Hepatitis B virus X protein (HBx)-induced abnormalities of nucleic acid metabolism revealed by 1H-NMR-based metabonomics

Hepatitis B virus X protein (HBx) plays an important role in HBV-related hepatocarcinogenesis; however, mechanisms underlying HBx-mediated carcinogenesis remain unclear. In this study, an NMR-based metabolomics approach was applied to systematically investigate the effects of HBx on cell metabolism. EdU incorporation assay was conducted to examine the effects of HBx on DNA synthesis, an important feature of nucleic acid metabolism. The results revealed that HBx disrupted metabolism of glucose, lipids, and amino acids, especially nucleic acids. To understand the potential mechanism of HBx-induced abnormalities of nucleic acid metabolism, gene expression profiles of HepG2 cells expressing HBx were investigated. The results showed that 29 genes involved in DNA damage and DNA repair were differentially expressed in HBx-expressing HepG2 cells. HBx-induced DNA damage was further demonstrated by karyotyping, comet assay, Western blotting, immunofluorescence and immunohistochemistry analyses. Many studies have previously reported that DNA damage can induce abnormalities of nucleic acid metabolism. Thus, our results implied that HBx initially induces DNA damage, and then disrupts nucleic acid metabolism, which in turn blocks DNA repair and induces the occurrence of hepatocellular carcinoma (HCC). These findings further contribute to our understanding of the occurrence of HCC.

HBx disrupted the metabolism of glucose, lipids and amino acids. PLS-DA, a multivariate statistical analysis, was performed on the NMR data to investigate the intrinsic differences in the metabolite levels between the groups at 48 h or 72 h post-infection. As shown in Fig. 1c, there was no distinct separation in scores plot of PLS-DA analysis between Ad-HBx-48 and Ad-N-48. However, Fig. 1d showed a clear separation between Ad-HBx-72 and Ad-N-72 with an R 2 of 0.997 and a Q 2 of 0.743.
The PLS-DA loadings plot can further identify spectral regions responsible for the separation observed in the scores plot. Fig. 1e showed loadings plot at 72 h post-infection, and points further from the center contributed most extensively to the variance. Metabolites such as glucose, choline, creatine, lactate, glutamate, glutamine, cysteine, cystine, acetate, serine, glycine and aspartate were important for the separation. The metabolites responsible for the classification of Ad-HBx-72 and Ad-N-72 were also identified using the variable importance in projection (VIP) scores. Metabolites with high VIP are more important in providing class separation, while those with small VIP provide less contribution 16 . The metabolites with VIP ≥ 1 were shown in Fig. 1f. Metabolites such as glucose, choline, glutamine, alanine, threonine, N-acetylaspartate and glucose-6-phosphate had higher VIP scores. Red or green on the right of Fig. 1f indicated the low or high concentration of metabolites by comparing the concentration of each metabolite in Ad-HBx-72 and Ad-N-72. The concentration of some metabolites such as glucose, choline and glucose-6-phosphate was reduced following Ad-HBx infection.
HBx induced abnormalities of nucleic acid metabolism. To further identify the metabolite biomarkers induced by HBx, we removed some metabolites and selected 45 components for PLS-DA analysis. As shown in Fig. 2a, the scores plots of Component 1 and Component 2 indicated that groups of Ad-HBx-48 and Ad-HBx-72 could be separated from Ad-N-48 and Ad-N-72, respectively. The parameters of the corresponding PLS-DA model were as follows: Ad-HBx-48 versus Ad-N-48: R 2 = 0.998, Q 2 = 0.608; Ad-HBx-72 versus Ad-N-72: R 2 = 0.997, Q 2 = 0.699. Figure 2b,c showed the higher VIP scores of identified metabolites. The concentration of nucleic acid components (data not shown) was relatively low, but uridine, inosine, guanosine, uracil and xanthine were important in providing class separation. As shown in Fig. 2b,c, at 48 h post-infection, the VIP scores of guanosine and uridine were ~1.1 and 1, respectively, and at 72 h post-infection, their scores were ~1.5 and 1.8, respectively. In addition, at 72 h post-infection, inosine and xanthine had higher VIP scores.
HBx inhibited DNA synthesis. DNA synthesis is an important part of nucleic acid metabolism 17 . To further confirm that HBx disrupted nucleic acid metabolism, we conducted an EdU incorporation assay. At 48 and 72 h after infection with Ad-HBx, EdU is an analog of thymidine, which is incorporated into DNA during DNA synthesis. EdU-labeled cells were respectively 27.17 ± 4.02% and 20.67 ± 1.66% for HepG2 (Fig. 4a), and 28.40 ± 1.58% and 22.84 ± 3.89% for SK-HEP-1 (Fig. 4b), which were significantly lower than those infected with Ad-N, or untreated cells (p < 0.05).
HBx affected the expression profiles of DNA damage-related genes. To disclose the mechanism of HBx-associated metabolic abnormalities, we detected the differential expression profiles in HepG2 cells infected with Ad-HBx or Ad-N by mRNA microarray analysis (Fig. 5a). The mRNA microarray analysis showed that the expression levels of 966 genes were remarkably altered, 381 of those genes were upregulated (≥2-fold) and 585 genes were downregulated (≥2-fold) in cells infected with Ad-HBx. Within the altered genes, 29 were DNA damage-related genes involved in DNA damage response, nucleotide-excision repair, signal transduction in response to DNA damage, double-strand break repair, and DNA repair, and were found to be downregulated under HBx induction (Fig. 5a). To validate the results, we performed qRT-PCR for a selection of 6 differentially Scientific RepoRts | 6:24430 | DOI: 10.1038/srep24430 expressed DNA damage-related genes in HepG2 cells infected with Ad-N or Ad-HBx (Fig. 5b). The results of qRT-PCR were consistent with those of mRNA microarray analysis.
HBx induced genomic instability. To examine whether HBx induced genome-wide chromosomal aberrations, we got metaphase chromosome spreads of cells infected with Ad-HBx or Ad-N, or untreated, counted 20 metaphase spreads per group, and conducted 3 independent experiments. As shown in Fig. 6, there were fewer metaphase spreads with broken chromosomes in cells infected with Ad-N and untreated, whereas in contrast, the percentage was significantly increased in cells infected with Ad-HBx (p < 0.05). In HepG2 and SK-HEP-1 cells infected with Ad-HBx, the average at 48 h and 72 h post-infection were respectively 21.67 ± 2.89% and 26.67 ± 2.89%, and 15.00 ± 5.00% and 16.67 ± 5.77%. The results suggested that HBx induced genomic instability. HBx induced DNA damage. DNA damage can be evaluated by comet assay 18,19 , which was employed in this study to determine whether HBx could result in DNA damage. As shown in Fig. 7, at 48 h and 72 h post-infection, the nuclei from both cell lines infected with Ad-HBx showed higher damage as evidenced by increased percentage of tail DNA.
When DNA is damaged, H2AX becomes phosphorylated at serine 139, which is then called γ -H2AX 20 . Therefore, the expression of γ -H2AX was detected by Western blotting and immunofluorescence. As shown in Fig. 8a,b, HBx induced a time-dependent increase in γ -H2AX levels, but had no effect on expression of total H2AX. Figure 8c,d showed the γ -H2AX foci induced by HBx. The average foci numbers in HepG2 and SK-HEP-1 cells infected with Ad-HBx at 48 h and 72 h post-infection were respectively 28.67 ± 3.06 and 40.67 ± 0.58, and 32 ± 3.61 and 34.67 ± 4.16, which were significantly different from those in cells infected with Ad-N. Similarly, higher levels of γ -H2AX were detected in liver tissues of individuals infected with HBV compared with those of individuals uninfected with HBV by immunohistochemistry analyses (Fig. 8e). Collectively, our results suggested that HBx induced DNA damage.

Discussion
HBx has been reported to be associated with HBV-related hepatocarcinogenesis, thus the identification of the underlying mechanisms of HBx-mediated carcinogenesis is an important research topic. Metabolomics is an approach for rapidly identifying global metabolic changes in biological systems, and has been widely used for the diagnosis and evaluation of diverse diseases and therapies 21 . Here, an NMR-based metabolomics approach was applied to identify the distinguishing metabolites under HBx induction.
Glucose is the main source of energy and precursor for biosynthesis in cells 22 . Our results showed that the levels of glucose and its phosphorylated product, glucose-6-phosphate, were disordered after Ad-HBx infection. Meanwhile, lactate as the end product of glycolysis was also found to be abnormal in cells infected with Ad-HBx. While previous studies have reported HBV-induced glucose metabolism abnormalities 6,23 , to the best of our knowledge, this is the first study that reports the association of HBx and glucose metabolism abnormalities. Meanwhile, levels of a variety of amino acids such as glutamate, glutamine, creatine, alanine, threonine, glutamate, serine and glycine, were disordered after Ad-HBx infection. These distinguishing metabolites were involved in amino acid metabolism, suggesting dysfunction of amino acid metabolism in cells infected with Ad-HBx. These results are in agreement with the amino acid metabolism abnormalities of HCC and liver cirrhosis 5,10 .
We did not detect lipids in the current study, because only water-soluble substances were extracted in the research investigating metabolic profiles of cells by an NMR-based metabolomics approach. However, we found that the levels of N-acetylaspartate (NAA) and choline were significantly different in cells infected with Ad-N versus Ad-HBx. NAA is involved in fatty acid metabolism 24 , while choline is a precursor of phosphatidylcholine and has a key role in systemic lipid metabolism 25,26 . Therefore, the abnormal levels of NAA and choline under HBx induction indicated that HBx might also disrupt lipid metabolism. Our result was consistent with previous studies that show that HBx resulted in lipid accumulation and activated lipogenic genes 7,13 . Combined, these findings indicated that HBx could disturb lipid metabolism.
Intriguingly, we found that in cells infected with Ad-HBx, uridine, guanosine and inosine had high VIP scores and showed a significant decrease in concentration (p < 0.05). Meanwhile, uracil and xanthine also were important in providing class separation. Their concentration from 1 H-NMR was relatively low; however, they had high VIP scores, which piqued our attention. Uridine, guanosine, inosine, uracil and xanthine as important metabolic intermediates of nucleic acid metabolism, are involved in the biosynthesis of DNA and RNA. The EdU incorporation assay showed that HBx inhibited DNA synthesis. Overall, the results suggested that HBx could induce abnormalities of nucleic acid metabolism. A recent study showed that adipocyte DNA damage could aggravate metabolic abnormalities, and the abnormalities could be ameliorated by reduction of DNA damage 27 .
As is well known, DNA damage affects all DNA metabolic processes, for example, DNA replication and transcription 28 . Pal et al. reported that amelioration of DNA damage by resveratrol restored protein and nucleic acid metabolism in the brain 29 . Thus, whether the metabolic abnormalities are related to DNA damage remains to be clarified in future studies.
Our previous study showed that HBx was able to induce G2/M phase arrest in HCC cells 30 . G2/M check-point has an important role in the replication of genome. When DNA lesions arise, mitosis is prevented to minimize the detrimental effects and provide an opportunity to repair these genomic lesions 31 . So, we speculated that HBx could induce DNA damage. In the current study, gene-expression profiles showed that 29 genes involved in DNA damage response, nucleotide-excision repair, signal transduction in response to DNA damage, double-strand break repair, and DNA repair, were differentially expressed. Analyzing the chromosomal aberrations in the whole genome showed that HBx induced genomic instability. In addition, HBx-induced DNA damage was further demonstrated by comet assay, Western blotting, immunofluorescence, and immunohistochemistry. Numerous studies have reported that DNA damage plays an important role in carcinogenesis 32,33 . Under normal physiological conditions, eukaryotes can repair damaged DNA through the signal pathways of DNA damage response 34 . Our study, using mRNA microarray analysis, suggested that genes of HBx-induced DNA damage response and DNA repair were downregulated. The decreased gene expression might be due to the abnormalities of nucleic acid synthesis.  35,36 , and the cytotoxic activities of HBx were maintained when forming a complex with DDB1 in the nucleus 37 . In our study, we also confirmed that HBx colocalized with DDB1 mainly in nucleus in cells infected with Ad-HBx (Supplementary Figure S1). DDB1 is a subunit of an E3 ubiquitin ligase complex, which can recognize DNA lesions and plays an important role in DNA repair 35 . HBx has the binding site of DDB1 38 . So, we speculate that HBx-induced DNA damage may be due to its interaction with DDB1, which disrupts the DNA repair activities of DDB1.
In conclusion, we speculate that HBx first induces DNA damage, and then disrupts nucleic acid metabolism. The metabolism abnormalities block DNA repair and induce the occurrence of HCC. However, further investigations are needed to study how HBx-induced DNA damage disrupts nucleic acid metabolism. Generation of recombinant adenovirus. HBx expressing recombinant adenovirus (Ad-HBx) was prepared as previously described 30 . Briefly, cDNA coding HBx was cloned into pENTR11 and then the plasmid expressing HBx (pENTR11-HBx) was recombined with pAd using Gateway system (Invitrogen). Ad-N as a control adenovirus did not express a foreign gene. Virus was produced in 293A cells and purified by cesium chloride gradient centrifugation. HBx in our study was from Hepatitis B virus genome, subtype ayr, and its sequence was shown in supplementary data.

Chemicals and reagents. Dulbecco's modified Eagle medium (DMEM) was purchased from Gibco
Metabolite extraction. Cells were seeded in 6 cm dishes and grew to 80%~90% confluence, then were infected with Ad-HBx or Ad-N at an MOI of 20 in DMEM supplemented with 2% FBS. After 2 h, the cells in each  Scientific RepoRts | 6:24430 | DOI: 10.1038/srep24430 1 H-NMR spectroscopy. All NMR experiments were performed on Bruker AV III 600 MHz spectrometer (Bruker Biospin, Milton, Canada) equipped with an inverse cryoprobe operating at 600.13 MHz. All NMR spectra of samples were acquired using a standard Bruker noesygppr1d pulse sequence, and a total of 256 scans were collected into 32768 data points over a spectral width of 8000.00 Hz.

NMR data analysis. Identification and quantification of individual metabolites were performed using
Chenomx NMR Suite software (version 7.7, Chenomx, Edmonton, Canada). 1 H-NMR spectra were compared against Chenomx library that contained the unique 1 H-NMR spectra of each standard compound quantified by a known reference signal (DSS). Comparisons of NMR spectra with this database produced a list of compounds and their concentration, and the absolute concentration of each compound was normalized based on the weight of samples. A supervised partial least squares-discriminant analysis (PLS-DA) approach was chosen to compare the variance of metabolite concentration between Ad-HBx-48 and Ad-N-48 or Ad-HBx-72 and Ad-N-72. Student's t test was used to compare between groups for the discriminant variables obtained from PLS-DA and a p value < 0.05 was considered significant.

EdU incorporation assay.
To further study the effect of HBx on nucleic acid metabolism, we analyzed DNA synthesis using 5-ethynyl-20-deoxyuridine (EdU) incorporation assay. HepG2 and SK-HEP-1 cells infected with Ad-HBx or Ad-N, or untreated were seeded in 96 well plates and allowed to grow for 46 h or 70 h, then labeled by incubation with 50 μM EdU for 2 h. After fixation, cells were stained with Apollo ® fluorescent dye, Apollo ® 488, then counterstained with Hoechst33342. Labeling of proliferating cells was imaged by fluorescence microscopy.
Gene expression profiles. Total RNA was isolated using Trizol reagent (Invitrogen) from HepG2 cells infected with Ad-N or Ad-HBx according to the manufacturer's instructions. RNA was reverse transcribed with Cy3-labelled-CTP and fluorescence quantified using NanoDrop ND-1000 (Thermo Scientific). After hybridization, GeneChips were scanned by Agilent Microarray Scanner (Agilent p/n G2565BA).

Quantitative real time PCR (qRT-PCR).
Total RNA was isolated with Trizol from HepG2 cells infected with Ad-N or Ad-HBx according to the manufacturer's instructions and was then reverse transcribed with a cDNA Synthesis Kit (TaKaRa). The resulting cDNA was used for qRT-PCR analysis using SYBR Green (TaKaRa) with gene-specific primers, and data were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The qRT-PCR primers were shown in Table 2. After hypotonic treatment with 0.075 M KCl for 10 min at 37 °C, cells were fixed with 3:1 mixture of methanol:acetic acid thrice. Then, cells were dropped on precooled glass slides to obtain metaphase chromosome spreads. Chromosomes were stained with Giemsa and imaged by a microscope with a 100X objective lens.
Comet assay Comet assay was performed as previously described 39,40 . Briefly, HepG2 and SK-HEP-1 cells were infected with Ad-HBx or Ad-N, or untreated. At 48 and 72 h post-infection, the cells were collected and resuspended in PBS, and 10 μl single-cell suspensions was mixed with 75 μl 0.7% low-melting point agarose, layered onto a glass microscope slide and left in ice-cold lysis buffer (1 M NaCl, 1 mM EDTA, 10 mM Tris-HCl, 30 mM NaOH, 1% Triton X-100 and 10% DMSO, pH 10.0) for 2 h. All slides were placed in alkaline electrophoresis buffer (300 mM NaOH and 1 mM EDTA, pH 10.0) for 40 min, and electrophoresis was performed at 25 V for 20 min. Then the slides were neutralized with buffer (0.4 M Tris-HCl, pH 7.5) thrice, stained with propidium iodide (2.5 μg/mL) for 10 min and imaged by a fluorescence microscope.   Western blotting. Western blotting was performed as previously described 41,42 . Briefly, HepG2 and SK-HEP-1 cells were infected with Ad-HBx or Ad-N, or untreated. At 48 and 72 h post-infection, the cells were collected, washed in PBS, and lysed in RIPA buffer containing 1 mM PMSF. Total protein concentration was quantified with the Enhanced BCA Protein Assay Kit. The cell lysates were separated by SDS-PAGE and transferred to polyvinylidene fluoride (PVDF) membranes. The membranes were blocked in Tris-buffered saline containing 0.1% Tween-20 (TBS-T) and 5% nonfat milk for 2 h at room temperature, and incubated with primary antibodies against HBx, γ -H2AX (phosphorylated on Ser139), H2AX and β -actin overnight at 4 °C, followed by incubation with horseradish peroxidase-conjugated secondary antibodies at 37 °C for 1h. The blots were detected using the enhanced chemiluminescence system (Millipore).
Immunofluorescence. Cells infected with Ad-HBx or Ad-N, or untreated were fixed with 4% paraformaldehyde in PBS for 20 min, then permeabilized with 0.5%Triton X-100 in PBS for 15 min and blocked with 5% goat serum for 30 min at room temperature as specified in a previous report 43 . For detection of the γ -H2AX expression, the fixed cells were incubated with anti-γ -H2AX antibody overnight at 4 °C followed by incubation with TRITC-conjugated goat anti-rabbit IgG for 1 h at 37 °C, and stained with DAPI to visualize the cell nucleus. The images were obtained by using a fluorescence microscope.
Immunohistochemistry. Human tissue arrays were provided by Shanghai Biochip Co., Ltd. (Shanghai, China) including liver tissues from 20 patients with or without HBV infection. DNA damage was evaluated by immunohistochemistry using anti-γ -H2AX antibody, and the ZSGB-BIO SPlink Detection Kit (Beijing, China). After standard dewaxing, washing, neutralization of endogenous peroxidase and heat-induced antigen-retrieval, tissues were blocked with goat serum for 15 min at room temperature and then incubated with anti-γ -H2AX antibody overnight at 4 °C followed by incubation with biotin-conjugated goat anti-rabbit IgG for 15 min at 37 °C . Subsequently, the sections were treated with horseradish peroxidase-conjugated streptavidin, visualized by incubation with 3, 3′-diaminobenzidine (DAB) and counterstained with hematoxylin.