The Journal of Immunology
Induction of Primary Biliary Cirrhosis in Guinea Pigs following Chemical Xenobiotic Immunization1 Patrick S. C. Leung,2* Ogyi Park,* Koichi Tsuneyama,‡ Mark J. Kurth,† Kit S. Lam,§ Aftab A. Ansari,¶ Ross L. Coppel,储 and M. Eric Gershwin* Although significant advances have been made in dissecting the effector mechanisms in autoimmunity, the major stumbling block remains defining the etiological events that precede disease. Primary biliary cirrhosis (PBC) illustrates this paradigm because of its high degree of heritability, its female predominance, and its extraordinarily specific and defined immune response and target destruction. In PBC, the major autoantigens belong to E2 components of the 2-oxo-acid dehydrogenase family of mitochondrially located enzymes that share a lipoylated peptide sequence that is the immunodominant target. Our previous work has demonstrated that synthetic mimics of the lipoate molecule such as 6-bromohexoanate demonstrate a high degree of reactivity with PBC sera prompted us to immunize groups of guinea pigs with 6-bromohexoanate conjugated to BSA. In this study, we provide serologic and immunohistochemical evidence that such immunized guinea pigs not only develop antimitochondrial autoantibody responses similar to human PBC, but also develop autoimmune cholangitis after 18 mo. Xenobiotic-immunized guinea pigs are the first induced model of PBC and suggest an etiology that has implications for the causation of other human autoimmune diseases. The data also reflect the likelihood that, in PBC, the multilineage antimitochondrial response is a pathogenic mechanism and that loss of tolerance and subsequent development of biliary lesions depends on either modification of the host mitochondrial Ag or a similar breakdown due to molecular mimicry. The Journal of Immunology, 2007, 179: 2651–2657.
T
he most problematic issue in autoimmunity is a detailed understanding of etiologic events that result in immunopathology. At present, etiology is conceived only in general terms of genetic and environmental factors that predispose to a breakdown in tolerance characterized by autoimmune responses of great specificity. Primary biliary cirrhosis (PBC)3 illustrates this enigma because of its high degree of concordance in identical twins, clinical uniformity, and the presence of a highly directed, specific serologic marker consisting of antimitochondrial Abs (AMA) directed at the E2 subunits of the 2-oxo-acid dehydroge-
*Division of Rheumatology, Allergy, and Clinical Immunology, University of California at Davis School of Medicine, and †Department of Chemistry, University of California, Davis, CA 95616; ‡Department of Pathology and 21st Century Center of Excellence Program, Toyama Medical and Pharmaceutical University, Toyama, Japan; §University of California at Davis Cancer Center, Division of Hematology and Oncology, University of California at Davis, Sacramento, CA 95817; ¶Department of Pathology, Emory University School of Medicine, Atlanta, GA 30322; and 储Department of Microbiology, Monash University, Clayton, Victoria, Australia Received for publication February 1, 2007. Accepted for publication June 8, 2007. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported in part by National Institutes of Health Grants DK39588 and DK037003. 2 Address correspondence and reprint requests to Dr. Patrick S. C. Leung, Division of Rheumatology, Allergy, and Clinical Immunology, University of California at Davis School of Medicine, 451 East Health Sciences Drive, Davis, CA 95616. E-mail address:
[email protected] 3 Abbreviations used in this paper: PBC, primary biliary cirrhosis; AMA, antimitochondrial Ab; 2-OADC-E2, E2 subunit of 2-oxo-acid dehydrogenase complex; PDCE2, E2 subunit of pyruvate dehydrogenase complex; 6BH-BSA, 6-bromohexanoate conjugated to BSA; BCOADC-E2, E2 subunit of branched chain 2-oxo-acid dehydrogenase complex; OGDC-E2, E2 subunit of 2-oxo-glutarate dehydrogenase complex; CLA-BSA, chloroacetate-conjugated BSA; CNSDC, chronic nonsuppurative destructive cholangitis; BEC, bile duct epithelial cell.
Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 www.jimmunol.org
nase complexes (2-OADC-E2) (1, 2). This signature of PBC, the AMA, is a useful probe to enable dissection of potential environmental triggers and has led to the proposition that the inciting event of PBC is a modification of the major autoantigen, the E2 subunit of pyruvate dehydrogenase (PDC-E2), by a xenobiotic via chemical conjugation. This theory is based on our demonstration that, when the lipoyl domain of PDC-E2 was modified with specific small molecule mimics, the ensuing structures displayed highly specific reactivity to PBC sera, at levels often higher than the native molecule (3, 4). Interestingly, the autoimmune response in terms of not only the CD4⫹ and CD8⫹ epitopes but also the autoantibodies in PBC all map within the lipoyl domain of the 2-OADC-E2 enzymes (5–7). We hypothesized that the restricted epitope recognition of the highly conserved lipoyl domain in PBC is initiated by the modification of the lipoid molecule by a chemical xenobiotic mimic. This structurally altered native protein is recognized as foreign by the immune system and primes the immune system to subsequently recognize self-proteins and induces loss of tolerance. To identify lipoyl molecular mimics that are recognized by AMA, a 12-residue synthetic peptide (173–184 of the inner lipoyl domain of PDC-E2) was coupled with a series of related synthetic lipoic acid mimics. Screening of these compounds for reactivity with AMA-positive PBC sera led to the identification of a number of organic chemicals that were recognized by AMA with higher relative affinity than that for parent lipoylated-PDC-E2 (3, 4, 8). Furthermore, immunization of rabbits with one of these xenobiotics, 6-bromohexanoate, conjugated to BSA (6BH-BSA) led to the induction of AMA with the same antigenic specificities and capacity to inhibit the activity of PDC enzymes that is displayed by sera from patients with PBC (9). These data suggest that the immune process in PBC can be triggered by exposure to PDC-E2 that had been modified by xenobiotics. However, the autoimmune response of xenobiotic-immunized rabbits
2652
INDUCTION OF PBC
was limited to the generation of AMA without any recognizable PBC-like liver pathology. We concluded that additional factors beyond generation of AMA were required for inducing a true PBC-like disease. We hypothesized that a particular biliary architecture and physiology may contribute to the pathologic process and therefore chose to study guinea pigs that are similar to humans for these characteristics (10). We report in this study the successful induction of AMA and autoimmune cholangitis and suggest that these data will provide a conceptual framework that will illuminate not only the etiology of PBC but also other autoimmune diseases.
Materials and Methods Choice of xenobiotics Our laboratory has previously examined whether the xenobiotic with or without the PDC-E2 peptide backbone, could induce autoimmunity in experimental animals. A pyramid scheme consisting of immunizing rabbits with a progressively decreasing number of mixtures of the reactive xenobiotic was used to identify the optimal structure that would lead to breaking of immune tolerance to the native OADC mitochondrial proteins in vivo. This work led to the identification of a BSA conjugate of 6BH (6BH-BSA) as producing a xenobiotic agent capable of inducing of AMA (9). The chemical structure of 6BH and its hypothetical alignment with the DKA amino acid residues of the PDC-E2 lipoyl domain is shown in Fig. 1.
Immunization of guinea pigs Female guinea pigs (n ⫽ 10) at 20 wk of age were s.c. immunized with 30 g of 6BH-BSA incorporated in CFA, and then boosted every 2 wk with 6BH-BSA incorporated in IFA throughout. Guinea pigs (n ⫽ 10) similarly immunized and boosted with BSA were included as controls. Sera were collected beginning 4 wk after the initial immunization and every 4 wk thereafter for analysis of AMA reactivity by ELISA against recombinant proteins of PDC-E2, the E2 subunits of branched chain 2-oxo-acid dehydrogenase complex (BCOADC-E2), the E2 subunits of 2-oxo-glutarate dehydrogenase complex (OGDC-E2), and shrimp tropomyosin as a negative control recombinant protein. Beginning 9 mo postimmunization, IL-1 fragment (Sigma-Aldrich) (10 g/animal) was i.p. administered to animals every 2 wk for 12 mo. Two animals were sacrificed at 12 mo postimmunization. Beginning at 18 mo postimmunization, two animals were sacrificed every 2 mo. At 24 mo postimmunization, all animals were sacrificed. Blood samples and tissues were collected immediately after animals were sacrificed. Animal protocols were approved by the Institutional Review Board at the University of California at Davis.
Detection of AMA AMA reactivity against recombinant proteins of PDC-E2, BCOADC-E2, and OGDC-E2 was analyzed by ELISA as described (11). Sera from AMA-positive (n ⫽ 5) and AMA-negative PBC patients (n ⫽ 5) were included throughout as controls (11, 12).
Cloning and expression of the guinea pig PDC-E2 lipoyl domain Guinea pig cDNA encoding the outer and inner PDC-E2 lipoyl domains was cloned and expressed in plasmid pRSET. Briefly, the guinea pig PDC-E2 was amplified from guinea pig liver cDNA by PCR. Because no guinea pig PDC-E2 nucleotide sequence was available from GenBank, the forward PCR primer (5⬘-TACAATGCAGGCAGGCACC-3⬘) and reverse PCR primer (5⬘-TCCTGCTGGACCAAAGGGAAGGGTG-3⬘) were designed by aligning human, mouse, rat, and pig PDC-E2 sequences. Nucleotide sequencing of PCR products were determined by dye-terminator reactions on an ABI 3730 capillary sequencer (Applied Biosystems). A cDNA sequence that corresponds to the guinea pig PDC-E2 lipoyl domain was subcloned into the EcoRI site of a pRSET expression vector. Plasmid DNA containing the correctly oriented guinea pig PDC-E2 in pRSET was transformed into Escherichia coli BL21 (DE3) pLysS competent cells. A single colony of the transformant was chosen, grown in LB broth containing ampicillin (50 g/ml) overnight at 37°C, and induced with isopropyl--D-thiogalactopyranoside at a final concentration of 1 mM. Guinea pig rPDC-E2 protein was purified using Ni column (Invitrogen Life Technologies).
Immunoreactivity against guinea pig PDC-E2 lipoyl domain Twenty micrograms of guinea pig rPDC-E2 was resolved by preparative SDS-PAGE, transferred onto nitrocellulose membrane, which were cut into
FIGURE 1. A and B, Structure of lipoic acid attached to the lysine residue of PDC-E2 inner lipoyl domain (A) and 6BH attached to the lysine residue of BSA (B).
3-mm strips, blocked with 3% nonfat dry milk in PBS (pH 7.4), and probed with sera from 6BH-BSA- and BSA-immunized guinea pigs (1/250 dilution) for 1 h. After three 5-min washes with PBS containing 0.05% Tween 20 (PBST), the membranes were incubated with HRP-conjugated antiguinea pig IgG, washed with PBST, and developed by chemiluminescence. Anti-lipoic acid Ab and PBC sera were used throughout as controls. A recombinant protein prepared in the same vector and expressing the shrimp tropomyosin was analyzed in parallel as a negative control (13).
Ag specificity of 6BH-BSA-immunized guinea pig sera Sera from 6BH-BSA-immunized guinea pigs (1/200 dilution) were separately absorbed overnight with 100 g/ml BSA, 6BH-BSA, rPDC-E2, rBCOADC-E2, chloroacetate-conjugated BSA (CLA-BSA) as an irrelevant xenobiotic control, and shrimp tropomyosin as an irrelevant protein control. Absorbed sera were then analyzed for reactivities against the above panel of Ags by ELISA. Percent reactivity was calculated as follows: 1 ⫺ (OD of absorbed sera/OD of unabsorbed sera) ⫻ 100.
PDC and citrate synthase enzyme activity inhibition assay A predetermined optimal dilution of sera at 8 mo postimmunization was incubated with PDC (Sigma-Aldrich) for 10 min at room temperature and added to a PDC reaction mixture containing 5 mM sodium pyruvate, 2.5 mM NAD⫹, 0.2 mM thiamine pyrophosphate, 0.1 mM coenzyme A, 0.3 mM DTT, 1 mM magnesium chloride, and 50 mM potassium phosphate buffer (pH 8.0). The change in absorbance per minute at 340 nm was monitored for 5 min. Enzyme activity of sera obtained before immunization was determined in parallel, and the values were defined as 100% activity. An irrelevant enzyme, citrate synthase, was analyzed in parallel as control. The reaction mixture contained 0.2 mM acetyl-CoA, 0.5 mM oxaloacetate (Sigma-Aldrich), and 0.1 mM 5,5⬘-dithiobis-(2-nitrobenzoate) in 100 mM Tris-HCl (pH 7.5) (14). The reaction was followed spectrophotometrically at 412 nm over 3 min. The inhibition of PDC and citrate synthase enzyme activity by sera from BSA-immunized guinea pigs and from PBC patients were analyzed in parallel.
Analysis of anti-lipoic acid reactivity Sera from 6BH-BSA-immunized guinea pigs at 0, 6, 12, and 24 mo postimmunization were analyzed for anti-lipoic acid reactivity by a microarray assay as described (4, 15). Briefly, lipoic acid-peptide-agarose were spotted onto glass slides (Mercedes Medical) using the Affymetrix 417 Microarrayer (Affymetrix). Each sample was spotted in triplicate. Microarrays were blocked with 3% nonfat dry milk in PBS buffer for 1 h at room temperature, and thereafter incubated with diluted Ab samples (guinea pig, 1/250; murine anti-PDC mAb, 1/1) in 1 ml of blocking buffer (3% nonfat dry milk in PBST) for 1 h at room temperature. After thorough washes with PBST, 1 ml of the Cy3-conjugated secondary Ab (1 g/ml) (Zymed Laboratories) in blocking buffer was added to each slide and incubated at room temperature for 30 min. Subsequently, slides were washed in PBST for 10 min and in water for 15 s. Abs to 6BH and an unrelated xenobiotic compound 11 ((2,2,2-trifluoroethoxy)acetic acid) were analyzed in parallel. Arrays were then dried and scanned using the Affymetrix 428 Array Scanner. Data analysis was performed using the ImageQuant software (Molecular Dynamics). Statistical analysis was performed using JMP software (SAS Institute). A paired Student’s t test was performed to compare differences of the signal intensity between sera obtained before and after immunization. A value of p ⬍ 0.05 is considered significant.
The Journal of Immunology
2653 Liver sections were deparaffinized, stained with H&E, and examined under light microscopy. Control tissues (salivary gland, thyroid, and parathyroid) were examined in parallel. To stain for T cells, deparaffinized liver sections from 6BH-BSA- and BSA-immunized guinea pigs were rehydrated, blocked with 3% normal horse serum in TBS at room temperature for 20 min, and then incubated for 1 h with a 1/100 dilution of mouse anti-guinea pig T cell (PAN) Ab (CT5 clone; Serotec). The slides were then washed in TBS for 5 min and thereafter incubated with biotinylated anti-mouse IgG for 30 min. After washing with TBS, the slides were incubated with Vectastain ABC (avidin/biotin complex) solution (Vector Laboratories) for 30 min, washed, and then developed with the alkaline phosphatase substrate for 20 min. The slides were washed in tap water for 5 min, mounted, and examined by light microscopy (Carl Zeiss MicroImaging).
Results Serological response of guinea pigs following 6BH-BSA immunization
FIGURE 2. A, Serial development of antimitochondrial autoantibodies in 6BH-BSA-immunized guinea pigs. Reactivity of 6BH-BSA-immunized guinea pig sera, before immunization and at serial time intervals up to 24 mo; sera were diluted 1/200 and analyzed by ELISA for AMA reactivity against recombinant proteins of PDC-E2, BCOADC-E2, OGDC-E2, and a negative control Ag. In 6BH-BSA-immunized guinea pigs, AMA to PDCE2, BCOADC-E2, and OGDC-E2 were detected as early as 4 wk after initial immunization; by 12 wk postimmunization, 100% of guinea pigs were PDC-E2 positive. All assays were performed in triplicate. B, Serologic reactivity of 6BH-BSA-immunized guinea pigs against self-PDC-E2. Sera from 6BH-BSA-immunized guinea pigs (lane A), from BSA-immunized guinea pigs (lane B), from a patient with PBC (lane C), from rabbit with anti-lipoic acid Abs (lane D), and from an irrelevant protein immunized guinea pig (lane E) were analyzed for reactivity against guinea pig rPDC-E2 lipoyl domain by immunoblotting; equal amounts of Ag were loaded on each lane. Positive reactivity was detected in lanes A, C, and D. An irrelevant control Ag (shrimp tropomyosin) prepared in the same recombinant construct was not reactive with sera from 6BH-BSA-immunized guinea pigs (lane F) and was also nonreactive using sera from a patient with PBC (lane G) but was reactive with a known positive control, sera from a subject with known Abs to shrimp tropomyosin (lane H).
Liver histopathology At 18 mo postimmunization and every 2 mo thereafter, liver tissues from guinea pigs immunized with 6BH-BSA or BSA were immediately fixed in 10% buffered formalin, embedded in paraffin, and cut into 5-m sections.
6BH-BSA-immunized guinea pigs produced Abs to PDC-E2, BCOADC-E2, and OGDC-E2 as early as 4 wk after the initial immunization, and 100% of animals were AMA positive by 12 wk (Fig. 2A). AMAs are typically detected using on various cell lines using fluorescence but induced Abs may react with mitochondria of some species but not of the source animal. To investigate whether 6BH-BSA-immunized guinea pigs produced autoreactive AMA, we first cloned and expressed the guinea pig PDC-E2 inner lipoyl domain in plasmid pRSET. Sera from 6BH-BSA-immunized guinea pigs react to guinea pig PDC-E2 (Fig. 2B). Furthermore, sera from 6BH-BSA and BSA-immunized guinea pigs inhibited PDC enzyme activity at 20.5 ⫾ 1.55 and 5.0 ⫾ 4.28%, respectively ( p ⬍ 0.05). The effect of 6BH-BSA- and BSAimmunized guinea pig sera on an irrelevant enzyme citrate synthase activity was analyzed in parallel. Sera from 6BH-BSAand BSA-immunized guinea pigs failed to inhibit citrate enzyme activity giving values of 1.17 ⫾ 0.05 and 2.74 ⫾ 0.02%, respectively ( p, nonsignificant), denoting the specificity of PDC enzyme inhibition. To characterize the antigenic specificity of the Abs present in the sera of 6BH-BSA-immunized guinea pigs, aliquots of sera were preabsorbed with a panel of reagents, including BSA, 6BHBSA, CLA-BSA, rPDC-E2, rBCOADC-E2, and tropomyosin, an irrelevant protein control. When sera from 6BH-BSA-immunized guinea pigs were absorbed with 6BH-BSA, reactivity to rPDC-E2 and rBCOADC-E2, was markedly decreased. Unrelated control Ags and an irrelevant xenobiotic CLA-BSA failed to absorb such reactivity (Table I). Interestingly, when 6BH-BSA-immunized sera were absorbed with recombinant mitochondrial Ags, their reactivity to 6BH-BSA remained virtually unchanged. One hundred percent of 6BH-BSA-immunized guinea pigs, but not BSA-immunized control, produced anti-lipoic acid Abs beginning at 6 mo
Table I. Percent reactivity following absorption of sera from 6BH-BSA-immunized guinea pigsa Absorbent Antigen
Unabsorbed
BSA
6BH-BSA
CLA-BSA
PDC-E2
BCOADC-E2
Control
BSA 6BH-BSA CLA-BSA PDC-E2 BCOADC-E2
100 100 100 100 100
59.5* 99.5 99.1 96.4 96.2
37.9 18.7* 71.9 46.8** 49.4**
85.3 97.2 31.5** 98.4 99
90.1 86.6 98.5 20.9** 35.2
86.3 97.5 98.6 40.1* 34.4
98.7 98.3 99.4 98.2 98.4
a Expressed as remaining reactivity (%) determined by ELISA. SEM of all assays were ⬍7.5%. ⴱ, p ⬍ 0.005 between unabsorbed and absorbed. ⴱⴱ, p ⬍ 0.05 between unabsorbed and absorbed.
2654
INDUCTION OF PBC
FIGURE 3. 6BH-BSA-immunized guinea pigs developed Abs to lipoic acid. Sera from 6BH-BSA-immunized guinea pigs were assayed for Ig reactivity against 6BH-BSA, lipoic acid (LA), and an irrelevant compound (11) by a microarray assay. Reactivity to 6BHBSA and lipoic acid were detected at 6, 12, and 24 mo postimmunization. p ⬍ 0.001 in 6BH-BSA-immunized sera against 6BH-BSA and LA between before immunization and at 6, 12, and 24 mo postimmunization.
postimmunization and remained seropositive to lipoic acid at 24 mo postimmunization (Fig. 3). 6BH-BSA-immunized guinea pigs exhibited PBC-like liver lesions Liver sections of 6H-BSA- and BSA-immunized guinea pigs did not show any significant PBC-like liver lesions until 18 mo postimmunization. After this age, 6BH-BSA-immunized guinea pigs demonstrated significant lymphoid cell infiltrates surrounding damaged bile ducts, which is characteristic of the liver histopathology seen in PBC (Fig. 4, A–D). Mild to moderate infiltration of lymphoplasmacytes and vacuolated histiocytes was frequently observed in portal areas. In affected portal tracts, interlobular bile
FIGURE 4. A–F, Histological analysis of from 6BH-BSA-immunized guinea pig liver specimen (A–E) and control (F). Liver sections were stained with H&E and examined under microscope. Magnification: A, ⫻100; B–F, ⫻400. A, A portal tract showing lymphoplasmacytes infiltration (black arrow) and neighboring portal tract without significant inflammation (white arrow). No significant change was observed in perivenular area (CV, central vein). B and C, Interlobular bile duct (arrow) surrounded by lymphoplasmacytes showed cellular damage in affected portal tracts. D, Bile duct loss in several portal tracts. E, Absence of inflammatory change in the large portal tracts containing septal bile ducts or large bile ducts. F, Liver specimens from the control animals did not show any inflammatory infiltration or bile duct destruction.
ducts were surrounded by variably swollen lymphoplasmacytes with a vacuolated cytoplasm and an irregular luminal border, or showed an eosinophilic shrunken appearance with pyknotic nuclei (Fig. 4, B and C). These bile duct lesions are quite similar to chronic nonsuppurative destructive cholangitis (CNSDC) of PBC. Portal inflammation was distributed heterogeneously within the liver (Fig. 4A). In some portal areas, interlobular bile ducts were not clearly observed (Fig. 4D) and confirmed by the analysis of serial sections reflecting bile duct loss. In contrast to the pathological changes of interlobular bile ducts, septal to large bile ducts did not show any significant pathological changes (Fig. 4E). Mild piecemeal necrosis (interface hepatitis) was observed in the affected portal tracts. Spotty to focal necrosis, sinusoidal lymphoid infiltration, and Kupffer cell hyperplasia was scattered. Other specific findings such as perivenular zonal necrosis with fibrosis, steatosis of the hepatocytes, and cholestasis were not observed in these animals up to 24 mo postimmunization. In liver tissues from the control animals, portal inflammation was minimal and bile duct damage was not detectable. Moreover, parenchymal histopathological changes were not observed except for occasional spotty necrosis (Fig. 4F). Control tissues from 6BH-BSA-immunized guinea pigs did not reveal lesions (Fig. 5). Whereas immunohistochemical analysis of liver sections using anti-guinea pig T cell
FIGURE 5. A–C, Histologic analysis of 6BH-BSA-immunized guinea pigs, including thyroid (A), parathyroid (B), and salivary gland (C) tissue sections from 6BH-BSA-immunized guinea pigs. Tissues were stained with H&E and examined under microscope (magnification, ⫻200). These tissues did not reveal inflammatory infiltration.
The Journal of Immunology
FIGURE 6. T cell infiltrates in the portal tracts of 6BH-BSA-immunized guinea pigs. A and B, Liver specimens from 6BH-BSA-immunized guinea pig (A) and the control (B) were immunostained by anti-guinea pig T cell (PAN) Ab using ABC-alkaline phosphatase method. Marked T cell infiltration was observed in the portal tract of A, whereas no T cell infiltration was observed in the portal tracts of B. Magnification, ⫻100.
(PAN) Ab demonstrated T cell infiltrates in portal areas of 6BHBSA-immunized guinea pigs, only rare T cells were seen in controls (Fig. 6).
Discussion In this study, we demonstrate that guinea pigs immunized with 6BH-BSA serve as a model that possesses many of the clinical parameters that are characteristic of PBC. 6BH-BSA-immunized guinea pigs, which are normally tolerant to their own 2-OADC, develop autoantibodies against PDC-E2, BCOADC-E2, and OGDC-E2. The serological AMA profile is similar to that of human PBC, with PDC-E2 as the immunodominant Ag (2). Of particular interest, 6BH-BSA-immunized guinea pigs react to selfPDC-E2, which clearly demonstrates that self-tolerance was broken by immunization with a xenobiotic coupled to a protein that lacks the PDC-E2 peptide backbone. Similar to human PBC (16, 17), guinea pig anti-PDC-E2 recognizes the PDC-E2 lipoyl domain (Fig. 2). Furthermore, the selective absorption of antiPDC-E2 reactivity by PDC-E2 or BCOADC-E2 and of antiBCOADC-E2 reactivity by PDC-E2 or BCOADC-E2 implies that there is an important hierarchy in the cross-reactivity of the immune response reminiscent of the concept of determinant spreading. The difference in the degree of absorption of BSA reactivity by 6BH-BSA and CLA-BSA in our data is most likely due to differences of structural changes in BSA upon modification by each of these compounds. Thus, the induction of AMA reactivity in 6BH-BSA-immunized guinea pigs is primarily influenced by the hapten. Interestingly, 6BH-BSA-immunized guinea pigs also developed anti-lipoic acid Abs, as observed in PBC patients (Fig. 3). Although the difference in inhibition of PDC enzyme function by 6BH-BSA-immunized guinea pig sera is considerably lower than that reported in human PBC sera (17), it does not underscore the specific inhibition of PDC enzyme function by 6BH-BSA-immunized guinea pig sera. The difference in PDC enzyme function inhibition between the 6BH-immunized guinea pig and human PBC sera is likely accountable by the differences in the epitopes and affinities of the guinea pig AMA and the human AMA to the PDC functional sites. It should also be noted that sera from 6BHBSA-immunized guinea pigs specifically inhibited the enzyme activity of PDC but not the activity of an irrelevant control enzyme citrate synthase. Although the precise physiological effect and mechanisms of AMA in the in vivo function of pyruvate dehydrogenase is unclear, it has been postulated that AMA can induce apoptosis of bile duct epithelial cells (BEC) through caspase activation (18). Interestingly, unlike other cell types in which autoantibody recognition of PDC-E2 was abrogated after apoptosis, the antigenicity of PDC-E2 persisted in apoptotic BEC (19). IL-1 in-
2655 creases the efficacy of Ag processing and presentation (20). Previous work has demonstrated that, in mice, coinjection of IL-1 with Coxsackie virus induced a more rigorous local inflammation and autoimmune myocarditis than virus alone (21). Thus, we envisioned that the IL-1 would enhance focal liver lesions in 6BHBSA-immunized guinea pigs. In early-stage PBC, the liver histopathology is characterized by infiltration of lymphocyte and eosinophils as well as destruction of interlobular bile ducts in the involved portal tracts. The distribution of the affected portal areas is patchy at this stage and located heterogeneously. Interlobular bile duct damage called CNSDC and bile duct loss are also observed. In late-stage PBC, the liver pathology is characterized by progressive bile duct damage, chronic cholestasis, extending fibrosis, and eventually cirrhosis of the liver. In this study, hepatic lesions such as lymphoplasmacytic inflammation, interlobular bile duct damages resembling to CNSDC, and occasional bile duct loss were seen in portal areas of 6BH-BSAimmunized guinea pigs (Fig. 4). The lesions were specific to small bile ducts because the large bile ducts and control tissues from 6BH-BSA-immunized guinea pigs were not affected (Figs. 4 and 5). We hypothesize that PBC may be induced by a chemical xenobiotic, which has the potential to form complexes with, or otherwise alter, self-proteins such that they become immunogenic (22). The loss of tolerance would involve the spreading of the immune responses from the modified protein to the unmodified native protein (23). The chronic presence of the self-protein would then serve to perpetuate the immune response initiated by the xenobiotically modified self-protein and lead to autoimmunity. In support of this hypothesis, patients exposed to halothane, whose trifluoroacetyl metabolite covalently links to lysine on cytochrome p450 2E1, develop anti-trifluoroacetyl Abs that are also cross-reactive to lipoylated PDC-E2 (24). In addition, we have identified a number of lipoic acid mimeotopes that are recognized by AMA-positive PBC sera (3, 4, 8). Furthermore, we have previously demonstrated that rabbits immunized with a xenobiotic-hapten, 6BH-BSA, produced self-AMA that had all the characteristics of human AMA in that they recognized PDC-E2, BCOADC-E2, and OGDC-E2 and inhibited PDC enzyme activity (9). Similarly, s.c. immunization of SJL/J mice with biotinylated murine PDC produced Ab and T cell reactivities to unmodified and modified murine PDC (25). Although mild inflammation in the portal tract was seen in immunized SJL/J mice, both the rabbit and the SJL mice model fall short of any evidence of the bile duct injury, cholestasis, or fibrosis that are characteristic of PBC livers (9). Therefore, it is likely that other factors such as architecture of the biliary system, inflammatory signals, and detoxification pathways are important in the development of PBC. It is important to note that IL-1 was administered to the guinea pigs every 2 wk beginning at 9 mo postimmunization. Immunization of guinea pigs with 6BH-BSA was not only able to reproduce the AMA response already seen in 6BH-BSA-immunized rabbits, but was also associated with histological evidence of T cell-mediated biliary damage (Fig. 5). Although previous studies have focused on unmodified peptides of the major autoantigens as targets of autoimmune CD4⫹ and CD8⫹ T cell responses in human PBC, our current data imply that xenobiotics, in particular lipoate mimics, serve a key role in the breakdown of self-tolerance to mitochondrial Ags in human PBC. Although the role of AMA in the immunopathology of PBC, or biliary destruction, is still not clear, the specificity of pathological changes localized to the bile ducts, the presence of lymphoid infiltration in the portal tracts, and the readily detectable expression of MHC Ags on the biliary epithelium suggest that autoantigenspecific T cell responses are directed against the BEC. T cells are
2656 known to recognize not only peptides but also haptens covalently attached to MHC class I- or class II-restricted peptides (26). Typical examples of protein-modifying haptens are chemicals such as trinitrochlorobenzene or trinitrobenzene sulfonic acid (27), drugs like penicillin (28), metal ions (29), or natural compounds like urushiol in poison ivy (30). Such reagents have raised particular attention as potent inducers of immune responses (31). Studies in several model systems have suggested that haptens can serve as part of the antigenic structure recognized by TCRs (32). Of particular interest, it has been shown in a mouse model that haptenspecific CTLs exhibit recognition specificities for not only the hapten but also the carrier peptides. In almost every case, this peptide reactivity was strong enough to also allow for lysis of target cells pulsed with homologous, unmodified peptides (33, 34). Based on these observations, it has been proposed that hapten modification can lead to the triggering of T cells that, once activated, would react to the unmodified carrier peptides, resulting in autoimmune T cell responses (32). PBC is characterized by T cell-mediated destruction of BEC, and portal tract inflammatory responses are central pathological events in PBC. CD4⫹ and CD8⫹ T cells reactive with PDC are present in the peripheral blood and liver of PBC (5, 7). Although data on PDC-specific autoreactive guinea pig T cells are not available, the close proximity of T cells to the damaged bile ducts suggests the possible involvement of self-PDC-specific T cells in these hepatic lesions. In 6BH-BSA-immunized guinea pigs, the presence of liver-specific histological lesions, the presence of lymphoid cell infiltrations surrounding damaged portal tracts together with mitochondrial autoantigen-specific AMA, implicates the involvement of cell-mediated immune responses in the liver. Future studies using reagents that demonstrate specificity for guinea pig CD4⫹ and CD8⫹ T cells will be designed to address the important role of T cell-mediated immune responses in this model (35, 36). However, we need to emphasize several limitations in the study of T cell responses and guinea pigs. First, these are outbred guinea pigs, making adoptive cell transfer experiments impossible. Second, there are considerably less reagents available in guinea pigs than in either humans or mice. This is best exemplified by the relative lack of CD4⫹- and CD8⫹-specific guinea pig reagents and their inability for utility in immunohistochemistry (37, 38). These deficiencies, notwithstanding, should not diminish the induction of biliary pathology in guinea pigs. There are very few examples of human autoimmune diseases associated with putative environmental inducers. Hence, our future directions must include the development of appropriate reagents to study the contribution of CD4⫹ and CD8⫹ T cell responses and the specific recognition of 6BH and PDC-E2 and their relative cross-reactivity. We also believe that future studies must also identify the mechanisms required for the long latency time between exposure and histologic changes. Previous attempts in generating an animal model of PBC have revealed that immunization by mitochondrial Ag alone or modified form of PDC is insufficient to generate both AMA production and PBC-like liver lesions (25, 39 – 44). The breaking of tolerance to mitochondrial autoantigens can be accomplished by immunization of xenobiotic small molecules that are not of mitochondrial origin (9). Although it appears that AMA can be induced by either self or foreign 2-OADC as well as xenobiotic agents, the successful generation of PBC liver pathology may require a combination of physiological, genetic, and environmental factors. The successful generation of AMA and PBC-like liver histopathology in 6BH-BSA-immunized guinea pigs is a powerful tool to investigate the mechanisms involved in the inductive phase of PBC. It suggests that the study of other organ-specific autoim-
INDUCTION OF PBC mune diseases may be advanced by the search for potential xenobiotic adducts on self-proteins.
Acknowledgments We thank Dr. Chih-Te Wu in the cloning of the guinea pig PDC-E2 and Dr. Andrew Han in the citrate synthase enzyme activity assay.
Disclosures The authors have no financial conflict of interest.
References 1. Selmi, C., M. J. Mayo, N. Bach, H. Ishibashi, P. Invernizzi, R. G. Gish, S. C. Gordon, H. I. Wright, B. Zweiban, M. Podda, and M. E. Gershwin. 2004. Primary biliary cirrhosis in monozygotic and dizygotic twins: genetics, epigenetics, and environment. Gastroenterology 127: 485– 492. 2. Gershwin, M. E., A. A. Ansari, I. R. Mackay, Y. Nakanuma, A. Nishio, M. J. Rowley, and R. L. Coppel. 2000. Primary biliary cirrhosis: an orchestrated immune response against epithelial cells. Immunol. Rev. 174: 210 –225. 3. Amano, K., P. S. Leung, R. Rieger, C. Quan, X. Wang, J. Marik, Y. F. Suen, M. J. Kurth, M. H. Nantz, K. S. Ansari, et al. 2005. Chemical xenobiotics and mitochondrial antigens in primary biliary cirrhosis: identification of antibodies against a common environmental, cosmetic and food additive, 2-octynoic acid. J. Immunol. 175: 5874 –5883. 4. Rieger, R., P. S. Leung, M. R. Jeddeloh, M. J. Kurth, M. H. Nantz, K. S. Lam, D. Barsky, A. A. Ansari, R. L. Coppel, I. R. Mackay, and M. E. Gershwin. 2006. Identification of 2-nonynoic acid, a cosmetic component, as a potential trigger of primary biliary cirrhosis. J. Autoimmun. 27: 7–16. 5. Kita, H., Z. X. Lian, J. Van de Water, X. S. He, S. Matsumura, M. Kaplan, V. Luketic, R. L. Coppel, A. A. Ansari, and M. E. Gershwin. 2002. Identification of HLA-A2-restricted CD8⫹ cytotoxic T cell responses in primary biliary cirrhosis: T cell activation is augmented by immune complexes cross-presented by dendritic cells. J. Exp. Med. 195: 113–123. 6. Kita, H., S. Matsumura, X. S. He, A. A. Ansari, Z. X. Lian, J. Van de Water, R. L. Coppel, M. M. Kaplan, and M. E. Gershwin. 2002. Quantitative and functional analysis of PDC-E2-specific autoreactive cytotoxic T lymphocytes in primary biliary cirrhosis. J. Clin. Invest. 109: 1231–1240. 7. Shimoda, S., M. Nakamura, H. Shigematsu, H. Tanimoto, T. Gushima, M. E. Gershwin, and H. Ishibashi. 2000. Mimicry peptides of human PDC-E2 163–176 peptide, the immunodominant T-cell epitope of primary biliary cirrhosis. Hepatology 31: 1212–1216. 8. Long, S. A., C. Quan, J. Van de Water, M. H. Nantz, M. J. Kurth, D. Barsky, M. E. Colvin, K. S. Lam, R. L. Coppel, A. Ansari, and M. E. Gershwin. 2001. Immunoreactivity of organic mimeotopes of the E2 component of pyruvate dehydrogenase: connecting xenobiotics with primary biliary cirrhosis. J. Immunol. 167: 2956 –2963. 9. Leung, P. S., C. Quan, O. Park, J. Van de Water, M. J. Kurth, M. H. Nantz, A. A. Ansari, R. L. Coppel, K. S. Lam, and M. E. Gershwin. 2003. Immunization with a xenobiotic 6-bromohexanoate bovine serum albumin conjugate induces antimitochondrial antibodies. J. Immunol. 170: 5326 –5332. 10. Yamamoto, K., M. M. Fisher, and M. J. Phillips. 1985. Hilar biliary plexus in human liver: a comparative study of the intrahepatic bile ducts in man and animals. Lab. Invest. 52: 103–106. 11. Moteki, S., P. S. Leung, R. L. Coppel, E. R. Dickson, M. M. Kaplan, S. Munoz, and M. E. Gershwin. 1996. Use of a designer triple expression hybrid clone for three different lipoyl domain for the detection of antimitochondrial autoantibodies. Hepatology 24: 97–103. 12. Leung, P. S., R. L. Coppel, A. Ansari, S. Munoz, and M. E. Gershwin. 1997. Antimitochondrial antibodies in primary biliary cirrhosis. Semin. Liver Dis. 17: 61– 69. 13. Leung, P., K. Chu, W. Chow, A. Ansari, C. Bandea, H. Kwan, S. Nagy, and M. Gershwin. 1994. Cloning, expression, and primary structure of Metapenaeus ensis tropomyosin, the major heat-stable shrimp allergen. J. Allergy Clin. Immunol. 94: 882– 890. 14. Morgunov, I., and P. A. Srere. 1998. Interaction between citrate synthase and malate dehydrogenase: substrate channeling of oxaloacetate. J. Biol. Chem. 273: 29540 –29544. 15. Amano, K., P. S. Leung, Q. Xu, J. Marik, C. Quan, M. J. Kurth, M. H. Nantz, A. A. Ansari, K. S. Lam, M. Zeniya, et al. 2004. Xenobiotic-induced loss of tolerance in rabbits to the mitochondrial autoantigen of primary biliary cirrhosis is reversible. J. Immunol. 172: 6444 – 6452. 16. Surh, C. D., A. Ahmed-Ansari, and M. E. Gershwin. 1990. Comparative epitope mapping of murine monoclonal and human autoantibodies to human PDH-E2, the major mitochondrial autoantigen of primary biliary cirrhosis. J. Immunol. 144: 2647–2652. 17. Van de Water, J., M. E. Gershwin, P. Leung, A. Ansari, and R. L. Coppel. 1988. The autoepitope of the 74-kD mitochondrial autoantigen of primary biliary cirrhosis corresponds to the functional site of dihydrolipoamide acetyltransferase. J. Exp. Med. 167: 1791–1799. 18. Matsumura, S., J. Van De Water, P. Leung, J. A. Odin, K. Yamamoto, G. J. Gores, K. Mostov, A. A. Ansari, R. L. Coppel, Y. Shiratori, and M. E. Gershwin. 2004. Caspase induction by IgA antimitochondrial antibody: IgA-mediated biliary injury in primary biliary cirrhosis. Hepatology 39: 1415–1422.
The Journal of Immunology 19. Odin, J. A., R. C. Huebert, L. Casciola-Rosen, N. F. LaRusso, and A. Rosen. 2001. Bcl-2-dependent oxidation of pyruvate dehydrogenase-E2, a primary biliary cirrhosis autoantigen, during apoptosis. J. Clin. Invest. 108: 223–232. 20. Huang, Q., J. Yang, Y. Lin, C. Walker, J. Cheng, Z. G. Liu, and B. Su. 2004. Differential regulation of interleukin 1 receptor and Toll-like receptor signaling by MEKK3. Nat. Immunol. 5: 98 –103. 21. Lane, J. R., D. A. Neumann, A. Lafond-Walker, A. Herskowitz, and N. R. Rose. 1993. Role of IL-1 and tumor necrosis factor in coxsackie virus-induced autoimmune myocarditis. J. Immunol. 151: 1682–1690. 22. Long, S. A., J. Van de Water, and M. E. Gershwin. 2002. Antimitochondrial antibodies in primary biliary cirrhosis: the role of xenobiotics. Autoimmun. Rev. 1: 37– 42. 23. Rose, N. R. 2000. Viral damage or “molecular mimicry”—placing the blame in myocarditis. Nat. Med. 6: 631– 632. 24. Frey, N., U. Christen, P. Jeno, S. J. Yeaman, Y. Shimomura, J. G. Kenna, A. J. Gandolfi, L. Ranek, and J. Gut. 1995. The lipoic acid containing components of the 2-oxoacid dehydrogenase complexes mimic trifluoroacetylated proteins and are autoantigens associated with halothane hepatitis. Chem. Res. Toxicol. 8: 736 –746. 25. Palmer, J. M., A. J. Robe, A. D. Burt, J. A. Kirby, and D. E. Jones. 2004. Covalent modification as a mechanism for the breakdown of immune tolerance to pyruvate dehydrogenase complex in the mouse. Hepatology 39: 1583–1592. 26. Pohlit, H., W. Haas, and H. von Boehmer. 1979. Cytotoxic T cell responses to haptenated cells. I. Primary, secondary and long-term cultures. Eur. J. Immunol. 9: 681– 690. 27. Shearer, G. M. 1974. Cell-mediated cytotoxicity to trinitrophenyl-modified syngeneic lymphocytes. Eur. J. Immunol. 4: 527–533. 28. Koponen, M., W. J. Pichler, and A. L. de Weck. 1986. T cell reactivity to penicillin: phenotypic analysis of in vitro activated cell subsets. J. Allergy Clin. Immunol. 78: 645– 652. 29. Sinigaglia, F. 1994. The molecular basis of metal recognition by T cells. J. Invest. Dermatol. 102: 398 – 401. 30. Kalish, R. S., J. A. Wood, and A. LaPorte. 1994. Processing of urushiol (poison ivy) hapten by both endogenous and exogenous pathways for presentation to T cells in vitro. J. Clin. Invest. 93: 2039 –2047. 31. de Weck, A. L. 1989. Drug allergy, immunotherapy, immune complexes and anaphylaxis. Curr. Opin. Immunol. 2: 548 –557. 32. Weltzien, H. U., C. Moulon, S. Martin, E. Padovan, U. Hartmann, and J. Kohler. 1996. T cell immune responses to haptens: structural models for allergic and autoimmune reactions. Toxicology 107: 141–151. 33. Martin, S., A. von Bonin, C. Fessler, U. Pflugfelder, and H. U. Weltzien. 1993. Structural complexity of antigenic determinants for class I MHC-restricted, hap-
2657
34. 35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
ten-specific T cells: two qualitatively differing types of H-2Kb-restricted TNP epitopes. J. Immunol. 151: 678 – 687. Martin, S., and H. U. Weltzien. 1994. T cell recognition of haptens, a molecular view. Int. Arch. Allergy Immunol. 104: 10 –16. Aoki, C. A., C. M. Roifman, Z. X. Lian, C. L. Bowlus, G. L. Norman, Y. Shoenfeld, I. R. Mackay, and M. E. Gershwin. 2006. IL-2 receptor ␣ deficiency and features of primary biliary cirrhosis. J. Autoimmun. 27: 50 –53. Shimoda, S., F. Ishikawa, T. Kamihira, A. Komori, H. Niiro, E. Baba, K. Harada, K. Isse, Y. Nakanuma, H. Ishibashi, et al. 2006. Autoreactive T-cell responses in primary biliary cirrhosis are proinflammatory whereas those of controls are regulatory. Gastroenterology 131: 606 – 618. Liversidge, J., A. W. Thomson, H. F. Sewell, and J. V. Forrester. 1987. EAU in the guinea pig: inhibition of cell-mediated immunity and Ia antigen expression by cyclosporin A. Clin. Exp. Immunol. 69: 591– 600. Schaecher, K., A. Rocchini, J. Dinkins, D. D. Matzelle, and N. L. Banik. 2002. Calpain expression and infiltration of activated T cells in experimental allergic encephalomyelitis over time: increased calpain activity begins with onset of disease. J. Neuroimmunol. 129: 1–9. Jones, D. E., J. M. Palmer, K. Bennett, A. J. Robe, S. J. Yeaman, H. Robertson, M. F. Bassendine, A. D. Burt, and J. A. Kirby. 2002. Investigation of a mechanism for accelerated breakdown of immune tolerance to the primary biliary cirrhosis-associated autoantigen, pyruvate dehydrogenase complex. Lab. Invest. 82: 211–219. Jones, D. E., J. M. Palmer, A. D. Burt, C. Walker, A. J. Robe, and J. A. Kirby. 2002. Bacterial motif DNA as an adjuvant for the breakdown of immune selftolerance to pyruvate dehydrogenase complex. Hepatology 36: 679 – 686. Jones, D. E., J. M. Palmer, J. A. Kirby, D. J. De Cruz, G. W. McCaughan, J. D. Sedgwick, S. J. Yeaman, A. D. Burt, and M. F. Bassendine. 2000. Experimental autoimmune cholangitis: a mouse model of immune-mediated cholangiopathy. Liver 20: 351–356. Masanaga, T., Y. Watanabe, J. Van de Water, P. S. Leung, T. Nakanishi, G. Kajiyama, B. H. Ruebner, R. L. Coppel, and M. E. Gershwin. 1998. Induction and persistence of immune-mediated cholangiohepatitis in neonatally thymectomized mice. Clin. Immunol. Immunopathol. 89: 141–149. Ohba, K., K. Omagari, K. Murase, H. Hazama, J. Masuda, H. Kinoshita, H. Isomoto, Y. Mizuta, M. Miyazaki, I. Murata, and S. Kohno. 2002. A possible mouse model for spontaneous cholangitis: serological and histological characteristics of MRL/lpr mice. Pathology 34: 250 –256. LeSage, G. D., A. Benedetti, S. Glaser, L. Marucci, Z. Tretjak, A. Caligiuri, R. Rodgers, J. L. Phinizy, L. Baiocchi, H. Francis, et al. 1999. Acute carbon tetrachloride feeding selectively damages large, but not small, cholangiocytes from normal rat liver. Hepatology 29: 307–319.
