Keywords

Introduction

Animal models are of great value in the evaluation of pathogenic mechanisms as well as novel and experimental treatments, which cannot be tested directly on patients. The ideal characteristics for an animal model are similarities to the human disease in terms of course, symptomatology, pathophysiology, and response to treatment. In addition, we would like the animal model to show reproducibility: a high rate of response to disease induction in the animals as well as homogeneity of the onset and disease manifestations between the animals. In this chapter, we survey the various experimental models that were introduced for Behçet’s disease (BD), and present some of our unpublished experience in this field. Basically, these models can be divided according to the proposed etiological paradigms.

Environmental Pollution Model

Prolonged oral administration of organic chlorides, organo-phosphate (DDT-trichloroethanediyl-bis-chlorobenzene, polychlorated-biphenyl (PCB), Sumithion™ – dimethyl-nitro-phosphorothioate), and inorganic copper to Pitman-Moor swine induced folliculitis, cutaneous nodules, genital ulcers, oral aphthae, and intestinal ulcers [1]. The clinical manifestations and histology resembled BD including changes in vascular endothelium, bleeding, hair follicles, and intestine mucosa necrosis. Microanalysis detected high levels of the above metals and low zinc concentrations in the peripheral neutrophils, infiltrating inflammatory cells and endothelial cells derived from the mucocutaneous lesions [1]. Although a study on BD patients sera had reported low levels of zinc and normal levels of magnesium, an X-ray spectroanalysis of BD skin lesions failed to detect the proposed offending elements [2]. To the best of our knowledge, this experimental model has not been used in further work. Nevertheless, it underlines the fact that prolonged exposure to certain chemical combinations can elicit multisystem inflammatory response.

Infectious Models

Bacterial Infectious Models

In the search for possible infectious causative agents of BD 4, species of the Streptococcus genus (S. salivarius, S. faecalis, S. pyogenes and S. sanguis) were isolated from lesions of patients with active BD. It was noted that crude extract of the bacterium and its superantigens induced higher immunoreactivity in BD lymphocytes in comparison to the healthy control immune cells. Animal experiments utilizing the whole bacteria or their capsular lipoteichoic acid induced acute multi organ infectious/inflammatory reactions, septic shock, and noninfiltrative short-term uveitis. The failure to reproduce an experimental model of BD led to search for other bacterial derived components as causative agents in BD. So far, no bacterial model has found uniform recognition [3].

Viral Infectious Models

Hulusi Behçet in his historical description of BD in 1937 proposed that the syndrome might be caused by a viral infection. For many decades, efforts were made to confirm this hypothesis. The results of extensive data collected regarding the significance of herpes simplex virus (HSV) in BD including the detection of anti-HSV antibodies, viral DNA expression, and antiherpetic therapeutic trials were controversial [4]. In 1998, Sohn et al. [5] reported that inoculation of HSV type I at the earlobe of ICR mice produced a BD-like disease in approximately 50% of the animals, including genital and oral ulcers, skin and eye lesions, arthritis, and gastrointestinal involvement. This model was induced in other mice strains including B10.BR (MHC H-2k), B10.RIII (H-2r), C57BL/6 (H-2b), C3H/He (H-2k), and Balb/c (H-2d) [6]. Symptoms developed in 40–50% of B10.BR, B10.RIII, and C57BL/6m but in only 2% of C3H/He and Balb/c. The lack of correlation between H-2 type and disease frequency cannot support the concept that MHC genetic phenotype is involved in the pathogenesis of this model of BD. This model has a high mortality rate; 30% of the infected mice and only 50% of the surviving mice develop some signs resembling BD. The disappointing results of therapeutic trials with antiherpes virus drugs in BD do not support the possibility that BD is a subtype of active chronic HSV infection [4]. This does not rule out the possibility that HSV infection can serve as a trigger to initiate the immunological dysregulation leading to the development of BD. Finally, this model provides extensive data about the inflammatory aspects of BD and serves as an experimental model to assess therapeutic modalities for human disease [7,8].

Autoimmune Models

Heat Shock Proteins

Heat shock proteins (HSP) are intracellular chaperone molecules with scavenger properties that are expressed in cells upon various stress stimuli [9]. The microbial HSP 65 kDa and the animal HSP 60 share a significant homology (over 50%). It was found that various antibodies directed to amino acid sequences of HSP 65 are cross-reactive with the human HSP 60 expressed in active lesions of BD. Moreover, T cells of BD patients from different ethnicity were highly reactive to HSP sequences, and the immunodominance hierarchy of these sequences differed from the pattern in healthy controls. Subcutaneous immunization of rats with human 60-kDa HSP-derived peptide 336-351 induced clinical and/or histological uveitis in 80% of rats. Subsequent experiments to prevent the development of uveitis by oral or nasal administration of the peptide have failed. Instead, uveitis was induced in 75% of rats when given the peptide orally, in 75% when given nasally, and 92% of those administered the peptide by both routes. Examination of mRNA from CD4-enriched splenic cells failed to yield significant differences in Th1 or Th2 cytokines. Treatment with monoclonal antibody (mAb) to CD4 yielded a dose-dependent decrease in uveitis from 82 to 25%. Similarly, treatment with IL-4 significantly decreased the development of uveitis from 68 to 30%. Conversely, treatment of the rats with mAb to CD8 greatly enhanced the onset of uveitis (from about 22 days in the controls to 11 days after immunization) and all the rats developed uveitis by day 24. Thus, CD4+ cells mediate, whereas CD8+ cells suppress the development of uveitis in this model. It was suggested that this experimental mucosal model of induction of uveitis by the human 60-kDa HSP-derived peptide is consistent with the oro-genital onset of BD and the development of uveitis [10]. This model can be categorized as an organ specific auto-immune model for BD.

S-Ag-Induced Uveitis

Retinal S-Ag is an immunologically sequestered protein existing mainly in the photoreceptor region of the retina. It is used for the induction of the classical model of experimental autoimmune uveitis [11]. The sera of BD patients among other patients with uveitis had antibodies directed against S-Ag, and their T cells were recognized and activated by this protein. Of high importance was the finding that an S-Ag epitope (aa 342–355) designated PDS-Ag shared homology to a conserved sequence in the HLA-B molecules (aa 125–138) designated B27PD. Immunization of rats with both peptides caused uveitis [12] supporting the concept of anti-HLA autoimmunity in the pathogenesis of BD. Activation of peripheral CD4+ T cells with these peptides occurred only in HLA B51 positive BD patients with posterior uveitis but not in patients without eye involvement. This implied that the normal tolerance to self-HLA class I epitopes is preserved in BD, and additional conditions are needed for its breakdown and for the development of posterior uveitis.

This model can be categorized together with the HSP model, as an organ specific auto-immune model.

Tropomyosin

Our group has shown that sera of patients with BD contain IgG antibodies directed to α-tropomyosin (TPM) protein, a component of the contractile apparatus of the muscles. Vaccination of Lewis rats with TPM emulsified in complete Freund’s adjuvant (CFA) caused an inflammatory disease with involvement of the skin, joints, and eyes. Infusion of an anti-TPM-directed T-cell line derived from the draining lymph node lymphocytes of the TPM-vaccinated rats induced a similar pathology [13]. The cytokine profile of pathogenic cells had a Th1 pattern. The model was used to test the therapeutic effects of lactobacillus GG [14]. In this model, we also analyzed the membranal fatty acids composition induced by probiotic bacteria consumption. We noticed a shift in the ratio of n−3/n−6 poly unsaturated fatty acids (PUFA) toward reduction in PUFA that can serve as precursors of pro-inflammatory prostaglandins (submitted for publication).

Transgenic Model

The paradigm of ethnic and genetic predisposition in BD is widely accepted. The discovery of the association between BD and the HLA Class I molecule B51 [15], led Takeno et al. [16] in 1995 to produce a transgenic (Tg) mouse model. The production of Tg was an important step in the attempt to elucidate the role of the genetic marker HLA B51 in the pathogenesis of BD. They inserted the human HLA B*5101 gene into C3H/He mice. The neutrophils of the Tg mice produced excessive superoxide similar to the documented phenomenon in BD patients. However, no clinical signs of the disease developed. Based on these results, it remains unknown whether HLA B51 is just a marker and other gene/s with linkage disequilibrium to its locus are involved in the pathogenesis of BD, whether the HLA B51 molecule is essential but insufficient for the development of BD, or if the mouse strain used was resistant for the development of active disease. We have failed to induce BD-like disease in these mice (Received from Prof. M. Takiguchi, Kumamoto University School of Medicine, Japan.) by vaccinating some antigens, mentioned above [17], which were proposed to be involved in the autoimmune aspects of BD (unpublished data). One should note that the Tg construct contained only the HLA-B*5101’s heavy chain without the coupled binding arm of the molecule, the β2-microglobulin. Thus, the fact that the Tg mice did not develop a BD like disorder spontaneously or upon antigenic challenges does not refute this hypothesis. The development of Tg mice with the complete molecule is warranted.

In an attempt to combine both assumptions – that BD is an HLA-B51-dependant autoimmune disorder – we used bio-informatic methods to estimate the potential role of B*5101 and the above antigens in the pathogenesis of BD based on the aforementioned animal models. Briefly, it is commonly accepted that the T cells recognize a particular peptide presented by the HLA molecule. For that purpose, three elements are required. Firstly, T-cell receptors (TCR) should have potential binding capacity to the amino acid sequences of the peptide. Secondly, the presented peptides should contain a particular motif that can be anchored by the HLA binding sites, and finally, the three dimensional-structure of the peptide should be recognized by the HLA molecule. Because of its steric structure, the binding sites of HLA class I molecules can bind only short peptides with a length of up to 10 amino acids. We searched for 9-mer peptide candidates for antigenic peptide motifs derived from the antigens proposed to induce experimental BD in animals. For that purpose, we used computerized programs that rank sequences of peptides according to their predicted half-time dissociation coefficient from the HLA and rat MHC class I molecules. The binding capacity of the human HLA-B*5101 and the corresponding rat class I molecule, designated MHC RT1.Al, to the following proteins or peptides, was studied: HSP-65, HSP-65 (aa 336–351), MICA, HLA-B*5101, Retinal S-Antigen, Sort sequence of Retinal S-Antigen (aa 342–355) designates PDS-Ag, short peptide HLA-B27 (aa 125–138) and human-Tropomyosin. In each protein examined, several short sequences with potential high binding capacity were found, with the exception of the short peptide B-27PD that has no binding capacity motif to HLA-B*5101 but only to rat molecule MHC RT1.Al. The peptide designated PDS-Ag has a potential binding capacity to HLA B*5101 but not to MHC RT1.Al. In addition, it was found that the Tropomyosin short peptide T2 has the highest predicted binding capacity to the human and rat class I molecules compared to the other TPM-derived peptides. This finding is in accordance with the clinical disease severity caused by TPM T2 in the animal model [13]. On the other hand, this information is still theoretical; in order to prove this concept, it should be tested by immunization of Tg animals bearing the complete HLA-B51molecule with each of the proteins and peptides mentioned above.

Comments (Summarized in Table 16.1)

The environmental pollution model is conceptually interesting since the pigs developed multisystem symptoms similar to BD, but the model has limitations to become utilized as a model for the disease since it is difficult to produce and the onset of symptoms appears erratically in a wide time range of 4–10 months. Moreover, the failure to show increased levels of the offending pollutants in BD patients raises questions as to its relevance. The Streptococcal models have similarity only to the eye involvement in BD. Eliciting autoreactivity to HSP in the animals contributed to the understanding the potential reactive autoimmune component of BD. This model is simple to induce with high rate of homogeneity. The HSV model has multisystem manifestations resembling BD; it has a moderate reproducibility since 30% of the inoculated mice die upon induction and low homogeneity. The use of human live virus demands special laboratory facilities. The autoimmune model utilizing S-Ag is a monosymptomatic model of BD-like uveitis. This model is easy to induce and extensive studies elucidated some of the immunological characteristics of BD including the paradigm of anti-HLA autoimmunity. The TPM model shares some clinical features of BD. This model has a potential to become a useful autoimmune model for BD. The only published trial to establish a Tg model for BD did not show any significant similarity to the human disease except hyper-responsiveness of neutrophils.

Table 16.1 Comparison summary table
Full size table