Department of Medicine and Molecular Science, Division of Frontier Medical Science, Programs for Biomedical Research, Graduate School of Biomedical Sciences, Hiroshima University, Hiroshima, Japan
Liver Research Project Center, Hiroshima University, Hiroshima, Japan
Laboratory for Digestive Diseases, Center for Genomic Medicine, RIKEN, Hiroshima, Japan
M.D., Ph.D., Department of Medical and Molecular Science, Division of Frontier Medical Science, Programs for Biomedical Research, Graduate School of Biomedical Science, Hiroshima University, 1-2-3 Kasumi, Minami-ku, Hiroshima 734-8551, Japan
Potential conflict of interest: The authors have no conflicts to disclose.
This study was supported, in part, by a Grant-in-Aid for Scientific Research from the Japanese Ministry of Labor, Health, and Welfare.
The necroinflammatory reaction plays a central role in hepatitis B virus (HBV) elimination. Cluster of differentiation (CD)8-positive cytotoxic T lymphocytes (CTLs) are thought to be a main player in the elimination of infected cells, and a recent report suggests that natural killer (NK) cells also play an important role. Here, we demonstrate the elimination of HBV-infected hepatocytes by NK cells and dendritic cells (DCs) using urokinase-type plasminogen activator/severe combined immunodeficiency mice, in which the livers were highly repopulated with human hepatocytes. After establishing HBV infection, we injected human peripheral blood mononuclear cells (PBMCs) into the mice and analyzed liver pathology and infiltrating human immune cells with flow cytometry. Severe hepatocyte degeneration was observed only in HBV-infected mice transplanted with human PBMCs. We provide the first direct evidence that massive liver cell death can be caused by Fas/Fas ligand (FasL) interaction provided by NK cells activated by DCs. Treatment of mice with anti-Fas antibody completely prevented severe hepatocyte degeneration. Furthermore, severe hepatocyte death can be prevented by depletion of DCs, whereas depletion of CD8-positive CTLs did not disturb the development of massive liver cell apoptosis. Conclusion: Our findings provide the first direct evidence that DC-activated NK cells induce massive HBV-infected hepatocyte degeneration through the Fas/FasL system and may indicate new therapeutic implications for acute severe/fulminant hepatitis B. (HEPATOLOGY 2012)
Between 4% and 32% of fulminant hepatitis cases, characterized by acute massive hepatocyte degeneration and subsequent development of hepatic encephalopathy and liver failure, are caused by acute hepatitis B virus (HBV) infection.1 Host2 and viral factors3 may influence the development of fulminant hepatitis, but these factors have not been fully elucidated.
Innate and adaptive immunity both play a role in the elimination of viral infections. In the innate immune response, cytoplasmic and membrane-bound receptors recognize viruses and induce interferon (IFN)-β production, which, in turn, up-regulates IFN-α and induces an antiviral state in surrounding cells.4 In the adaptive immune response, viruses are recognized by dendritic cells (DCs), which activate cluster of differentiation (CD)8-positive T cells to reduce viral replication through cytolytic5 and noncytolytic mechanisms.6 The role of immune cells, especially HBV-specific cytotoxic T lymphocytes (CTLs), is crucial in the development of fulminant hepatitis.7, 8 CTLs can kill target cells using two distinct lytic pathways: the degranulation pathway, in which perforin is used to puncture the membranes of infected cells, and the Fas-based pathway, in which the interaction between Fas ligand (FasL) expressed on cytolytic lymphocytes and Fas on target cells triggers apoptosis and target cell death.9 However, the role of innate immune cells, especially natural killer (NK) cells, in fulminant hepatitis remains obscure. NK cells have recently been reported to contribute to the pathogenesis of human hepatitis and animal models of liver injury.10, 11 Replication of HBV is host cell dependent, and the study of cellular immune response in hepatitis B has long been hampered by the lack of a small animal model that supports the replication of HBV and elimination of infected cells by immune response. Before the advent of human hepatocyte chimeric mice,12, 13 only chimpanzees had been used as a model for HBV infection and inflammation, although fulminant hepatitis B (FHB) had never been reported, and severe liver inflammation is rare in chimpanzees.14 We previously established an HBV-infection animal model using chimeric mice, in which the livers were extensively repopulated with human hepatocytes.15-17 In this study, we attempted to establish an animal model of HBV-infected human hepatocytes with human immunity by transplanting human peripheral mononuclear cells (PBMCs) to HBV-infected human hepatocyte chimeric mice.
APC, allophycocyanin; asialo GM1, ganglio-N-tetraosylceramide; CD, cluster of differentiation; CHB, chronic hepatitis B; CTLs, cytotoxic T lymphocytes; DC, dendritic cell; FasL, Fas ligand; FHB, fulminant hepatitis B; HBcAg, hepatitis B core antigen; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; HLA, human leukocyte antigen; HSA, human serum albumin; IFN, interferon; IP, intraperitoneally; ISG, interferon-stimulated gene; mAb, monoclonal antibody; mDC, myeloid DC; mRNA, messenger RNA; NK, natural killer; PBMCs, peripheral blood mononuclear cells; PCR, polymerase chain reaction; pDC, plasmacytoid DC; SCID, severe combined immunodeficiency; TUNEL, terminal deoxynucleotidyl transferase dUTP nick end labeling; uPA, urokinase-type plasminogen activator.
Materials and Methods
Generation of Human Hepatocyte Chimeric Mice.
Generation of the urokinase-type plasminogen activator (uPA)+/+/severe combined immunodeficiency (SCID)+/+ mice and transplantation of human hepatocytes with human leukocyte antigen (HLA)-A0201 were performed as described previously.15, 16 All mice were transplanted with frozen human hepatocytes obtained from the same donor. Infection, extraction of serum samples, and euthanasia were performed under ether anesthesia. Concentration of human albumin, which is correlated with the repopulation index,15 was measured in mice as described previously.16 All animal protocols described in this study were performed in accord with the Guide for the Care and Use of Laboratory Animals and the local committee for animal experiments, and the experimental protocol was approved by the Ethics Review Committee for Animal Experimentation of the Graduate School of Biomedical Sciences at Hiroshima University (Hiroshima, Japan).
Human Serum Samples.
Human serum samples, containing high titers of genotype C HBV DNA (5.3 × 106 copies/mL), were obtained from patients with chronic hepatitis who provided written informed consent. Individual serum samples were divided into aliquots and stored in liquid nitrogen. Six weeks after hepatocyte transplantation, chimeric mice were injected intravenously with 50 μL of HBV-positive human serum.
Analysis of HBV.
DNA was extracted using SMITEST (Genome Science Laboratories, Tokyo, Japan) and dissolved in 20 μL of H2O. HBV DNA was measured by real-time polymerase chain reaction (PCR) using a light cycler (Roche, Mannheim, Germany). Primers used for amplification were 5′-TTTGGGCATGGACATTGAC-3′ and 5′-GGTGAACAATGTTCCGGAGAC-3′. Amplification conditions included initial denaturation at 95°C for 10 minutes, followed by 45 cycles of denaturation at 95°C for 15 seconds, annealing at 58°C for 5 seconds, and extension at 72°C for 6 seconds. The lower detection limit of this assay was 300 copies.
Preparation of Human Blood Mononuclear Cells and Transplantation of Human PBMCs Into Human Hepatocyte Chimeric Mice.
PBMCs were isolated from healthy blood donors with HLA-A0201 and successfully vaccinated with recombinant yeast-derived hepatitis B surface antigen (HBsAg) vaccine (Bimmugen; Chemo-Sero Therapeutic Institute, Kumamoto, Japan) using Ficoll-Hypaque density gradient centrifugation. Neither monocytes nor macrophages were observed in the isolated PBMCs (Supporting Fig. 1). PBMCs isolated from 3 healthy, unvaccinated blood donors were also transplanted. Eight weeks after HBV inoculation, human PBMCs were transplanted into human hepatocyte chimeric mice. To deplete mouse NK cells and prevent the elimination of human PBMCs from human hepatocyte chimeric mice, 200 μL of phosphate-buffered saline, containing 120 μL of anti–ganglio-N-tetraosylceramide (asialo GM1) antibody (Wako, Osaka, Japan), were administered intraperitoneally (IP) 1 day before (day 0; Fig. 1) the initial IP transplantation (day 1) of human PBMC. Then, 10 μL/g of liposome-encapsulated clodronate (Sigma-Aldrich, St. Louis, MO) were also administered 4 days before PBMC transplantation (day −2) to deplete mouse macrophages and DC cells. The second PBMC administration (4 × 107 cells/mouse) was performed 2 days after the initial administration (day 3).
To assess the effect of the depletion of human DC, NK, or CD8-positive CTL cells from administered PBMCs on hepatitis formation, the BD IMag separation system (BD Biosciences, Franklin Lakes, NJ) was used. Alternatively, mice were treated with an IP administration of clodronate, as described above, 1 day before PBMC transplantation.
To analyze the effect of inhibition of the Fas/FasL system, IFN-γ, IFN-α, antihuman FasL monoclonal antibody (mAb) (1.5 mg/mouse; R&D Systems, Minneapolis, MN), antihuman IFN-γ mAb (1.5 mg/mouse; R&D Systems), and antihuman IFN-α mAb (1.5 mg/mouse; PBL Biomedical Laboratories, Piscataway, NJ) were injected 1 day before transplantation of human PBMCs.
Reconstructed human PBMC proliferation in mice was determined by flow cytometry with the following mAbs used for PBMC surface staining: allophycocyanin (APC)-H7 antihuman CD3 (clone SK7); APC-conjugated anti-CD4 (clone SK); BD Horizon V450 antihuman CD8 (clone RPA-T8); APC-conjugated antihuman CD11c (clone B-ly6); HU HRZN V500 MAB-conjugated antihuman CD45 (clone H130); Alexa Fluor 488–conjugated antihuman CD56 (clone B159); PerCP-Cy5.5 antihuman CD123 (clone 7G3); fluorescein isothiocyanate–conjugated Lineage cocktail 1 (Lin-1) (anti-CD3, CD14, CD16, CD19, CD20, and CD56); APC-H7 antihuman HLA-DR (clone L243); phycoerythrin (PE)-conjugated antihuman FasL (clone NOK-1); and biotin-conjugated antimouse H-2Db (clone KH95). The biotinylated mAbs were visualized using PE-Cy7-streptavidin. Each of the above mAbs were purchased from BD Biosciences. PE-conjugated HBV core-derived immunodominant CTL epitope (HBcAg93)18 (Medical & Biological Laboratories Co., Ltd., Nagoya, Japan). Dead cells identified by light scatter and propidium iodide staining were excluded from the analysis. Flow cytometry was performed using a FACSAria II flow cytometer (BD Biosciences), and results were analyzed with FlowJo software (Tree Star, Inc., Ashland, OR).
DCs can be classified into two main subsets: plasmacytoid DCs (pDCs) and myeloid DCs (mDCs).19, 20 pDCs were defined as CD45+Lin-1−HLA-DR+CD123+ cells, whereas mDCs were defined as CD45+ Lin-1−HLA-DR+CD11c+ cells.
Histochemical Analysis of Mouse Liver and Terminal Deoxynucleotidyl Transferase dUTP Nick End Labeling Assay.
Histochemical analysis and immunohistochemical staining using an antibody against human serum albumin (HSA; Bethyl Laboratories, Inc., Montgomery, TX), an antibody against hepatitis B core antigen (HBcAg) (Dako Diagnostika, Hamburg, Germany) and antibody against Fas (BD Biosciences, Tokyo, Japan) were performed as described previously.16 Immunoreactive materials were visualized using a streptavidin-biotin staining kit (Histofine SAB-PO kit; Nichirei, Tokyo, Japan) and diainobenzidine. For the terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay in sliced tissues, we used an in situ cell death detection kit (POD; Roche Diagnostics Japan, Tokyo, Japan).
Dissection of Mouse Livers and Isolation of RNA and Measurement of Messenger RNAs of Fas by Reverse-Transcription PCR.
Mice were sacrificed by anesthesia with diethyl ether, and livers were excised, dissected into small sections, and then snap-frozen in liquid nitrogen. Total RNA was extracted from cell lines using the RNeasy Mini Kit (Qiagen, Valencia, CA). One microgram of each RNA sample was reverse transcribed with ReverseTra Ace (Toyobo Co., Tokyo, Japan) and Random Primer (Takara Bio Inc., Kyoto, Japan). We analyzed the messenger RNA (mRNA) levels of Fas by reverse-transcription PCR, as previously reported, using Fas forward primer 5′- GGGCATCTGGACCCTCCTA-3′ and Fas reverse primer 5′- GGCATTAACACTTTTGGACGATAA-3′.
mRNA expression levels of Fas and interferon-stimulated genes (ISGs) were compared using Mann-Whitney's U test and unpaired t tests. A P value less than 0.05 was considered statistically significant.
Establishment of an Animal Model of Fulminant Hepatitis Using HBV-Infected Human Hepatocyte Chimeric Mice and Human PBMC Transplantation.
Administration of 2 × 107 PBMCs twice after suppression of mice NK cells by anti–asialo GM1 antibody21 and macrophages and DCs by liposome-encapsulated clodronate22 before transplantation enabled us to establish a human PBMC chimerism in uPA-SCID mice. We observed an up to 7% human mononuclear cell chimerism among the liver-resident mononuclear cells of uninfected and HBV-infected mice 2-14 days after the initial injection of PBMC (Fig. 1A; Table 1). Chimerism was most prominent 4 days after initial PBMC administration and almost undetectable by day 14 (Fig. 1A). Histological examination of chimeric mice livers showed extensive human liver cell death, comparable to the massive liver cell death observed in fulminant hepatitis, only in HBV-infected and PBMC-treated mice liver (Fig. 1B). Human hepatocytes were almost completely eliminated and replaced by human albumin-negative mouse hepatocytes at days 7 and 14. Consistent with these histological changes, we observed a rapid decline of HSA levels and HBV DNA only in HBV-infected and PBMC-treated mice (Fig. 1C). The decline of mice HSA levels and HBV DNA was also observed in 2 of 3 HBV-infected mice transplanted with PBMCs isolated from healthy blood donors without HBsAg vaccination (Fig. 1D and Supporting Fig. 2).
Table 1. Analysis of Liver-Infiltrating Cells by Flow Cytometry
Human NK (%)
Fas (+) NK (%)
Human NK (%)
FasL (+) NK (%)
Abbreviations: NA, not analyzed; ND, not detectable.
Analysis of Liver-Infiltrating Human Lymphocytes Necessary to Establish Massive Hepatocyte Degeneration.
We then analyzed liver-infiltrating cells with flow cytometry. Unexpectedly, we did not detect CD8-positive and tetramer-positive CTLs, as reported previously (Fig. 2A). Instead, we observed substantial numbers of CD3-negative and CD56-positive NK cells (Fig. 2B) and small numbers of pDCs and mDCs (Fig. 2C). The majority of NK cells of HBV-infected mice were FasL positive (Fig. 2D). In contrast, such FasL-positive NK cells were not detected in uninfected mice livers (Table 1; Fig. 2D), suggesting that these NK cells were activated in HBV-infected mice. These activated NK cells and DCs were detectable in mice livers only 4 days after the initial PBMC injection, but were undetectable after 2 and 7 days (Supporting Figs. 3 and 4, respectively).
Effect of DC Depletion on Establishment of Massive Hepatocyte Degeneration.
To confirm the necessity of both DCs and NK cells to complete hepatocyte destruction, we depleted DCs or NK cells with negative selection using antibody-coated magnetic beads before the administration of PBMC. Depletion of either DCs or NK cells completely abolished the decline of human albumin as well as HBV DNA (Supporting Fig. 5A). However, analysis of liver-infiltrating cells revealed that chimerism with human PBMC was poorly established in these animals, probably the result of the loss or damage of human cells by bound antibodies during separation and/or subsequent incubation in mice (Supporting Fig. 5B; Supporting Table 1).
To overcome possible confounding resulting from poor chimerism resulting in poor human hepatocyte degeneration in mice, we attempted to remove DCs from transplanted human PBMCs by alternate means. We attempted to deplete human DCs by administering clodronate 1 day before PBMC transplantation, because we thought that clodronate remaining in the mouse body would impair transplanted human DCs. As expected, we observed an almost complete elimination of DCs by this procedure without impairing PBMC chimerism (Supporting Figs. 6A and 7A; Supporting Table 1). Activation of NK cells was not observed in this setting (Supporting Figs. 6B and 7B; Supporting Table 1). Depletion of DCs completely abolished the decline of both human albumin and HBV DNA (Fig. 3). Histological examination showed that hepatocyte degeneration was absent, and that there were no TUNEL-staining–positive cells (data not shown). Clodronate lyposomes may also nonspecifically deplete macrophages and monocytes in addition to DCs, but no monocytes or macrophages were observed when transplanted PBMCs were analyzed using Ficoll-Hypaque density gradient centrifugation, indicating that the clodronate administration was specifically associated with DC depletion in this study.
Analysis of Fas/FasL System in Massive HBV-Infected Hepatocyte Degeneration Model.
We then assessed the importance of the Fas/FasL system and the occurrence of apoptosis in NK-cell–mediated human hepatocyte degeneration. Only HBV-infected human hepatocytes positive for HSA were positive for Fas antibody staining (Fig. 4A). TUNEL staining was also positive only in mice infected with HBV and inoculated with PBMCs (days 4 and 7). Measurement of mRNA levels in infected and uninfected livers showed that expression levels of Fas mRNA increased significantly upon HBV infection (Fig. 4B). To confirm that apoptosis of human hepatocytes was mediated by the Fas/FasL pathway and to determine whether IFN-α or IFN-γ played a role in the establishment of liver cell degeneration, we administered a blocking mAb against FasL, IFN-α, and IFN-γ 1 day before PBMC transplantation. Treatment of mice with antibody against FasL before PBMC completely abolished the decline of human albumin and HBV DNA (Fig. 5A). This abolishment of human albumin decline in mouse serum suggests that the Fas/FasL pathway almost exclusively eliminated infected hepatocytes in this model, which also suggests that Fas-mediated apoptosis could play an important role in FHB. Antibodies against IFN-α and IFN-γ inhibited IFN-induced ISG expression in mice livers (Supporting Fig. 8); however, these antibodies did not disturb the decline of HSA levels (Fig. 5A) and histological inflammation (Fig. 5B). Contact-dependent and -independent activation of NK cells by DCs has been reported previously.23-25 Although IFN-α and IFN-γ play a role in their activation,23, 25, 26 our results indicate that the effects of IFN-α are almost negligible in our experiments (Fig. 5A), suggesting that direct contact among these cells, or cytokines other than IFN-α and IFN-γ, are necessary to activate NK cells in this setting. NK cells have also been reported to exert antiviral effects by secreting IFN-γ. However, our results suggest that this mechanism does not work well in our model (Fig. 5A).
In this study, we established a small animal model in which massive hepatocyte degeneration similar to FHB in humans is observed. Our initial attempts to detect human PBMCs in blood or any organ in transplanted mice failed even after injecting 2 × 107 cells, which is sufficient to establish human PBMC chimerism in SCID mice.27 We assumed that failure to develop chimerism was the result of the activity of NK cells and macrophages because the activity of these cells in uPA-SCID mice is higher than in SCID mice.28, 29 Therefore, we attempted to eliminate these effects by administering clodronate and anti–asialo GM1 antibody, which are known to effectively eliminate these cells.30, 31 This assumption appears to be valid, because we were able to establish human PBMC chimerism and massive hepatocyte degeneration by suppressing these cells (Fig. 1).
HBV-specific CTLs have been reported to play an important role in eliminating the virus.32-34 Accordingly, we attempted to detect HBV-specific CTLs in mice with massive hepatocyte degeneration. Unexpectedly, we failed to detect HBV-specific CTLs (Fig. 2A and Supporting Fig. 9) and instead found that infiltrating cells in the liver were CD3-negative NK cells (Fig. 2B,D and Supporting Fig. 10). The reason for the absence of CTLs in our experiment is unknown, but this suggests that massive hepatocyte degeneration resembling fulminant hepatitis can be caused by NK cells as a main player, and recent reports demonstrating that NK cells contribute to severe acute and chronic hepatitis B (CHB) support this assertion.11, 35 We attempted to collect CTLs from HBV-infected patients and to establish hepatitis in chimeric mice. However, we rarely detected tetramer-positive CTLs in blood samples from chronically infected patients and were therefore unable to establish hepatitis using CD8-positive T cells. Consequently, a limitation of this study is that differential roles of NK cells and CTLs in massive liver cell death could not be examined.
Although it is not clear in this study how profoundly DC and NK cell activity plays a role in patients with FHB, our results suggest that the immune system can trigger severe hepatocyte degeneration. The importance of the activation of NK cells by DCs was evident, because depletion of DCs almost completely abolished the massive hepatocyte degeneration in this model (Supporting Fig. 10; Table 1). The interaction between NK cells and DCs is not well characterized, although it has been established that antigen-presenting accessory cells provide both indirect (i.e., soluble) and direct (i.e., contact-dependent) signals to T cells. Experiments in which NK cells are separated from pathogens and antigen-presenting cells by semipermeable membranes are cultured with supernatants from pathogen-activated DCs or in which cytokines are neutralized with blocking antibodies. These reports indicate that both soluble and contact-dependent signals may contribute to the activation of NK cells.23, 25, 26
The importance of the Fas/FasL system in hepatocyte damage in acute and chronic HBV infection has been reported previously.37, 38 However, the extent to which this system plays a role in human hepatitis B, especially fulminant hepatitis, is unknown. As shown in this study (Fig. 5A), inhibition of the Fas/FasL system by anti-Fas antibody dramatically reduced the effect of human PBMC transplantation. This showed the possibility that the Fas/FasL system plays an important role in the degeneration of infected hepatocytes in FHB. Further studies should be conducted to evaluate what immunological responses play important roles in human hepatitis B.
The importance of NK-cell activity suggests that the suppression of DCs and NK-cell activity or the Fas/FasL system might have therapeutic implications for FHB.11, 35 If DCs and NK-cell activity or Fas/FasL activity could be controlled in the early stages of severe acute or fulminant hepatitis, we might be able to control hepatitis activity and prevent subsequent liver failure. Of course, it would be necessary to monitor the development of chronic hepatitis after such treatment because DCs and NK cells contribute to early host defenses and shape subsequent adaptive immune response through complex cross-talk regulating the early phase of the immune response.19, 24, 39, 40
We analyzed liver damage using HBV genotype C–infected mice in this study. However, HBV genotype C is associated with more severe histological liver damage than genotype B,41 and future studies should compare immunological differences between genotypes B and C.
In summary, we established an animal model of FHB using highly repopulated human hepatocyte chimeric mice and transplanted human PBMCs. Modifications of this model will facilitate further research into acute and CHB using human immune cells, including HBV-directed CTL clones, suppressor and regulatory T cells, as well as immunological experiments to study interactions between DCs and NK cells. Such models may be useful to develop and evaluate new therapeutic strategies against HBV infection.
The authors thank Rie Akiyama and Yoko Matsumoto for their expert technical assistance. This work was carried out at the Analysis Center of Life Science, Natural Science Center for Basic Research and Development, Hiroshima University.