This article is mentioned in:

Introduction

HBV-related HCC cases are not necessarily characterized by chronic inflammation, the mechanism of hepatocarcinogenesis is believed to involve both host and viral factors, particularly those related to sites wherein the viral genome has integrated into the host genomes (1–4). Integration of HBV genomic elements reportedly leads to carcinogenesis via destabilization of human chromosomes, altered expression of genes in the vicinity of integration sites, and production of chimeric human-HBV proteins via expression and translation of integrant-proximal sequences. Furthermore, the fusion transcript of HBV X protein (HBx) and long-interspersed nuclear element 1 reportedly functions as a long noncoding RNA and affects Wnt signaling (5).

Next-generation sequencing (NGS) has revolutionized human genomic analysis in a variety of fields, including disease genomics. However, given that host integration sites often contain numerous LINE repeats, short interspersed nuclear elements (SINE), and other transposable elements, short-read sequencers sometimes fail to accurately identify integration sites, making integration analysis more challenging. Moreover, because host integration sites include pseudogenes, failure to obtain sufficiently long read lengths and accurate sequence information makes proper identification of these sites difficult, regardless of high read number yields and the use of paired-end sequencing technology (6).

In our previous study using custom-made HBV-specific bait, we selectively captured genomic fragments containing HBV integrants from DNA extracted from HBV-infected HCC cell lines, allowing for the development of an efficient NGS-based integration-analysis methodology (NGS-based structural methylation analysis of virus genome integration; G-Navi) (7). G-Navi analysis using HBV-infected liver cancer cell lines [PLC/PRF/5 and Hep2.2.15 (HepG2 cells infected with HBV)] enabled the discovery of integrated HBV fragments not only in the nuclear genome but also in the mitochondrial genome of HepG2.2.15 cells. Unfortunately, previous primary NGS technologies were limited in their ability to provide sufficiently detailed analyses, even when combined with the G-Navi method. In particular, these methods exhibited a limited capacity to distinguish pseudogenes containing repetitive or highly similar sequences. However, advanced NGS methods do not require polymerase chain reaction (PCR) pretreatment of sequenced samples and employ single-molecule real-time (SMRT) sequencing technology that can read single bases at the molecular level in real-time. Using these ‘third-generation’ NGS machines capable of achieving high-fidelity reads without PCR amplification and the attendant GC bias that it introduces, we aimed to test the hypothesis that their incorporation into G-Navi analysis would facilitate detailed integration analysis of the HepG2.2.15 cell line and especially its mtDNA.

Materials and methods

Cell lines

The PLC/PRF/5 (Alexander) and HepG2 human liver cancer cell lines were purchased from the Japanese Collection of Research Bioresources (JCRB, Tokyo, Japan) (8). These cell lines have been authenticated using STR profiling in the JCRB. HepG2.2.15 cells (genotype D) authenticated using STR profiling were provided by Professor Stephan Urban of the University Hospital Heidelberg (9,10).

SMRT DNA sequencing-based HBV DNA-integration analysis

DNA extraction was performed using the standard phenol-chloroform method. The integrity of the extracted DNA was assessed by 0.8% agarose gel electrophoresis, and concentration was measured with a Qubit 2.0 Fluorometer (Thermo Fisher Scientific K.K., Waltham, Ma, USA). We used the SureSelect target enrichment system (Agilent Technologies) along with 12 391 custom baits covering the DNA sequences of HBV genotypes A through J and PLC/PRF/5 HBV sequences. HBV DNA fragments were selectively captured using unique baits in a sequence-dependent manner. The resulting HBV integrant (HBV plus human genome) DNA fragments were used for downstream analysis via SMRT DNA sequencing. The concentration of the final library was determined on an Agilent 2100 Bioanalyzer. Libraries were prepared with the PacBio DNA Template Prep Kit 1.0 (Pacific Biosciences, Menlo Park, CA, US, 100-259-100) and the PacBio DNA/Polymerase Binding Kit P6 (Pacific Biosciences,100-372-700), and Sequencing was performed using C4 chemistry (DNA sequencing Reagent 4.0, Pacific Biosciences) and PacBio RSII platform (Pacific Biosciences) with an on-plate loading concentration of 0.15 pM. Data analysis was performed by CLC Genomics Workbench software 10.1.1 (Qiagen, Hilden, Germany, http://digitalinsights.qiagen.com/).

Reverse transcription (RT)-PCR

RNA extraction was performed using TRIzol RNA Isolation Reagents (Thermo Fisher). For cDNA synthesis, 100 ng of total RNA was used with SuperScript™ III First-Strand Synthesis System (Thermo Fisher). Real-time quantitative RT-PCR was performed using SYBR green and targeting mitochondrially encoded cytochrome C oxidase III (MT-CO3), hepatitis B pre-S1 protein (HB PreS1), hepatitis B core/capsid protein (HBc), and HBx. The PCR amplification conditions were: An initial denaturation step of 20 s at 95°C, followed by 45 cycle of 3 s at 95°C and 30 s at 60°C. All qPCR reactions were performed in triplicate on an ABI 7500 fast (Applied Biosystems, Foster City, Ca, USA). The qPCR data analysis was performed using the 2-ΔΔCq (Livak) method, a widely used method for relative gene expression analysis, corrected for the housekeeping gene Actin Beta (ACTB) (11). Three cell lines were analyzed using the following primers: MT-CO3 forward 5-CCCACCAATCACATGCCTAT-3 and reverse 5-GTGGCCTTGGTATGTGCTTT-3; HB PreS1 forward 5-GGGTCACCATATTCTTGGGAAC-3 and reverse 5-CCTGAGCCTGAGGGCTCCAC-3: HBc forward 5-CTGGGTGGGTGTTAATTTGG-3 and reverse 5-TAAGCTGGAGGAGTGCGAAT-3; and HBx forward 5-CACTTCGCCTCACCTCTG-3 and reverse 5-TCGGTCGTTGACATTGCT-3 and ACTB forward 5-TCCTTCCTGGGCATGGAGT-3 and reverse 5-CAGGAGGAGCAATGATCTTGAT-3.

Immunofluorescence analysis

Cell fixation was performed using 4% Paraformaldehyde Phosphate Buffer Solution for 10 min at room temperature (RT). Blocking was performed with 2.5% normal horse serum (Vector Laboratories, Inc., Burlingame, CA, USA) for 1-h at RT. Immunofluorescence analysis of MT-CO3, hepatitis B surface protein (HBs), and HBc was performed using 4-well Millicell® EZ slides (Merck Millipore, Ltd., Carrigtwohill, Ireland) with one of the following primary antibodies for overnight at 4°C: rabbit polyclonal anti-MT-CO3 (1:200; Cat#55082-1-AP; Proteintech, Rosemont, IL, USA), rabbit polyclonal anti-HBsAg (1:200; Cat#NB100-62652; NOVUS Biologicals, Littleton, CO, USA), and mouse anti-HBcAg (1:100; Cat#ab8638; Abcam, Cambridge, MA, USA). As secondary antibodies, Vecta Fluor Excel Amplified Anti-Rabbit IgG, DyLight 488 Antibody Kit(Vector Laboratories, Inc.), and Vecta Fluor Excel Amplified Anti-Mouse IgG, DyLight 488 Antibody Kit(Vector Laboratories, Inc.) were used. Incubated at room temperature for 15 min with Amplifier Antibody, followed by 30 min at room temperature with VectaFluor Reagent. Cells were observed using an all-in-one Fluorescence Microscope (KEYENCE Corp., Osaka, Japan).

Western blotting

Cells were lysed in a 0.5% NP40 Lysis Buffer (final concentration 50 mM Tris-HCL, 150 mM NaCl, 0.5% NP-40 50 mM NaF) in the presence of the complete protease inhibitor cocktail (Roche, Basel, Switzerland). Protein concentrations were determined with the BCA protein assay kit (Thermo Fisher Scientific K.K., Waltham, Ma, USA). For SDS-PAGE, samples were loaded 20 µg using 4 × NuPAGE LDS Sample Buffer and NuPAGE 10% Bis-Tris Gel(Thermo Fisher). Western blot analysis was performed using the iBlot2 Gel Transfer Device (Thermo Fisher) and PVDF membranes(Thermo Fisher). The blocking reagent was Blocking One (NACALAI TESQUE, INC., Kyoto, Japan) and incubated at RT 1-h. Primary antibodies were rabbit polyclonal anti-MT-CO3 (1:300; Cat#55082-1-AP; Proteintech) and mouse monoclonal anti-cytochrome c (1:2,000; Cat#NB100-56503SS; NOVUS Biologicals) with COX IV Antibody (1:1,000; Cat#4844; Cell Signaling Technology, Inc.) was used as the reference proteins, mouse monoclonal anti-HBsAg (1:1,000; Cat#ab20758; abcam), and mouse anti-HBcAg (1:1,000; Cat#ab8638; Abcam) with Tubulin Antibody (1:1,000; Cat#4844; Cell Signaling Technology, Inc.) was used as the reference proteins and incubated at 4°C overnight. ECL Anti-Mouse IgG, Horseradish Peroxidase Linked Whole Antibody (1:10,000; cytiva, MEL, USA) and ECL Anti-Rabbit IgG, Horseradish Peroxidase Linked Whole Antibody (1:10,000; cytiva) was used as a secondary antibody and incubated at RT 1-h. Detected with LAS3000 (Fuji Photo Film Co., Ltd. Tokyo, Japan) using ECLSelect™ Western Blotting Detection Reagent (cytiva).

Cell viability assay

HepG2, HepG2.2.15, and PLC/PRF/5 cells were seeded in a 96-well plate at a density of 4000 cells/well. Cells were evaluated using a Cell Counting Kit-8 (Dojindo Molecular Technologies, Rockville, MD, USA) according to the manufacturer's instructions (12). Absorbance was measured at 450 nm after 4-h incubation at 37°C in a CO2 incubator.

Lactate assay

Measurement of lactate concentrations in the medium of HepG2, HepG2.2.15, and PLC/PRF/5 cultures was performed using a lactate assay kit (WST; Dojindo Molecular Technologies) and a microplate reader (Multiskan Ex; Cat#51118230; Thermo Fisher Scientific) according to the manufacturer's instructions (13).

Quantitative pyrosequencing methylation analysis

DNA methylation of the integrated HBV genome as well as the adjacent human mitochondrial genome was analyzed by bisulfite pyrosequencing (allele-specific and orthologous loci DNA methylation analysis). Bisulfite PCR was performed using the EpiTect Bisulfite Kit (Qiagen N.V., Venlo, NLD) according to the manufacturer's protocol. One microliter of bisulfite-treated DNA was used as the template. Primers used for methylation analysis of mitochondrial genome-integration sites in HepG2.2.15 cells were as follows: biotinylated forward primer 5-TGTAGTATGGTGAGGTGAATAATGT-3 and reverse primer 5-CCCRCTAAATCCCCTAAAAATCCCACTC-3; sequencing primer-1 5-GTTTAGGAGATTTTAAGGTTTT-3 and sequencing primer-2 5-AGGTGATTGATATTTTTGATG-3. Methylation levels of orthologous mitochondrial genome loci in HepG2.2.15 cells at the same (empty) mitochondrial target sites as those in HepG2 and PLC/PRF/5 cells were analyzed using bisulfite pyrosequencing. Primers used for methylation analysis were as follows: biotinylated forward primer 5-TAGATTATGGTGAGTTTAGGTGATTGATAT-3 and reverse primer 5-ATTAAAAAAACACTAACCCCCAACAAACA-3; and sequencing primer 5-GTTTAGGTGATTGATATTTTTG-3. Analyses were performed using touchdown PCR, with denaturation at 95°C for 30 s, annealing at the 95°C for 30 s, and extension at 72°C for 30 s. The PCR products were confirmed by electrophoresis using a 2% agarose gel. Ten microliters of biotinylated strands were then captured on streptavidin-coated beads (cytiva) and incubated with sequencing primers. The pyrosequencing reactions were performed using the PyroMark Q24 Advanced (Qiagen). The resulting pyrogram was analyzed with the PyroMark Q24 software version 3.0.0 (Qiagen).

Statistical analysis

All statistical analyses were performed using SPSS for Windows (v.12.0; SPSS, Inc., Chicago, IL, USA) and PRISM for Windows (v.7.0; GraphPad Software, San Diego, CA, USA), and free software R (v. 4.2.1; R Development Core Team, Vienna, Austria). Integration-site analysis and tree-view analysis were performed using Geneious Prime software (v.2019.2.3; (Biomatters Ltd) and data are presented as the means ± standard error of the mean.

For comparisons involving the three groups, the Bonferroni correction to the Kruskal-Wallis test (multiple comparison test) was performed. For MT-CO3 expression analysis in the four groups of HepG2, HepG2.2.15, PLC/PRF/5, and HepG2.2.15 (HBV integrant), the Kruskal-Wallis test was also performed and a Bonferroni correction (multiple comparison test) was performed. All reported P-values were two-sided, and a P<0.05 was considered significant.

Results

NGS combined with G-Navi increases the accuracy of integration-site identification

We used our previously described method (G-Navi)7 along with the PacBio RSII sequencer (Pacific Biosciences) to identify HBV-integration sites. For efficient genome analysis, we synthesized 12 391 custom baits based on the sequences of HBV genotypes A through J (7). The DNA fragment length obtained by SureSelect Target enrichment system was approximately 1500 bp.

NGS analysis using genomic DNA from three cell lines (HepG2, HepG2.2.15, and PLC/PRF/5) revealed a total read number ranging from 0.61 to 0.97 million, with an average read-quality score of 0.848 (raw: 0.532), corresponding to a >99.9% accuracy. We first constructed a consensus sequence using both polymerase reads and subreads, followed by mapping of the sequencing data using the UCSC Genome Browser (https://genome.ucsc.edu/index.html). The results showed that the use of G-Navi combined with the PacBio RSII sequencer revealed 49-fold more bases than conventional NGS analysis.

Identification of an HBV integrant in MT-CO3 from a similar mitochondrial pseudogene

Using an assembly of NGS data containing the human genome (Human GRCh38/hg38; http://genome.ucsc.edu/index.html), the HBV genome (AB205126 genotype D), and the mitochondrial genome (NC_012920), we determined that an HBV integrant in the HepG2.2.15 cell line (HBV genotype D) was present in MT-CO3 (NC-012920; chromosome M, 9652), representing an HBV gene fragment (AB205126; preCore, X gene, 1080–1804) with a common homology sequence of ‘CACCA’ (Fig. 1A-C). Alignment analysis revealed that the HBV fragment integrated into the mitochondrial genome was not the full-length genome, but rather the contig from HBx (1804) to the fragment encoding the hepatitis B preCore protein (HB preCore; 1080) (Fig. 1A and C). We did not observe any other HBV genome fragments integrated into the mitochondrial genome. Moreover, due to the high subread (965 820 reads) and base-pair (1 699 868 394 bp) yields combined with the high mean read length (19 102 bp) returned by PacBio RSII sequencing, we were able to distinguish among several highly similar MT-CO3 pseudogenes (MT-CO3Pxx) and the true MT-CO3 sequence by manipulating the consensus sequences used for NGS analysis (Figs. 1D, and 2A and B).

Figure 1. Schematic representation of NGS data. (A) Overview of NGS data from the HepG2.2.15 cell line (inner circle represents the HBV genome, and the outer diagram represents the human genome). Although integration into various nuclear genes is apparent, it occurred at only one locus in the mitochondrial genome. The boundary region between the mitochondrial HBV integrant and the human mitochondrial genome is depicted at the sequence and gene levels. (B) Boundary between the HBV integrant and the human mitochondrial genome shares a common ‘CACCA’ motif. (C) In the mitochondrial genome, integrant insertion initiates in the middle of exon 1 of the MT-CO3 gene, and in the HBV fragment, in the middle of the HBx gene. (D) Difference in first-generation NGS analysis (using the GS FLX Titanium system) and third-generation NGS analysis (using a PacBio RSII sequencer). HBV, hepatitis B virus; HBx, HBV X protein; MT-CO3, mitochondrially encoded cytochrome C oxidase III; NGS, next-generation sequencing.

Figure 2. Existence of nuclear pseudogenes of MT-CO3 at various loci. (A) Representation of differences in sequence homology between MT-CO3 and the pseudogenes according to TreeView. These pseudogenes include both the MT-CO3 gene and other mitochondrial genes. Letters A-O represent adjacent mitochondrial genes. (B) Primary mitochondrial genes are designated as A-O. (C) A cell containing mitochondrial genomes into which HBV has integrated, as well as non-integrated mitochondrial genomes. Pyrosequencing analysis of MT-CO3 methylation in the non-integrated mitochondrial genomes of HepG2.2.15, HepG2 and PLC/PRF/5 cells. (D) Pyrosequencing analysis of methylation on both sides (the HBV side and the human mitochondrial side) of the integrant border. CpG sites at which MT-CO3 methylation was measured were the same in (C and D) HBV, hepatitis B virus; MT-CO3, mitochondrially encoded cytochrome C oxidase III.

MT-CO3 expression levels are higher in HepG2.2.15 than HepG2 cells

MT-CO3 expression levels in HepG2.2.15 were significantly higher than in HepG2 cells and also higher in PCL/PRF/5 cells (P<0.0001) (Fig. 3A). However, the expression level at a specific integration site in HepG2.2.15 cells was lower than the difference in expression levels between HepG2.2.15 and HepG2 cells (Fig. 3A). Immunofluorescence staining confirmed MT-CO3 levels in the mitochondria of all three cell lines (Fig. 3C). Furthermore, a comparison of HBV expression levels in PLC/PRF/5 and HepG2.2.15 cells indicated higher levels of HB PreS1 and HBc in HepG2.2.15 cells (P<0.0001) (Fig. 3B). However, these results did not show mitochondrial expression, instead revealing levels throughout the somatic cells, including the nucleus (Fig. 3B).

Figure 3. Expression analyses in three different hepatic cell lines. (A) There was a significant difference in MT-CO3 expression between HepG2 and HepG2.2.15 cells. Three groups were compared using the Kruskal-Wallis test followed by the Bonferroni multiple comparison test. (B) In HepG2.2.15 cells, HBV-integrant expression included HB PreS1 and HBc. HepG2 cells without HBV integration were used as a control. Gene expression of HBVs and HBc was analyzed in triplicate, and since the Kruskal-Wallis test showed significant differences among the three groups (HepG2, HepG2.2.15 and PLC/PEF/5 cells), multiple comparisons were then performed using the Bonferroni post hoc test. Asterisks indicate significant differences (*P<0.05, **P<0.01, ****P<0.0001). (C) Expression and localization of MT-CO3 in each cell line according to immunofluorescence staining. MT-CO3 was highly expressed in the mitochondria of HepG2.2.15 compared with other cell lines. Although HBs and HBc expression was apparent in the HepG2.2.15 cell line, they localized overlapping with mitochondria. Scale bars, 100 µm. HBc, hepatitis B core/capsid protein; HBs PreS1, hepatitis B surface protein; HBV, hepatitis B virus; MT-CO3, mitochondrially encoded cytochrome C oxidase III.

HepG2.2.15 cells display higher proliferative capacity relative to HepG2 cells

Cell viability assays revealed that HepG2.2.15 cells showed higher proliferative capacity than HepG2 cells (P=0.01) (Fig. 4A). Additionally, mitochondrial MT-CO3 levels were higher in HepG2.2.15 cells (P=0.0002) and cytochrome C levels were also higher in HepG2.2.15 (P=0.0007) (Fig. 4B). HB PreS1 and HBc levels were not significantly different between HepG2.2.15 and PLC/PRF/5 cells (P=0.06 for HB PreS1 and P=0.1 for HBc). These results may reflect the fact that we used total protein extracts but not mitochondrial protein only (Fig. 4B). Moreover, lactate assays indicated that HepG2.2.15 cells displayed significantly greater lactic acid production relative to that observed in HepG2 cells (P=0.0003) (Fig. 4C).

Figure 4. (A) Viability analysis of the three cell lines. HepG2.2.15 cells exhibited higher proliferation than HepG2 cells. Cell viability counts were analyzed in triplicate and statistical analysis was performed to compare the HepG2, HepG2.2.15 and PLC/PEF/5 groups on day4. (B) MT-CO3 protein levels in mitochondria were compared by western blotting and were found to be higher in HepG2.2.15 cells than in HepG2 cells. Cytochrome C levels in mitochondria of HepG2.2.15 cells were higher than those of HepG2 cells. There were no significant differences between HepG2.2.15 and PLC/PRF/5 cells for HB PreS1 and HBc in somatic cells, including the nucleus (HBc; P=0.066, HBs; P=0.079). Protein levels of MT-CO3 and cytochrome C in mitochondria and HB PreS1 and HBc in somatic cells, including the nucleus, were analyzed in triplicate. (C) Analysis of lactate acid production. The color of the culture medium was more yellow in HepG2.2.15 cells compared with HepG2 cells and PLC/PRF/5 cells, indicating that HepG2.2.15 produced the most lactic acid when analyzed for lactic acid production. Data are presented as the mean ± standard error of the mean. In experiments with three groups, data were analyzed using the Kruskal-Wallis test followed by the Bonferroni multiple comparison test. Asterisks indicate significant differences (*P<0.05, **P<0.01, ***P<0.001). COXIV, cytochrome c oxidase subunit IV; HBc, hepatitis B core/capsid protein; HBs, hepatitis B surface protein; MT-CO3, mitochondrially encoded cytochrome C oxidase III.

Identification of MT-CO3 pseudogenes (MT-CO3Pxx) in nuclear DNA

Using information obtained from the UCSC Genome Browser (Human GRCh38/hg38; http://genome.ucsc.edu/index.html), we identified the existence of several nuclear MT-CO3 pseudogenes with sequences highly similar to that of the original gene. MT-CO3 pseudogenes were found at a total of 46 loci (Fig. 2A and B), ranging from MT-CO3 P6 (chromosome 3) exhibiting the highest sequence similarity with MT-CO3 to single CO3 sequences and constructs existing as mitochondrial genome complexes. The data obtained from G-Navi-NGS analysis enabled the distinction between MT-CO3 and pseudogenes.

The MT-CO3/HBV integrant shows concordance and high methylation levels in HepG2.2.15 cells

G-Navi-NGS analysis was able to not only detect HBV integrants at MT-CO3 (MT-CO3/HBV) in HepG2.2.15 cells, it also confirmed the results of direct sequencing analysis. The GC percentage of the integrated MT-CO3/HBV boundary (500 bp) in HepG2.2.15 was less than 50% (GC: 48.4%; A: 28.2%; C: 17.0%; G: 31.4%; T: 23.4%) located in non-promoter regions of both MT-CO3 and HBx preCore genes. We also found mitochondrial heteroplasmy in HepG2.2.15 cells, containing both nonintegrated and MT-CO3/HBV integrated mitochondria. DNA methylation analysis demonstrated that nonintegrated HepG2.2.15 MT-CO3 loci exhibited low DNA methylation (3.0%), similar to levels in HepG2 and PLC/PRF/5 cells (4.3 and 4.0%, respectively) (Fig. 2C). In contrast, MT-CO3/HBV loci in HepG2.2.15 cells showed high levels of methylation throughout both HBV and MT-CO3 (92.3 and 95.5%, respectively) based on semiquantitative pyrosequencing analysis (Fig. 2D).

Discussion

Underlying persistent HBV infection, chronic hepatitis, and cirrhosis that precede initial infections along with chronic inflammation associated with long-term infections are suggested as factors affecting hepatocarcinogenesis. These processes are believed to foment the accumulation of genetic and epigenetic multistage gene alterations via the involvement of reactive oxygen species produced inside hepatocytes as a response to chronic inflammation (14–16). However, because HBV-related hepatocarcinogenesis is not necessarily characterized by long-term chronic inflammation, it is widely accepted that the carcinogenic mechanism involves viral factors, particularly the HBV integrants in the human genome (17–21). Development of NGS technologies has enabled comprehensive detection of integration sites in host genomes. However, these methods have lacked the sophistication to facilitate efficient and detailed analysis, especially for repetitive and approximated sequences (22,23). We previously performed NGS-based HBV-integration analysis in an HBV-infected liver cancer cell line (HepG2.2.15) and found integrants in the mitochondrial genome as well as the nucleus (4). However, many mitochondrial pseudogenes, which are highly similar in sequence to the original genes on the mitochondrial genome, are located on human chromosomes overlapping each other by as long as 30,000 bp. Therefore, conventional short-lead NGS, with its average read length of 300 bp, has precluded a more detailed analysis of the mitochondrial genome and the nuclear genome. Recently, studies focusing on the relationship between mtDNA damage and carcinogenesis clarified that most of the proteins constituting mitochondria have been encoded by nuclear DNA (pseudogenes) (7,24). Consequently, we next tried to analyze HBV/mtDNA integrants using high specification of NGS following the G-Navi method.

Recent advances in NGS technology have led to the development of machines capable of generating read depth, length, and bases sufficient for de novo genome sequencing. To investigate the integration of HBV genetic material into the mitochondrial genome of HepG2.2.15 cells, we altered our NGS protocol originally designed for the Roche 454 GS FLX Titanium system (discontinued in 2013 by Roche, Basel, Switzerland) to work with the PacBio RSII sequencer and applied this platform with a focus on the mitochondrial genome.

Our G-Navi-NGS analysis revealed only one integration locus in the mitochondrial genome (NC_012920) of HepG2.2.15 cells. The integrant was located in exon 1 of MT-CO3 and contained a shared microhomology sequence ‘CACCA’ (bases 9648–9652 in MT-CO3) as the boundary between the mitochondrial and HBV genomes. The integrant contained portions of the HBV preCore, X gene sequences, with the immediate vicinity (~50 bp) of the boundary highly similar across all contigs obtained from NGS. We also found both non-integrant and HBV/MT-CO3 integrants of mitochondria in HepG2.2.15 by pyrosequencing analysis (mitochondrial heteroplasmy). This suggested that in HepG2.2.15 mitochondria, the HBV/MT-CO3 integration site was not repaired and partially retained the integrants.

Mitochondria are organelles that generate energy through oxygen absorption and possess a 37-gene genome that encodes enzymes responsible for respiratory function. In addition to mitochondrial diseases, genome abnormalities not only cause diabetes, neurodegenerative diseases, and cancer, but are also implicated in age-related tissue alterations (15,16). However, investigation of the mitochondrial genome is difficult, hampering detailed investigations necessary to elucidate the impact of alterations and the mechanisms by which diseases of this organelle arise. Furthermore, mutations in mitochondrial DNA from HCC cells have been identified and implicated in mitochondrial dysfunction, although the associated mechanisms have not yet been determined (25).

To clarify this mechanism, we next compared cell viability and lactate assays between HepG2 and HepG2.2.15, given the significantly increased cell viability and lactic acid production of HepG2.2.15 cells. Thus, we hypothesized that HBV genome integrants (especially HBV/mtDNA integrants) in mtDNA might functionally affect the mitochondrial genome; however, we were unable to clarify the association due to discrepancies in gene expression and the result of epigenetic modifications (DNA methylation in the boundary).

Immunofluorescence staining showed higher MT-CO3 staining in HepG2.2.15 than in HepG2 and PLC/PRF/5 cells; protein levels showed higher MT-CO3 and cytochrome C in mitochondria in HepG2.2.15 cells relative to HepG2 cells. MT-CO3 expression in HepG2.2.15 was significantly higher than that in HepG2 cells; however, it is not a comparable different levels for HBV/MT-CO3 integrant specific expression levels in HepG2.2.15.

A variety of spontaneous mutations and abnormal DNA-methylation states are present in cancer cells, including those associated with hepatocellular carcinoma (26–28). Moreover, a previous study discussed the extent of epigenetic modifications caused by gene integration (29). To elucidate the effects of the MT-CO3/HBV integrant on the mitochondria of HepG2.2.15 cells, we analyzed epigenetic modifications in somatic cells. We hypothesized that these modifications might be similar to those occurring in hepatic cells undergoing HBV integration (7). Given that no reports discussing this process in the mitochondrial genome exist, we performed this analysis and found that non-integrated MT-CO3 was unmethylated across three cell lines (HepG2, HepG2.2.15, and PLC/PRF/5), whereas MT-CO3/HBV was highly methylated on both the HBV and MT-CO3 sides.

Many MT-CO3 pseudogenes exist on various chromosomes. To account for their possible effects on CO3 levels, we statistically analyzed the sequence homology of all 46 pseudogene loci (MT-CO3Pxx on autosomal chromosomes and chromosome Y) according to information accessed via the human UCSC genome browser graphical viewing tool (https://genome.ucsc.edu/).

Taken together, our results identify continuous preservation of the integration of HBV genetic material into the mitochondrial MT-CO3 gene of HepG2.2.15 cells as a mitochondrial heteroplasmy. HepG2.2.15 cells showed higher expression, immunostaining, and protein levels of MT-CO3, as well as higher proliferative capacity and lactate production. Thus, we hypothesized that HBV/MT-CO3 integrants may functionally affect the mitochondrial genome; however, the HBV/MT-CO3 boundary was already highly methylated on both the HBV and MT-CO3 sides. We speculate that epigenetic modification (DNA methylation) might be affected HBV/MT-CO3 silent; however, the detailed mechanism of the epigenetic mitochondrial modifications remains unknown.

Acknowledgements

The authors would like to thank Professor Stephan Urban (Department of Infectious Diseases, Molecular Virology at Heidelberg University, Heidelberg, Germany) for the cell lines.

Funding

This work was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (#16K09295, #18K15764 and #19H03568), the Suzuken Memorial Foundation (#17-032) and a Glaxosmithkline Research Grant 2018 (#E-7).

Availability of data and materials

The raw sequence datasets generated and/or analyzed during the current study are available in the DDBJ Sequence Read Archive (DDBJ; http://ddbj.nig.ac.jp/public/ddbj_database/dra/fastq/DRA010/DRA010456/). The other datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Authors' contributions

RO substantially contributed to the conception and the design of the study, and was involved in data acquisition and analysis. YW performed SMRT DNA sequencing-based HBV DNA-integration analysis and the bioinformatics analysis, and was a major contributor in writing the manuscript. HYa substantially contributed to the conception and the design of the study, and contributed to manuscript drafting. HYo and FI were involved in analysis and interpretation of the data, and critically revised the intellectual content. RO, YW, HYo, HYa, and FI confirm the authenticity of all the raw data. All authors have read and approved the final manuscript.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References