Abstract

Background. As a tumor suppressor, p16 can competitively block the cyclin D1-CDK4/6 complex to arrest the cell cycle in the G1 phase. Lack of p16 gene expression can lead to infinite cell proliferation and malignant transformation.

Objectives. To determine whether the hepatitis B virus X protein (HBx) regulates the methylation and expression of the p16 gene.

Materials and methods. We constructed a eukaryotic expression vector carrying the HBx gene and green fluorescent protein (GFP), and transfected it into HepG2 cells to build cell lines stably expressing GFP and GFP-HBx. The p16 protein level and p16 gene methylation were measured in these cell lines. We further detected the mRNA expression of DNA methyltransferases (DNMTs) 1, 3A and 3B. The DNMT1, DNMT3A and DNMT3B gene promoter sequences were inserted into a reporter vector (pGL3-Basic) to build recombinant vectors, which were then transiently transfected into different cell lines. After 48 h, the luciferase activity was measured.

Results. The level of p16 protein in GFP-HBx/HepG2 cells was significantly lower than in HepG2 and GFP/HepG2 cells. The CpG methylation was present in the p16 gene promoter region of GFP-HBx/HepG2 cells. The DNMT1 and DNMT3A mRNA levels in GFP-HBx/HepG2 cells were significantly elevated compared to that in the HepG2 cells (p = 0.0495). The luciferase activity in GFP-HBx/HepG2 cells transfected with the pGL3-DNMT1/3A pro plasmid was significantly higher than in HepG2 and GFP/HepG2 cells (both p < 0.05).

Conclusions. The HBx can induce p16 gene promoter methylation and inhibit the expression of p16 in HepG2 cells. This occurs due to HBx activation of DNMT1 and DNMT3A promoters and the induction of p16 promoter methylation, which downregulates the expression of p16 protein.

Key words: DNA methylation, p16, hepatitis B X protein, DNA methylation transferase

Background

Hepatocellular carcinoma (HCC) is one of the most common malignant tumors in the world. Chronic hepatitis B virus (HBV) infection is the main risk factor for HCC. There are about 296 million HBV-infected persons around the world and approx. 820,000 deaths in this group every year.¹ The risk of HCC is significantly higher in hepatitis B surface antigen (HBsAg)-positive individuals than in uninfected persons. In China, 80% of HCC cases are related to HBV infection.² Hepatitis B virus is the smallest double-stranded DNA virus known to infect humans, and it has been discovered that hepatitis B virus X protein (HBx) plays an important role in the development and progression of HBV-related HCC.^{3, 4, 5}

Tumor suppressor gene p16 is an important cell cycle negative regulatory factor, and lack of p16 gene expression can lead to infinite cell proliferation and malignant transformation.⁶ Studies have shown that there is a correlation between the regulation of p16 gene expression and HBx.⁷ There are currently 2 proposed mechanisms how HBx regulates the expression of p16. One is that HBx can induce p16 mutation, which leads to loss of heterozygosity and of expression.⁸ The other is that HBx can activate various transcription factors and signal pathways through transactivation; this process can abnormally activate the enzyme system regulating DNA methylation, induce hypermethylation of the tumor suppressor gene and downregulate its expression.⁹ Recent studies have shown that HBx is a very effective epigenetic modifier that can induce methylation and inactivation of the promoters of tumor suppressor genes, including p53, PTEN and E-cadherin, and this is closely related to the development and progression of HBV-related HCC.^{10, 11}

As a transactivator, HBx can combine many factors related to transcription and gene regulation. The DNA methylation is maintained under the action of DNA methyltransferase (DNMT). Currently, 3 DNMTs that exhibit transmethylation activity have been identified: DNMT1, DNMT3A and DNMT3B. The main function of DNMT1 is to maintain methylation. The DNMT3A and DNMT3B interact to cause de novo methylation of DNA sequences. At present, the role of HBx transactivation with respect to DNMTs is unclear. Jung et al. reported that HBx can upregulate the expression of DNMT1 and DNMT3A, which induce methylation of the retinoic acid receptor (RAR) β-promoter and thus inhibit its expression.¹² The inhibition of its expression results in a failure of the tumor suppressor protein RA to combine with RAR-β, which results in the malignant transformation of hepatocytes.¹² In our prior work, we reported that HBx upregulates DNMT1 and DNMT3A mRNA and protein expression; however, the underlying molecular mechanism of this effect is unclear.¹³

Objectives

The current study aims to investigate the molecular mechanism underlying HBx regulates the expression of the p16 gene and if HBx regulates p16 expressions indirectly through methylation.

Materials and methods

Plasmid construction

The pHBV plasmid containing 1.3 copies of the HBV genome (adr serum type) that can cause the development of HBV virus particles was previously developed by our group and is maintained in our laboratory. Green fluorescent protein (GFP) eukaryotic expression vector pEGFP-C1 was purchased from Biotech Company (Bergisch Gladbach, Germany). The pGL3-Basic and pRL-TK vectors were purchased from Promega (Madison, USA). To construct the pEGFP-HBx recombinant plasmid, the HBx gene was amplified using the pHBV plasmid as a template and then cloned into the BgIII and KpnI sites of the pEGFP-C1 vector. To construct the pGL3-DNMT1 pro-recombinant plasmid, a human DNMT1 promoter sequence (ENST00000340748)¹⁴ was amplified from normal human genomic DNA (Bio-Chain, Newark, USA), and then cloned into the XhoI and HindIII sites of the pGL3-Basic vector. To construct pGL3-DNMT3A/B pro-recombinant plasmids, human DNMT3A and DNMT3B promoter sequences (ENST00000264709, ENST00000328111)^{15, 16} were amplified and cloned into the KpnI and BglII sites of the pGL3-Basic vector.

Primers used for all of the constructs are shown in Table 1. All constructs were confirmed using 1.2% agarose gel electrophoresis (Biowest, Nuaillé, France) and DNA sequencing (Takara, Kusatsu, Japan).

Cell cultures and transfections

The human hepatoma cell line HepG2 was maintained in our laboratory. The HepG2 cell line was propagated and maintained in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, Waltham, USA) supplemented with 10% fetal bovine serum (FBS) (HyClone, Logan, USA). Cultures were incubated in a humidified atmosphere at 37°C with 5% CO₂.

All transfections were performed with Lipofectamine™ 2000 Reagent (Invitrogen, Carlsbad, USA), according to the manufacturer’s recommendations. Transfected cells were observed with an inverted fluorescence microscope (Nikon Eclipse TS100/100F; Nikon Corp., Tokyo, Japan) 6 h after transfection.

Screening of stably transfected cell lines – G418 (geneticin) screening experiment

The HepG2 cells (2000 cells/mL) were placed in 96-well plates. The G418 was added on days 1, 2, 3, and 4, respectively, at concentrations of 0, 200, 400, 600, 800, and 1000 μg/mL. At 20 days of culture, the lowest concentration that resulted in the death of all cells was 600 μg/mL. Therefore, 800 μg/mL was selected for day 2 screening. The HepG2 cells were transfected with pEGFP-HBx, and after 24 h of transfection the culture medium was removed, the cells were washed with phosphate-buffered saline (PBS) 3 times, treated with trypsin, and centrifuged at 1500 rpm for 5 min. The G418 (800 μg/mL) was added to 5 mL of complete DMEM culture medium and placed in a 25-mL ventilated culture bottle; the solution was changed daily. The HepG2 cells transfected with empty vector pEGFP-C1 were used as a control. The 800 μg/mL concentration of G418 was screened for 20 days.

An inverted fluorescence microscope (Nikon Eclipse TS100/100F) was used to visualize monoclonal cells expressing GFP. The cells were washed 3 times with PBS and treated with trypsin. Under microscopic visualization, the monoclonal cells were carefully aspirated with a pointed pipette and placed into a 25-mL ventilated culture bottle. Complete DMEM with a G418 concentration of 800 μg/mL was added, and after the culture bottle was full of cell clones, it was amplified in vitro. The cells were named GFP/HepG2 (transfected with pEGFP-C1) and GFP-HBx/HepG2 (transfected with pEGFP-HBx).

Reverse transcription polymerase
chain reaction (RT-PCR)

The RNA was isolated from harvested cells using an RNA Isolation Kit (TianGen, Beijing, China), according to the manufacturer’s instructions. The cDNA was then synthesized from the RNA using a PrimeScript® RT Reagent Kit with gDNA Eraser (Takara). Polymerase chain reaction was performed at 95°C for 5 min, followed by 35 cycles of 98°C for 10 s, 68°C for 10 s, and 72°C for 5 min. The subsequent PCR products were visualized with 1.2% agarose gel electrophoresis.

An SYBR® Premix ExTaq™ II Kit (Takara) was used to perform quantitative PCR (qPCR) of human DNMT1, DNMT3A and DNMT3B. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used for normalization. The primers are shown in Table 1. The PCR conditions were 95°C for 30 s, followed by 40 cycles of 95°C for 5 s and then 60°C for 20 s.

Methylation-specific PCR (MSP)

The promoter sequence of the human p16 gene (ENST000004494) was was obtained University of California Santa Cruz Genome Browser (http://genome.ucsc. edu). The CpG islands were predicted using MetaPrime (Supplementary Figure 1). Methylation and non-methylation primers were designed using the Methyl Primer Express v. 1.0 software (Applied Biosystems Inc., Carlsbad, USA), MethPrimer (http://www.urogene.org/methprimer/), Primer Premier 5.0 (PREMIER Biosoft, San Francisco, USA), and Oligo 7.0 (Molecular Biology Insights, Inc., Cascade, USA) (Table 1). Methylation-specific PCR was conducted using an EpiTect MSP Kit (Qiagen, Hilden, Germany), according to the manufacturer’s instructions. The PCR conditions were 95°C for 10 min, followed by 40 cycles of 94°C for 15 s, 55°C for 30 s, 72°C for 30 s, and 72°C for 10 min. The subsequent PCR products were visualized using 1.2% agarose gel electrophoresis.

Western blotting

The M-PER™ Mammalian Protein Extraction Reagent (Pierce, Rockford, USA) was used to extract the total protein from the transfected cells, and protein concentration was determined using BCA Protein Assay Reagent A (Pierce) according to the manufacturer’s instruction. Proteins were separated with 10% sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE), transferred to nitrocellulose membranes, and then detected using the appropriate antibodies, with 12 h incubation time for primary antibody and 2 h for secondary antibody. Rabbit anti-GAPDH was purchased from Xianzhi Company (Guangzhou, China), anti-HBx was purchased from Santa Cruz Biotechnology (Dallas, USA) and anti-p16 was purchased from Cell Signaling Technology (Danvers, USA).

Determination of DNMT1, DNMT3A and DNMT3B gene promoter activity

The stably transfected cells GFP/HepG2, GFP-HBx/HepG2 and HepG2 were transiently transfected with pGL3-DNMT1 pro, pGL3-DNMT3A pro, pGL3-DNMT3B pro, and pGL-3-Basic (control) together with the Renilla luciferase plasmid phRL-tk. The double-luciferase reporter assay system (Promega) was used to detect the activity of luciferase.

Statistical analyses

With a limited sample size (3–9), the assumption of normality was not adopted to the continuous variables in this study, and only nonparametric statistical tests were used. For all comparisons among HepG2, GFP/HepG2 and GFP-HBx/HepG2 groups, the results were indicated using a dot-plot with medians, and the significance was tested using the Kruskal–Wallis test and Dunn’s test with Bonferroni adjustment as post hoc test. The Mann–Whitney U test was used in the comparison between the 2 groups. The detailed statistics, p-value, sample size, and testing method were reported in Supplementary Tables. All analyses were done using IBM SPSS v. 25 software (IBM Corp., Armonk, USA). The statistical significance level for all the tests was set at a p 0.05 (two-tailed).

Results

Construction and identification of the pEGFP-HBx eukaryotic expression vector

Based on the designed target gene primers, the HBx gene fragment was approx. 470 bp long and this finding was consistent with the prediction (Figure 1A). The recombinant plasmid pEGFP-HBx was treated with rapid restriction endonucleases BglI and KpnI. The products were subjected to 1% agarose gel electrophoresis, and the empty plasmid (pEGFP-C1) was used as a control. The DNA bands of 4700 bp were observed for both, while a DNA band of 470 bp was only observed for the recombinant vector (Figure 1B). Sequencing results of the recombinant plasmid confirmed that HBx was inserted between the BglII and KpnI digestion sites of the pEGFP-C1 empty vector. Blast homology comparison using the National Center for Biotechnology Information (NCBI) database showed that the HBx gene was forwardly connected to the pEGFP-C1 eukaryotic vector, and was named pEGFP-HBx for this study.

Construction of pGL3-DNMTs
pro-recombinant plasmids

The electrophoresis showed a band of approx. 850 bp for DNMT1, 1000 bp for DNMT3A and 900 bp for DNMT3B, which implied that the amplification was successful (Figure 2A,B). The digestion of the pGL3-DNMT pro-recombinant plasmids by relative restriction endonucleases showed a size of 850 bp for DNMT1, 1000 bp for DNMT3A and 900 bp for DNMT3B, indicating that the connections were successful (Figure 2C,D). The sequence analysis of the recombinant plasmids confirmed that the DNMT1, DNMT3A and DNMT3B promoters were positively inserted into the pGL3-Basic vector, and were respectively named pGL3-DNMT1 pro, pGL3-DNMT3A pro and pGL3-DNMT3B pro.

Establishment of HepG2 cell lines stably expressing GFP and GFP-HBx protein

The HepG2 cells were transfected with pEGFP-C1 or pEGFP-HBx and screened for about 20 days with G418. All HepG2 cells in the blank control died. Resistant cell clones were observed in the pEGFP-C1 and pEGFP-HBx plasmid transfection groups. Cell clones expressing GFP were carefully selected under the inverted fluorescence microscope. Cells stably expressing GFP were screened out under white light and fluorescence, and named GFP/HepG2 and GFP-HBx/HepG2, respectively (Figure 3).

Detection of HBx gene expression
using RT-PCR and western blotting

Reverse transcription polymerase chain reaction was used to detect the expression of HBx based on the mRNA isolated from the different stable cell lines. The results showed that the gene fragments of GFP-HBx/HepG2 cells were the expected size, while the RT-PCR products of HepG2 and GFP/HepG2 cells exhibited no bands on electrophoresis, indicating that HBx gene transcription was present in GFP-HBx/HepG2 cells (Figure 4A). The detection of HBx protein expression in stably transfected cells using western blotting showed that HBx could be detected in GFP-HBx/HepG2 cells but not GFP/HepG2 or HepG2 cells (Figure 4B).

p16 protein levels in different cell types

After the Kruskal–Wallis test, no significant difference of p16 relative gray value was found among HepG2, GFP/HepG2 and GFP-HBx/HepG2 groups (as indicated in Supplementary Table 1,2). However, the level of p16 protein in GFP-HBx/HepG2 cells was significantly lower than in GFP/HepG2 cells and HepG2 cells (p = 0.049, p = 0.049) in separate Mann–Whitney U tests, while there was no significant difference in the p16 protein level between GFP/HepG2 and HepG2 cells (Figure 5A,B). Since the authors believe that the differences in p16 protein levels were worthy of interpretation, both the Kruskal–Wallis test and the Mann–Whitney U test results were reported.

The methylation products and demethylation products in the promoter region of the p16 gene were detected with MSP. Only the demethylation products were amplified in HepG2 cells and GFP/HepG2 cells, while both the demethylation and methylation products were amplified in GFP-HBx/HepG2 cells (Figure 6). Methylation products were not detected in GFP-HBx/HepG2 cells 3 days after the addition of 5-Aza-2′-deoxycytidine (20 µmol/L) (Figure 6).

The impact of HBx on the transcription of DNMTs

The expression level of DNMT mRNA in HepG2 cells was set at 1.00. The expressions of DNMT1 and DNMT3A mRNA in GFP-HBx/HepG2 cells were significantly higher than in the GFP/HepG2 cells (p = 0.0495), while there was no difference in DNMT3B mRNA (p > 0.05) (Figure 7). Detailed statistics and p-values were reported in Supplementary Table 3.

Impact of HBx on DNMT promoter activity

The relative activity of luciferase in each group was compared among the different cell lines transfected with different plasmids. The activity of luciferase in HepG2 cells transfected with pGL3-Basic was set at 1.0. There was no statistical difference in luciferase activity among the 3 blank groups transfected with the pGL3-Basic vector. The relative luciferase activity in GFP-HBx/HepG2 cells transfected with the pGL3-DNMT1 pro and pGL3-DNMT3A pro plasmids was significantly higher than in the GFP/HepG2 and HepG2 cells (p < 0.05). There was no difference in activity among the 3 groups when transfected with pGL3-Basic or pGL3-DNMT3B plasmids (the relative activity was transformed using log10 to level down the disparity of scale) (Figure 8 and Supplementary Table 1,2).

Discussion

In this study, we found that HBx can induce p16 gene promoter methylation and inhibit the expression of the p16 gene. Importantly, the results showed that the mechanism by which HBx regulates the expression of the p16 gene is through the upregulation of DNMT1 and DNMT3A gene promoter activity, which enhances their transcription and expression. The DNMT1 and DNMT3A belong to a family of DNA methyltransferase enzymes that catalyze the transfer of methyl groups to specific CpG structures in DNA, alter chromatin structure and regulate gene expression.¹⁷ This subsequently results in methylation of the p16 gene promoter and the downregulation of gene expression.

The HBx gene is the smallest and most conservative gene among the 4 open reading frames of the HBV genome.¹⁸ The HBx can activate a variety of oncogenes and transcription factors, including C-myc, AP-1, NF-κB, and AP-2.¹⁸ The HBx is involved in tumor cell metabolism, proliferation, apoptosis, invasion, and metastasis.^{19, 20} Current research has also suggested that HBx is an effective epigenetic modifier, which may activate various transcription factors via transactivation, and downregulate the expression of tumor suppressor genes by hypermethylation.

The fusion expression of an exogenous target protein and GFP can facilitate real-time monitoring and intracellular localization of a target gene²¹; hence, GFP expression of the pEGFP-C1 vector was used in this study. The HBx gene was inserted into pEGFP-C1 to construct the recombinant plasmid pEGFP-HBx. The pEGFP-C1 empty vector and pEGFP-HBx were transfected into the human hepatic carcinoma cell line HepG2. The results showed that both transfected cells effectively expressed GFP. After G418 screening, HepG2 cells with stable expression of GFP and GFP-HBx were obtained, and after 40 generations of culture, the cells still expressed strong fluorescence. The RT-PCR and western blotting showed that HepG2 cells transfected with pEGFP-HBx expressed the HBx gene. These results showed that HepG2 cell lines stably expressing GFP and GFP-HBx fusion protein were successfully established, and provided a platform for subsequent studies.

As the most important negative regulator of the cell cycle, tumor suppressor gene p16 regulates cell growth and differentiation by binding and inhibiting cell cycle-dependent protein kinases CDK4 and CDK6, and reduces retinoblastoma protein phosphorylation. The loss of p16 gene expression can cause infinite cell proliferation and induce malignant transformation of cells. Our results showed that there was a significant negative correlation between p16 and HBx expression in GFP-HBx/HepG2 cells. Therefore, we speculated that there might be a mechanism by which HBx downregulates p16 expression. Many viruses, including HBV, can induce methylation of tumor suppressor genes, and thus downregulate their expression.²² It has been reported that p16 promoter methylation in HCC is significantly related to HBV infection.^{23, 24} There are several causes of p16 gene inactivation, such as gene loss, gene mutation, promoter hypermethylation, and homozygous deletion. Prior studies have focused on tumor suppressor gene mutations, while recent studies have shown that epigenetic abnormalities of tumor suppressor gene methylation are common causes for the development and progression of malignant tumors.²⁵ Importantly, HBx is a multi-functional viral protein closely related to the development of HCC, and recent studies have suggested that HBx is an effective epigenetic modifier that can inactivate tumor suppressor genes by inducing methylation. Our results showed that the exogenous transfection of the HBx gene resulted in the methylation of some CpG sites in the p16 gene promoter sequence in HepG2 cells. The treatment of GFP-HBx/HepG2 cells with 5-Aza-2′-deoxycytidine resulted in the recovery of the demethylated state of the p16 gene promoter, indicating that DNA methylation was one of the reasons for the downregulation of the p16 gene by HBx.

The upregulation of the expression of DNMTs has been found in many malignant tumors, including HCC.²⁶ Abnormally high expression of DNMT1 is closely related to the occurrence of ovarian cancer, cervical cancer, lung cancer, gastric cancer, and liver cancer.²⁶ The DNMT1 participates in characteristics of tumor cells such as unlimited proliferation, migration and invasion, and plays an important role in the maintenance of a continuous methylation state of tumor-related genes.^{27, 28} However, the specific mechanisms by which DNMTs promote liver cancer have not been elucidated. Recent studies have shown that many viruses, including HBV, can upregulate the expression of DNMTs, thus causing the methylation of tumor suppressor genes and downregulation of their expression.^{29, 30} Jung et al. reported that HBx can induce the methylation of the RAR-β promoter by upregulating the expression of DNMT1 and DNMT3A.¹² This results in the failure of tumor suppressor RA to combine with RAR-β, which leads to the failure of tumor suppressor activity and the subsequent malignant transformation of hepatocytes.¹² At present, the role of DNMTs in the process of HBx-induced p16 methylation is not clear. Therefore, we designed specific primers to amplify DNA fragments, including DNMT1, DNMT3A and DNMT3B promoter regions, and constructed corresponding promoter reporter vectors, in order to study the effect of HBx on DNMTs at the promoter level.

Limitations

We found that HBx significantly upregulated the promoter transcription activity of DNMT1 and DNMT3A. In addition, the mRNA expression of the DNMT1 and DNMT3A genes was also increased, while HBx had no significant effect on DNMT3B at the promoter level and the mRNA level. Therefore, we speculate that DNMT1 and DNMT3A may play a catalytic role in the methylation of the p16 gene promoter induced by HBx, probably via the mechanism of extensive transactivation. Previous studies have shown that HBx is a powerful transactivator.³¹ Although HBx in the nucleus cannot directly bind to DNA, it can interact with a variety of transfer factor proteins, resulting in the increase of transcription activity of specific genes. However, how HBx regulates the expression of DNMTs is still unclear, and further research is needed.

Conclusions

In conclusion, we successfully constructed the eukaryotic expression vector pEGFP-HBx carrying the HBx gene, and stably transfected it into HepG2 cells. We found that HBx can induce p16 gene promoter methylation and inhibit the expression of the p16 gene. The mechanism by which HBx regulates the expression of the p16 gene is the upregulation of DNMT1 and DNMT3A gene promoter activity, which enhances their transcription and expression. This subsequently results in the methylation of the p16 gene promoter and the downregulation of gene expression.

Supplementary data

The supplementary files are available at https://doi.org/10.
5281/zenodo.7332522. The package consists of the following files:

Supplementary Figure 1. p16 CpG island prediction through Metaprime.

Supplementary Table 1. Kruskal–Wallis test results for Figure 5 and Figure 8.

Supplementary Table 2. Dunn’s test with Bonferroni adjustment as post hoc comparison results for Figure 5 and Figure 8.

Supplementary Table 3. Mann–Whitney U test results for Figure 7.