Background: Fistulas represent a frequent complication in Crohn's disease (CD) and surgical resection is often required. Despite some progress in the understanding and treatment of inflammatory bowel disease (IBD), more effective medical treatments are still required, especially for CD patients with fistula formation. Previously we demonstrated that epithelial-to-mesenchymal transition (EMT) plays a critical role for fistula development in CD patients. Preceding upregulation of TGF-beta, IL-13, TNF & their receptors along fistula tracts in CD patients seems to orchestrate a number of events contributing to the onset of fistulas, by inducing EMT. Due to a lack of a reliable in vivo model, new drug developments are complicated. Here, we are describing a new xenograft (XGR) mouse model of intestinal fistula, resembling the human condition.
Methods: 12–18 weeks (w) old human fetal small intestine was transplanted subcutaneously onto the backs of SCID mice. After 12–16w, ∼15% of the mature xenografts spontaneously developed enterocutaneous fistulas. Using systemic LPS treatment followed by mild skin irritation adjacent to the transplant, we established a reproducible model system, resulting in enterocutaneous fistulas 2–4w later. Tissue specimens were immunohistochemically stained (IHC) for EMT & immune cell markers.
Results: Morphological analysis of the fistulating XGR samples revealed flattening of the intestinal epithelial cells lining the fistula tract, resembling transitional cells described in human patient samples. IHC stainings for various EMT markers (e.g. SLUG, β-6 Integrin) revealed similar expression patterns like for human fistulating CD patient samples. The expression of the mesenchymal marker alpha-smooth muscle actin confirmed the hypothesis that EMT plays a critical role for the fistula development in the XGR samples, as well.
H&E staining also showed inflammation in the gut XGR up- & downstream to the fistulous tracts. Most of this inflammatory response consisted of human CD45+ round cells & very few murine CD45+ cells, mostly polymorphonuclear. Collagen staining revealed these inflammatory regions were also associated with massive fibrosis, suggesting extracellular matrix remodeling.
Besides we observed many potential necroptotic paneth cells in the XGR samples & a loss of this cells in the crypts adjacent to the fistula. RNA sequencing showed significant upregulation of genes related to IBD, necroptosis, ripoptosome, & NF-κB signaling in inflamed LPS-treated XGRs.
Conclusions: Our data demonstrate that the in vivo model recapitulates both morphologically & mechanistically, the human disease. Necroptosis might be the underlying molecular mechanism driving inflammation, EMT and finally resulting in fistula formation.
M. Scharl and N.Y. Shpigel contributed equally.