Different types of viral‑host junction found in HBV integration breakpoints in HBV‑infected patients.

The present study surveyed the characteristics of hepatitis B virus (HBV) integration in the liver genomes of patients with acute hepatitis B (AHB), carriers of inactive hepatitis B surface antigen (HBsAg), and patients with chronic hepatitis B (CHB) receiving antiviral treatment. 'Short‑read' whole genome sequencing (WGS) with an average of 4,879x coverage for HBV integration was performed in three patients with AHB, two carriers of inactive HBsAg, and 13 patients with CHB receiving antiviral treatment. Conventional polymerase chain reaction and Sanger sequencing were used to verify integration breakpoints supported by at least two paired‑end reads, and viral‑host chimeric transcripts were surveyed simultaneously. HBV integration breakpoints were 100% identified with an average of 138.2±379.9 breakpoints per sample. The numbers of HBV integration breakpoints were positively associated with the sequencing depth coverage numbers and levels of intrahepatic covalently closed circular DNA, respectively (P<0.0001 and P<0.0001). Four types of viral‑host junction in 14 HBV integration breakpoints were detected (two viral junctions mapped in the HBs gene, one in the Precore gene, and others within the HBx gene): Forward simple junction, reverse simple junction, forward and reverse complicated junction, and microhomology were found in many of the junctions. Expression of viral‑human chimeric transcripts was observed in several breakpoints, including the HBs gene. As a result, HBV can integrate into the host gene in the same manner as non‑homologous end joining and microhomology‑mediated end joining with numerous sites, and a close association may exist between HBV integration and patient prognosis. HBx integration may be indispensable for viral‑host chimeric transcription and HBsAg may be produced from integrated DNA.

Introduction

Hepatitis B virus (HBV) infection represents a major risk factor for the development of hepatocellular carcinoma (HCC), ranking fifth in global cancer incidence and representing the third leading cause of cancer-associated mortality (1). Following the infection of hepatoctyes, HBV relaxed-circle DNA (rcDNA) is transferred to the nucleus, where it forms covalently closed circular DNA (cccDNA). Within infected cells, the pregenomic RNA is then transcribed from the cccDNA and transported to the cytoplasm, where the mature capsids of the rcDNA are reverse transcribed and either secreted from the cells or returned to the nucleus to form the cccDNA pool. During this process, some HBV DNA genes integrate into the host chromosomal DNA. HBV integration is suspected to be one of the most important etiological events in HBV-induced HCC (2).

Previous isolation of HBV integration sites using polymerase chain reaction (PCR)-based methods, including Alu-PCR or ligation-mediated PCR, has suggested that the HBV insertional sites occur randomly throughout the genome, leading to the suggestion that there are no preferential integration sites with little oncogenic annotation (3–5). Through the application of whole genome sequencing (WGS), a large cohort of integration sites have been identified and, among them, a number of hotspots have been found, including TERT, MLL4, KMT2B, CCNE1, and FN1. This suggests that HBV may have preferential integration sites associated with distinct biological consequences, for example, altering the function of endogenous genes, causing genetic damage and chromosomal instability, which may lead to tumorigenesis in HBV-infected patients (6–9). In the present study, according to WGS detection in patients with HBV infection, several types of viral-host junction were we found in HBV integration breakpoints. Certain expression characteristics of viral-human chimeric transcripts may assist in further understanding the molecular mechanisms of HBV integration and its role in hepatocarcinogenesis.

Materials and methods

Patients and samples

Liver biopsy specimens were collected from a total of 18 patients from Renmin Hospital of Wuhan University (Wuhan, China) between March and December 2016, comprising 16 men and two women, aged between 21 and 43 (32.16±7.11) years old. The patients included three patients with acute hepatitis B (AHB) achieving virological seroclearance (VR) spontaneously (group 1), two inactive HBsAg carriers (group 2), 13 patients with chronic hepatitis B (CHB) receiving nucleos(t)ide analogs (adefovir, 10 mg/day or lamivudine, 100 mg/day) either as monotherapy or with pegylated-interferon (IFN)α (100 µg/week). Among the patients with CHB, six patients had primary treatment failure (group 3), five patients had achieved VR (group 4), and two patients had achieved both VR and HBsAg seroclearance and had not relapsed for >6 months (group 5) (10). Every enrolled patient signed an informed consent form approved by the Ethics Committee of Renmin Hospital of Wuhan University. The biopsy specimens were frozen in liquid nitrogen and then stored at −80°C until further experimental analysis.

HBV integration detection

Liver DNA was extracted from biopsy specimens using the QIAamp DNA Mini kit (Qiagen GmbH, Hilden, Germany). A ‘short-read’ WGS for HBV integration was performed by the Beijing Genomics Institute (Shenzhen, China), as previously reported (7). A cluster of multiple read pairs was considered to be a candidate HBV integration breakpoint when it was identified with close mapping positions, linking an end of the human genome and an end of the HBV genome.

PCR and Sanger sequencing validation

In order to confirm the newly identified events and detect the unknown HBV sequences within the chimeric fragments, conventional PCR and Sanger sequencing were used to verify the HBV integration breakpoints with reads N≥2 in the partial HBV integration breakpoints detected by WGS, the strategy of which is shown in Fig. 1. The PCR primers were designed based on WGS-assembled fragments, in which one primer was located in the human genome and the other in the HBV genome. The PCR mix was prepared as follows: 1 µl of DNA; 2 µl of 10X Taq buffer; 11.5 µl of H2O; 2.5 µl of dNTPs; 1 µl of forward and reverse primers (10 µM, respectively) and 1 µl of Taq™ enzyme (Takara Bio, Inc., Otsu, Japan). PCR was subjected to the following cycling conditions: Initial denaturation for 10 min at 95°C; 40 cycles of denaturation for 10 sec at 95°C, annealing for 10 sec and extension at 72°C, final extension for 7 min at 72°C. The forward and reverse primer sequences the denaturation temperature and the duration of extension were determined by preliminary tests. The PCR products were electrophoresed on a 1% agarose gel and were then extracted and sequenced by Sanger sequencing (Shanghai Sangon Biology Engineering Technology & Service Co., Ltd., Shanghai, China). Finally, the results of sequencing were compared with the HBV and human genomes using the Basic local Alignment Search Tool (BLAST; blast.ncbi.nlm.nih.gov/Blast.cgi).

Figure 1.

Polymerase chain reaction and Sanger sequencing validation. HBV1, HBV sequences detected by WGS; HBV2, unknown HBV sequences within the chimeric fragments; Human, host gene sequences detected by WGS; P1 and P2, designed primers. Black highlights the viral sequence and white highlights the human sequence. The open arrows represent the orientation of the primers (5′-3′). HBV, hepatitis B virus; WGS, whole genome sequencing.

Viral-host chimeric transcripts

Total RNA was extracted as previously described (11). Reverse transcription-PCR (RT-PCR) was used to synthesize cDNA according to the manufacturer's protocol (PrimeScript RT™ Reagent kit with gDNA Eraser, Takara Bio, Inc.). The expression of viral-human chimeric transcripts was surveyed by conventional PCR, the conditions of which were designed by the preliminary tests above, and Sanger sequencing. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene was used as a control. The primer sequence for GAPDH was GAPDH-A1, sense 5′-ACCACAGTCCATGCCATCAC-3′ and antisense 5′-TCCACCACCCTGTTGCTGTA-3′. The PCR amplification conditions of GAPDH consisted of initial denaturation at 95°C for 10 min, followed by 95°C for 10 sec, 60°C for 10 sec, and 72°C for 20 sec for 30 cycles. The PCR product was 452 bp.

Intrahepatic (IH) HBV covalently closed circular (cccDNA) quantification

The IH HBV cccDNA levels were measured using quantitative PCR analysis as described previously (12). The cccDNA copy number for the extracted liver samples is calculated by dividing the copies/µl. In order to illuminate the influence of variation in the amount of liver tissue among samples, β-globin, the housekeeping gene (LightCycler Control Kit DNA, Roche Diagnostics) used to calculate the number of cells based on one copy of β-globin per genome, was used to allow for the standardization of the extracted DNA and expression of HBV cccDNA as copies per cell (copies/cell) (12).

Statistical analysis

The statistical analysis was performed using SPSS 13.0 statistical software (SPSS, Inc., Chicago, IL, USA). Continuous variables are expressed as the mean ± standard deviation. Correlations were evaluated using Pearson's correlation test. P<0.05 (two-sided) was considered to indicate a statistically significant difference.

Results

IH cccDNA quantification

With a lower limit of 0.00024 copies/cell, IH cccDNA levels were detected in two patients with AHB, six patients with CHB with primary treatment failure, five patients achieving VR, and one patient achieving both VR and HBsAg seroclearance (38.41±106.18 copies/cell).

Identification of global HBV integrations in HBV-infected patients

The average sequencing depth coverage of WGS for HBV integration was 4,879×. HBV integration breakpoints were identified within all 18 patients with a total of 2,083 and with an average of 138.2±379.9 breakpoints per sample (1–1,596 breakpoints per sample), whereas the higher number was 248.5±57.3 in group 3 and 18.6±13.7 in group 4, respectively (Fig. 2). The number of HBV integration breakpoints was positively associated with the sequencing depth coverage and the IH cccDNA levels, respectively (P<0.0001 and P<0.0001; Fig. 3).

Figure 2.

Number of HBV integration sites in HBV-infected patients. 1, patients with acute HBV; 2, carriers of inactive HBsAg; 3, patients with CHB with primary treatment failure; 4, patients with CHB achieving VR; 5, patients with CHB achieving both VR and HBsAg seroclearance. HBV, hepatitis B virus; HBsAg, hepatitis B surface antigen; CHB, chronic hepatitis B; VR, virological response.

Figure 3.

Correlation analysis between the (A) number of integration sites and the average coverage of HBV integration and the (B) number of integration sites and IH cccDNA levels. HBV, hepatitis B virus; IH cccDNA, intrahepatic covalently closed circular DNA.

Characteristics of HBV integration breakpoints

A total of 14 putative insertions (Table I) with at least two supporting paired-end reads were selected for PCR analysis and successful validation of 100% of these integration sites was achieved (Fig. 4); the unknown HBV sequences (HBV2) were detected within the chimeric fragments of numbers 1, 2, 3, 4, 7, 8, 9, 10, 11, 13 and 14 (Fig. 5). The designed primer sequences of conventional PCR for these breakpoints and their detected characteristics are shown in Table II. There were four types of viral-host junction: Forward simple junction, found in the breakpoints of numbers 2, 3, 8 and 12; reverse simple junction, found in those of numbers 4, 5, 6, 9, 10, 11, 13 and 14; and forward and reverse complicated junctions, found in those of numbers 1 and 7, respectively (Fig. 5). Microhomology was found in several viral-cellular junctions. For example, 3 bp (GCT) of microhomology were found in breakpoint of number 1, 5 bp (AAAAG) of microhomology were found in number 2, 2 bp (GC) of microhomology were found in number 8, and 7 bp (GACCTTC) of microhomology were found in number 11. A detailed analysis of the inserted viral fragments revealed that, with the exception of the breakpoints of number 4 and number 7 mapped in the S gene and that of number six mapped in the Precore gene, the others were localized within the HBx gene. The 3′-end of HBx in several of these breakpoints was found to be deleted.

Table I.

Viral-host junctions of 14 HBV integration breakpoints with at least two supporting paired-end reads.

Number	Code	HBV (nt)	Gene	Host	Location (nt)	Gene
1	92096	1,782	X	chr9	100,940,295	CORO2A|intron
2	92096	1,820	PreC/X	chr9	101,030,026	Intergenic
3	73958	1,471	P/X	chr3	2,002,381	Intergenic
4	24454	689	S/P	chr2	216,281,034	Intergenic
5	32827	1,807	X	chr16	51,320,015	Intergenic
6	32827	1,878	PreC	chr16	51,320,070	Intergenic
7	68892	2,360	P/S	chr10	34,836,578	PARD3|intron
8	68892	1,651	X	chr14	19,828,316	Intergenic
9	68892	1,754	X	chr14	57,417,625	Intergenic
10	68892	1,831	X	chr14	57,417,636	Intergenic
11	68892	1,760	X	chr16	34,990,970	Intergenic
12	94220	1,546	X/P	chr9	98,928,137	Intergenic
13	95658	1,465	X/P	chrX	36,919,408	Intergenic
14	43985	1,513	X/P	chr5	42,009,760	Intergenic

HBV, hepatitis B virus; chr, chromosome.

Figure 4.

Validation of integration sites. Numbers 1–14 represent the 14 hepatitis B virus integration sites. M, DNA marker (DL 100 bp).

Figure 5.

Sequences of viral-host junctions. Numbers 1–14 represent HBV sequences of 14 viral-host junctions. HBV1, HBV sequences detected by WGS. HBV2, HBV sequences detected by conventional PCR. Human, host gene sequences detected by WGS. There were four types of viral-host junction: Forward simple junction (2, 3, 8 and 12); reverse simple junction (4, 5, 6, 9, 10, 11, 13 and 14); forward complicated junction (1); and reverse complicated junction (7). Viral sequences are shown in gray, human sequences are shown in black. The open arrows represent the orientation of the genes (increasing bases). HBV, hepatitis B virus; WGS, whole genome sequencing.

Table II.

Sequences of primers and reaction conditions of 14 HBV integration breakpoints with at least two supporting paired-end reads.

Number	Primer sequence	Denaturation temperature (°C)	Extension (sec)	Amplification fragment length (bp)
1	5′-ACTTCGCTTCACCTCTGC-3′	64	14	321
	5′-TATGATGGCACCACTGCA-3′
2	5′-GAGGCGGTGTCTAGGAGA-3′	54	17	415
	5′-GGTCAAATTGTTTGGATAAA-3′
3	5′-GCAAAACTCATCGGGACT-3′	56	16	360
	5′-TTGTGATGACTTGCTGGA-3′
4	5′-AAGACCTGCACGATTC-3′	50	14	310
	5′-TATGCTCACTTCCACA-3′
5	5′-GTCTTGCCCAAGGTCTTA-3′	56	14	305
	5′-CAGATGGCGCACTAACAA-3′
6	5′-TGAGTAACTCCACAGAAGC-3′	56	10	136
	5′-CAAGAAATAGCCCCAACT-3′
7	5′-CCCGATACAGAGCAGAGG-3′	62	12	270
	5′-GGTCATGGCATGGGAAGA-3′
8	5′-CGCTTCTCCGCCTATTGT-3′	58	18	435
	5′-ACCCTGGCATCCCTGGTTC-3′
9	5′-ATGGCTGCTAGGCTGTGC-3′	58	25	622
	5′-CCATTCCCAACTTGAAGATTTA-3′
10	5′-AGTATAGCTTGCCTGAGT-3′	56	16	354
	5′-CTGCTCTGAGGCAATTAA-3′
11	5′-GCTTGGAGGCTTGAACAG-3′	56	12	253
	5′-AGAGCCCCTTGGAAAATA-3′
12	5′-GAGGTGGCAATGAGGTGAG-3′	56	10	103
	5′-CAGAGGTGAAGCGAAGTG-3′
13	5′-AAAACTCATCGGGACTGA-3′	48	15	319
	5′-CATTTTGTTGACCTGGAA-3′
14	5′-CTGTGCTGCCAACTGGAT-3′	60	12	254
	5′-GCTAAGTGGAGCTTATTTCA-3′

Expression of viral-human chimeric transcripts

Among the 14 viral-host junctions with at least two supporting paired-end reads, chimeric transcripts were observed in breakpoints of numbers 1, 3, 4, 5, 6, 8, 12 and 14 (Fig. 6).

Figure 6.

Expression of viral-human chimeric transcripts. Numbers 1–14 represent the 14 hepatitis B virus integration sites. M, DNA marker (DL 100 bp); GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Discussion

HBV, a notorious DNA virus using cccDNA as its stable transcriptional template for HBV mRNAs, can insert its genome into the human genome and induce multiple hepatocarcinogenesis events. Current antiviral drugs, including INF-α and tenofovir, cannot completely eliminate intrahepatic cccDNA, which has significant involvement in HBV infection relapse and the pathogenesis of HCC and cirrhosis (13). However, HBV integration, the process by which underlying cancer-driving genetic events, including somatic mutation, structural rearrangement and clonal expansion, and its in-depth mechanisms and timing remain to be fully elucidated.

HBV DNA integration events occur with low frequency (0.01–0.1% of infected hepatocytes) in transient infections (14). Compared with traditional PCR-based methods, which do not have the required sensitivity for the detection of virus-cell junctions with such a low frequency of occurrence, WGS may provide insights into HBV integration detection by yielding greater specificity and sensitivity (6–9). In the present study, using ‘short reads’ WGS with an average sequencing depth of 4,879×, it was found that the rates of HBV integration were all 100% in patients with AHB, carriers of inactive HBsAg, and patients with CHB achieving different prognoses following antiviral treatment. The average number of integration breakpoints was 138.2±379.9 per sample. There may be several reasons for detecting a greater number of integration breakpoints per sample in the present study compared with other studies (15). Firstly, a previous study indicated that the ability to identify HBV insertion events depended on the HBV insertion allele frequency and the sequencing depth and coverage (16). This was supported by the positive correlation between the number of HBV integration sites and the sequencing depth coverage in the present study. Secondly, it was not possible to acquire significant differences in the number of integration sites among different groups due to a small number of patients in each group. The positive association between the number of HBV integration sites and the IH cccDNA level, which is associated with the overall survival rate of patients (17), suggested that there exists a close association between the frequency of HBV integration and the prognosis of HBV-infected patients, which requires validation in further investigations with larger sample sizes.

During HBV infection, host chromosomal DNA double stranded breaks (DSBs), provoking cellular responses with the most deleterious effects, may be caused by oxidative DNA damage (18,19). Under this condition and in order to maintain DNA integrity, a mixture of DNA repair mechanisms for host chromosomes may be involved in HBV DNA integration, for example, homologous recombination repair (HRR), classical non-homologous end joining (c-NHEJ), alternative end joining [including microhomology-mediated end joining (MMEJ)] and single strand annealing (20). HRR is an evolutionarily conserved, error-free repair mechanism, using an undamaged sister chromatid as a template to accurately repair the damage. By contrast, c-NHEJ is an error-prone repair pathway, repairing DSBs by joining two non-homologous DNA segments together, which may lead to the potential risks of gene deletion, insertion, indirect or direct repeats, and the phenomenon that HBV DNA molecules integrate into the host chromosomal DNA. Enriched microhomology (MH) exists between human and HBV genome sequences, and MMEJ may be another important mechanism mediating virus integration processes (21).

In the present study, several forms of junction were found between viral and host genes: Forward and reverse simple junctions, and forward and reverse complicated junctions, suggesting that they may be formed by NHEJ, as NHEJ is typically associated with deletions and sometimes insertions of different sequences at the termini (22). Secondly, the finding of MH in viral-host junctions in the present study, for example, 3 bp (GCT) in number 1, 5 bp (AAAAG) in number 2, and 7 bp (GACCTTC) in number 11, indicates that MMEJ may be key in their formation (23). This finding is similar to the result of a previous study, where significant enrichment of MH sizes of 2 and 5 bp was found in the integration junction (24). Thirdly, several HBV integration breakpoints were observed within the HBx genome in the present study, confirming that the HBx gene had more integration opportunities than other regions of HBV, due to the existence of a large number of viral transcriptional regulators in the HBx gene (8).

In the viral-human chimeric transcription analysis, it was found that several transcribed viral-human sequences were located within the X gene, demonstrating that chimeric transcripts were observed only when the site of integration was at 3′-end of HBx and often when its deletion occurred (7). Several independent lines of evidence have demonstrated that HBsAg is not only expressed from the episomal cccDNA minichromosome, but also from transcripts arising from HBV DNA integrated into the host genome (25). In the present study, this was demonstrated by the expression of chimera transcription of number 4, where the breakpoint of the inserted viral fragment was mapped in the S gene. This finding was also in accordance with a previous report, in which serum HBsAg and IH cccDNA levels were not correlated in patients with HBeAg-negative CHB (26), whereas patients with HBeAg-negative CHB usually have a longer, mostly perinatal HBV infection history and are expected to have more extensive HBV DNA integration than HBeAg-positive cases (27).

There were several limitations in the present study. Firstly, the full HBV integration sequences of these sites were not detected, the left end of which may not exactly match the double-stranded linear DNA ends, but may instead include terminal truncations of ~100 bp (14,28). Secondly, in order to further validate the hypothesis that HBV integration sites may have another origin for HBsAg production, chimeric protein expression requires detection from the integrated S gene. Finally, further experiments with larger sample sizes are required to further elucidate the molecular mechanisms and tumorigenesis of HBV integration.

In conclusion, the present study showed that HBV integration was detected in 100% of HBV-infected patients with a high number of integration sites. A close association existed between HBV integration and the prognoses of patients. HBx integration may be indispensable for viral-host chimeric transcripts and the expression of the HBs-host chimeric transcript suggests that HBsAg may be produced from integrated DNA.