We appreciate the comments by by Lapierre et al. on our article on experimental murine autoimmune hepatitis (emAIH), because the group is known for their expertise on autoimmune hepatitis (AIH).
The major advantage of our model is the genetic background predisposing to (1) a more rapid development of AIH and (2) a disease phenotype closely resembling the human disease. As discussed in the article, nonobese diabetic (NOD) mice share genetic predispositions also relevant in human AIH (e.g., major histocompatibility complex and cytotoxic T-lymphocyte antigen 4), whereas such associations are not known for C57Bl/6 and FVB/N mice. However, even after follow-up for 30 weeks, we did not see signs of AIH in the latter two strains. Our model offers the additional advantage that mechanisms leading to a break of tolerance and therapeutic interventions could be studied. Besides this, the more severe course of the disease enabled us to study disease end-points such as fibrosis. Fibrosis was verified by a pathologist in a blinded way with a standard staining used in clinical routine. Others' models did not see any fibrosis[2, 3] or unphysiological subcapsular fibrosis just after intraperitoneal virus injection.
Other researchers have also described the need for strong “danger signals” to break tolerance against liver antigens[3-5]; however, we could show that the mere overexpression of the autoantigen is not sufficient to cause AIH. This is well in line with the experience of Lapierre et al. employing a repeated prime boost protocol with DNA vaccination with a plasmid encoding for a secreted chimeric antigen of cytochrome P4502D6 (CYP2D6) and formiminotransferase cyclodeaminase (FTCD), with the addition interleukin (IL)-12 or expression with an adenovirus. Of note, we have also tested intramuscular DNA injection with cytomegalovirus-promoter-driven FTCD and IL-12, but did not see meaningful AIH within 6 months in NOD mice (see Fig. 1).
The important publication by Bowen et al. did not use danger signals, but rather induced a self-limited hepatitis, followed by immune tolerance, thereby, rather supporting our statement.
Regarding the nature of the inducing autoantigen, we could show that similar antigens can induce AIH as well as identical antigens. However, we did not try CYP2D6 as an orthologous autoantigen. The amount of virus could be excluded as a variable, because we tried 1 × 109 to 1 × 1010 plaque-forming units of our adenovirus, as reported by Holdener et al.
As recently discussed, we think that all the mentioned models are important and can be used to explain various aspects of AIH and hepatic immune tolerance.
Matthias Hardtke-Wolenski, Ph.D.
Michel P. Manns, M.D.
Elmar Jäckel, M.D.
Department of Gastroenterology, Hepatology, and Endocrinology, Hannover Medical School, Hannover, Germany