Animal models of human autoimmune disease: The NOD mouse and  Schistosoma Mansoni infection

Figure 2: Immunofluorescent staining of a NOD pancreas showing mononuclear cell infiltration. Section of NOD pancreas showing insulin producing β cell mass (orange) and mononuclear cells (green) stained with CD3. Pancreas from young mice show no infiltrate whereas (a) older mice show a spontaneous infiltration around the islet (b) initiating β cell destruction.

Since the 1970s the NOD (Non-Obese Diabetic) mouse has provided a good model for the study of T1D. Initially generated in Japan by Makino and co-workers, the NOD mouse has became one of the most popular models to study T1D (25). NOD mice spontaneously develop T1D, with features similar to the human disease. NOD T1D is under polygenic control and, much like the human disease, associates with particular Class II major histocompatibility (MHC) polymorphisms (26). The pancreas of NOD mice become infiltrated with mononuclear cells around 6 weeks of age, with cells appearing chiefly around the islets and pancreatic ducts. By 8–12 weeks of age the infiltrate progresses to the islets, causing destruction of the β cell mass. Pathology is primarily cell-mediated, with dendritic cells (DC), macrophages (MΦ) and B cells responsible for the initiation of the autoimmune process by presentation of pancreatic antigen and secretion of inflammatory mediators (27,28). Subsequently, CD8+ and CD4+ T cells enter the pancreas, infiltrate the islet area and β cell destruction arises through a Th1-mediated immune response (29) (Figure 2). After 12 weeks of age the clinical signs of disease start to manifest with polydypsia and glycosuria and by the age of 30 weeks 80–100% of female mice are diabetic. NOD mice show a distinct gender difference with female NOD mice developing T1D at a much higher incidence than the males (10–20%). There are variations between colonies and the conditions under which NOD mice are kept appear to greatly influence the rate and frequency of onset of diabetes. It rapidly became clear that NOD mice kept under germ-free conditions developed diabetes at a much faster rate and higher incidence than mice kept under conventional conditions (30). This observation, made independently in many different laboratories, provoked immunologists to consider the possibility that infection and/or exposure to microbial products was responsible for the reduction of T1D incidence in some animal colonies. Experiments designed to test this hypothesis and elucidate the mechanisms of T1D prevention, revealed that infections triggering both Th1- and Th2-like responses could delay or abolish autoimmune pathology in NOD mice (click to see Table 3).

The NOD mouse appears to be a good model for testing the predictions of the Hygiene Hypothesis, therefore we have used it to study the effects of bacterial and helminth infection on the onset of T1D. Schistosoma mansoni infection, or even exposure to antigens derived from this helminth, results in long-lasting prevention of diabetes characterized by a strong Th2 response (31,32). S. mansoni protection appears to stem largely from a shift to a non-pathological Th2 response, although there is also evidence for the generation of immunosuppressive regulatory cells (Treg) (32). Essentially similar results have been observed using two other species of helminth, Heligmosomoides polygyrus and Trichinella spiralis (C. Lawrence, unpublished data). Similarly, infection of NOD mice with live attenuated Salmonella bacteria induces a long-lasting protection from T1D (35). Prevention of a Th1-mediated autoimmune disease such as T1D by infection with a classic Th1-stimulating pathogen appears rather paradoxical. Potentially a generalized Salmonella-induced IFN-γ release may mediate suppression through its effects on the innate immune system, particularly DC, but the mechanism awaits full characterization (our unpublished observations). At any rate, the invocation of a simple Th1 to Th2 shift is unable to explain all the immunomodulatory effects of microbial infection that lead to prevention of T1D, and may instead require a more complex paradigm incorporating, for example, the action of Treg.