SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. The Convenient and Inconvenient Truths
  4. Deficits in Immune Tolerance as a Component of the Cause of PBC
  5. Summary
  6. References

The most difficult issue in autoimmunity remains etiology. Although data exist on effector mechanisms in many autoimmune diseases, the underlying cause or causes are still generically ascribed to genetics and environmental influences. Primary biliary cirrhosis (PBC) is considered a model autoimmune disease because of its signature antimitochondrial autoantibody (AMA), the homogeneity of clinical characteristics, and the specificity of biliary epithelial cell (BEC) pathology. Twenty years ago, we reported the cloning and identification of the E2 component of pyruvate dehydrogenase (PDC-E2) as the immunodominant autoantigen of PBC, allowing for vigorous dissection of T and B lymphocyte responses against PDC-E2 and development of several valid experimental models. There has also been considerable study of the biology of BECs, which has included the unique properties of apoptosis in which there is exposure of PDC-E2 to the effector processes of the immune system. In this review, we present these data in the context of our proposal that the proximal cause of PBC is autoimmunity directed against well-identified mitochondrially located autoantigens in individuals with inherited deficits of immune tolerance. We present these data under the umbrella of convenient truths that support this thesis as well as some inconvenient truths that are not readily accommodated by current theory. Conclusion: We emphasize that the potential initiator of PBC includes inter alia particular environmental xenobiotics; pathogenesis is aided and abetted by genetic weaknesses in mechanisms of immune regulation; and subsequent multilineage immunopathology impacts upon uniquely susceptible BECs to culminate clinically in the chronic autoimmune cholangiolitis of PBC. (HEPATOLOGY 2007.)

It is now 20 years since we reported the successful cloning and identification of pyruvate dehydrogenase E2 subunit (PDC-E2) as the immunodominant mitochondrial autoantigen of primary biliary cirrhosis (PBC). We then predicted rapid progress in defining the etiology of PBC,1 and there is now a better understanding of the origins of anti–PDC-E2 and of CD4, CD8, and B lymphocyte responses to mitochondrial autoantigens. Most recently, progress has accelerated with the generation of 3 separate genetically based mouse models of PBC2–4 and the availability of epidemiological data on large PBC population groups in the United States5 indicative of environmental influences. Further, in laboratory studies, there has been emphasis on the species conservation of the mitochondrial autoantigens including bacteria6 and clear identification of the major epitope site, for both B-cell and CD4 and CD8 T-cell responses, within the inner lipoyl domain of PDC-E2.7–12 Thus, there are now substantial data that encourage us to propose a credible cause for PBC within a general concept of a provocative autoantigenic stimulus and definable deficits in immune tolerance and that allow an assessment of the relevance of these ideas to the cause of other complex diseases of an autoimmune nature.13 We will begin, however, with a contemporary summary of established features of PBC—convenient and inconvenient truths—before engaging the specific thesis of etiology.

The Convenient and Inconvenient Truths

  1. Top of page
  2. Abstract
  3. The Convenient and Inconvenient Truths
  4. Deficits in Immune Tolerance as a Component of the Cause of PBC
  5. Summary
  6. References

Clinical Presentations

Convenient Truths.

PBC14 is an autoimmune liver disease with a high female predilection characterized by progressive immune-mediated destruction of intrahepatic biliary ductules, resulting in decreased bile flow, “obstructive” liver functional indices, hepatic fibrosis and cirrhosis, and eventual liver failure. The clinicopathologic picture is relatively well defined, and diagnostic uncertainties are rare. PBC is clearly associated, within individuals and among family members, with other autoimmune diseases, either organ-specific or multisystem, reflecting the “clustering” so characteristic of autoimmunity.15

Inconvenient Truths.

There is occurrence among cases of PBC of “overlap” syndromes, particularly with autoimmune hepatitis (AIH),7 although most cases that we have evaluated are essentially PBC. Perhaps others represent either de novo PBC in a patient with AIH or a dual disease. Clinical PBC-AIH overlap extends to histopathology noting the occurrence of interface hepatitis in cases of PBC undergoing transition to stage 3 and to autoimmune specificity. However, in the latter case, the specificities of the antinuclear antibodies (ANAs) seen in PBC (cited later) differ from the antinucleosome reactivity seen in AIH. Interestingly, whereas PBC and AIH overlap, as do AIH and primary sclerosing cholangitis (PSC), there is no overlap between PBC and PSC. Additionally, PBC is not highly responsive to conventional immunosuppressive drugs, yet therapeutic benefit (unique to PBC) is conferred by ursodeoxycholic acid, which is not regarded as an immunomodulatory or anti-inflammatory agent.

Pathologic Features

Convenient Truths.

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 major histocompatibility complex (MHC) antigens on the biliary epithelium indicate that autoantigen-specific T cell responses are directed against biliary epithelial cells (BECs), also referred to as cholangiocytes. PBC lesions occur only in bile ducts of small to medium caliber, so the exact nature of the BECs (cholangiocytes) lining these calibers may be relevant, given that cholangiocyte gene expression and phenotypes vary along the proximal to distal bile ducts. Our laboratory has accumulated substantial data suggesting that the destruction of BECs in PBC is mediated by liver-infiltrating autoreactive CD4+ T cells with specificity for the immunodominant PDC-E2 autoantigen, and MHC class II–restricted target epitopes have been mapped16, 17 (see below).

Inconvenient Truths.

In early histologic lesions in PBC, there is eosinophilia, and there are also noncaseating granulomas as seen in certain other liver diseases including sarcoidosis,14 drug reactions, and infections, but they are unique to PBC versus autoimmune pathologies overall. Granulomas have led to suspicions of an microbial basis for PBC,18 but this has not been established, nor can retroviral infection be substantiated.19

Immunologic Features

Convenient Truths.

The serologic signature of PBC is the anti–PDC-E2 (AMA) response, arguably the most highly disease-specific autoantibody in human immunopathology. The targets of the disease-specific multilineage antimitochondrial response are all members of a family of enzymes, the 2-oxo-acid dehydrogenase complexes (2-OADC) with the predominant antigenic site on their E-2 subunits. Thus, the autoantigens include the E2 subunits of the pyruvate dehydrogenase complex (PDC-E2), the branched chain 2-oxo-acid dehydrogenase complex (BCOADC-E2), the 2-oxo-glutaric acid dehydrogenase complex (OGDC-E2), and additionally the dihydrolipoamide dehydrogenase binding protein (E3BP).20 Some 90%-95% of PBC sera are AMA-positive.21 AMA can also be detected in bile, urine, and saliva in PBC, and levels correlate with those in sera.22–24 The 2-OADC target antigens, all located within the matrix of the inner mitochondrial membrane, catalyze the oxidative decarboxylation of keto acid substrates. The E2 enzymes have a common structure consisting of an N-terminal domain that contains a single attachment site or multiple attachment sites to lysine (173K in mammalian PDC-E2) of a lipoic acid cofactor, also called a lipoyl group. The dominant epitope sites recognized by AMA are in contiguity with lipoic acid attachment site(s) in what are designated as the lipoyl domains of these target antigens.21, 25–27 Interestingly, of the mere 5 proteins in mammals that contain lipoic acid, 4 are autoantigens in PBC. Moreover, when recombinant forms of the mitochondrial 2-OADC-E2 proteins are used diagnostically, a positive test is virtually diagnostic or at least indicative that the person is at substantial risk of future development of PBC,28, 29 but more data on this would be useful. Indeed, the antimitochondrial response is so overwhelmingly directed against PDC-E2 as to predict that future nosology could well discard the term AMA in favor of anti–PDC-E2.

In regard to T lymphocytes, there is a 100- to 150-fold increase, in the hilar lymph nodes and the liver compared to peripheral blood, in the number of CD4+ T cells recognizing the human leukocyte antigen (HLA) DRB4*0101–restricted PDC-E2163-176 lipoyl domain peptide.16 We have also characterized an MHC class 1 (HLA-A2)–restricted epitope for CD8+ T cells as PDC-E2159-167.30 Using tetramer technology, there is a 10-fold increase in the frequency of such CD8+ cytotoxic T lymphocytes (CTLs) specific for this epitope in liver compared to blood.8 Interestingly, the dominant autoreactive B cell epitope16 and the MHC (HLA) class I–restricted and class II–restricted T cell epitopes (peptides 159-167 and 163-176) all appear to localize to the same region of the inner lipoyl domain of the autoantigen PDC-E2.8, 21 Furthermore, CTL cell lines specific for PDC-E2 can be efficiently generated from peripheral blood mononuclear cells of patients with PBC using soluble PDC-E2 complexed with anti–PDC-E2 autoantibodies as a result of cross-presentation of the PDC-E2 epitope by antigen-presenting cells (APCs).30 Antibody complexing to the lipoyl domain could protect this sequence until it is loaded onto MHC class I molecules intracellularly and subsequently expressed on the surface of the APC; if so, there would be defined a unique role for anti–PDC-E2 antibodies in the pathogenesis of PBC.30

Inconvenient Truths.

In PBC, there are highly disease-specific and antigen-specific antinuclear activities directed at a variety of antigens, including centromere, nucleoporins gp210 and p62, and Sp100.31 Perhaps these nuclear specificities depend upon unique reactivity to apoptotic biliary cells, but this remains speculative. However, there is a significant correlation of disease severity with antibodies to gp210, and this clearly warrants further research.32

Genetic Influences

Convenient Truths.

PBC has a strong genetic basis, with a higher risk for PBC among relatives than that seen in almost all other autoimmune diseases.33 The relative risk for a first-degree relative is 50- to 100-fold higher than that for the general population. Recently, we evaluated the concordance of PBC in a genetically defined population of twin sets34 and identified 16 pairs of twins within a 1400-family cohort including 8 sets each of monozygotic and dizygotic twins. In 5 of 8 sets of the monozygotic twins, both individuals had PBC (0.63 concordance), but among the 8 dizygotic twin pairs, none were concordant for PBC. Interestingly, the age at onset of disease was similar in 4 of the 5 concordant monozygotic twin pairs. Hence, while the concordance rate of PBC in identical twins is among the highest reported for any autoimmune disease, some discordant pairs were identified.

Inconvenient Truths.

An MHC (HLA) bias, so characteristic of autoimmune disease in general, is lacking in PBC35; there are only weak and regional associations between PBC and MHC class I or class II alleles.36, 37 There is insufficient information available on the MHC of intrafamilial cases for comment on whether such cases are haploidentical. In respect to the enigmatically high female predominance, several hypotheses have been offered. A higher rate of X-monosomy in PBC has been ascertained,38, 39 suggestive of a gene for an immunoregulatory protein on the X chromosome. It is of interest to note that patients with Turner's syndrome or XO have a higher incidence of autoimmune disease.40 Perhaps a key immune response gene on the X chromosome is subject to an epigenetic influence. We should also note that unlike scleroderma, there is refutation of the potential role of microchimerism in PBC.41

Environmental Influences

Convenient Truths.

Environmental factors are confidently proposed as triggering or exacerbating PBC.42–44 We previously reported a large-scale epidemiological study that evaluated risk factors and comorbidities in PBC, including an interview-based study of 1032 patients and controls matched for number, sex, age, race, and geographical location.5 The results indicated, as expected, a risk in having a first-degree relative with PBC (see above), and other risk factors included a history of urinary tract infections, past cigarette smoking, use of reproductive hormone replacement, and, to a minor degree, frequent use of nail polish. Another study showed an increased prevalence of PBC and PSC near superfund (“rubbish-tip”) sites and significant clustering of cases of PBC surrounding such toxic sites.45 Minifocal clustering was identified in large-scale epidemiologic studies from Newcastle, United Kingdom46, 47; this could be explained by local exposures to unidentified environmental agents but may also reflect inbreeding within a local susceptible population group. There are also now detailed data on the incidence of PBC in other countries, including, for example, the United Kingdom, Denmark, Australia, and Estonia.15, 43, 44, 48, 49 Finally, there are intriguing epidemiologic clues, interpretable on either an environmental or genetic basis, including ethnic-racial differences in prevalence even within one country, such as the United States,5 and by a lower prevalence of PBC in Australia among Australian-born than among Caucasian migrant populations.48

Xenobiotics claim attention because many environmental chemicals are metabolized primarily in the liver and, during metabolism, may form reactive metabolites that can modify cellular proteins to form neoantigens. Several mechanisms could be involved. First, a direct toxic effect of the xenobiotics or the neoantigens may induce cell death by apoptosis or necrosis, so favoring “availability” of immunogenic autoepitopes. Second, there may exist in the immune repertoire neoantigen-specific T cells and B cells that, once primed, will cross-react with the less immunogenic native autoantigen. Third, chemical modification of the native cellular protein, such as removal and/or exchange of a hapten, has been shown to change processing in APCs, so leading to the presentation of cryptic peptides.50 The direct toxic effect of xenobiotics is usually dose-dependent and thus becomes evident shortly after drug intake as frank drug-induced liver injury51; hence, such cases are relatively easily identified. On the other hand, the immune-mediated effects that follow such drug toxicity may have latency prior to attaining threshold levels of toxicity for clinical expression.52 During this period, the autoimmune reactivity may undergo substantial evolution both quantitatively and qualitatively, so masking the identity of an original causative agent. Notwithstanding, there has been progress by study of the reactivity of chemically modified PDC-E2 peptides with PBC patient sera that provide a “footprint” of the original neoantigens, thus unmasking the original environmental chemical(s) responsible for the generation of these immunogenic neoantigens.

In our studies, a 12–amino acid residue peptide (amino acids 173 to 184 of the inner lipoyl domain of PDC-E2) was initially utilized, with the lipoic acid cofactor that is attached to 173lysine being replaced with a series of similar but distinct synthetic structures.51–53 The finding was that PDC-E2–specific antibodies from patients with PBC reacted to higher affinity with the modified autoepitopes than the unmodified PDC-E2 peptide. Importantly, several unique compounds, commonly used in perfumes and food flavorings, were recognized by sera from patients with PBC but not by sera from patients with control autoimmune diseases or normal controls.53, 54 Complementary to this, rabbits immunized with 6-bromohexanoate (6-BH)–bovine serum albumin (BSA) in complete Freund's adjuvant (CFA; without PDC-E2 in the immunogen) produced antibodies with a binding specificity and enzyme inhibitory activity to PDC-E2 similar to that of anti–PDC-E2 in human PBC55; these data are supported by recent results in guinea pigs and in a congenic derivative of autoimmune-prone nonobese diabetic (NOD) mice (NOD.1101). We immunized guinea pigs with 6-BH conjugated to BSA in CFA, and these not only developed an anti–PDC-E2 autoantibody (AMA) response to PDC-E2 similar to that seen in human PBC but also developed autoimmune cholangitis after 18 months. Such xenobiotic-immunized guinea pigs have provided the first inducible animal model of PBC and indeed suggest an etiology that has causal implications for human autoimmune diseases other than PBC.56 The data also reinforce the likelihood that in PBC the multilineage anti–PDC-E2 response is itself integral to the pathogenetic mechanisms. We have similar unpublished data for NOD.1101 mice that do not (unlike NOD.c3c4 mice, see below) spontaneously develop AMA or biliary duct pathology but, when immunized with 2-octynamido–BSA, do develop anti–PDC-E2 and autoimmune cholangitis as soon as 8 weeks after immunization, illustrating the important contribution of a genetically based deficit in immunological tolerance as a facilitator of autoimmunity. Taken together, the findings support the thesis that exposure to xenobiotic-modified PDC-E2 can initiate autoimmune responses to PDC-E2 and PBC-like pathology in the liver.

Inconvenient Truths.

The inconvenient truth is that induction of PBC in animals requires CFA for the immunization protocol. CFA has a profound influence on bypassing normal tolerance mechanisms, and the significance of the necessity for CFA in experimental systems to simulate spontaneous PBC in humans remains problematic.

BECs as Disease-Specific Targets in PBC

Convenient Truths.

A striking feature of PBC is the specificity of the immune attack on the small intrahepatic bile ducts, despite the fact that the mitochondrial targets are ubiquitous proteins expressed in all nucleated cells. This apparent paradox can be resolved to some degree if we consider some immunopathological characteristics that explain the unique vulnerability of BECs. Staining of small bile ducts with a panel of monoclonal antibodies (mAbs) against the PDC-E2 mitochondrial autoantigen showed an intense staining at the apical surface of BECs lining the lumen, specifically in PBC livers but not in controls.57, 58 This pattern is distinct from the normal cytoplasmic pattern seen with mAbs against other non–PBC-related mitochondrial proteins.7 The fact that such apical staining is seen only with selected and not all mAbs that react with PDC-E2 and that distinct epitopes can be identified with these apically staining mAbs led to the initial hypothesis that a target molecule cross-reactive with PDC-E2 could be located at the apical surface of BECs in PBC.57–59 However, subsequent research demonstrated that the apical staining was due to a new complex between AMA of immunoglobulin A (IgA) isotype and PDC-E2, and this raised the potential role of IgA as a participant in the immune-mediated destruction of BECs.60, 61

In considering reasons for the specific targeting of the BEC in PBC, we note that this is often conferred in autoimmune diseases by an affinity of tissue-specific peptides for MHC molecules, although PBC has only weak HLA associations, and so we must look further. There are clearly no individually specific predisposing biliary cell phenotypes since features of destruction of BEC re-emerge after a liver transplant for PBC.62 Rather, the involvement in PBC of bile duct cells or occasionally salivary ductular cells, or even other epithelial cells, may depend on particular biological properties of such cells. Such properties may include unique processes of apoptosis, mechanisms related to mitochondrial autophagy, presence of poly Ig receptors, and, especially in the case of BECs, a capacity to elicit intense mucosal responses.63

BECs are active participants rather than innocent victims in the autoimmune pathology of PBC, and our preferred explanation is an anomaly of apoptosis by which a lack of glutathiolation exposes PDC-E2 in such cells to autoimmune effector agents.64, 65 Apoptosis of BEC has been proposed as a potential source of “neoantigens” that could be responsible for activation of autoreactive T lymphocytes or a target for effector cells or antibodies. Unlike other cell types65 for which autoantibody recognition of PDC-E2 was abrogated after apoptosis, the antigenicity of PDC-E2 persists in the apoptotic BEC. The suggested reason is that glutathiolation of PDC-E2 prevents autoantibody recognition of the PDC-E2 autoantigen in non-BECs undergoing apoptosis, but notably, glutathiolation does not occur in the BEC.64, 65 These data take on significance in light of the mechanisms proposed for cardiomyopathy development in infants born of anti-Ro–positive mothers, another autoantibody for Sjogren's syndrome or lupus. In this scenario, anti-Ro binds onto the Ro antigen expressed on the bleb of an apoptotic monocyte and prevents its clearance with subsequent generation of transforming growth factor-β (TGF-β) and inflammatory mediators.66–69

Inconvenient Truths.

There is the inconvenient truth that BECs have been very difficult to study and a considerable amount of information on biliary cell destruction is based on extrapolation of data. Another inconvenient truth is that the most effective drug in PBC is ursodeoxycholic acid. It is assumed that in PBC the mechanism of action of this drug is at the effector rather than the inductive stages of disease not only by providing for less toxic bile but also by inhibiting apoptosis of BECs so as to limit exposure to the immune system of the target PDC-E2 autoantigen. However, this remains speculative.

Deficits in Immune Tolerance as a Component of the Cause of PBC

  1. Top of page
  2. Abstract
  3. The Convenient and Inconvenient Truths
  4. Deficits in Immune Tolerance as a Component of the Cause of PBC
  5. Summary
  6. References

Deficits in immune tolerance are seen as a critical component of the cause of PBC. Furthermore, such tolerance deficits are not simply consequential to the disease but rather are an expression of a genotype that is highly permissive for an aberrant immune response. However, the solution to the problem of self-tolerance, predicted to be impending as far back as 1960,70 remains elusive, despite substantial advances. There are two major bulwarks against immune reactivity to autologous molecules. First, there is deletion of high-affinity self-reactive lymphocytes in the primary lymphoid organs, thymus and bone marrow, during lymphopoiesis in early life. Second, there is activity in the periphery of a subset of T lymphocytes, dedicated to regulatory function [T regulatory cells (Tregs)], that express the CD4 and CD25 surface markers and the transcription factor forkhead box P3 (FOXP3). Additionally, there are various other backups along the road from innate to adaptive immune responses, such that any potentially self-reactive lymphocytes must navigate a series of “checkpoints” that can limit the induction and expression of autoimmunity. Accordingly, failure of self-tolerance could involve multiple faults, mostly of genetic origin, and notably involve apoptosis pathways, cytokines and their receptors, chemokine signaling, T cell–T cell interactions exemplified by CD28/CTLA-4, and intracellular signal transduction. This idea is well supported by various autoimmune disease mutations in inbred strains of mice, whether of spontaneous origin (NZB, lpr, NOD, or others) or induced experimentally.

The idea of underlying deficits in immune tolerance in PBC is based on clinical and experimental evidence. PBC exhibits clustering with various other “tolerance-deficient” autoimmune disorders, either within individuals or among family members, exemplified by thyroiditis, rheumatoid arthritis, Sjogren's syndrome, and the limited (CREST) form of systemic sclerosis, and some 40% of PBC sera show autoimmune reactivity to particular nuclear antigens, centromere, the Sp-100 molecule, and nucleoporins.31, 71, 72 Moreover, there are in PBC reduced levels of Tregs that express the transcription factor FOXP3.73 The experimental evidence is that in 2 genetically manipulated mouse strains, there is spontaneous occurrence, after several weeks of life, of a PBC-like lymphoid cholangiolitis together with positivity for anti–PDC-E2 (AMA).3, 4 The particular genetic manipulation in these mice involved disruption of the receptor for the tolerogenic cytokine TGF-β or the CD25 molecule that is highly expressed on Tregs. The latter manipulation is particularly interesting in light of a PBC-like disease in a child with inborn deficiency of interleukin-2 receptor α (CD25).74 In our outbred model of PBC in the guinea pig, tolerance deficits are replicated by the potent immunoactivating effects of CFA on innate immunity via APCs and their Toll-like receptors (TLRs).56

Finally, we can comment on the appearance of “autoimmune biliary disease” in the NOD.c3c4 mouse in which critical (but as yet unidentified) diabetes susceptibility loci of the parental NOD were replaced with segments of normal mouse chromosomes from BALB/c and C57BL/6 mice.2 This latter model is somewhat different from human PBC because of the initial appearance of dilation of the common bile duct with subsequent disease affecting biliary ductules. Microarray analysis of cholangiocytes from this model showed a deficiency of Fas and FasL, suggesting that these biliary cells failed to die.75 Hence, it is possible that PDC-E2, which is not degraded during biliary apoptosis, would then be presented as an autoantigen in the genetically susceptible NOD.c3c4 mouse, resulting in periductular lymphoid infiltrates and AMA reactivity.

Tolerance-deficit phenotypes are well identified in the various spontaneously autoimmune inbred strains of mice. In the case of outbred humans, equivalent deficits are often evident, although their origins may be more subtle. In a recent analysis, Goodnow76 estimated that the likely frequency of germline gene mutation, multiplied over the large number of known “tolerance genes,” would result in virtually all individuals carrying multiple rare but deleterious mutations in a heterozygous state. However, clinical effects would become evident only when, as a result of a “disadvantageous” mating, there was produced a homozygous autoimmune-prone phenotype, or if somatic mutations were to occur in precursor cells of hemopoietic lineages, with inactivation of a single copy of an intact germline tolerance gene. The characteristic latency of many autoimmune diseases and notably PBC suggests that several mutations may need to occur in a stepwise stochastic manner for clinical effects to become evident.76

Spontaneous germline mutation can be simulated in mice by chemically induced mutagenesis, which is exemplified by mutation of the Roquin gene (of which the gene product represses follicular helper T cells) with ensuing autoimmune expressions such as systemic lupus erythematosus, or by congenic introduction of particular chromosomal insertions.77

Summary

  1. Top of page
  2. Abstract
  3. The Convenient and Inconvenient Truths
  4. Deficits in Immune Tolerance as a Component of the Cause of PBC
  5. Summary
  6. References

We emphasize the very high degree of specificity of antibodies to PDC-E2 and the subsequent development of clinical PBC and our belief that immunoreactivity with recombinant PDC-E2, after a long latency, is strongly indicative of development of clinical disease. Table 1 and Fig. 1 illustrate these data, including our schematic presentation of the steps that lead to PBC.

Table 1. Significant Etiological Observations in PBC
1. PBC has a disease phenotype that has well-defined stage-specific characteristics evolving from a lymphoid cholangiolitis with BEC destruction to ductopenia and cirrhosis.
2. There is a strong genetic predisposition albeit with only a minor MHC (HLA) contribution.
3. Epidemiologic gradients point to susceptible and nonsusceptible population groups.
4. A highly focused multilineage immune response is directed to the nominal autoantigenic epitope, the inner lipoyl domain of the E2 subunit of the pyruvate dehydrogenase complex (PDC-E2) and related enzymes, expressed by antibody (AMA) and T lymphocytes.
5. Numbers of PDC-E2–specific CD4+ T cells are much higher in liver and regional (hepatic) lymph nodes than in blood, and there are higher numbers of PDC-E2–specific tetramer-positive CD8+ T cells in liver than in blood, particularly in stages I and II.
6. Laboratory studies implicate xenobiotics in the initiation of PBC:
 (a) A chemically synthesized lipoic acid mimic, 2-octynoic acid, reacts with AMA +ve PBC sera at a higher affinity than does the “parental” PDC-E2 lipoyl domain.
 (b) Octynoic acid coupled to BSA given with CFA elicits antibody to PDC-E2 (AMA) in rabbits but without liver pathology.
 (c) Octynoic acid coupled with BSA given with CFA elicits antibodies to PDC-E2 (AMA) and autoimmune cholangiolitis in guinea pigs with a latency of 18-24 months.
 (d) Octynoic acid coupled with BSA given with CFA induces antibodies to PDC-E2 (AMA) and autoimmune cholangitis in NOD.1101 mice with a latency of only 8 weeks.
7. Small bile ducts are not “innocent victims” since their unique biology determines their targeting, but no unique PBC biliary phenotype is involved, as PBC reoccurs following transplantation.
thumbnail image

Figure 1. A schematic pathogenesis of PBC. In “natural” apoptosis of epithelial cells, PDC-E2 remains intact because of non-glutathiolation (step 1). PDC-E2 can provide an autoantigenic stimulus as unmodified lipoylated PDC-E2 (step 2a), as xenobiotic/hapten-modified PDC-E2 (step 2b), or as a microbial mimic of PDC-E2 (step 2c). PDC-E2 is endocytosed by an APC, wherein it is degraded to potentially T-cell immunogenic peptides (step 3). Host genetic susceptibility is necessary and critically depends on inherited deficits of immune tolerance (step 4). If the APC is activated via stimulation of TLR (see text), self peptides will be presented in immunogenic mode via MHC class II to autoreactive CD4+ T cells (step 5) and via MHC class I to CD8+ T cells (cross-priming; step 6). CD4+ T cells provide help to autoreactive B cells that produce AMA (step 7). AMA can form complexes with PDC-E2 that are “captured” by APCs via Fc receptors as another source for antigen presentation (step 8). Thus, a multilineage anti–PDC-E2 response is generated (step 9). The BEC is vulnerable because of expression of intact PDC-E2 (see step 1), expression of MHC molecules, and receptors for anti-PDC of IgA isotype (step 10). Intact PDC-E2 released from damaged BEC maintains self-perpetuating disease (step 11).

Download figure to PowerPoint

In recent work using a bead assay, less than 5% of patients with PBC remain AMA-negative.78 Further, in the limited numbers of AMA-negative PBC patients, autoreactive (anti–PDC-E2) CD4+ T cells are readily identifiable (Shimoda S, personal communication, 2007). There is recurrence of PBC after liver transplantation62, 79, 80 with no convincing evidence that the anti–PDC-E2 multilineage reactivity (that is, serology, CD4+ T cells, and CD8+T cells) changes significantly after the allograft.

PBC, like most polygenic autoimmune diseases, clearly belongs to the “complex disease” category that is attributable to combined effects of multiple environmental and behavioral influences, genetic elements, and perhaps chance. We here deconstruct these components into (1) chemical-xenobiotic–dependent initiation, (2) an initial deficit in natural immune tolerance that includes most of the genetic contribution and is permissive for maintenance of the disease, and (3) a vulnerability of the primarily affected BEC by reason of its particular biology.

We appreciate that any hypothesis should accommodate all the facts, not just selective facts, and we have attempted to accomplish this, albeit with some “inconvenient truths.” Our final thought would be that our model of autoimmunity requires that natural immune tolerance be broken, but in autoimmune-prone individuals, tolerance may be sufficiently “labile” that it fails without any provocation. Hence, there is disease clustering as seen also in the spontaneous autoimmune diseases in the NOD and NZB mouse.

The thesis presented herein on PBC has notable implication for immune tolerance and autoimmune disease in general. Tolerance is essentially the absence in the periphery of T and B cells with receptors of sufficient affinity to recognize/react with self, including the presence and effects of Tregs. If one or the other of these is deficient, the individual, whether a mouse or man, is on a knife edge of anti-self reactivity that (1) might not ever happen; (2) happens regularly in “highly” predisposed individuals whenever a native (unaltered) cell fragment encounters an undeleted anti-self lymphocyte under immunogenic conditions, as in autoimmune NOD or NZB mice; or (3) happens occasionally when the immune system of a “moderately” predisposed individual is confronted with a near-self antigen under immunogenic conditions. Indeed, we submit that, whether it is a chemical synthetic xenobiotic or a bacterial mimic, immunization under appropriate circumstances (genetically tolerance-deficient host or use of CFA) can lead to autoimmune pathology including disease of a nature similar to human PBC.

References

  1. Top of page
  2. Abstract
  3. The Convenient and Inconvenient Truths
  4. Deficits in Immune Tolerance as a Component of the Cause of PBC
  5. Summary
  6. References