• Potential conflict of interest: Nothing to report.


We thank the Newcastle group for their generous comments, and we are very appreciative of their recognition of our work. We do acknowledge their informative development since 1999 of various adjuvant-facilitated models of primary biliary cirrhosis (PBC) in the J substrain of inbred “Swiss Jim Lambert” (SJL) mice, either by use of complete Freund's adjuvant (CFA) or bacterial motif of DNA. There did seem earlier some call to question specificity,1 given that SJL mice are well known for diffuse tissue infiltrates, and that administration of CFA alone has long been known2 and often noted subsequently3 to induce granulomatous infiltrates in tissues including liver.

Our recent observations on the induction of autoimmune cholangitis in mice following xenobiotic immunization follows up on a several year effort on our thesis that loss of tolerance to the E-2 subunit of pyruvate dehydrogenase complex (PDC-E2), the immunodominant autoantigen of PBC, is the central and critical requirement for disease development.4–6 Moreover, the antimitochondrial response is more than merely a serologic marker but rather reflects a multilineage series of effector pathways that include antibody but especially autoreactive CD4+ and CD8+ T lymphocytes; the T cell precursor frequency in the liver of autoreactive CD4 and CD8 cells is between 10-fold and more than 100-fold higher than it is in blood.7 Also, T cell reactivity to mitochondrial antigens is demonstrable even in patients negative for antimitochondrial antibody. The development of autoimmune cholangitis following xenobiotic immunization indicates that not only is loss of tolerance the primary event, but also that the biliary epithelium is a unique victim due to its particular properties of apoptosis, thus explaining why PBC recurs following liver transplantation. In addition to our xenobiotic-induced model, there are described several other mouse models of “spontaneous” autoimmune cholangitis independent of adjuvant usage8–11; such mice often share in common a depletion in T regulatory pathways, similar to that seen in a child described with autoimmune cholangitis born with interleukin-2 receptor α deficiency.12

At this stage, it now seems important to proceed onward from these models to focus on the molecular basis of disease including epitope mimics, the role of microorganisms noting the induction of autoimmune cholangitis following infection with Novosphingobium aromaticivorans,13 and use of these models for dissecting the complex molecular genetics of autoimmune predisposition. Adjuvant is a critical issue that should not be overlooked; in our model, too little or too much CFA will modulate the result. We are certain there will be, additional to SJL/J, other PBC-susceptible mouse strains, and we will be presenting data on at least two further strains that are susceptible, as well as newer xenobiotics based upon our qualitative structure analysis relationships of antimitochondrial antibody reactivity.

M. Eric Gershwin*, Ian R. Mackay†, * Division of Rheumatology, Allergy, and Clinical Immunology, University of California at Davis School of Medicine, Davis, CA, † Departments of Biochemistry and Molecular Biology, Monash University, Clayton, Australia.