Antimitochondrial antibodies may be insufficiently specific to define primary biliary cirrhosis-like disease in mouse models

Abstract

To the Editor: Primary biliary cirrhosis (PBC) is considered to be an autoimmune disorder. When not adequately responding to medical treatment, PBC often progresses to liver cirrhosis, necessitating liver transplantation.1 This constitutes a need for the development of animal models, both to unravel pathogenetic pathways and to test innovative therapeutic agents. Gershwin and his group have greatly contributed to this endeavor and in their most recent and intriguing manuscript successfully test a novel therapeutic approach in a well-established PBC model.2 In humans, antimitochondrial antibodies (AMAs, especially anti–PDC-E2) are the diagnostic hallmark of PBC.3 Accordingly, AMAs are being used to define PBC-like disease also in mice,4 even though alterations in serum liver tests or histological changes are sometimes minimal. In our view, however, use of AMAs to define PBC in mice is potentially misleading when insufficiently quantified. Using recombinant PDC-E2170-313,6 we established an enzyme-linked immunosorbent assay (ELISA) to quantify AMA reactivity in mouse and human serum, determining the half maximal effective concentration. We found significantly increased AMAs in dnTgfβ-R2 mice, a proposed PBC mouse model, at 3 months of age, in line with previous reports.4 However, their AMA titer was only 3.6-fold increased compared with wild-type littermates (Fig. 1A,B). In contrast, AMA reactivity in sera of human PBC patients was more than 2,500-fold increased compared with age-matched healthy controls (Fig. 1C,D). Subsequently, we studied AMA reactivity in a cohort of 24 wild-type female C57Bl/6 mice by comparing optical density in single dilutions (1:1,000) and found a significant increase with age from 0.35 ± 0.11 to 0.55 ± 0.30 and 1.05 ± 0.72 (optical density) at 3, 6, and 12 months of age, respectively (Fig. 1E). This age dependency was not found in a cohort of 116 female human controls (Fig. 1F). We conclude from these observations that (1) AMAs do not adequately define PBC-like disease in mice, (2) other immunologic and histologic features of PBC must instead be carefully evaluated in PBC models, and (3) the value of purely AMA-based PBC animal models to test therapeutic compounds should be re-evaluated. Anti–PDC-E2 signals of sera from (A) dnTgfβ-R2 mice (n = 14) and wild-type littermates (n = 7) and (C) PBC patients (n = 15) and healthy controls (n = 14) were determined following stepwise dilution with phosphate-buffered saline by ELISA. The half maximal effective concentration (EC50) was calculated from these titration curves for precise quantification of anti-AMA reactivity (B and D). *P < 0.05 (analysis of variance). Applying single dilutions (1:1,000), anti-PDC-E2 signals of (E) wild-type mice (n = 24) and (F) healthy human females (n = 116) were determined by ELISA, and optical density (OD) was correlated with age. *P < 0.05 (analysis of variance). Simon Hohenester M.D. Ulrich Beuers M.D. Juan F. Medina M.D., Ph.D. Ronald P. Oude Elferink Ph.D. Tytgat Institute for Liver and Intestinal Research, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands Division of Gene Therapy and Hepatology, CIMA, and School of Medicine, University of Navarra and Ciberehd, Pamplona, Spain