Immunity in Primary Biliary Cirrhosis

REVIEW Genes and (auto)immunity in primary biliary cirrhosis

C Selmi1,2, P Invernizzi1, M Zuin1, M Podda1, MF Seldin2,3 and ME Gershwin2 1Division of Internal Medicine, San Paolo School of Medicine, University of Milan, Milan, Italy; 2Division of Rheumatology, Allergy and Clinical Immunology, University of California at Davis, Davis, CA 95616, USA; 3Biochemistry and Rowe Program in Human Genetics, University of California at Davis, School of Medicine, Davis, CA, USA

Primary biliary cirrhosis (PBC) is a chronic autoimmune cholestatic liver disease most commonly encountered in postmenopausal women; it is characterized by high-titer serum autoantibodies to mitochondrial antigens, elevated serum IgM, progressive destruction of intrahepatic bile ducts, and ultimately liver cirrhosis and failure. The cytopathic mechanisms leading to the selective destruction of intrahepatic cholangiocytes are still largely unknown. The current theory on the pathogenesis of PBC indicated that environmental factors might trigger autoimmunity in genetically susceptible individuals. In fact, genetic predisposition is critical to disease onset and progression, yet peculiar among autoimmune diseases, as indicated by the lack of a strong association with major histocompatibility complex haplotypes. Further, the recently reported concordance rate among monozygotic twins strenghtens the importance of genetic factors, while also indicating that additional factors, possibly infectious agents or xenobiotics, intervene to trigger the disease. In this review, the available data regarding the genetic factors associated with PBC susceptibility and progression, as well as the available evidence regarding the immunomediated pathogenesis of PBC, will be critically illustrated and discussed. Genes and Immunity (2005) 6, 543–556. doi:10.1038/sj.gene.6364248; published online 21 July 2005

Keywords: autoimmune cholangitis; complex disease; association study; environmental factors; pathogenesis

Introduction we will concentrate on the crosstalk between these two apparently independent compartments of PBC etiopatho- Primary biliary cirrhosis (PBC) is a chronic cholestatic liver genesis, and will draw possible future scenarios. disease of unknown etiology that leads to a progressive nonsuppurative destruction of small and medium-sized intrahepatic bile ducts, and eventually to liver cirrhosis and failure. PBC is currently considered as presenting Genetics of PBC autoimmune features, despite the lack of direct evidence Tackling the genetic bases of complex diseases to support this assumption. Indeed, PBC has significant Similar to other autoimmune diseases, the pathogenesis similarities with other autoimmune diseases, such as the of PBC should be regarded as complex or multifactorial, presence of highly specific autoantibodies directed against with multiple genetic and environmental factors inter- mitochondrial (AMA) or nuclear (ANA) antigens and playing to determine disease susceptibility and onset.2 In autoreactive T cells, the striking female predominance fact, it is now clear that PBC genetic components are not (estimated as a 9 : 1 female to male ratio), and the related to a single gene mutation, nor the disease coexistence of other autoimmune diseaseas in approxi- presents a complete penetrance, as indicated by data 1 mately one-third of patients with PBC. It is currently from monozygotic twin sets. Further, PBC is a complex believed, based on experimental and clinical evidence, that disease that should be defined as rare, based on its PBC onset and progression are determined by environ- prevalence in the general population (the highest mental factors acting on genetically susceptible inidivi- reported being 402 cases per million3). So far, data on duals. Both these components, in fact, appear to be the genetics of PBC are derived only from case–control necessary, albeit not being self-sufficient, to induce the studies designed to determine associations between autoimmune events that characterize PBC. In this article, specific polymorphic genes and the disease.4 This we will attempt a critical approach to the available data on approach, albeit meritory, raises some concerns due to the genetic and immunological bases of PBC. In particular, its limits. In fact, association studies present several degrees of potential bias, including those caused by sample size and selection and control-matching criteria. Correspondence: Dr ME Gershwin, Division of Rheumatology, Allergy For example, patients are often enrolled through tertiary and Clinical Immunology, University of California at Davis School of referral centers and possibly over-represent the more Medicine, GBSF Suite 6510, Davis, CA 95616, USA. E-mail: [email protected] advanced spectrum of PBC. As for controls, it is critical Received 17 June 2005; accepted 17 June 2005; published online 21 that these are matched also by race and geographical July 2005 background to cases to minimize false-positive results Genes and immunity of PBC C Selmi et al 544 based on population structure differences. The choice of terns, while only one affected relative could be identified candidate genes warrants further discussion; in most among 85 first-degree relatives.16 cases it is indicated by immunological features of the Several studies have addressed the hypothesis that less disease, as in the case of the major histocompatibility specific immunological disturbances rather than a complex (MHC), or is often derived from data in other definitive diagnosis of PBC might in fact be shared by autoimmune conditions. Both these methods present family members of patients. Interestingly, as many as several strengths; however, we note how the selection of 70% of PBC relatives had some degree of immunological genes with a solid rationale should in all cases be a impairment in studies from the early 1970s,17,18 while priority in association studies. The overinterpretation of somewhat conflicting results came, more recently, from data is an additional critical issue in the interpretation of two studies of patterns of immunological disorders in association data. To avoid false associations, all putative family members. First, the Liver Group in Newcastle, UK association results should be reproduced. It is disap- found that 14% of first-degree relatives of 160 PBC cases pointing to consider that the vast majority of proposed had at least one autoimmune condition, with systemic associations in complex diseases would fail this test.5 sclerosis being the most frequent diagnosis.19 In a smaller Two more issues need to be addressed. First, there needs study from Brasil, only 5/100 first-degree relatives of 26 to be correction for multiple testing. Recent utilization of patients with PBC referred an autoimmune disease (only statistical approaches such as false discovery rate can in one case PBC), while serum autoantibodies were obviate some of the problems with the conservative detected in 11% of relatives, in most cases not accom- Bonferroni correction that may cause type 1 errors.6 panied by clinical signs of autoimmunity.20 Lastly, one should also consider the publication bias that favors weakly positive associations over negative, yet Twin studies rigorous, studies, thus often causing a duplication of Disease concordance rate in monozygotic and dizygotic efforts that could be easily avoided. twin pairs estimates the weight of genetic and environmental factors in the susceptibility to complex diseases. Familial PBC The concordance rates in monozygotic twins for auto- Family members of patients with PBC have a higher risk immune diseases range between 12% for rheumatoid 21 22 of developing the disease, and the occurrence of multiple arthritis and 83% for celiac disease, and are on cases within the same family is referred to as ‘familial average well below 50%. Until recently, only two case PBC’7 presenting variable frequency rates, possibly reports addressed the issue for PBC, with one describing 23 24 following a geoepidemiological pattern (Table 1). Studies a concordant and the other a discordant pair defined from the UK indicate that such frequency ranges as monozygotic, although in both cases zygosity was not between 1 and 2.4%,8,9 possibly increasing over the past genetically proven. We have recently reported that decade, likely for the awareness of physicians and novel concordance rates for PBC are 63% in monozygotic and diagnostic tools. In the United States, Bach and Schaff- null in dizygotic twins, with three out of eight mono- ner10 observed a 4.3% familial prevalence in an analysis zygotic twin sets being discordant for disease occur- 25 of 396 PBC cases both retrospectively and prospectively. rence. Further, in some cases, the natural history of PBC Familial PBC was studied also in other countries and as well as the comorbidities could vary significantly data indicated frequencies of 4.5% in Sweden,11 3.8% in between concordant twins. The phenotypical discor- Italy,12 and 5.1% in Japan,13 in all cases calculated on the dance is particularly interesting as it could be caused number of unrelated index PBC cases. More recently, a by epigenetic factors, differences in exposure to environ- 26 27 population-based study from Minnesota3 identified 46 mental factors such as infections, or xenobiotics, as cases of PBC and reported a prevalence among family well as serendipity. members higher than the general population. In the most recent series from the UK, 6.4% of patients with PBC had MHC and PBC one familial case; in 1% of cases first-degree relatives Strong associations with specific MHC alleles have been were studied, leading to a 10.5 sibling relative risk.14,15 reported in many autoimmune diseases, in some cases constituting the gold standard for the diagnosis of Finally, we have recently demonstrated that patients 28 with PBC from the genetically peculiar population of undetermined cases. Table 2 shows the results of Iceland did not present similar antimitochondrial pat- association studies for human leukocyte antigens (HLA) in PBC onset and progression.

Class I HLA. In the past, studies performed on small Table 1 Prevalence rates of familial PBC series of subjects (between 21 and 75) have examined the association between HLA class I molecules and PBC Country, year Familial Unrelated index Reference 29–35 prevalence (%) cases (n) susceptibility, but the conclusions of such earlier reports were affected by several major flaws. First, the US, 1994 4.3 396 10 small patient populations evaluated in these studies may UK, 1983 1 117 8 account for a limited statistical power of comparisons. UK, 1990 2.4 347 9 Second, technical methods available at the time only UK, 1995 1.33 736 7 allowed the analysis of limited members of HLA class I UK, 1999 6.4 157 14 11 alleles, therefore raising the possibility of existing Sweden, 1990 4.5 111 associations being underestimated. Nevertheless, our Italy, 1997 3.8 156 12 Japan, 1999 5.1 156 13 group reported that PBC is associated with various Iceland, 2004 7.1 14 16 HLA-B alleles in a small proportion of the patients studied.36 It is possible that positive association might be

Genes and Immunity Genes and immunity of PBC C Selmi et al 545 Table 2 Synopsis of HLA association studies in PBC

Country, year HLA investigated Significant association Prevalence in Prevalence in P (corrected) Cit with PBC (n) controls (n)

Spain, 1979 A, B, C, DRw DRw3 57.1% (12/21) 14.8% (11/74) o0.004 29 Japan, 1983 A, B, DR DR2 68% (15/22) 30% (15/50) o0.042 30 UK, 1985 A, B, DR No Associations — — — 32 UK, 1987 A, B, DR, C4A, C4B, C4B2 45% (15/33) 17% (53/307) 0.014 33 Bf, C3 US, 1987 DR, DQ DR8 30.1% (35/114) 4.7% (8/171) o0.0001 37 DR5 decreased in PBC 9.6% (11/114) 25.2% (43/171) 0.0118 US, 1987 A, B, DR(1-7) No associations (DR8 ———175 not tested) US, 1990 A, B, C, DR7, DRw8, DRw8 18.4% (6/35) 4.7% (73/1546) 0.02 176 DRw17, DQw2, DQw3 DQw3 decreased in 26.3% (9/35) 53.5% (827/1546) o0.001 PBC Germany, 1991 A, B, C, DR, C4A, DRw8 36% (9/25) 3.6% (6/169) 0.00013 34 C4B, Bf C4A-Q0 72% (18/25) 34.5% (51/148) 0.0056 DRw8/C4A-Q0 20% (5/25) 0.7% (1/148) 9.7E À7* UK, 1992 DRB, DQB DR8 11% (18/159) 4% (6/162) o0.01 177 DR8/DQB1*0402 11% (10/89) 2.2% (4/181) o0.001 Denmark, 1992 A, B, C, DRB, DQA, DR3 52.2% (12/23) 24.6% (296/1204) o0.05 35 DQB, DPA, DPB Japan, 1993 DR, DQ, DPB1 DQ3 80.9% (38/47) 51.3% (77/150) o0.05 45 DPB1*0501 85.1% (40/47) 55.3% (83/150) o0.01 DPB1*0402 decreased 2.2% (1/47) 23.3% (35/150) o0.05 in PBC DR52 (DRB3) 34% (16/47) 54% (81/150) o0.05 decreased in PBC

UK, 1993 DRB1, DRB3, DQA, DR8 18.5% (24/130) 9.1% (33/363) o0.005** 46 DQB Japan, 1994 A, B, C DRB1*0803 35.5% (22/62) 7.4% (32/430) o0.0001 31 DRB1, B3, B5 DQA1*0103 43.5% (27/62) 18.4% (79/430) o0.0001 DQA1, B1 DQB1*0601 43.5% (27/62) 17.9% (77/430) o0.0001 DQA1*0102 decreased 3.2% (2/62) 16.3% (70/430) 0.0288 in PBC UK, 1994 DRB1, DQB1, DPB1 No association 42 US, 1994 DRB1, DQA1, DQB1 DRB1*0801 7.8% (8/102) 1.9% (9/480) o0.05 38 DRB1*0901 3.9% (4/102) 0.8% (4/480) o0.05 DQA1*0401/0601 9.8% (10/102) 2.7% (13/480) o0.001 DQA1*0102 decreased 5.9% (6/102) 18.8% (90/480) o0.001 in PBC DQB1*0602 decreased 2.9% (3/102) 12.1% (58/478) o0.025 in PBC

Germany, 1995 DPB1 DPB1*0301 50% (16/32) 12.8% (6/47) o0.015 44 UK, 1995 DPB1 No association — — — 178 UK, 2001 DRB, DQA, DQB DRB1*0801 15% (25/164) 2.9% (3/102) 0.0014 41 DRB1*0801/ 23% (21/88) 2.9% (3/102) 4.4E À6 DQA1*0401/ DQB1*0402 (late stage) Sweden, 2002 DRB1, DQB1, DPB1 DRB1*08 29.3% (29/99) 11.4% (18/158) 0.001 39 DQB1*0402 28.6% (28/98) 10.8% (17/158) 0.001 Italy, 2003 A, B, DRB1 B*15 8% (9/112) 3.4% (19/558) 0.0039 36 B*41 3.1% 0.3% 8.8E À15 B*55 3.6% (4/112) 1.3% 0.034 B*58 1.8% (2/112) 0.3% 0.00042 DRB1*11 10.7% (16/149) 27.6% (154/558) 1.6E À6 Brazil, 2003 DR, DQ No Associations 43 US, 2005 DRB1, DQA1, DQB1 DRB1*08 19.4% (14/72) 8.7% (33/381) 0.011 40 DRB1*1501 13.9% 26.6% 0.022 DQA1*0102 16.7% 37.9% 0.0014 DQB1*0602 12.7% 27.6% 0.012

secondary to linkage disequilibrium. Based on the available Class II HLA. HLA class II alleles have been widely data, we should therefore regard PBC associations with studied in association with PBC in Caucasian and Asian HLA class I genes as weak. series of patients. In studies from Germany, Spain,

Genes and Immunity Genes and immunity of PBC C Selmi et al 546 Sweden, and the United States, HLA-DR8 (DRB1*08) was more focused on coding variants (ie with demonstrated found with significantly higher frequency in patients effects on phenotype) of genes, although this might with PBC compared to controls; cumulatively, data exclude other SNPs possibly in linkage disequilibrium aggregation indicates that DR8 might constitute a risk with genes important for disease onset.5 A summary of factor for PBC among Caucasians.34,37–40 In 2001, data association studies in PBC is illustrated in Table 3. from Newcastle, UK demonstrated that the linkage of Most studies of SNPs in PBC have been dedicated to DQA1*0401 and DR8-DQB1*0402 is associated with PBC molecules involved in regulating the immune response, progression and not susceptibility.41 This association was thus hypothesizing that genomic differences at these not confirmed in several other studies from different levels might confer susceptibility to the loss of tolerance non-British European populations.32,35,36,42,43 Moreover, or to an aberrant immune response. Based on the other European studies suggested significant associa- expression of cytotoxic T lymphocyte antigen-4 (CTLA- tions of PBC with DR329,35 and DPB1*0301,44 while the 4) by T cells following activation and the regulatory effect most recent study from the United States demonstrated of this molecule on peripheral T-cell responses, SNPs of an association between DRB1*08-DQA1*0401-DQB1*04 CTLA-4 were suggested as factors facilitating the break- haplotype and PBC, albeit in a minority of patients.40 down of tolerance. Accordingly, the coding A49G SNP Finally, studies from Japan failed to provide a consistent was found associated with PBC in a large British study58 picture of HLA class II associations with PBC.30,31,45 and in 77 Chinese patients with PBC,59 while a smaller Interestingly, in a large series of Italian patients with PBC study from Brazil failed to confirm the association.43 and controls, we observed a protective effect of the Several studied were further dedicated to SNPs of DRB1*11 allele against PBC.36 Transporters associated interleukins (IL), based on their active and critical role with antigen processing (TAP1 and TAP2) also belong to in the regulation of the immune response. Prompted by the HLA class II family and have been studied in patients experimental data such as its dysregulated production with PBC, since their products determine the transport of by monocytes in PBC,60 SNPs of IL-1 were studied in antigenic peptides across the endoplasmic reticulum PBC. First, a study from the UK reported a significantly prior to being loaded onto MHC molecules. However, higher frequency of the IL-1B*1,1 genotype in patients possibly also due to the limited polymorphic nature, no with PBC compared to controls. The difference in the IL- association with PBC onset was found.46 In summary, we 1B*1,1 genotype distribution was even more marked in can conclude that the picture of HLA class II involve- patients with early-stage disease, thus possibly indicat- ment in PBC is quite complex. We could somehow ing that IL-1 alleles might influence disease progression, assume that, similar to epidemiological data, the genetic albeit in a cross-sectional analysis.41 The lack of associa- background in PBC should also be regarded as following tion with PBC onset was also independently confirmed a geographical pattern. by Hungarian61 and Chinese62 researchers. The latter group, however, more recently described an association Class III HLA. Data from association studies of poly- of PBC with the IL1-RN intron genotype comparing morphisms of tumor necrosis factor alpha (TNF-a)in frequencies in 77 patients with PBC and 160 controls.63 PBC are conflicting and a cautious interpretation is Further, based on experimental evidence of cytokine encouraged.47 A polymorphism of the gene promoter profiles and its involvement in the development of T region produces the more frequent variant TNF1, and the helper 1 (Th1) cell responses, SNPs of the promoter less frequent variant TNF2,48 with TNF2 associated with region of the IL-10 gene were also analyzed in patients increased transcription.49–51 The prevalence of the TNF2 with PBC and controls.55,63,64 Data from Italian and allele was reportedly protective against PBC onset,52 Japanese series demonstrated that both groups presented while two other studies independently failed to detect a higher prevalence of the À1082G/G genotype.64 Such any difference in genotype distributions between pa- association was not confirmed in 77 Chinese patients tients and controls.15,53 We note that, in the study from with PBC.62,63 Tanaka et al, heterozygous patients had a significantly SNPs of the 1,25-dihydroxyvitamin D receptor (VDR) worse prognosis compared to homozygous TNF1/TNF1 gene have been investigated in several studies, based on patients,53 as indicated by higher Mayo score value, the dual role of vitamin D in the regulation of bone currently the only validated index for PBC.54 However, a metabolism and inflammation. We note in fact that study from Newcastle, UK did not confirm this alleged accelerated bone loss rates in patients with prolonged association.15 Similarly, in the data obtained from cholestasis (as in PBC) have been repeatedly reported, Scottish, Brazilian, and Chinese series of patients with sometimes with conflicting results and in some cases PBC,43,55,56 and from a small population of patients with less than rigorous experimental designs. A sig- undergoing liver transplantation for end-stage PBC,57 no nificant association between BsmII polymorphisms of association of TNF genotypes with disease susceptibility VDR and PBC was reported in patients with PBC from or onset was observed. Germany, Hungary, and China,61,65–67 while the proposed association with bone loss68 was not reproduced.69 We Non-HLA genetic candidates in PBC believe that differences in the VDR gene might unravel High-throughput novel technologies have made the further potential scenarios to help explain the infre- study of single-nucleotide polymorphisms (SNPs) of quency of PBC in African-American women,70 and could candidate genes the method of choice for association in turn support a possible role for sunlight exposure in studies in PBC. In fact, the analysis of SNPs can define PBC onset. Although fascinating, this assumption re- the linkage of specific loci or neighboring regions with mains at present only a theory yet to be demonstrated. disease traits. We note that, alongside the general Molecules responsible for bile acid transport and considerations expressed above on the choice of candi- excretion in the biliary tree have been obvious targets date genes and populations, the study of SNPs should be in the search for genomic determinants of PBC onset.

Genes and Immunity Genes and immunity of PBC C Selmi et al 547 Table 3 Proposed associations between non-MHC genes and susceptibility to PBC

Gene Country, reference PBC patients/ P Value for Further associations controls (n) association with susceptibility

CTLA-4 England17989 200/200 0.000063 Brazil43,44 50/67 NS IL-1 England41,42 164/101 0.00078 IL-1B*1,1 genotype associated with earlier PBC stages (Po0.0001) Hungary61 33/160 NS China63,64 77/160 0.03 MBL Japan64,65 65/218 0.003 NRAMP1 England18090 53/78 o0.024 IL-10 England18191 171/141 0.49 (NS) Italy64 94/72 o0.041 Japan64,65 65/71 NS China63,64 77/160 NS Vitamin D Germany65,66 74/214 0.009 Association with lower bone mineral density receptor in PBC (P ¼ 0.01) Hungary61,62 33/160 o0.03 Hungary66,67 31/51 0.01 China63,64 58/160 0.01 Osteopontin Japan78 50/34 0.16–0.72 CD14 France76 30/27 NS CD40 ligand Japan75 24/— — T-cell receptor Italy79 70/70 NS TLRCb2 heterozygous genotype associated with PBC stages (P ¼ 0.032) eNOS Italy77 109/242 NS Association with early stage disease (P ¼ 0.032) and rapid progression (P ¼ 0.038) ABC B11 Switzerland71,72 76/93 NS ABC B4 Switzerland71,72 76/56 NS CYP2E1 Italy74 165/225 NS CYP2E1 c2 allele associated with advanced disease in a minority of patients (P ¼ 0.015) CYP2D6 Italy74 165/225 NS MDR1 Italy74 165/225 NS PXR Italy74 165/225 NS TLR9 Italy81 90/90 NS

Interestingly, SNPs and mutation of ATP-binding cas- disease severity was observed in a small subgroup of sette (ABC) transporters involved in the secretion of bile patients. from the hepatocyte71 have been associated with intra- Based on different rationales, several other genes have hepatic cholestasis of pregnancy,72 somehow reprodu- been investigated in the susceptibility to PBC in cing the clinical picture observed in PBC. Pauli-Magnus association studies. Among these, the distribution of et al71 have recently investigated by gene sequencing the specific alleles of CD40 ligand,75 CD14,76 endothelial variants of genes coding for the two main ABC nitric oxide synthase,77 early T-lymphocyte activation 1/ transporters with identification of 45 ABCB11 and 46 osteopontin,78 T-cell receptor C-beta2,79 insulin-like ABCB4 variants. Data, however, demonstrated that no growth factor 1, collagen-Ia1,80 and toll-like receptor 981 mutation was associated with PBC. did not differ between patients and controls. In some As explained in depth below, recent studies from our cases, however, a significant association with disease group, combined with epidemiological evidence such as severity77,79 or bone loss80 seemed evident. the geographical pattern of PBC prevalence, have suggested that PBC might result from the combination Sex chromosomes in PBC: towards a new frontier? of a genetic background, with the necessary contribution X-linked genes are well known to determine gender- of environmental factors, either infectious agents or related characteristics at different levels, while also xenobiotics (ie chemicals).73 Prompted by this theory, regulating the immune function, particularly to maintain we have recently investigated whether genetic variants tolerance. We were intrigued that major defects of the X leading to different xenobiotic metabolism or transport chromosome, such as those leading to Turner’s syn- might in turn account for an increased risk of developing drome82 or premature ovarian failure,83 are commonly the disease. We therefore genotyped several polymorph- characterized as autoimmune conditions (particularly isms of enzymes involved in the transport and metabo- thyroid disease) and intrahepatic cholestasis. We then lism of xenobiotics in 169 patients with PBC and 225 studied peripheral X monosomy rates in white blood healthy controls.74 Genes analyzed coded for cytochrome cells from women with PBC and controls, including P450 (CYP) 2E1 and 2D6, multidrug resistance 1 (MDR1), women with viral chronic hepatitis. We found a and pregnane X receptor (PXR). Data demonstrated that significantly higher frequency of monosomy of the X no polymorphism was associated with PBC susceptibil- chromosome in peripheral leukocytes (particularly those ity, while a weak association of the CYP2E1 c2 allele with of the adaptive immune response, ie T and B cells) in

Genes and Immunity Genes and immunity of PBC C Selmi et al 548 PBC84 compared to controls, and that such frequency was a large number of representative families with more than correlated with age in all groups, since it is well one member affected. Second, we encourage the use of established that a certain degree of X chromosome loss novel gene microarray technologies in PBC samples, is a physiological phenomenon.85 Importantly, we ruled with particular attention paid to technical limitations and out that monosomic cells were in fact (microchimeric) bias of the methods in use.92 Lastly, a whole genome male cells.86 Moreover, these findings were also repro- association scanning will definitively provide the best duced in two other autoimmune diseases characterized answer to our questions93 if performed on a large series by female predominance and different organ specificity, of well-selected cases and matched controls. such as systemic sclerosis and autoimmune thyroiditis,87 thus disclosing different theoretical scenarios. First, autoimmunity in PBC could arise from a polygenic Immunopathogenesis of PBC model with an X-linked major locus of susceptibility with genes escaping inactivation being involved; the AMA female susceptibility might therefore as well derive from High-titer serum AMA are sensitive markers for PBC, 1 genes mapping on the Y chromosome. The second thesis being detected in up to 95% of patient sera; their states that PBC susceptibility might arise from a multi- prevalence and specificity depend largely on the diag- genic complex inheritance model in which specific Y- nostic method used. Indirect immunofluorescence (IIF) linked genes are protective. Conversely, X-linked genes on tissue sections is used for screening purposes, would interact with other genes and be involved via a providing a limited sensitivity and, particularly at lower 94 dose-dependent effect. dilutions of sera, low specificity for PBC. When AMA Based on the fact that autoimmunity affects predomi- are tested with recombinant mitochondrial antigens nantly postmenopausal women,88 a role for the persis- (using immunoblotting or ELISA at research centers), tence of nonself (male) cells, likely of fetal origin the sensitivity and specificity of the test approaches in 94 (microchimerism), has been proposed in PBC, as in other selected series 100%. This is particularly important, autoimmune diseases. Fanning et al89 demonstrated that since AMA positivity is one of the three diagnostic detectable fetal DNA in liver samples years after criteria for PBC, thus proving crucial in the clinical pregnancy was associated with PBC, while other groups management of many cases. Serum AMA specifically failed to recapitulate the hypothesis.90,91 The hypothesis, recognize components of the 2-oxoacid dehydrogenase mostly based on data in other conditions, still awaits (2-OADC) family of enzymes located on the inner 95,96 definitive confirmation or denial. mitochondrial membrane, including epitopes within the E2 and E3BP components of the pyruvate dehydro- Is PBC a unique autoimmune disease? Where do we go genase complex (PDC) and the E2 subunits of the 2- from here? oxoglutarate dehydrogenase complex (OGDC-E2) and It should be clear, at this point, that PBC should be branched-chain oxoacid dehydrogenase complex regarded as a ‘peculiar, yet representative,’ autoimmune (BCOADC-E2). Interestingly, AMA epitopes on different condition from a genetic standpoint. First, PBC appears molecules are conformationally closely related and 97,98 to have a unique genetic background compared to other include in all cases lipoic acid attached to the lysine human autoimmune diseases in which, for example, (K) residue within the amino-acid motif DKA. Whether sound evidence of associations with specific MHC alleles lipoic acid is in fact necessary for the epitope recognition 97,99 has been provided. Data on alleles of class I, II, and III by AMA is not understood. MHC in patients with PBC and controls have shown From a clinical standpoint, AMA pattern or titer do not 100,101 some degree of association, albeit weak, but was seldom correlate with disease severity of prognosis; simi- independently confirmed or were limited to specific larly, patients with PBC lacking detectable AMA present geographical areas. Second, we can surmise that PBC clinical features similar to their AMA-positive counter- 102 recognizes a multi-hit genetic background, with specific parts. These observations, alongside the numerous genes predisposing to the development of autoimmunity approaches that lead to the development of AMA not (first ‘hit’) and others determining natural history accompanied by PBC-like liver pathology in animal (second ‘hit’).41 Third, our twin study indicates that models, seem to militate against a direct pathogenic role 103 PBC concordance rates among monozygotic twins are of AMA in PBC. Further, serum AMA seem to appear 104 the highest reported for autoimmunity,25 with the several years the onset of PBC in affected individuals. exception of celiac disease,22 which is also characterized by a strong genetic component as well as a defined ANA environmental stimulus such as gliadin. On the other PBC-specific ANA IIF patterns are found in up to a third hand, PBC also presents some similarities with the of patients and sometimes more commonly in AMA- genetics of other autoimmune diseases. First, as in the negative cases. Such PBC-specific patterns include case of most autoimmune conditions, family members of ‘nuclear rim’ and ‘multiple nuclear dots’, secondary to patients with PBC have an increased risk of being ANA directed against gp210 and nucleoporin 62 (within diagnosed with the disease, or to present other auto- the nuclear pore complex, NPC) and the nuclear body immune features as systematically reported in 1995.7 protein sp100 and PML, respectively.105 More recently, it More recently, the common finding of enhanced X has been suggested that small ubiquitin-like modifiers monosomy also indicates common ground between (SUMO) also constitute independent autoantigens spe- PBC and autoimmunity in general. cific for PBC, although anti-SUMO antibodies are found We believe that future evidence should be derived only in anti-Sp100-positive sera.106 Indeed, cross-sec- from three possibly more extensive approaches. First, a tional and longitudinal studies on anti-NPC antibodies worldwide effort should be dedicated to the collection of in PBC suggested that patients presenting such ANA

Genes and Immunity Genes and immunity of PBC C Selmi et al 549 positivity might present a more severe and rapidly secreting proinflammatory cytokines that influence the progressing disease.107,108 adaptive response following stimulation by pathogen- associated molecular patterns (PAMPs). Data supporting Cellular immunity a possible role of bacterial lipopolysaccharide (LPS) in Autoreactive T cells. The study of liver samples from PBC come from both human and animal studies. First, patients with PBC reveal several types of mononuclear Ballot et al121 studied sera from 36 patients with PBC and cells in the proximity of affected bile ducts with CD4 þ 560 controls, including several autoimmune conditions, and CD8 þ T cells as the most represented subpopula- for the presence of anti-lipid A (the immunogenic and tions.109 toxic component of LPS) IgM and soluble CD14, and The main epitope for autoreactive CD4 þ T cells in observed that 64% of PBC sera were positive for such PBC has been mapped within PDC-E2,110 particularly antibodies in a PBC-specific fashion, correlating also amino-acid residues 163–176 spanning across the inner with more florid histological lesions. Interestingly, lipid lipoylated domain of the subunit.111 Further, the PDC-E2 A accumulates in hepatocytes, Kupffer cells, and amino acids E, D, and K are essential for the recognition biliocytes in PBC livers.122 These results confirmed a of the epitope by HLA-restricted T-cell clones, similar to previous report conducted on a smaller series of patients 123 that observed for AMA. PDC-E2163–176-specific CD4 þ T- and controls, which supported an innate mechanism cell clones also react with other AMA-specific mitochon- in the involvement of bacteria in the pathogenesis of drial autoantigens, including the E2 subunits of OGDC PBC. These data have also further implications based on and BCOADC, as well as E3BP.112,113 In all these animal experiments. Masanaga et al124 induced auto- molecules, interestingly, the epitopes include a lipoylated immune cholangitis in mice by immunization with LPS domain. The study of TCR b gene rearrangements of cell and the mitochondrial autoantigens identified in PBC. clones obtained from PBC liver biopsies demonstrated The proposed role of innate alterations in PBC is different degrees of oligoclonality of the infiltrating T supported also by a recent study demonstrating the 114,115 cells. The use of TCR Vb or Jb genes by PDC-E2163– enhanced expression of toll-like receptor 3 (belonging to

176-reactive T-cell clones is also highly diverse with GXG the family of receptors binding PAMPs) in PBC liver cell or GXS motifs found within the complementarity- subpopulations isolated by laser-capture microdissec- determining region 3 (CDR3) of all T-cell clones.116 When tion.125 Other studies have concentrated on specific cell autoreactive T-cell frequencies are compared among types constituting the innate response. Natural killer T different tissues of patients with PBC, PDC-E2163–176- (NKT) CD1d-restricted cells are characterized by the reactive precursor T cells are 100–150-fold more repre- expression of an invariant Va24JaQ TCR and the sented in the hilar lymph nodes and the liver compared secretion of both IFN-g and IL-4 in response to stimula- to the peripheral blood, possibly with changes in tion with either anti-CD3 or a-galactosylceramide (a- different disease stages.112 GalCer) presented by the nonpolymorphic, nonclassical Several lines of experimental evidence117 suggest that MHC class I-like CD1 molecules.126 Although NKT cells CD8 þ cytotoxic T cells (CTL) participate in the immuno- do not strictly belong to the innate compartment, their pathogenesis of PBC. CTL lines were generated from innate-like pattern and magnitude of activation indicate patients with PBC using autologous immature dendritic their inclusion in this part of our review. Human NKT cells pulsed with peptides selected by analyzing the cells have been implicated in a variety of immune mitochondrial antigens for amino-acid sequences con- responses, including human autoimmune diseases as taining HLA-A*0201-binding motifs;118 further, these well as their murine models.127 NKT cells were quantita- CTL lines produced IFN-g after restimulation with tively studied in PBC using a tetramer composed of PDC-E2 inner lipoylated domain. In particular, Kita human CD1d and a-GalCer,126 and tetramer-restricted et al118 indicate that the PDC-E2 amino-acid motif cells were significantly more represented in liver tissue at positions 159–167 is an HLA-A*0201-restricted CTL compared to peripheral blood in PBC. Moreover, the epitope, and that autoreactive PDC-E2159–167-specific frequency of NKT cells was higher in both these CD8 þ CTL are present in the peripheral blood of PBC compartments when patients with PBC were compared patients. The same authors then used PDC-E2159–167- to controls. At immunohistochemistry, NKT cells were HLA-A*0201 tetramers and later reported that the commonly observed within portal tracts and liver frequency of PDC-E2159–167-specific CTL is approximately parenchyma in patients with PBC or other liver diseases 10-fold higher in the liver than in the peripheral blood of and healthy controls.128 The involvement of natural patients with PBC, and that the precursor frequency of immunity can be addressed as well, due to an activation

PDC-E2159–167-specific CTL in the peripheral blood was of Toll-like Receptor 9 (TLR9) in B cells if stimulated by enhanced in early PBC stages.118 Another HLA-A*0201- bacterial DNA CpG islands,129 exiting in an increased restricted CTL epitope was identified in a different IgM production, a common unexplained finding in PBC sequence (amino acids 165–174) located within the inner sera. These data, not secondary on the presence of TLR9 lipoyl domain of PDC-E2 and partially overlapping with genetic polymorphisms,130 also support the role of the previously described CTL and CD4 þ T-cell epitopes.119 bacteria in the induction of PBC. No data are currently available regarding other cells of the innate compart- Innate immunity. Most research efforts have so far ment, such as monocytes/macrophages or polymorpho- focused on the characterization of the adaptive immune nuclear cells. response in patients with PBC. Important links between the adaptive and innate immune compartments have Cytokines been recently described, thus making the study of the The cytokine pattern characteristic of PBC is still innate response critical to the understanding of auto- debated. When tested by in situ hybridization, the PBC immunity.120 For example, macrophages are capable of liver presents a significantly higher frequency of IFN-g-

Genes and Immunity Genes and immunity of PBC C Selmi et al 550 and IL-4 mRNA-positive cells compared to controls with, respiratory chain enzymes such as PDC) has been on the other hand, considerably fewer cells, with reported for several viral and bacterial species, including detectable levels of IL-4 mRNA than cells expressing Escherichia coli (the most widely studied species).141 IFN-g mRNA.131 Moreover, IFN-g mRNA-positive cells Interestingly, all these strains presented an ExDK motif are detected primarily around damaged bile ducts and in within their mimicry epitopes. Once again, these cross- areas of piecemeal necrosis. We also note that mitogen- reactivities might be secondary to the highly conserved stimulated T cells infiltrating the liver of PBC produce sequence of PDC across all species, from eubacteria to significantly higher levels of IL-4 and IL-10 compared to mammals,142 while the observation that human mito- control T cells, but little IFN-g.132 A consistent upregula- chondria likely originate from uptaken bacteria prompts tion of IFN-g mRNA expression is also demonstrated by further interest. Recently, we have proposed that the the study of RNA extracted from PBC liver.133–135 Overall, newly defined bacterial strain Novosphingobium aromati- these results appear to indicate that type 1 cytokines are civorans should be regarded as the best candidate yet for the dominant pattern in PBC. However, an upregulation the induction of PBC through molecular mimicry. Several of specific type 2 cytokines, such as IL-5, IL-6, and IL-10, lines of evidence support our theory. First, the amino- in PBC has also been postulated with conflicting acid sequences of four different proteins from N. results.133–135 The mechanisms leading to T-cell recruit- aromaticivorans present one of the highest degrees of ment in the liver of patients with PBC remain enigmatic. homology with the main AMA autoepitopes.143 This has Both selective recruitment and selective retention appear been confirmed by the cloning, sequencing, and expres- to contribute to the resulting tissue enrichment. It is also sion of these four proteins by Padgett et al144 that also possible that, at different stages, the migration of T cells defined the recombinant proteins using well-defined into the liver is controlled by different pathways. Further, monoclonal antibodies. Second, although first isolated a process similar to that recently observed for a closely from sewage samples, N. aromaticivorans can be found in related disease such as primary sclerosing cholangitis several environmental media, including water. Further, should be verified in PBC.136 the bacterium does not cause disease in humans. Interestingly, moreover, a bacterium sharing a sequence homology over 99% with N. aromaticivorans has been Mechanisms of tolerance breakdown and demonstrated to be capable of modulating 17b-estra- pathogenesis diol.145 Lastly, and perhaps most importantly, the bacterium can metabolize a growing number of chemi- In a dissertation about pathogenesis theories proposed in cals (xenobiotics);146 for this reason, it is under study for PBC, it should be remembered that all such theories are bioremediation purposes. A role for halogenated xeno- not independent from individual susceptibility (ie biotics in the induction of PBC has been supported by genetic background) and in some cases are not mutually several studies from our group since the first report in exclusive. According to this view, PBC would then result 2001.147 Based on these lines of evidence, we hypothe- from the interplay between genetic predisposition and sized that N. aromaticivorans or other strains sharing its one or more environmental factors leading to tolerance major characteristics might be the missing link between breakdown, and one or more of the proposed pathoge- xenobiotics and bacteria in the induction of PBC. Our netic mechanisms leading to tissue-specific chronic serological study demonstrated a reactivity against two inflammation and injury. Further, recent evidence on lipoylated proteins of 47 and 50 kDa molecular weight in the possible mechanisms of tolerance breakdown has 100% of anti-PDC-E2-positive and in a fraction of AMA- been provided using the murine strain SJL/J.137 negative sera, regardless of the disease stage. Further, the Although we maintain some concerns about the speci- immunoblotting positivity pattern observed was highly ficity of this model,138 the data produced using these specific for PBC sera compared to sera from first-degree mice are intriguing and deserve to be discussed in the relatives of patients or controls with other autoimmune following sections. diseases. Importantly, the reactivity was 100- to 1000-fold higher than against E. coli, and still detectable at Molecular mimicry dilutions as high as 10À6 in 23% of anti-PDC-E2-positive Molecular mimicry is the mechanism mediating auto- sera.143 We also searched for the presence of the bacterial immunity induction by environmental factors that has genome in feces from patients with PBC and non-PBC received most research attention.139 Briefly, we can define controls living in the same household, and were able to molecular mimicry as the sharing of epitopes on proteins detect evidence of the specific 16S rRNA in approxi- from unrelated species that elicit an immune response mately 25% of patients and controls. For purpose of crossreacting with human self antigens. In most cases, completeness, we also note that researchers from one direct proof of causality cannot be achieved and hence group have suggested that the presence of a novel mere crossreactivity cannot be ruled out, particularly human beta retrovirus in the liver and the lymph nodes when an animal model is not obtained. However, might induce PBC, and reported that the culture of evidence of the possibility of this mechanism in PBC normal biliary epithelial cells (BECs) in the presence of a has been obtained in the SJL/J murine model by Jones homogenate of PBC lymph nodes induced the expression et al,140 who were able to induce a humoral and cellular of a PDC-E2-like antigen on cell membrane.148 Although autoimmune response by coimmunizing mice with both intriguing, we could not recapitulate this retroviral self and nonself (bovine) PDC-E2. Further, the possibility hypothesis.149 Lastly, we note that auxiliary mechanisms that molecular mimicry might lead to AMA and/or have been postulated that might enhance the response to autoreactive T-cell appearance has been widely investi- infectious agents, thus making mimicking proteins more gated, and humoral and cellular crossreactivity with immunogenic or the host more sensitive towards an prokaryotic antigens (particularly with the microbial aberrant immune response. A similar process can be

Genes and Immunity Genes and immunity of PBC C Selmi et al 551 hypothesized from data obtained using bacterial CpG mRNA expression analyses.163 Indeed, the hypothesis motifs that can accelerate the breakdown of tolerance in that these observations underlie an active role for BECs SJL/J mice,150 as well as induce the characteristic hyper- in their own destruction remains fascinating, yet to be IgM observed in PBC sera.129 proven. The role of xenobiotics in the scenario of molecular The role of apoptosis in bile duct loss in PBC has been mimicry leading to PBC warrants further discussion. As widely investigated, and, in most cases, data have mentioned above, xenobiotics are chemical compounds indicated programmed cell death as a prominent which might either modify or bind to self-molecules, mechanism involved in histological lesions of PBC. First, thus altering their recognition by the immune system Harada et al164 reported that BECs from patients with and possibly leading to the ultimate aggression of the PBC are prone to apoptosis via the perforine/granzyme similar native molecule. The presence of this molecule, in B pathway, and this process is enhanced by CD95/ turn, is responsible for the chronic autoimmune reaction CD95L binding. Further, BECs undergoing apoptosis and injury in a self-alimenting process; this also implies overexpress CD40, Fas, and FasL.165 Critical evidence for that the trigger molecule might be no longer present at the role of apoptosis in PBC came from Joseph Odin the moment of diagnosis. The main detoxifying organ is et al,166 who demonstrated the glutathionylation of the the liver, and this makes hepatocytes and BECs particu- lysine lipoic acid moiety of PDC-E2 to be reduced by larly prone to the action of ingested, injected, and patient AMA and that this process is related to the inhaled chemicals. This paradigm is illustrated by an expression of Bcl-2. Bcl-2 concentration in the cell was example. The appearance of serum autoantibodies also correlated with the observation that PCD-E2 is following the intravenous administration of halotane, released intact from caspase cleavage in cells undergoing an inhalatory narcotic no longer used in anesthesiology, apoptosis, thus preserving its immunogenicity.119 We was described,151 and these antibodies crossreacted with should also note that human and rat BECs express higher human PDC-E2. These findings seem to indicate that levels of Bcl-2 compared to other cells.167,168 chemically induced modifications of the lipoylated PBC has been defined as a mucosal disease, based on domains of PDC-E2 could generate antibody production. the common involvement of other epithelia (mainly In 2001, Long et al147 replaced the inner lipoyl domain of salivary glands) besides BECs. BECs, similar to other PDC-E2 molecules with chemicals and tested the epithelial cells, secrete IgA and AMA-IgA have been serologic reactivity of patients with PBC, and reported detected in bile and other body fluids (including urine that serum AMA recognized three of the 18 novel and saliva) from patients with PBC to react with PDC- molecules to a significantly higher extent than the native E2.169,170 More recently, our group demonstrated the form. Further, our group demonstrated that the immu- colocalization of AMA-IgA and PDC-E2, or a mimicking nization of an animal model with a halogenated molecule, on the apical surface and in the cytoplasm of compound conjugated with a non-PDC-E2 amino-acid PBC-BECs.171 To assess the direct pathogenicity of AMA- backbone led to serum AMA appearance,152 and such IgA to BECs, Matsumura et al172 incubated highly serological finding was reversible when immunization purified AMA- IgAs with canine kidney cells transfected was discontinued.153 Importantly, the immunization of with the human polymeric Ig receptor, inducing caspase SJL/J mice with covalently modified self PDC-E2 could upregulation and thus providing evidence of the direct elicit the breakdown of both T- and B-cell tolerance.154 proapoptotic effect. Lastly, and most recently, one common everyday The presence of aberrantly expressed PDC-E2-like additive, 2-octynoic acid, was identified by Amano antigens on the cell membrane of affected tissues was et al155 as a common antigen in PBC sera. described in several studies; however, its precise definition and nature remain enigmatic. In fact, the mimicking Pathogenesis of the tissue injury protein is neither over-expressed in PBC liver, as shown The major paradox of PBC is that the damage is highly by in situ hybridization of PDC-E2 mRNA,131 nor has localized and targets BECs lining only the small and their expression on BEC membranes been definitely medium-sized intrahepatic bile ducts (and the salivary identified.173 An apoptosis-induced modification can be gland epithelial cells in patients with also Sjogren’s hypothesized in the mechanism of apical expression, syndrome) despite the ubiquitous expression of the possibly mediated by a resulting alteration in cellular autoantigens. To address this issue, many efforts have trafficking. been carried out to determine the unique features of this cell population. In fact, the epithelium of affected bile ducts presents a distinct morphology if compared to Concluding remarks and future directions their larger counterparts,156 also presenting different reactivity patterns when challenged with proinflamma- The genetic and immunological bases leading to PBC tory cytokines.157 Further characterization indicated that, onset, perpetuation, and progression remain enigmatic, following stimulation with proinflammatory cytokines, despite the growing body of evidence produced in the BECs from patients and controls share the capability to past decade. The autoimmune paradigm mutuated from increase the expression of the adhesion molecules ICAM- other conditions seems to apply only in part to PBC, 1, as well as the expression of MHC class I and II, TNF-a, based, for example, on the lack of a consistent association INF-g, and IL-1158 or VCAM-1 and LFA-1.159,160 In this with MHC haplotypes or the inefficacy of immunosup- case, BECs are similar to endothelial cells stimulated pressive treatments. To further complicate the picture, with atherogenic factors.161 BECs could indeed act as we cannot rule out at present that the genetic and antigen-presenting cells, since they express secondary etiological factors might be numerous, possibly specific costimulatory signals for T cells such as CD28, B7-1, or to geographical areas, and separated from disease onset B7-2;162 these data, however, were not supported by by a long period of time. The possible approaches to

Genes and Immunity Genes and immunity of PBC C Selmi et al 552 unraveling the genetic factors are several,174 but in all 18 Galbraith RM, Smith M, Mackenzie RM, Tee DE, Doniach D, cases the common factor must be an international Williams R. High prevalence of seroimmunologic abnormal- collaboration effort to collect a large number of cases ities in relatives of patients with active chronic hepatitis or and controls or representative families. As for a better primary biliary cirrhosis. N Engl J Med 1974; 290: 63–69. understanding of the immunopathogenesis of PBC, the 19 Watt FE, James OF, Jones DE. Patterns of autoimmunity in discovery of representative animal models should be the primary biliary cirrhosis patients and their families: a priority to confirm existing or future theories. In population-based cohort study. QJM 2004; 97: 397–406. particular, we encourage scientists from other disciplines 20 Bittencourt PL, Farias AQ, Abrantes-Lemos CP et al. Pre- valence of immune disturbances and chronic liver disease in to participate in our effort and to investigate their animal family members of patients with primary biliary cirrhosis. models for serum or histological signs of autoimmune J Gastroenterol Hepatol 2004; 19: 873–878. cholestasis. Lastly, researchers in PBC should always be 21 MacGregor AJ, Snieder H, Rigby AS et al. Characterizing the well aware of the clinical features of the disease, since quantitative genetic contribution to rheumatoid arthritis their peculiarities might hold the key to the solution of using data from twins. Arthritis Rheum 2000; 43: 30–37. the enigma. In conclusion, we have gone a long way in 22 Greco L, Romino R, Coto I et al. The first large population the study of PBC in the past years. The answer might be based twin study of coeliac disease. Gut 2002; 50: 624–628. already out there; what remains to do is ask the right 23 Chohan MR. Primary biliary cirrhosis in twin sisters. Gut question to the right person. 1973; 14: 213–214. 24 Kaplan MM, Rabson AR, Lee YM, Williams DL, Montaperto PA. Discordant occurrence of primary biliary cirrhosis in monozygotic twins. N Engl J Med 1994; 331: 951. References 25 Selmi C, Mayo MJ, Bach N et al. Primary biliary cirrhosis in monozygotic and dizygotic twins: genetics, epigenetics, and 1 Talwalkar JA, Lindor KD. Primary biliary cirrhosis. Lancet environment. Gastroenterology 2004; 127: 485–492. 2003; 362: 53–61. 26 Selmi C, Gershwin ME. Bacteria and human autoimmunity: 2 Risch N. Searching for genes in complex diseases: lessons the case of primary biliary cirrhosis. Curr Opin Rheumatol from systemic lupus erythematosus. J Clin Invest 2000; 105: 2004; 16: 406–410. 1503–1506. 27 Kita H, He XS, Gershwin ME. Autoimmunity and environ- 3 Kim WR, Lindor KD, Locke III GR et al. Epidemiology and mental factors in the pathogenesis of primary biliary natural history of primary biliary cirrhosis in a US commu- cirrhosis. Ann Med 2004; 36: 72–80. nity. Gastroenterology 2000; 119: 1631–1636. 28 McCluskey J, Peh CA. The human leucocyte antigens 4 Cardon LR, Bell JI. Association study designs for complex and clinical medicine: an overview. Rev Immunogenet 1999; diseases. Nat Rev Genet 2001; 2: 91–99. 1: 3–20. 5 Hirschhorn JN, Lohmueller K, Byrne E, Hirschhorn K. A 29 Ercilla G, Pare´s A, Arriaga F et al. Primary biliary cirrhosis comprehensive review of genetic association studies. Genet associated with DLA-DRw3. Tissue Antigens 1979; 14: 449–452. Med 2002; 4: 45–61. 30 Miyamori H, Kato Y, Kobayashi K, Hattori N. HLA antigens 6 Sabatti C, Service S, Freimer N. False discovery rate in in Japanese patients with primary biliary cirrhosis and linkage and association genome screens for complex dis- autoimmune hepatitis. Digestion 1983; 26: 213–217. orders. Genetics 2003; 164: 829–833. 31 Onishi S, Sakamaki T, Maeda T et al. DNA typing of HLA class 7 Brind AM, Bray GP, Portmann BC, Williams R. Prevalence II genes; DRB1*0803 increases the susceptibility of Japanese to and pattern of familial disease in primary biliary cirrhosis. primary biliary cirrhosis. JHepatol1994; 21: 1053–1060. Gut 1995; 36: 615–617. 32 Bassendine MF, Dewar PJ, James OF. HLA-DR antigens in 8 Hamlyn AN, Macklon AF, James O. Primary biliary cirrhosis: primary biliary cirrhosis: lack of association. Gut 1985; 26: geographical clustering and symptomatic onset seasonality. 625–628. Gut 1983; 24: 940–945. 33 Briggs DC, Donaldson PT, Hayes P, Welsh KI, Williams R, 9 Myszor M, James OF. The epidemiology of primary biliary Neuberger JM. A major histocompatibility complex class III cirrhosis in north-east England: an increasingly common allotype (C4B 2) associated with primary biliary cirrhosis disease? QJMed1990; 75: 377–385. (PBC). Tissue Antigens 1987; 29: 141–145. 10 Bach N, Schaffner F. Familial primary biliary cirrhosis. J 34 Manns MP, Bremm A, Schneider PM et al. HLA DRw8 and Hepatol 1994; 20: 698–701. complement C4 deficiency as risk factors in primary biliary 11 Danielsson A, Boqvist L, Uddenfeldt P. Epidemiology of cirrhosis. Gastroenterology 1991; 101: 1367–1373. primary biliary cirrhosis in a defined rural population in the 35 Morling N, Dalhoff K, Fugger L et al. DNA polymorphism of northern part of Sweden. Hepatology 1990; 11: 458–464. HLA class II genes in primary biliary cirrhosis. Immunoge- 12 Floreani A, Naccarato R, Chiaramonte M. Prevalence of netics 1992; 35: 112–116. familial disease in primary biliary cirrhosis in Italy. J Hepatol 36 Invernizzi P, Battezzati PM, Crosignani A et al. Peculiar HLA 1997; 26: 737–738. polymorphisms in Italian patients with primary biliary 13 Tsuji K, Watanabe Y, Van De Water J et al. Familial primary cirrhosis. J Hepatol 2003; 38: 401–406. biliary cirrhosis in Hiroshima. J Autoimmun 1999; 13: 171–178. 37 Gores GJ, Moore SB, Fisher LD, Powell FC, Dickson ER. 14 Jones DE, Watt FE, Metcalf JV, Bassendine MF, James OF. Primary biliary cirrhosis: associations with class II major Familial primary biliary cirrhosis reassessed: a geographi- histocompatibility complex antigens. Hepatology 1987; 7: 889– cally-based population study. J Hepatol 1999; 30: 402–407. 892. 15 Jones DE, Watt FE, Grove J et al. Tumour necrosis factor-alpha 38 Begovich AB, Klitz W, Moonsamy PV, Van de Water J, Peltz promoter polymorphisms in primary biliary cirrhosis (see G, Gershwin ME. Genes within the HLA class II region comments). J Hepatol 1999; 30: 232–236. confer both predisposition and resistance to primary biliary 16 Olafsson S, Gudjonsson H, Selmi C et al. Antimitochondrial cirrhosis. Tissue Antigens 1994; 43: 71–77. antibodies and reactivity to N. aromaticivorans proteins in 39 Wassmuth R, Depner F, Danielsson A et al. HLA class II Icelandic patients with primary biliary cirrhosis and their markers and clinical heterogeneity in Swedish patients with relatives. Am J Gastroenterol 2004; 99: 2143–2146. primary biliary cirrhosis. Tissue Antigens 2002; 59: 381–387. 17 Feizi T, Naccarato R, Sherlock S, Doniach D. Mitochondrial 40 Mullarkey ME, Stevens AM, McDonnell WM et al. Human and other tissue antibodies in relatives of patients with leukocyte antigen class II alleles in Caucasian women with primary biliary cirrhosis. Clin Exp Immunol 1972; 10: 609–622. primary biliary cirrhosis. Tissue Antigens 2005; 65: 199–205.

Genes and Immunity Genes and immunity of PBC C Selmi et al 553 41 Donaldson P, Agarwal K, Craggs A, Craig W, James O, Jones 61 Lakatos LP, Bajnok E, Hegedus D, Toth T, Lakatos P, Szalay F. D. HLA and interleukin 1 gene polymorphisms in primary Vitamin D receptor, oestrogen receptor-alpha gene and biliary cirrhosis: associations with disease progression and interleukin-1 receptor antagonist gene polymorphisms in disease susceptibility. Gut 2001; 48: 397–402. Hungarian patients with primary biliary cirrhosis. Eur J 42 Zhang L, Weetman AP, Bassendine M, Oliveira DB. Major Gastroenterol Hepatol 2002; 14: 733–740. histocompatibility complex class-II alleles in primary biliary 62 Fan LY, Zhu Y, Zhong RQ et al. Genetic association between cirrhosis. Scand J Immunol 1994; 39: 104–106. interleukins gene polymorphisms with primary biliary 43 Bittencourt PL, Palacios SA, Farias AQ et al. Analysis of cirrhosis in Chinese population. Zhongguo Yi Xue Ke Xue major histocompatibility complex and CTLA-4 alleles in Yuan Xue Bao 2004; 26: 505–509. Brazilian patients with primary biliary cirrhosis. J Gastro- 63 Fan LY, Tu XQ, Zhu Y et al. Genetic association of cytokines enterol Hepatol 2003; 18: 1061–1066. polymorphisms with autoimmune hepatitis and primary 44 Mella JG, Roschmann E, Maier KP, Volk BA. Association of biliary cirrhosis in the Chinese. World J Gastroenterol 2005; 11: primary biliary cirrhosis with the allele HLA-DPB1*0301 in a 2768–2772. German population. Hepatology 1995; 21: 398–402. 64 Matsushita M, Tanaka A, Kikuchi K et al. Association of 45 Seki T, Kiyosawa K, Ota M et al. Association of primary single nucleotide polymorphisms of the interleukin-10 biliary cirrhosis with human leukocyte antigen DPB1*0501 in promoter gene and susceptibility to primary biliary cirrhosis: Japanese patients. Hepatology 1993; 18: 73–78. immunogenetic differences in Italian and Japanese patients. 46 Gregory WL, Daly AK, Dunn AN et al. Analysis of HLA- Autoimmunity 2002; 35: 531–536. class-II-encoded antigen-processing genes TAP1 and TAP2 in 65 Vogel A, Strassburg CP, Manns MP. Genetic association of primary biliary cirrhosis. QJMed1994; 87: 237–244. vitamin D receptor polymorphisms with primary biliary 47 Donaldson PT. TNF gene polymorphisms in primary biliary cirrhosis and autoimmune hepatitis. Hepatology 2002; 35: cirrhosis: a critical appraisal. J Hepatol 1999; 31: 366–368. 126–131. 48 Wilson AG, di Giovine FS, Blakemore AI, Duff GW. Single 66 Halmos B, Szalay F, Cserniczky T et al. Association of primary base polymorphism in the human tumour necrosis factor biliary cirrhosis with vitamin D receptor BsmI genotype alpha (TNF alpha) gene detectable by NcoI restriction of PCR polymorphism in a Hungarian population. Dig Dis Sci 2000; product. Hum Mol Genet 1992; 1: 353. 45: 1091–1095. 49 Wilson AG, Symons JA, McDowell TL, McDevitt HO, Duff 67 Fan LY, Zhong RQ, Tu XQ et al. Genetic association of vitamin GW. Effects of a polymorphism in the human tumor necrosis D receptor polymorphisms with primary biliary cirrhosis factor alpha promoter on transcriptional activation. Proc Natl and autoimmune liver diseases on Chinese. Zhonghua Yi Xue Acad Sci USA 1997; 94: 3195–3199. Za Zhi 2003; 83: 1852–1855. 50 Kroeger KM, Carville KS, Abraham LJ. The À308 tumor 68 Springer JE, Cole DE, Rubin LA et al. Vitamin D-receptor necrosis factor-alpha promoter polymorphism effects tran- genotypes as independent genetic predictors of decreased scription. Mol Immunol 1997; 34: 391–399. bone mineral density in primary biliary cirrhosis. Gastro- 51 Wu WS, McClain KL. DNA polymorphisms and mutations of enterology 2000; 118: 145–151. the tumor necrosis factor-alpha (TNF-alpha) promoter in 69 Pares A, Guanabens N, Alvarez L et al. Collagen type Ialpha1 Langerhans cell histiocytosis (LCH). J Interferon Cytokine Res and vitamin D receptor gene polymorphisms and bone mass 1997; 17: 631–635. in primary biliary cirrhosis. Hepatology 2001; 33: 554–560. 52 Gordon MA, Oppenheim E, Camp NJ, di Giovine FS, Duff 70 Vong S, Bell BP. Chronic liver disease mortality in the United GW, Gleeson D. Primary biliary cirrhosis shows association States, 1990-1998. Hepatology 2004; 39: 476–483. with genetic polymorphism of tumour necrosis factor alpha 71 Pauli-Magnus C, Kerb R, Fattinger K et al. BSEP and MDR3 promoter region (see comments). J Hepatol 1999; 31: 242–247. haplotype structure in healthy Caucasians, primary biliary 53 Tanaka A, Quaranta S, Mattalia A et al. The tumor necrosis cirrhosis and primary sclerosing cholangitis. Hepatology 2004; factor-alpha promoter correlates with progression of primary 39: 779–791. biliary cirrhosis. J Hepatol 1999; 30: 826–829. 72 Pauli-Magnus C, Lang T, Meier Y et al. Sequence analysis of 54 Dickson ER, Grambsch PM, Fleming TR, Fisher LD, bile salt export pump (ABCB11) and multidrug resistance p- Langworthy A. Prognosis in primary biliary cirrhosis: model glycoprotein 3 (ABCB4, MDR3) in patients with intrahepatic for decision making. Hepatology 1989; 10: 1–7. cholestasis of pregnancy. Pharmacogenetics 2004; 14: 91–102. 55 Bathgate AJ, Pravica V, Perrey C, Hayes PC, Hutchinson IV. 73 Selmi C, Invernizzi P, Keefe EB et al. Epidemiology and Polymorphisms in tumour necrosis factor alpha, interleukin- pathogenesis of primary biliary cirrhosis. J Clin Gastroenterol 10 and transforming growth factor beta1 genes and end-stage 2004; 38: 264–271. liver disease. Eur J Gastroenterol Hepatol 2000; 12: 1329–1333. 74 Kimura Y, Selmi C, Leung PS et al. Genetic polymorphisms 56 Fan LY, Zhong RQ, Tu XQ et al. Genetic association of tumor influencing xenobiotic metabolism and transport in patients necrosis factor (TNF)-alpha polymorphisms with primary with primary biliary cirrhosis. Hepatology 2005; 41: 55–63. biliary cirrhosis and autoimmune liver diseases in a Chinese 75 Higuchi M, Horiuchi T, Kojima T et al. Analysis of CD40 population. Zhonghua Gan Zang Bing Za Zhi 2004; 12: 160–162. ligand gene mutations in patients with primary biliary 57 Jazrawi SF, Zaman A, Muhammad Z et al. Tumor necrosis cirrhosis. Scand J Clin Lab Invest 1998; 58: 429–432. factor-alpha promoter polymorphisms and the risk of 76 Corpechot C, Poupon R. Promoter polymorphism of the rejection after liver transplantation: a case control analysis CD14 endotoxin receptor gene and primary biliary cirrhosis. of 210 donor-recipient pairs. Liver Transplant 2003; 9: 377–382. Hepatology 2002; 35: 242–243. 58 Agarwal K, Jones DE, Daly AK et al. CTLA-4 gene 77 Selmi C, Zuin M, Biondi ML et al. Genetic variants of polymorphism confers susceptibility to primary biliary endothelial nitric oxide synthase in patients with primary cirrhosis (in process citation). J Hepatol 2000; 32: 538–541. biliary cirrhosis: association with disease severity. J Gastro- 59 Fan LY, Tu XQ, Cheng QB et al. Cytotoxic T lymphocyte enterol Hepatol 2003; 18: 1150–1155. associated antigen-4 gene polymorphisms confer suscept- 78 Kikuchi K, Tanaka A, Miyakawa H et al. Eta-1/osteopontin ibility to primary biliary cirrhosis and autoimmune hepatitis genetic polymorphism and primary biliary cirrhosis. Hepatol in Chinese population. World J Gastroenterol 2004; 10: Res 2003; 26: 87–90. 3056–3059. 79 Selmi C, Invernizzi P, Tripputi P et al. T-cell receptor 60 Kershenobich D, Rojkind M, Quiroga A, Alcocer-Varela J. polymorphism in primary biliary cirrhosis. Ann Ital Med Int Effect of colchicine on lymphocyte and monocyte function 2003; 18: 149–153. and its relation to fibroblast proliferation in primary biliary 80 Lakatos PL, Bajnok E, Tornai I et al. Insulin-like growth factor cirrhosis. Hepatology 1990; 11: 205–209. I gene microsatellite repeat, collagen type Ialpha1 gene Sp1

Genes and Immunity Genes and immunity of PBC C Selmi et al 554 polymorphism, and bone disease in primary biliary cirrhosis. 101 Vleggaar FP, van Buuren HR. No prognostic significance of Eur J Gastroenterol Hepatol 2004; 16: 753–759. antimitochondrial antibody profile testing in primary biliary 81 Kikuchi K, Lian ZX, Kimura Y et al. Genetic polymorphisms cirrhosis. Hepatogastroenterology 2004; 51: 937–940. of toll-like receptor 9 influence the immune response to CpG 102 Invernizzi P, Crosignani A, Battezzati PM et al. Comparison and contribute to hyper-IgM in primary biliary cirrhosis. J of the clinical features and clinical course of antimitochon- Autoimmun 2005; 24: 344–352. drial antibody-positive and -negative primary biliary cirrho- 82 Ranke MB, Saenger P. Turner’s syndrome. Lancet 2001; 358: sis. Hepatology 1997; 25: 1090–1095. 309–314. 103 Krams SM, Dorshkind K, Gershwin ME. Generation of 83 Davis CJ, Davison RM, Payne NN, Rodeck CH, Conway GS. biliary lesions after transfer of human lymphocytes into Female sex preponderance for idiopathic familial premature severe combined immunodeficient (SCID) mice. J Exp Med ovarian failure suggests an X chromosome defect: opinion. 1989; 170: 1919–1930. Hum Reprod 2000; 15: 2418–2422. 104 Metcalf JV, Mitchison HC, Palmer JM, Jones DE, Bassendine 84 Invernizzi P, Miozzo M, Battezzati PM et al. Frequency of MF, James OF. Natural history of early primary biliary monosomy X in women with primary biliary cirrhosis. Lancet cirrhosis. Lancet 1996; 348: 1399–1402. 2004; 363: 533–535. 105 Invernizzi P, Selmi C, Ranftler C, Podda M, Wesierska-Gadek 85 Guttenbach M, Koschorz B, Bernthaler U, Grimm T, Schmid J. Antinuclear antibodies in primary biliary cirrhosis. Semin M. Sex chromosome loss and aging: in situ hybridization Liver Dis 2005, in press. studies on human interphase nuclei. Am J Hum Genet 1995; 106 Janka C, Selmi C, Gershwin ME, Will H, Sternsdorf T. Small 57: 1143–1150. ubiquitin-related modifiers: A novel and independent class 86 Kaplan MM, Bianchi DW. Primary biliary cirrhosis: for want of autoantigens in primary biliary cirrhosis. Hepatology 2005; of an X chromosome? Lancet 2004; 363: 505–506. 41: 609–616. 87 Invernizzi P, Miozzo M, Selmi C et al. X chromosome 107 Invernizzi P, Podda M, Battezzati PM et al. Autoantibodies monosomy: a common mechanism for autoimmune diseases. against nuclear pore complexes are associated with more J Immunol 2005; 175: 575–578. active and severe liver disease in primary biliary cirrhosis. 88 Lambert N, Nelson JL. Microchimerism in autoimmune J Hepatol 2001; 34: 366–372. disease: more questions than answers? Autoimmun Rev 108 Invernizzi P, Wesierska-Gadek J, Battezzati PM et al. Prog- 2003; 2: 133–139. nostic value of autoantibodies against proteins of nuclear 89 Fanning PA, Jonsson JR, Clouston AD et al. Detection of male pore complexes (anti-NPCs) in early primary biliary cirrhosis DNA in the liver of female patients with primary biliary (PBC). J Hepatol 2004; 40: 159–160 (abstract). cirrhosis. J Hepatol 2000; 33: 690–695. 109 Hashimoto E, Lindor KD, Homburger HA et al. Immunohis- 90 Tanaka A, Lindor K, Gish R et al. Fetal microchimerism alone tochemical characterization of hepatic lymphocytes in pri- does not contribute to the induction of primary biliary mary biliary cirrhosis in comparison with primary sclerosing cirrhosis. Hepatology 1999; 30: 833–838. cholangitis and autoimmune chronic active hepatitis. Mayo 91 Invernizzi P, De Andreis C, Sirchia SM et al. Blood fetal Clin Proc 1993; 68: 1049–1055. microchimerism in primary biliary cirrhosis. Clin Exp 110 Van de Water J, Ansari A, Prindiville T et al. Heterogeneity of Immunol 2000; 122: 418–422. autoreactive T cell clones specific for the E2 component of the 92 Oertelt S, Selmi C, Invernizzi P, Podda M, Gershwin ME. pyruvate dehydrogenase complex in primary biliary cirrho- Genes and goals: an approach to microarray analysis in sis. J Exp Med 1995; 181: 723–733. autoimmunity. Autoimmun Rev 2005, in press. 111 Shimoda S, Nakamura M, Ishibashi H, Hayashida K, Niho Y. 93 Hinds DA, Stuve LL, Nilsen GB et al. Whole-genome patterns HLA DRB4 0101-restricted immunodominant T cell auto- of common DNA variation in three human populations. epitope of pyruvate dehydrogenase complex in primary Science 2005; 307: 1072–1079. biliary cirrhosis: evidence of molecular mimicry in human 94 Miyakawa H, Tanaka A, Kikuchi K et al. Detection of autoimmune diseases. J Exp Med 1995; 181: 1835–1845. antimitochondrial autoantibodies in immunofluore- 112 Shimoda S, Van de Water J, Ansari A et al. Identification and scent AMA-negative patients with primary biliary cirrhosis precursor frequency analysis of a common T cell epitope using recombinant autoantigens. Hepatology 2001; 34: motif in mitochondrial autoantigens in primary biliary 243–248. cirrhosis. J Clin Invest 1998; 102: 1831–1840. 95 Gershwin ME, Mackay IR, Sturgess A, Coppel RL. Identifica- 113 Shimoda S, Nakamura M, Ishibashi H et al. Molecular tion and specificity of a cDNA encoding the 70 kd mitochon- mimicry of mitochondrial and nuclear autoantigens in drial antigen recognized in primary biliary cirrhosis. J primary biliary cirrhosis. Gastroenterology 2003; 124: 1915– Immunol 1987; 138: 3525–3531. 1925. 96 Gershwin ME, Ansari AA, Mackay IR et al. Primary biliary 114 Moebius U, Manns M, Hess G, Kober G, Meyer cirrhosis: an orchestrated immune response against epithelial zum Bu¨ schenfelde KH, Meuer SC. T cell receptor gene cells. Immunol Rev 2000; 174: 210–225. rearrangements of T lymphocytes infiltrating the liver in 97 Bruggraber SF, Leung PS, Amano K et al. Autoreactivity to chronic active hepatitis B and primary biliary cirrhosis (PBC): lipoate and a conjugated form of lipoate in primary biliary oligoclonality of PBC-derived T cell clones. Eur J Immunol cirrhosis. Gastroenterology 2003; 125: 1705–1713. 1990; 20: 889–896. 98 Mao TK, Davis PA, Odin JA, Coppel RL, Gershwin ME. 115 Tsai SL, Lai MY, Chen DS. Analysis of rearranged T cell Sidechain biology and the immunogenicity of PDC-E2, the receptor (TCR) Vb transcripts in livers of primary biliary major autoantigen of primary biliary cirrhosis. Hepatology cirrhosis: preferential Vb usage suggests antigen-driven 2004; 40: 1241–1248. selection. Clin Exp Immunol 1996; 103: 99–104. 99 Quinn J, Diamond AG, Palmer JM, Bassendine MF, James OF, 116 Ichiki Y, Shimoda S, Hara H et al. Analysis of T-cell receptor Yeaman SJ. Lipoylated and unlipoylated domains of beta of the T-cell clones reactive to the human PDC-E2 163- human PDC-E2 as autoantigens in primary biliary cirrhosis: 176 peptide in the context of HLA-DR53 in patients with significance of lipoate attachment. Hepatology 1993; 18: primary biliary cirrhosis. Hepatology 1997; 26: 728–733. 1384–1391. 117 Meuer SC, Moebius U, Manns MM et al. Clonal analysis of 100 Van Norstrand MD, Malinchoc M, Lindor KD et al. Quanti- human T lymphocytes infiltrating the liver in chronic active tative measurement of autoantibodies to recombinant mito- hepatitis B and primary biliary cirrhosis. Eur J Immunol 1988; chondrial antigens in patients with primary biliary cirrhosis: 18: 1447–1452. relationship of levels of autoantibodies to disease progres- 118 Kita H, Lian ZX, Van de Water J et al. Identification of HLA- sion. Hepatology 1997; 25: 6–11. A2-restricted CD8(+) cytotoxic T cell responses in primary

Genes and Immunity Genes and immunity of PBC C Selmi et al 555 biliary cirrhosis: T cell activation is augmented by immune in experimental autoimmune cholangitis: a mouse model of complexes cross-presented by dendritic cells. J Exp Med 2002; primary biliary cirrhosis. Hepatology 1999; 30: 65–70. 195: 113–123. 138 Sasaki M, Allina J, Odin JA et al. Autoimmune cholangitis in 119 Matsumura S, Van De Water J, Kita H et al. Contribution to the SJL/J mouse is antigen non-specific. Dev Immunol 2002; 9: antimitochondrial antibody production: cleavage of pyruvate 103–111. dehydrogenase complex-E2 by apoptosis-related proteases. 139 Christen U, von Herrath MG. Induction, acceleration or Hepatology 2002; 35: 14–22. prevention of autoimmunity by molecular mimicry. Mol 120 Bachmann MF, Kopf M. On the role of the innate immunity Immunol 2004; 40: 1113–1120. in autoimmune disease. J Exp Med 2001; 193: F47–F50. 140 Jones DE, Palmer JM, Bennett K et al. Investigation of 121 Ballot E, Bandin O, Chazouilleres O, Johanet C, Poupon R. a mechanism for accelerated breakdown of immune toler- Immune response to lipopolysaccharide in primary biliary ance to the primary biliary cirrhosis-associated autoantigen, cirrhosis and autoimmune diseases. J Autoimmun 2004; 22: pyruvate dehydrogenase complex. Lab Invest 2002; 82: 153–158. 211–219. 122 Sasatomi K, Noguchi K, Sakisaka S, Sata M, Tanikawa K. 141 Shimoda S, Nakamura M, Shigematsu H et al. Mimicry Abnormal accumulation of endotoxin in biliary epithelial peptides of human PDC-E2 163-176 peptide, the immuno- cells in primary biliary cirrhosis and primary sclerosing dominant T-cell epitope of primary biliary cirrhosis. Hepatol- cholangitis. J Hepatol 1998; 29: 409–416. ogy 2000; 31: 1212–1216. 123 Ide T, Sata M, Nakano H, Suzuki H, Tanikawa K. Increased 142 Van de Water J, Ishibashi H, Coppel RL, Gershwin ME. serum IgM class anti-lipid A antibody and therapeutic effect Molecular mimicry and primary biliary cirrhosis: premises of ursodeoxycholic acid in primary biliary cirrhosis. Hepato- not promises. Hepatology 2001; 33: 771–775. gastroenterology 1997; 44: 1569–1573. 143 Selmi C, Balkwill DL, Invernizzi P et al. Patients 124 Masanaga T, Watanabe Y, Van de Water J et al. Induction and with primary biliary cirrhosis react against a ubiquitous persistence of immune-mediated cholangiohepatitis in neo- xenobiotic-metabolizing bacterium. Hepatology 2003; 38: natally thymectomized mice. Clin Immunol Immunopathol 1250–1257. 1998; 89: 141–149. 144 Padgett KA, Selmi C, Kenny TP et al. Phylogenetic and 125 Takii Y, Nakamura M, Ito M et al. Enhanced expression of immunological definition of four lipoylated proteins from type I interferon and toll-like receptor-3 in primary biliary Novosphingobium aromaticivorans, implications for primary cirrhosis. Lab Invest 2005; 85: 908–920. biliary cirrhosis. J Autoimmun 2005; 24: 209–219. 126 Kita H, Naidenko OV, Kronenberg M et al. Quantitation and 145 Fujii K, Kikuchi S, Satomi M, Ushio-Sata N, Morita N. phenotypic analysis of natural killer T cells in primary biliary Degradation of 17beta-estradiol by a gram-negative bacter- cirrhosis using a human CD1d tetramer. Gastroenterology ium isolated from activated sludge in a sewage treat- 2002; 123: 1031–1043. ment plant in Tokyo, Japan. Appl Environ Microbiol 2002; 68: 127 Hammond KJ, Kronenberg M. Natural killer T cells: natural 2057–2060. or unnatural regulators of autoimmunity? Curr Opin Immunol 146 Shi F, Ljunggren HG, Sarvetnick N. Innate immunity and 2003; 15: 683–689. autoimmunity: from self-protection to self-destruction. Trends 128 Harada K, Isse K, Tsuneyama K, Ohta H, Nakanuma Y. Immunol 2001; 22: 97–101. Accumulating CD57+ CD3+ natural killer T cells are related 147 Long SA, Quan C, Van de Water J et al. Immunoreactivity of to intrahepatic bile duct lesions in primary biliary cirrhosis. organic mimeotopes of the E2 component of pyruvate Liver Int 2003; 23: 94–100. dehydrogenase: connecting xenobiotics with primary biliary 129 Kikuchi K, Lian ZX, Yang GX et al. Bacterial CpG induces cirrhosis. J Immunol 2001; 167: 2956–2963. hyper-IgM production in CD27(+) memory B cells in primary 148 Xu L, Shen Z, Guo L et al. Does a betaretrovirus infection biliary cirrhosis. Gastroenterology 2005; 128: 304–312. trigger primary biliary cirrhosis? Proc Natl Acad Sci USA 130 Kikuchi K, Lian ZX, Kimura Y et al. Genetic polymorphisms 2003; 100: 8454–8459. of toll-like receptor 9 influence the immune response to CpG 149 Selmi C, Ross SR, Ansari AA et al. Lack of immunological or and contribute to hyper-IgM in primary biliary cirrhosis. J molecular evidence for a role of mouse mammary tumor Autoimmun 2005; in press. retrovirus in primary biliary cirrhosis. Gastroenterology 2004; 131 Harada K, Van de Water J, Leung PS et al. In situ nucleic acid 127: 493–501. hybridization of cytokines in primary biliary cirrhosis: 150 Jones DE, Palmer JM, Burt AD, Walker C, Robe AJ, Kirby JA. predominance of the Th1 subset. Hepatology 1997; 25: 791– Bacterial motif DNA as an adjuvant for the breakdown of 796. immune self-tolerance to pyruvate dehydrogenase complex. 132 Lo¨hr HF, Schlaak JF, Gerken G, Fleischer B, Dienes HP, Meyer Hepatology 2002; 36: 679–686. zum Bu¨ schenfelde KH. Phenotypical analysis and cytokine 151 Hastings KL, Thomas C, Hubbard AK, Gandolfi AJ. Screen- release of liver-infiltrating and peripheral blood T lympho- ing for antibodies associated with halothane hepatitis. Br J cytes from patients with chronic hepatitis of different Anaesth 1991; 67: 722–728. etiology. Liver 1994; 14: 161–166. 152 Leung PS, Quan C, Park O et al. Immunization with a 133 Martinez OM, Villanueva JC, Gershwin ME, Krams SM. xenobiotic 6-bromohexanoate bovine serum albumin con- Cytokine patterns and cytotoxic mediators in primary biliary jugate induces antimitochondrial antibodies. J Immunol 2003; cirrhosis. Hepatology 1995; 21: 113–119. 170: 5326–5332. 134 Shindo M, Mullin GE, Braun-Elwert L, Bergasa NV, Jones EA, 153 Amano K, Leung PS, Xu Q et al. Xenobiotic-induced loss of James SP. Cytokine mRNA expression in the liver of patients tolerance in rabbits to the mitochondrial autoantigen of with primary biliary cirrhosis (PBC) and chronic hepatitis B primary biliary cirrhosis is reversible. J Immunol 2004; 172: (CHB). Clin Exp Immunol 1996; 105: 254–259. 6444–6452. 135 Nagano T, Yamamoto K, Matsumoto S et al. Cytokine profile 154 Palmer JM, Robe AJ, Burt AD, Kirby JA, Jones DE. Covalent in the liver of primary biliary cirrhosis. J Clin Immunol 1999; modification as a mechanism for the breakdown of immune 19: 422–427. tolerance to pyruvate dehydrogenase complex in the mouse. 136 Eksteen B, Grant AJ, Miles A et al. Hepatic endothelial CCL25 Hepatology 2004; 39: 1583–1592. mediates the recruitment of CCR9+ gut-homing lymphocytes 155 Amano K, Leung PS, Rieger R et al. Chemical xenobiotics and to the liver in primary sclerosing cholangitis. J Exp Med 2004; mitochondrial autoantigens in primary biliary cirrhosis: 200: 1511–1517. identification of antibodies against a common environmental, 137 Jones DE, Palmer JM, Yeaman SJ, Kirby JA, Bassendine MF. cosmetic, and food additive, 2-octynoic acid. J Immunol 2005; Breakdown of tolerance to pyruvate dehydrogenase complex 174: 5874–5883.

Genes and Immunity Genes and immunity of PBC C Selmi et al 556 156 Katayanagi K, Kono N, Nakanuma Y. Isolation, culture and in acute and chronic hepatitis, cirrhosis and hepatocellular characterization of biliary epithelial cells from different carcinoma. Pathol Res Pract 1999; 195: 19–24. anatomical levels of the intrahepatic and extrahepatic biliary 169 Tanaka A, Nalbandian G, Leung PS et al. Mucosal immunity tree from a mouse. Liver 1998; 18: 90–98. and primary biliary cirrhosis: presence of antimitochondrial 157 Ueno Y, Alpini G, Yahagi K et al. Evaluation of differential antibodies in urine. Hepatology 2000; 32: 910–915. gene expression by microarray analysis in small and large 170 Reynoso-Paz S, Leung PS, Van De Water J et al. Evidence for a cholangiocytes isolated from normal mice. Liver Int 2003; 23: locally driven mucosal response and the presence of 449–459. mitochondrial antigens in saliva in primary biliary cirrhosis. 158 Ayres RC, Neuberger JM, Shaw J, Joplin R, Adams DH. Hepatology 2000; 31: 24–29. Intercellular adhesion molecule-1 and MHC antigens on 171 Fukushima N, Nalbandian G, Van De Water J et al. human intrahepatic bile duct cells: effect of pro-inflamma- Characterization of recombinant monoclonal IgA anti-PDC- tory cytokines. Gut 1993; 34: 1245–1249. E2 autoantibodies derived from patients with PBC. Hepatol- 159 Dienes HP, Lohse AW, Gerken G et al. Bile duct epithelia as ogy 2002; 36: 1383–1392. target cells in primary biliary cirrhosis and primary scleros- 172 Matsumura S, Van De Water J, Leung P et al. Caspase ing cholangitis. Virchows Arch 1997; 431: 119–124. induction by IgA antimitochondrial antibody: IgA-mediated 160 Yasoshima M, Nakanuma Y, Tsuneyama K, Van de Water J, biliary injury in primary biliary cirrhosis. Hepatology 2004; 39: Gershwin ME. Immunohistochemical analysis of adhesion 1415–1422. molecules in the micro-environment of portal tracts in 173 Migliaccio C, Nishio A, Van de Water J et al. Monoclonal relation to aberrant expression of PDC-E2 and HLA-DR on antibodies to mitochondrial E2 components define autoepi- the bile ducts in primary biliary cirrhosis. J Pathol 1995; 175: topes in primary biliary cirrhosis. J Immunol 1998; 161: 5157– 319–325. 5163. 161 Hansson GK, Jonasson L, Holm J, Claesson-Welsh L. Class II 174 Donaldson PT. Genetics of liver disease: immunogenetics MHC antigen expression in the atherosclerotic plaque: and disease pathogenesis. Gut 2004; 53: 599–608. smooth muscle cells express HLA-DR, HLA-DQ and the 175 Johnston DE, Kaplan MM, Miller KB, Connors CM, Milford invariant gamma chain. Clin Exp Immunol 1986; 64: 261–268. EL. Histocompatibility antigens in primary biliary cirrhosis. 162 Tsuneyama K, Harada K, Yasoshima M, Kaji K, Gershwin Am J Gastroenterol 1987; 82: 1127–1129. ME, Nakanuma Y. Expression of co-stimulatory factor B7-2 176 Prochazka EJ, Terasaki PI, Park MS, Goldstein LI, Busuttil on the intrahepatic bile ducts in primary biliary cirrhosis and RW. Association of primary sclerosing cholangitis with HLA- primary sclerosing cholangitis: an immunohistochemical DRw52a. N Engl J Med 1990; 322: 1842–1844. study. J Pathol 1998; 186: 126–130. 177 Underhill J, Donaldson P, Bray G, Doherty D, Portmann B, 163 Leon MP, Bassendine MF, Wilson JL, Ali S, Thick M, Kirby Williams R. Susceptibility to primary biliary cirrhosis is JA. Immunogenicity of biliary epithelium: investigation of associated with the HLA-DR8-DQB1*0402 haplotype. Hepa- antigen presentation to CD4+ T cells. Hepatology 1996; 24: tology 1992; 16: 1404–1408. 561–567. 178 Underhill JA, Donaldson PT, Doherty DG, Manabe K, 164 Harada K, Ozaki S, Gershwin ME, Nakanuma Y. Enhanced Williams R. HLA DPB polymorphism in primary sclerosing apoptosis relates to bile duct loss in primary biliary cirrhosis. cholangitis and primary biliary cirrhosis. Hepatology 1995; 21: Hepatology 1997; 26: 1399–1405. 959–962. 165 Afford SC, Ahmed-Choudhury J, Randhawa S et al. CD40 179 Agarwal K, Czaja AJ, Jones DE, Donaldson PT. Cytotoxic T activation-induced, Fas-dependent apoptosis and NF-kap- lymphocyte antigen-4 (CTLA-4) gene polymorphisms and paB/AP-1 signaling in human intrahepatic biliary epithelial susceptibility to type 1 autoimmune hepatitis. Hepatology cells. FASEB J 2001; 15: 2345–2354. 2000; 31: 49–53. 166 Odin JA, Huebert RC, Casciola-Rosen L, LaRusso NF, Rosen 180 Graham AM, Dollinger MM, Howie SE, Harrison DJ. A. Bcl-2-dependent oxidation of pyruvate dehydrogenase-E2, Identification of novel alleles at a polymorphic microsatellite a primary biliary cirrhosis autoantigen, during apoptosis. J repeat region in the human NRAMP1 gene promoter: Clin Invest 2001; 108: 223–232. analysis of allele frequencies in primary biliary cirrhosis. J 167 Charlotte F, L’Hermine A, Martin N et al. Immunohisto- Med Genet 2000; 37: 150–152. chemical detection of bcl-2 protein in normal and patholo- 181 Zappala F, Grove J, Watt FE et al. No evidence for gical human liver. Am J Pathol 1994; 144: 460–465. involvement of the interleukin-10 À592 promoter poly- 168 Nakopoulou L, Stefanaki K, Vourlakou C, Manolaki N, morphism in genetic susceptibility to primary biliary Gakiopoulou H, Michalopoulos G. Bcl-2 protein expression cirrhosis. J Hepatol 1998; 28: 820–823.

Genes and Immunity