• Potential conflict of interest: Nothing to report.

We appreciate the kind comments regarding our work on murine models of primary biliary cirrhosis (PBC) by Hohenester et al., but, with due respect, there are a number of points the authors should consider. First, our position has always been that the antimitochondrial antibody (AMA) alone is insufficient to make a diagnosis of PBC in either humans or murine models; the diagnosis relies on evidence of autoimmune cholangitis identified on coded slides by a “blinded” pathologist.[1] Importantly, cholangitis can be transferred to naive Rag recipients using CD8 T cells from dnTGFβRII mice demonstrating an autoimmune component, data consistent with flow cytometry of liver subpopulations and cytokine analysis. Third, AMA standardization uses not only enzyme-linked immunosorbent assay (ELISA), but also enzyme inhibition, immunoblotting, absorption, peptide microarrays, affinity-purified antisera, and generation of large panels of mAbs,[1] not just the ELISA data of Hohenester et al. In fact, the definition and measurement of AMAs always includes reactivity to MIT3, the only recombinant autoantigen approved by the Food and Drug Administration (FDA) for liver autoimmune disease. Fourth, it is of concern that the AMA ELISA values reported by Hohenester et al. in control mice are extraordinarily high, higher than we have seen in three decades of screening dozens of strains for AMAs. Hohenester et al. do not provide reactivity to a control recombinant autoantigen and we suspect their recombinant autoantigen is contaminated with proteins of E. coli origin. Fifth, in our view, cross-species comparison of titers of AMAs from humans with many years of established disease with 3-month-old dnTGFβRII mice is inappropriate. It is unknown whether humans in a very early stage of disease (corresponding to 3-month-old dnTGFβRII mice) would have high AMA titers, and there is no immunological basis for excluding any mechanistic role for AMA simply on the basis of titers at this timepoint. The titers of autoantibodies in other murine models of autoimmunity, i.e., New Zealand mice, are not as high as in the kaleidoscope of established human systemic lupus erythematosus. We view PBC as a multi-hit orchestrated immune response, involving CD4, CD8, B cell, and innate immune responses and have documented that intense inflammation can be induced in vitro when normal human biliary cells are incubated with macrophages and AMAs.[2] Finally, we should not forget the excitement after nearly three decades of searching that there are now mice that demonstrate a breach of tolerance against the fundamental PBC immunological sin, mitochondrial autoantigens. We have always offered hepatologists the opportunity to send us sera for free diagnostic testing in cases of questionable patient diagnosis. Thus, we would be pleased to exchange reagents with Hohenester and his group.

“Honest disagreement is a good sign of progress” — Mahatma Gandhi

  • M. Eric Gershwin, M.D.1

  • Patrick S.C. Leung, PH.D.1

  • William M. Ridgway, M.D.2

  • Ross L. Coppel, PH.D.3

  • Aftab A. Ansari, PH.D.4

  • 1Division of Rheumatology, Allergy and Clinical Immunology, University of California at Davis School of Medicine, Davis, CA

  • 2Department of Rheumatology and Clinical Immunology, University of Cincinnati College of Medicine, Cincinnati, OH

  • 3Department of Microbiology, Monash University, Clayton, Victoria, Australia

  • 4Department of Pathology, Emory University School of Medicine, Atlanta, GA