SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

The innate immunopathogenesis responsible for the susceptibility to hepatocyte injury in chronic hepatitis B surface antigen carriers is not well defined. In this study, hepatitis B virus (HBV) transgenic mice (named HBs-Tg) were oversensitive to liver injury after immunologic [polyinosinic:polycytidylic acid or concanavalin A (ConA)] or chemical (CCl4) triggering. It was then found that the nonhepatotoxic low dose of ConA for wild-type mice induced severe liver injury in HBs-Tg mice, which was dependent on the accumulated intraheptic natural killer (NK) cells. Expressions of NKG2D ligands (Rae-1 and Mult-1) in hepatocytes were markedly enhanced upon ConA stimulation in HBs-Tg mice, which greatly activated hepatic NK cells via NKG2D/Rae-1 or Mult-1 recognition. Interestingly, the presence of NK T cells was necessary for NK cell activation and worked as positive helper cell possibly by producing interferon-γ and interleukin-4 in this process. Conclusion: Our findings for the first time suggested the critical role of NKG2D recognition of hepatocytes by NK cells in oversensitive liver injury during chronic HBV infection. (HEPATOLOGY 2007.)

The complications of hepatitis B virus (HBV) infection such as acute and chronic hepatitis, cirrhosis, and hepatocellular carcinoma are among the most important human health problems.1, 2 Noticeably, healthy hepatitis B surface antigen (HBsAg) carriers are susceptible to biochemical and histological activation of hepatocyte injury, which is characterized by programmed responses of inflammation and regeneration, as a serious risk for development of liver cirrhosis and hepatocellular carcinoma.3 It is important to identify the pathogenesis for the susceptibility to hepatocyte injury in chronic HBsAg carriers. From studies on HBV transgenic mice mimicking human healthy HBsAg carriers, it was reported that sensitive liver injury was related to the retained HBsAg within the endoplasmic reticulum of hepatocytes, which caused the hepatocytes to be extraordinarily sensitive to the cytolytic effect of interferon-γ (IFN-γ).4, 5 It was assumed that most of the liver cell injury was mediated by nonspecific inflammatory cells. As components of the overwhelming innate immune system in the liver, natural killer (NK) cells have been documented to be involved in the sensitive liver injury process in the HBV transgenic mice triggered by α-galactosylceramide (α-GalCer).4 It has been reported that NK cells did not contribute to α-GalCer–induced hepatocyte injury in wild-type mice, which was mediated by natural killer T (NKT) cells through the Fas-Fas ligand pathway.6 So, the actual mechanisms by which NK cells become the effectors of hepatocyte injury remain elusive.

The ligands of NKG2D (natural killer group 2, member D), an activating killer cell receptor, are known to be “stress-inducible” molecules, triggered by transformation or infection with viral and bacterial pathogens.7, 8 NKG2D serves a fundamental role in the surveillance against microbial infection and cancer, but may also be deleterious. Inappropriate or dysregulated expression of their ligands on target cells may induce NK-mediated or T cell–mediated autoimmune responses. The involvement of NKG2D and its ligands has been revealed in autoimmune diseases such as rheumatoid arthritis, celiac disease, and autoimmune diabetes.9–12 These findings prompted us to investigate whether interaction between NKG2D receptor and its ligands is involved in the sensitive liver injury in the form of autoimmune injury.

In this study, we first identified that NKG2D-ligand interaction was essential for the oversensitive liver injury of chronic HBsAg carriers (HBs-Tg mice) in response to the immunomodulator concanavalin A (ConA). Higher NKG2D activation of NK cells via recognition with Rae-1 or Mult-1 on HBsAg-expressed hepatocytes accounted for the stronger direct cytolysis-mediated liver injury after triggering with a low dose of ConA. Meanwhile, NKT cells, which are responsive to ConA stimulation, worked as helpers for NK cell activation in HBs-Tg mice after low-dose ConA injection.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Animals.

Eight-week-old to 10-week-old male HBV transgenic mice C57BL/6J-TgN (AlblHBV) 44Bri (named as HBs-Tg mice in our study), which harbor the gene encoding S, PreS1, and PreS2 domains of HBV and express HBsAg in serum, liver, and kidney tissues, but with no virus replication, were purchased from Department of Laboratory Animal Science of Peking University who obtained the mice from the Jackson Laboratory (Bar Harbor, ME) and bred them for us. The littermate C57BL/6J mice were obtained as the control. All mice were maintained under specific pathogen-free and controlled conditions (22°C, 55% humidity, and 12-hour day/night rhythm) and received care in compliance with the guidelines outlined in the Guide for the Care and Use of Laboratory Animals.

Reagents.

ConA (type IV) (Sigma Chemical Co., St. Louis, MO) and polyinosinic:polycytidylic acid [poly(I:C)] (Sigma Chemical Co.) were dissolved in pyrogen-free phosphate-buffered saline (PBS) at a concentration of 1 mg/mL and injected to mice intravenously and intraperitoneally at specific doses. Mice were administered CCl4 (2 mg/kg body weight) intraperitoneally. The monoclonal antibodies (mAbs) used for flow cytometry included fluorescein isothiocyanate (FITC)–conjugated anti-NK1.1, FITC–anti-CD25, FITC–anti-MHC class I molecules (H-2Db), FITC–anti-IgG2a (isotype), phycoerythrin (PE)–conjugated anti-NK1.1, PE–anti-CD69, PE–anti-CD122, PE–anti-retinoic acid early transcript 1 (Rae-1), PE–anti-IgG2a (isotype), PE–anti-IFN-γ (eBioscience, San Diego, CA), PE–anti-interleukin-4 (IL-4) (PharMingen, San Diego, CA), and CyCrome (CY) –conjugated-anti-CD3e (PharMingen).

Assay for Serum Aminotransferase Activities.

Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) activities were determined by the standard photometric method using the serum aminotransferase test kit (Rong Sheng, Shanghai, China) following the supplier's protocol.

Histological Examination.

Liver samples were fixed in 10% neutral-buffered formalin and embedded in paraffin. Sections of 5-μm thickness were affixed to slides, deparaffinized, dehydrated, and then stained with hematoxylin and eosin using routine methods.

Isolation of Liver Mononuclear Cells.

Liver mononuclear cells (MNCs) were prepared as described.13 Briefly, the liver was pressed through a 200-gauge stainless steel mesh and then suspended in RPMI-1640 medium (Gibco BRL) containing 5% fetal bovine serum. After 1 washing, the cells were resuspended in 40% Percoll (Gibco BRL) solution containing 100 U/mL heparin, and the cell mixture was gently overlaid onto 70% Percoll solution and then centrifuged at 750g for 20 minutes at room temperature. The interface cells between the Percoll solutions were aspirated and washed twice with RPMI-1640 medium containing 5% fetal bovine serum.

Cell Depletion.

Anti-NK1.1 mAb (PK136) was obtained from partially purified hybridoma culture supernatant by ammonium sulfate precipitation (American Type Culture Collection, Manassas, VA). Mice were given 2 intraperitoneal injections of the indicated mAb (50 μg per mouse) 48 and 24 hours before subsequent treatment to deplete NK1.1-positive cells (NK and NKT cells). A dose of 50 μg anti-ASGM-1 antibody (Wako Pure Chemical, Osaka, Japan) was injected intravenously to mice 24 hours before subsequent treatment to deplete NK cells. These protocols resulted in a ≥90% decrease in the number of indicated cells.

Adoptive Transfer of Liver MNCs.

The recipient was treated with anti-NK1.1 mAb 72 and 48 hours before adoptive transfer to deplete NK and NKT cells but not to affect the transferred MNCs. Under ether anesthesia, hepatic MNCs (5 × 106 cells) suspended in 100 μL pyrogen-free PBS were injected into the lateral left lobe of the liver at a rate of 10 μL/second using a 29-gauge needle attached to a 1-mL syringe, followed by intravenous injection of ConA, as described.14

Flow Cytometry Analysis.

After being blocked with normal rat immunoglobulin to saturate rat Fc receptor, cells were stained with saturating amount of the indicated fluorescence-labeled mAbs at 4°C for 30 minutes in darkness for the surface antigens, and then washed 3 times and acquired by FACSCalibur (Becton Dickinson) and analyzed with WinMDI version 2.8 software. For the intracellular cytokine assay, after the surface antigens became stained, cells were fixed and permeabilized using 100 μL each of cytofix and cytoperm solution (eBioscience, San Diego, CA), and then were stained with PE–anti-IFN-γ or PE–anti-IL-4.

Purification of NK and NKT Cells.

Hepatic MNCs were stained with PE-anti-NK1.1 (or FITC-anti-NK1.1) mAb and CY-anti-CD3 mAb at 4°C for 30 minutes, and then washed 3 times. Subsequently, the stained MNCs were automatically sorted by flow cytometry (Becton Dickinson) in PBS with a total volume of 1 mL/1 × 107 cells. The separated cells were ≥95% pure.

Isolation of Mouse Hepatocytes.

Mice were anesthetized with sodium pentobarbital (intraperitoneally with 30 mg/kg body weight), and then the portal vein was cannulated. The liver was subsequently perfused with ethylene glycol tetraacetic acid (EGTA) solution (5.4 mM KCl, 0.44 mM KH2PO4, 140 mM NaCl, 0.34 mM Na2HPO4, 0.5 mM EGTA, and 25 mM tricine, pH 7.2) and digested with 0.075% collagenase solution.15 Then the viable hepatocytes suspended in Dulbecco's modified Eagle medium (Life Technologies, Gaithersburg, MD) solution were separated by 40% Percoll (Gibco BRL) solution with centrifugation at 420g for 10 minutes at 4°C.

Reverse Transcription PCR Analysis.

RNA was extracted from hepatocytes by Trizol Reagent (Invitrogen, Carlsbad, CA). Cellular RNA (1 μg) was used for complementary DNA synthesis. Polymerase chain reaction (PCR) primers for detecting Qa-1, mouse UL16-binding protein-like transcript 1 (Mult-1), and β-actin were as follows: Qa-1, sense, 5′-CATTCGCTGCGGTATTT-3′, antisense, 5′-GGTATGCCCTCTGTTGGTG-3′; Mult-1, sense, 5′-GGGAGCCTTCCATCAGC-3′, antisense, 5′-GTGACGGGCAAGCAGTA-3′; β-actin, sense, 5′-ATGGATGACGATATCGCT-3′, antisense, 5′-ATGAGGTAGTCTGTCAGGT-3′. Electrophoretic bands were then analyzed by Scion Image software for the relative density compared with β-actin.

NKG2D Blockade.

Functional grade-purified anti-mouse NKG2D (blocking) (clone: CX5) was purchased from eBioscience (San Diego, CA). When injected in vivo, anti-NKG2D mAb CX5 either blocked or modulated NKG2D, but did not deplete NK cells.11, 12 A dose of 250 μg anti-NKG2D mAb was injected intravenously to the mouse 5 hours before ConA treatment for the blockade in vivo. A dose of 20 μg/mL anti-NKG2D mAb was added into the culture for the blockade in vitro.

Cytotoxicity Assay.

To assay the cytotoxicity of hepatic NK cells against hepatocytes, a 4-hour AST release assay was performed.15 Hepatic NK cells purified from 2-hour ConA-treated HBs-Tg mice were added to the freshly isolated hepatocytes from 2-hour ConA-treated HBs-Tg mice at the indicated effector/target (E/T) cell ratios. Hepatocytes (1 × 104) were used as target cells in the assay. After 4 hours, the supernatant was harvested, and AST activity was measured. The specific cytotoxicity was calculated as: [(ASTexperimental − ASTspontaneous)/ (ASTmaximum − ASTspontaneous)] × 100%.

Cytokine Neutralization.

Anti-IFN-γ mAb (R4-6A2) and anti-IL-4 mAb (11B11) were obtained from partially purified hybridoma culture supernatant by ammonium sulfate precipitation (American Type Culture Collection, Manassas, VA). Mice were given 2 injections of the indicated mAb (50 μg per mouse) intraperitoneally 48 hours and 24 hours before ConA treatment to neutralize the cytokine in vivo (see supplementary figures). The Materials and Methods was for the results of Supplementary Figures 2 and 3.

Statistical Analysis.

The results were analyzed by Student t test or analysis of variance where appropriate. All data were shown as mean ± standard error of the mean (SEM). P value < 0.05 was considered to be statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

HBs-Tg Mice Were Oversensitive to Liver Injury Triggered by Exoteric Stimulations.

As shown by the levels of serum ALT, injection of poly(I:C), ConA, or CCl4 induced much more severe hepatocyte injury in HBs-Tg mice than that of C57BL/6 mice (Fig. 1A). As reported previously, injection of ConA induced liver injury in C57BL/6 mice in a dose-dependent manner.16 A low dose of ConA (3 μg/g body weight) injection could not induce liver injury in C57BL/6 mice, but induced severe liver injury in HBs-Tg mice, demonstrated by serum ALT levels (Fig. 1B) and liver histopathological changes (Fig. 1C). A high dose of ConA (15 μg/g body weight) injection induced the peak level of serum ALT but seldom caused death in wild-type mice, but caused death of most HBs-Tg mice (Fig. 1D). To look for the cells which correlated to liver injury induced by the low dose of ConA in HBs-Tg mice, adoptive hepatic mononuclear cell transfer was performed in mice depleted of NK and NKT cells. Treatment with anti-NK1.1 mAb (PK136) completely depleted both NK and NKT cells (data not shown) and abolished the elevation of serum ALT induced by the low dose of ConA (Fig. 2). With injection of the low dose of ConA (3 μg/g body weight), adoptive transfer of hepatic MNCs from HBs-Tg mice caused severe liver injury as shown by serum ALT levels (4390 ± 219.5 U/L) in HBs-Tg mice, but only relatively weak injury (380 ± 19 U/L) in C57BL/6 mice. On the other hand, adoptive transfer of hepatic MNCs from C57BL/6 mice could only cause mild liver injury (1350 ± 67.5 U/L) in HBs-Tg mice, but almost could not induce any injury in wild-type mice. These results suggested that hepatic MNCs became more cytotoxic and hepatocytes more sensitive to cytotoxicity from HBs-Tg mice after stimulation with a low dose of ConA.

thumbnail image

Figure 1. HBs-Tg mice were oversensitive to liver injury. (A) Poly(I:C) (5 μg/g body weight, intraperitoneally), ConA (3 μg/g body weight, intravenously), CCl4 (2 μg/g body weight, intraperitoneally) were injected to HBs-Tg mice and their littermates C57BL/6 mice. At 24 hours after injection, serum ALT activity was determined. *P < 0.05 as comparison with C57BL/6 mice. (B) Different doses of ConA (1 μg/g, 3 μg/g, 9 μg/g, and 15 μg/g body weight) were injected intravenously to HBs-Tg mice and their littermates C57BL/6 mice. At 0, 6, 12, 18, 24, and 48 hours after ConA injection, serum ALT activity was determined. Data were shown as mean ± SEM from 5 mice in each group. (C) Liver samples were collected 18 hours after ConA (3 μg/g body weight) or PBS injection for hematoxylin and eosin staining. “N” indicated the necrosis area (original magnification ×200). (D) After stimulation of ConA (15 μg/g or 20 μg/g body weight), survival rate of HBs-Tg mice and C57BL/6 mice were measured. There were 6 mice in each group. Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; ConA, concanavalin A; PBS, phosphate-buffered saline; wt, body weight.

Download figure to PowerPoint

thumbnail image

Figure 2. Hepatic MNCs were more cytotoxic and hepatocytes more sensitive to cytotoxicity from HBs-Tg mice after stimulation with a low dose of ConA. Hepatic MNCs (5 × 106) from HBs-Tg mice or C57BL/6 mice were adoptively transferred into the liver of PK136-treated recipient mice followed by intravenous injection of ConA (3 μg/g body weight). Serum ALT was measured 18 hours after cell transfer. Data are shown as mean ± SEM from 4 mice in each group. *P < 0.05. Abbreviations: ALT, alanine aminotransferase; ConA, concanavalin A; MNC, mononuclear lymphocyte; Tg, HBs-Tg mice.

Download figure to PowerPoint

The Liver Injury Induced by the Low Dose of ConA Was Dependent on NK Cells in HBs-Tg Mice.

After the low dose of ConA injection, more NK cells accumulated in the liver of HBs-Tg mice, and these NK cells were fully activated as shown by up-regulated CD69 and CD122 expressions (Fig. 3A). No significant difference in other lymphocyte populations such as NKT and T cells was observed between HBs-Tg mice and C57BL/6 mice (data not shown). Depletion of NK cells by pretreatment with anti-ASGM-1 mAb in HBs-Tg mice absolutely alleviated liver injury induced by the low dose of ConA, shown by significantly decreased serum ALT and AST levels (Fig. 3B) and ameliorated histopathological changes (Fig. 3C). Because the alleviating functions of depletion of both NK and NKT cells by pretreatment with anti-NK1.1 mAb and depletion of NK cells by pretreatment with anti-ASGM-1 mAb were almost the same, we concluded that liver injury induced by the low dose of ConA in HBs-Tg mice was dependent on NK cells (Fig. 3B,C).

thumbnail image

Figure 3. NK cells played a critical role during low-dose ConA-induced liver injury in HBs-Tg mice. (A) At the indicated time points after ConA (3 μg/g body weight) injection, hepatic MNCs were isolated and analyzed by flow cytometry using FITC-anti-NK1.1 and CY-anti-CD3 mAbs. According to the percentage of NK cells (CD3NK1.1+), NK cell number was calculated by total viable MNCs. Data are shown as mean ± SEM from 4 mice in each group. *P < 0.05 as comparison with C57BL/6 mice. CD69, CD25, and CD122 expression on hepatic NK cells were analyzed by flow cytometry 18 hours after ConA (3 μg/g body weight) or PBS injection. The open histogram is for HBs-Tg mice and the gray histogram is for C57BL/6 mice. (B) Cell depletion was conducted by anti-NK1.1 mAb (PK136) and anti-ASGM-1 antibody for HBs-Tg mice. Mice were then treated with ConA (3 μg/g body weight). Serum ALT and AST levels were determined 18 and 48 hours after ConA injection, respectively. Data are shown as mean ± SEM from 4 mice in each group. *P < 0.05. (C) Liver samples were collected for hematoxylin and eosin staining from PK136-treated and anti-ASGM-1–treated HBs-Tg mice 18 and 48 hours after ConA (3 μg/g body weight) injection, respectively (original magnification ×200). Abbreviations: Ab, antibody; ALT, alanine aminotransferase; AST, aspartate aminotransferase; ConA, concanavalin A; Ig, immunoglobulin; NK, natural killer; PBS, phosphate-buffered saline.

Download figure to PowerPoint

NKG2D Activation Played a Critical Role in NK Cell–Mediated Liver Injury in HBs-Tg Mice after Administration of the Low Dose of ConA.

NK cell receptors and their ligands on hepatocytes were then investigated. Expression of H-2Db (Fig. 4A) and Qa-1 (Fig. 4B), ligands of NK cell inhibitory receptors, was down-regulated in hepatocytes of HBs-Tg mice as compared with wild-type C57BL/6 mice. ConA injection did not significantly affect their expression. In contrast, expression of Rae-1 (Fig. 4C) and Mult-1 (Fig. 4D), 2 important ligands of activating receptor NKG2D, were significantly up-regulated in hepatocytes of HBs-Tg mice, but not in hepatocytes of wild-type C57BL/6 mice, upon stimulation with the low dose of ConA. Blockade of NKG2D recognition by anti-NKG2D blocking mAb dramatically inhibited the liver injury (Fig. 4E), and the effect of NKG2D blockade was almost the same as the depletion of NK cells by anti-ASGM-1 mAb (Fig. 3B). The dependence of NKG2D-ligand interaction between NK cells and hepatocytes was then confirmed by direct cytotoxicity of the purified NK cells against hepatocytes from ConA-treated HBs-Tg mice in a 4-hour AST release assay in vitro (Fig. 4F). This result was further confirmed by our observation that adoptive transfer of the activated hepatic NK cells from ConA-treated HBs-Tg mice could deliver liver injury in NK/NKT-depleted HBs-Tg mice (Fig. 5).

thumbnail image

Figure 4. NK cells mediated liver injury through NKG2D recognition in HBs-Tg mice after injection with low-dose ConA. (A and C) Hepatocytes were prepared for flow cytometry analysis of H-2Db and Rae-1 at 6 hours after ConA (3 μg/g body weight) or PBS injection. These results were from a single experiment representative of 3 independent experiments. (B and D) Hepatocytes were prepared at 0, 1, and 3 hours after ConA (3 μg/g body weight) injection for reverse transcription PCR analysis of Qa-1 and Mult-1 mRNA. “+” indicates HBs-Tg mice and “−” indicates C57BL/6 mice. Results were then analyzed by Scion Image software for the relative density compared with β-actin. Values are shown as mean ± SEM from 3 independent experiments. *P < 0.05 as comparison with C57BL/6 mice. (E) A dose of 250 μg anti-NKG2D (blocking) mAb was administrated to HBs-Tg mice 5 hours before ConA (3 μg/g body weight) injection. Serum ALT and AST levels were determined 18 hours after ConA injection. Data are shown as means ± SEM from 4 mice in each group. *P < 0.05 as comparison with the control. (F) The 4-hour AST release assay was performed to test cytotoxicity of hepatic NK cells against hepatocytes. Hepatic NK cells purified from HBs-Tg mice treated for 2 hours with ConA (3 μg/g body weight) were added to the freshly isolated hepatocytes from HBs-Tg mice treated for 2 hours with ConA (3 μg/g body weight) at the effector cell/target cell ratio (E/T) of 8/1. Hepatocytes (1 × 104) were used as target cells. A dose of 20 μg/mL anti-NKG2D (blocking) mAb was added in the culture for the blockade. Data are shown as mean × SEM from triplicates in each group. *P < 0.05 as comparison with the control. Abbreviations: PBS, phosphate-buffered saline; ConA, concanavalin A; ALT, alanine aminotransferase; AST, aspartate aminotransferase; NK, natural killer; Rae-1, retinoic acid early transcript 1; Mult-1, mouse UL16-binding protein-like transcript 1; Ig, immunoglobulin.

Download figure to PowerPoint

thumbnail image

Figure 5. Activated NK cells delivered liver injury in NK/NKT cell–depleted HBs-Tg mice. (A) The procedures of adoptive NK cell transfer into liver of NK/NKT cell–depleted HBs-Tg mice are shown. Hepatic NK cells were sorted by flow cytometry from HBs-Tg mice treated for 6 hours with ConA (3 μg/g body weight). Purified NK cells (1 × 106) were adoptively transferred into the liver of PK136-treated HBs-Tg mice which had been treated with ConA (3 μg/g body weight) for 6 hours for the similar condition as donor mice. Twelve hours after NK cell transfer, liver injury was observed by analysis of serum ALT/AST level and tissue pathology. (B) CD69 expression on hepatic NK cells from 6-hour ConA-treated HBs-Tg mice was analyzed by flow cytometry. The gray histogram is for hepatic NK cells from untreated HBs-Tg mice as control. The purification of sorted hepatic NK cells from 6-hour ConA-treated HBs-Tg mice is shown. Serum ALT and AST levels were measured 12 hours after NK cell transfer. Data are shown as mean ± SEM from 4 mice in each group. *P < 0.05 as comparison with the control. Abbreviations: ConA, concanavalin A; NK, natural killer; ALT, alanine aminotransferase; AST, aspartate aminotransferase.

Download figure to PowerPoint

Help Function of NKT Cells in NK Cell–Mediated Liver Injury of HBs-Tg Mice.

Because high doses of ConA might induce acute hepatitis of mice by directly activating hepatic NKT cells,14, 17 we then studied the role of a low dose of ConA in triggering NKT cells in NK cell–mediated oversensitive liver injury in HBs-Tg mice. Adoptive cell transfer was performed as shown in Fig. 6A. NK cell-depletion alleviated the ConA-induced liver injury of HBs-Tg mice (Fig. 3B,C). Surprisingly, we found that adoptive transfer of resting NK cells from HBs-Tg mice could not mediate liver injury in NK/NKT cell–depleted HBs-Tg mice even after stimulation with the low dose of ConA (Fig. 6B), suggesting the presence of NKT cells was necessary for NK cell–mediated liver injury. When adoptively transferring resting NK cells and NKT cells together (NK/NKT cells), ConA induced severe liver injury (Fig. 6B), further demonstrating the necessity of NKT cell presence. The reason why NK cells did not deliver liver injury in NK/NKT cell–depleted HBs-Tg mice was possibly explained by inactivation of NK cells without NKT cell help, shown by CD69 expression in Fig. 6C. Although NKT cells rapidly disappeared after ConA stimulation for activation-induced cell death,14, 18 cytokines secreted by NKT cells might exert a function (Fig. 6C). As reported by others19,20 and us (Supplementary Fig. 1), hepatic NKT cells produced IFN-γ and IL-4 upon ConA stimulation, and then the function of these 2 cytokines was neutralized with anti-IFN-γ or anti-IL-4 mAb, respectively. The results showed that the low dose of ConA could not induce liver injury even in the presence of NK cells in HBs-Tg mice (Supplementary Fig. 2), in which NK cell activation was inhibited (Supplementary Fig. 3). These results suggest that NKT cells may possibly function as helpers for NK cell activation by producing IFN-γ and IL-4 during liver injury in HBs-Tg mice.

thumbnail image

Figure 6. The presence of NKT cells was necessary for NK cell–mediated liver injury in HBs-Tg mice after injection of low-dose ConA. (A) The procedures of adoptive cell transfer to liver of NK/NKT cell–depleted HBs-Tg mice are shown. Hepatic NK cells or NKT cells were sorted from untreated HBs-Tg mice by flow cytometry. Purified NK cells (1 × 106) were then adoptively transferred to the liver of PK136-treated HBs-Tg mice to obtain only NKT cell–depleted HBs-Tg mice, which were followed by intravenous injection of ConA (3 μg/g body weight). At 18 hours after ConA injection, (B) liver injury was observed and (C) hepatic MNCs were analyzed by flow cytometry. In the positive control, both NK and NKT cells (NK/NKT) sorted from untreated HBs-Tg mice were transferred to the liver of PK136-treated HBs-Tg mice. (B) Serum ALT was measured 18 hours after ConA injection. Data are shown as mean ± SEM from 4 mice in each group. *P < 0.05. (C) At 18 hours after ConA injection, liver MNCs were isolated from PK136-treated HBs-Tg mice transferred with NK cells or NK/NKT cells as demonstrated in (A), and then analyzed for CD69 expression on transferred NK cells. The open histogram indicates the transferred NK cells in PK136-treated HBs-Tg mice injected with ConA and the gray histogram indicates the NK cells without ConA stimulation. Abbreviations: ConA, concanavalin A; ALT, alanine aminotransferase; NK, natural killer.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Approximately 10% of adults and 90% of children become chronic virus-carriers after HBV infection, around 350 million persistent HBV carriers worldwide.1 Susceptibility to liver injury is a serious problem in this huge population of chronic HBV carriers,3 which was demonstrated in HBs-Tg mice triggered by stimulations such as poly(I:C), ConA, or CCl4 injection in our study (Fig. 1A). However, the mechanisms of susceptibility have not been well explored. In subjects with hepatitis C virus infection, it is suggested that hepatitis C virus core and envelope proteins may be one causative agent of spontaneous liver cell injury by themselves.21 When filamentous HBsAg particles accumulate in HBV-infected hepatocytes to extremely high levels, the death of hepatocytes will be induced by severely compromised endoplasmic reticulum function.22 Noticeably, functionally normal HBsAg-positive hepatocytes are exquisitely hypersensitive to destruction, as reported that HBsAg-positive hepatocytes of HBV transgenic mice were hypersensitive to cytolytic effect of IFN-γ.5 Further, NKT cells triggered by α-GalCer induced more severe liver injury in HBs-Tg mice, which was verified to be dependent on NK cells.4 Though it has been demonstrated that innate immune cells such as NK or NKT cells are involved in oversensitive injury of virus-infected liver, the precise immunologic mechanisms remain unclear. In this study, we demonstrated that NKG2D activation of NK cells via recognition with NKG2D ligands (Rae-1 and Mult-1) on hepatocytes accounted for the oversensitive hepatocyte injury in HBs-Tg mice.

As immunomodulators, ConA and α-GalCer were used to investigate the innate immunologic mechanisms of acute liver injury. It has been reported that NKT cells, independent of NK cells, played the critical role in ConA-induced or α-GalCer-induced liver injury in wild-type mice,6, 14, 17, 23 though only α-GalCer was verified to directly activate NKT cells. However, it was observed that NKT cell–induced liver injury was more sensitive in an NK cell–dependent manner in HBs-Tg mice after α-GalCer triggering,4 which leaves the controversy on NK cell function during innate lymphocyte-mediated liver injury between wild-type mice and HBs-Tg mice. In our study, we also observed this controversy in wild-type mice and HBs-Tg mice by triggering with ConA (Fig. 3). Most interesting was that we observed the actual function of NK cells and their relation to NKT cells (Figs. 4, 5, and 6). Our results indicate that ConA-activated hepatic NK cells directly effectively attacked HBsAg-positive hepatocytes, but NKT cells did not induce any liver injury when NK cells were depleted (Figs. 3B,C, 4, and 5); meanwhile, when both NK and NKT cells were depleted by anti-NK1.1 mAb and the purified NK cells adoptively transferred back into mice (a likely NKT cell–deficient mouse), ConA administration did not induce any liver injury (Fig. 6B). Thus, NK cells were effectors but their function was dependent on the presence of NKT cells (Fig. 6B,C). NKT cells worked as helpers for NK cell activation in the process (Fig. 6C). Because Trobonjaca et al. did not delineate the relationship between NK and NKT cells in their α-GalCer–induced liver injury in HBs-Tg mice,4 we could not compare our conclusion with theirs. Why NKT cells, but not NK cells, play an important role in ConA-induced liver injury in wild-type mice, and both NK and NKT cells exert injury effects on liver in HBs-Tg mice, needs further investigation.

NKG2D has been implicated in autoimmune diseases by studies in human and mouse. It was first revealed in rheumatoid arthritis that profound dysregulation of NKG2D on CD4+CD28 T cells and abnormal expression of MICA/B on synoviocytes caused autoreactive T cell stimulation.9 In celiac disease, IELs (intestinal intraepithelial lymphocytes) had changed their characteristics from the typical of antigen-specific T cells to a phenotype of NK-like cells, which mediated epithelial cell damage through NKG2D recognition of MICA/B on epithelial cells in an antigen-independent pathway.10, 24 Additionally, autoreactive CD8+ T cells infiltrating the pancreas expressed NKG2D and islet cells in the prediabetic NOD (non-obese diabetic) mice expressed Rae-1, which caused the development of autoimmune diabetes.11, 12 A role for NKG2D in mouse hepatitis virus (MHV) was suggested by studies of Dandekar et al.,25 who showed that the pathology caused by MHV infection of the central nervous system was partly due to an NKG2D-dependent mechanism mediated by γδT cells. Here, we report that NKG2D on NK cells and abnormal expression of its ligands on hepatocytes accounted for the development of autoimmune hepatocyte injury during HBV infection. In fact, there was no significant difference in the expression of NKG2D on intrahepatic lymphocytes between HBs-Tg mice and wild-type mice, even after ConA injection (data not shown). The enhanced expression of NKG2D ligands (Rae-1 and Mult-1) in hepatocytes accounted for NK cell cytotoxicity against hepatocytes in HBs-Tg mice triggered by the low dose of ConA (Fig. 4). Why Rae-1 and Mult-1 genes were expressed preferentially in HBsAg-expressed hepatocytes but not in wild-type hepatocytes when stimulated with ConA was unresolved. One possibility was a consequence of viral gene integration, which made NKG2D ligands easily inducible. It has been reported that infection with murine cytomegalovirus also leads to the up-regulated expression of NKG2D ligand Rae-1.26 As a consequence of the NOD genetic background, pancreatic cells also inadvertently express Rae-1 genes.11 On the other hand, viral gene integration in the HBV transgenic mice caused major histocompatibility complex (MHC) class I molecule (H-2Db and Qa-1) expression decreased in hepatocytes (Fig. 4A,B). A mechanism has been reported for virus infection, such as adenoviruses, herpes simplex virus, and human and murine cytomegaloviruses,27 or even in human HBV-infected hepatocyte cell lines,28 in which down-regulation of MHC class I molecules on the surface of infected cells was observed, which was explained as immune escape of virus from adaptive T cell attack by limiting the number of antigen-presenting MHC class I molecules. Here, we found that down-regulation of MHC class I molecules on hepatocytes might activate innate immune cells by the “missing-self” mechanism, which has been also observed in Pseudomonas aeruginosa exotoxin A–induced hepatotoxicity.29 The underlying molecular mechanisms of down-regulation of MHC class I molecules by virus remain unknown. The balance in the expression of activating and inhibitory ligands will determine whether a cell becomes a target for NK cell–mediated killing.30 Diminishing the ligands of inhibitory receptors (H-2Db and Qa-1) and increasing the ligands of activating receptor NKG2D (Rae-1 and Mult-1) on hepatocytes may possibly lead hepatocytes to become a target of NK cell–mediated killing in the development of autoimmune liver injury in HBs-Tg mice.

Notably, Radaeva et al. reported hepatic NK cell-mediated killing through NKG2D recognition in the development of liver fibrosis. High levels of Rae-1 were induced on activated hepatic stellate cells, whereas Rae-1 was undetectable on quiescent hepatic stellate cells, by which NK cell activation induced cell death to activated hepatic stellate cells and attenuated the severity of liver fibrosis.31 During immune disorders in the liver, NKG2D ligands might be inducible in several types of liver cells. After ConA stimulation, expression of NKG2D ligands (Rae-1 and Mult-1) in other types of liver cells in addition to hepatocytes deserves further investigation.

In this study, we demonstrated that NK cell activation by a low dose of ConA was assisted by NKT cells in HBs-Tg mice (Fig. 6C), in which the mechanism was speculated that NKT cells regulated NK cell activation by producing cytokines such as IFN-γ and IL-4 (Supplementary Figs. 1-3). It has been reported that NK cells could not be directly activated by ConA, which required other types of immune cells and was critically dependent on IFN-γ.19 As extensively reported, IFN-γ is an important link between NK and NKT cells, through which NK cells were activated to function.19, 32–35 Although HBs-Tg mice were extremely hypersensitive to the cytolytic effect of IFN-γ,5 NKT cells producing IFN-γ only acted as helpers for NK cells in oversensitive liver injury of HBs-Tg mice induced by a low dose of ConA. The precise mechanism(s) in NKT cell help for NK cells needs more work to define.

Oversensitive liver injury induced by a low dose of ConA in murine chronic HBsAg carriers provided us an interesting platform to investigate the precise mechanisms of hepatocyte susceptibility to injury during chronic HBV infection. Our findings suggest NKG2D activation of NK cells via recognition with Rae-1 or Mult-1 on hepatocytes induced by ConA stimulation accounted for the oversensitive hepatocyte injury, and NKT cells worked as helpers necessary for NK cell activation possibly via secreting IFN-γ and IL-4 in this condition. This conclusion would be helpful to interpret the immunological mechanisms of hepatocyte susceptibility to injury in human chronic HBsAg carriers.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Supplementary material for this article can be found on the H EPATOLOGY Web site ( http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html ).

FilenameFormatSizeDescription
jws-hep.21872.fig1.pdf119K Supplementary Fig. 1. Cytokine production of intrahepatic NKT cells in HBs-Tg mice after a low dose of Con A injection.After 1 h stimulation of Con A (3 μg/g body weight) or PBS, intrahepatic NKT cells were analyzed by flow cytometry for the production of IFN-γ and IL-4. These were from a single experiment representative of three independent experiments. PBS, phosphatebuffered saline; Con A, concanavalin A; IFN-γ, interferon-gamma; IL-4, interleukin-4.
jws-hep.21872.fig2.pdf98K Supplementary Fig. 2. Neutralization of IFN-g or IL-4 alleviated over-sensitive liver injury in HBs-Tg mice after a low dose of Con A injection.IFN-γ(A) or IL-4 (B) was neutralized by the indicated antibody in HBs-Tg mice before Con A (3 μg/g body weight) injection respectively. Serum ALT and AST were determined 18 h after Con A injection. Data were shown as mean ± SEM from four mice in each group. Con A, concanavalin A; ALT, alanine aminotransferase; AST, aspartate aminotransferase; Ig, immunoglobulin; Ab, antibody; IFN-γ, interferon-gamma; IL-4, interleukin-4.
jws-hep.21872.fig3.pdf115K Supplementary Fig. 3. Neutralization of IFN-g or IL-4 inhibited NK cell activation in HBs-Tg mice after a low dose of Con A injection.IFN-γ or IL-4 was neutralized by the indicated antibody in HBs-Tg mice before Con A (3 μg/g body weight) injection respectively. At 18 h of Con A injection, CD69 expression on intrahepatic NK cells was analyzed by flow cytometry. The gray histogram was for the NK cells from control HBs-Tg mice with PBS injection. Ig, immunoglobulin; Ab, antibody; IFN-γ, interferon-gamma; IL-4, interleukin-4.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.