.7		Ser		Ser	0.007*
nt 378	C	→	U/A	16.5	→	34.9		Leu		Leu	0.002*
nt 417	C	→	U	5.9	→	20.6		Leu		Leu	0.001*
nt 435	G	→	A/C	10.1	→	34.9		Gly		Gly	<0.001*
nt 446	G	→	A	0.5	→	7.9		Arg	→	Lys	0.004†
nt 456	G	→	A	24.5	→	47.6		Ala		Ala	0.001*
nt 462	C	→	U	25	→	49.2		Gly		Gly	<0.001*
nt 471	U	→	C/G	38.3	→	20.6	↓	Val		Val	0.01*
nt 472	C	→	G/U	2.1	→	9.5		Leu	→	Val/Leu	0.026†
nt 481	G	→	A	2.7	→	20.6		Gly	→	Ser	<0.001†
nt 549	C	→	U	14.4	→	30.2		Ser		Ser	0.005*

*Pearson χ²‐tests; †continuity correction; ‡Fisher’s exact test.

Gene polymorphisms and amino acid substitutions associated with HCC  The 24 significant nucleotide mutations found by χ²‐tests were further analyzed by logistic regression. Eight polymorphisms were significantly related to the increased HCC risk: A028C, G209A, C219U/A, U264C, A271C/U, C378U/A, G435A/C, and G481A; and U303C/A was significantly associated with decreased HCC risk (Table 2). Furthermore, four polymorphisms brought about amino acid substitutions which might change the structure and functions of core protein and lead to HCC. For the other five synonymous mutations, it is possible that they might lead to amino acid substitutions of ARFP (alternative ribosomal frame shift protein) which has been reported to be able to affect cell proliferation and apoptosis^{( 15 )} (Table 3). However, HCC is thought to be a consequence of long‐term infection of HCV and chronic hepatitis. Although the gene polymorphisms and amino acid substitutions in the HCV core region identified in this analysis were proven to be significantly associated with HCC, whether they are “causes” or “effects” of HCC remains to be determined.

Table 2.

 Significant nucleotide mutations identified in logistic regression

Gene position	Mutation types	Odds ratio	95% CI		P‐values
			Lower	Upper
nt 028	A→C	14.65	1.31	163.51	0.029
nt 209	G→A	3.30	1.40	7.78	0.006
nt 219	C→U/A	4.37	1.49	12.77	0.007
nt 264	U→C	4.13	1.62	10.52	0.003
nt 271	A→C/U	5.50	2.26	13.37	<0.001
nt 303	U→C/A	0.42	0.18	0.98	0.045
nt 378	C→U/A	2.69	1.08	6.73	0.034
nt 435	G→A/C	4.16	1.61	10.72	0.003
nt 481	G→A	10.484	2.77	39.62	0.001

Nagelkerke R ² = 0.572.

Table 3.

 Amino acid substitutions of core protein and ARFP due to nucleotide mutations

Codon	Gene position	Amino acid position	Nucleotide			Amino acid
			Wild type		Mutant	Wild type		Mutant
Normal	nt 028	aa 010	AAA	→	CAA	Lys	→	Gln
	nt 209	aa 070	CGG	→	CAG	Arg	→	Gln
	nt 219	aa 073	GGC	→	GGU/A	Gly		Gly
	nt 264	aa 088	AAU	→	AAC	Asn		Asn
	nt 271	aa 091	AUG	→	C/UUG	Met	→	Leu
	nt 303	aa 101	CGU	→	CGC/A	Arg		Arg
	nt 378	aa 126	CUC	→	CUU/A	Leu		Leu
	nt 435	aa 145	GGG	→	GGA/C	Gly		Gly
	nt 481	aa 161	GGC	→	AGC	Gly	→	Ser
+1 Codon	nt 028	aa 010	GAA	→	GAC	Glu	→	Asp
	nt 209	aa 070	GGC	→	AGC	Gly	→	Ser
	nt 219	aa 073	GCA	→	GU/AA	Ala	→	Val/Glu
	nt 264	aa 088	AUG	→	ACG	Met	→	Thr
	nt 271	aa 091	GCA	→	GCC/U	Ala		Ala
	nt 303	aa 101	GUG	→	GC/AG	Val	→	Ala/Glu
	nt 378	aa 126	UCA	→	UU/AA	Ser	→	Leu/Stop
	nt 435	aa 145	GGG	→	GA/CG	Gly	→	Glu/Ala
	nt 481	aa 161	ACG	→	ACA	Thr		Thr
−1 Codon	nt 028	aa 010	AAA	→	ACA	Lys	→	Thr
	nt 209	aa 070	CCG	→	CCA	Pro		Pro
	nt 219	aa 073	CAG	→	T/AAG	Gln	→	Stop/Lys
	nt 264	aa 088	UGA	→	CGA	Stop	→	Arg
	nt 271	aa 091	CAU	→	CC/UU	His	→	Pro/Leu
	nt 303	aa 101	UGG	→	C/AGG	Trp	→	Arg
	nt 378	aa 126	CAC	→	U/AAC	His	→	Tyr/Asn
	nt 435	aa 145	GGG	→	A/CGG	Gly	→	Arg
	nt 481	aa 161	CGG	→	CAG	Arg	→	Gln

ARFP, alternative ribosomal frame shift protein.

Recently, a similar study^{( 16 )} determined four amino acid substitutions associated with increased HCC risk, and three of them were the same as our results (G209A, A271C/U, and G481A). Moreover, up to now, the G209A polymorphism has been thought to be related with both IFN/Ribavirin (RBV) treatment resistance and HCC.^{( 10} , ^{11 )} The relationship between them is an interesting question. We guess that apoptosis might make an important function in this process. Apoptosis is considered to be a common pathway of virus clearance. HCV core protein could suppress apoptosis and escape this clearance mechanism.^{( 9 )} On the other hand, inhibition of apoptosis might be able to make the genetically damaged hepatocytes survive and lead to neoplastic transformation. Whether the amino acid substitution of R70Q could enhance these effects is worth studying.

Alteration of the RNA and protein secondary structures of the HCV‐1b core gene  Gene mutations of A271C/U, G435A/C, and G481A, were found to be able to cause the changes of the RNA secondary structure of HCV‐1b core gene using RNAdraw software^{( 17 )} (Karolinska Institute, Stockholm, Sweden). Furthermore, Chou–Fasman protein secondary prediction with the gene analysis software Genetyx (Genetyx, Tokyo, Japan) showed that A028C might turn the coil structure of AA8‐12 into β‐sheet and G209A might turn the coil stucture of AA71 into α‐helix (Fig. 1). The results indicated that the point‐mutations of the core gene might change the secondary structure of not only RNA but also protein. As a result, the functions of both RNA and protein of the core region, such as interaction with other DNA/RNA or proteins (lymphotoxin B receptor, heterogeneous‐nuclear ribonucleoprotein, tumor necrosis factor receptor 1, etc.),^{( 18} , ¹⁹ , ^{20 )} might change and lead to HCC.

Figure 1.

 Secondary structure of wild‐type (a) and mutant (b) hepatitis C virus (HCV)‐1b core protein predicted by Chou–Fasman analysis with Genetyx. The straight, bent, looped, and zigzag lines represent coil, β‐sheet, α‐helix, and turn structures, respectively. The numbers indicate the amino acid positions. A028C might turn the coil stucture of AA8‐12 into β‐sheet and G209A might turn the coil stucture of AA71 into α‐helix. (The secondary structure of the wild type was predicted based on the following consensus amino acid sequence of 1336 HCV‐1b core, except 70R: MSTNPKPQRKTKRNTNRRPQDVKFPGGGQIVGGVYLLPRRGPRLGVRATRKTSERSQPRGRRQPIPKARRPEGRAWAQPGYPWPLYGNEGMGWAGWLLSPRGSRPSWGPTDPRRRSRNLGKVIDTLTCGFADLMGYIPLVGAPLGGAARALAHGVRVLEDGVNYATGNLPGCSFSIFLLALLSCLTIPASA.)

In conclusion, eight characteristic mutations in the HCV‐1b core gene related to the occurrence of HCC were identified: A028C, G209A, C219U/A, U264C, A271C/U, C378U/A, G435A/C, and G481A. Meanwhile, U303C/A was identified to be significantly associated with the decreased HCC risk. These mutations could bring about four amino acid substitutions: K10Q, R70Q, M91L, and G161S. The structural and functional alterations of core protein due to these mutations and the relationship with the occurrence of HCC need to be further studied.