Primary biliary cirrhosis (PBC) is a chronic cholestatic liver disease that affects mainly middle-aged women and is characterized by an immune-mediated inflammatory destruction of the small intrahepatic bile ducts, progressing to cirrhosis and subsequent liver failure.1 The diagnostic hallmark of PBC is the presence of high-titer antibodies directed to mitochondria (AMA) and in particular the E2 component of the pyruvate dehydrogenase complex (PDC-E2).1 The principal B2 and T (both CD43 and CD84) cell epitopes on PDC-E2 have been defined and shown to colocalize within the inner lipoyl-binding domain of the subunit, overlapping amino acids (aa) 208-237 (PDC-E2208–237). PDC-E2208–237 is physically exposed on the molecule's surface, and this may partially explain its particular antigenicity.5 The mechanism by which this short sequence becomes the focus of PBC-specific antimitochondrial immune responses remains obscure.6

Interplay of genetic and environmental factors is invariably invoked to justify the emergence of the autoimmune aggression in PBC.1 There is an increasing body of circumstantial evidence implicating an external trigger—be it an infectious agent or other environmental factor—in the pathogenesis of PBC (see reviews7, 8). This includes clustering of cases within families with the affected members of the same family acquiring the disease at the same time rather than at the same age; clustering of cases within geographical areas; geographical variation of disease's prevalence; effect of migration with migrating populations acquiring the prevalence of the disease of the host population; absence of disease's equivalent in childhood; poor response to immunosuppressive therapy; weak human leucocyte antigens (HLA) associations; granulomatous lesions resembling those due to mycobacteria; epidemiological association with infectious agents.

The evidence in support of a viral trigger is indirect7, 8: the elevated levels of IgM may be due to virus- induced polyclonal activation; AMA (and in particular anti-PDC-E2 antibodies) belong to the IgG3 isotype as antiviral antibody responses frequently do; akin to viral hepatitis, PBC recurs after liver transplantation, the recurrence, as in viral hepatitis, occurring earlier and being more aggressive if higher doses of immunosuppressive are used; a transmissible factor from lymph nodes of patients with PBC can promote in vitro aberrant expression of the PDC-E2 on biliary epithelial cells.

Despite the evidence, whether PBC has an infective component, bacterial or viral, remains to be defined. We are dealing with an “enigmatic” disease, the pathogenesis of which is “poorly understood” as emphasized by two papers9, 10 published, respectively, in the November issue and this issue of HEPATOLOGY. Both papers provide suggestive evidence that a microbe may be the driving force leading to PBC; both emphasize the role of the genetic background. The similarity between the two papers ends here.

Selmi et al.9 link the ubiquitous microorganism Novosphingobium aromaticivorans (N. aromaticivorans) to the development of PBC. Having found a striking amino acid similarity between human PDC-E2208–237, the major mitochondrial autoepitope and the corresponding sequence on PDC-E2 of N. aromaticivorans, these authors9 provide evidence that the newly identified sequence may act as a bridge to autoimmunity through a mechanism of molecular mimicry.

Molecular mimicry (Fig. 1) requires that recognition of an immunogenic determinant on an exogenous agent, bacterium, or virus leads to the formation of antibodies or effector T cells which can react with homologous epitopes on a host protein.11, 12 The sharing of a linear amino acid sequence or a conformation fit between a microorganism and a host “self” determinant is the initial step of this process. Autoimmunity provoked by molecular mimicry should occur only when the microbial and host determinants are similar enough to cross-react, yet different enough to break immunological tolerance.11, 12 Antibodies or T lymphocytes may then cross-react with a self-protein, thereby causing cellular injury leading to cell destruction. Once the infectious agent initiates this process, it need not be present during the autoimmune destruction that follows. In this situation, the microbial agent may have been cleared, but elements of the immune response mounted against it would lead to tissue injury that, in turn, would release more self antigens, thus strengthening the autoimmune response, and so on. Therefore, although microbes may initiate disease, the likelihood of their recovery from tissue sites is small. Sequence similarity, however, does not necessarily lead to structural/conformational similarity and, hence, need not equate with actual cross-reactivity (antigenic mimicry), particularly in the case of B-cell epitopes. In other words, amino acid similarity does not have biological significance if the implicated microbial and self sequences are not targets (epitopes) of immune responses.11, 12

thumbnail image

Figure 1. (A) The receptor of a B lymphocyte, depicted here in its soluble form as an antibody, recognizes an amino acid sequence on a microbe that strongly resembles a sequence on a self antigen. Exposure to a sequence mimicking self, especially if repeated, expands a potentially cross-reactive lymphocyte population, ultimately leading to cross-reactivity and autoimmunity.12 (B) A T cell receptor specific for a microbial peptide presented by HLA- molecule 1 can cross-react with a similar peptide presented by the same HLA molecule or by HLA- molecule 2 if the final conformation fits the TCR. Of note, the T cell receptor recognizes as antigen not only the peptide but also the “shoulders” of the HLA-molecule.31

Download figure to PowerPoint

The findings of Selmi's work add support to the notion that a microbial exposure may be concerned with the appearance and/or maintenance of AMA responses by a cross-reactive mechanism.13–15 The first observation that PBC sera cross-react with bacterial antigens was made almost 30 years ago when Sayers and Baum demonstrated cross-reactivity of AMA positive sera with membrane vesicles of Paracoccus denitrificans.16 In view of the remarkable sequence similarities of PDC-E2 across species, Burroughs et al.,13 suggested that molecular mimicry between microbial PDC-E2, especially that of Escherichia coli, and PDC-E2 of man could lead to a cross-reactive immune response manifesting itself as AMA. This suggestion arose in the light of early epidemiological considerations, such as an association of AMA with urinary tract infection and the fact that sera and T cell clones from patients with PBC are capable of human/microbial PDC-E2 cross-recognition.14, 17 Others have challenged this view arguing that cross-reactive immunity can readily be predicted, taking into account the highly conserved nature of PDC-E2 among species, and in particular of the inner lipoyl domain.18 Antibodies reacting with Escherichia coli PDC-E2 have indeed been found in PBC sera, but the fact that their titers are 100-fold lower than those of mammalian PDC-E2 has been used as an argument against cross-reactivity being at the origin of AMA production.19

The debate has been rekindled by the findings of Selmi et al.9 who have discovered the best ever similarity between microbial and human PDC-E2. Of 13 contiguous amino acids comprising the core sequence of human PDC-E2 (GDLLAEIETDKAT), 12 are identical and only 1 is different in the corresponding sequence of N. aromaticivorans. What makes this finding even more interesting is the fact that 100% of anti-PDC-E2 positive PBC patients reacted with 2 lipoylated proteins of N. aromaticivorans with titers up to 10−6, these titers being 100 to 1000 fold higher than those to the PDC-E2 of Escherichia coli, a bacterium frequently implicated in the cross-reactive induction of AMA. Of interest, even though the authors do not draw our attention to it, is that the N. aromaticivorans PDC-E2 mimic contains motifs critical for CD4 and CD8 T cell recognition, making this mimic a potential target not only for B cells but also for T cells.17, 20 The findings of Selmi's work and those of recent studies demonstrating the existence of PDC-E2 cross-reactive microbial mimics in non-PDC-E2 microbial sequences of Escherichia coli,21 Mycobacteria,22Lactobacillus,23 and other common pathogens support the concept of a “multiple hit” mechanism of molecular mimicry (Fig. 1) whereby a short sequence on human PDC-E2 becomes a cross-reactive target of several microbial mimics, a process that, in susceptible individuals, may culminate in tolerance breakdown.12

The immune response to N. aromaticivorans proteins was found to be highly disease specific as none of the 195 healthy and pathological controls investigated had antibodies to N. aromaticivorans.

This observation warrants a caveat. The total correlation of reactivity to the microorganism and to PDC-E2 might be interpreted as no more than the recognition by anti-PDC-E2 antibodies of almost identical sequences in another set of molecules. However, N. aromaticivorans not only has a high degree of homology with the critical lipoylated inner domain of human PDC-E2, but is also capable of metabolizing xenobiotics to species such as those shown as modifying the lipoyl residues of PDC-E2 to generate forms very reactive with sera of PBC patients, and giving rise to autoimmunity in experimental animals. The authors speculate that asymptomatic presence of N. aromaticivorans and the chemical modification of PDC-E2 inner domain in a genetically susceptible host may lead to a vigorous self-response.9

In contrast to the immune response to N. aromaticivorans that is universally present in patients with PBC and highly specific for the disease, the microorganism itself can be found in the feces of ¼ of PBC patients and in an identical proportion of pathological and normal controls. The reason as to why an immune response directed to N. aromaticivorans is restricted to patients with PBC when the bacterium is present in both PBC and controls is still unclear and invokes the contribution of additional factors, with the genetic background being the most likely candidate. In support of a strong genetic component to PBC, are the results contained in a communication to the 2002 meeting of the American Association for the Study of Liver Diseases, showing a concordance rate of 75% for PBC in monozygotic twin pairs and of 25% in dizygotic twin pairs.24 In line with other diseases characterized by marked autoimmune manifestations, the studies of the genetic component of PBC have largely focused on the HLA system, with unrewarding results. There are no strong HLA associations in PBC, in contrast to autoimmune hepatitis, or non-liver autoimmune diseases such as multiple sclerosis and type 1 diabetes. Equally uninfluencing appear to be other genes involved as the HLA in shaping the immune response such as those coding for cytokines or chemokines.

Approximately 10% of the human genome is encoded by retroviral elements entering the human genome during evolution (reviewed by refs. 25, 26). Once inserted into the germline, these elements are vertically transmitted as genomic components. Retroviral fragments are retained as genetic fossils, superantigens, or possible regulatory elements in the MHC class I and II, Fas and complement genes. Several studies have indicated that endogenous retroviruses may be responsible for the induction of an autoimmune process leading to autoimmune disease, such as systemic lupus erythematosus and Sjögren's syndrome, through a variety of mechanisms.25, 26 The role of retroviruses in PBC has mainly been investigated by Mason's group.27–29 These investigators were able to isolate products with homology to retroviruses and found the sequences to be derived from human endogenous retroviruses.30 In subsequent studies they found that up to 50% of PBC patients had antibodies to human intracisternal A-type particle, a retrovirus isolated from Sjögren's syndrome patients; observed virus-like particles by electron microscopy in the biliary epithelial cells of PBC patients; and produced the aberrant expression of the PDC-E2 like protein, a phenotypic manifestation of PBC, on normal biliary epithelial cells.27–29 The transmissible factor promoting this PBC phenotype is gamma radiation sensitive, has the hydrodynamic properties of an enveloped retrovirus, has the morphological features of a B-type particle by electron microscopy, and demonstrates reverse transcriptase activity.29

In their present study,10 these workers cloned a complete proviral genome of a betaretrovirus, the name deriving from its strong similarity with the mouse mammary tumor betaretrovirus (MMTV), from the lymph nodes of PBC patients and provide evidence to support the exogenous origin of the retrovirus. The question remains as to whether the cloned retrovirus is human or murine in nature, since it shares 93–97% with MMTV isolates. This uncertainty grows with the authors' demonstration that pure mouse-derived isolates MMTV can induce the same PBC phenotype of PDC-E2 expression in biliary epithelial cells.29 The described retrovirus appears to be more lymphotropic than hepatotropic and by reverse transcription-polymerase chain reaction was only seen in ⅕ of the serum samples that presented antiretroviral reactivity.29 The disease specificity needs to be further addressed, since the retrovirus is present in pathological controls and an antiretroviral immune response, equalling or exceeding that found in PBC, is present in patients with viral hepatitis, primary sclerosing cholangitis, biliary atresia, and systemic lupus erythematosus.27, 29

In the present papers, both groups of workers provide hints as to how their microbe may explain female preponderance of the disease: thus the betaretrovirus contains an element that can be activated by female hormones, while Novosphingobium can degrade 17-β estradiol, increasing its active fraction.

In conclusion, both papers provide evidence favoring the involvement of a microbe in the pathogenesis of PBC, but by default emphasize the importance of a genetic component. It is becoming clear that there are several pieces to the PBC jigsaw, and only their fitting together is likely to result in clinical disease.


  1. Top of page
  2. References
  • 1
    Kaplan MM. Primary biliary cirrhosis. N Engl J Med 1996; 335: 15701580.
  • 2
    Van de Water J, Gershwin ME, Leung P, Ansari A, Coppel RL. The autoepitope of the 74-kD mitochondrial autoantigen of primary biliary cirrhosis corresponds to the functional site of dihydrolipoamide acetyltransferase. J Exp Med 1988; 167: 17911799.
  • 3
    Shimoda S, Nakamura M, Ishibashi H, Hayashida K, Niho Y. HLA DRB4 0101-restricted immunodominant T cell autoepitope of pyruvate dehydrogenase complex in primary biliary cirrhosis: evidence of molecular mimicry in human autoimmune diseases. J Exp Med 1995; 181: 18351845.
  • 4
    Kita H, Lian ZX, Van de Water J, He XS, Matsumura S, Kaplan M, Luketic V, et al. Identification of HLA-A2-restricted CD8(+) cytotoxic T cell responses in primary biliary cirrhosis: T cell activation is augmented by immune complexes cross-presented by dendritic cells. J Exp Med 2002; 195: 113123.
  • 5
    Howard MJ, Fuller C, Broadhurst RW, Perham RN, Tang JG, Quinn J, Diamond AG, et al. Three-dimensional structure of the major autoantigen in primary biliary cirrhosis. Gastroenterology 1998; 115: 139146.
  • 6
    Bogdanos DP, Baum H, Vergani D. Antimitochondrial and other autoantibodies. Clin Liver Dis 2003; 7: 759777.
  • 7
    Haydon GH, Neuberger J. PBC: an infectious disease? Gut 2000; 47: 586588.
  • 8
    Joplin R. Thoughts on the infectious aetiology of primary biliary cirrhosis. In: MannsMP, PaumgartnerU, LeuschnerU, eds. Falk Symposium 114, Immunology and Liver. Dordrecht: Kluwer Academic Publishers, 2000: 268278.
  • 9
    Selmi C, Balkwill DL, Invernizzi P, Ansari A, Coppel R, Podda M, Leung P, et al. Patients with primary biliary cirrhosis react against a ubiquitous xenobiotic metabolizing bacterium. HEPATOLOGY 2003; 38: 12501257.
  • 10
    Xu L, Sakalian M, Zhiwei S, Loss G, Neuberger J, Mason A. Cloning the human betaretrovirus proviral genome from patients with primary biliary cirrhosis. HEPATOLOGY 2004; 39: 151156.
  • 11
    Oldstone MB. Molecular mimicry and immune-mediated diseases. Faseb J 1998; 12: 12551265.
  • 12
    Bogdanos DP, Choudhuri K, Vergani D. Molecular mimicry and autoimmune liver disease: virtuous intentions, malign consequences. Liver 2001; 21: 225232.
  • 13
    Burroughs AK, Butler P, Sternberg MJ, Baum H. Molecular mimicry in liver disease. Nature 1992; 358: 377378.
  • 14
    Shimoda S, Nakamura M, Shigematsu H, Tanimoto H, Gushima T, Gershwin ME, Ishibashi H. Mimicry peptides of human PDC-E2 163-176 peptide, the immunodominant T-cell epitope of primary biliary cirrhosis. HEPATOLOGY 2000; 31: 12121216.
  • 15
    Bogdanos DP, Baum H, Sharma UC, Grasso A, Ma Y, Burroughs AK, Vergani D. Antibodies against homologous microbial caseinolytic proteases P characterise primary biliary cirrhosis. J Hepatol 2002; 36: 1421.
  • 16
    Sayers TJ, Baum H. Possible cross-reactivity of human anti-mitochondrial antibodies with membrane vesicles of Paracoccus denitrificans. Biochem Soc Trans 1976; 4: 138139.
  • 17
    Kita H, Matsumura S, He XS, Ansari AA, Lian ZX, Van de Water J, Coppel RL, et al. Analysis of TCR antagonism and molecular mimicry of an HLA-A0201-restricted CTL epitope in primary biliary cirrhosis. HEPATOLOGY 2002; 36: 918926.
  • 18
    Van de Water J, Ishibashi H, Coppel RL, Gershwin ME. Molecular mimicry and primary biliary cirrhosis: premises not promises. HEPATOLOGY 2001; 33: 771775.
  • 19
    Fussey SP, Lindsay JG, Fuller C, Perham RN, Dale S, James OF, Bassendine MF, et al. Autoantibodies in primary biliary cirrhosis: analysis of reactivity against eukaryotic and prokaryotic 2-oxo acid dehydrogenase complexes. HEPATOLOGY 1991; 13: 467474.
  • 20
    Shimoda S, Van de Water J, Ansari A, Nakamura M, Ishibashi H, Coppel RL, Lake J, et al. Identification and precursor frequency analysis of a common T cell epitope motif in mitochondrial autoantigens in primary biliary cirrhosis. J Clin Invest 1998; 102: 18311840.
  • 21
    Bogdanos DP, Baum H, Grasso A, Okamoto M, Butler P, Ma Y, Rigopoulou EI, et al. Microbial mimics are major targets of crossreactivity with human pyruvate dehydrogenase in primary biliary cirrhosis. J Hepatol 2004; 40: 4351.
  • 22
    Bogdanos DP, Pares A, Sharma UC, Caballeria L, Ma Y, Burroughs AK, Rodes J, et al. Crossreactivity between mycobacterium gordonae and PDC-E2, the major target of antimitochondrial antibody, is present in Spanish but not British patients with primary biliary cirrhosis. J Hepatol 2001; 34( suppl): 19.
  • 23
    Bogdanos DP, Grasso A, Okamoto M, Montalto P, Butler P, Williams R, Burroughs AK, et al. Cross-reactivity to Lactobacillus delbrueckii/human pyruvate dehydrogenase complex-E2 characterizes primary biliary cirrhosis. HEPATOLOGY 2000; 32( suppl): 300A.
  • 24
    Selmi C, Mayo MJ, Bach N, Gish RG, Wright HI, Ishibashi H, et al. Occurrence of primary biliary cirrhosis in monozygotic twins. HEPATOLOGY 2002; 36( suppl): 387A.
  • 25
    Urnovitz HB, Murphy WH. Human endogenous retroviruses: nature, occurrence, and clinical implications in human disease. Clin Microbiol Rev 1996; 9: 7299.
  • 26
    Adelman MK, Marchalonis JJ. Endogenous retroviruses in systemic lupus erythematosus: candidate lupus viruses. Clin Immunol 2002; 102: 107116.
  • 27
    Mason AL, Xu L, Guo L, Munoz S, Jaspan JB, Bryer-Ash M, Cao Y, et al. Detection of retroviral antibodies in primary biliary cirrhosis and other idiopathic biliary disorders. Lancet 1998; 351: 16201624.
  • 28
    Sadamoto T, Joplin R, Keogh A, Mason A, Carman W, Neuberger J. Expression of pyruvate-dehydrogenase complex PDC-E2 on biliary epithelial cells induced by lymph nodes from primary biliary cirrhosis. Lancet 1998; 352: 15951596.
  • 29
    Xu L, Shen Z, Guo L, Fodera B, Keogh A, Joplin R, O'Donnell B, et al. Does a betaretrovirus infection trigger primary biliary cirrhosis? Proc Natl Acad Sci U S A 2003; 100: 84548459.
  • 30
    Xu L, Guo L, Joplin R, Neuberger J, Rixon F, Aitkin J, Carman W, et al. Viral induction of primary biliary cirrhosis. J Hepatol 1997; 26( suppl 1): 61.
  • 31
    Wekerle H, Hohlfeld R. Molecular mimicry in multiple sclerosis. N Engl J Med 2003; 349: 185186.