Experimental Evidence on the Immunopathogenesis of Primary Biliary Cirrhosis

REVIEW

Experimental evidence on the immunopathogenesis of primary biliary cirrhosis

Carlo Selmi1,2,3,FrancescaMeda1,2,AnaidKasangian2, Pietro Invernizzi2,ZhigangTian4,ZhexiongLian4, Mauro Podda1,2 and M Eric Gershwin3

Primary biliary cirrhosis (PBC) is a chronic cholestatic liver disease for which an autoimmune pathogenesis is supported by clinical and experimental data, including the presence of autoantibodies and autoreactive T cells. The etiology remains to be determined, yet data suggest that both a susceptible genetic background and unknown environmental factors determine disease onset. Multiple infectious and chemical candidates have been proposed to trigger the disease in a genetically susceptible host, mostly by molecular mimicry. Most recently, several murine models have been reported, including genetically determined models as well as models induced by immunization with xenobiotics and bacteria. Cellular & Molecular Immunology (2010) 7, 1–10; doi:10.1038/cmi.2009.104; published online 23 December 2009

INTRODUCTION A probable diagnosis is made when two of three criteria are Primary biliary cirrhosis (PBC) is a chronic cholestatic liver disease of present. unknown etiology characterized by autoimmune-mediated destruc- The most common symptoms include fatigue, which is present in tion of small- and medium-sized intrahepatic bile ducts.1 about 19% of patients at diagnosis, pruritus in about 20% of patients Serologically, PBC is characterized by the presence of increased levels at diagnosis, and jaundice,5 but because of the changing disease scen- of immunoglobulin M (IgM), a high titer of serum antimitochondrial ario jaundice is now a very rare sign at presentation.6 autoantibodies (AMAs) and, in some of patients, PBC-specific anti- Although the complete pathways of PBC pathogenesis remain nuclear antibodies (ANAs).1 From both a clinical and a pathogenetic unknown, several clinical and experimental findings strongly support point of view, PBC is considered a peculiar, yet representative, auto- autoimmune mechanisms for bile duct damage.7 However, the ques- immune disease.2 PBC preferentially affects women, with one of the tion of what causes the disease remains unanswered. highest female/male ratios (10:1) described in autoimmunity,3 and The most accepted hypothesis states that PBC results from an envir- most cases present within the fifth and sixth decades of life, with only onmental insult on a genetically susceptible background. In this scen- exceptional cases reported in teenagers. AMAs are present in about ario, adaptive, both humoral and cellular (CD4 and CD8 T cells), and 95% of PBC cases, with a disease specificity close to 100%, and are innate immunity have been proposed as coplayers in immune- therefore considered the serological hallmark of the disease. mediated liver damage. Histologically, PBC presents bile duct inflammation with consequent This review includes a critical discussion of what is known and what destruction/loss of intrahepatic bile ducts and development of fibrosis we hope will soon be known regarding the causes of onset and per- and biliary cirrhosis. There are four PBC histological stages: (i) portal petuation of liver damage in PBC. tract inflammation with bile duct obliteration and granulomas; (ii) In this regard, we will first discuss the somehow overlooked role of extension of inflammation to the periportal area; (iii) septal or female predominance in autoimmunity in general and in PBC in bridging fibrosis, with ductopenia (over half of the visible interlobular particular. We will then review what is known of the genetic basis of bile ducts having vanished); and (iv) estab-lished cirrhosis virtually PBC susceptibility and the contribution of environmental factors in its undistinguishable from end stage liver diseases of different etiologies. development. Third, we will illustrate the established evidence on the Epithelioid granulomas with no sign of caseous necrosis are the char- immunobiology of PBC with specific mention of the new line of acteristic lesions of PBC and can be found at any stage around research on innate immunity. damaged bile ducts.4 The definitive diagnosis of PBC is made when all of the following The sex ratio of autoimmune disease three criteria are fulfilled: the presence of serum AMA, increased Similar to Sjogren’s syndrome, systemic lupus erythematosus, auto- enzymes indicating cholestasis (i.e., alkaline phosphatase) for longer immune thyroid disease and scleroderma, PBC manifests the highest than 6 months and a compatible or diagnostic liver histology. predominance in females, with over 80% of patients being women. To

1Department of Translational Medicine, Universita`degli Studi di Milano, Milan, Italy; 2Department of Medicine, IRCCS Istituto Clinico Humanitas Rozzano, Milan, Italy; 3Division of Rheumatology, Allergy, and Clinical Immunology, University of California at Davis, Davis, CA, USA and 4Institute of Immunology, University of Science and Technology of China, Hefei, China Correspondence: Dr C Selmi, Department of Medicine, IRCCS Istituto Clinico Humanitas, via A. Manzoni 56, 20089 Rozzano, Milan, Italy E-mail: [email protected] Received 5 October 2009; accepted 12 October 2009 Experimental evidence on immunopathogenesis of PBC C Selmi et al 2

explain this sex bias observed in autoimmunity, three major working We have recently provided experimental data supporting a new hypotheses have been investigated so far, i.e., the role of sex hormones, fascinating hypothesis on the female predominance of autoimmunity fetal microchimerism and X-chromosome defects. based on major sex chromosome defects.17 This theory is based mainly Sex-associated hormones, e.g., estrogens, androgens and prolactin, on two observations: first, several genes that are key factors in the which not only differ between males and females but also can vary maintenance of immune functions and tolerance are located on the according to age, have been the first candidates taken into account, X chromosome and, second, diseases associated with constitutive X mainly because of their ability to modulate immune responses. monosomy or its major structural abnormalities, such as Turner’s Estrogen-mediated modulation of the immune response can act at syndrome, are frequently associated with autoimmune features and different levels, including regulation of lymphocyte homing to a target in some cases with chronic cholestasis. Quite surprisingly, only a few organ and antigen presentation, thus potentially influencing both studies have so far investigated the genetics of sex chromosomes in organ specificity of autoimmunity and breakdown of tolerance. autoimmunity. X-chromosome inheritance displays a peculiar pattern Estrogens are also capable of directly modulating both pro- and compared with autosomal chromosomes, since women are functional anti-inflammatory activities of CD4 T cells and are therefore capable mosaics for X-linked genes. In females, most genes on one X chro- of influencing the outcome of the CD4 T-cell-mediated immune res- mosome are silenced as a result of X-chromosome inactivation (XCI). ponse. Finally, sex hormones can have activational effects on the The result of XCI is to achieve equivalent levels of X-linked gene hypothalamic–pituitary–adrenal axis because of their ability to modu- products between males and females. More recent data have mined late stress responses. Based on these observations, it should be clear this dogmatic view by demonstrating that at least 15% of X-linked that sex hormones, particularly estrogens, may be central to the T genes are capable of escaping XCI in healthy women and are thus helper 1/T helper 2 balance within sites of inflammation, thus deter- expressed from both X chromosomes. Up to 10% of total X-linked 18 mining an appropriate or inappropriate inflammatory response in genes manifest variable XCI patterns in different individuals. As a infection, tolerance development or autoimmunity.8 In past decades, result of these observations, a role for the X chromosome was first various studies have reported that PBC is associated with the occur- proposed based on experimental evidence that women with auto- rence of sex hormone dysfunction. Epidemiological studies have immune diseases have a significantly higher frequency of peripheral shown a presence of menstrual dysfunctions, a higher frequency of blood cells with a single X chromosome (i.e., X monosomy) compared hysterectomy and a higher risk of osteoporosis, indirectly suggesting with healthy women. Importantly, this was observed in diseases with 19 an active role of estrogens in the pathogenesis of disease. In a recent different organ specificities, including PBC, scleroderma and auto- 20 study, taking hormone replacement therapies following menopause immune thyroid disease. Loss of an X chromosome is indeed pref- 21 was significantly associated with PBC, although this may be secondary erential and more frequently involves a parentally inherited one, to the proposed enhanced rate of bone loss in chronic cholestasis.9 suggesting a possible critical involvement of X-chromosome gene Besides their immunological functions, estrogens could have direct product defects in the female preponderance of PBC and other auto- 22 effects on cholangiocytes, the specific target of immune-mediated immune diseases, although new factors, such as microRNAs or epi- 23,24 damage in PBC. Cholangiocytes have indeed been reported to be able genetics, should not be overlooked. Other authors have suggested to express estrogen receptors. Estrogens could, therefore, directly tar- that women affected with specific female-preponderant autoimmune get cholangiocytes through their receptors, inducing proliferation diseases, i.e., scleroderma, manifest a skewed XCI pattern in their peripheral white blood cells.25 In PBC, however, we failed to dem- (also in response to damage) and/or secretory activities of cholangio- 21 cytes. It has been shown that cholangiocytes from patients with PBC, onstrate such preferential inactivation. especially in the advanced histological stages, do not express estrogen receptors, thus suggesting a role of estrogen deficiency in the develop- WHAT CAUSES THE DISEASE? ment of ductopenia in PBC. The possible influence of estrogens on Genetics factors damaging cholangiocytes in PBC still needs to be confirmed and fur- A decisive role of genetic factors in conferring PBC susceptibility has 10 ther investigated. been widely demonstrated, as in other autoimmune diseases.26 As with A second hypothesis on female predominance in autoimmune dis- other autoimmune diseases, genetic components of PBC susceptibility eases is the persistence of fetal genome parts in women, a phenomenon are not related to a single gene and do not have a Mendelian inher- known as fetal microchimerism. Persistence in the maternal circula- itance. First insights into a genetic component come from early epi- tion of fetal cells, even for decades after pregnancy, creates a micro- demiological studies. These studies reported that the risk of chimeric status in which fetal cells are semiallogeneic to the maternal developing PBC was significantly increased if a first-degree relative immune system. Many studies in autoimmunity have tried to associate suffered from the disease, thus suggesting some level of heritability. the presence of fetal microchimeric cells with an increased risk of In general, previous data indicate that 1–6% of PBC cases27,28 developing autoimmunity in women. Microchimeric cells were indeed (Table 1) have at least one family member manifesting the disease reported in peripheral blood mononuclear cells from patients with and our most recent study on 1032 patients throughout the United autoimmune diseases, such as scleroderma, thus suggesting that States confirmed such data, showing that 6% of cases also had a first- non-autologous cells might mediate a graft-versus-host disease-like degree relative affected.9 reaction in these patients. Nevertheless, subsequent studies have failed A more recent study has also shown an increased incidence of AMA to recapitulate these findings.11,12 In PBC, most of the studies failed to without any sign of disease in offspring and first-degree relatives of find a significant difference in the frequency of male microchimerism PBC patients. Such familial prevalence rates are significantly higher in female PBC compared with controls.13–16 We are convinced that than general population prevalence estimates, thus indirectly indi- fetal microchimerism may be involved in the pathogenesis of some cating the existence of a genetic disposition.29 However, the difficulty autoimmune diseases, although available data are still controversial, and therefore a major limitation of such studies is that prevalence rates but it does not seem to play a major role in the pathogenesis of in the general population are still uncertain, and control groups are PBC. not always included in the family studies.

Cellular & Molecular Immunology Experimental evidence on immunopathogenesis of PBC C Selmi et al 3

Table 1 Reported prevalence rates of PBC in families with one index studies. Urinary tract infections have been reported by several authors case to be more frequent in patients with PBC than in controls, with 36–38 Unrelated index cases (n) Familial prevalence (%) Escherichia coli being the main etiological agent. More recently, our 2005 epidemiological study on 1032 patients with PBC and 1041 United States, 1994 396 4.3 rigorously matched controls9 confirmed the association of an United Kingdom, 1983 117 1 United Kingdom, 1990 347 2.4 increased risk of PBC not only with urinary tract infections but also United Kingdom, 1995 736 1.33 with vaginal infections, lifestyle factors such as smoking, and previous United Kingdom, 1999 157 6.4 pregnancies. We also showed that frequent use of nail polish was Sweden, 1990 111 4.5 associated with a slightly increased risk of having PBC. Italy, 1997 156 3.8 The most widely accepted mechanism by which infectious agents Japan, 1999 156 5.1 can trigger autoimmune responses in several autoimmune diseases is Iceland, 2004 14 7.1 known as molecular mimicry. The molecular mimicry hypothesis states that infectious agents There is a more powerful tool that provides a means to estimate the contain short amino-acid sequences (epitopes) that share a sufficient genetic component in complex/multifactorial diseases, i.e., the com- degree of similarity with self-protein(s) and are therefore able to trig- parison between disease incidence in monozygotic (identical) and ger a promiscuous immune response, both antibody- and cell- dizygotic (non-identical) twins. In autoimmunity, the concordance mediated. This aberrant immune response, in turn, can recognize both rate in monozygotic twins for late-onset diseases has been shown to be, microbial and self-epitopes due to crossreactivity. In PBC, in which on average, well below 50%. We first reported that concordance rates the main autoantigens are of mitochondrial origin, crossreactivity is for PBC are as high as 63% in eight monozygotic sets but null in not surprising due to the conserved sequence of mitochondrial dizygotic twins,30 thus strongly indicating a genetic contribution in enzymes across all species, from eubacteria to mammals. It has been PBC susceptibility. Of note, however, is the observation that in some proposed that mitochondria originated following uptake of bacteria concordant sets, the PBC phenotype as well as comorbidities varied into the precursors of eukaryotic cells and maintenance as intracellular significantly within one pair. We can hypothesize that the phenotype/ symbionts. Thus, it becomes difficult to tease out a causal role for genotype discordance could be due to different factors, including microbial proteins in pathogenesis given their phylogenetic relation- epigenetic factors (perhaps of the X chromosome, as suggested ship to the human autoantigen. One line of argument that we have below),17 exposure to environmental factors or serendipity (see taken is that the breakdown of immune tolerance with consequent below). induction of autoimmunity would be more likely to occur when the The challenge to identify susceptibility gene(s) that predispose for microbial protein is extremely similar in sequence and that it would the development of PBC is still open. The majority of such studies have not be necessary for tolerance breakdown to take place in the disease not only been derived solely from case-control designs,31 but also were target organ. In this scenario, T-cell activation produces crossreacting limited by poor-control-matching criteria and sample size or selec- T cells leading to self-tissue destruction, which ultimately perpetuates tion. A plethora of association studies have been conducted (reviewed the autoimmune injury, possibly through the degeneracy of T-cell elsewhere), mainly focused on immune genes that affect the immune receptors and cross-priming. Several bacterial strains in addition to system belonging to both the human leukocyte antigen (HLA) family E. coli have been suggested as crossreactive agents in PBC, including and non-HLA immune modulator genes, including cytotoxic T- Klebsiella pneumoniae, Proteus mirabilis, Staphylococcus aureus, lymphocyte-associated antigen 4, interleukin (IL)-1, IL-10 and vit- Salmonella minnesota, Mycobacterium gordonae, Neisseria meningitidis amin-D receptor. The discussion of these data goes beyond the aims and Trypanosoma brucei, and the yeast Saccharomyces cerevisiae has 31,32 of this article and details have been reviewed elsewhere. also been investigated in PBC with nonspecific results.39 Conflicting Of interest, a more recent multicenter study reported the first gen- evidence has been obtained on the role of Chlamydia pneumoniae in ome-wide association study and identified IL-12 and its relative recep- the pathogenesis of PBC since its antigens were first detected in 100% 33 tor as susceptibility genes for PBC. While we welcome the results of of PBC explanted liver sections compared with 8.5% of controls, but this enormous study, we also submit our concerns on the need to this was not independently recapitulated.40,41 recapitulate these associations in independent (and possibly more Recently, our group has provided serological and molecular data homogeneous) cohorts of patients. suggesting that a ubiquitous xenobiotic-metabolizing Gram-negative bacterium, Novosphingobium aromaticivorans, is the best candidate Environmental factors thus far for the induction of PBC. Two main characteristics of this Although it is almost certain that PBC susceptibility is conferred by an bacterium render it an ideal candidate for the induction of PBC: (i) it unknown genetic factor(s), gene alteration is not sufficient to trigger contains two proteins sharing amino-acid sequences with the immu- the disease. Exposure to certain environmental factor(s), even not nodominant epitope of pyruvate dehydrogenase complex E2 (PDC- harmful per se, could result in immune tolerance breakdown and are E2), with the highest degree of homology ever described; and (ii) it can therefore necessary for PBC onset. metabolize organic compounds and estrogens. We reported not only Two main environmental factors have been investigated in PBC, that is N. aromaticivorans present in human fecal specimens (25%), i.e., infectious agents (bacteria and viruses) and xenobiotics (chemical but also that it is able to elicit a specific antibody reaction (up to 1000- compounds).34,35 Epidemiological and experimental evidence, as well fold higher than against E. coli) in patients with PBC, but not in as animal models, supports the key role of environmental factors and, controls.42 Most recently, Mattner et al. supported the pathogenic role in particular, xenobiotics in the development of PBC. of N. aromaticivorans being able to induce serum autoantibodies and PBC-like liver lesions in mice following immunization.43 Infectious agents. A first insight on the role of infectious agents/che- A novel human beta-retrovirus has been recently reported in peri- micals as initial triggers of PBC comes from early epidemiological hepatic lymph nodes and other samples from patients with PBC, thus

Cellular & Molecular Immunology Experimental evidence on immunopathogenesis of PBC C Selmi et al 4

suggesting its possible role in the pathogenesis of PBC.44 However, in In PBC, recent experimental data reported significantly lower fre- an independent experimental study conducted with a more compre- quencies of CD41 CD25high Tregs as percentages of total T-cell antigen hensive molecular and immunological approach on a larger series of receptor-ab1/CD41 T cells in PBC patients compared with controls, a patients and controls, we could not confirm such a hypothesis.45 More fact that may contribute to the immune tolerance breakdown in the recently, human beta-retrovirus has been found by real-time reverse disease.53 The possible role in the loss of immune tolerance of such transcriptase polymerase chain reaction in the plasma of patients with cells49 is also supported by PBC animal models (see below). different liver diseases, including PBC, autoimmune hepatitis and 46 viral hepatitis. Although these data have not yet been confirmed, WHAT CAUSES LIVER DAMAGE? they suggest the involvement of human beta-retrovirus in the patho- Immunobiology genesis of etiologically different liver diseases and are therefore not The assumption that PBC is an autoimmune disease is supported by specific for PBC. We therefore strongly discourage the use of antire- 47 several lines of clinical and experimental evidence that make this con- troviral drugs eventually proposed as PBC treatment. dition somehow a model and a paradox for autoimmunity.54,55 In fact, PBC shares features commonly found in autoimmunity, such as Xenobiotics. Xenobiotics are understood as substances that are foreign female predominance, multifactorial genetic predisposition, the pres- to the biological human system. The mechanism through which these ence of autoreactive T cells (CD4 and CD8) and the presence of dis- compounds could induce an aberrant immune response to self-pro- ease-specific autoantibodies, i.e., AMAs and PBC-specific ANAs teins is based on the fact that they are believed to alter or to complex (Table 2). Such autoantibodies, however, represent the basis for either to self- or non-self-proteins, changing their molecular struc- PBC being a paradox, as their direct pathogenetic role is still poorly tures. The altered protein could therefore induce an immune response. defined.55 Serum autoantibodies are detected in approximately 90% of As is the case for molecular mimicry, such an immune response may in patients, yet seronegative cases manifest similar histological features turn lead to cross-recognition of the self form, ultimately ending in and disease progression.56 Further, in the autoimmunity paradigm, immune tolerance breakdown and chronic autoimmunity. the passive transfer of autoantibodies should reproduce disease-spe- Important evidence indicating a possible pathogenic role of an cific clinical features, and experimental immunization with disease- organic compound in PBC comes from recent experimental data specific antigens should reproduce a model disease. These aspects have obtained in our laboratory. Long et al. have demonstrated that specific been only partially reproduced in PBC. halogenated organic compounds, once attached to the major mito- Moreover, in autoimmune diseases, the reduction of disease-spe- chondrial epitope backbone, were able to elicit AMA production by cific autoantibody titers correlates with disease amelioration; this cri- sera from patients with PBC. Antibodies against such a modified terion is also poorly met in PBC, in which there is no correlation mitochondrial epitope had a higher affinity than antibodies to the between pattern or titer of AMA and progression or severity of dis- native form.48 Later, using a multiplex approach, a subsequent study ease.57 Finally, most autoimmune diseases are responsive to immuno- determined that 2-nonynoic acid is recognized by antibodies that do suppressive therapy, while no such agent has proven effective for not crossreact with the PDC-E2 native form present in PBC sera with PBC.58 high affinity. This is particularly interesting since this compound does However, the breakdown of immune tolerance seems crucial to PBC not occur naturally and is found in several cosmetic products, includ- initiation,59,60 possibly secondary to the unique apoptotic features of ing nail polish.9 bile duct cells61–63 and the unique tolerogenic liver function.64,65

Regulatory T cells Autoantibodies CD41 CD25high regulatory T cells (Tregs) are a subset of T cells that Circulating AMAs are highly specific for PBC and are detected in are able to suppress inappropriate immune responses or regulate the nearly 100% of patients when tested for using techniques based on cellular immune response.49 In this scenario, a quantitative or qual- recombinant mitochondrial antigens (immunoblotting or enzyme- itative impairment in Tregs could represent an important key factor in linked immunosorbent assay). The high sensitivity and specificity of the breakdown of self-tolerance.50 An important role of Tregs in trig- AMAs make them one of the most specific diagnostic tests of human gering aberrant immune responses against self has been demonstrated immunopathology,66 different from other autoimmune condi- not only in autoimmunity in general but also in liver autoimmune tions,67,68 including autoimmune hepatitis.69 diseases such as chronic autoimmune hepatitis.51 Moreover, animal In most clinical settings for initial screening, AMAs are usually studies have demonstrated that transfer of T cells lacking the CD41 detected by indirect immunofluorescence using rodent liver, kidney CD25high Treg subset into athymic nude mice results in the develop- and stomach sections as a substrate, leading in some cases to false ment of various T-cell-mediated autoimmune diseases.52 positive or negative results.

Table 2 Summary of the autoantibodies detected in PBC sera using indirect immunofluorescence (IIF), immunoblotting (IB) or other serological methods Molecular targets (pattern) Detection systems PBC-specific Prevalence in PBC (%)

ANA Sp100, PML, SUMO (nuclear dots) IIF 11 30–50 gp210, NUP62 (rim-like) IIF 11 30–50 Anticentromere IIF 2 20 AMA E2 subunits of 2-oxoacid dehydrogenase complexes, mainly PDC-E2 IIF 111 85–95 ELISA, IB

AMA, antimitochondrial autoantibody; ANA, antinuclear antibody; ELISA, enzyme-linked immunosorbent assay; NUP62, nuclear pore glycoprotein p62; PBC, primary biliary cirrhosis; PDC-E2, pyruvate dehydrogenase complex E2; PML, promyelocytic leukemia proteins; SUMO, small ubiquitin-like modifiers.

Cellular & Molecular Immunology Experimental evidence on immunopathogenesis of PBC C Selmi et al 5

AMA specifically recognizes lipoylated domains within compo- technology has shown a 10-fold higher frequency of PDC-E2159– nents of the 2-oxoacid dehydrogenase family of enzymes within the 167-specific CD8 T cells within the liver compared with the peripheral mitochondrial respiratory chain. Among its constituents, the major blood of patients with PBC. Moreover, the precursor frequency of autoantigen recognized is the E2 subunit of PDC-E2. Less frequent PDC-E2-specific autoreactive CD8 T cells is significantly higher in autoantigens are the E2 components of 2-oxoglutarate dehydrogenase the early rather than late stage of the disease. Functionally, autoreac- and branched-chain 2-oxoacid dehydrogenase complexes and the E3 tive CD8 T cells in PBC have specific cytotoxicity against PDC-E2 binding protein.70,71 All immunodominant epitopes contain a aspartic antigen as well as the ability to produce interferon-c rather than acid - lysine - alanine motif, with lipoic acid attached to lysine (K), IL-4/IL-10 cytokines,83 while IL-17 has been recently suggested to play which is necessary and/or sufficient for antigen recognition.72 a role in PBC.84 Highly PBC-specific serum ANAs have been detected in as many as 50% of patients and are named after their immunofluorescence pat- Innate immunity tern, i.e., multiple nuclear dots and rim-like pattern. Multiple nuclear The adaptive immune system recognizes and responds to antigens via dots reactivity is based on the recognition of Sp100 and promyelocytic highly specific T-cell receptors, while innate immunity recognizes leukemia proteins (possibly also crossreacting with small ubiquitin- distinct evolutionarily conserved structures usually shared by invading 73 like modifiers). Rim-like reactivity reacts against proteins of the pathogens, termed pathogen-associated molecular patterns (PAMPs), nuclear pore complexes (NPCs) that include gp210 (a 210-kDa trans- allowing a high efficiency for rapid recognition and elimination of membrane glycoprotein involved in the attachment of NPC constitu- viruses, bacteria and fungi. PAMPs engage pattern-recognition recep- ents within the nuclear membrane), nucleoporin p62 (a glycoprotein tors, such as Toll-like receptors (TLRs), expressed on cells of the innate located in the core of NPCs) and the inner nuclear membrane protein immune system, i.e., monocytes, dendritic cells and natural killer cells, lamin B receptor. Serum anti-gp210 antibodies are detected in about all of which are therefore able to modulate the function of adaptive 25% of patients with PBC (range: 10–40%), while anti-p62 and anti- humoral and cellular immunity.85 Only recently has the study of lamin B receptor antibodies are found in about 13 and 1%, respect- innate immunity as a potential activator of autoimmune responses ively. Unlike AMA, cross-sectional and longitudinal data demon- received a significant impetus and is no longer overlooked by clinical strated an association between the presence of PBC-specific ANAs, immunologists.85 The liver has been recognized both structurally especially anti-NPC antibodies, and a poorer disease prognosis. The and functionally as a major organ of innate immunity. One of the pathogenic role of these antibodies, however, has been poorly inves- largest resident populations of cellular components of the innate tigated so far and remains unknown. immune system, including natural killer and natural killer T (NKT) cells, resides in the liver. Functionally, the liver constantly expresses Autoreactive T cells effective immune responses against a wide range of pathogens from T helper (CD41) T-cell antigen receptor-ab1 and CD81 T cells are viruses to multicellular parasites. In that sense, the ambience of an most commonly seen in portal tracts, in particular around damaged inflammatory milieu seems to be critical for effective immunogenic bile ducts in the liver tissue of PBC patients, strongly suggesting the signals to be delivered to intrahepatic T cells. Mounting evidence involvement of cellular immune mechanisms in biliary damage,74–81 shows that innate immunity likely contributes to the initiation and/ as suggested by liver tolerance.82 In past decades, the nature and the or progression of liver damage. PBC exhibits specific immunological role of cellular adaptive immune responses in PBC have been exten- features, such as the presence of epithelioid granulomas, elevated sively characterized, showing that both CD4 and CD8 T cells particip- levels of polyclonal IgM, hyper-responsiveness to CpG oligodeoxy- ate in liver tissue damage. Autoreactive CD4 T cells specifically nucleotides, increased levels of natural killer cells and cytokine res- targeting PDC-E2 self-antigen have been reported both in the peri- ponses, which strongly suggest a crucial role of innate immunity in the pheral blood and in the liver of PBC patients, but not in controls. pathogenesis of PBC (Table 3). There is a specific 100- to 150-fold increase in the number of PDC- IgMs are commonly elevated in PBC, independent of AMA or ANA 2 86 E2-specific CD4 T cells in the hilar lymph nodes and liver when com- status, and their reduction is usually observed during treatment. pared with peripheral blood of patients with PBC, thus further sup- Our group reported that polyclonal hyperIgM is secondary to a porting their role in liver damage. Shimoda et al. characterized their chronic polyclonal innate immune response of memory B cells to 87 antigen specificity, showing that in HLA DR4*0101-positive PBC specific bacterial PAMPs, such as unmethylated CpG motifs. In this patients, a single epitope, i.e., a 163–176 amino-acid sequence that study, following stimulation with synthetic oligodeoxynucleotides 1 encompasses the lipoic acid-binding residue of the inner lipoyl containing CpG motifs, CD27 memory B cells in cultured domain of PDC-E2, is the immunodominant epitope recognized by peripheral blood mononuclear cells from patients with PBC secreted autoreactive CD4 T cells. Functionally, these cells in PBC patients, but not in controls, are of a proinflammatory nature, as demonstrated by their ability to produce proinflammatory cytokines, such as inter- Table 3 Synopsis of recent findings from the study of innate immun- feron-c.76 ity changes in PBC Autoreactive CD8 T cells have also been well characterized in PBC Peculiarity and are currently considered the major effectors of adaptive immunity Monocytes Increase in absolute number in tissue injury encountered in PBC. The HLA class I restricted epitope Increase in proinflammatory cytokines upon infectious for CD8 T cells, namely, the 159–167 amino-acid sequence, maps in challenge close vicinity to the epitopes recognized by CD4 T cells, as well as by Elevated IgM B cell response to bacterial stimuli AMA. It is of note that the autoepitope for T cells, both CD4 and NKT cells Increased NKT cells in PBC peripheral blood and liver CD8 T cells, overlaps with the B cell (AMA) counterpart and Increased NKT cytotoxic activity Liver histology Focal duct obliteration with granuloma formation includes the lipoylated amino acid of the inner lipoylated domain. Similar to CD4 autoreactive T cells, the recent use of tetramer IgM, immunoglobulin M; NKT, natural killer T; PBC, primary biliary cirrhosis.

Cellular & Molecular Immunology Experimental evidence on immunopathogenesis of PBC C Selmi et al 6

significantly higher amounts of polyclonal IgM compared with con- Animal models trols. In a subsequent study, our group reported that B cells in peri- Similar to all complex diseases, such as autoimmune disease, the pheral blood mononuclear cells exposed to CpG motifs express development of an animal model is of obvious importance in elucid- increased amounts of TRL9 and CD86 and greatly enhance the pro- ating the mechanism(s) responsible for the initiation and progression duction of AMAs. This evidence strongly suggests a profound disease- of disease. specific dysregulation of B cells and supports the proposed link Several models, mostly murine,91 have been proposed for PBC and between bacteria and PBC pathogenesis. In this scenario, B cells are are illustrated in Table 4. Several mouse strains have been reported to hyper-responsive to innate stimuli, such as microbial CpG motifs, be spontaneous PBC model animals and will be briefly described. and therefore contribute to the perpetuation of the autoimmune Two animal models, i.e., dnTGFbRII and an IL-2 receptor-a knock- process.88 out mouse, indicate the possible crucial role of Treg deficiency in loss Monocytes have also been implicated in the pathogenesis of PBC, of immune tolerance with the consequent development of an auto- with their proinflammatory activity greatly enhanced in PBC. immune response against PDC-E2 in PBC. A mouse with dnTGFbRII Functionally, once activated by PAMPs through TLRs, monocytes showed PBC-like liver disease, i.e., 100% AMA positivity against PDC- are able to release proinflammatory cytokines, including IL-1, IL-6, E2.92 Transforming grown factor-b receptor II is essential for signal IL-18, IL-12 and tumor-necrosis factor-a, which are able to amplify transduction of transforming grown factor-b, which is a key regulator adaptive T-cell-mediated immune responses against pathogens. Our of lymphocyte activation.93 data have shown that peripheral monocytes from patients with PBC A mouse deficient for IL-2 receptor-a, which is highly expressed on challenged with different ligands for TLR2, TLR3, TLR4, TLR5 and Tregs, developed 100% AMA positivity against PDC-E2, 80% ANA TLR9 produce a significantly increased level of all proinflammatory positivity, and lymphocyte infiltration around the portal tracts assoc- cytokines compared with healthy controls.89 From the innate immun- iated with cholangiocyte injury.94 ity perspective, these findings suggest that peripheral monocytes from Animal models support our hypothesis that xenobiotics can induce patients with PBC are more sensitive to infectious stimuli, resulting in loss of tolerance. First, our group demonstrated loss of tolerance in the secretion of proinflammatory cytokines. The mechanisms for such rabbits immunized with 6-bromohexonate, a xenobiotically modified increased sensitivity are currently unknown but might reflect or be hapten mimicking lipoic acid, coupled with bovine serum albumin. secondary to the higher frequency of recurrent Gram-negative bac- The immunized rabbits were able to produce not only antibodies terial infections (e.g., urinary tract infections) in PBC. against the xenobiotic, but also a high titer of anti-PDC-E2 antibodies. Therefore, both B cells and monocytes constantly exposed to bac- Anti-PDC-E2 antibodies in this model were not, however, sufficient to 95 terially derived products from portal blood could participate in the induce specific hepatic lesions, at least in the short term. More modulation of the adaptive cellular immune response and possibly recently, induction of PBC-like lesions was obtained in a non-obese also in its priming. diabetic (NOD) background by Wakabayashi et al. and in guinea-pigs 96,97 The role of NKT cells in autoimmunity is also attracting growing exposed to xenobiotic immunization by Leung et al. attention.90 NKT cells are innate effector cells, which are regulated by An additional animal model is a variant of the NOD mouse model self- and non-self-glycolipid antigens presented by the antigen- (NOD.c3c4). It has been described that NOD.c3c4 presents auto- presenting molecule CD1d. This activation allows a rapid NKT cell immune cholestasis and PBC-specific serology, showing AMA posi- production of effector cytokines and chemokines, thus modulating tivity of 50–60% and ANA positivity of 80–90%. Histologically, both innate and adaptive immune responses. The first evidence of a NOD.c3c4 presents lymphocyte infiltration around portal tracts with possible involvement of NKT cells in the pathogenesis of PBC comes from our study reporting a higher frequency of CD1d-restricted NKT Table 4 Similarities of the innate and acquired immunity compart- cells in PBC patients compared with healthy individuals and, as was ments between proposed animal models and human PBC the case for autoreactive T cells, CD1d-restricted NKTs were more Mouse model Adaptive immunity Innate immunity frequent in the liver compared with peripheral blood in PBC Ae2(a.b)-deficient AMA — patients.81 Chuang et al. more recently confirmed the increased num- Lymphocytic CD81 infiltrates ber of CD1d-restricted NKT cells in the liver of a dominant-negative Decreased Tregs transforming growth factor receptor beta II (dnTGFbRII) mice, our PBC-like liver lesions PBC mouse model, compared with controls. They further investigated dnTGFbRII AMA NKT cells worsen liver the function of such CD1d-restricted NKTs, showing increased inter- Deficient Treg function injury feron-c production in hepatic CD1d-restricted NKTs after exposure IL-2Ra2/2 AMA — to a-galactosylceramide, which represents an NKT-specific ligand. Portal tract CD41 and CD81 They also reported that CD1d-deficient dnTGFbRII mice had cells decreased hepatic lymphoid cell infiltrates and milder cholangitis NOD.c3c4 Lymphocytic infiltrate — 90 AMA, ANA compared with controls. N. aromaticivorans on AMA NKT cells are required We are well aware that innate immune system hyper-responsiveness NOD.1101 PBC-like liver lesions is likely not sufficient for immune tolerance breakdown. We can hypo- Disease transfer by T cells thesize, however, that these alterations might play a role in the initiation Xenobiotic on C57BL/6 Lymphocytic CD81 infiltrate — and/or perpetuation of autoimmune injury. This is particularly intri- AMA guing considering the previously mentioned study by Mattner et al., PBC-like liver lesions who demonstrated that in a murine model of PBC, N. aromaticivorans AMA, antimitochondrial autoantibody; ANA, antinuclear antibody; dnTGFbRII, a was capable of inducing autoreactive AMAs and chronic T-cell- dominant negative form of transforming growth factor receptor-b type II; IL-2Ra,IL- mediated autoimmunity against small bile ducts in an NKT-dependent 2 receptor-a; Treg, regulatory T cells; NKT, natural killer T; NOD, non-obese fashion.43 diabetic; PBC, primary biliary cirrhosis.

Cellular & Molecular Immunology Experimental evidence on immunopathogenesis of PBC C Selmi et al 7

Figure 1 A mechanistic view of the etiopathogenetic hypothesis of PBC development. In PBC, the autoimmune response against biliary epithelial cells is secondary to the triggered immune tolerance breakdown to PDC-E2. (a) Different factors have been implicated for immune tolerance breakdown. Infectious agent(s) sharing sequence homology with PDC-E2, i.e., N. aromaticivorans, can induce immune crossreactivity with PDC-E2. (b) Apoptosis of cholangiocytes can lead to the accumulation of structurally intact PDC-E2, which therefore retains its immunogenicity. (c) Xenobiotics (2-nonynoic acid) disposed of by the liver can accumulate within cholangiocytes and potentially modify PDC-E2 locally. (d) Defective Tregs may lead to an inappropriate control of T-cell response against PDC-E2. (e) Major sex chromosome defects, e.g., X monosomy, have been implicated to explain female predominance of immune tolerance breakdown in autoimmunity in general and in PBC in particular. (f) Genetic factors participate in conferring individual susceptibility, predisposing immune tolerance breakdown, while epigenetic factors can explain phenotype/genotype discordance. Once breakdown of tolerance has been established, innate immunity and adaptive immunity, both humoral and cellular, are thought to be involved in liver tissue damage. PBC, primary biliary cirrhosis; PDC-E2, pyruvate dehydrogenase complex E2. chronic non-suppurative destructive cholangitis and epithelioid gran- should be aberrantly exposed on the cholangiocyte cell surface, uloma formations; nevertheless, the morphological features of bile where it may be recognized by AMA and/or antigen-specific T cells. ducts differ somewhat from those in human PBC.98 Studies based on in situ hybridization of PDC-E2 mRNA failed to demonstrate significant differences in its amount in PBC liver com- PBC pathogenesis pared with other liver diseases. PDC-E2 may be selectively overex- Once we have illustrated the available evidence on the immune pressed in small bile duct cholangiocytes, as indirectly suggested by mechanisms involved in PBC pathogenesis, it should be clear that early experimental evidence showing positive staining of a murine the etiology of the tissue injury and, in particular, the outstanding anti-PDC-E2 monoclonal antibody selectively on the surface of biliary organ specificity of the immune-mediated tissue damage, remains to epithelial cells in the livers of patients with PBC, but not in normal be elucidated, and several hypotheses have been proposed. We are controls. convinced that these should not be regarded as mutually exclusive On the other hand, co- or post-translational modifications of PDC- from any causative factor, such as susceptible genes, but rather as E2 may cause its abnormal turnover, leading to its accumulation. terminal mechanisms of the pathway leading to the clinical manifesta- Chemicals (i.e., xenobiotics) disposed of by the liver may have a role tions, possibly mediated by a specific cytokine pattern.99 in this scenario by accumulating in the biliary epithelial cells and One major hypothesis for the selective destruction of biliary epithe- modifying PDC-E2 locally. Solid data to support these fascinating lial cells implies that the immunodominant autoantigen PDC-E2 mechanisms are lacking or weak and we cannot rule out that the

Cellular & Molecular Immunology Experimental evidence on immunopathogenesis of PBC C Selmi et al 8

molecules expressed and identified on the ductular surface and recog- FUTURE DEVELOPMENTS nized by AMA may not be PDC-E2 itself, but possibly unrelated PDC- What does the future hold for PBC etiology and pathogenesis? We are E2 mimics crossreacting with human PDC-E2. convinced that efforts should be dedicated to overcoming some of the As mentioned before, the pathogenetic role of AMA is not yet clear. conceptual and logistic difficulties. AMA belongs mainly to the immunoglobulin G isotype, in particular First, we encourage the collection of representative families through immunoglobulin G3 subclasses, and thus is potentially pathogenic a worldwide effort to further confirm reported associations. This will through different mechanisms, e.g., complement activation and anti- empower definition of the genetic basis associated with PBC through body-dependent cytotoxicity. There is no direct experimental evid- the collection of a large series of patients and controls and the use of ence, however, supporting the involvement of these mechanisms in genome-wide analysis on thousands of single nucleotide polymorph- the pathogenesis of PBC. isms. Second, the role of xenobiotics and bacteria in the onset of PBC AMA can be also of the immunoglobulin A (IgA) isotype. The role should be further studied by means of new molecular multiplex tools. of such AMA-IgA has long been ignored, but may be critical in the Third and most important, it is time to prove the pathogenic role of pathogenesis of PBC. AMA-IgA can indeed be detected not only in AMA in PBC. In this view, the proposed animal models may be a good sera, but also in bile, saliva and urine of patients with PBC, in some starting point to achieve this goal. Finally, we are convinced that future cases correlating with disease severity. Moreover, IgA represents the progress in PBC will be possible only through a multidisciplinary principal immunoglobulin isotype on epithelial surfaces, including approach uniting clinicians, basic immunologists, geneticists, che- biliary epithelia. AMA-IgA has been reported to colocalize with mists and microbiologists. PDC-E2 both inside the cell cytoplasm and in the apical membrane of cholangiocytes in PBC, but not in controls. Thus, AMA-IgA and in particular AMA-IgA bound to a mitochondrial antigen could disrupt 1 Hirschfield GM, Heathcote EJ. Cholestasis and cholestatic syndromes. Curr Opin cell metabolism and may also induce cellular dysfunction and damage, Gastroenterol 2009; 25: 175–179. thus leading to a tissue-specific injury. We cannot preclude the pos- 2 Kaplan MM, Gershwin ME. Primary biliary cirrhosis. N Engl J Med 2005; 353: sibility that the apical staining obtained with anti-PDC-E2 monoclo- 1261–1273. 3 Lleo A, Battezzati PM, Selmi C, Gershwin ME, Podda M. Is autoimmunity a matter of nal antibodies may be secondary to the presence of immune complexes sex? Autoimmun Rev 2008; 7: 626–630. formed by secreted IgA and AMA antigens, as one line of evidence 4 Ludwig J, Dickson ER, McDonald GS. Staging of chronic nonsuppurative destructive 100 cholangitis (syndrome of primary biliary cirrhosis). Virchows Arch A Pathol Anat Histol seems to suggest. 1978; 379: 103–112. Apoptosis of biliary epithelial cells in PBC warrants further discus- 5 Pares A, Rodes J. Natural history of primary biliary cirrhosis. Clin Liver Dis 2003; 7: sion and may prove crucial for immune tolerance breakdown,63,101 as 779–794. 6 Lee YM, Kaplan MM. The natural history of PBC: has it changed? Semin Liver Dis illustrated under other conditions. Quite surprisingly, Odin et al. first 2005; 25: 321–326. demonstrated that PDC-E2 remains intact and retains its immuno- 7 Selmi C, Zuin M, Gershwin ME. The unfinished business of primary biliary cirrhosis. J Hepatol 2008; 49: 451–460. genicity during cholangiocyte apoptosis because of a cell-specific lack 8 Bouman A, Heineman MJ, Faas MM. Sex hormones and the immune response in 61 of glutathionylation of biliary epithelial cells. The intact PDC-E2 in humans. Hum Reprod Update 2005; 11: 411–423. apoptotic fragments could then be taken up by local antigen-present- 9 Gershwin ME, Selmi C, Worman HJ, Gold EB, Watnik M, Utts J et al. Vierling, risk factors and comorbidities in primary biliary cirrhosis: a controlled interview-based ing cells and transferred to regional lymph nodes for priming of cog- study of 1032 patients. Hepatology 2005; 42: 1194–1202. nate T cells, thus initiating PBC. This is indeed an attractive possibility; 10 Alvaro D, Invernizzi P, Onori P, Franchitto A, de Santis A, Crosignani A et al. Estrogen receptors in cholangiocytes and the progression of primary biliary cirrhosis. J Hepatol however, solid data of such antigen presentation are lacking, and we 2004; 41: 905–912. cannot exclude the possibility that the reported mechanisms are not 11 Nelson JL, Furst DE, Maloney S, Gooley T, Evans PC, Smith A et al. Microchimerism PBC-specific. and HLA-compatible relationships of pregnancy in scleroderma. Lancet 1998; 351: 559–562. In conclusion, we may now recapitulate our current working hypo- 12 Murata H, Nakauchi H, Sumida T. Microchimerism in Japanese women patients with thesis (Figure 1). Three major events in PBC are represented, i.e., bile systemic sclerosis. Lancet 1999; 354: 220. 13 Invernizzi P, de Andreis C, Sirchia SM, Battezzati PM, Zuin M, Rossella F et al. Blood duct cell apoptosis, female predominance and genetic susceptibility. A fetal microchimerism in primary biliary cirrhosis. Clin Exp Immunol 2000; 122: mimicking microorganism (possibly the ubiquitous N. aromaticivor- 418–222. ans) enters the human system through the digestive mucosa, and its 14 Nomura K, Sumida Y, Yoh T, Morita A, Matsumoto Y, Taji S et al. Lack of evidence for leukocyte maternal microchimerism in primary biliary cirrhosis. World J Gastroenterol PDC-E2-like proteins are modified within the liver by xenobiotics to 2004; 10: 2415–2416. form immunoreactive adducts. These modifications could then be 15 Schoniger-Hekele M, Muller C, Ackermann J, Drach J, Wrba F, Penner E et al. Lack of evidence for involvement of fetal microchimerism in pathogenesis of primary biliary sufficient to trigger the innate immune system to initiate a cascade cirrhosis. Dig Dis Sci 2002; 47: 1909–1914. of local inflammatory events, resulting in local dendritic cell activation 16 Corpechot C, Barbu V, Chazouilleres O, Poupon R. Fetal microchimerism in primary and antigen processing. Mucosal antigen-presenting cells could in biliary cirrhosis. J Hepatol 2000; 33: 696–700. 17 Selmi C. The X in sex: how autoimmune diseases revolve around sex chromosomes. turn activate autoreactive T and B cells that are directed to the liver Best Pract Res Clin Rheumatol 2008; 22: 913–922. through the portal system. T cells, therefore, could participate directly 18 Carrel L, Willard HF. X-inactivation profile reveals extensive variability in X-linked not only in autoimmune injury, but also in its amplification and gene expression in females. Nature 2005; 434: 400–404. 19 Invernizzi P, Miozzo M, Battezzati PM, Bianchi I, Grati FR, Simoni G et al. Frequency perpetuation. B cells, on the other hand, could secrete AMA, particu- of monosomy X in women with primary biliary cirrhosis. Lancet 2004; 363: 533–535. larly of the IgA type. AMA-IgA could then be transported to the 20 Invernizzi P, Miozzo M, Selmi C, Persani L, Battezzati PM, Zuin M et al. X chromosome monosomy: a common mechanism for autoimmune diseases. J Immunol 2005; 175: vascular side of biliary epithelial cells, where they could recognize 575–578. PDC-E2-like molecules located on the lumenal surface cell membrane. 21 Miozzo M, Selmi C, Gentilin B, Grati FR, Sirchia S, Oertelt S et al. Preferential X AMA-IgA/PDC-E2-like molecule engagement could initiate an apop- chromosome loss but random inactivation characterize primary biliary cirrhosis. Hepatology 2007; 46: 456–462. totic signaling cascade. Ultimately, the immune complexes of posta- 22 Padgett KA, Lan RY, Leung PC, Lleo A, Dawson K, Pfeiff J et al. Primary biliary poptotic PDC-E2 and immunoglobulin G-AMA and the direct cirrhosis is associated with altered hepatic microRNA expression. JAutoimmun 2009; 32: 246–253. cytopathic effects of autoreactive T cells (and possibly AMA) lead to 23 Sanchez-Pernaute O, Ospelt C, Neidhart M, Gay S. Epigenetic clues to rheumatoid selective bile duct destruction. arthritis. J Autoimmun 2008; 30:12–20.

Cellular & Molecular Immunology Experimental evidence on immunopathogenesis of PBC C Selmi et al 9

24 Hewagama A, Richardson B. The genetics and epigenetics of autoimmune diseases. J 56 Invernizzi P, Crosignani A, Battezzati PM, Covini G, de Valle G, Larghi A et al. Autoimmun 2009; 33:3–11. Comparison of the clinical features and clinical course of antimitochondrial 25 Ozbalkan Z, Bagislar S, Kiraz S, Akyerli CB, Ozer HT, Yavuz S et al. Skewed X antibody-positive and -negative primary biliary cirrhosis. Hepatology 1997; 25: chromosome inactivation in blood cells of women with scleroderma. Arthritis 1090–1095. Rheum 2005; 52: 1564–1570. 57 Benson GD, Kikuchi K, Miyakawa H, Tanaka A, Watnik MR, Gershwin ME. Serial 26 Invernizzi P, Gershwin ME. The genetics of human autoimmune disease. J Autoimmun analysis of antimitochondrial antibody in patients with primary biliary cirrhosis. Clin 2009; in press. Dev Immunol 2004; 11: 129–133. 27 Bach N, Schaffner F. Familial primary biliary cirrhosis. JHepatol1994; 20: 58 Heathcote EJ. Management of primary biliary cirrhosis. The American Association 698–701. for the Study of Liver Diseases practice guidelines. Hepatology 2000; 31: 28 Jones DE, Watt FE, Metcalf JV, Bassendine MF, James OF. Familial primary biliary 1005–1013. cirrhosis reassessed: a geographically-based population study. J Hepatol 1999; 30: 59 Shimoda S, Miyakawa H, Nakamura M, Ishibashi H, Kikuchi K, Kita H et al. CD4 T-cell 402–407. autoreactivity to the mitochondrial autoantigen PDC-E2 in AMA-negative primary 29 Lazaridis KN, Talwalkar JA. Clinical epidemiology of primary biliary cirrhosis: biliary cirrhosis. J Autoimmun 2008; 31: 110–115. incidence, prevalence, and impact of therapy. J Clin Gastroenterol 2007; 41: 60 Sasaki M, Ikeda H, Nakanuma Y. Activation of ATM signaling pathway is involved in 494–500. oxidative stress-induced expression of mito-inhibitory p21WAF1/Cip1 in chronic non- 30 Selmi C, Mayo MJ, Bach N, Ishibashi H, Invernizzi P, Gish RG et al. Primary biliary suppurative destructive cholangitis in primary biliary cirrhosis: an immuno- cirrhosis in monozygotic and dizygotic twins: genetics, epigenetics, and environment. histochemical study. J Autoimmun 2008; 31:73–78. Gastroenterology 2004; 127: 485–492. 61 Odin JA, Huebert RC, Casciola-Rosen L, LaRusso NF, Rosen A. Bcl-2-dependent 31 Selmi C, Invernizzi P, Zuin M, Podda M, Gershwin ME. Genetics and geoepidemiology oxidation of pyruvate dehydrogenase-E2, a primary biliary cirrhosis autoantigen, of primary biliary cirrhosis: following the footprints to disease etiology. Semin Liver Dis during apoptosis. J Clin Invest 2001; 108: 223–232. 2005; 25: 265–280. 62 Allina J, Stanca CM, Garber J, Hu B, Sautes-Fridman C, Bach N et al. Anti-CD16 32 Invernizzi P, Gershwin ME. The genetic basis of primary biliary cirrhosis: premises, not autoantibodies and delayed phagocytosis of apoptotic cells in primary biliary promises. Gastroenterology 2008; 135: 1044–1047. cirrhosis. J Autoimmun 2008; 30: 238–245. 33 Hirschfield GM, Liu X, Xu C, Lu Y, Xie G, Gu X et al. Primary biliary cirrhosis associated 63 Lleo A, Selmi C, Invernizzi P, Podda M, Coppel RL, Mackay IR et al. Apotopes and the with HLA, IL12A, and IL12RB2 Variants. N Engl J Med 2009; 360: 2797–2798. biliary specificity of primary biliary cirrhosis. Hepatology 2009; 49: 871–879. 34 van de Water J, Ishibashi H, Coppel RL, Gershwin ME. Molecular mimicry and primary 64 Volkmann M, Luithle D, Zentgraf H, Schnolzer M, Fiedler S, Heid H et al. SLA/LP/ biliary cirrhosis: premises not promises. Hepatology 2001; 33: 771–775. tRNP((Ser)Sec) antigen in autoimmune hepatitis: identification of the native protein 35 Christen U, Hintermann E, Holdener M, von Herrath MG. Viral triggers for in human hepatic cell extract. J Autoimmun 2009; in press. autoimmunity: is the ‘glass of molecular mimicry’ half full or half empty? J Autoimmun 2009; in press. 65 Tiegs G, Lohse AW. Immune tolerance: what is unique about the liver. J Autoimmun 36 Burroughs AK, Rosenstein IJ, Epstein O, Hamilton-Miller JM, Brumfitt W, Sherlock S. 2009; in press. Bacteriuria and primary biliary cirrhosis. Gut 1984; 25: 133–137. 66 Oertelt S, Rieger R, Selmi C, Invernizzi P, Ansari AA, Coppel RL et al. A sensitive bead 37 Butler P, Hamilton-Miller JM, McIntyre N, Burroughs AK. Natural history of assay for antimitochondrial antibodies: chipping away at AMA-negative primary biliary bacteriuria in women with primary biliary cirrhosis and the effect of antimicrobial cirrhosis. Hepatology 2007; 45: 659–665. therapy in symptomatic and asymptomatic groups. Gut 1995; 36: 931–934. 67 Picha L, Pakas I, Guialis A, Moutsopoulos HM, Vlachoyiannopoulos PG. Comparative 38 Parikh-Patel A, Gold EB, Worman H, Krivy KE, Gershwin ME. Risk factors for primary qualitative and quantitative analysis of scleroderma (systemic sclerosis) serologic biliary cirrhosis in a cohort of patients from the united states. Hepatology 2001; 33: immunoassays. J Autoimmun 2008; 31: 166–174. 16–21. 68 Meroni PL. Pathogenesis of the antiphospholipid syndrome: an additional example of 39 Selmi C, Gershwin M. Bacteria and human autoimmunity: the case of primary biliary the mosaic of autoimmunity. J Autoimmun 2008; 30:99–103. cirrhosis. Curr Opin Rheumatol 2004; 16: 406–410. 69 Longhi MS, Ma Y, Mieli-Vergani G, Vergani D. Aetiopathogenesis of autoimmune 40 Leung PS, Park O, Matsumura S, Ansari AA, Coppel RL, Gershwin ME. Is there a hepatitis. J Autoimmun 2009; in press. relation between Chlamydia infection and primary biliary cirrhosis? Clin Dev 70 Gershwin ME, Mackay IR, Sturgess A, Coppel RL. Identification and specificity of a Immunol 2003; 10: 227–233. cDNA encoding the 70 kd mitochondrial antigen recognized in primary biliary 41 Abdulkarim AS, Petrovic LM, Kim WR, Angulo P, Lloyd RV, Lindor KD. Primary biliary cirrhosis. J Immunol 1987; 138: 3525–3531. cirrhosis: an infectious disease caused by Chlamydia pneumoniae? J Hepatol 2004; 71 Selmi C, Zuin M, Bowlus CL, Gershwin ME. Anti-mitochondrial antibody-negative 40: 380–384. primary biliary cirrhosis. Clin Liver Dis 2008; 12: 173–185, ix. 42 Selmi C, Balkwill DL, Invernizzi P, Ansari AA, Coppel RL, Podda M et al. Patients with 72 Bruggraber SF, Leung PS, Amano K, Quan C, Kurth MJ, Nantz MH et al. Autoreactivity primary biliary cirrhosis react against a ubiquitous xenobiotic-metabolizing to lipoate and a conjugated form of lipoate in primary biliary cirrhosis. bacterium. Hepatology 2003; 38: 1250–1257. Gastroenterology 2003; 125: 1705–1713. 43 Mattner J, Savage PB, Leung P, Oertelt SS, Wang V, Trivedi O et al. Liver autoimmunity 73 Invernizzi P, Selmi C, Ranftler C, Podda M, Wesierska-Gadek J. Antinuclear antibodies triggered by microbial activation of natural killer T cells. Cell Host Microbe 2008; 3: in primary biliary cirrhosis. Semin Liver Dis 2005; 25: 298–310. 304–315. 74 Shimoda S, van de Water J, Ansari A, Nakamura M, Ishibashi H, Coppel RH et al. 44 Xu L, Shen Z, Guo L, Fodera B, Keogh A, Joplin R et al. Does a betaretrovirus infection Identification and precursor frequency analysis of a common T cell epitope motif in trigger primary biliary cirrhosis? Proc Natl Acad Sci USA 2003; 100: 8454–8459 mitochondrial autoantigens in primary biliary cirrhosis. JClinInvest1998; 102: 45 Selmi C, Ross SR, Ansari AA, Invernizzi P, Podda M, Coppel RL et al.Lackof 1831–1840. immunological or molecular evidence for a role of mouse mammary tumor retrovirus 75 Shimoda S, Harada K, Niiro H, Yoshizumi T, Soejima Y, Taketomi A et al. Biliary in primary biliary cirrhosis. Gastroenterology 2004; 127: 493–501. epithelial cells and primary biliary cirrhosis: the role of liver-infiltrating 46 McDermid J, Chen M, Li Y, Wasilenko S, Bintner J, McDougall C et al.Reverse mononuclear cells. Hepatology 2008; 47: 958–965. transcriptase activity in patients with primary biliary cirrhosis and other 76 Shimoda S, Ishikawa F, Kamihira T, Komori A, Niiro H, Baba E et al. Autoreactive T- autoimmune liver disorders. Aliment Pharmacol Ther 2007; 26: 587–595. cell responses in primary biliary cirrhosis are proinflammatory whereas those of 47 Selmi C, Gershwin ME. The retroviral myth of primary biliary cirrhosis: is this (finally) controls are regulatory. Gastroenterology 2006; 131: 606–618. the end of the story? J Hepatol 2009; 50: 548–554. 77 Shimoda S, Miyakawa H, Nakamura M, Ishibashi H, Kikuchi K, Kita H et al. CD4 T-cell 48 Long SA, Quan C, van de Water J, Nantz MH, Kurth MJ, Barsky D et al. autoreactivity to the mitochondrial autoantigen PDC-E2 in AMA-negative primary Immunoreactivity of organic mimeotopes of the E2 component of pyruvate biliary cirrhosis. J Autoimmun 2008; 31: 110–115. dehydrogenase: connecting xenobiotics with primary biliary cirrhosis. JImmunol 78 Shimoda S, Nakamura M, Ishibashi H, Hayashida K, Niho Y. HLA DRB4 0101- 2001; 167: 2956–2963. restricted immunodominant T cell autoepitope of pyruvate dehydrogenase complex 49 Scaglione BJ, Salerno E, Gala K, Pan M, Langer JA, Mostowski HS et al. Regulatory T in primary biliary cirrhosis: evidence of molecular mimicry in human autoimmune cells as central regulators of both autoimmunity and B cell malignancy in New Zealand diseases. J Exp Med 1995; 181: 1835–1845. Black mice. J Autoimmun 2009; 32:14–23. 79 Shimoda S, Nakamura M, Ishibashi H, Kawano A, Kamihira T, Sakamoto N et al. 50 Jordan MA, Baxter AG. The genetics of immunoregulatory T cells. J Autoimmun 2008; Molecular mimicry of mitochondrial and nuclear autoantigens in primary biliary 31: 237–244. cirrhosis. Gastroenterology 2003; 124: 1915–1925. 51 Longhi MS, Meda F, Wang P, Samyn M, Mieli-Vergani G, Vergani D et al. Expansion and de novo generation of potentially therapeutic regulatory T cells in patients with 80 Shimoda S, Nakamura M, Shigematsu H, Tanimoto H, Gushima T, Gershwin ME et al. autoimmune hepatitis. Hepatology 2008; 47: 581–591. Mimicry peptides of human PDC-E2 163–176 peptide, the immunodominant T-cell 52 Kessel A, Bamberger E, Masalha M, Toubi E. The role of T regulatory cells in human epitope of primary biliary cirrhosis. Hepatology 2000; 31: 1212–1216. sepsis. J Autoimmun 2009; 32: 211–215. 81 Kita H, Naidenko OV, Kronenberg M, Ansari AA, Rogers P, He XS et al. Quantitation 53 Lan RY, Cheng C, Lian ZX, Tsuneyama K, Yang GX, Moritoki Y et al. Liver-targeted and and phenotypic analysis of natural killer T cells in primary biliary cirrhosis using a peripheral blood alterations of regulatory T cells in primary biliary cirrhosis. human CD1d tetramer. Gastroenterology 2002; 123: 1031–1043. Hepatology 2006; 43: 729–737. 82 Carambia A, Herkel J. CD4 T cells in hepatic immune tolerance. J Autoimmun 2009; 54 Gershwin ME. The mosaic of autoimmunity. Autoimmun Rev 2008; 7: 161–163. in press. 55 Gershwin ME, Ansari AA, Mackay IR, Nakanuma Y, Nishio A, Rowley MJ et al. Primary 83 Kita H, Matsumura S, He XS, Ansari AA, Lian ZX, van de Water J et al. Quantitative and biliary cirrhosis: an orchestrated immune response against epithelial cells. Immunol functional analysis of PDC-E2-specific autoreactive cytotoxic T lymphocytes in Rev 2000; 174: 210–225. primary biliary cirrhosis. J Clin Invest 2002; 109: 1231–1240.

Cellular & Molecular Immunology Experimental evidence on immunopathogenesis of PBC C Selmi et al 10

84 Lan RY, Salunga TL, Tsuneyama K, Lian ZX, Yang GX, Hsu W et al. Hepatic IL-17 93 Moritoki Y, Zhang W, Tsuneyama K, Yoshida K, Wakabayashi K, Yang GY et al. B cells responses in human and murine primary biliary cirrhosis. J Autoimmun 2009; 32: suppress the inflammatory response in a mouse model of primary biliary cirrhosis. 43–51. Gastroenterology 2009; 136: 1037–1047. 85 Selmi C, Lleo A, Pasini S, Zuin M, Gershwin ME. Innate immunity and primary biliary 94 Wakabayashi K, Lian ZX, Moritoki Y, Lan RY, Tsuneyama K, Chuang YH et al.IL-2 cirrhosis. Curr Mol Med 2009; 9:45–51. receptor alpha(2/2) mice and the development of primary biliary cirrhosis. 86 Pares A, Caballeria L, Rodes J, Bruguera M, Rodrigo L, Garcia-Plaza A et al. Long-term Hepatology 2006; 44: 1240–1249. effects of ursodeoxycholic acid in primary biliary cirrhosis: results of a double-blind 95 Amano K, Leung PS, Xu Q, Marik J, Quan C, Kurth MJ et al. Xenobiotic-induced loss of controlled multicentric trial. UDCA-Cooperative Group from the Spanish Association tolerance in rabbits to the mitochondrial autoantigen of primary biliary cirrhosis is for the Study of the Liver. J Hepatol 2000; 32: 561–566. reversible. J Immunol 2004; 172: 6444–6452. 87 Kikuchi K, Lian ZX, Yang GX, Ansari AA, Ikehara S, Kaplan M, Miyakawa H, Coppel RL, 96 Wakabayashi K, Lian ZX, Leung PS, Moritoki Y, Tsuneyama K, Kurth MJ et al. Loss of Gershwin ME. Bacterial CpG induces hyper-IgM production in CD27(1) memory B tolerance in C57BL/6 mice to the autoantigen E2 subunit of pyruvate dehydrogenase cells in primary biliary cirrhosis. Gastroenterology 2005; 128: 304–312. by a xenobiotic with ensuing biliary ductular disease. Hepatology 2008; 48: 88 Moritoki Y, Lian ZX, Wulff H, Yang GX, Chuang YH, Lan RY et al. AMA production in 531–540. primary biliary cirrhosis is promoted by the TLR9 ligand CpG and suppressed by 97 Wakabayashi K, Yoshida K, Leung PS, Moritoki Y, Yang GX, Tsuneyama K et al. potassium channel blockers. Hepatology 2007; 45: 314–322. Induction of autoimmune cholangitis in non-obese diabetic (NOD).1101 mice 89 Mao TK, Lian ZX, Selmi C, Ichiki Y, Ashwood P, Ansari A et al. Altered monocyte following a chemical xenobiotic immunization. Clin Exp Immunol 2009; 155: responses to defined TLR ligands in patients with primary biliary cirrhosis. 577–586. Hepatology 2005; 42: 802–808. 98 Irie J, Wu Y, Wicker LS, Rainbow D, Nalesnik MA, Hirsch R et al. NOD.c3c4 congenic 90 Chuang YH, Lian ZX, Yang GX, Shu SA, Moritoki Y, Ridgway WM et al. Natural killer T mice develop autoimmune biliary disease that serologically and pathogenetically cells exacerbate liver injury in a transforming growth factor beta receptor II dominant- models human primary biliary cirrhosis. J Exp Med 2006; 203: 1209–1219. negative mouse model of primary biliary cirrhosis. Hepatology 2008; 47: 571–580. 99 Oo YH, Adams DH, The role of chemokines in the recruitment of lymphocytes to the 91 Oertelt S, Ridgway WM, Ansari AA, Coppel RL, Gershwin ME. Murine models of liver. J Autoimmun 2009; in press primary biliary cirrhosis: comparisons and contrasts. Hepatol Res 2007; 37(Suppl 100 Fukushima N, Nalbandian G, van de Water J, White K, Ansari AA, Leung P et al. 3): S365–S369. Characterization of recombinant monoclonal IgA anti-PDC-E2 autoantibodies 92 Oertelt S, Lian ZX, Cheng CM, Chuang YH, Padgett KA, He XS et al. Anti-mitochondrial derived from patients with PBC. Hepatology 2002; 36: 1383–1392. antibodies and primary biliary cirrhosis in TGF-beta receptor II dominant-negative 101 Lleo A, Selmi C, Invernizzi P, Podda M, Gershwin ME. The consequences of apoptosis mice. J Immunol 2006; 177: 1655–1660. in autoimmunity. J Autoimmun 2008; 31: 257–262.

Cellular & Molecular Immunology