Article Text

Download PDFPDF

Original article
Prolonged suppression of HBV in mice by a novel antibody that targets a unique epitope on hepatitis B surface antigen
  1. Tian-Ying Zhang1,2,
  2. Quan Yuan1,2,
  3. Jing-Hua Zhao1,2,
  4. Ya-Li Zhang1,2,7,
  5. Lun-Zhi Yuan1,2,
  6. Ying Lan1,2,
  7. Yu-Chieh Lo4,
  8. Cheng-Pu Sun4,
  9. Chang-Ru Wu5,
  10. Jun-Fang Zhang1,2,
  11. Ying Zhang1,2,
  12. Jia-Li Cao1,2,
  13. Xue-Ran Guo1,2,
  14. Xuan Liu1,2,
  15. Xiao-Bing Mo3,6,
  16. Wen-Xin Luo1,2,
  17. Tong Cheng1,2,
  18. Yi-Xin Chen1,2,
  19. Mi-Hua Tao4,
  20. James WK Shih1,2,
  21. Qin-Jian Zhao1,2,
  22. Jun Zhang1,2,
  23. Pei-Jer Chen5,
  24. Y Adam Yuan1,2,3,6,
  25. Ning-Shao Xia1,2
  1. 1State Key Laboratory of Molecular Vaccinology and Molecular Diagnostics, School of Life Science & School of Public Health, Xiamen University, Xiamen, China
  2. 2National Institute of Diagnostics and Vaccine Development in Infectious Diseases, School of Life Science & School of Public Health, Xiamen University, Xiamen, China
  3. 3Department of Biological Sciences and Center for Bioimaging Sciences, National University of Singapore, Singapore, 117543, Singapore
  4. 4Academia Sinica, Institute of Biomedical Sciences, Taipei, Taiwan
  5. 5Department of Internal Medicine, National Taiwan University Hospital, National Taiwan University College of Medicine, Taipei, Taiwan
  6. 6National University of Singapore (Suzhou) Research Institute, Suzhou 215123, China
  7. 7Xiamen Blood Services, Xiamen 361002, China
  1. Correspondence to Quan Yuan and Ning-Shao Xia, State Key Laboratory of Molecular Vaccinology and Molecular Diagnostics, Xiamen University, Xiamen 361102, People's Republic of China; yuanquan{at}xmu.edu.cn, nsxia{at}xmu.edu.cn Y Adam Yuan, Department of Biological Sciences and Center for Bioimaging Sciences, National University of Singapore, Singapore, 117543, Singapore; dbsyya{at}nus.edu.sg

Abstract

Objective This study aimed to investigate the therapeutic potential of monoclonal antibody (mAb) against HBV as a novel treatment approach to chronic hepatitis B (CHB) in mouse models.

Methods Therapeutic effects of mAbs against various epitopes on viral surface protein were evaluated in mice mimicking persistent HBV infection. The immunological mechanisms of mAb-mediated viral clearance were systematically investigated.

Results Among 11 tested mAbs, a novel mAb E6F6 exhibited the most striking therapeutic effects in several HBV-persistent mice. Single-dose administration of E6F6 could profoundly suppress the levels of hepatitis B surface antigen (HBsAg) and HBV DNA for several weeks in HBV-transgenic mice. E6F6 regimen efficiently prevented initial HBV infection, and reduced viral dissemination from infected hepatocytes in human-liver-chimeric mice. E6F6-based immunotherapy facilitated the restoration of anti-HBV T-cell response in hydrodynamic injection (HDI)-based HBV carrier mice. Immunological analyses suggested that the Fcγ receptor-dependent phagocytosis plays a predominant role in E6F6-mediated viral suppression. Molecular analyses suggested that E6F6 recognises an evolutionarily conserved epitope (GPCK(R)TCT) and only forms a smaller antibody–viral particle immune complex with limited interparticle crosslinking when it binds to viral particles. This unique binding characteristic of E6F6 to HBV was possibly associated with its effective in vivo opsonophagocytosis for viral clearance.

Conclusions These results provided new insight into understanding the therapeutic role and mechanism of antibody against persistent viral infection. The E6F6-like mAbs may provide a novel immunotherapeutic agent against human chronic HBV infection.

  • HEPATITIS B
  • ANTIBODY TARGETED THERAPY
  • ANTIVIRAL THERAPY
  • IMMUNOTHERAPY
  • INFECTIOUS DISEASE
View Full Text

Statistics from Altmetric.com

Significance of this study

What is already known on this subject?

  • Achievement of hepatitis B surface antigen (HBsAg) loss is an ideal endpoint of treatment, and is the closest outcome to cure chronic hepatitis B (CHB).

  • Currently available anti-HBV drugs, nucleos(t)ide analogues or pegylated interferon, appear to have very limited effects in inducing HBsAg clearance in patients with CHB. HBsAg loss is rarely achieved with rates of 3%–7% in patients who received regular antiviral regimens. Therefore, novel therapeutic agents with more potent capability for HBsAg clearance are urgently required.

  • Monoclonal antibodies (mAbs) already have a well-established therapeutic role for cancer and autoimmune diseases, and also showed promising therapeutic potency against persistent viral infection in recent HIV studies.

  • Early studies of mAb-based treatments for CHB only demonstrated short-term viraemia suppression effects that were very similar to the effects of treatments based on hepatitis B immune globulin, which is prepared from the plasma of donors who have high counts of HBsAg antibodies. More potent antibodies are certainly required.

What are the new findings?

  • E6F6-like mAbs, which recognise the evolutionarily conserved sA epitope (GPCK(R)TCT) on HBsAg, display more profound and prolonged HBV suppression effects than mAbs binding to other epitopes we tested so far.

  • Single-dose administration of E6F6 can profoundly suppress the levels of HBsAg and HBV DNA for several weeks. E6F6 regimen efficiently prevents initial HBV infection, and blocks viral spreading from infected hepatocytes in human-liver-chimeric mice. E6F6-based immunotherapy can facilitate the restoration of anti-HBV T-cell response in hydrodynamic injection (HDI)-based HBV carrier mice.

  • Fcγ receptor-dependent phagocytosis plays a predominant role in E6F6-mediated HBV viraemia suppression.

  • E6F6 only forms a smaller antibody–viral particle immune complex with limited interparticle crosslinking when it binds to HBV viral particles. This unique binding characteristic of E6F6 to HBV possibly plays an important role in its potent HBV-clearance ability.

How might it impact on clinical practice in the foreseeable future?

  • E6F6 is a superior candidate for developing therapeutic antibodies against HBV. Humanised E6F6 will next be tested in the clinic alone or in combination with currently available anti-HBV drugs to treat patients with CHB, and aims to increase the HBsAg loss rate.

  • If successful, E6F6-based immunotherapy will provide a novel anti-HBV strategy to improve the clinical management of CHB.

Introduction

HBV infection causes chronic hepatitis and places patients at high risk of death due to liver cirrhosis and hepatocellular carcinoma.1 Worldwide, each year, approximately 600 000 people die from the consequences of acute or chronic HBV infection. Despite the successful development of preventive hepatitis B vaccines that have effectively reduced new cases of HBV infection globally,2 there are still more than 240 million people who are persistently infected by HBV, and some of whom need effective anti-HBV therapy to prevent the complications of the disease.3 The approved anti-HBV drugs, which are either interferon or nucleos(t)ide analogues, can only induce disease remission, but not effectively eradicate the virus. The ideal endpoint for anti-HBV therapy is the loss of hepatitis B surface antigen (HBsAg); however, treatment options based on current drugs rarely result in HBsAg clearance.3 Development of more effective new drugs and design of new treatment strategies are, therefore, needed to improve the clinical management of this disease.4

Immunotherapy through antibody infusion has a well-established therapeutic role for cancer, autoimmune diseases and acute viral infections, but not for chronic viral infections.5 Recent HIV treatment studies have demonstrated that some special antibodies exhibit rapid and profound virus-clearance abilities in humanised mice and macaques.6–9 These results suggest that antibody-mediated immunotherapy may provide an alternative approach for treating HIV-like persistent viral infection, although the molecular mechanisms underlying viral infection clearance remain unclear. In vivo administration of virus-specific antibodies may have multiple therapeutic functions. Neutralising antibodies that bind to and inactivate viral envelope proteins block virus entry, and therefore, prevent the spread of infection.5 Moreover, antibodies might have intrinsic effector functions that facilitate direct clearance of circulatory viruses, viral antigens or virus-producing cells via antibody-dependent cell-mediated cytotoxicity, complement lysis or phagocytosis.5 In addition, antibody–virus complexes bind to Fc receptors that are expressed by immune effector cells that can trigger a multitude of innate and adaptive responses against viruses.10 Therefore, we proposed infusion of monoclonal antibodies (mAbs) as a potential therapeutic strategy against chronic HBV infection. In this study, we systematically investigated the therapeutic efficacy of mAbs against various epitopes on viral surface in mice mimicking persistent HBV infection. We identified a novel E6F6 mAb, which binds to a unique epitope on HBsAg and shows potent viral suppression effect in vivo. Moreover, we described the mechanisms of E6F6-mediated viral clearance, and provided a deep understanding of antibody-mediated immunotherapy against persistent viral infection.

Materials and methods

The mAbs

MAbs against the HBV surface protein were produced using hybridoma technology, and were characterised as described previously.11 ,12 Eleven representative mAbs (4D11, 7H11, 2B2, E6F6, E7G11, G12F5, A2C1, 22F10, 42B6, A13A2 and 129G1—detailed description of these mAbs is shown in the online supplementary materials and methods section) were selected for therapeutic evaluation in mice. An mAb (16G12) that is specific to the HIV-1 p24 protein was selected as a control.

Mouse models

The HBV-transgenic (HBV-Tg) mice were provided by Pei-Jer Chen (NTU, Taiwan).13 The human-liver-chimeric mice supporting HBV infection were constructed through transplantation of human hepatocytes into FRG mice as previously described.14 ,15 The hydrodynamic injection (HDI)-based HBV-carrier models were constructed as previously described by using pAAV-HBV1.2 plasmid.16 All mice were maintained under specific pathogen-free conditions in the Laboratory Animal Centre of Xiamen University. The experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals.

More details of the mAb characterisation, animal studies, virological and immunological assays, electron microscopy (EM), and statistical analyses are provided in the online supplementary materials and methods section.

Results

Associations between epitope recognition and therapeutic efficacy of mAbs

In this study, a total of 11 mAbs that target different regions (sA, sB, sE and preS; figure 1A) of the viral envelope proteins were selected for anti-HBV evaluations. Among these mAbs, the 4D11 and 7H11 mAbs recognise aa 21–47 of the pre-S1 region, whereas the 2B2 mAb recognises aa 33–52 of the pre-S2 region. E6F6, E7G11 and G12F5 mAbs bind to a linear epitope (aa 119–125 of HBsAg, designated ‘sA epitope’) on the ‘first loop N-terminus’ located at the major hydrophilic region of HBsAg, whereas A2C1 and 22F10 mAbs specifically recognise the conformation-dependent ‘a’ determinant (around aa 124–147 of HBsAg, designated ‘sB epitope’). 42B6, A13A2 and 129G1 mAbs recognise the ‘second loop’ linear epitope (aa 137–151 of HBsAg, designated ‘sE epitope’). The binding activities to viral antigens and in vitro infection-neutralising capabilities of these mAbs were quantitatively measured, and were presented in figure 1B and C, respectively.

Figure 1

In vitro characterisation of various mAbs against HBV surface proteins. (A) A schematic diagram depicted the binding sites of mAbs that target the HBV surface protein. (B) Binding activities of mAbs to recombinant HBV antigens. The minimal concentrations that yielded detectable binding signals were derived from the binding curve to rank the relative affinity of these mAbs. (C) The infection-neutralising potential of these mAbs was assessed by testing their blocking effects of HBV infection in differentiated HepaRG cells. The data were expressed as the mean±SD. The upper dashed line indicates the value of 100% infectivity, and the lower dashed line indicates the value of 50% infectivity. HBeAg, hepatitis B e antigen; mAb, monoclonal antibody.

The anti-HBV efficacies of various mAbs were evaluated in HBV-Tg C57/B6 mice. The HBV-Tg mice persistently expressed high serum levels of HBV DNA (on average 7.55±0.44 and 7.02±0.44 log10 copies/ml for males and females, respectively) and HBsAg (on average 4.00±0.29 and 3.37±0.28 log10 IU/mL for males and females, respectively), which are as high as the levels that are commonly observed in human beings with chronic HBV infection. For each mAb, four HBV-Tg mice received a single intravenous infusion at a dose of 20 mg/kg. To avoid potential interference of exogenous mAbs in the immunoassays of viral antigens, the HBsAg levels were measured using a denatured-HBsAg quantification assay as described in the online supplementary figure S1. Interestingly, HBV-Tg mice that received different mAbs targeting identical epitopes exhibited a similar viraemia profile (figure 2A and see online supplementary figure S2). Significant decreases in serum HBsAg and HBV DNA were observed in mice that received mAbs that recognised the sA, sB and sE epitopes, whereas little or no viraemia suppression was observed in mice that received mAbs that recognised the preS epitope (figure 2A). Notably, among the tested mAbs, the most precipitous and durable declines of both serum HBsAg and HBV DNA were observed in mice that received three mAbs that recognised the sA epitope. Mice treated with the sA mAbs (E6F6, E7G11 and G12F5) exhibited a more remarkable viraemia decrease in each time interval than the mice in other groups treated by other mAbs (figure 2B).

Figure 2

The therapeutic efficacy of mAbs for HBV-Tg mice. (A) Serum HBsAg and HBV DNA decline profiles of HBV-Tg mice after infusions of mAbs. Data were expressed as medians with IQRs. Four animals were tested for each mAb. (B) Statistics regarding the maximal decreasing effects and therapeutic durability displayed by the mAbs involved in this report were summarised. The mAb-induced maximal decline for each mouse was calculated as the difference in the lowest level after mAb infusion and the baseline level. The durability was calculated as the number of days lasting for an over 90% reduction to baseline level. (C) Serum HBsAg and HBV DNA profiles of HBV-Tg mice after infusions of pAbs. The pAbs of anti-sA and anti-HBsAg were injected into HBV-Tg mice (n=4) at the same dosage (E6F6 equivalent dose: 0.5 mg/kg). Mice that received anti-HBc pAb or were untreated served as controls. The data were expressed as the mean±SD. Dpi, days post-injection; HBsAg, hepatitis B surface antigen; HBV-Tg, HBV-transgenic mice; mAb, monoclonal antibody; pAb, polyclonal antibody.

To further validate our findings, we inserted the sA epitope into an HBc149 protein to construct the chimeric HBc-sA virus-like particle protein as an immunogen to generate anti-sA polyclonal antibodies (pAb) in Balb/C mice (figure 3A). In addition, anti-HBsAg pAb and anti-HBc149 pAb were generated in Balb/C mice as controls. As expected, treatment with purified anti-sA pAb led to more significant and durable decreases in HBsAg and HBV DNA than anti-HBsAg pAb infusion, whereas anti-HBc pAb regimens did not exhibit any suppression effects with regard to serum HBV markers (figure 3A). To compare the in vivo viral suppression efficiency between sA mAbs and the commercial hepatitis B immune globulin (HBIG), we constructed a recombinant human-Fc chimeric sA mAb, which had a variable region of mouse E6F6 (the most potent one of three sA mAbs generated) and human IgG1 Fc region, designated as cE6F6. Dose–response analyses (figure 3B) indicated that the HBsAg binding activity of 1 IU/mL of HBIG was equal to 0.625 μg/mL of cE6F6. As the results of the treatment experiment shown in figure 3C, cE6F6 infusion induced more significant and prolonged suppression either on serum HBsAg or on HBV DNA than HBIG administration at an equivalent dose. In summary, these results demonstrated that the sA epitope on HBsAg might be a superior target for developing therapeutic antibodies against HBV.

Figure 3

Comparison of therapeutic efficacies of anti-sA antibodies and anti- HBsAg polyclonal antibodies in HBV-Tg mice. (A) Serum HBsAg and HBV DNA profiles of HBV-Tg mice after infusions of pAbs. The pAbs of anti-sA and anti-HBsAg were injected into HBV-Tg mice (n=4) at the same dosage (E6F6 equivalent dose: 0.5 mg/kg). Mice that received anti-HBc pAb or were untreated served as controls. The data were expressed as the mean±SD. (B) Dose–response analyses of binding activity of cE6F6 and HBIG. According to the correlation analyses, the binding activity of 1 IU/mL of HBIG was equal to 0.625 μg/mL of cE6F6. (C) Comparison of therapeutic efficacy of cE6F6 and HBIG at an equivalent dose. The data were expressed as the mean±SD. Dpi, days post-injection; HBIG, hepatitis B immune globulin; HBsAg, hepatitis B surface antigen; HBV-Tg, HBV-transgenic mice; pAb, polyclonal antibody; VLP, virus-like particle.

Prolonged viraemia suppression by E6F6 in HBV-Tg mice

Among all mAbs, E6F6 exhibited the strongest suppression on HBV DNA and HBsAg: for the HBV-Tg mice, the levels decreased by more than 2 log10 for a time period of greater than 20 days (figure 4A). E6F6 treatment neither changed hepatitis B e antigen (HBeAg) level significantly nor induced obvious alanine aminotransferase (ALT) elevations (figure 4A) or brought toxicity in other organs (data not shown). Moreover, we compared the efficacy on HBV treatment by E6F6 administration with a daily regimen of oral entecavir. The superiority of E6F6 was represented by its parallel suppression effect on HBsAg and HBV DNA, whereas entecavir did not decrease the HBsAg level (figure 4A). A dose–effect analysis (see online supplementary figure S3) revealed that the maximum suppression effect of E6F6 on HBsAg can be achieved at a dose of greater than 1.2 mg/kg, and the response durability was dose dependent.

Figure 4

The therapeutic effects of E6F6 for HBV-Tg mice. (A) Serum HBsAg, HBV DNA, HBeAg and ALT profiles of HBV-Tg mice after a single E6F6 infusion. 16G12 was an isotype control. Daily oral entecavir was used as a positive control. Antibodies were used at a dosage of 20 mg/kg. The data were expressed as the mean±SD. (B) Profiles of intrahepatic HBV markers of HBV-Tg mice after E6F6 infusion. HBsAg and HBcAg were analysed using western blot, and the HBV DNA was analysed using Southern blot, while the HBV RNA was analysed using northern blot. All assays were performed 3 days after monoclonal antibody (mAb) infusion. (C) Immunohistochemical staining of HBsAg and HBcAg in the livers of HBV-Tg mice after mAb infusion. Assays were performed 3 days after mAb infusion. (D) Quantitative analyses for HBsAg (left panel) and HBcAg (right panel) in the specimens of liver tissue lysates and perfusion purified hepatocyte lysates from HBV-Tg mice that received mAb treatments (3 day after mAb infusion). p Values were calculated using a two-sided unpaired t test, and **indicated p<0.01. ALT, alanine aminotransferase; HBcAg, hepatitis B core antigen; HBeAg, hepatitis B e antigen; HBsAg, hepatitis B surface antigen; HBV-Tg, HBV-transgenic mice; mAb, monoclonal antibody.

We also examined the E6F6-mediated anti-HBV effect on intrahepatic HBV markers on postinfusion day 3. The results indicated that E6F6 treatment did not change intrahepatic levels of viral RNA, DNA and hepatitis B core antigen (HBcAg) (figure 4B). However, both western blot and immunohistochemistry assays indicated that E6F6-treated mice exhibited significantly lower levels of intrahepatic HBsAg than those that received control IgG (16G12) or phosphate-buffered saline (PBS) control (figure 4B, C). In addition, a quantitative analysis by using the liver lysates and perfusion purified hepatocytes also revealed that the intrahepatic HBsAg level of E6F6-treated mice decreased to about 30% of that of 16G12-treated or PBS-treated mice (figure 4D, p<0.001), whereas intrahepatic HBcAg was not significantly changed.

E6F6 suppresses HBV in human-liver-chimeric mice

The experiment was performed in two parts to investigate the anti-HBV effects of E6F6 in human-liver-chimeric FRG (Hu-FRG) mice, which support in vivo HBV infection. The first part of the experiment aimed to evaluate the therapeutic role of E6F6 against established HBV infection. As shown in figure 5A and online supplementary figure S4A, eight Hu-FRG mice with stably established HBV-infection (6–8 weeks after infection by 5×107 copies of HBV) were divided into three groups and infused with E6F6 (n=3), 129G1 (n=3) or 16G12 (n=2). As expected, treatment with E6F6 and 129G1 presented a significant decrease in the levels of HBsAg and HBV DNA, but did not have any obvious influence on the levels of HBeAg and ALT (figure 5B). The viral suppression was averagely maintained, 7 days for 129G1 and 14 days for E6F6 (figure 5B). To investigate whether viral rebound was associated with the development of viral genetic resistance to mAbs, we then performed consecutive mAb infusions once every 10 days from day 58 to day 98 in two mice from the E6F6-treated group and three mice from the 129G1-treated group. Sustained viral suppression was observed in both mice that received E6F6, whereas this effect was observed in only one of three mice that received 129G1 (see online supplementary figure S4B). We also analysed the viral sequence of the virus binding sites of E6F6 and 129G1 in serum samples taken at days 21, 58, 88 and 98. No characteristic escape mutations were detected.

Figure 5

Evaluation of E6F6-mediated viral clearance effects in Hu-FRG mice with established HBV infection. (A) A schematic diagram depicted the experimental procedure of this study. Three animals were used in E6F6 and 129G1 groups, while two were used in 16G12 group. (B) Dynamic change of HBsAg, HBV DNA, HBeAg and ALT in HBV-infected Hu-FRG mice after a single mAb treatment. The data were plotted as mean±SEMs. ALT, alanine aminotransferase; HBeAg, hepatitis B e antigen; HBsAg, hepatitis B surface antigen; Hu-FRG, human-liver-chimeric FRG; mAb, monoclonal antibody.

The second part of the Hu-FRG mouse experiment aimed to investigate the in vivo capacity of E6F6 to prevent initial HBV infection and viral spreading from the initially infected hepatocytes. As depicted in figure 6A, nine Hu-FRG mice were infected with 1×107 copies of HBV at day 0, three of them (group B) were infused a single dose of E6F6 at 1 day before the infection, another three of them received twice-infusion of E6F6 at day 28 and 35 after the infection (group C) and the remaining three mice were untreated as control (group A). While a significant increase in HBsAg and HBeAg levels was determined in the control group, preinfusion of E6F6 (group B) greatly blocked the rise of viral antigen, and treatment with E6F6 starting in the viral ramp-up phase (group C) could also have significantly hindered viraemia increase (figure 6B). To assess the intrahepatic HBV profiling, animals were sacrificed at 42 days postinfection for histological profiling. Human albumin (hAlb) immunostaining revealed that chimerism did not differ significantly among the three groups (figure 6C). The great majority (>80%) of hAlb-positive cells were simultaneously stained HBsAg positive in the control group, very few (about 1%–3%) human hepatocytes stained HBsAg positive in mice that received E6F6 immunoprophylaxis (group B) and only 5%–10% of the human hepatocytes appeared HBV positive in liver sections of mice (group C) that received two doses of E6F6 treatment, respectively (figure 6C).

Figure 6

Evaluation of E6F6-mediated in vivo blockage of de novo infection and virus spreading of HBV in Hu-FRG mice. (A) A schematic diagram depicted the experimental procedure of this study. Three animals were used in each group. (B) Dynamic change of HBsAg and HBeAg in Hu-FRG mice after HBV infection. The data were expressed as the mean±SD. (C) Immunohistochemical staining of human Alb and HBsAg of serial sections of Hu-FRG mouse liver specimens from different groups. Human hepatocytes were visualised using a human-specific Alb staining (left two columns), while HBV-infected hepatocytes were indicated using HBsAg staining (right two columns). The pictures in the second and fourth columns showed a close-up of the region marked by red boxes in the first and third columns, respectively. The sample of non-chimeric (NC) control was presented the liver sample of non-human hepatocyte transplanted FRG mouse. HBeAg, hepatitis B e antigen; HBsAg, hepatitis B surface antigen; Hu-FRG, human-liver-chimeric FRG; mAb, monoclonal antibody.

E6F6 enhance virus-specific CD8+ T-cell responses in HBV-tolerant mice

HDI-based HBV carrier mouse model, which mimics HBV-induced ‘acquired’ tolerance of HBV infection in humans beings,16 ,17 was used to examine whether E6F6-based immunotherapy facilitates the restoration of virus-specific T-cell response. The viral suppression characteristics of E6F6 treatment in HBV-HDI mice were very similar to those observed in HBV-Tg mice: for serological viral markers, E6F6 infusion profoundly suppressed the HBsAg and HBV DNA for several weeks, while inducing little influence on the HBeAg and ALT (figure 7A); for intrahepatic viral makers, a single administration of E6F6 could partly reduce intrahepatic HBsAg whereas it did not reduce intrahepatic hepatitis B core antigen (HBcAg) significantly (figure 7B). To investigate the immune-modulating effects of E6F6, a total of 14 HBV-HDI mice that exhibited high levels of serum HBsAg for >2 months were enrolled for a long-term treatment experiment. After the initial infusion, the HBsAg levels of the E6F6-treated mice rapidly decreased to undetectable levels (figure 7C). However, a serum HBsAg rebound occurred in most mice between days 60 and 80; this rebound was correlated with a decline in serum anti-HBsAg titres to undetectable levels (figure 7C). However, one mouse (E6F6-6) spontaneously recleared serum HBsAg after viral rebound. In contrast, no HBsAg suppression was observed in mice that received 16G12. Subsequently, we conducted three additional antibody infusions for all mice at 119, 209 and 265 days after the initial injection. The results indicated that additional E6F6 infusions always induced effective HBsAg suppression. After four infusion doses (by day 285), an ELISPOT assay indicated that compared with the 16G12 treatment, the HBcAg-specific IFN-γ-secreting T cells in mouse peripheral blood lymphocytes (PBLs) were greatly increased after the E6F6 regimen (figure 7F, p<0.001). Moreover, the CD8+ T cells specific to HBsAg (aa 208–226) or HBcAg (aa 93–100) in mouse PBL were also significantly increased in E6F6-treated mice (figure 7F, p<0.001). These data suggested that E6F6-based immunotherapy suppresses viraemia and facilitates significant restoration of anti-HBV T-cell response.

Figure 7

The therapeutic effects of E6F6 in HBV-HDI mice. (A) The profiles of HBsAg, HBV DNA, HBeAg and ALT in HBV-HDI mice after mAb treatments. Seven animals were used in each group, and three of them were sacrificed and analysed intrahepatically at 3 days after mAb infusion. The data were expressed as the mean±SD. (B) Immunohistochemical staining of HBsAg and HBcAg of the livers of representative mAb-treated HBV-HDI mice (3 days after mAb infusion). (C) The HBsAg (top panel) and anti-HBsAg (bottom panel) in HDI-HBV carrier mice after mAb treatments. A total of 14 new HBV-HDI mice that exhibited high levels of serum HBsAg for >2 months were enrolled for this experiment (n=7 per group). (D) Anti-HBV T-cell response in mice after mAb treatments. The top panel showed the frequencies of the HBcAg-specific IFN-γ cell-secreting cells in the PBLs of HBV-tolerant mice treated using E6F6 and 16G12. The bottom panel showed the frequencies of HBsAg (aa93-aa100)-specific and HBcAg (aa190-aa197)-specific CD8+ T cells in the PBLs of HBV-tolerant mice treated using E6F6 and 16G12. p Values were calculated using a two-sided Fisher's exact test. ALT, alanine aminotransferase; HBcAg, hepatitis B core antigen; HBeAg, hepatitis B e antigen; HBsAg, hepatitis B surface antigen; HDI, hydrodynamic injection; mAb, monoclonal antibody; PBL, peripheral blood lymphocyte.

E6F6 clears HBV via Fcγ receptor-dependent phagocytosis

We investigated the kinetics of the initial viraemia decline after E6F6 infusion for HBV-Tg mice. The E6F6-mediated viral immunoclearance was surprisingly rapid. An obvious decrease occurred as soon as 10 min after infusion, and the maximal suppression plateau was achieved within 24 h after infusion (figure 8A). This rapid response should be associated with an innate immune pathway and not an adaptive immune pathway. The evidence that E6F6 was effective even in immunodeficient FRG mice supported this hypothesis. Additional therapeutic data from different mouse strains, including nude, SCID, Rag2−/− and NOD-SCID, confirmed that the direct antiviral effects of E6F6 were independent of adaptive immune pathways (figure 8B).

Figure 8

The mechanism of E6F6-mediated viral clearance. (A) The kinetics of the initial decline of serum HBsAg and HBV DNA after E6F6 infusion. The data were expressed as the mean±SD. (B) The roles that different immune components played in E6F6-mediated viral clearance. HDIs of pAAV-HBV plasmid were performed on nude, SCID, Rag2−/−, NOD-SCID and C57/B6 mice. The depletions of complement and macrophage were performed using intraperitoneal injections of CVF and λ-Carrageenan in HBV-Tg mice, respectively. The data were expressed as the mean±SD. (C) The antiviral effects of E6F6 variants for HBV-Tg mice. The data were expressed as the mean±SD. (D) Flow cytometric analyses of the mAb-HBV ICs in Kupffer cells, neutrophils, phagocytes and NK cells of mice. The left column showed the populations of Kupffer cells in LMNCs, neutrophils, phagocytes and NK cells in PBLs from mAb-treated mice. The middle and right columns showed the cell populations positive for mAb and HBV input in different lymphocytes from the E6F6-treated and 16G12-treated groups, respectively. (E) Immunofluorescence staining for internalisation of the mAb–HBV immune complex on primary mouse peritoneal macrophages. E6F6 was labelled with dylihght488, whereas the virus was indicated by dylihght594-labelled 42B6-F(ab′)2 against the non-E6F6 epitope. (F) The serum cytokine profile, obtained with Luminex-based MILLIPLEX MAP Mouse Cytokine/Chemokine Magnetic Bead Panel and the Immunology Multiplex Assay, of HBV-HDI carrier mice (n=4) that received E6F6 compared with those that received 16G12 at 6 h after infusion. The data were expressed as medians with IQRs. p Values were calculated using a two-sided Fisher's exact test. *p<0.05. CVF, cobra venom factor; HBsAg, hepatitis B surface antigen; HBV-Tg, HBV-transgenic; HDI, hydrodynamic injection; LMNC, liver mononuclear cell; mAb, monoclonal antibody; PBL, peripheral blood lymphocyte; PBS, phosphate-buffered saline.

Significantly, complement depletion via cobra venom factor did not influence the E6F6-mediated virus suppression, whereas phagocyte depletion via λ-Carrageenan significantly reduced the antiviral effects of E6F6 (figure 8B). Therefore, antibody-mediated phagocytosis was the likely mechanism in E6F6-mediated viral clearance. Because antibody-mediated phagocytosis is Fcγ-receptor dependent, the modification of the Fc region of E6F6 should significantly affect E6F6-mediated HBV clearance. To investigate this possibility, we generated a human Fc chimeric E6F6 (cE6F6) mAb and a cE6F6 variant by introducing two mutations into the Fc region of E6F6 (D265A and N297A, dubbed ‘DANA’)18 and repeated the virus suppression experiments. As expected, and in contrast to wild-type mouse E6F6, the binding affinities of wild-type cE6F6 and DANA-mutated cE6F6 to mouse FcγR I, IIB, III and IV decreased by ∼1 log10 and ∼2 log10, respectively (see online supplementary figure S5). The DANA mutation did not affect the binding of E6F6 to the neonatal Fc receptor (FcRn), which is essential for maintaining the proper antibody half-life in plasma, and it did not affect the binding of E6F6 to HBsAg. Consistent with the binding affinity data, cE6F6 displayed less effective HBV suppression activity, and the introduction of DANA mutations completely disabled the antiviral ability of E6F6 in HBV-Tg mice (figure 8C). Moreover, the Fab and F(ab’)2 fragments derived from E6F6 also exhibited little or no antiviral activity (figure 8C). Hence, Fc-FcγRs interaction was required for the antiviral effects of E6F6. To determine the types of effector cells involved in E6F6-mediated immunoclearance, PBLs and liver mononuclear cells (LMNCs) derived from HBV-Tg mice that received mAb treatment were analysed. Notably, compared with mice treated with 16G12, both intracellular E6F6 and HBV were significantly increased in Kupffer cells (LMNCs), neutrophils (PBLs) and phagocytes (PBLs), but not in the NK cells (PBLs) of mice that received E6F6 (figure 8D). Moreover, an in vitro opsonophagocytosis assay on thioglycollate-eliciting primary mouse peritoneal macrophages also demonstrated that E6F6 could effectively induce HBV internalisation in macrophages (figure 8E). When evaluating the change of mouse serum cytokine profiles after antibody infusion, among the 28 cytokine profiles examined, only the granulocyte colony-stimulating factor (G-CSF) level in the E6F6 group had a transient induction at 6 h after infusion (figure 8F), which returned to baseline within 1 week (see online supplementary figure S6). The elevated G-CSF level was most likely produced by activated phagocytes to stimulate neutrophil maturation. In summary, our data demonstrated that Fcγ receptor-dependent opsonophagocytosis is the predominant mechanism of E6F6-mediated immunoclearance.

Epitope and the binding characteristics of E6F6-HBsAg recognition

Our data demonstrated that Fcγ receptor-dependent phagocytosis primarily mediates E6F6-mediated HBV clearance, which suggests that the binding affinity of the Fc-FcγRs interaction may be important for the antibody-mediated antiviral efficacy. However, an in vitro binding assay did not reveal a positive correlation between the Fc–FcγRs binding activities and antiviral efficacies of the mAb used in this study (see online supplementary figure S7). This result argues against the hypothesis that the Fc–FcγRs binding affinity plays a key role in determining the anti-HBV efficacy of the different antibodies. We, therefore, speculated that the unique epitope and binding properties of the E6F6–HBsAg complex could play important roles in determining the superior antiviral efficacy of E6F6. Peptide mapping revealed that E6F6 binds to the GPCK(R)TCT motif, which is highly conserved among all HBV genotypes and non-human primate HBV strains (figure 9A). Treatments of E6F6 in HBV-HDI mice that carried different HBV strains also demonstrated that this mAb exhibits a universal anti-HBV efficacy against a broad range of HBV genotypes (figure 9B). Additional mutational analyses revealed that the two cysteines in GPCK(R)TCT were crucial for the E6F6 binding (figure 9C), and the minimal binding sequence consisted of CK(R)TC, analogous to the epitope of H166 mAb previously reported.19 However, quantitative analysis revealed that the binding activity of E6F6 to dithiothreitol (DTT)-treated HBsAg protein was maintained at about 70% of that to native HBsAg protein. The minimal loss in binding activity suggested that E6F6 binding is either C121-C124 disulfide independent, or this disulfide bond was re-formed readily from the vicinal cysteine residues on DTT removal under the subsequent assay conditions (figure 9D).

Figure 9

Characterisation of the binding epitope of E6F6. (A) The amino acid variability of the E6F6-binding epitope for various genotypes and subtypes. The E6F6 binding site was marked by red box. (B) Effective reduction in hepatitis B surface antigen (HBsAg) levels of 10 different HBV strains by E6F6 in the hydrodynamic injection mouse model. The data were expressed as the mean±SD. (C) Evaluation of the binding activity of E6F6 to proteins harbouring amino acid mutation in its binding site. Experimental details of this assay were described in the online supplementary materials and methods section. (D) Quantitative analysis of the binding activities between E6F6 and disulfide reduced HBsAg protein (treated with 50 mM DTT) or native HBsAg protein. The binding activity of 22F10 was highly sensitive to disulfide bond reduction, which had been described in a previous study, and was served as a control.

To characterise the features of the viral particle recognition by antibodies for different epitopes, we investigated the morphology of antibody–viral particle immune complexes (ICs) using EM. Interestingly, these data demonstrated that the HBV ICs formed by sA epitope mAbs were smaller and more dispersed than those formed by other mAbs (figure 10A). Consistent with the EM observations, a low-speed centrifugation analysis (figure 10B) indicated that sA epitope mAbs barely induced any viral particle aggregation, even at a concentration of 5.0 log10 ng/mL, whereas other mAbs effectively induced viral particle aggregation (the EC50 ranged from 3.7 to 4.4 log10 ng/mL). Previous studies have demonstrated that the size of antibody-opsonised particles strongly affects their phagocytosis efficacy.20 ,21 Macrophages phagocytose smaller ICs more efficiently than larger ones because the cell membrane takes more time to enclose larger particles than small particles.20 ,22 Consequently, better therapeutic efficacy was exhibited by the sA epitope mAbs than the other mAbs because of the small ICs that these antibodies induced.

Figure 10

Characterisation of the antibody–viral particle immune complexes induced by different anti-HBsAg mAbs. (A) TEM images of the antibody–viral particle ICs formed by mAbs. (B) Dose–response titration curves of mAbs that triggered viral particle aggregation and precipitation. The blue dots and line indicate the amount of HBsAg in the supernatant, and the red dots and line indicate the HBsAg in precipitation. A dashed line indicates greater than 90% precipitation. HBsAg, hepatitis B surface antigen; IC, immune complex; mAb, monoclonal antibody.

Discussion

In this study, we identified a novel mAb E6F6, which showed striking therapeutic effect in mouse models mimicking HBV persistence and infection. We provided evidence to demonstrate three types of anti-HBV effects of E6F6 in animal models, including promoting HBsAg/virion clearance, blocking viral infection and immune-modulation benefit. Among all 11 mAbs and 2 pAbs (mouse anti-HBs pAb and HBIG) tested in our study, the viral clearance effect mediated by E6F6 was the most potent and prolonged. A single infusion of E6F6 profoundly suppressed the HBsAg and HBsAg-enveloped virus (E6F6 binding targets) in HBV carrier animals for several weeks. These viraemia suppression profiles resembled those described in recent studies of antibody-mediated anti-SHIV experiments in macaques.7 ,8 We further provided in vitro and in vivo evidence that Fcγ-receptor-mediated opsonophagocytosis (figure 8B, C)—neither antibody-dependent neutralisation nor complement deposition or cellular cytotoxicity—is crucial for E6F6-mediated viral clearance. Similar mechanisms have been suggested to play an important role in antibody-mediated infection protection against the influenza virus and West Nile Virus.23–25 Taken together, we propose that Fcγ-receptor-dependent phagocytosis is a general pathway in antibody-mediated viral clearance. By using Hu-FRG mice based in vivo HBV infection model, we showed that E6F6 regimens could efficiently prevent initial viral infection and also block HBV dissemination postinfection when treatment started in the early ramp-up infection phase (figure 6). The viral infection blocking function of E6F6 was possibly attributed to both its potent viral clearance capacity and inhibiting activity for viral entry. Moreover, we demonstrated that E6F6-based immunotherapy facilitated the restoration of anti-HBV T-cell response in acquired HBV-tolerant mice (figure 7C). This result is consistent with the observation that antibody-mediated HIV treatment enhanced the HIV-specific immunoresponse of macaques,7 and suggests the presence of immune-modulation benefits due to antibody treatment. In contrast to the approved anti-HBV agents, the most important advantage of E6F6 is its ability to greatly reduce the HBsAg level, which current drugs fail to do. Because a high level of HBsAg can exhaust HBsAg-specific T-cell response, and is proposed as an important factor for viral immunotolerance in patients with chronic hepatitis B (CHB),26–29 the reduction of HBsAg could allow the immune system to tame the viral infection. Although the exact mechanism by which the antibody stimulates virus-specific immunoresponses requires additional investigation, this stimulation might be derived from promoting antigen presentation via opsonophagocytosis.

Notably, previous studies regarding the antibody treatment of chronic HBV-infected human beings and animals suggested that antibodies targeted at the ‘a’ determinant (sB epitope), the ‘second-loop’ (sE epitope) of HBsAg or both exhibited only short-term viraemia suppression effects.30 ,31 Our results are the first to show that the virus-clearance efficacies exhibited by antibodies could be associated with their binding epitopes, rather than positively correlated with the mAb's binding activity or viral neutralising capability (based on the results shown in figures 1B, C and 2B). Moreover, E6F6-like antibodies, which recognise the evolutionarily conserved sA epitope, had more profound and more durable viral suppression effects than those that bind to other epitopes. The sA epitope carried residues between 119 and 125 of the HBsAg and contained a CXXC motif, which had been demonstrated to be the most important sequence required for the infectivity of HBV and HDV.32 Unfortunately, we are not yet able to provide mechanistic evidence to explain how and why the binding epitope difference impacts the viral clearance potency. However, our preliminary analyses by EM and centrifugation indicated that E6F6 only forms a smaller antibody–viral particle IC with limited interparticle crosslinking when it binds to viral particles. This unique binding characteristic of E6F6 to HBV was possibly associated with its effective in vivo opsonophagocytosis for viral clearance.

Taken together, the E6F6 antibody we generated is a superior candidate for further development towards therapeutic antibody for CHB treatment. Our findings provide new insight into understanding the therapeutic role and mechanism of antibody against persistent viral infection. Considering the distinct anti-HBV effects, antibody-mediated immunotherapy alone or in combination with interferon, nucleos(t)ide analogues, the recently reported LTβR agonists targeting HBV covalently closed circular DNA degradation33 and/or other novel anti-HBV agents may provide revolutionary anti-HBV strategies to increase the CHB cure rate.

Acknowledgments

We thank Dr Ding Xue at the University of Colorado, Dr Tong-Ming Fu at Merck Research Laboratories and Dr Jiahuai Han at the Xiamen University for their careful review and constructive comments on our manuscript.

References

View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • T-YZ and QY contributed equally.

  • Contributors T-YZ, QY, YAY and N-SX had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: T-YZ, QY, YAY and N-SX. Acquisition of data: T-YZ, QY, J-HZ, Y-LZ, L-ZY, YL, Y-CL, C-RW, J-FZ, YZ, J-LC, X-RG, XL and X-BM. Analysis and interpretation of data: T-YZ, QY, M-HT, JZ, P-JC, JWKS and N-SX. Drafting of the manuscript: T-YZ, QY and YAY. Critical revision of the manuscript for important intellectual content: M-HT, Q-JZ, JZ, P-JC, JWKS and N-SX. Statistical analysis: T-YZ and QY. Technical or material support: C-PS, W-XL, TC and Y-XC. Obtained funding: QY, W-XL, YAY and N-SX. Study supervision: QY, YAY, JZ and N-SX. Approval of the final version of the manuscript: QY, YAY, P-JC, JZ and N-SX.

  • Funding The National Scientific and Technological Major Project (2013ZX10002002-001 and 2012ZX10004503-005), the National Science Fund (81371819), the ‘863’ project (2012AA02A307) and the Excellent Youth Foundation of Fujian Scientific Committee (2015J06018) supported this work.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles