Mechanisms underlying delayed loss of HBeAg and HBV DNA following HBsAg seroclearance in PEG-IFNα treated patients of chronic hepatitis B

ABSTRACT

Background & Aims

A notable proportion of CHB patients undergoing PEG-IFNα based therapy experience lagged serum HBeAg and/or HBV DNA disappearance in patients achieving HBsAg loss. In this study, we explored the molecular mechanisms behind this clinical phenomenon, offering novel insights into the sustainability of chronic HBV infection.

Methods

Two independent clinical cohorts were enrolled to validate this phenomenon. Then comprehensive analysis was performed using public datasets, coupled with a series of molecular biology experiments.

Results

Approximately 17-20% CHB patients underwent PEG-IFNα based therapy experienced seroclearance of HBsAg, while serum HBeAg and/or HBV DNA remained positive. These patients are more prone to serum HBsAg reappearance compared to those achieving complete virological response. Analysis of public datasets revealed that compared to the PC/BCP, the SP1/SP2 promoter displayed more pronounced inhibitory epigenetic modifications in HBeAg-negative patients and SP1/SP2 in-frame mutation peaked in immune active patients. In vitro experiments demonstrated that introduced SP1/SP2 inactive mutations would enhance PC/BCP transcriptional activity by a mechanism known as adjacent transcriptional interference. Furthermore, the deletion of L-HBsAg facilitated intracellular cccDNA replenishment.

Conclusion

This study elucidates that under IFNα treatment and low viral load, transcriptional suppression of SP1/SP2 promoters through mutations and/or epigenetic changes would favour the maintenance of sustain chronic HBV infection, via enhancing the transcription activity of BCP to promote cccDNA replenishment.

Impact and implications

In clinical practice with IFNα antiviral treatment for CHB patients, a “paradoxical” phenomenon is observed where serum HBsAg disappears while HBV DNA or/and HBeAg remains at low positive levels, with delayed disappearance. Our study confirms this clinical phenomenon using two independent clinical cohorts and explores the potential mechanisms behind the persistence of chronic HBV infection under IFNα treatment and low viral load. Transcriptional suppression of SP1/SP2 promoters through mutations and/or epigenetic changes supports the maintenance of chronic HBV infection by enhancing the transcriptional activity of the BCP, which in turn promotes cccDNA replenishment.

Highlights

1. Approximately 20% of patients with CHB who have just achieved HBsAg loss under IFNα treatment show positive serum HBV DNA and/or HBeAg.

2. During disease progression, in frame indel mutations accumulate in the HBV genome’s SP1 and SP2 promoters, with epigenetic modifications contributing to their suppression.

3. In frame indel mutations in the HBV genome’s SP1 and SP2 promoters inhibit the transcription of HBV S mRNA and promote the transcription of 3.5 kb HBV RNA.

4. The loss of L-HBs and envelop proteins leads to an increase in intracellular cccDNA, promoting the maintenance of chronic infection.

GRAPHICAL ABSTRACT

Introduction

Chronic hepatitis B virus (HBV) infection continues to pose a substantial public health challenge despite advancements in vaccination and antiviral therapies [1]. As the disease advances, an estimated 20% to 40% of individuals with chronic hepatitis B (CHB) will progress to end-stage liver diseases, including cirrhosis, decompensated cirrhosis, and hepatocellular carcinoma [2,3]. In China, approximately 86 million individuals are afflicted with CHB [4]. The seroclearance of hepatitis B surface antigen (HBsAg) is considered an optimal treatment goal, indicative of the transcriptional inactivation of covalently closed circular DNA (cccDNA) and the integrated HBV DNA within the infected liver cell [5]. Current frontline antiviral therapies encompass nucleos(t)ide analogues (NAs) and PEGylated interferon-alpha (PEG-IFNα) based therapy [6,7], and seroclearance of HBsAg is rarely achieved with oral NAs which is the cornerstone of current HBV therapy, with a pooled annual rate of only 0.8% in on-treatment patients with CHB [8]. Therefore, various PEG-IFNα based strategies, enhancing both innate and adaptive immune responses and targeting both infected hepatocytes and those harbouring integrated viral DNA, have been explored to enhance the rate of HBsAg loss [9].

The HBV partially double-stranded relaxed circular DNA (rcDNA) in Dane particle has only 3.2 kb in length. It is a very compact genome containing four overlapping open reading frames (ORFs), four promoters and two enhancers and other elements, all of which are important for efficient HBV replication. After infection, the rcDNA is delivered into the nucleus of the cell and repaired to form covalently closed circular DNA (cccDNA), which serves as the transcription template for five viral mRNAs. Previous researches have demonstrated that HBsAg can originate from both cccDNA and integrated HBV DNA, and in HBeAg-negative chronic hepatitis B patients, integrated HBV DNA is the main source of HBsAg [10,11]. Therefore, theoretically, seroclearance of HBsAg means the achieving transcriptional silencing or elimination of cccDNA, followed by the clearance or silencing of hepatocytes with integrated HBV DNA, and typically, before the achievement of seroclearance of HBsAg, the patient’s intrahepatic cccDNA is inactive and serum HBeAg negative. However, a notable proportion of CHB patients undergoing PEG-IFNα based therapy still exhibit low positive levels of HBeAg and/or HBV DNA at the first time of serum HBsAg loss. This phenomenon has been documented in prior studies [12,13], which presents a clinical conundrum and apparent paradox: The seroclearance of HBsAg is associated with the transcriptional silencing of cccDNA, as well as the clearance and silencing of HBV DNA-integrated hepatocytes. However, the persistence of HBeAg and/or HBV DNA positivity suggests ongoing transcriptional activity of the precore (PC) and basal core promoter (BCP) within cccDNA, complicating the interpretation of seroclearance outcomes. In clinical practice, sustained antiviral treatment in such cases typically leads to a clinical cure.

In this study, we clarified the prevalence of this phenomenon in two independent clinical cohorts based on PEG-IFNα based antiviral therapy and explored the impact of HBeAg or/and HBV DNA positivity on HBsAg reappearance. Moreover, considering that the 3.2 kb HBV genome is highly compact, and over half of the nucleotides encode proteins in overlapping reading frames, we attempt to elucidate the potential molecular biological mechanisms underlying this phenomenon by exploring transcriptional interference between adjacent genes in cccDNA, thereby contributing to the persistence of chronic HBV infection. In other words, delayed loss of HBeAg and HBV DNA following serum HBsAg undetectability in PEG-IFNα treated patients, which suggests that the transcriptional activity on cccDNA’s preS1 (SP1) and preS2 (SP2) promoters may be diminished due to mutations or histone epigenetic modifications during antiviral therapy. This change alleviates interference with BCP transcription and facilitated intracellular cccDNA replenishment, maintaining low but detectable levels of serum HBeAg or HBV DNA. The occurrence of HBV SP1/SP2 mutations and epigenetic modifications may be one of the potential mechanisms for the maintenance of chronic HBV infection.

Materials and methods

Clinical cohorts

In our study, we compiled two clinical cohorts of CHB patients who achieved HBsAg seroclearance with PEG-IFNα based antiviral treatments. Treatments included sequential and combination therapies with NAs. The first loss of serum HBsAg marks the starting point of the clinical cohort study in this research. Cohort A included 1658 patients who have achieved serum HBsAg loss from Beijing YouAn Hospital, antiviral treated from 2008 to 2023. Cohort B comprised 58 patients who achieved HBsAg seroclearance from Fudan University’s Huashan Hospital, treated between 2017 and 2021. Notably, the enrolled clinical cohort B is from a recently published study [14]. Due to differing research objectives, we conducted distinct analyses. The study was approved by the ethics committee of Youan Hospital (No. JYK2018-050) and Huashan Hospital (No. KY2015-212) and written informed consent was obtained from all participants.

Downloading and analysing data from public datasets

1. For the comparative analysis of histone modification levels in HBeAg-positive versus HBeAg-negative CHB Patients, the dataset GSE113879 was procured, containing WIG files that delineate histone modification levels at specific HBV DNA loci, standardized to RPM values. We selected untreated samples that were infected with HBV genotype A or D and had complete clinical information for subsequent analysis. A bespoke script was deployed to recalibrate the initial position coordinates to align with the EcoRI site. Subsequently, the WIG files underwent a conversion process to the BigWig format via the UCSC wigToBigWig utility. Replicate experiments were systematically analysed to derive average signal metrics, utilizing the bigwigCompare function of deepTools configured with the “–operation mean” setting. This function was also instrumental in ascertaining the fold enrichment relative to the mean HBV RPM, applying a log2 ratio transformation for comparison. The analytical focus was on histone modification levels proximal to the viral integration site. The deepTools plotProfile utility was engaged to compute histone modification profiles within the vicinity of the integration locus.

2. For the analysis of HBV in-frame mutations, we first downloaded the transcriptome data of chronic hepatitis B patients from the GEO database (GSE230397), and aligned the sequencing data to the HBV genome (ayw) using bwa-mem. Afterwards, we extracted HBV mutation data using scripts we developed. We visualized the mutation data using ggplot2.

Full-length transcriptome sequencing

Huh-7 cells were transfected with the prcccDNA/pCMV-cre recombined plasmid system in three replicates, and these samples were derived from previous research in our lab [15]. In addition, liver tissue samples from three liver transplant patients were also included in the sequencing. The experimental procedure was performed according to the standard protocol provided by Oxford Nanopore Technologies (ONT). After quality control, reads with a length <600 bp and a Qscore <6 were excluded. Then, we adjusted the read orientation and removed the barcodes using a script from https://github.com/guanguiwensy/HBV-nanopore-work-flow. Next, the clean reads were mapped to the reference genome (hg19 with HBV genotype D) using minimap2 (2.24-r1122) [16]. Reads containing HBV sequences were extracted using samtools (1.3.1) [17]. The data were then visualized using R software.

Plasmids construction

The prcccDNA/pCMV-Cre (genotype D and C) recombinant plasmid system was generously provided by Professor Qiang Deng from Fudan University. Mutations in SP1, SP2 and concurrent SP1/SP2 were introduced into the precccDNA/pCMV-Cre recombinant plasmid system by homologous recombination. The SP1 mutation site coincides with the TATA box, a binding site for hepatocyte nuclear factor 1 (HNF1) [18]. The SP2 mutation is situated at the CCAAT box, its essential functional domain [19] (Supplementary Figure 2A). To construct the precccDNA/pCMV-Cre recombinant plasmid, nonsense mutations were introduced at the initial start codons of the preS1, preS2, and S ORFs. This was achieved by substituting thymine with cytosine (ATG to ACG) at the HBsAg start codons at positions 2851nt, 3208nt, and 158nt. The primer sequences used are as follows:

• PreS1-F: 5′-CTTGGGAACAAGATCTACAGCACGGGGCAGAATCTTTCCAC-3′

• PreS1-R: 5′-GTTGTGGAATTCCACTGCGTGGCCTGAGGATGAGTGTTTC-3′

• PreS2-F: 5′-GAAACACTCATCCTCAGGCCACGCAGTGGAATTCCACAAC-3′

• PreS2-R: 5′-TCCTGATGTGATGTTCTCCGTGTTCAGCGCAGGGTCCCCA-3′

• S ORF-F: 5′-TGGGGACCCTGCGCTGAACACGGAGAACATCACATCAGGA-3′

• S ORF-R: 5′-GTGGAAAGATTCTGCCCCGTGCTGTAGATCTTGTTCCCAAG-3′

The pCMV-HBV plasmid was constructed using the pHY106 plasmid, incorporating the HBV genotype D sequence 1816-3182/1-1840nt (NCBI: NC_003977.2) into the pHY106 plasmid. Detailed information about the pHY106 plasmid and the construction strategy can be referenced from the previous literature [20]. In the pCMV-HBV plasmid, the codon TTG for leucine at position 9 of the HBV L protein was mutated to the stop codon TAG to terminate the expression of the L protein, resulting in the construction of the pCMV-HBV-ΔL plasmid. Additionally, the codons TTGTTG for leucine at positions 21 and 22 of the HBV S protein were mutated to stop codons TAGTAG to terminate the expression of the L/M/S surface proteins, thereby constructing the pCMV-HBV-ΔEnv plasmid. The pCMV-HBV-dHAe, pCMV-HBV-dHAe-ΔL, and pCMV-HBV-dHAe-ΔEnv plasmids were constructed based on the cccDNA-dependent fusion HAe expression strategy as described in the previous literature [21]. In these plasmids, the start codon ATG of HBeAg was mutated to TG to abolish potential plasmid-derived HBeAg expression. On this basis, a sequence carrying the HA tag (GTGGACATCTACCCATACGACGTTCCAGATTACGCTGGC) was inserted upstream of the HBV core protein start codon ATG to achieve the fusion expression of the HA tag with HBeAg.

Cell lines and transfection

The HepG2 human hepatoma cell line, procured from the American Type Culture Collection, was propagated in Dulbecco’s Modified Eagle Medium sourced from Corning, California, USA, enriched with 10% fetal bovine serum from Gibco, California, USA, and a combination of 100 IU/mL penicillin and 100 μg/mL streptomycin, also obtained from Gibco. For cellular adhesion and growth, the cells were plated on surfaces coated with rat tail collagen I provided by Corning and incubated for 16 hours. Transfections were carried out using Lipofectamine 2000 reagent from Invitrogen, California, USA, according to the manufacturer’s protocol. Following transfection, both supernatants and cells were harvested at predetermined time points for subsequent assays as detailed later.

Quantification of HBsAg and HBeAg in cell culture supernatants

The concentrations of HBsAg and HBeAg in the cell culture supernatant were quantitatively assessed using a chemiluminescence detection kit (Shenzhen New Industries Biomedical Engineering Co., Ltd.). Measurements were performed strictly according to the manufacturer’s protocol.

Extraction and quantification of HBV RNA and DNA

HBV DNA was extracted from the supernatant using a Viral DNA/RNA Kit (TransGen Biotech) and quantified with the 2×RealStar Power SYBR qPCR Mix (GenStar), employing a standard curve for normalization. The primers for HBV DNA detection were: forward, 5′-TAGGACCCCTTCTCGTGTT-3′, and reverse, 5′-GTGATTGGAGGTTGGGGACT-3′. Total cellular RNA was isolated using TRIzol reagent, while encapsidated HBV RNA was extracted with core lysis buffer, micrococcal nuclease, and TRIzol LS. The RNA was reverse-transcribed into cDNA using the HiScript III 1st Strand cDNA Synthesis Kit (Nanjing Vazyme Biotech Co.). Quantitative polymerase chain reaction (qPCR) was then employed to measure the levels of β-actin and precore messenger RNA (mRNA) (3.5 kb) using the 2× RealStar Power SYBR qPCR Mix (GenStar). The primer sequences used were as follows:

• β-Actin-F: 5′-TTGTTACAGGAAGTCCCTTGCC-3′,

• β-Actin-R: 5′-ATGCTATCACCTCCCCTGTGTG-3′;

• 3.5 kb mRNA-F: 5′-AGACCACCAAATGCCCCTATC-3′,

• 3.5 kb mRNA-R: 5′-TCTGCGAGGCGAGGGAGTTC-3′;

Western blot analysis

Cells were lysed using 1×Lysis Buffer containing a protease inhibitor cocktail, and the lysates were incubated at 4°C for 30 minutes. The supernatants were then collected, supplemented with loading buffer, and heated at 100°C for 10 minutes. Proteins of varying molecular weights were separated via SDS-PAGE and transferred to a PVDF membrane (Invitrogen). The membrane was blocked using 5% nonfat milk in TBST (Tris-buffered saline with 0.1% Tween-20) at room temperature for 1–2 hours and incubated with primary antibodies overnight at 4°C. Secondary antibodies were either fluorescent-conjugated or HRP-linked anti-rabbit/mouse IgG. Protein detection was conducted using an Odyssey Infrared Imager (LI-COR Biosciences, Nebraska, USA) and a Tanon-5200 Chemiluminescent Imaging System (Tanon Science & Technology, Shanghai, China). The primary antibodies were as follows: anti-β-Action (Applygen, Beijing, China), anti-HBc (gifted by Prof. Ningshao Xia of Xiamen University), anti-HBs (Abcam, Cambridge, UK), anti-PreS1(Santa Cruz Biotechnology, CA, USA).

Extraction of core-associated DNA and hirt DNA and detection by Southern blotting

The extraction of core-associated DNA and Hirt DNA and detection by Southern blotting [22]. The core-associated DNA extraction method is as follows: the transfected cells were washed with PBS and lysed in lysis buffer (50 mM Tris-HCl [pH 7.4], 1 mM EDTA, and 1% NP-40) for10 min on ice. The cellular debris and nuclei were removed by centrifugation for 1 min at 10,000 g and the supernatants were digested with 10 mM MgCl₂ and 10 mg/mL DNase I (Sigma-Aldrich) for 30 min at 37°C for eliminating the contaminated transfected plasmid DNA. The reaction was ceased by the addition of 25 mM EDTA, following which 0.5 mg/mL proteinase K (Qiagen) and 1% sodium dodecyl sulphate (SDS) were added, and the mixture was incubated at 55°C for 2 h. The HBV-DNA from the intracellular core particles was purified by phenol–chloroform (1:1) extraction and ethanol precipitation. The DNA was washed with 75% ethanol and dissolved in Tris-EDTA (TE) buffer (10 mM Tris-HCl [pH 8.0] and 1 mM EDTA). The DNA samples were resolved in 1% agarose gels and transferred to positively charged nylon membranes (GE Healthcare). Hirt DNA was extracted and detected by Southern blotting, Briefly, cells inoculated with HBV were lysed in TE buffer (10:10) (10 mm Tris–HCl [pH 7.5] and 10 mm EDTA) supplemented with sodium dodecyl sulphate (SDS) at a final concentration of 0.64% for 30 min at room temperature. Then, the lysate was collected. The proteins and protein-associated DNA were precipitated by adding NaCl to a final concentration of 1 M for at least 16 h at 4°C. The protein precipitate was removed by centrifugation. DNA was extracted by phenol and phenol/chloroform extractions and dissolved in TE buffer (10:1) (10 mm Tris–HCl [pH 7.5] and 1 mm EDTA). The DNA sample was subjected to 1.2% (weight (wt)/volume (vol)) agarose gel electrophoresis overnight at 25 V. The gel was subsequently treated with depurination buffer (0.2 M HCl), denaturing buffer (0.5 M NaOH and 1.5 M NaCl), and neutralization buffer (1.5 M NaCl and 1 M Tris–HCl [pH 7.4]) in turn. The nylon membrane was used for DNA transfer via overnight adsorption. HBV cccDNA was detected by hybridization with a ³²P-labeled HBV DNA probe. Hybridization signals were visualized and analysed by Typhoon FLA 9500 imager (GE Healthcare Life Sciences).

ChIP-qPCR

A SimpleChIP^®Plus Enzymatic Chromatin IP Kit (Cell Signaling Technology, 9005) was used to perform the ChIP experiment. A total of 4 × 10⁶ cells were crosslinked and prepared as an IP sample. After nuclei preparation and chromatin digestion, the lysates were centrifugated at 9400 g for 10 min at 4℃. Supernatant were diluted to 500 μL by 1 × ChIP buffer containing protease inhibitor, and 10 μL of antibody targeting H3K4me3 and H3K9me3 were added followed by incubation, elution, reverse crosslinking, and purification steps. The target DNA fragment was then collected. The modification level of H3K4me3 and H3K9me3 were analysed in BCP region using qPCR, and presented as fold enrichment, which was calculated as 2^{(Ct IgG sample−Ct IP sample)}. Primer sequences are listed as follows:

• BCP-F: 5′-ACGACCGACCTTGAGGCATA-3′

• BCP-R: 5′-TCAACCCCCTCCTCTAATCCA-3′

Animals and immunohistochemistry (IHC) staining

Human hepatocyte chimeric mice (named as Hu-URG) were gifted by Beijing Vitalstar Biotechnology Co.Ltd. The Hu-URG infected with wild-type HBV and BCP + PC double mutants were derived from our previous work in the laboratory [23]. Liver samples were fixed with 4% paraformaldehyde for 24 h and embedded with paraffin. Paraffin sections were used for IHC staining. The primary antibodies utilized were anti-HBcAg (Maixin Biotech, Fuzhou, China), anti-HBsAg (Maixin Biotech), and anti-8OHdG (Santa Cruz Biotechnology, Texas, USA).

Statistical analysis

Continuous variables were expressed as median (interquartile range) or mean (±SD), and compared by ANOVA, Mann–Whitney U/Kruskal–Wallis test as appropriate. Categorical variables were expressed as number (%) and compared by Chi-squared test. Statistical significance was defined as P < 0.05. Analysis was performed on SPSS Version 28.0.1.0 (IBM).

In this study, CHB patients with HBsAg loss were identified as those who showed the first loss of serum HBsAg during IFNα-based antiviral treatment, specifically quantitative HBsAg < 0.05 IU/mL. At this testing point, other virological and biochemical markers were also collected. It should be noted that this study includes HBV DNA test results obtained up to four weeks after HBsAg serumclearance, as HBsAg and HBV DNA tests are not always simultaneous in clinical settings.

Results

Lagged serum HBeAg and/or HBV DNA disappearance in patients achieving HBsAg loss after receiving Peg-IFN-based therapy

We first conducted a retrospective screening of 1658 CHB patients who achieved serum HBsAg clearance after receiving Peg-IFNα based treatment at Youan Hospital in Beijing from 2008 to 2023 (Figure 1(A)). In this study, we investigated the serum HBeAg and HBV DNA levels at the point when serum HBsAg first became undetectable during Peg-IFN-based antiviral therapy. Given that HBsAg and HBV DNA tests were not always simultaneous in clinical practice, HBV DNA test data were collected within four weeks post firstserum HBsAg loss. The serum HBeAg and HBsAg levels in this study were measured simultaneously. Unexpectedly, 20.9% (347/1658) of them remain test positive for serum HBeAg and/or HBV DNA. Of them, 7.2% (120/1658) were serum HBeAg positive alone, 12.4% (206/1658) were serum HBV DNA positive alone, and 1.3% (21/1658) were positive for both of them. Additional clinical data for CHB patients with HBsAg seroclearance yet still testing positive for HBeAg and/or HBV DNA are presented in Table 1. These patients achieved serum HBsAg loss at the age of 36.3 ± 11.0 years with ALT and AST of 39.3 (24.0-69.5) U/L and 34.6 (25.5-55.0) U/L, respectively. As expected, 161 (46.4%) and 195 (56.2%) patients achieved HBsAb and HBeAb positivity, respectively. Further analysis revealed that the HBeAb positivity rate was significantly lower in the HBeAg-positive group compared to those positive for HBV DNA alone and negative for both HBV DNA and HBeAg (P < 0.001). Unexpectedly, the rate of HBsAg seroconversion was higher in serum HBeAg-positive group than in those patients positive for HBV DNA alone or both HBV DNA and HBeAg (P = 0.005).

Figure 1.

Patient flow chart. Selection of participants for the current analysis. Serum HBsAg reappearance defined as the occurrence of HBsAg (≥0.05 IU/mL). (A) Cohort A. (B) Cohort B.

Table 1.

CHB patients with seroclearance yet still testing positive for HBeAg and/or HBV DNA demographics.

Characteristics	Total (n = 1658)	HBeAg and HBV DNA negative (n = 1311)	HBeAg and/or HBV DNA positive (n = 347)	P-value	HBeAg positive (n = 120)	HBV DNA positive (n = 206)	HBeAg and HBV DNA positive (n = 21)	P-value
Age (years)	37.7 ± 11.2	38.1 ± 11.3	36.3 ± 11.0	0.228	32.3 ± 10.8	39.1 ± 10.8	32.1 ± 5.3	<0.001
Male sex	1042 (62.8%)	831 (63.4%)	211 (60.8%)	0.205	68 (56.7%)	134 (65.0%)	9 (40.9%)	0.075
HBsAb positivity	813 (49.0%)	652 (49.7%)	161 (46.4%)	0.148	70 (58.3%)	82 (39.8%)	9 (40.9%)	0.005
HBeAb positivity	1116 (67.3%)	921 (70.3%)	195 (56.2%)	<0.001	16 (13.3%)	176 (85.4%)	3 (13.6%)	<0.001
HBV DNA (Log10 IU/mL)						1.4 (1.2–1.7)	1.3 (1.1–1.4)	0.086
ALT (U/L)	34.0 (21.4–57.0)	31.9 (20.3–53.4)	39.3 (24.0–69.5)	<0.001	30.3 (19.9–51.2)	45.9 (27.8–83.4)	37.7 (23.8–50.4)	<0.001
AST (U/L)	31.6 (23.3–45.8)	30.6 (22.7–43.0)	34.6 (25.5–55.0)	<0.001	30.5 (23.5–41.3)	39.9 (27.2–66.0)	35.0 (25.3–54.2)	<0.001
GGT (U/L)	30.2 (18.1–52.0)	29.6 (18.2–50.0)	32.0 (25.5–55.0)	0.293	24.5 (15.0–40.1)	39.5 (20.3–67.0)	28.5 (16.4–40.5)	<0.001
Albumin (g/L)	45.6 (43.8–47.2)	45.6 (43.8–47.2)	45.4 (43.7–47.1)	0.279	45.3 (43.9–47.1)	45.7 (43.6–47.2)	44.9 (42.8–46.8)	0.646
HBsAg reappearance rate	456 (27.5%)	344 (26.2%)	112 (32.3%)	0.030	43 (35.8%)	59 (28.6%)	10 (47.6%)	0.122

Note: ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, γ-glutamyl transferase.,

Data presented as mean (± standard deviation); median (interquartile range) or number (%) and compared by ANOVA, Mann–Whitney U/Kruskal–Wallis test or Chi-squared test.

To further clarify this clinical phenomenon, we conducted analysis using a recently published prospective clinical cohort composed of 58 CHB patients who achieved serum HBsAg clearance following Peg-IFNα based treatment at Huashan Hospital in Shanghai from 2017 to 2021, with focus on CHB functional cure through Peg-IFNα based antiviral therapy (as illustrated in Figure 1(B)) [14]. The clinical data for enrolled CHB patients are shown in Supplementary Table 1. A retro-prospective analysis reviewed, among these patients, 17.2% (10/58) of patients remained positive for HBeAg at the time point of HBsAg seroconversion. Among them, 3 of 4 patients who received consolidation therapy with Peg-IFNα achieved HBeAg loss, none of them experienced HBsAg reappearance, while the one remained HBeAg positive had HBsAg tested positive, during the follow-up period of 48 weeks post Peg-IFN withdrawn. In contrast, 14.6% (7/48) patients who achieved all virological markers disappearance experienced serum HBsAg reappearance. The binary Logistic regression analysis showed that serum HBeAg positivity, ALT and AST at the time of treatment discontinuation emerged as an independent predictive factor for HBsAg reappearance post-medication cessation (Supplementary Figure1 B).

Epigenetic modifications and mutations reduce or eliminate SP1/SP2 promoter transcriptional activity in cccDNA

The HBV genome is a highly compact 3.2 kb partially double-stranded relaxed circular DNA (rcDNA) in the Dane particle, containing four overlapping ORFs, promoters, enhancers, and other critical elements for efficient replication. After infection, rcDNA enters the nucleus and is repaired into covalently closed circular DNA (cccDNA), which serves as the transcription template for five viral mRNAs. The diagram of HBV cccDNA genomic structure is illustrated in Figure 2(A). Consistent with previous studies [15], results from our long-read RNA sequencing analysis of prcccDNA/Cre cccDNA mimicking system in Huh-7 indicated the transcriptional activity of the SP1/SP2 promoter in cccDNA is significantly stronger than that of the PC/BCP promoter, which was confirmed by sequencing data of liver tissue specimens biopsied for 3 HBeAg patients (Supplementary Figure 1). The observed clinical phenomenon and these laboratory data looked paradoxical, as the products of pgRNA/PreC RNA originating from cccDNA should not remain detectable following the disappearance of HBsAg.

Figure 2.

Distribution of histone modifications on the core promoter and SP1/SP2 promoters’ regions of cccDNA in HBeAg-Positive and HBeAg-Negative CHB Patients. (A) the diagram of HBV cccDNA genomic structure. (B) The deposition of H3K4me3 (blue) and H3K9me3 (pink) on HBV genome were compared between HBeAg-positive patients (n = 2, upper) and HBeAg-negative patients (n = 2, lower). The peak size represented the histone modification levels of HBV DNA, as detailed in materials and methods. The numbers in the left corner represented the range of peak sizes. The bottom showed the position and transcriptional regulatory elements of the HBV genome.

To investigate the potential underlying causes of the aforementioned “anomalous” clinical phenomenon that occurred in a certain proportion (17–20%) of cases, we first consider that some of these individuals may be infected with viral variants producing an antigenically modified HBV S protein undetectable by the available HBsAg assays [24]. However, this rare HBsAg immune escape mutation site, along with the increasingly sensitive commercial HBsAg assay targeting multiple epitopes of HBsAg, cannot account for the nearly 20% incidence rate of the clinical phenomenon [25,26]. We then postulated if epigenetic modifications or mutation mechanisms may lead to the reduced or absent selectively the transcriptional activity of the SP1/SP2 promoters in cccDNA. Utilizing ChIP-seq data from patients with CHB based on the Gene Expression Omnibus (GEO) dataset GSE113879 [27], an increase in inhibitory histone modifications (H3K9me3) and a decrease in activating histone modifications (H3K4me3) at the SP1/SP2 promoter during CHB progression were observed (Figure 2B, Supplementary Figure 3 and 4). This concurrent reduction suggests that while the SP1/SP2 promoters undergo selective transcriptional silencing, the PC/BCP region may retain some transcriptional activity, enabling HBV to continue transcription and replication at a low but detectable level.

SP1/SP2 mutation enhancing PC/BCP transcriptional activity by reducing the adjacent transcriptional interference

In addition to the epigenetic modifications leading to the attenuation or loss of SP1/SP2 transcriptional activity, we are also concerned about the transcriptional silencing caused by SP1/SP2 promoter mutations. Therefore, we conducted an analysis of in-frame mutations within the PreS/S ORF and SP1/SP2 promoter regions throughout the natural history of CHB, by retrieving the expression profile matrix and clinical data for the dataset GSE230397 comprised 81 CHB patients segmented into 27 in the immune tolerance (IT), 15 in the immune active (IA), 23 in the immune control (IC), and 16 in the HBeAg-negative hepatitis (ENEG) from the GEO database (https://www.ncbi.nlm.nih.gov/geo/) [28]. After quality control, we ultimately included 65 samples of this dataset. Considering that the PreS/S ORF and SP1/SP2 promoter are entirely embedded within the ORF of the Pol protein within the highly compact cccDNA, the mutational potential of the PreS/S ORF and SP1/SP2 promoter should be constrained to prevent any possible loss of Pol protein functionality. As depicted in Figure 3(A), in-frame mutations relative to those in the Pol protein are present in the SP1 and preS/S coding and PC/BCP regions throughout the natural history of CHB. Furthermore, we evaluated the variance in the frequency of P-ORF in-frame mutations within SP1 and preS/S coding region across the natural course of CHB. As illustrated in Figure 3(B), the deletion mutation rates in the SP1 and preS/S coding region are significantly higher in the IA and ENEG phases under host antiviral immune pressure, compared to that in the IT and IC phases. These rates are not influenced by the different HBV replication capacities at different stages.

Figure 3.

Database analysis of in-frame mutations in the SP1/SP2 regions during the progression of CHB. (A) Mapping the occurrence of in-frame indel mutations across the natural progression of CHB, the area between the red, blue and orange dashed lines represents the preS1, preS2 promoter and PC/BCP regions. (B) Differences in the frequency of preS1/S2 in-frame mutation rates across various stages of the natural history of CHB. Upper: the level of HBV RNA; Below: the rate of SP1/SP2 in-frame mutation.

Based on the compact structure of cccDNA and reported transcriptional interference between promoters within cccDNA [29], we speculated whether these pol ORF in-frame SP1/SP2 promoter mutations might have alleviated their inhibitory effects on the adjacent PC/BCP promoter transcriptional activity, since we suspect that such interactive interference among various promoter transcriptional activities on cccDNA would enhance HBV DNA replication eventually. Then, point mutations were introduced into the core functional region of the SP promoters to study their effects on PC/BCP transcriptional activity, these mutations are in-frame relative to the Pol protein, ensuring that Pol activity is not affected. Plasmids were engineered harbouring Pol protein in-frame mutations in SP1, SP2, and both in a dual-mutant construct on the prcccDNA plasmid (Supplementary Figure 2A). After the confirming successful plasmid construction, the results indicated that mutation in the SP1 promoter region reduced the expression of L-HBs protein, while SP2 mutations decreased S-HBs protein levels. In the case of dual mutations, the levels of both L and S-HBs proteins were reduced, and there was a significant decrease in HBsAg levels in the supernatant containing SP2 and SP1/SP2 dual mutants. The reduction of L-HBs due to the SP1 mutation resulted in an increase in the efficiency of HBsAg secretion, leading to a significant increase in M- and S-HBsAg levels in the supernatant (Figure 4(A&B)) [30]. These wild-type and mutant plasmids were transiently transfected into HepG2 cells. Comparative analysis revealed that the SP2 and SP1/2 dual-mutation groups significantly upregulated the levels of 3.5 kb HBV RNA and core protein relative to the wild-type prcccDNA transfection group, demonstrating transcriptional interference exerted by the PreS1/S2 promoters on the PC/BCP (Figure 4(D&A)). Additionally, we also observed a significant increase in serum HBeAg levels in the SP2 mutant compared to the wild-type (Figure 4(C)). Notably, the upregulation effect of the SP2 mutation on the 3.5 kb HBV RNA and core protein levels was significantly greater than that induced by the SP1 mutation. Such a more pronounced transcriptional interference effect of the SP2 on the PC/BCP is likely due to the much higher transcriptional activity of SP2 promoter than SP1 [31]. To ascertain whether HBsAg alleviate transcriptional interference on the PC/BCP promoter, next a mutation was introduced at the HBsAg start codon “ATG” in precccDNA/pCMV-cre, rendering HBsAg untranslatable (Supplementary Figure 2B). This modification did not significantly affect the levels of HBeAg and HBV DNA in the supernatant, nor the expression of intracellular core protein (Figure 4(E–H)). To further validate, we have added the HBV genotype-C in our experiments to explore the impact of SP1/SP2 mutations on HBV replication. It was also confirmed that SP1/SP2 mutations can eliminate the impact on the transcriptional activity of the BCP promoter. leading to increased levels of HBV DNA and HBeAg in the supernatant (Supplementary Figure 4).

Figure 4.

Mutations in the SP1/SP2 promoter partially alleviate interference with PC/BCP promoter transcriptional activity. (A) Westernblot of HBsAg and core level in HepG2 cells tranfected with SP1/SP2 promoter mutants. (B) Supernatant HBsAg level in HepG2 cells. (C) Non-denaturing gel analysis of the assembly status of the core protein (D) HBV DNA level (E) HBeAg level (F) 3.5 kb HBV RNA relative levels. (G) Westernblot of HBsAg and core level in HepG2 cells tranfected with the deletion of HBsAg. (H&J) Supernatant HBsAg, HBeAg and HBV DNA levels in HepG2 cells.

Then ChIP-qPCR was performed to examine how SP1/SP2 mutations affect histone methylation (H3K27me3 and H3K9me3) in the BCP region. The results showed that SP1, SP2, and double mutations reduce H3K4me3 and H3K9me3 in the BCP region (Supplementary Figure 3). This reduction in transcriptional activation and repression signals is consistent with our discovery of database analysis (Figure 2(B)), which indicated a decrease in these signals as viral load transitions from high to low. However, the BCP region retains enough transcriptional activity to support chronic infection.

The loss of L-HBs and envelop proteins leads to an increase in intracellular cccDNA, promoting the maintenance of chronic infection

To further clarify the impact of HBsAg absence on HBV replication, we transfected HBV replication-competent plasmids with the deletion of L-HBsAg and envelop proteins (all HBsAg proteins) expression (Figure 5(A&B)). In brief, the 9th leucine codon of the HBV L-HBsAg was mutated from TTG to the stop codon TAG in pCMV-HBV plasmid, terminating L-HBsAg protein expression and creating the pCMV-HBV-ΔL plasmid. Similarly, leucine codons at positions 21 and 22 of the HBV S protein were changed from TTG-TTG to stop codons TAG-TAG, halting the expression of the L, M, and S surface proteins and resulting in the pCMV-HBV-ΔEnv plasmid (Supplementary 3C&D). As expected, the above mutations leaded to an increase in the levels of HBV replication intermediates (rcDNA, dslDNA, ssDNA) and enhances the accumulation of intracellular cccDNA (Figure 5(C)). Interestingly, though the total HBsAg in the supernatant increase in L-HBsAg deletion as expected, the amount of HBeAg did not increase in parallel with the increased cccDNA in these L-HBsAg and envelope deletion.

Figure 5.

Deletion of HBsAg and L-HBsAg enhances intracellular cccDNA accumulation. (A) Huh-7 cells were transfected with pCMV-HBV, pCMV-HBV-ΔEnv, and pCMV-HBV-ΔL plasmids, respectively. Western blot of core and HBsAg. (B) Supernatant HBsAg/HBeAg level. (C) Southern blot of the cellular cccDNA and HBV DNA levels.

Furthermore, we constructed a cccDNA-dependent HA-tagged HBeAg expression system as previously reported [21], which can eliminate the confounding signal from HBcAg, to explore the impact of HBsAg absence on cccDNA accumulation (Figure 6(A)). Also, the mutations involving the expression deletion of HBsAg and L-HBsAg in this system (Figure 6(B&C)). Following transfection into Huh-7 cells, HBV replication-competent plasmids with the deletions of HBsAg or L-HBsAg both enhanced HA-HBeAg levels in the supernatant (Figure 6(D)), suggesting an increase in intracellular cccDNA content.

Figure 6.

Both HBsAg and L-HBsAg Deficiencies Increase the Levels of HA-HBeAg in Supernatants. (A) Schematic representation of cccDNA-dependent HA-tagged HBeAg. Based on this principle, three HBV replicative plasmids were constructed: pCMV-HBV-dHAe, pCMV-HBV-dHAe-ΔEnv, and pCMV-HBV-dHAe-ΔL-HBsAg. (B&D) Culture supernatants were collected on days 1, 3, 5, 7, 9, and 11 post-transfection and ELISA was used to detect HBeAg, HBsAg and HA-HBeAg.

In addition to the accumulation of intracellular cccDNA caused by L-HBs deletion, our previous experiments also found that BCP mutations (A1762T/G1764A) and BCP combined with PC (G1896A) mutations occurred frequently in the natural history of CHB will selectively bring about the transcriptional up-regulation of pgRNA and the down-regulation of the expression of HBsAg decreases [15], a phenomena would be conducive to the maintenance of chronic HBV infection under host immune pressure. Here in this study, we performed HBV cycle system to explore the impact of BCP mutations and BCP/PC combined mutations on HBV replication. The results showed a significant decrease in HBeAg and HBsAg in the cell supernatant due to BCP mutations and BCP combined with PC mutations (Supplementary Figure 4A&B). Western blot analysis revealed a decrease in intracellular HBsAg expression, particularly S-HBsAg, along with an increase in core protein expression (Supplementary Figure 4C). Immunohistochemistry results in humanized mouse liver models also confirmed that the combined BCP and PC mutations lead to increased core protein expression and decreased HBsAg in the liver (Supplementary Figure 4D), as described in our previous work [23].

In summary, the accumulation of BCP mutations and BCP/PC combined mutations during the progression of CHB under host immune pressure, results in selective enhancement of pgRNA transcriptional activity, increased core expression, and decreased HBsAg expression. Additionally, mutations and epigenetic modifications of SP1/SP2 lead to the transcriptional silencing of promoter activity and the loss of L-HBsAg and HBsAg. This, in turn, promotes the intracellular replenishment of cccDNA, facilitating the maintenance of chronic infection under low viral load conditions despite host immune pressure.

Discussion

Serum HBsAg loss is the ideal target for CHB treatment, essential for immune recovery and slowing liver disease progression [32]. IFNα can enhance immune responses and target infected hepatocytes, thus offering a higher cure rate compared to NAs treatment [7]. Since the SP1/SP2 promoters are generally more active compared to the PC/BCP promoters, patients undergoing IFNα treatment usually experience serum HBV DNA and HBeAg negative before HBsAg seroclearance. However, in this study, we observed that 20.9% (347/1658) of HBsAg seroclearance patients undergoing PEG-IFNα based therapy remain positive for serum HBeAg and/or HBV DNA by using a large number of patient samples. By in vitro experiment, we proved that the SP1/SP2 promoter in-frame indel mutations can enhance PC/BCP transcriptional activity and inhibit the 2.8 and 2.4 kb mRNA transcription. This explains the “abnormal” clinical phenomena observed during interferon treatment and suggests that antiviral therapy should aim to achieve negative results for all virological markers to attain a functional cure.

In theory, the seroclearance of HBsAg suggests the elimination or transcriptional silencing of integrated HBV DNA and intrahepatic cccDNA. This study reports that the presence of HBeAg/HBV DNA in patients who have achieved HBsAg seroclearance indicates that the cccDNA PC/BCP promoter retains some transcriptional activity. Regarding this “anomalous” clinical phenomenon, we propose three possible explanations. Firstly, some individuals may be infected with viral variants producing an antigenically modified HBV S protein undetectable by the available HBsAg assays [24]. Secondly, there is the possibility that epigenetic histone modifications or mutation mechanisms may lead to the reduced or absent selective transcriptional activity of the SP1/SP2 promoters in cccDNA. Lastly, mutations in the SP1/SP2 promoters may reduce their activity, leading to the HBsAg seroclearance before the undetectable HBeAg and HBV DNA. In this study, we focused on the third possible reason: we found that SP1/SP2 promoters experience frequent in-frame mutations that alleviate transcriptional interference with the nearby PC/BCP promoter. Building on this foundation, we engineered a precccDNA plasmid with in-frame indel mutations in the SP1/SP2 promoters that relieve transcriptional interference with the PC/BCP promoter, thereby facilitating the transcription and expression of 3.5 kb HBV RNA, the core protein, and the loss of L-HBs and envelope, leading to intracellular accumulation of cccDNA. These results highlight the importance of mutations in the SP1/SP2 promoters in maintaining chronic HBV infection under host immune pressure.

Additionally, utilizing ChIP-seq data from patients with CHB based on the Gene Expression Omnibus (GEO) dataset GSE113879 [27], we observed an increase in inhibitory histone modifications (H3K9me3) and a decrease in activating histone modifications (H3K4me3) at the SP1/SP2 promoter during CHB progression (Figure 2(B)), although the number of samples was relatively small and it was impossible to conduct effective statistical analysis. This concurrent reduction suggests that while the SP1/SP2 promoters undergo selective transcriptional silencing, the PC/BCP region may retain some transcriptional activity, enabling HBV to continue transcription and replication at a low but detectable level. Our experiments also found that when SP1/SP2 promoters undergo in-frame indel mutations, inhibitory histone modifications in the PC/BCP region decrease, while activating histone modifications increase, suggesting a link between SP1/SP2 promoter mutations and promoter histone modifications. In existing literature, approximately 5% of baseline liver samples from HBeAg + participants showed HBV core and HBsAg staining, whereas HBeAg- liver samples demonstrated high HBsAg staining but little to no HBV core staining, and few cells were observed in which both HBV core and HBsAg were detected simultaneously [33]. It is noteworthy that HBsAg-/HBV core + staining is less well defined, as infected hepatocytes containing cccDNA should be able to co-express HBV core and HBsAg. However, the fact that HBsAg-/HBV core + cells were most common in untreated HBeAg + participants is informative. This may imply that in most cccDNA + cells, the transcriptional activity of the SP1/SP2 promoters is diminished or even silenced due to SP1/SP2 mutations and histone epigenetic modifications.

The Hepatitis B Virus (HBV) genome is extremely compact, with substantial overlap between genes. This results in a pronounced phenomenon where the genetic elements of adjacent genes interact strongly. A consequence of the overlapping genomic structure, lead to a situation where the functions and expression patterns of adjacent genes are closely interwoven. Moreover, mutations in the PC and BCP have a notable impact on HBV replication.

Our previous studies have revealed the critical role of PC and BCP mutations in sustaining chronic HBV infection under host immune pressure. A1762T/G1764A BCP mutations increased the transcription of pregenomic RNA (pgRNA) while decreasing that of preC RNA to enhance the replication efficiency of cccDNA [15]. Here in this study, we found a significant decrease in HBeAg and HBsAg in the cell supernatant due to BCP mutations and BCP combined with PC mutations, suggesting that mutations in the SP1/SP2 promoter region may also play a role in this “abnormal” clinical phenomenon. HBV has undergone these changes under selective pressure to maintain its most basic life activities. What does this change mean for the maintenance of HBV infection? Based on our findings, we hypothesize that: (i) the reduced expression of HBsAg, particularly L-HBsAg, impedes the timely translocation of newly assembled nucleocapsids into MVBs to acquire vesicle membranes. In this state, the bare nucleocapsid without a vesicular membrane allows prolonged free entry of dNTPs, enhancing the synthesis of positive-stranded DNA. (ii) According to Ju-Tao Guo [34] and Jianming Hu et al. [35], mature virus particles with an extended positive strand display advantages in phosphorylation modification and capsid disassembly, owing to increased rigidity and accumulation of negative charge. This promotes the selective transport of mature HBV capsids by nuclear pore complexes into the nucleus and the release of rcDNA to sustain the cccDNA pool level (Figure 7).

Figure 7.

Effect of HBsAg expression level on cccDNA maintaining in HBV infected hepatocytes. (Created with Biorender.com).

Due to the very low HBV load in both the liver and serum of patients after treatment, the detection of mutations in the HBV SP1/SP2 region was limited. Although we attempted to include the longitudinal cohort of patients with chronic hepatitis B undergoing interferon treatment, in the end, no fragments with HBV mutations were captured after treatment. In the future, it will be necessary to explore the situation of mutations in the HBV SP1/SP2 region after interferon treatment through sequencing methods with higher depth. Although this study has yielded valuable insights, it has limitations. A primary shortcoming is the absence of direct clinical case evidence. The current research predominantly depends on in-vitro experiments and theoretical analyses. Without a sufficient number of clinical cases, the generalizability and practical application of our findings are inevitably constrained. In the future, the focus will be on the in-depth development of sequencing technology and the acquisition of clinical samples to enrich the viewpoints of this study. Furthermore, an ever-growing body of research has been placing emphasis on the influence of host immune factors on the functional cure of CHB patients. The most important findings were the emergence of a novel atypical adaptive immune landscape in FC patients as defined by CD4-CTLs [36]. it is interesting to note that the loss of viral antigens correlated both with the emergence of CD4-CTLs and memory-like NK cells.

In conclusion, during the progression of CHB under the pressure of host immune defences, this study focuses on how SP1/SP2 mutations and epigenetic modifications reduce the transcriptional interference of SP1/SP2 promoters on the PC/BCP promoters, thereby enhancing the transcription of 3.5 kb RNA and the translation of the core protein. Additionally, the absence of L-HBsAg can lead to an increase in intracellular cccDNA replenishment, maintaining the state of chronic HBV infection. This also explains the “abnormal” clinical phenomena observed during interferon treatment, which supports the recommendation that antiviral therapy should aim to achieve negative results for all virological markers to attain functional cure.

Abbreviations: hepatitis B virus: HBV; chronic hepatitis B: CHB; nucleos(t)ide analogues: NAs; PEGylated interferon-alpha: PEG-IFNα; hepatitis B surface antigen: HBsAg; covalently closed circular DNA: cccDNA; hepatitis B core–related antigen: HBcrAg; pregenomic RNA: pgRNA; deoxynucleoside triphosphates: dNTPs; precure: PC; basal core promoter: BCP; preS1 and preS2 promoters: SP1 and SP2; hepatocyte nuclear factor 1: HNF1; immunohistochemistry: IHC; immune tolerance: IT; immune active: IA; immune control: IC; HBeAg-negative hepatitis: ENEG; Quantitative polymerase chain reaction: qPCR; open reading frames: ORFs; Gene Expression Omnibus: GEO

Funding Statement

This study was supported by the National Key Research and Development Program of China [Grant/Award Number: 2022YFA1303600 and 2023YFC2306900] National Natural Science Foundation of China [82272315 and 62301362].

Disclosure statement

No potential conflict of interest was reported by the author(s).

Author contributions

These authors were involved with this manuscript: Fengmin Lu, Man-fung Yuen, Zhenhuan Cao, Jiming Zhang, Xiangmei Chen, Yuchen Xia (study concept and design and interpretation of data); Fengmin Lu, Man-fung Yuen, Bei Jiang, Guiwen Guan, Zhiqiang Gu, Weilin Gu, Lin Wang, Minghui Li (drafting of the manuscript); Bei Jiang, Guiwen Guan, Kaitao Zhao, Zhiqiang Gu, Weilin Gu (experiment); Bei Jiang and Guiwen Guan (data analysis); Bei Jiang, Minghui Li, Yifei Guo, Zhenhuan Cao (acquisition of data); Fengmin Lu, Xiangmei Chen, Lin Wang, Bei Jiang (obtained funding).

Data availability

The long-read RNA sequencing data and individual participant data reported in this publication will be available and can be reviewed upon request.

Supplemental Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/22221751.2025.2475847.