Infection of mice with an Adenovirus expressing the major human autoantigen cytochrome P450 2D6 (hCYP2D6) recognized by sera of patients suffering from type 2 autoimmune hepatitis results in a persistent form of autoimmune-mediated liver disease characterized by extensive hepatitis, fibrosis and generation of a CYP2D6-specific immune response.
Autoimmune hepatitis is a rare but life threatening autoimmune disease of the liver of unknown etiology1,2. In the past many attempts have been made to generate an animal model that reflects the characteristics of the human disease 3-5. However, in various models the induction of disease was rather complex and often hepatitis was only transient3-5. Therefore, we have developed a straightforward mouse model that uses the major human autoantigen in type 2 autoimmune hepatitis (AIH-2), namely hCYP2D6, as a trigger6. Type 1 liver-kidney microsomal antibodies (LKM-1) antibodies recognizing hCYP2D6 are the hallmark of AIH-27,8. Delivery of hCYP2D6 into wildtype FVB or C57BL/6 mice was by an Adenovirus construct (Ad-2D6) that ensures a direct delivery of the triggering antigen to the liver. Thus, the ensuing local inflammation generates a fertile field9 for the subsequent development of autoimmunity. A combination of intravenous and intraperitoneal injection of Ad-2D6 is the most effective route to induce a long-lasting autoimmune damage to the liver (section 1). Here we provide a detailed protocol on how autoimmune liver disease is induced in the CYP2D6 model and how the different aspects of liver damage can be assessed. First, the serum levels of markers indicating hepatocyte destruction, such as aminotransferases, as well as the titers of hCYP2D6 antibodies are determined by sampling blood retroorbitaly (section 2). Second, the hCYP2D6-specific T cell response is characterized by collecting lymphocytes from the spleen and the liver. In order to obtain pure liver lymphocytes, the livers are perfused by PBS via the portal vein (section 3), digested in collagen and purified over a Percoll gradient (section 4). The frequency of hCYP2D6-specific T cells is analyzed by stimulation with hCYP2D6 peptides and identification of IFNγ-producing cells by flow cytometry (section 5). Third, cellular infiltration and fibrosis is determined by immunohistochemistry of liver sections (section 6). Such analysis regimen has to be conducted at several times after initiation of the disease in order to prove the chronic nature of the model. The magnitude of the immune response characterized by the frequency and activity of hCYP2D6-specific T and/or B cells and the degree of the liver damage and fibrosis have to be assessed for a subsequent evaluation of possible treatments to prevent, delay or abrogate the autodestructive process of the liver.
1. Intravenous injection of Ad-2D6 into ophthalmic sinus
Alternatively, the intravenous injection can be executed via the tail vein. However, injection via the ophthalmic sinus is much easier to perform and minimizes the loss of virus solution. It has been demonstrated that these two techniques can be used interchangeably and that both routes are equally effective 10.
Infected mice are monitored for adverse effects and obvious signs of suffering and pain, such as loss of appetite, loss of weight, loss of mobility, failure to groom, rough hair coat, hunched appearance, as well as licking, biting, scratching, or shaking a particular area (i.e. site of injection). Although the mice display a massive damage to the liver in the CYP2D6 model, the mice seem healthy and show no signs of suffering, pain or stress. Nevertheless, indicators of severe liver failure, such as serum aminotransferases are determined on a regular basis or are part of the performed experiments. Serum aminotransferase levels of >1000 U/l should only appear as a direct result of the virus infection, but not chronically. If the serum aminotransferase levels are chronically above 1000 U/l (severity limit) the mice are sacrificed.
2. Mouse eye bleeding
Alternatively, blood sampling can be done via the tail vein or via the superficial temporal vein11.
3. Mouse liver perfusion
Remark: This step is important to avoid a contamination of isolated liver lymphocytes by blood lymphocytes
4. Isolation of liver lymphocytes
5. Intracellular cytokine staining (ICCS)
5.1 Stimulation:
5.2. Staining:
6. Immunohistochemistry
6.1. H&E staining for the assessment of general liver pathology:
6.2. Immunohistochemistry for collagen I deposition:
7. Representative Results
Infection of mice with Ad-2D6 induced two distinct stages of liver damage. In the first days after infection, virus infection of the liver causes an acute increase of serum aminotransferase levels, an indicator of hepatocyte death6. This first, acute stage occurs independent of the expression of hCYP2D6 and can also be observed after infection with an empty control virus (Ad-Control) or a virus expressing green fluorescence protein (Ad-GFP). In contrast, the second stage is autoimmune-mediated and dependent on the expression of the triggering molecule hCYP2D6. This autoimmune stage occurs within 2-4 weeks post-infection and persists for several months6.
Figure 1. The CYP2D6 model. Wildtype FVB or C57BL/6 mice are injected with two doses of Ad-2D6 (intravenous and intraperitoneal). At a time of interest the liver and the serum are collected and assessed for liver damage and formation of hCYP2D6-specific antibodies and T cells.
The following features are characteristic for the persistent autoimmune hepatitis developing after Ad-2D6-infection of wildtype FVB or C57BL/6 mice in the CYP2D6 model: First, the liver shows an aberrantly changed morphology. In particular, the liver lobes are fused, hyperplasic nodules appear scattered over the entire liver and a massive capsular fibrosis is visible (figure 2).
Figure 2. Liver morphology. Typical Ad-2D6-induced liver damage as detected at week 4 post-infection. Note that the individual liver lobes are fused (arrowheads). Hyperplasic nodules appear scattered over the entire liver (arrows). Infection with a control adenovirus not expressing hCYP2D6 has no effect on the liver morphology.
Second, extensive and persistent cellular infiltrations appear in the peri-portal and parenchymal regions of the livers of Ad-2D6-infected but not Ad-Control-infected mice (figure 3A). Fibrosis predominantly develops in the subcapsular region with some collagen bundles protruding into the parenchyma (figure 3B).
Figure 3. Liver histology. A: H&E staining of liver tissue section of FVB mice infected with either Ad-Control or Ad-2D6 at week 4 post-infection. Note that significant infiltrations of mononuclear cells are only present in Ad-2D6 infected mice (single arrows). Triple arrows indicate larger cellular infiltrations in the liver parenchyma bridging between neighboring portal tracts. B: Immunohistochemical representation of liver fibrosis at week 4 after infection with Ad-2D6 or Ad-control. The liver sections have been stained for collagen using an anti-collagen I antibody. Note multiple layer of collagen disposition under the liver capsule (triple arrows) with some bundles protruding into the parenchyma (single arrows) and the presence of large clusters of infiltrating cells (arrowheads). Two representative liver sections are displayed for each condition. Size bars: 100μm.
A third feature is the generation of high titers of anti-CYP2D6 antibodies, which have a similar epitope specificity to human LKM-1 antibodies 6,13. Last, hCYP2D6-specific CD4 and CD8 T cells are generated that predominantly home to the liver. Such hCYP2D6-specific T cells can be detected by stimulation with immunodominant hCYP2D6 peptides and subsequent measurement of IFNγ expression by an intracellular cytokine staining (ICCS) (figure 4). hCYP2D6-specific CD8 T cells kill Ad-2D6-infected target cells in vivo 12.
Figure 4. hCYP2D6-specific T cells. Flow cytometric analysis of hCYP2D6-specific T cells. Liver lymphocytes have been isolated at week 4 after Ad-2D6-infection and were stimulated with either the immunodominant CD4 or CD8 hCYP2D6 epitope overnight. Generation of IFNγ was detected by intracellular cytokine staining and analyzed by flow cytometry using a FACS Canto II (BD Biosciences, Heidelberg, Germany). The percentage of hCYP2D6-specific CD4 and CD8 cells is indicated.
In previous studies, experimental hepatitis was often reported to be only transient and many current models for autoimmune liver disease depend on a rather complex disease induction protocols (for reviews see 3,5). For example, several models use transgenic mice expressing specific target antigens and adoptively transferred target antigen-specific, mostly TcR-transgenic, T cells14,15. Often an additional infection with livertropic viruses, bacteria or parasites is necessary to induce disease16-18. Alternatively, DNA-vaccination with plasmids encoding for target antigens, including CYP2D619, is used to induce hepatitis. However, an additional vaccination with plasmids encoding for pro-inflammatory cytokines, such as IL-12, is required20. Unfortunately, with a few exceptions15,20 hepatitis is only transient.
The CYP2D6 model6,21 uses a straightforward method to induce autoimmune hepatitis by simply infecting mice with an Adenovirus expressing hCYP2D6, the major autoantigen in AIH-27,8. Delivery of CYP2D6 by an adenovirus construct guarantees both a direct targeting of the liver as well as a local inflammation that promotes the breakdown of tolerance. It is important to note however that both the virus titer as well as the route of administration is crucial to this model. On the one hand, Ad-2D6 is a replication deficient virus, thus, a sufficient high titer of virus is required to generate a critical amount of antigen within the liver. On the other hand, too high a titer might result in a fatal acute liver failure. In addition, it has been demonstrated previously that infection of mice by high titers of adenovirus leads to a functional exhaustion of the immune response22. In our model, we used a combination of intravenous and intraperitoneal infection in order to obtain both chronic cellular infiltration as well as extensive fibrosis. We have observed that intravenous infection alone still causes massive cellular infiltration, but no fibrosis of the subcapsular region, indicating that peritoneal inflammation due to the intraperitoneal injection might be involved in the continuous subcapsular accumulation of collagen (Hintermann & Christen, manuscript in preparation). However, it is important to note that the observed hepatic fibrosis is antigen-specific, since we did not detect activation of hepatic stellate cells and collagen deposition after infection with equal virus-titers of Ad-GFP or Ad-Control23.
The CYP2D6 model reflects many aspects of human AIH1,2, such as chronic hepatitis characterized by persistent cellular infiltration and liver fibrosis6. In addition, the presence of high-titer anti-CYP2D6 antibodies with similar epitope specificity and spreading13 to LKM-1 antibodies found in AIH-2 patients indicates the presence of a common chronic autoimmune reactivity. CYP2D6 is the major antigen in AIH-2, but it is not recognized by patients diagnosed with the more frequent type 1 AIH. In addition, patients suffering from other liver diseases with autoimmune etiology, such as primary biliary cirrhosis (PBC) or primary sclerosing cholangitis (PSC), generate hallmark autoantibodies with specificity to distinct liver autoantigens and their livers show different pathological features centering on small (PBC) or large (PSC) bile ducts. Nevertheless, the CYP2D6 models offers the opportunity to analyze immunopathogenic mechanisms involved in chronic hepatic inflammatory processes as occurring during autoimmune liver diseases, to identify key players that drive the autoimmune destruction of hepatocytes and to evaluate possible therapeutic interventions to cure the disease.
The authors have nothing to disclose.
This work is supported by the Goethe University Hospital Frankfurt and a grant of the German Research Foundation to U.C.
Name of the equipment | Company | Catalogue number | Comments |
23G needle | BD Biosciences | 300800 | |
27½G needle | VWR | 612-0151 | |
30½ G needle | BD Biosciences | 304000 | |
Alanine aminotransferase test strips (GPT/ALT) | Roche Diagnostics | 10 745 138 202 | |
Aspartate aminotransferase test strips (GOT/AST) | Roche Diagnostics | 10 745 120 202 | |
Anaesthesia Unit Univentor 400 | AgnTho’s | ||
Base molds, disposable 37x24x10 mm | VWR | 720-0208 | |
Cell strainer 70mm | VWR | 734-0003 | |
Cryostat | Leica | CM1850 UV | |
Heparinized capillary tubes | Fisher | 3123987 | |
Microscope slides Superfrost Plus | Menzel Gläser | J1800AMNZ | |
Microtainer SST tubes | BD Biosciences | 365951 | |
Microtiter plates, V-bottom | VWR | 391-1924 | |
Reflotron Plus | Roche Diagnostics | Determiniation of serum AST / ALT | |
Name of the reagent | Company | Catalogue number | Comments |
Rabbit anti-mouse anti-Collagen type I IgG, | Chemicon | AB765P | |
Biotinylated goat anti-rabbit IgG | Vector (AXXORA) | VC-BA-1000-MC15 | |
FITC-conjugated anti mouse CD8a antibody | Southern Biotech | 1550-02 | |
PE-conjugated anti mouse IFNγ antibody | BD Biosciences | 554412 | |
PE-conjugated anti-mouse CD16/CD32 antibody (FcR block) | BD Biosciences | 553141 | |
Aquatex | VWR | 1.08562.0050 | |
Avidin/Biotin Blocking kit | Vector (AXXORA) | VC-SP-2001-KI01 | |
Brefeldin A | Sigma | B6542 | |
Collagenase IV | Sigma | C5138-100MG | |
DAB substrate kit 3’3’diaminobenzidine | Vector (AXXORA) | VC-SK-4100-KI01 | |
DNase | Sigma | DN-25 | |
Elite ABC Reagent | Vector (AXXORA) | VC-PK-7100-L050 | |
Eosin G/Y solution | Roth | X883.1 | |
Fetal calf serum (FCS) | Biochrom AG | S 0115 | |
Isofluran | Forene | B506 | |
Meyer’s Hematoxylin solution | AppliChem | A4840,1000 | |
OCT Compound | Sakura Finetek Europe | 4583 | |
PBS-buffered formaldehyde 10% | Roth | A146.3 | |
Percoll | Amersham | 17-0891-01 | |
Roti-Histokit | Roth | 6638.1 | |
RPMI | Invitrogen | 61870 | |
Saponin from quillaja bark | Sigma | S7900 |