Liver International ISSN 1478-3223
VIRAL HEPATITIS
Hepatitis B virus X protein promotes P3 transcript expression of the insulin-like growth factor 2 gene via inducing hypomethylation of P3 promoter in hepatocellular carcinoma Shaohui Tang1*, Wei Hu2*, Jianjun Hu1, Shenglan Wu1, Junfeng Li1, Yuhong Luo3, Mingrong Cao3, Hongke Zhou1 and Xiangwu Jiang1 1 Department of Gastroenterology, The First Affiliated Hospital, Jinan University, Guangzhou, China 2 Department of Internal Medicine, The Hospital of Wuhan University, Wuhan, China 3 Department of General Surgery, The First Affiliated Hospital, Jinan University, Guangzhou, China
Keywords hepatitis B virus X protein – hepatocellular carcinoma – hypomethylation – insulin-like growth factor 2 – promoters Abbreviations AFP, a-foetoprotein; AU, arbitrary units; BNC-HCC, HBV infection-positive/HCV infection-negative HCC; HBV, hepatitis B virus; HBx, hepatitis B virus X protein; HCC, hepatocellular carcinoma; IGF2, insulin-like growth factor 2; NBNC-HCC, both HBV and HCV infection-negative HCC; TEPV, tumour embolus of portal vein. Correspondence Shaohui Tang, Department of Gastroenterology, The First Affiliated Hospital, Jinan University, Guangzhou, China Tel/Fax: +86 020 38688039 e-mail: [email protected] Received 5 September 2013 Accepted 8 January 2014 DOI:10.1111/liv.12469
Abstract Background & Aims: Hepatitis B virus (HBV) X protein (HBx) contributes to hepatocarcinogenesis. The overexpression of transcripts from P3 and P4 promoters of the insulin-like growth factor 2 (IGF2) gene is observed in hepatocellular carcinoma (HCC). Here, we aimed to explore the involvement of HBx in P3-driven mRNA overexpression and underlying epigenetic mechanism. Methods: P3 mRNA, P3 methylation status, HBx mRNA and HBx protein were analysed in human HCC samples with and without HBV infection using quantitative RT-PCR, bisulphite sequencing and Western blotting. The effects of HBx on P3 mRNA expression, and P3 transcriptional activity and methylation were further evaluated in HCC cell lines. Results: P3 mRNA level was higher and P3 methylation level was lower in HBV-positive HCC specimens compared with those of HBV-negative HCC specimens. P3 transcript abundance was positively correlated with HBx expression and negatively correlated with P3 methylation in HCC specimens. The stable expression of HBx upregulated P3 mRNA expression and reduced P3 methylation level in HepG2-HBx cells. The transient expression of HBx stimulated P3 promoter activity and decreased P3 methylation level of P3 promoter-luciferase construct in a dose-dependent manner in HepG2 and Huh-7 cells. Furthermore, HBx mRNA expression was found to be independent predictive factors for both shorter disease-free survival time and shorter overall survival time of HCC patients. Conclusion: HBx may promote IGF2-P3 transcript expression by inducing hypomethylation of P3 promoter and may be associated with an inferior clinical outcome of HBV-related HCC patients. This study provides useful information for understanding the mechanism of HBxmediated HCC.
Chronic hepatitis B virus (HBV) infection has been identified as a major global aetiological factor of human hepatocellular carcinoma (HCC), accounting for 55% of cases worldwide and 80% or more of those in the eastern Pacific region and sub-Saharan Africa (1). The HBV X gene encodes a protein that is termed HBx. HBx is required for the virus infection and has been believed to play a pivotal role in the pathogenesis of HCC (2). It is a multifunctional regulator that promotes cell cycle progression, activates signal transduction pathway, modulates apoptosis and binds to and inhibits the expression of tumour suppressor genes (3). However, *These authors contributed equally to this work.
Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
the exact molecular mechanisms of HBx-induced HCC are incompletely understood. The human insulin-like growth factor 2 (IGF2) is a potent mitogen, which plays an important role in fetal growth and development (4, 5). The human IGF2 gene is located on chromosome 11p15.5, spanning 30 kb in length and contains nine exons and four promoters (P1–P4). The exons 1–6 are noncoding leader exons, of which the exons 1, 4, 5, 6 are preceded by a different promoter, respectively, and the exons 7, 8 and the first part of the exon 9 code for the IGF2 precursor protein. Each of these promoters initiates distinct transcription, yielding a promoter-specific transcript that has a unique 5′-untranslated region. Although varied in length, all of
1
Tang et al.
HBx promotes IGF2 expression in HCC
the promoter-specific transcripts appear to produce finally the same mature protein, a 67-amino acid polypeptide (6–8). The four promoters are activated in a development-dependent and tissue-specific manner. In fetal liver promoters P2, P3 and P4 are active, of which P3 is the most active promoter, and P1 is inactive. In adult liver, however, P1 becomes dominant and produces about 50% of the total IGF2 transcripts, and the activities of P2–P4 are decreased remarkably or lost (9, 10). An elevated expression of IGF2 has been observed in a variety of malignant tumours, including cancers of the breast, colon, stomach, ovary and liver (11–15). There is compelling clinical and experimental evidence that IGF2 is involved in hepatocarcinogenesis (16). IGF2 transgenic mice (20–30 fold increased level in serum IGF2) developed HCC after a long latency (17). Lin et al. reported that there was overexpression of IGF2 in human hepatoma cell lines Huh-7 and HepG2 and that antisense oligonucleotides complementary to IGF2 mRNA reduced both IGF2 mRNA and protein levels in association with decreased cell proliferative activity (18). The reactivation of IGF2 expression was found in primary HCC specimens (19–21). The overexpression of IGF2 in HCC is associated with reexpression of the fetal transcripts driven by P3 and P4 promoters (20, 21). However, the reasons and mechanisms responsible for the reactivation of the two fetal transcripts during hepatocarcinogenesis are not well clarified. The two reports by Park et al. and Su et al. showed that the high expression of IGF2 peptide was observed in all the HBV-positive HCC samples (22) and all the HCC samples with HBx expression (23) tested by means of immunohistochemical staining. Similarly, in a previous study, we found that there was a significantly increased IGF2 expression of the transcripts from fetal P3 and P4 promoters in a large majority of HBV-associated HCC samples (21). These findings suggest that HBV or its product HBx might facilitate IGF2 overexpression originating from the fetal P3 and P4 promoters in hepatic carcinogenesis. There is accumulating evidence suggesting that aberrant DNA methylation of CpG islands located around promoter regions changes gene transcription and is implicated in tumourigenesis (24–26). Because the P3 promoter of IGF2 gene is the most active promoter (10, 21) and its methylation status has not yet been studied in HCC, the present study was done to determine whether HBx can induce P3 promoter hypomethylation and then enhances its transcription expression in human HCC. Material and methods Expression vectors and retrovirus-pBABE-puro-HBx
pCMV-tag2B-HBx plasmid expressing HBx (HBx gene, GenBank accession no. AF223955.1), pCMV-tag2B control blank plasmid and pHBV3.8 plasmid encoding
2
the whole transcript of HBV DNA were kindly provided by Dr Ming Liang (Southern Medical University, Guangzhou, China). pCMV-tag2B-HBp, pCMV-tag2B -HBe, pCMV-tag2B-HBc, pCMV-tag2B-HBpreS1, pCMVtag2B-HBpreS2 and pCMV-tag2B-HBs plasmids were constructed by insertion of the P, pre-C, C, pre-S1, pre-S2 and S gene regions of HBV DNA respectively. pGL3-P3 vector was constructed by insertion of the P3 promoter ( 1251/+152, GenBank accession no. NT_009308) of human IGF-II gene into luciferase reporter pGL3-Basic vector (Promega, Madison, WI, USA). pBABE-puro-HBx plasmid was constructed by insertion of HBx gene into pBABE-puro retrovirus vector (Addgene, Cambridge, MA, USA). pBABE-puroHBx and control vector pBABE-puro were transfected into 293FT package cells to generate retrovirus-pBABEpuro-HBx and retrovirus-pBABE-puro respectively. All constructs were verified by sequencing. Cell lines and tissue specimens
The human HCC cell lines HepG2 (ATCC, HB-8065, Manassas, VA, USA) and Huh-7 (JCRB Cell Bank, Osaka, Japan) were cultured in Dulbecco’s modified Eagle medium (Gibco BRL, Rockville, MD, USA) supplemented with heat-inactivated 10% fetal bovine serum (Gibco BRL). HepG2-HBx cell line stably overexpressing HBx and HepG2-control cell line were established by using retrovirus-pBABE-puro-HBx and retrovirus-pBABE-puro to infect HepG2 cell line respectively. Tumour tissue specimens were obtained from 43 patients with HBV infection-positive/hepatitis C virus (HCV) infection-negative HCC (BNC-HCC) (33 men, 10 women) and nine patients with both HBV and HCV infection-negative HCC (NBNC-HCC) (eight men, one woman) who underwent curative resection at the First Affiliated Hospital of Jinan University (Guangzhou, China) from January 2004 to December 2011. The diagnosis of HCC was made by a pathological examination, and all the patients did not receive any preoperative treatment before admission. The patients were followed up for 5–86 months with a median follow-up time of 28 months. The patients with recurrence were submitted to standard treatments, such as re-resection, ablation or transarterial chemoembolization according to their condition. As control, normal adult liver tissue samples were obtained from six subjects (five cases of hepatic haemangioma and one case of hepatic rupture; five men, one woman). HBV infection was defined as positive detection of serum HBsAg and HBV DNA, and no HCV infection was defined as negative detection of serum anti-HCV and HCV RNA. These tissues were snap frozen and kept at 80°C until use. This study was approved by the medical ethics committee of the First Affiliated Hospital of Jinan University, and written informed consent was obtained from each participant. Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Tang et al.
Semiquantitative RT-PCR and quantitative RT-PCR
Total RNA was extracted from HCC tissue specimens and HCC cell lines using Trizol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. RNA (2 lg) was reversely transcribed to obtain cDNA using 10 units of Reverse Transcriptase XL (AMV) (TaKaRa, Kyoto, Japan). PCR was performed in a 25-ll reaction mixture and PCR products were run on 2% agarose gels. b-actin served as a internal control. Real time quantitative RT-PCR was carried out using Power SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA). Each experiment was repeated thrice and all reactions were carried out in triplicate. The data were analysed according to previous description and the △CT method was used to quantify the results (27). The primers used for the PCRs were as following: P3 transcript (P3 mRNA) sense 5′-ATT ACA CGC TTT CTG TTT CTC TCC-3′ and antisense 5′-AAA TGA GGT CAG CTG TTG TAT CAA G-3′ (172 bp) (28); HBx mRNA sense 5′-GCA CTT CGC TTC ACC TCT-3′ and antisense 5′-TAT GCC TAC AGC CTC CTA-3′ (211 bp) (29); b-actin mRNA (internal control) sense 5′-TTA AGG AGA AGC TGT GCT ACG-3′and antisense 5′-TTG AAG GTA GTT TCG TGG ATG-3′ (205 bp) (GenBank accession no. NM_001101). In vitro Methylation of pGL3-P3 vector
pGL3-P3 plasmid was methylated in vitro by incubating 10 lg of plasmid DNA with 20 units of SssI CpG methyltransferase (New England BioLabs Inc., Ipswich, MA, USA) in the recommended buffer containing 160 lM S-adenosylmethionine for 3 h at 37°C. The above procedure repeated once again and complete methylation was confirmed by observing full protection from HpaII digestion (30). Western blot analysis
Protein extraction from tumour tissues or cultured HCC cells was performed as described previously (31). Equal amounts of protein were subjected to SDS-PAGE followed by transfer of protein to nitrocellulose membranes (Bio-Rad, Hercules, CA, USA). The membranes were first incubated with antibodies against HBx (Abcam, Cambridge, UK) and b-actin (Abcam), then followed by incubation with HRP-conjugated secondary antibodies. The results were visualized using an enhanced chemiluminescence detection system (Amersham Biosciences, Piscataway, NJ, USA). Each experiment was repeated thrice and all reactions were carried out in triplicate. Transient transfection and dual luciferase assay
HepG2 and Huh-7 cells were plated at a density of 2 9 105/well in a 24-well dish and transiently Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
HBx promotes IGF2 expression in HCC
cotransfected with 0.5 lg luciferase reporter vectors (0.455 lg premethylated pGL3-P3 vector and 0.045 lg pRL-TK vector) and pCMV-tag2B-HBx, pCMV-tag2BHBp, pCMV-tag2B-HBe, pCMV-tag2B-HBc, pCMVtag2B-HBpreS1, pCMV-tag2B-HBpreS2, pCMV-tag2BHBs or pCMV-tag2B control plasmid using TransFast transfection reagent (Promega). At 48 h after transfection, the luciferase assay was done according to the manufacturer’s instructions. Each experiment was repeated three times in triplicate. Bisulphite sequencing
Two micrograms DNA underwent sodium bisulphite modification using the CpGenome DNA Modification kit (Serologicals Corp., Norcross, GA, USA) according to the manufacturer’s protocols. The modified DNA samples isolated from HepG2-HBx cells, HepG2-control cells or HCC tissue specimens were subjected to PCR amplification using the sense and antisense primers for the genomic P3 promoter (amplified fragment of 250 bp, P3 amplification region of 540/ 290 bp), which were 5′-TTT TTT GTT GGG GTA GGT G-3′ and 5′-AAA TTC AAA AAC CCC ATC C-3′ respectively. The modified DNA samples isolated from HepG2 cells transiently cotransfected with the above plasmids were subjected to PCR amplification using a first set of primers, and the PCR products were used as templates for a subsequent PCR utilizing nested primers. The outside sense (spanning the junction region between pGL3-Basic vector and P3 promoter) and antisense primers used for pGL3-P3 vector (GenBank accession no. U47295.2 for pGL3-Basic vector, NT_009308 for P3 promoter) were 5′-AAT AGG TTG TTT TTA GTG TAA GTG-3′ and 5′-TAA TTT TTA CCA AAT CAA AAA CCA C-3′ respectively; the nested sense and antisense primers used for pGL3-P3 vector (amplified fragment of 286 bp, P3 amplification region of 1251/ 1003 bp) were 5′-AGT GTA AGT GTA GGT GTT AG-3′ and 5′ACC TAA CCA AAA AAA AAA AAC C-3′ respectively. The above primer pairs for PCR were designed using CpG Ware Primer Design Software (Serologicals). The PCR products for the genomic P3 promoter and the PCR products of the second reaction for pGL3-P3 vector were subcloned into pMD19-T vectors (TaKaRa) for DNA sequencing (Invitrogen, Shanghai, China). The eight positive clones of each sample were sequenced to ascertain the methylation patterns of each CpG locus. Statistical analysis
Categorical data were evaluated by v2 or Fisher’s exact tests, depending on the absolute numbers included in the analysis, quantitative data were analysed by independent sample t-test followed by Mann–Whitney U test, and linear correlation was evaluated by Pearson correlation coefficient. The Kaplan–Meier method and logrank test were used for survival analysis. Overall survival
3
Tang et al.
HBx promotes IGF2 expression in HCC
time was defined as the duration from the date of resection to the date of death from HCC. Disease-free survival time was defined as the duration from the date of resection to the date of tumour recurrence or metastasis. Independent prognostic factors were evaluated by Cox regression analysis. SPSS 13.0 software (SPSS, Inc., Chicago, IL, USA) was used for all statistical analysis. Results were considered statistically significant at P < 0.05. Results Effects of HBx expression on IGF2-P3 mRNA and P3 promoter methylation level in HCC specimens
To determine if P3 mRNA expression is associated with HBx expression, normal adult liver tissue and HCC tissue specimens were analysed. Figure 1A and B shows the upregulation of P3 mRNA level in HBx-positive HCC specimens as compared with that of HBx-negative (A)
HBx-positive (HCC)
HCC specimens and normal adult liver specimens by using semiquantitative RT-PCR and Western blot analysis. Bisulphite sequencing result revealed that the methylation of the 17 CpG sites in the P3 promoter fragment of 540/ 290 bp was lower in one HBx-positive HCC sample (case 10) than that in one HBx-negative HCC sample (case 51) and one normal adult liver sample (case 58) (Fig. 1C). Further quantitative RT-PCR revealed that P3 mRNA level was higher in BNC-HCC specimens than that in NBNC-HCC specimens (Fig. 2A, Tables 1 and S1) and that P3 mRNA abundance was positively correlated with HBx mRNA and protein levels (Fig. 2B, C, Table S1). On the other hand, bisulphite sequencing of the 17 CpG sites in the P3 promoter fragment of 540/ 290 bp indicated that the mean methylation of the 17 CpG sites in BNC-HCC specimens was lower than that in NBNC-HCC specimens (Fig. 2D, Tables 1 and S1) and that P3 mRNA level was negatively correlated with P3 promoter methylation (Fig. 2E, Table S1). These results suggest that HBV infection or HBx-negative (HCC)
HBx-negative (Normal liver)
P3 mRNA HBx mRNA β-actin mRNA HBx protein β-actin protein (B)
Case No. P3 mRNA
HBx mRNA
10
51
58
(C)
+1
IGF2-P3 –540
–290
Unmethylated
IGF2
TSS Methylated
β-actin mRNA HBx protein β-actin protein Case 10 (19.9%)
Case 51 (31.6%)
Case 58 (79.4%)
P = 0.026 P = 0.000 Fig. 1. P3 mRNA and P3 promoter methylation levels in HCC specimens with and without HBx expression and normal adult liver specimens. (A) Expression levels of IGF2-P3 mRNA were evaluated by semiquantitative RT-PCR in five HBx mRNA and protein expression-positive HCC specimens, five HBx mRNA and protein expression-negative HCC specimens and four normal adult liver specimens. b-actin mRNA and protein served as internal controls. (B) Semiquantitative RT-PCR and Western blot results of case 10 (HBx-positive HCC specimen), case 51 (HBxnegative HCC specimen) and case 58 (normal adult liver specimens). (C) Bisulphite sequencing results of methylation levels of P3 promoter fragment ( 540/ 290 bp) from case 10, case 51 and case 58 samples. Each row of circles represents sequencing results of a clone, filled circles are methylated CpG dinucleotides and empty circles are unmethylated CpG dinucleotides. The percentage of methylated CpG dinucleotides is indicated in parentheses. TSS, transcription start site.
4
Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Tang et al.
HBx promotes IGF2 expression in HCC
(D)
P = 0.006
P3 promoter methylation levels (%)
–Δ CT(P3 mRNA level)
(A) –4
–12
BNC–HCC
30
BNC–HCC
NBNC–HCC
(E) –2
–Δ CT(P3 mRNA level)
–Δ CT(P3 mRNA level)
50
NBNC–HCC
(B) γ = 0.739, P = 0.000
–10
–18 –20
–15
–10
–5
–Δ CT(HBx mRNA level)
–Δ CT(P3 mRNA level)
P = 0.014
10
–20
(C)
70
–2 γ = 0.882, P = 0.000
–10
–18 15
35
55
75
P3 promoter methylation(%)
2
γ = 0.732, P = 0.000
–2 –6 –10 –14 –18 –0.5
0
0.5
1
1.5
2
HBx protein level (AU) Fig. 2. Effects of HBx expression on P3 mRNA and P3 promoter methylation levels in HCC specimens. P3 mRNA, and HBx mRNA and protein levels were measured in BNC-HCC and NBNC-HCC specimens. The MCT method was used to quantify quantitative RT-PCR results and the signal densitometric ratio (AU) of HBx vs. b-actin Western blot results was used to quantify HBx protein level. The methylation level of P3 promoter fragment ( 540/ 290 bp) was analysed using bisulphite sequencing. (A) P3 mRNA level in BNC-HCC samples was higher than that in NBNC-HCC samples. (B, C) P3 mRNA abundance was positively correlated with HBx mRNA and protein levels. (D) The methylation level of P3 promoter in BNC-HCC samples was lower than that in NBNC-HCC samples. (E) P3 mRNA abundance and P3 promoter methylation level were negatively correlated. AU, arbitrary units; BNC-HCC, HBV infection-positive/HCV infection-negative HCC; NBNC-HCC, both HBV and HCV infection-negative HCC.
HBx expression may decrease the P3 promoter methylation level and then upregulate P3 mRNA expression at tissue level. HBx promotes the expression of IGF2 gene by inducing DNA hypomethylation of the P3 promoter in HCC cells
To further investigate the possible role of HBx in the regulation of P3 transcript expression, P3 mRNA level and the methylation profile of the 17 CpG sites in the P3 promoter fragment of 540/ 290 bp were analysed in HCC cells by quantitative RT-PCR and bisulphite Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
sequencing respectively. P3 mRNA level in HepG2-HBx cells stably overexpressing HBx was significantly higher than that in HepG2-control cells not expressing HBx (Fig. 3A, B). Meanwhile, the methylation of the 17 CpG sites was lower in HepG2-HBx cells than that in HepG2-control cells (Fig. 3C). The findings indicate that HBx overexpression may promote P3 transcript expression via inducing DNA hypomethylation of IGF2P3 promoter. In order to determine whether the difference in the P3 transcript expression might have an additional explanation between HepG2-HBx cells and HepG2-control
5
Tang et al.
HBx promotes IGF2 expression in HCC
Table 1. Comparisons of clinicopathological data and experimental results between BNC-HCC and NBNC-HCC groups
(A)
HepG2-control HepG2-HBx
HBx protein
Mean age (years) Mean tumour diameter (cm) No. of female patients No. of patients with AFP ≥ 400 lg/L No. of patients with Edmondson grade III– IV No. of patients with TEPV No. of patients with multiple tumour nodularity P3 mRNA levels ( MCT) P3 promoter methylation levels (%)
NBNC-HCC (n = 9)
51.1 ± 12.7 7.1 ± 3.4
60.8 ± 10.4 8.7 ± 4.0
P 0.038 0.226
10 22
1 3
0.717 0.108
33
3
0.030
28
2
0.046
8
2
1.000
9.59 ± 3.22 36.9 ± 15.5
12.97 ± 3.08 52.1 ± 19.1
0.006 0.014
AFP, a-foetoprotein; BNC-HCC, HBV infection-positive/HCV infectionnegative HCC; NBNC-HCC, both HBV and HCV infection-negative HCC; TEPV, tumour embolus of portal vein.
cells, the relative stabilities of P3 mRNAs were evaluated using the in vitro system for studying mRNA stability described by Pei et al. (32). No significant difference was found in the relative stabilities of the two mRNAs (results not shown). These data suggest that the stabilities of them may not be responsible for the discrepancy in the P3 transcript expression from the above two kinds of HCC cells and that HBx may not act at the level of P3 mRNA stability. HBx enhances the activity of IGF2-P3 promoter via inducing DNA hypomethylation in vitro
To further confirm whether HBx can upregulate the activity of IGF2-P3 promoter, in vitro-premethylated P3 promoter-luciferase construct (pGL3-P3) was transiently cotransfected with increasing amounts of HBx expression plasmid (pCMV-tag2B-HBx) or control plasmid (pCMV-tag2B) into HepG2 and Huh-7 cells. Luciferase assays revealed that P3 promoter activity was stimulated by the ectopic expression of HBx in a dosedependent manner in the two cell lines (Fig. 4A). Western blot and RT-PCR analyses indicated that the transient expression of HBx resulted in upregulation of endogenous P3 mRNA level in a dose-dependent manner in the two cell lines (Fig. 4B). Hepatitis B virus has four open reading frames that encode HBx, HBp, HBe, HBc, HBpreS1, HBpreS2 and HBs proteins. To investigate the specificity of HBx in the regulation of P3 promoter activity, the above seven protein expression plasmids (pCMV-tag2B-HBx, pCMV-tag2B-HBp, pCMV-tag2B-HBe, pCMV-tag2B-
6
β-actin protein
(B) P3 mRNA expression level
Parameters
BNC-HCC (n = 43)
4 3
P = 0.000
2 1 0
HepG2-HBx
(C)
+1
IGF2-P3 –540
HepG2-control
–290
Unmethylated
IGF2 TSS
Methylated
HepG2-control (28.7%)
HepG2-HBx (14.7%) P = 0.005
Fig. 3. P3 mRNA and P3 promoter methylation levels in HepG2HBx cell lines stably overexpressing HBx vs. HepG2-control cell lines not expressing HBx. (A) Detection of HBx protein expression in HepG2-HBx cell lines and HepG2-control cell lines by Western blot analysis. (B) P3 mRNA was upregulated in HepG2-HBx cells compared with that in HepG2-control cells detected by quantitative RT-PCR and the results are the average of three independent experiments carried out in triplicate. (C) The methylation level of the 17 CpG sites in the P3 promoter fragment ( 540/ 290 bp) was lower in HepG2-HBx cells than that in HepG2-control cells by bisulphite sequencing analysis. Each row of circles represents sequencing results of a clone, filled circles are methylated CpG dinucleotides and empty circles are unmethylated CpG dinucleotides. The percentage of methylated CpG dinucleotides is indicated in parentheses. TSS, transcription start site.
HBc, pCMV-tag2B-HBpreS1, pCMV-tag2B-HBpreS2 and pCMV-tag2B-HBs) were cotransfected with in vitro-premethylated pGL3-P3 into HepG2 cells. As shown in Figure 4C, HBx significantly enhanced P3 promoter activity, whereas HBp, HBe, HBc, HBpreS1, HBpreS2 and HBs as well as control had no effect on its activity. Next, we investigated the mechanism by which HBx enhances the activity of IGF2-P3 promoter. Bisulphite sequencing was performed to analyse the methylation profile of the 29 CpG sites located within P3 promoter fragment of 1251/ 1003 bp from in vitro-premethyLiver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Tang et al.
HBx promotes IGF2 expression in HCC
(A) Relative luciferase activity
60
HepG2
tion of the 29 CpG sites (Fig. 5). These results suggest that the ectopic expression of HBx may enhance the transcriptional activity of the P3 promoter by lowering DNA methylation level.
Huh-7
50 40 30
*
*
Characteristics and prognostic factors of subjects
20 10
0 Premethylated HBx expression plasmid
+ –
+ +
+ ++
+ +++
Control plasmid
+
–
–
–
(B) – +
HBx expression plasmid Control plasmid
++ +++ – –
+ –
– +
+ –
++ +++ – –
P3 mRNA β-actin mRNA HBx protein
P = 0.000
30 25 20
HepG2
15 10 5
H Bs
H Be H Bc H Bp re S1 H Bp re S2
nt co
H Bp
0
ro l H Bx
(C)
Relative luciferase activity
β-actin protein
Fig. 4. The ectopic expression of HBx increases the activity and transcription of IGF2-P3 promoter in vitro. (A) HepG2 and Huh-7 cells were transiently cotransfected with increasing amounts of HBx expression plasmid or control plasmid and premethylated pGL3-P3 vector. The observed firefly luciferase activity was normalized to Renilla luciferase activity and the results were the average of three independent experiments carried out in triplicate. The transfected quantities of HBx expression plasmid were 0.25 lg (+), 0.5 lg (+ +) and 1.0 lg (+ + +) per 1 9 106 cells. *Indicates that the mean values in HepG2 or Huh-7 cells cotransfected with HBx expression plasmid are significantly different from those of mock-transfected controls (P = 0.000). (B) Expression analyses of P3 mRNA and HBx protein in HepG2 and Huh-7 cells transiently transfected with increasing amounts of HBx expression plasmid or control plasmid. b-actin mRNA and protein served as internal controls. The transiently transfected HBx expression plasmid DNA quantities were as follows: 0.25 lg (+), 0.5 lg (+ +) and 1.0 lg (+ + +) per 1 9 106 cells. (C) In vitro-premethylated pGL3-P3 was cotransfected into HepG2 cells with control, HBx, HBp, HBe, HBc, HBpreS1, HBpreS2 and HBs expression plasmid. The transfected quantities of each plasmid were 0.25 lg. The observed firefly luciferase activity was normalized to Renilla luciferase activity and the results were the average of three independent experiments carried out in triplicate.
lated pGL3-P3 vector transiently cotransfected into HepG2 and Huh-7 cells. The addition of HBx expression plasmids resulted in a dose-dependent demethylaLiver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
The clinicopathological characteristics of the subjects are presented in Table S2. The patients with BNC-HCC were much younger and were involved more often with poor tumour differentiation (Edmondson grade III–IV) and tumour embolus of portal vein (TEPV) compared with those of the patients with NBNC-HCC, whereas there were no differences between the BNC-HCC and NBNC-HCC groups with respect to sex, tumour size, tumour nodularity and a-foetoprotein level (Table 1). Further, HBx mRNA and protein levels, and P3 mRNA level were higher in HCC specimens with poor tumour differentiation and with TEPV than those in HCC specimens with well tumour differentiation and without TEPV, but there were no differences in HBx mRNA and protein levels, and P3 mRNA level between HCC patients with single tumour nodularity and with multiple tumour nodularity (Table 2). These results suggest that HBx and P3 mRNA expression were closely associated with poor tumour differentiation and TEPV of HBV-related HCC patients. Kaplan–Meier method and the log-rank test showed that the patients with BNC-HCC had both shorter disease-free survival time and shorter overall survival time compared with those of the patients with NBNC-HCC (Fig. 6). Univariate Cox regression analysis revealed that a-foetoprotein (AFP), tumour differentiation, TEPV and HBx mRNA expression were significantly associated with both disease-free survival time and overall survival time. Using multivariate Cox regression analysis, poor tumour differentiation and HBx mRNA expression were found to be independent predictive factors for both shorter disease-free survival time and shorter overall survival time of HCC patients (Table 3). Discussion
Evidence has shown that the overexpression of IGF2 plays a key role in the pathogenesis of HCC (16–18, 33–35). In a previous study, we showed that the increased IGF2 expression from the fetal P3 and P4 promoters was observed in most of HCC tissues and could be involved in the transformation of a premalignant liver lesion to HCC (21). However, the mechanisms that may contribute to upregulation of the fetal P3 and P4 transcripts are not well understood. Subsequently, in another study, we further found that the overexpression of IGF2-P4 transcript was closely associated with P4 promoter hypomethylation in HCC tissues (36). In this study, we explored the effect of HBx on IGF2-P3 transcript expression and epigenetic modification-related molecular mechanisms associated
7
Tang et al.
HBx promotes IGF2 expression in HCC
+1
pGL3-P3 promoter –1251
–1003
Luciferase
TSS Huh-7
HepG2 Control (73.7%)
Control (70.3%)
HBx(+) (55.6%)
HBx(+) (50.9%)
HBx(++) (46.1%)
HBx(++) (40.5%)
HBx(+++) (32.3%)
HBx(+++) (27.6%)
Unmethylated
Methylated
Fig. 5. The ectopic expression of HBx lowers the methylation level of IGF2-P3 promoter in vitro. The methylation profile of P3 promoter fragment ( 1251/ 1003 bp) from in vitro-premethylated pGL3-P3 vector transiently cotransfected into HepG2 and Huh-7 cells with increasing amounts of HBx expression plasmid or control plasmid. The transfected quantities of HBx expression plasmid were 0.25 lg (+), 0.5 lg (+ +) and 1.0 lg (+ + +) per 1 9 106 cells. Each row of circles represents sequencing results of a clone, filled circles are methylated CpG dinucleotides, and empty circles are unmethylated CpG dinucleotides. The percentage of methylated CpG dinucleotides is indicated in parentheses. TSS, transcription start site.
Table 2. Correlation of HBx expression and P3 mRNA expression levels with clinicopathological features of HCC Parameters Differentiation degree Poorly differentiated (III–IV) Well differentiated (I–II) TEPV Positive Negative Tumour nodularity Uninodular Multinodular
N
HBx mRNA level ( MCT)
P-value
HBx protein level (AU)
P-value
P3 mRNA levels( MCT)
P-value
36 16
10.98 ± 3.00 14.33 ± 3.17
0.001
0.93 ± 0.50 0.39 ± 0.49
0.001
9.03 ± 2.87 12.77 ± 3.19
0.000
30 22
10.28 ± 2.72 14.37 ± 2.79
0.000
1.05 ± 0.43 0.37 ± 0.44
0.000
8.43 ± 2.76 12.56 ± 2.75
0.000
42 10
11.86 ± 3.35 12.66 ± 3.74
0.507
0.79 ± 0.54 0.64 ± 0.60
0.459
9.92 ± 3.31 11.26 ± 3.82
0.270
AU, arbitrary units; TEPV, tumour embolus of portal vein.
with the expression variation of this transcript in human HCC. Our data showed that the higher expression level of P3 mRNA was detected in the HCC samples with HBV infection compared with that of the HCC samples without HBV infection, and that P3 mRNA level was positively correlated with HBx mRNA and protein levels in all HCC specimens examined. Moreover, we further
8
found that the stable or transient expression of HBx upregulated P3 mRNA level in HCC cells. These in vivo and in vitro findings suggest that HBx may promote IGF2-P3 transcript expression in human HCC. These results are concordant with that reported by Tian et al., who found that HBx could increase the expression of Raf1, a member of the Ras/Raf/MEK/ERK kinase Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Tang et al.
HBx promotes IGF2 expression in HCC
(A)
(B)
P = 0.031
P = 0.038
Fig. 6. Kaplan–Meier curves for disease-free survival and overall survival time in 43 BNC-HCC patients and 9 NBNC-HCC patients. (A) The patients with BNC-HCC had a shorter disease-free survival time than those with NBNC-HCC. (B) The patients BNC-HCC had a shorter overall survival time than those with NBNC-HCC.
Table 3. Univariate and multivariate Cox regression analysis showing disease-free survival and overall survival in 52 patients with HCC Disease-free survival Parameters Univariate Gender M: n = 41 F: n = 11 Age (years)
VIRAL HEPATITIS
Hepatitis B virus X protein promotes P3 transcript expression of the insulin-like growth factor 2 gene via inducing hypomethylation of P3 promoter in hepatocellular carcinoma Shaohui Tang1*, Wei Hu2*, Jianjun Hu1, Shenglan Wu1, Junfeng Li1, Yuhong Luo3, Mingrong Cao3, Hongke Zhou1 and Xiangwu Jiang1 1 Department of Gastroenterology, The First Affiliated Hospital, Jinan University, Guangzhou, China 2 Department of Internal Medicine, The Hospital of Wuhan University, Wuhan, China 3 Department of General Surgery, The First Affiliated Hospital, Jinan University, Guangzhou, China
Keywords hepatitis B virus X protein – hepatocellular carcinoma – hypomethylation – insulin-like growth factor 2 – promoters Abbreviations AFP, a-foetoprotein; AU, arbitrary units; BNC-HCC, HBV infection-positive/HCV infection-negative HCC; HBV, hepatitis B virus; HBx, hepatitis B virus X protein; HCC, hepatocellular carcinoma; IGF2, insulin-like growth factor 2; NBNC-HCC, both HBV and HCV infection-negative HCC; TEPV, tumour embolus of portal vein. Correspondence Shaohui Tang, Department of Gastroenterology, The First Affiliated Hospital, Jinan University, Guangzhou, China Tel/Fax: +86 020 38688039 e-mail: [email protected] Received 5 September 2013 Accepted 8 January 2014 DOI:10.1111/liv.12469
Abstract Background & Aims: Hepatitis B virus (HBV) X protein (HBx) contributes to hepatocarcinogenesis. The overexpression of transcripts from P3 and P4 promoters of the insulin-like growth factor 2 (IGF2) gene is observed in hepatocellular carcinoma (HCC). Here, we aimed to explore the involvement of HBx in P3-driven mRNA overexpression and underlying epigenetic mechanism. Methods: P3 mRNA, P3 methylation status, HBx mRNA and HBx protein were analysed in human HCC samples with and without HBV infection using quantitative RT-PCR, bisulphite sequencing and Western blotting. The effects of HBx on P3 mRNA expression, and P3 transcriptional activity and methylation were further evaluated in HCC cell lines. Results: P3 mRNA level was higher and P3 methylation level was lower in HBV-positive HCC specimens compared with those of HBV-negative HCC specimens. P3 transcript abundance was positively correlated with HBx expression and negatively correlated with P3 methylation in HCC specimens. The stable expression of HBx upregulated P3 mRNA expression and reduced P3 methylation level in HepG2-HBx cells. The transient expression of HBx stimulated P3 promoter activity and decreased P3 methylation level of P3 promoter-luciferase construct in a dose-dependent manner in HepG2 and Huh-7 cells. Furthermore, HBx mRNA expression was found to be independent predictive factors for both shorter disease-free survival time and shorter overall survival time of HCC patients. Conclusion: HBx may promote IGF2-P3 transcript expression by inducing hypomethylation of P3 promoter and may be associated with an inferior clinical outcome of HBV-related HCC patients. This study provides useful information for understanding the mechanism of HBxmediated HCC.
Chronic hepatitis B virus (HBV) infection has been identified as a major global aetiological factor of human hepatocellular carcinoma (HCC), accounting for 55% of cases worldwide and 80% or more of those in the eastern Pacific region and sub-Saharan Africa (1). The HBV X gene encodes a protein that is termed HBx. HBx is required for the virus infection and has been believed to play a pivotal role in the pathogenesis of HCC (2). It is a multifunctional regulator that promotes cell cycle progression, activates signal transduction pathway, modulates apoptosis and binds to and inhibits the expression of tumour suppressor genes (3). However, *These authors contributed equally to this work.
Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
the exact molecular mechanisms of HBx-induced HCC are incompletely understood. The human insulin-like growth factor 2 (IGF2) is a potent mitogen, which plays an important role in fetal growth and development (4, 5). The human IGF2 gene is located on chromosome 11p15.5, spanning 30 kb in length and contains nine exons and four promoters (P1–P4). The exons 1–6 are noncoding leader exons, of which the exons 1, 4, 5, 6 are preceded by a different promoter, respectively, and the exons 7, 8 and the first part of the exon 9 code for the IGF2 precursor protein. Each of these promoters initiates distinct transcription, yielding a promoter-specific transcript that has a unique 5′-untranslated region. Although varied in length, all of
1
Tang et al.
HBx promotes IGF2 expression in HCC
the promoter-specific transcripts appear to produce finally the same mature protein, a 67-amino acid polypeptide (6–8). The four promoters are activated in a development-dependent and tissue-specific manner. In fetal liver promoters P2, P3 and P4 are active, of which P3 is the most active promoter, and P1 is inactive. In adult liver, however, P1 becomes dominant and produces about 50% of the total IGF2 transcripts, and the activities of P2–P4 are decreased remarkably or lost (9, 10). An elevated expression of IGF2 has been observed in a variety of malignant tumours, including cancers of the breast, colon, stomach, ovary and liver (11–15). There is compelling clinical and experimental evidence that IGF2 is involved in hepatocarcinogenesis (16). IGF2 transgenic mice (20–30 fold increased level in serum IGF2) developed HCC after a long latency (17). Lin et al. reported that there was overexpression of IGF2 in human hepatoma cell lines Huh-7 and HepG2 and that antisense oligonucleotides complementary to IGF2 mRNA reduced both IGF2 mRNA and protein levels in association with decreased cell proliferative activity (18). The reactivation of IGF2 expression was found in primary HCC specimens (19–21). The overexpression of IGF2 in HCC is associated with reexpression of the fetal transcripts driven by P3 and P4 promoters (20, 21). However, the reasons and mechanisms responsible for the reactivation of the two fetal transcripts during hepatocarcinogenesis are not well clarified. The two reports by Park et al. and Su et al. showed that the high expression of IGF2 peptide was observed in all the HBV-positive HCC samples (22) and all the HCC samples with HBx expression (23) tested by means of immunohistochemical staining. Similarly, in a previous study, we found that there was a significantly increased IGF2 expression of the transcripts from fetal P3 and P4 promoters in a large majority of HBV-associated HCC samples (21). These findings suggest that HBV or its product HBx might facilitate IGF2 overexpression originating from the fetal P3 and P4 promoters in hepatic carcinogenesis. There is accumulating evidence suggesting that aberrant DNA methylation of CpG islands located around promoter regions changes gene transcription and is implicated in tumourigenesis (24–26). Because the P3 promoter of IGF2 gene is the most active promoter (10, 21) and its methylation status has not yet been studied in HCC, the present study was done to determine whether HBx can induce P3 promoter hypomethylation and then enhances its transcription expression in human HCC. Material and methods Expression vectors and retrovirus-pBABE-puro-HBx
pCMV-tag2B-HBx plasmid expressing HBx (HBx gene, GenBank accession no. AF223955.1), pCMV-tag2B control blank plasmid and pHBV3.8 plasmid encoding
2
the whole transcript of HBV DNA were kindly provided by Dr Ming Liang (Southern Medical University, Guangzhou, China). pCMV-tag2B-HBp, pCMV-tag2B -HBe, pCMV-tag2B-HBc, pCMV-tag2B-HBpreS1, pCMVtag2B-HBpreS2 and pCMV-tag2B-HBs plasmids were constructed by insertion of the P, pre-C, C, pre-S1, pre-S2 and S gene regions of HBV DNA respectively. pGL3-P3 vector was constructed by insertion of the P3 promoter ( 1251/+152, GenBank accession no. NT_009308) of human IGF-II gene into luciferase reporter pGL3-Basic vector (Promega, Madison, WI, USA). pBABE-puro-HBx plasmid was constructed by insertion of HBx gene into pBABE-puro retrovirus vector (Addgene, Cambridge, MA, USA). pBABE-puroHBx and control vector pBABE-puro were transfected into 293FT package cells to generate retrovirus-pBABEpuro-HBx and retrovirus-pBABE-puro respectively. All constructs were verified by sequencing. Cell lines and tissue specimens
The human HCC cell lines HepG2 (ATCC, HB-8065, Manassas, VA, USA) and Huh-7 (JCRB Cell Bank, Osaka, Japan) were cultured in Dulbecco’s modified Eagle medium (Gibco BRL, Rockville, MD, USA) supplemented with heat-inactivated 10% fetal bovine serum (Gibco BRL). HepG2-HBx cell line stably overexpressing HBx and HepG2-control cell line were established by using retrovirus-pBABE-puro-HBx and retrovirus-pBABE-puro to infect HepG2 cell line respectively. Tumour tissue specimens were obtained from 43 patients with HBV infection-positive/hepatitis C virus (HCV) infection-negative HCC (BNC-HCC) (33 men, 10 women) and nine patients with both HBV and HCV infection-negative HCC (NBNC-HCC) (eight men, one woman) who underwent curative resection at the First Affiliated Hospital of Jinan University (Guangzhou, China) from January 2004 to December 2011. The diagnosis of HCC was made by a pathological examination, and all the patients did not receive any preoperative treatment before admission. The patients were followed up for 5–86 months with a median follow-up time of 28 months. The patients with recurrence were submitted to standard treatments, such as re-resection, ablation or transarterial chemoembolization according to their condition. As control, normal adult liver tissue samples were obtained from six subjects (five cases of hepatic haemangioma and one case of hepatic rupture; five men, one woman). HBV infection was defined as positive detection of serum HBsAg and HBV DNA, and no HCV infection was defined as negative detection of serum anti-HCV and HCV RNA. These tissues were snap frozen and kept at 80°C until use. This study was approved by the medical ethics committee of the First Affiliated Hospital of Jinan University, and written informed consent was obtained from each participant. Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Tang et al.
Semiquantitative RT-PCR and quantitative RT-PCR
Total RNA was extracted from HCC tissue specimens and HCC cell lines using Trizol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. RNA (2 lg) was reversely transcribed to obtain cDNA using 10 units of Reverse Transcriptase XL (AMV) (TaKaRa, Kyoto, Japan). PCR was performed in a 25-ll reaction mixture and PCR products were run on 2% agarose gels. b-actin served as a internal control. Real time quantitative RT-PCR was carried out using Power SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA). Each experiment was repeated thrice and all reactions were carried out in triplicate. The data were analysed according to previous description and the △CT method was used to quantify the results (27). The primers used for the PCRs were as following: P3 transcript (P3 mRNA) sense 5′-ATT ACA CGC TTT CTG TTT CTC TCC-3′ and antisense 5′-AAA TGA GGT CAG CTG TTG TAT CAA G-3′ (172 bp) (28); HBx mRNA sense 5′-GCA CTT CGC TTC ACC TCT-3′ and antisense 5′-TAT GCC TAC AGC CTC CTA-3′ (211 bp) (29); b-actin mRNA (internal control) sense 5′-TTA AGG AGA AGC TGT GCT ACG-3′and antisense 5′-TTG AAG GTA GTT TCG TGG ATG-3′ (205 bp) (GenBank accession no. NM_001101). In vitro Methylation of pGL3-P3 vector
pGL3-P3 plasmid was methylated in vitro by incubating 10 lg of plasmid DNA with 20 units of SssI CpG methyltransferase (New England BioLabs Inc., Ipswich, MA, USA) in the recommended buffer containing 160 lM S-adenosylmethionine for 3 h at 37°C. The above procedure repeated once again and complete methylation was confirmed by observing full protection from HpaII digestion (30). Western blot analysis
Protein extraction from tumour tissues or cultured HCC cells was performed as described previously (31). Equal amounts of protein were subjected to SDS-PAGE followed by transfer of protein to nitrocellulose membranes (Bio-Rad, Hercules, CA, USA). The membranes were first incubated with antibodies against HBx (Abcam, Cambridge, UK) and b-actin (Abcam), then followed by incubation with HRP-conjugated secondary antibodies. The results were visualized using an enhanced chemiluminescence detection system (Amersham Biosciences, Piscataway, NJ, USA). Each experiment was repeated thrice and all reactions were carried out in triplicate. Transient transfection and dual luciferase assay
HepG2 and Huh-7 cells were plated at a density of 2 9 105/well in a 24-well dish and transiently Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
HBx promotes IGF2 expression in HCC
cotransfected with 0.5 lg luciferase reporter vectors (0.455 lg premethylated pGL3-P3 vector and 0.045 lg pRL-TK vector) and pCMV-tag2B-HBx, pCMV-tag2BHBp, pCMV-tag2B-HBe, pCMV-tag2B-HBc, pCMVtag2B-HBpreS1, pCMV-tag2B-HBpreS2, pCMV-tag2BHBs or pCMV-tag2B control plasmid using TransFast transfection reagent (Promega). At 48 h after transfection, the luciferase assay was done according to the manufacturer’s instructions. Each experiment was repeated three times in triplicate. Bisulphite sequencing
Two micrograms DNA underwent sodium bisulphite modification using the CpGenome DNA Modification kit (Serologicals Corp., Norcross, GA, USA) according to the manufacturer’s protocols. The modified DNA samples isolated from HepG2-HBx cells, HepG2-control cells or HCC tissue specimens were subjected to PCR amplification using the sense and antisense primers for the genomic P3 promoter (amplified fragment of 250 bp, P3 amplification region of 540/ 290 bp), which were 5′-TTT TTT GTT GGG GTA GGT G-3′ and 5′-AAA TTC AAA AAC CCC ATC C-3′ respectively. The modified DNA samples isolated from HepG2 cells transiently cotransfected with the above plasmids were subjected to PCR amplification using a first set of primers, and the PCR products were used as templates for a subsequent PCR utilizing nested primers. The outside sense (spanning the junction region between pGL3-Basic vector and P3 promoter) and antisense primers used for pGL3-P3 vector (GenBank accession no. U47295.2 for pGL3-Basic vector, NT_009308 for P3 promoter) were 5′-AAT AGG TTG TTT TTA GTG TAA GTG-3′ and 5′-TAA TTT TTA CCA AAT CAA AAA CCA C-3′ respectively; the nested sense and antisense primers used for pGL3-P3 vector (amplified fragment of 286 bp, P3 amplification region of 1251/ 1003 bp) were 5′-AGT GTA AGT GTA GGT GTT AG-3′ and 5′ACC TAA CCA AAA AAA AAA AAC C-3′ respectively. The above primer pairs for PCR were designed using CpG Ware Primer Design Software (Serologicals). The PCR products for the genomic P3 promoter and the PCR products of the second reaction for pGL3-P3 vector were subcloned into pMD19-T vectors (TaKaRa) for DNA sequencing (Invitrogen, Shanghai, China). The eight positive clones of each sample were sequenced to ascertain the methylation patterns of each CpG locus. Statistical analysis
Categorical data were evaluated by v2 or Fisher’s exact tests, depending on the absolute numbers included in the analysis, quantitative data were analysed by independent sample t-test followed by Mann–Whitney U test, and linear correlation was evaluated by Pearson correlation coefficient. The Kaplan–Meier method and logrank test were used for survival analysis. Overall survival
3
Tang et al.
HBx promotes IGF2 expression in HCC
time was defined as the duration from the date of resection to the date of death from HCC. Disease-free survival time was defined as the duration from the date of resection to the date of tumour recurrence or metastasis. Independent prognostic factors were evaluated by Cox regression analysis. SPSS 13.0 software (SPSS, Inc., Chicago, IL, USA) was used for all statistical analysis. Results were considered statistically significant at P < 0.05. Results Effects of HBx expression on IGF2-P3 mRNA and P3 promoter methylation level in HCC specimens
To determine if P3 mRNA expression is associated with HBx expression, normal adult liver tissue and HCC tissue specimens were analysed. Figure 1A and B shows the upregulation of P3 mRNA level in HBx-positive HCC specimens as compared with that of HBx-negative (A)
HBx-positive (HCC)
HCC specimens and normal adult liver specimens by using semiquantitative RT-PCR and Western blot analysis. Bisulphite sequencing result revealed that the methylation of the 17 CpG sites in the P3 promoter fragment of 540/ 290 bp was lower in one HBx-positive HCC sample (case 10) than that in one HBx-negative HCC sample (case 51) and one normal adult liver sample (case 58) (Fig. 1C). Further quantitative RT-PCR revealed that P3 mRNA level was higher in BNC-HCC specimens than that in NBNC-HCC specimens (Fig. 2A, Tables 1 and S1) and that P3 mRNA abundance was positively correlated with HBx mRNA and protein levels (Fig. 2B, C, Table S1). On the other hand, bisulphite sequencing of the 17 CpG sites in the P3 promoter fragment of 540/ 290 bp indicated that the mean methylation of the 17 CpG sites in BNC-HCC specimens was lower than that in NBNC-HCC specimens (Fig. 2D, Tables 1 and S1) and that P3 mRNA level was negatively correlated with P3 promoter methylation (Fig. 2E, Table S1). These results suggest that HBV infection or HBx-negative (HCC)
HBx-negative (Normal liver)
P3 mRNA HBx mRNA β-actin mRNA HBx protein β-actin protein (B)
Case No. P3 mRNA
HBx mRNA
10
51
58
(C)
+1
IGF2-P3 –540
–290
Unmethylated
IGF2
TSS Methylated
β-actin mRNA HBx protein β-actin protein Case 10 (19.9%)
Case 51 (31.6%)
Case 58 (79.4%)
P = 0.026 P = 0.000 Fig. 1. P3 mRNA and P3 promoter methylation levels in HCC specimens with and without HBx expression and normal adult liver specimens. (A) Expression levels of IGF2-P3 mRNA were evaluated by semiquantitative RT-PCR in five HBx mRNA and protein expression-positive HCC specimens, five HBx mRNA and protein expression-negative HCC specimens and four normal adult liver specimens. b-actin mRNA and protein served as internal controls. (B) Semiquantitative RT-PCR and Western blot results of case 10 (HBx-positive HCC specimen), case 51 (HBxnegative HCC specimen) and case 58 (normal adult liver specimens). (C) Bisulphite sequencing results of methylation levels of P3 promoter fragment ( 540/ 290 bp) from case 10, case 51 and case 58 samples. Each row of circles represents sequencing results of a clone, filled circles are methylated CpG dinucleotides and empty circles are unmethylated CpG dinucleotides. The percentage of methylated CpG dinucleotides is indicated in parentheses. TSS, transcription start site.
4
Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Tang et al.
HBx promotes IGF2 expression in HCC
(D)
P = 0.006
P3 promoter methylation levels (%)
–Δ CT(P3 mRNA level)
(A) –4
–12
BNC–HCC
30
BNC–HCC
NBNC–HCC
(E) –2
–Δ CT(P3 mRNA level)
–Δ CT(P3 mRNA level)
50
NBNC–HCC
(B) γ = 0.739, P = 0.000
–10
–18 –20
–15
–10
–5
–Δ CT(HBx mRNA level)
–Δ CT(P3 mRNA level)
P = 0.014
10
–20
(C)
70
–2 γ = 0.882, P = 0.000
–10
–18 15
35
55
75
P3 promoter methylation(%)
2
γ = 0.732, P = 0.000
–2 –6 –10 –14 –18 –0.5
0
0.5
1
1.5
2
HBx protein level (AU) Fig. 2. Effects of HBx expression on P3 mRNA and P3 promoter methylation levels in HCC specimens. P3 mRNA, and HBx mRNA and protein levels were measured in BNC-HCC and NBNC-HCC specimens. The MCT method was used to quantify quantitative RT-PCR results and the signal densitometric ratio (AU) of HBx vs. b-actin Western blot results was used to quantify HBx protein level. The methylation level of P3 promoter fragment ( 540/ 290 bp) was analysed using bisulphite sequencing. (A) P3 mRNA level in BNC-HCC samples was higher than that in NBNC-HCC samples. (B, C) P3 mRNA abundance was positively correlated with HBx mRNA and protein levels. (D) The methylation level of P3 promoter in BNC-HCC samples was lower than that in NBNC-HCC samples. (E) P3 mRNA abundance and P3 promoter methylation level were negatively correlated. AU, arbitrary units; BNC-HCC, HBV infection-positive/HCV infection-negative HCC; NBNC-HCC, both HBV and HCV infection-negative HCC.
HBx expression may decrease the P3 promoter methylation level and then upregulate P3 mRNA expression at tissue level. HBx promotes the expression of IGF2 gene by inducing DNA hypomethylation of the P3 promoter in HCC cells
To further investigate the possible role of HBx in the regulation of P3 transcript expression, P3 mRNA level and the methylation profile of the 17 CpG sites in the P3 promoter fragment of 540/ 290 bp were analysed in HCC cells by quantitative RT-PCR and bisulphite Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
sequencing respectively. P3 mRNA level in HepG2-HBx cells stably overexpressing HBx was significantly higher than that in HepG2-control cells not expressing HBx (Fig. 3A, B). Meanwhile, the methylation of the 17 CpG sites was lower in HepG2-HBx cells than that in HepG2-control cells (Fig. 3C). The findings indicate that HBx overexpression may promote P3 transcript expression via inducing DNA hypomethylation of IGF2P3 promoter. In order to determine whether the difference in the P3 transcript expression might have an additional explanation between HepG2-HBx cells and HepG2-control
5
Tang et al.
HBx promotes IGF2 expression in HCC
Table 1. Comparisons of clinicopathological data and experimental results between BNC-HCC and NBNC-HCC groups
(A)
HepG2-control HepG2-HBx
HBx protein
Mean age (years) Mean tumour diameter (cm) No. of female patients No. of patients with AFP ≥ 400 lg/L No. of patients with Edmondson grade III– IV No. of patients with TEPV No. of patients with multiple tumour nodularity P3 mRNA levels ( MCT) P3 promoter methylation levels (%)
NBNC-HCC (n = 9)
51.1 ± 12.7 7.1 ± 3.4
60.8 ± 10.4 8.7 ± 4.0
P 0.038 0.226
10 22
1 3
0.717 0.108
33
3
0.030
28
2
0.046
8
2
1.000
9.59 ± 3.22 36.9 ± 15.5
12.97 ± 3.08 52.1 ± 19.1
0.006 0.014
AFP, a-foetoprotein; BNC-HCC, HBV infection-positive/HCV infectionnegative HCC; NBNC-HCC, both HBV and HCV infection-negative HCC; TEPV, tumour embolus of portal vein.
cells, the relative stabilities of P3 mRNAs were evaluated using the in vitro system for studying mRNA stability described by Pei et al. (32). No significant difference was found in the relative stabilities of the two mRNAs (results not shown). These data suggest that the stabilities of them may not be responsible for the discrepancy in the P3 transcript expression from the above two kinds of HCC cells and that HBx may not act at the level of P3 mRNA stability. HBx enhances the activity of IGF2-P3 promoter via inducing DNA hypomethylation in vitro
To further confirm whether HBx can upregulate the activity of IGF2-P3 promoter, in vitro-premethylated P3 promoter-luciferase construct (pGL3-P3) was transiently cotransfected with increasing amounts of HBx expression plasmid (pCMV-tag2B-HBx) or control plasmid (pCMV-tag2B) into HepG2 and Huh-7 cells. Luciferase assays revealed that P3 promoter activity was stimulated by the ectopic expression of HBx in a dosedependent manner in the two cell lines (Fig. 4A). Western blot and RT-PCR analyses indicated that the transient expression of HBx resulted in upregulation of endogenous P3 mRNA level in a dose-dependent manner in the two cell lines (Fig. 4B). Hepatitis B virus has four open reading frames that encode HBx, HBp, HBe, HBc, HBpreS1, HBpreS2 and HBs proteins. To investigate the specificity of HBx in the regulation of P3 promoter activity, the above seven protein expression plasmids (pCMV-tag2B-HBx, pCMV-tag2B-HBp, pCMV-tag2B-HBe, pCMV-tag2B-
6
β-actin protein
(B) P3 mRNA expression level
Parameters
BNC-HCC (n = 43)
4 3
P = 0.000
2 1 0
HepG2-HBx
(C)
+1
IGF2-P3 –540
HepG2-control
–290
Unmethylated
IGF2 TSS
Methylated
HepG2-control (28.7%)
HepG2-HBx (14.7%) P = 0.005
Fig. 3. P3 mRNA and P3 promoter methylation levels in HepG2HBx cell lines stably overexpressing HBx vs. HepG2-control cell lines not expressing HBx. (A) Detection of HBx protein expression in HepG2-HBx cell lines and HepG2-control cell lines by Western blot analysis. (B) P3 mRNA was upregulated in HepG2-HBx cells compared with that in HepG2-control cells detected by quantitative RT-PCR and the results are the average of three independent experiments carried out in triplicate. (C) The methylation level of the 17 CpG sites in the P3 promoter fragment ( 540/ 290 bp) was lower in HepG2-HBx cells than that in HepG2-control cells by bisulphite sequencing analysis. Each row of circles represents sequencing results of a clone, filled circles are methylated CpG dinucleotides and empty circles are unmethylated CpG dinucleotides. The percentage of methylated CpG dinucleotides is indicated in parentheses. TSS, transcription start site.
HBc, pCMV-tag2B-HBpreS1, pCMV-tag2B-HBpreS2 and pCMV-tag2B-HBs) were cotransfected with in vitro-premethylated pGL3-P3 into HepG2 cells. As shown in Figure 4C, HBx significantly enhanced P3 promoter activity, whereas HBp, HBe, HBc, HBpreS1, HBpreS2 and HBs as well as control had no effect on its activity. Next, we investigated the mechanism by which HBx enhances the activity of IGF2-P3 promoter. Bisulphite sequencing was performed to analyse the methylation profile of the 29 CpG sites located within P3 promoter fragment of 1251/ 1003 bp from in vitro-premethyLiver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Tang et al.
HBx promotes IGF2 expression in HCC
(A) Relative luciferase activity
60
HepG2
tion of the 29 CpG sites (Fig. 5). These results suggest that the ectopic expression of HBx may enhance the transcriptional activity of the P3 promoter by lowering DNA methylation level.
Huh-7
50 40 30
*
*
Characteristics and prognostic factors of subjects
20 10
0 Premethylated HBx expression plasmid
+ –
+ +
+ ++
+ +++
Control plasmid
+
–
–
–
(B) – +
HBx expression plasmid Control plasmid
++ +++ – –
+ –
– +
+ –
++ +++ – –
P3 mRNA β-actin mRNA HBx protein
P = 0.000
30 25 20
HepG2
15 10 5
H Bs
H Be H Bc H Bp re S1 H Bp re S2
nt co
H Bp
0
ro l H Bx
(C)
Relative luciferase activity
β-actin protein
Fig. 4. The ectopic expression of HBx increases the activity and transcription of IGF2-P3 promoter in vitro. (A) HepG2 and Huh-7 cells were transiently cotransfected with increasing amounts of HBx expression plasmid or control plasmid and premethylated pGL3-P3 vector. The observed firefly luciferase activity was normalized to Renilla luciferase activity and the results were the average of three independent experiments carried out in triplicate. The transfected quantities of HBx expression plasmid were 0.25 lg (+), 0.5 lg (+ +) and 1.0 lg (+ + +) per 1 9 106 cells. *Indicates that the mean values in HepG2 or Huh-7 cells cotransfected with HBx expression plasmid are significantly different from those of mock-transfected controls (P = 0.000). (B) Expression analyses of P3 mRNA and HBx protein in HepG2 and Huh-7 cells transiently transfected with increasing amounts of HBx expression plasmid or control plasmid. b-actin mRNA and protein served as internal controls. The transiently transfected HBx expression plasmid DNA quantities were as follows: 0.25 lg (+), 0.5 lg (+ +) and 1.0 lg (+ + +) per 1 9 106 cells. (C) In vitro-premethylated pGL3-P3 was cotransfected into HepG2 cells with control, HBx, HBp, HBe, HBc, HBpreS1, HBpreS2 and HBs expression plasmid. The transfected quantities of each plasmid were 0.25 lg. The observed firefly luciferase activity was normalized to Renilla luciferase activity and the results were the average of three independent experiments carried out in triplicate.
lated pGL3-P3 vector transiently cotransfected into HepG2 and Huh-7 cells. The addition of HBx expression plasmids resulted in a dose-dependent demethylaLiver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
The clinicopathological characteristics of the subjects are presented in Table S2. The patients with BNC-HCC were much younger and were involved more often with poor tumour differentiation (Edmondson grade III–IV) and tumour embolus of portal vein (TEPV) compared with those of the patients with NBNC-HCC, whereas there were no differences between the BNC-HCC and NBNC-HCC groups with respect to sex, tumour size, tumour nodularity and a-foetoprotein level (Table 1). Further, HBx mRNA and protein levels, and P3 mRNA level were higher in HCC specimens with poor tumour differentiation and with TEPV than those in HCC specimens with well tumour differentiation and without TEPV, but there were no differences in HBx mRNA and protein levels, and P3 mRNA level between HCC patients with single tumour nodularity and with multiple tumour nodularity (Table 2). These results suggest that HBx and P3 mRNA expression were closely associated with poor tumour differentiation and TEPV of HBV-related HCC patients. Kaplan–Meier method and the log-rank test showed that the patients with BNC-HCC had both shorter disease-free survival time and shorter overall survival time compared with those of the patients with NBNC-HCC (Fig. 6). Univariate Cox regression analysis revealed that a-foetoprotein (AFP), tumour differentiation, TEPV and HBx mRNA expression were significantly associated with both disease-free survival time and overall survival time. Using multivariate Cox regression analysis, poor tumour differentiation and HBx mRNA expression were found to be independent predictive factors for both shorter disease-free survival time and shorter overall survival time of HCC patients (Table 3). Discussion
Evidence has shown that the overexpression of IGF2 plays a key role in the pathogenesis of HCC (16–18, 33–35). In a previous study, we showed that the increased IGF2 expression from the fetal P3 and P4 promoters was observed in most of HCC tissues and could be involved in the transformation of a premalignant liver lesion to HCC (21). However, the mechanisms that may contribute to upregulation of the fetal P3 and P4 transcripts are not well understood. Subsequently, in another study, we further found that the overexpression of IGF2-P4 transcript was closely associated with P4 promoter hypomethylation in HCC tissues (36). In this study, we explored the effect of HBx on IGF2-P3 transcript expression and epigenetic modification-related molecular mechanisms associated
7
Tang et al.
HBx promotes IGF2 expression in HCC
+1
pGL3-P3 promoter –1251
–1003
Luciferase
TSS Huh-7
HepG2 Control (73.7%)
Control (70.3%)
HBx(+) (55.6%)
HBx(+) (50.9%)
HBx(++) (46.1%)
HBx(++) (40.5%)
HBx(+++) (32.3%)
HBx(+++) (27.6%)
Unmethylated
Methylated
Fig. 5. The ectopic expression of HBx lowers the methylation level of IGF2-P3 promoter in vitro. The methylation profile of P3 promoter fragment ( 1251/ 1003 bp) from in vitro-premethylated pGL3-P3 vector transiently cotransfected into HepG2 and Huh-7 cells with increasing amounts of HBx expression plasmid or control plasmid. The transfected quantities of HBx expression plasmid were 0.25 lg (+), 0.5 lg (+ +) and 1.0 lg (+ + +) per 1 9 106 cells. Each row of circles represents sequencing results of a clone, filled circles are methylated CpG dinucleotides, and empty circles are unmethylated CpG dinucleotides. The percentage of methylated CpG dinucleotides is indicated in parentheses. TSS, transcription start site.
Table 2. Correlation of HBx expression and P3 mRNA expression levels with clinicopathological features of HCC Parameters Differentiation degree Poorly differentiated (III–IV) Well differentiated (I–II) TEPV Positive Negative Tumour nodularity Uninodular Multinodular
N
HBx mRNA level ( MCT)
P-value
HBx protein level (AU)
P-value
P3 mRNA levels( MCT)
P-value
36 16
10.98 ± 3.00 14.33 ± 3.17
0.001
0.93 ± 0.50 0.39 ± 0.49
0.001
9.03 ± 2.87 12.77 ± 3.19
0.000
30 22
10.28 ± 2.72 14.37 ± 2.79
0.000
1.05 ± 0.43 0.37 ± 0.44
0.000
8.43 ± 2.76 12.56 ± 2.75
0.000
42 10
11.86 ± 3.35 12.66 ± 3.74
0.507
0.79 ± 0.54 0.64 ± 0.60
0.459
9.92 ± 3.31 11.26 ± 3.82
0.270
AU, arbitrary units; TEPV, tumour embolus of portal vein.
with the expression variation of this transcript in human HCC. Our data showed that the higher expression level of P3 mRNA was detected in the HCC samples with HBV infection compared with that of the HCC samples without HBV infection, and that P3 mRNA level was positively correlated with HBx mRNA and protein levels in all HCC specimens examined. Moreover, we further
8
found that the stable or transient expression of HBx upregulated P3 mRNA level in HCC cells. These in vivo and in vitro findings suggest that HBx may promote IGF2-P3 transcript expression in human HCC. These results are concordant with that reported by Tian et al., who found that HBx could increase the expression of Raf1, a member of the Ras/Raf/MEK/ERK kinase Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Tang et al.
HBx promotes IGF2 expression in HCC
(A)
(B)
P = 0.031
P = 0.038
Fig. 6. Kaplan–Meier curves for disease-free survival and overall survival time in 43 BNC-HCC patients and 9 NBNC-HCC patients. (A) The patients with BNC-HCC had a shorter disease-free survival time than those with NBNC-HCC. (B) The patients BNC-HCC had a shorter overall survival time than those with NBNC-HCC.
Table 3. Univariate and multivariate Cox regression analysis showing disease-free survival and overall survival in 52 patients with HCC Disease-free survival Parameters Univariate Gender M: n = 41 F: n = 11 Age (years)
