An autoimmune biliary disease mouse model for primary biliary cirrhosis: Something for everyone


  • Potential conflict of interest: Nothing to report.


Primary biliary cirrhosis (PBC) is an autoimmune disease with a strong genetic component characterized by biliary ductular inflammation with eventual liver cirrhosis. The serologic hallmark of PBC is antimitochondrial antibodies that react with the pyruvate dehydrogenase complex, targeting the inner lipoyl domain of the E2 subunit (anti-PDC-E2). Herein we demonstrate that NOD.c3c4 mice congenically derived from the non-obese diabetic strain develop an autoimmune biliary disease (ABD) that models human PBC. NOD.c3c4 (at 9-10 wk, before significant biliary pathology) develop antibodies to PDC-E2 that are specific for the inner lipoyl domain. Affected areas of biliary epithelium are infiltrated with CD3+, CD4+, and CD8+ T cells, and treatment of NOD.c3c4 mice with monoclonal antibody to CD3 protects from ABD. Furthermore, NOD.c3c4-scid mice develop disease after adoptive transfer of splenocytes or CD4+ T cells, demonstrating a central role for T cells in pathogenesis. Histological analysis reveals destructive cholangitis, granuloma formation, and eosinophilic infiltration as seen in PBC, although, unlike PBC, the extrahepatic biliary ducts are also affected. Using a congenic mapping approach, we define the first ABD (Abd) locus, Abd1. These results identify the NOD.c3c4 mouse as the first spontaneous mouse model of PBC.

Irie J, Wu Y, Wicker LS, Rainbow D, Nalesnik MA, Hirsch R, Peterson LB, Leung PS, Cheng C, Mackay IR, Gershwin ME, Ridgway WM. NOD.c3c4 congenic mice develop autoimmune biliary disease that serologically and pathogenetically models human primary biliary cirrhosis. J Exp Med. 2006 May 15;203(5):1209-19. (Reprinted with permission from Rockefeller University Press.)


A new autoimmune biliary disease (ABD) mouse model has been serendipitously created by Bill Ridgway's group as a result of genetic manipulation of the non-obese diabetic (NOD) mouse strain.1, 2 The appearance of this model is timely as previous attempts to develop an animal model for PBC have been disappointing. Earlier studies employed transfer of peripheral blood mononuclear cells (PBMC) from patients into SCID mice but bile duct damage was observed in mice receiving PBMC from PBC patients and control subjects, likely as a result of graft-versus-host disease.3 Efforts to break tolerance to pyruvate dehydrogenase (PDC)-E2 protein did not fair any better. Autoantigen and adjuvant administration in mice and other animals merely produced AMA production without bile duct disease.4 Initial reports using the lupus prone SJL mice suggested this maneuver may actually work in the correct genetic background5 but follow up studies reported spontaneous bile duct disease in the SJL/J mice irrespective of antigenic challenge.6

In contrast, the NOD mice require no manipulation and develop spontaneous type 1 diabetes (T1D) associated with autoimmune responses to antigens associated with human T1D. The mice exhibit multiple immune defects that include functional abnormalities of T lymphocytes, NKT cells, and macrophages; deficiencies in regulatory CD4+CD25+ T cells and NK cells as well as the absence of C5a and hemolytic complement.7 One major genetic factor in the development of diabetes in NOD mice is homozygosity for MHC I-Ag7, which encodes a structurally unstable class II protein.8 The protein binds peptide poorly and it is thought that autoreactive T cells with high avidity to self peptides are preferentially activated in the thymus.8 The NOD mice provide a useful model for the study of T1D, as patients and mice tend to share homozygous for class II chain B proteins, have autoimmunity to the same islet derived antigens, possess a range of immune defects and propensity for the development of other autoimmune syndromes.

Extensive genetic susceptibility studies have identified specific loci for insulin dependant diabetes (idd) in NOD-mice and this knowledge has been applied to construct diabetes resistant strains using congenic segments from chromosomes 3 to 4 to replace specific idd loci. Bill Ridgway and his colleagues found that the NOD.c3c4 line are completely resistant to diabetes but develop ABD.2 In this model, component immunophenotypes interact to produce the autoimmune biliary disease. For example, partial substitution of chromosome 3 loci in NODc3 mice caused a mild histological biliary disease and partial substitution of chromosome 4 loci in NODc4 lead to generation of anti-nuclear antibodies. The combined substitutions in the NOD.c3c4 mouse produced full blown disease with hepatomegaly, biliary disease, liver failure in 50% of females and 25% of males within a year.1, 2

The NOD.c3c4 mouse and PBC patients share comparable clinical and autoimmune phenotypes. Histological appearance of biliary disease appears early on in all NOD.c3c4 mice characterized by a mononuclear infiltrate around small bile ducts with evidence of non-suppurative destructive cholangitis; portal granulomas were found in only 1 of 7 mice, however (Fig. 1). Furthermore, 55% of the NOD.c3c4 mice make AMA from an early age that react with the inner lipoyl domain of PDC-E2 and inhibit enzyme function. However, the mice develop additional biliary features that are not found in patients with PBC. For example, common bile duct dilatation is observed as early as 3 weeks of age and occurs without any reported evidence of biliary obstruction.2 Furthermore, up to 90% of mice develop progressive intra-hepatic biliary cyst formation accompanied by adjacent lymphocytic infiltration by the age of 30 weeks.1, 2 It is unknown why the mice develop proliferative disease in the intra and extraheptic bile ducts with lesions atypical of other human biliary disease, such as primary sclerosing cholangitis and Caroli's syndrome. Despite these anomalies, however, it is fair to say that the NOD.c3c4 model exhibits the most of the phenotypic features of PBC, with additional frills on the side. No doubt, unraveling the molecular process that promotes the dysregulated growth of the biliary epithelium in the NOD.c3c4 mouse will provide invaluable insight into the pathogenesis of PBC.

Figure 1.

Histological sections of (a, b, c) 20- and (d, e, f) 30-wk-old NOD.c3c4 mice livers showing lesions comparable to PBC. (a) Granulomatous lesion in a portal area, (b) chronic NSDC in portal area, and (c) peribiliary lymphoplasmacytic infiltration. Progression of the liver pathology in 30-wk-old NOD.c3c4 mice is shown by (d) pronounced biliary polycystic changes with infiltration of inflammatory cells, (e) fibrosis in the portal areas with eosinophilic infiltration (blue arrows), (f) hyperplasia of biliary epithelium with dilation of bile duct (black arrows) and degenerative biliary ductules with macrophage aggregates in the lumen (green arrows). All bars, 50, m. (Reproduced from The Journal of Experimental Medicine, 2006 May 15;203(5):1209-19, by copyright permission of Rockefeller University Press).

Another recent report suggests that T cell TGF-β receptor II dominant-negative mice can develop spontaneous periportal infiltrates and AMA reactive with PDC-E2, as well.9 Taken together, these 2 models provide additional support for the idea that activation of self-reactive lymphocytes has more to do with triggering autoimmunity than molecular mimicry. As compared to the NOD.c3c4 mouse, the clinical picture seen in the TGF-β receptor II dominant-negative mouse provides less of a parallel with PBC, in as much as the mice present with a wasting syndrome associated with inflammatory lung and bowel disease at 3 to 4 months with only minimal infiltration in the liver and kidneys.10 In fact, the pathology in the TGF-β receptor II dominant-negative mouse is more in keeping with the scurfy mouse model and human IPEX syndrome, where mutations in the FoxP3 gene result in dysregulation of peripheral tolerance as a result of diminished T regulatory (Treg) cell function.11 In contrast, the NOD mice are thought to lack mechanisms that regulate central tolerance.8 However, both models should provide considerable resources for investigating the generation of AMA in PBC.

Autoimmunity has been studied extensively in the NOD mouse for over 20 years directly implicating T lymphocytes in propagating and transferring diabetes. For example, anti-CD4+ abrogates diabetes and the transfer of CD4+ T cells induces disease in SCID mice.7, 12 Similar findings have been observed in the NOD.c3c4 mouse, where anti-CD3+ therapy prevents ABD and transfer of splenocytes or CD4+ into NOD.c3c4-SCID strain triggers disease. These studies directly implicate T lymphocytes in generating autoimmune disease but curiously, the mice exhibit an immunodeficient phenotype as well.7 With respect to a purely autoimmune etiology, manipulations of NOD mice have produced contradictory data as interventions used to stimulate immune response can paradoxically abrogate diabetes and attempts at reducing an immune response can lead to deterioration. For example, measures that boost NKT cells with α-galactosylceramide prevent the development of diabetes and NOD-mice bred on a CD1 (NKT receptor ligand) deficient background develop aggravated disease.7 These observations are not so surprising when you consider other diseases, such as HIV infection, where autoimmunity and immunodeficiency go hand in hand and autoimmunity improves with immune reconstitution.13

In the present study, the ability to transfer disease in leukocytes from the NOD.c3c4 mouse was only achieved in SCID mice with an intact abd 1 region on chromosome 4; this region will undoubtedly be scrutinized to determine disease modifier gene(s) associated with ABD.1 The abd 1 region contains immune response genes, autoimmune disease susceptibility genes, genes interacting with TGF-β as well as several endogenous retrovirus sequences of the murine leukemia virus (MuLV) and mouse mammary tumor virus (MMTV). It is notable that endogenous retroviral sequences in the germline can either be harmful or protective for the host. Infectious retroviruses can be transmitted vertically following Mendelian rules of inheritance and activation of replication competent virus can cause neoplastic and inflammatory disease, especially so in inbred strains of mice that have lost genetic mechanisms that prevent retroviral infection. Alternatively the expression of endogenous retroviral sequences may actually be protective. In this case, the expression of MTV superantigen in utero deletes specific T cell subsets required for viral replication, which in turn provides protection in the adult by preventing MMTV multiplying in lymphoid tissue.14

While the transfer studies implicate autoreactive lymphocytes in the development of biliary disease, should we be thinking of any microbial passengers in the lymphocytes that can cause disease in this immunodeficient mouse model as well? The role of bacterial infection and xenobiotics as environmental factors is probably irrelevant in the NOD.c3c4 mouse, as mice bred in germ free facility develop spontaneous ABD. However, the MMTV-related human betaretrovirus has been linked with PBC and the virus is associated with aberrant expression of PDC-E2 in vivo and in vitro.15 This observation brings up the question of whether endogenous MMTV is expressed in the NOD.c3c4 mouse, which is a frequent finding in many mouse models.14 If MMTV were linked with ABD, then abrogation of ABD with lymphocyte depleting antibodies and the passage of disease in lymphocytes to SCID mice would be equally consistent with passage of an endogenous retroviral infection in lymphocytes in this NOD.c3c4 mouse model.

The role of endogenous retroviral infection in inflammatory, degenerative and neoplastic disease is well recognized. For example, endogenous MMTV infection is central to the development of breast cancer and lymphoma in mice.14 Indeed, the SJL mice with spontaneous biliary disease5 are also known to develop lymphoma as a result of expressing endogenous MTV2916 NOD mice also express endogenous retroviruses.17 In the T1D model, MuLV is expressed in the pancreas at12 weeks at the height of inflammation in the pancreas.17, 18 The mice make antibodies to MuLV but mount a poor cytotoxic T lymphocyte response to virus. So, when faced with a viral infection that should provoke an antiviral immune response, the NOD mouse mounts an autoimmune response as a result of multiple immune defects. Similar observations have been found in other mouse models with organ specific expression of MuLV, such as the lupus prone NZB mouse that expresses MuLV in sites of inflammation such as the kidneys, lung and brain19 and the senescence accelerated SAMP mouse that develops dementia, liver, and inflammatory bowel disease.20 However, proving cause and effect is a different matter.

In summary, the NOD mouse model has served endocrinologists well in the study of genetics, immunology, and therapeutics of T1D for many years. The NOD.c3c4 mouse should fulfill the same role for PBC and lead immunologists, geneticists, and even virologists and lead to a better understanding of a poorly understood biliary disease.