Antibodies and primary biliary cirrhosis – piecing together the jigsaw

Abstract

Despite major advances in understanding the epidemiology, natural history and immunology of primary biliary cirrhosis (PBC), both the cause and the definitive treatment remains as elusive in 2002 as it did in 1851 when the syndrome was first described by Addison and Gull. PBC is usually classified as an autoimmune disease, so a clue to the etiology of PBC may lie in the autoantibodies associated with the disease. Autoantibodies may, as in the case of Graves’ disease or myasthenia gravis, mediate the disease or they may be simply phenomenological events [[1]Ermann J. Fathman C.G. Autoimmune diseases: genes, bugs and failed regulation.Nat Immunol. 2001; 2: 799-851Crossref Scopus (153) Google Scholar]1. PBC and AMAThe very close association between PBC and AMA was first recognised by Doniach and colleagues in 1966; the subsequent identification in 1988, of the major antigen recognised by the AMA as the dihydrolipoamide acetyl transferase by Yeaman and by Gershwin was a major achievement [[2]Van de Water J. Gershwin M.E. Leung P. Ansari A. Coppel R.L. The autoepitope of the 74 kD mitochondrial autoantigen of primary biliary cirrhosis corresponds to the functional site of dihydrolipoamide acetyl transferase.J Exp Med. 1988; 167: 1791-1799Crossref PubMed Scopus (233) Google Scholar]. Subsequent studies defined the dominant B and T cell epitope to amino acids 212–226 of the inner lipoyl domain of the E2 component of pyruvate dehydrogenase – a well conserved enzyme located on the inner mitochondrial membrane [[3]Tuaillon N. Andre C. Briand J.P. Penner E. Muller S. A lipoyl synthetic octadecapeptide of dihydrolipoamide acetyltransferase specifically recognised by anti-M2 antibodies in primary biliary cirrhosis.J Immunol. 1992; 148: 445-450PubMed Google Scholar]. Yet, these discoveries have so far done little to help elucidate the pathogenesis of PBC.Several observations may help clarify the implications of the very closed association between the disease and the AMA [4Thomson R.K. Neuberger J.M. PBC and AMA – what is the connection?.Hepatology. 1999; 29: 271-276Crossref PubMed Scopus (85) Google Scholar, 5Haydon G.K. Neuberger J. PBC: an infectious disease?.Gut. 2000; 47: 586-588Crossref PubMed Scopus (71) Google Scholar]:•AMA may be found in normal individuals at low titre;•AMA are found in over 95% patients with PBC;•AMA may also be associated with some infectious diseases, such as tuberculosis or viral infection [[6]Klein R. Weibel M. Engelhardt S. Berg P.A. Sera from patients with tuberculosis recognise the M2a-epitope (E2 sub-unit of pyruvate dehydrogenase) specific for primary biliary cirrhosis.Clin Exp Immmunol. 1993; 92: 308-316Crossref PubMed Scopus (63) Google Scholar];•high titre AMA may precede clinical, biochemical and histological features of PBC;•although E2 is located in all nucleated cells, the immune targets in PBC are the middle-sized biliary epithelial cells and possibly other epithelial cells;•in PBC and normal subjects, there is increased amount of E2 in biliary epithelial cells compared with the density of mitochondria;•although there is over-expression of E2 in BECs in patients with PBC, increased mRNA cannot be detected;•the target antigen is found aberrantly on the plasma membrane on biliary epithelial cells (BECs) both in vivo and in vitro in PBC but in no other condition;•lymphocytes, both liver derived and intra-hepatic, recognise E2;•AMA titre does not correlate with severity or rate of progression of the disease;•treatment of patients with ursodeoxycholic acid (which affects the biochemistry and may slow disease progression) is associated with a reduction or abolition of AMA in serum;•E2 (unlike other mitochondrial enzymes) is not degraded during apoptosis of BECs in vitro [[7]Odin J.A. Huebert R.C. Cascila-Rosen L. LaRusso N.F. Rosen A. Bcl-2-dependent oxidatin of pyruvate dehydrogenase-E2, a primary biliary cirrhosis autoantigen, during apoptosis.J Clin Invest. 2001; 108: 223-232Crossref PubMed Scopus (186) Google Scholar];•recurrent PBC is seen in some but not all liver allograft recipients;•AMA persist after transplantation, and do not correlate with histological evidence of disease recurrence in the allograft;•aberrant membrane expression of E2 by BECs is seen in the allograft of patients grafted for PBC, whether or not there is histological evidence of recurrent disease;•immunisation of animals with recombinant E2 is not associated with histological evidence of PBC although AMA are generated.However, none of these observations really define the relationship between the diagnostic antibodies and disease. BECs, the target cells of PBC, are immunologically active [[8]Howell C.D. Nie X. Antigen presentation by cytokine-stimulated mouse bile duct epithelial cells.Hepatology. 2001; 34 (abstract): 478ACrossref PubMed Scopus (27) Google Scholar]. There is no evidence to date that the AMA are pathogenetic or whether they are true auto-antibodies or antibodies cross-reacting with bacterial, viral or xenobiotic-associated antigens.2. PBC without AMAAutoimmune cholangitis (AIC) resembles PBC with respect to the clinical, serological and histological findings, the natural history and the response to treatment [[9]Heathcote J. Autoimmune cholangitis.Gut. 1997; 40: 440-442PubMed Google Scholar]. The major distinction between patients with autoimmune cholangitis and PBC (other than absence of AMA) is that in AIC, serum IgG and transaminases levels tend to be higher, alkaline phosphatase levels lower, and anti-nuclear antibodies present at greater titres and antSLA/LP antibodies [10Czaja A.J. Carpenter H.A. Santrach P.J. Moore S.B. Autoimmune cholangitis within the spectrum of autoimmune liver disease.Hepatology. 2000; 31: 1231-1238Crossref PubMed Scopus (148) Google Scholar, 11Nakajima M. Shimiza H. Miyazaki A. Watanabe S. Kitami N. Sato N. detection of IGA, IgM and IgG sub-classes of anti-M2 antibody by immunoblotting in autoimmune cholangitis: is autoimmune cholangitis an early stage of primary biliary cirrhosis?.J Gastroenterol. 1999; 34: 657-658Crossref PubMed Scopus (37) Google Scholar, 12Kanzler S. Bozkurt S. Herkel J. Galle P.R. Dienes H.P. Lohse A. Presence of SLA/LP-autoantibodies in patients with primary biliary cirrhosis as a marker of secondary autoimmune hepatitis (overlap syndrome).Hepatology. 2001; 34 (abstract): 377ACrossref PubMed Scopus (59) Google Scholar]. Classification of these patients as autoimmune cholangitis must be done with caution: some of these patients do have AMA, not detected by immunofluorescence but detectable by immunoblotting [11Nakajima M. Shimiza H. Miyazaki A. Watanabe S. Kitami N. Sato N. detection of IGA, IgM and IgG sub-classes of anti-M2 antibody by immunoblotting in autoimmune cholangitis: is autoimmune cholangitis an early stage of primary biliary cirrhosis?.J Gastroenterol. 1999; 34: 657-658Crossref PubMed Scopus (37) Google Scholar, 13Miyakawa H. Tanaka A. Kikuchi K. Mtsushita M. Kitazawa E. Kawaguchi N. et al.Detection of antimitochondrial auto-antibodies in immunofluorescence negative patients with primary biliary cirrhosis using recombinant autoantigens.Hepatology. 2001; 34: 243-248Crossref PubMed Scopus (173) Google Scholar] and others develop AMA during the course of their illness, although some remain resolutely AMA-negative. These patients also show the aberrant expression of E2 in the BECs, similar to that seen in PBC.The inflammatory infiltrate may be phenotypically different, with those with PBC having fewer plasma cells but more lymphocytic cells in the portal inflammatory cell infiltrate [[14]Watanabe S. Deguchi A. Uchida N. Kurokohchi K. Arima K. Nishioka M. et al.Histopathologic comparison of anti-mitochondrial antibody-positive primary biliary cirrhosis and autoimmune cholangiopathy.Hepatol Res. 2001; 19: 41-51Crossref PubMed Scopus (16) Google Scholar]; the functional significance of this difference is unclear. Indeed, Mayo failed to identify any functional differences in oligoclonally expanded T cells isolated from the livers of patients with PBC and AIC [[15]Mayo M.J. Lipsky P.E. Miller S.N. Stastny P. Coombes B. Similar T-cells oligoclonality in anti-mitochondrial antibody-positive and -negative primary biliary cirrhosis.Dig Dis Sci. 2001; 46: 345-351Crossref PubMed Scopus (5) Google Scholar].The data on HLA associations are conflicting [[12]Kanzler S. Bozkurt S. Herkel J. Galle P.R. Dienes H.P. Lohse A. Presence of SLA/LP-autoantibodies in patients with primary biliary cirrhosis as a marker of secondary autoimmune hepatitis (overlap syndrome).Hepatology. 2001; 34 (abstract): 377ACrossref PubMed Scopus (59) Google Scholar]: whereas Czaja [[10]Czaja A.J. Carpenter H.A. Santrach P.J. Moore S.B. Autoimmune cholangitis within the spectrum of autoimmune liver disease.Hepatology. 2000; 31: 1231-1238Crossref PubMed Scopus (148) Google Scholar], looking at 22 patients with AIC found no difference in HLA associations, when compared with PBC. Donaldson suggested that DRB1*0801 was associated with disease progression rather than disease susceptibility [[16]Donaldson P.T. Agarwal K. Craggs A. Craig W. James O.F.W. Jones D. HLA and interleukin 1 gene polymorphisms in primary biliary cirrhosis: association with disease progression and disease susceptibility.Gut. 2001; 48: 397-402Crossref PubMed Scopus (115) Google Scholar]. However, Stone and colleagues [[17]Stone J. Wade J.A. Cauch-Dudek K. Ng C. Lindor K.D. Heathcote E.J. Human leukocyte antigen Class II associations in serum antimitochondrial antibodies (AMA)-positive and AMA-negative primary biliary cirrhosis.J Hepatol. 2002; 36: 8-13Abstract Full Text Full Text PDF PubMed Scopus (45) Google Scholar], in this issue of the Journal, report significant differences between PBC and AIC, with respect to the class II antigens. Although those with PBC had higher frequencies of HLA DRβ1*08 and DQβ1*04, both these antigens were absent in a cohort of 26 patients with AIC, compared with 15% and 14% respectively in 154 patients with PBC.Current evidence suggests no significant differences between the two conditions, implying that AMA are just markers of PBC and are not pathogenetic. Host factors (such as HLA) may therefore affect some of the immunological aspects of PBC such as antibody production or the severity of interface hepatitis.3. Why do patients with PBC have AMA?It still remains uncertain whether AMA are recognising autoantigens or cross-reactive antigens. For example, Long and colleagues [[18]Long S.A. Quan C. Van de Water J. Nantz M. Kurth M. Barsky D. et al.Immunoreactvity of organic mimeotopes of the E2 component of pyruvate dehydrogenase: connecting xenobiotics with primary biliary cirrhosis.J Immunol. 2001; 167: 2956-2963PubMed Google Scholar] have recently shown that modification of the inner lipoyl domain of E2 with halide or ethyl halide, results in increased reactivity of AMA from PBC patients, suggesting that xenobiotics could make cellular components antigenic.An infectious trigger for PBC has been suggested [[5]Haydon G.K. Neuberger J. PBC: an infectious disease?.Gut. 2000; 47: 586-588Crossref PubMed Scopus (71) Google Scholar]: although some evidence for bacteria, mycobacteria and viruses has been published, none has yet been convincingly implicated. There are several mechanisms whereby infection could lead to autoimmunity [[19]Marrack P. Kappler J. Kotzin B.L. Autoimmune disease: why and where it occurs.Nat Med. 2001; 7: 899-905Crossref PubMed Scopus (423) Google Scholar]: products of infection such as lipopolysaccharide, bacterial DNA or viruses could act as adjuvants and improve the immune response to unrelated antigens, invading organisms could trigger the innate immune system and increase the immune response or lead to autoimmunity by molecular mimicry. Molecular mimicry between bacterial or viral antigens and mitochondrial antigens could trigger PBC [20Sutton I. Neuberger J. Primary biliary cirrhosis: seeking the silent partner of autoimmunity.Gut. 2002; (in press)PubMed Google Scholar, 21Bogdanos D.P. Choudhuri K. Vergani D. Molecular mimicry and autoimmune liver disease: virtuous intentions, malign consequences.Liver. 2001; 21: 225-232Crossref PubMed Scopus (141) Google Scholar]; sequencing antigen on the plasma membrane of the BECs could resolve the issue.3.1 Bacteria and PBCThe involvement of Escherichia coli in the pathogenesis of PBC is suggested by several pieces of evidence: females (but not males) with PBC have a greater risk of having urinary tract infections (UTI) [[22]Parikh-Patel A. Gold E.B. Worman H. Krivey K.E. Gershwin M.E. Risk factors for primary biliary cirrhosis in a cohort of patients from the United States.Hepatology. 2001; 33: 16-21Crossref PubMed Scopus (197) Google Scholar]; Butler has shown that some women with recurrent UTIs have low titre AMA [[23]Butler P. Hamilton-Miller J. Baum H. Burroughs A.K. Detection of M2 antibodies in patients with recurrent urinary tract infection using an ELISA and purified PBC specific antigens.Biochem Mol Biol Int. 1995; 35: 473-485PubMed Google Scholar] although there is controversy whether these really do have the same specificity as in PBC [[24]Floreani A. Bassendine M.E. Mitchison H.C. Freeman R. James O.F. No specific association between primary biliary cirrhosis and bacteriuria?.J Hepatol. 1989; 8: 201-207Abstract Full Text PDF PubMed Scopus (32) Google Scholar]. Hopf [[25]Hopf U. Moller B. Stermerowicz R. Lobeck H. Rodloff A. Freudenberg M. et al.Relation between Escherischia coli (Rough)-forms in gut, lipid A in liver and primary biliary cirrhosis.Lancet. 1989; ii: 1419-1422Abstract Scopus (115) Google Scholar] reported increased amounts of rough forms of E. coli in the faeces of patients with PBC, compared with controls; these findings have not been confirmed by others.These clinical pointers of E. coli as a trigger for PBC have been supported by some laboratory evidence: For instance, Shigematsu [[26]Shigematsu H. Shimoda S. Nakamura M. Matsushita H. Nishimura Y. Sakamoto N. et al.Fine specificity of T cells reactive to human PDC-E2 163–172 peptide, the immunodominant autoantigen in primary biliary cirrhosis: implications for molecular mimicry and cross-recognition amomg mitochondrial autoantigens.Hepatology. 2000; 32: 901-909Crossref PubMed Scopus (61) Google Scholar] used T cell clones to analyse a series of single amino acid substituted peptides. Some of these cell lines reacted with various microbial proteins carrying an ExDK motif. Autoreactive T cell clones from patients with PBC show cross-reactivity between human AMA and E. coli PDC-E2 and with OGDC-E2 [[27]Tanimoto H. Shimoda S. Kawano S. Nakamura M. Hayashida K. Gershwin M.E. Autoreactive T cell clones in patients with primary biliary cirrhosis: evidence of cross reactivity with E. coli OGDC-E2 and human mitochondrial autoantigens.Heptology. 2001; 34 (abstract): 366AAbstract Full Text Full Text PDF Scopus (87) Google Scholar]. In this issue of the Journal, Bogdanos and colleagues [[28]Bogdanos D-P Baum H Sharma U.C. Grasso A Ma Y Burroughs A.K. et al.Antibodies against homologous microbial caseinolytic proteases P characterise primary biliary cirrhosis.J Hepatol. 2002; 36: 14-21Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar] have investigated further the possibility of a cross-reaction. Following the observations of Mayo [[29]Mayo I. Arizti P. Pares A. Oliva J. Doforno R.A. de Sagarra M.R. et al.Antibodies against the COOH-terminal region of E. coli ClpP protease in patients with primary biliary cirrhosis. J Hepatol. 2000; 33: 528-536Google Scholar] that antibodies to the proteolytic sub-unit of the ATP-dependent Clp protease of E. coli are present in some patients with PBC, Baum and colleagues observed a close similarity between a peptide of Clp protease complex and the dominant T cell epitope of PDC-E2 [[30]Baum H. Bogdanos D.P. Vergani D. Antibodies to Clp protease in primary biliary cirrhosis: possible role of a mimicking T-cell epitope.J Hepatol. 2001; 34: 785-787Abstract Full Text Full Text PDF PubMed Scopus (27) Google Scholar]. As there is a high degree of homology between the ClpP of E. coli and other bacteria, they tested the hypothesis that there is cross-reaction with other bacterial antigens; they found evidence of sensitisation not only to E. coli ClpP but also that of Yersinia enterocolitica and Haemophilus influenzae. Whether this means that more than one bacterial species can trigger PBC or whether this is merely cross reaction, remains to be confirmed. Certainly, other bacterial infections have been implicated in inducing PBC [[31]Harada K. Tuneyama K. Sudo Y. Masuda S. Nakanuma Y. Molecular identification of bacterial 16S ribosomal RNA gene in liver tissue of primary biliary cirrhosis: is Propionibacterium acnes involved in granuloma formation?.Hepatology. 2001; 33: 530-536Crossref PubMed Scopus (96) Google Scholar].It is difficult to distinguish whether antibodies are a cause or a consequence of the liver disease. Using antibodies to calreticulin as a marker of infection with gut-derived antigens, Kreisel [[32]Kreisel W. Siegel A. Bahler A. Spamer C. Schlitz E. Kist M. et al.High prevalence of antibodies to calreticulin of the IgA class in primary biliay cirrhosis: a possible role of gut-derived bacterial antigens in its aetiology?.Scand J Gastroenterol. 1999; 34: 623-628Crossref PubMed Scopus (27) Google Scholar] found that IgA antibodies to calreticulin were found in two thirds of patients with PBC, similar frequencies were found in those with yersiniosis; nearly half of those with alcoholic liver disease also had these IgA antibodies. Thus, antibody responses to bacterial or viral antigens may reflect the consequences of liver disease. Morshed [[33]Morshed S.A. Nishioka M. Saito I. Komiyama K. Mori I. Increased expression of Epstein-Barr virus in primary biliary cirrhosis patients.Gastroenterol Jpn. 1992; 27: 751-758PubMed Google Scholar] found increased expression of Epstein-Barr virus in peripheral blood cells, liver and saliva in patients with PBC, compared with normal controls and those with other liver diseases, suggesting that these patients have a depressed immune response to the EB virus.Evidence against the theory of molecular mimicry comes from the work of Potter and colleagues [[34]Potter K.N. Thomson R.K. Hamblin A. Richards S.D. Lindsay J.G. Stevenson F.K. Immunogenetic analysis reveals that epitope shifting occurs during B-cell affinity maturation I primary biliary cirrhosis.J Mol Biol. 2001; 306: 37-46Crossref PubMed Scopus (13) Google Scholar] who investigated the structure of a human monoclonal antibody with specificity against the inner lipoyl domain of E2: they studied reactivities of mutants of a monoclonal antibody from a patient with PBC against recombinant domains of the E2 subunit. Fab in which the V(H)-encoded portion were reverted to germline, lost reactivity against the inner lipoyl domain alone, but recognised a different domain in a didomain including the inner lipoyl domain, E1/E3 binding domain and the hinge region. Since the complete V(H) and V(L) germline was unreactive to the human lipoyl domain, E. coli didomain and whole pyruvate dehydrogenase complex, they suggested that the B cell first recognises antigen, as yet, unrecognised,.. Furthermore, they postulated that accumulation of somatic mutations results in an intermolecular epitope shift directed towards an epitope involving the E1/E3 binding domain and further mutations result in that specificity being re-directed to the inner lipoyl domain. Studies in the XenoMouse, on the other hand, suggest that the autoantibodies react with an immunogen which leads to somatic mutation without subsequent development of determinant spreading [[35]Sasaki M. Van de Water J. Kenny T.P. Gallo M.L. Leung P.S. Nakanuma Y. et al.Immunoglobulin gene usage and immunohistochemical characteristics of human monoclonal antibodies to the mitochondrial autoantigens of primary biliary cirrhosis induced in the XenoMouse.Hepatology. 2001; 34: 631-637Crossref PubMed Scopus (19) Google Scholar]. Furthermore, the response to the inner lipoyl domain, found at the membrane of the BECs, is heterogeneous [[36]Migliaccio C. van de Water J. Ansari A.A. Kaplan M.M. Coppel R.L. Lam K.S. et al.Heterogeneous response of antimitochondrial autoantibodies and bile duct apical staining monoclonal antibodies to pyruvate dehydrogenase complex E2: the molecule versus the mimic.Hepatology. 2001; 33: 792-801Crossref PubMed Scopus (56) Google Scholar].4. Anti-nuclear antibodies associated with PBCAlthough most researchers have focused their attention on the AMA, it must not be forgotten that other autoantibodies are associated with PBC, with a specificity similar to that of AMA. The antibodies which have a specificity similar to AMA for PBC (although a lower sensitivity) are the antibodies against the nuclear pore complex (gp210 and Sp100). These antibodies, which are not associated with AMA tend to be associated with more active and severe disease [[37]Invernizzi P. Podda M. Battezzati P.M. Crosignani A. Zuin M. Hitchman E. et al.Autoantibodies against nuclear pore complexes are associated with more active and severe liver disease in primary biliary cirrhosis.J Hepatol. 2001; 34: 366-372Abstract Full Text Full Text PDF PubMed Scopus (123) Google Scholar]. Recently, Shimoda [[38]Shimoda S. Gershwin M.E. Ishibashi H. Molecular mimicry between nuclear autoantigens and PDC-E2 in patients with primary biliary cirrhosis.Hepatology. 2001; 34 (abstract): 354AAbstract Full Text Full Text PDF Scopus (101) Google Scholar] showed that T cells clones, reactive to the immunodominant T cell epitope of PDC-E2 had reactivity with some peptides present in nuclear pore proteins. Thus cross-presentation and determinant spreading may contribute to the spectrum of antibodies seen in patients with PBC.5. ConclusionsThus, while PBC is associated with a unique profile of antibody responses, characterisation of these antibodies has, as yet, failed to reveal the pathogenesis of the disease. Although these antibodies must be pointers to the pathogenesis of PBC, these clues point in various directions. Separation of primary and secondary events is difficult. The number of hypotheses as to the pathogenesis remains a good testament to our ignorance. Many of these are based on limited information and often ignore well-described features of the disease. Many of the pieces of the jig-saw are in place but no one yet has been able to complete the puzzle. Despite major advances in understanding the epidemiology, natural history and immunology of primary biliary cirrhosis (PBC), both the cause and the definitive treatment remains as elusive in 2002 as it did in 1851 when the syndrome was first described by Addison and Gull. PBC is usually classified as an autoimmune disease, so a clue to the etiology of PBC may lie in the autoantibodies associated with the disease. Autoantibodies may, as in the case of Graves’ disease or myasthenia gravis, mediate the disease or they may be simply phenomenological events [[1]Ermann J. Fathman C.G. Autoimmune diseases: genes, bugs and failed regulation.Nat Immunol. 2001; 2: 799-851Crossref Scopus (153) Google Scholar] 1. PBC and AMAThe very close association between PBC and AMA was first recognised by Doniach and colleagues in 1966; the subsequent identification in 1988, of the major antigen recognised by the AMA as the dihydrolipoamide acetyl transferase by Yeaman and by Gershwin was a major achievement [[2]Van de Water J. Gershwin M.E. Leung P. Ansari A. Coppel R.L. The autoepitope of the 74 kD mitochondrial autoantigen of primary biliary cirrhosis corresponds to the functional site of dihydrolipoamide acetyl transferase.J Exp Med. 1988; 167: 1791-1799Crossref PubMed Scopus (233) Google Scholar]. Subsequent studies defined the dominant B and T cell epitope to amino acids 212–226 of the inner lipoyl domain of the E2 component of pyruvate dehydrogenase – a well conserved enzyme located on the inner mitochondrial membrane [[3]Tuaillon N. Andre C. Briand J.P. Penner E. Muller S. A lipoyl synthetic octadecapeptide of dihydrolipoamide acetyltransferase specifically recognised by anti-M2 antibodies in primary biliary cirrhosis.J Immunol. 1992; 148: 445-450PubMed Google Scholar]. Yet, these discoveries have so far done little to help elucidate the pathogenesis of PBC.Several observations may help clarify the implications of the very closed association between the disease and the AMA [4Thomson R.K. Neuberger J.M. PBC and AMA – what is the connection?.Hepatology. 1999; 29: 271-276Crossref PubMed Scopus (85) Google Scholar, 5Haydon G.K. Neuberger J. PBC: an infectious disease?.Gut. 2000; 47: 586-588Crossref PubMed Scopus (71) Google Scholar]:•AMA may be found in normal individuals at low titre;•AMA are found in over 95% patients with PBC;•AMA may also be associated with some infectious diseases, such as tuberculosis or viral infection [[6]Klein R. Weibel M. Engelhardt S. Berg P.A. Sera from patients with tuberculosis recognise the M2a-epitope (E2 sub-unit of pyruvate dehydrogenase) specific for primary biliary cirrhosis.Clin Exp Immmunol. 1993; 92: 308-316Crossref PubMed Scopus (63) Google Scholar];•high titre AMA may precede clinical, biochemical and histological features of PBC;•although E2 is located in all nucleated cells, the immune targets in PBC are the middle-sized biliary epithelial cells and possibly other epithelial cells;•in PBC and normal subjects, there is increased amount of E2 in biliary epithelial cells compared with the density of mitochondria;•although there is over-expression of E2 in BECs in patients with PBC, increased mRNA cannot be detected;•the target antigen is found aberrantly on the plasma membrane on biliary epithelial cells (BECs) both in vivo and in vitro in PBC but in no other condition;•lymphocytes, both liver derived and intra-hepatic, recognise E2;•AMA titre does not correlate with severity or rate of progression of the disease;•treatment of patients with ursodeoxycholic acid (which affects the biochemistry and may slow disease progression) is associated with a reduction or abolition of AMA in serum;•E2 (unlike other mitochondrial enzymes) is not degraded during apoptosis of BECs in vitro [[7]Odin J.A. Huebert R.C. Cascila-Rosen L. LaRusso N.F. Rosen A. Bcl-2-dependent oxidatin of pyruvate dehydrogenase-E2, a primary biliary cirrhosis autoantigen, during apoptosis.J Clin Invest. 2001; 108: 223-232Crossref PubMed Scopus (186) Google Scholar];•recurrent PBC is seen in some but not all liver allograft recipients;•AMA persist after transplantation, and do not correlate with histological evidence of disease recurrence in the allograft;•aberrant membrane expression of E2 by BECs is seen in the allograft of patients grafted for PBC, whether or not there is histological evidence of recurrent disease;•immunisation of animals with recombinant E2 is not associated with histological evidence of PBC although AMA are generated.However, none of these observations really define the relationship between the diagnostic antibodies and disease. BECs, the target cells of PBC, are immunologically active [[8]Howell C.D. Nie X. Antigen presentation by cytokine-stimulated mouse bile duct epithelial cells.Hepatology. 2001; 34 (abstract): 478ACrossref PubMed Scopus (27) Google Scholar]. There is no evidence to date that the AMA are pathogenetic or whether they are true auto-antibodies or antibodies cross-reacting with bacterial, viral or xenobiotic-associated antigens. The very close association between PBC and AMA was first recognised by Doniach and colleagues in 1966; the subsequent identification in 1988, of the major antigen recognised by the AMA as the dihydrolipoamide acetyl transferase by Yeaman and by Gershwin was a major achievement [[2]Van de Water J. Gershwin M.E. Leung P. Ansari A. Coppel R.L. The autoepitope of the 74 kD mitochondrial autoantigen of primary biliary cirrhosis corresponds to the functional site of dihydrolipoamide acetyl transferase.J Exp Med. 1988; 167: 1791-1799Crossref PubMed Scopus (233) Google Scholar]. Subsequent studies defined the dominant B and T cell epitope to amino acids 212–226 of the inner lipoyl domain of the E2 component of pyruvate dehydrogenase – a well conserved enzyme located on the inner mitochondrial membrane [[3]Tuaillon N. Andre C. Briand J.P. Penner E. Muller S. A lipoyl synthetic octadecapeptide of dihydrolipoamide acetyltransferase specifically recognised by anti-M2 antibodies in primary biliary cirrhosis.J Immunol. 1992; 148: 445-450PubMed Google Scholar]. Yet, these discoveries have so far done little to help elucidate the pathogenesis of PBC. Several observations may help clarify the implications of the very closed association between the disease and the AMA [4Thomson R.K. Neuberger J.M. PBC and AMA – what is the connection?.Hepatology. 1999; 29: 271-276Crossref PubMed Scopus (85) Google Scholar, 5Haydon G.K. Neuberger J. PBC: an infectious disease?.Gut. 2000; 47: 586-588Crossref PubMed Scopus (71) Google Scholar]:•AMA may be found in normal individuals at low titre;•AMA are found in over 95% patients with PBC;•AMA may also be associated with some infectious diseases, such as tuberculosis or viral infection [[6]Klein R. Weibel M. Engelhardt S. Berg P.A. Sera from patients with tuberculosis recognise the M2a-epitope (E2 sub-unit of pyruvate dehydrogenase) specific for primary biliary cirrhosis.Clin Exp Immmunol. 1993; 92: 308-316Crossref PubMed Scopus (63) Google Scholar];•high titre AMA may precede clinical, biochemical and histological features of PBC;•although E2 is located in all nucleated cells, the immune targets in PBC are the middle-sized biliary epithelial cells and possibly other epithelial cells;•in PBC and normal subjects, there is increased amount of E2 in biliary epithelial cells compared with the density of mito