Portal de Periódicos da CAPES

Transfer of HBV Genomes Using Low Doses of Adenovirus Vectors Leads to Persistent Infection in Immune Competent Mice

2012; Elsevier BV; Volume: 142; Issue: 7 Linguagem: Inglês

10.1053/j.gastro.2012.03.006

1528-0012

Li–Rung Huang, Yvonne A. Gäbel, Steffi Graf, Silke Arzberger, Christian Kurts, Mathias Heikenwälder, Percy A. Knolle, Ulrike Protzer,

Viral gastroenteritis research and epidemiology

Studies of mechanisms responsible for the persistence of hepatitis B virus (HBV) infection have been hindered by a lack of appropriate animal models. HBV genomes can be delivered to livers of mice using hydrodynamic injection or high doses of an adenoviral vector; these lead to clearance of HBV. We found that infection of immunocompetent mice with low doses of an adenoviral vector resulted in persistent HBV infection; the mice neither underwent seroconversion to production of antibodies against HBV nor developed a strong HBV-specific effector T-cell response. As in patients with chronic HBV infection, DNA vaccination failed to generate T cells that cleared infection. This model of persistent HBV infection could be used to study the pathogenesis of chronic HBV infection and develop new therapeutic strategies. Studies of mechanisms responsible for the persistence of hepatitis B virus (HBV) infection have been hindered by a lack of appropriate animal models. HBV genomes can be delivered to livers of mice using hydrodynamic injection or high doses of an adenoviral vector; these lead to clearance of HBV. We found that infection of immunocompetent mice with low doses of an adenoviral vector resulted in persistent HBV infection; the mice neither underwent seroconversion to production of antibodies against HBV nor developed a strong HBV-specific effector T-cell response. As in patients with chronic HBV infection, DNA vaccination failed to generate T cells that cleared infection. This model of persistent HBV infection could be used to study the pathogenesis of chronic HBV infection and develop new therapeutic strategies. Hepatitis B virus (HBV) infection is restricted to its natural hosts (ie, human beings and chimpanzees), but the host barrier can be overcome by generation of HBV-transgenic mice or delivery of HBV genomes into murine hepatocytes in vivo through either adenoviral gene transfer (AdHBV) or hydrodynamic injection (HDI-HBV).1Sprinzl M.F. et al.J Virol. 2001; 75: 5108-5118Crossref PubMed Scopus (133) Google Scholar, 2Yang P.L. et al.Proc Natl Acad Sci U S A. 2002; 99: 13825-13830Crossref PubMed Scopus (324) Google Scholar Although both HDI-HBV into immunocompromised mice or HDI of mutant HBV lacking core (HBc) as the key target of cytotoxic T lymphocyte (CTL) immunity led to persistence of HBV,2Yang P.L. et al.Proc Natl Acad Sci U S A. 2002; 99: 13825-13830Crossref PubMed Scopus (324) Google Scholar, 3Lin Y.J. et al.Proc Natl Acad Sci U S A. 2010; 107: 9340-9345Crossref PubMed Scopus (58) Google Scholar a model for persistent HBV infection is lacking in immunocompetent mice. Transfer of wild-type HBV genomes by HDI, leading to transfection of approximately 10% of hepatocytes, or high-dose infection with adenoviral vectors targeting more than 90% of hepatocytes (109 infectious units/mouse) resulted in acute self-limiting viral hepatitis.2Yang P.L. et al.Proc Natl Acad Sci U S A. 2002; 99: 13825-13830Crossref PubMed Scopus (324) Google Scholar, 4von Freyend M.J. et al.J Viral Hepat. 2011; 18: 216-226Crossref PubMed Scopus (32) Google Scholar Nevertheless, adenoviral genome transfer has been shown to allow for persistence of HBV genomes in vivo despite an initial vigorous virus-specific immune response.4von Freyend M.J. et al.J Viral Hepat. 2011; 18: 216-226Crossref PubMed Scopus (32) Google Scholar We here examined the course of HBV infection of a nonsusceptible host, the mouse, after low-dose AdHBV. This approach was based on the observation that smaller rather than large virus inocula lead to persistent HBV infection in chimpanzees.5Asabe S. et al.J Virol. 2009; 83: 9652-9662Crossref PubMed Scopus (220) Google Scholar To our surprise, we observed that low-dose AdHBV infection using 108 infectious units per mouse led to persistent infection that was characterized by persistence of HBsAg and hepatitis B e antigen in serum and lack of seroconversion to antibodies to HBsAg (anti-HBs) until day 98 after infection (Figure 1A–D); increasing the dose of infectious AdHBV to 3 × 108, 109, and 3 × 109 dose-dependently induced anti-HBs seroconversion over time (Figure 1D, inset, and Supplementary Figure 1), indicating that less than 3 × 108 AdHBV should be used for induction of persistent infection. As expected,4von Freyend M.J. et al.J Viral Hepat. 2011; 18: 216-226Crossref PubMed Scopus (32) Google Scholar high-dose but not low-dose AdHBV infection led to increased serum alanine aminotransferase (ALT) levels (Supplementary Figure 1). Early after low-dose AdHBV infection (day 4), only a few HBc-expressing hepatocytes were detected but their numbers increased up to day 21 (Figure 1E), and expression persisted up to day 98 after infection. Because there is no evidence for spread of HBV in mice,4von Freyend M.J. et al.J Viral Hepat. 2011; 18: 216-226Crossref PubMed Scopus (32) Google Scholar, 6Yang P.L. et al.Proc Natl Acad Sci U S A. 2010; 107: 798-802Crossref PubMed Scopus (173) Google Scholar the increase in HBc-expressing hepatocytes may result from immunohistochemical detection of accumulating levels of HBcAg. HDI-HBV mice showed strong HBV antigen expression in a higher number of hepatocytes at day 4, reflecting higher levels of antigen expression, which rapidly decreased, becoming undetectable after day 14 (Figure 1A–E). Persistent antigen expression in low-dose AdHBV-infected mice is owing to persistence of the AdHBV inoculum4von Freyend M.J. et al.J Viral Hepat. 2011; 18: 216-226Crossref PubMed Scopus (32) Google Scholar and not to establishment of HBV covalently closed circular DNA, the natural transcription template. It correlated with lymphocyte infiltration in the liver, visualized by H&E staining of liver tissue sections (Figure 1E). Because HBV clearance in HDI-HBV–treated6Yang P.L. et al.Proc Natl Acad Sci U S A. 2010; 107: 798-802Crossref PubMed Scopus (173) Google Scholar as well as in high-dose Ad-HBV–infected mice4von Freyend M.J. et al.J Viral Hepat. 2011; 18: 216-226Crossref PubMed Scopus (32) Google Scholar, 7Isogawa M. et al.Virology. 2005; 333: 293-300Crossref PubMed Scopus (26) Google Scholar depends on potent virus-specific CTL immunity, we assumed that low-dose AdHBV infection does not elicit functional HBV-specific CTL. Indeed, neither HBc- nor HBs-specific CTLs were detected in AdHBV-infected mice until day 21, in contrast to rapid generation of such CTLs in HDI-HBV–treated mice (Figure 2A). Although we observed a small yet significant increase in HBV-specific CTLs at day 21 after low-dose AdHBV infection, it apparently was not able to control viral antigen expression (Figure 1A and B). CTLs induced by introduction of HBV into hepatocytes via HDI were fully functional: upon antigen re-encounter ex vivo, HBc-specific CTLs up-regulated lysosomal marker CD107a (Figure 2B, left) and produced interferon γ (Figure 2B, upper right panel), indicating cytotoxic degranulation and effector function. In addition, killing of HBV-peptide–loaded target cells in vivo showed potent cytotoxic activity in HDI-HBV mice (Figure 2B, lower right). In contrast, HBc- and HBs-specific CTLs generated in low-dose AdHBV-infected mice did not show any effector function ex vivo or in vivo. Absence of an ALT peak at the time of seroconversion despite the presence of significant numbers of functional HBV-specific CTLs in HDI-HBV–treated mice can be explained by our recent observation that CTL-mediated elimination of low numbers of virus-infected hepatocytes escapes detection by measuring serum ALT level.8Stabenow D. et al.Hepatology. 2010; 51: 1430-1437Crossref PubMed Scopus (32) Google Scholar Thus, viral antigenemia in low-dose AdHBV-infected mice resembles the initial stages of natural HBV infection in which significant levels of circulating HBV-antigens but no virus-specific CTL immunity is detected.9Guidotti L.G. et al.Science. 1999; 284: 825-829Crossref PubMed Scopus (1041) Google Scholar To determine whether viral persistence in AdHBV-infected mice also reflects exhaustion of CTL immunity during persistent human hepatic infection, we subjected these mice to therapeutic DNA vaccination with HBc-encoding or control plasmid DNA. Compared with naive mice, we found lower numbers of HBV-specific CTLs after HBc-DNA vaccination in low-dose AdHBV-infected mice (Figure 2C), indicating T-cell tolerance. HBV-specific CTLs generated by HBc-DNA vaccination homed into the livers of low-dose AdHBV-infected animals (Figure 2D), causing mild hepatitis (serum ALT level, <100 U/L). These CTLs, however, up-regulated T-cell immunoglobulin domain and mucin domain 3 (TIM3) and programmed death 1 (PD1) as markers of T-cell dysfunction, and failed to control HBV infection (Figure 2E and F). This recapitulates findings in persistent murine LCMV infection10Jin H.T. et al.Proc Natl Acad Sci U S A. 2010; 107: 14733-14738Crossref PubMed Scopus (554) Google Scholar and in persistently HBV-infected human beings.11Fisicaro P. et al.Gastroenterology. 2010; 138 (e1–4): 682-693Abstract Full Text Full Text PDF PubMed Scopus (353) Google Scholar Taken together, we describe here a model for persistent HBV infection in immunocompetent mice resembling many of the immunologic features observed in patients with chronic HBV infection, such as lack of HBV-specific B-cell/CTL responses and failure to respond to therapeutic DNA vaccination. Low-dose AdHBV infection seems to mimic underwhelming of the CTL response by low-dose HBV12Bocharov G. et al.J Virol. 2004; 78: 2247-2254Crossref PubMed Scopus (86) Google Scholar or the stealth strategy of HBV to enter the liver without causing innate immune defense or hepatocyte apoptosis,5Asabe S. et al.J Virol. 2009; 83: 9652-9662Crossref PubMed Scopus (220) Google Scholar, 13Hosel M. et al.Hepatology. 2009; 50: 1773-1782Crossref PubMed Scopus (288) Google Scholar which may fail to provide sufficient innate immune stimulation to mount CTL immunity. In contrast, HDI elicits danger signals from dying hepatocytes14Racz Z. et al.Nucleic Acid Ther. 2011; 21: 215-224Crossref PubMed Scopus (15) Google Scholar and high-dose AdHBV infection triggers innate immune responses15Hartman Z.C. et al.J Virol. 2007; 81: 1796-1812Crossref PubMed Scopus (126) Google Scholar that likely provide sufficient immunogenic stimuli to elicit potent antiviral immunity. Our model suggests that cross-presentation of circulating HBV antigens in the absence of strong innate immune stimulation fails to promote CTL immunity, which remains to be formally proven. Prolonged circulation of HBV antigens did not lead to depletion of HBV-specific CTLs because we were able to induce HBV-specific CTLs by DNA vaccination, which however, became dysfunctional. Because of its similarities with persistent HBV infection in human beings, this model will serve as an important tool for evaluating novel therapeutic approaches aiming to overcome HBV persistence related to immune-mediated mechanisms. The authors thank Theresa Asen and the Institute of Virology in Bonn for providing help in measuring viral parameters and Ruth Hillermann and Daniel Kull for help with immunohistochemical staining. P.A.K. and U.P. contributed equally to this manuscript. Plasmid, pENTRY HBV1.3, contained an 1.3-fold overlength genome of HBV, subtype ayw, with a 5′ terminal redundancy encompassing enhancers I and II, direct repeats DR1 and DR2, the X- and pregenomic-/core-promoter regions, the transcription initiation site of the pregenomic RNA, the unique polyadenylation site, a duplication of the entire X open reading frame, and a single copy of every other open reading frame as described previously.1von Freyend J.M. et al.J Viral Hepat. 2011; 18: 216-226Crossref PubMed Scopus (36) Google Scholar The pENTRY HBV1.3 was prepared using the EndoFree plasmid Maxi or Mega Kit (Qiagen, Hilden, Germany) and used for HDI and adenoviral vector preparation, respectively. The HBV1.3 genome from pENTRY HBV1.3 was inserted into the E1 region of adenovirus (Ad5ΔE1/E3) backbone plasmid pAd/PL-DEST through Gateway recombination following the manufacturer's instructions (Gateway System; Invitrogen, Karlsruhe, Germany). The recombinant adenoviral genome was transfected into human embryonic kidney 293 cells for production of AdHBV. The AdHBV was purified twice by CsCl gradient centrifugation and further dialyzed and quantified for its titer as described previously.2Dumortier J. et al.Gene Ther. 2005; 12: 668-677Crossref PubMed Scopus (28) Google Scholar C57BL/6j mice were obtained from Janvier (St Berthevin Cedex, France) and maintained under specific pathogen-free conditions according to the guidelines of the Federation of Laboratory Animal Science Associations. All the animal experiments were performed in accordance with German legislation governing animal studies and the Principles of Laboratory Animal Care guidelines (National Institutes of Health publication 85–23, 1996 revision). For HDI of pENTRY HBV1.3 plasmid in mice, 6- to 8-week-old male C57BL/6 mice were anesthetized with ketamine (Medistar Arzneimittelvertrieb, Ascheberg, Germany) and xylazine (Sigma, Munich, Germany). Ten micrograms of HBV plasmid DNA was injected into the tail veins of mice in a volume of phosphate-buffered saline equivalent to 8% of the mouse body weight. The total volume was delivered within 5 seconds. For AdHBV infection, the indicated dose of AdHBV in 200 μL phosphate-buffered saline was delivered into mice intravenously. The serum or blood specimens for assays of hepatitis B surface antigen, hepatitis B e antigen, anti-HBs, and serum ALT were collected at the indicated time points after injections. The livers of mice were fixed in 10% buffered formalin (Sigma) for 24 hours for histologic and immunohistochemical analysis. Serum levels of HBsAg and hepatitis B e antigen and anti-HBs of the infected mice were determined using the Architect System (for antigens) or the AXSYM system (for anti-HBs) (Abbott Laboratories, Abbott Park, IL). Serum alanine aminotransferase activity was measured using specific bioreaction strips on a Reflovet Plus reader (Roche Diagnostics, Mannheim, Germany). Fixed, pretreated, 2-μm thick, paraffin-embedded liver sections were decorated with a monoclonal antibody against HBcAg (NCL-HBcAg-506; clone #LF161, dilution 1:200; Leica Microsystems, Wetzlar, Germany). Pretreatment (H2O2) and staining were performed on a Leica Bond-Max, and detection was performed with a Bond Polymer refine kit (DS9800) using a rabbit anti-mouse secondary antibody, coupled to horseradish peroxidase and 3,3′-diaminobenzidine tetra hydrochloride as a substrate. Livers were harvested from infected mice after perfusion with phosphate-buffered saline via portal vain for 30 seconds at a flow rate of 4 mL/min at indicated time points after infections and minced through a 250-μm cell strainer. Cell suspension from one liver was resuspended in 3 mL of 40% Percoll (GE Healthcare, Munich, Germany) and underlaid with 3 mL of 80% Percoll. After centrifugation for 20 minutes at 1400 × g at room temperature, liver-associated lymphocytes were collected from the interface between 40% and 80% layers and treated with erythrocyte lysis buffer (0.83 % NH4Cl, 0.1% KHCO3). One-tenth of the liver-associated lymphocytes from one liver were stained with allophycocyanin-conjugated dextramers composed of H-2Kb and HBV-derived peptides, HBc93–100 (MGLKFRQL) and HBs190–197 (VWLSVIWM), respectively (Immudex, Copenhagen, Denmark), for 10 minutes, followed by staining with PerCp-Cy5.5–conjugated CD8α (e-Bioscience, San Diego, CA) in the presence of anti-Fc receptor antibody (clone 2.4G2) for another 20 minutes, and subjected to flow cytometric analysis for detection of HBV-specific CD8 T cells. During flow cytometric analysis, a specific number of CountBright absolute counting beads (Invitrogen) were added to all the samples for calculation of the absolute number of Ag-specific CD8 T cells in the liver. Liver-associated lymphocytes (5 × 105) were cultured in RPMI medium 1640 (Invitrogen) with 8% fetal calf serum in the in the presence of 10 μmol/L of peptides, HBc93-100 and HBs190–197 (Pineda, Berlin, Germany), respectively, for 5 hours. For intracellular interferon-γ staining, Brefeldin A and Monensin (0.1% for each; e-Bioscience) was added to the culture 1 hour after the culture had been set up. The cells were harvested for CD8α surface staining as described previously, subjected to intracellular staining of interferon-γ using allophycocyanin-conjugated anti–IFN-γ antibody (clone XMG1.2; e-Bioscience) and a cell fixation/permeabilization kit (BD Biosciences, Heidelberg, Germany), and flow cytometric analysis. For degranulation assay, the cells were restimulated with peptide in the presence of Monensin and phycoerythrin-conjugated anti-CD107a antibody (clone 1D4B; BD Biosciences) for 5 hours and subjected to dextramer and surface marker staining of CD8α as described previously. All the samples were fixed in 4% paraformaldehyde before flow cytometric analysis and dead cells were excluded from the analysis using the LIVE/DEAD Fixable Violet Dead cell stain kit (Invitrogen). Splenocytes from C57BL/6 mice were loaded with 20 μmol/L HBc93-100 and labeled with high-dose carboxyfluorescein succinimidyl ester (CFSE) (2 μmol/L; Invitrogen), with controls labeled with low-dose CFSE (0.2 μmol/L). The other target cells were loaded with 20 μmol/L HBs190-197 and labeled with high-dose CellTrace Far Red (0.5 μmol/L; Invitrogen), with controls with low-dose CellTrace Far Red (0.05 μmol/L). The conditions for peptide pulse and florescence label were at 37°C for 30 minutes and for 15 minutes, respectively. The 4 labeled cell batches were mixed equally and 2 × 107 of the mixed cells were injected intravenously into each mouse previously infected with HBV by HDI or AdHBV at day 19 after infection. Cytotoxicity was determined after 14 hours using splenocytes from the recipients for flow cytometry. The percentage of specific killing was calculated using the following equation: 100* [1 − (dyehigh/dyelow)infected mouse/(dyehigh/dyelow)control mouse]. Naive or AdHBV-infected C57BL/6j mice at day 11 after infection were injected intramuscularly into the quadriceps muscle with 100 μg of pCI/HBcAg or pCI/ovalbumin plasmid dissolved in 50 μL of phosphate-buffered saline, followed by in vivo electroporation to increase the expression level of the injected plasmids.3Widera G. et al.J Immunol. 2000; 164: 4635-4640Crossref PubMed Scopus (449) Google Scholar Serum hepatitis B e antigen and ALT levels in these mice were monitored at 4 days before DNA immunization, and at 1, 7, 14, and 28 after DNA immunization. Liver-associated lymphocytes and splenocytes were isolated from the mice at days 14 and 28 after immunization following the protocol mentioned earlier for the detection of HBc93-100–specific CTLs. The expression of PD-1 and TIM-3 on these CTLs from livers were detected by flow cytometric analysis using fluorescein isothiocyanate–conjugated anti–PD-1 antibody (clone J43; e-Bioscience) and phycoerythrin-conjugated anti–TIM-3 antibody (clone RMT3-23; e-Bioscience). The Student t test was used for analysis. Data are depicted as the mean ± SD, and P values less than .05 were considered significant.