Chronic thyroiditis histologically identical to that in Hashimoto's thyroiditis occurs spontaneously in Obese strain (OS) chickens (172), beagles (173), mice, and rats. It can be induced in dogs (174), mice, rats, hamsters, guinea pigs, rabbits, monkeys (175), and baboons (176) by immunization with autologous or allogenic thyroid homogenate mixed with adjuvants, or by using heterologous TG (177), or TG that has been arsenylated or otherwise chemically modified (1). An important thyroiditogenic epitope includes a thyroxine residue (aa 2553) in human TG (178, 179) but the role of iodination at this site is unclear and may depend on the type of T cell assay system used, as well as other parameters (180).

Induced thyroiditis leads to formation of humoral antibodies and T cell- mediated immunity. Usually the histologic pattern conforms to that of T cell-mediated immunity (181). The role of TG antibodies is unclear but minor. An idiotype-anti-idiotype network exists for TG antibodies in mice but the induction of those antibodies does not lead to thyroiditis (182). Furthermore, the intensity of the thyroiditis correlates better with T cell mediated immunity than with antibody levels, and can be transferred by T cells but not antibodies, and both CD4 ⁺and CD8 ⁺T cells are usually needed for transfer (183). In normal mice, thyroiditis can be produced by immunization with mouse TG in adjuvant, and transferred to isogenic animals by sensitized Ly-1+ T cells. The same cells, given before immunization, vaccinate against the development of thyroiditis during subsequent immunization (184).

However, a subpopulation of CD4 ⁺T cells has an important regulatory role in tolerance to murine TG, keeping in check those TG-reactive T cells which escape thymic deletion and peripheral anergy-inducing mechanisms (185). Amelioration of thyroiditis by oral administration of TG (186) operates through enhancing the activity of these regulatory T cells although other mechanisms are possible. Recent studies have emphasised the importance of regulatory T cells in suppression of thyroiditis in animals immunised with TG. In particular, semi-mature dendritic cells, which can be induced with granulocyte-macrophage colony stimulating factor, can induce the function of TG-specific CD4 ⁺,CD25 ⁺T cells which can suppress thyroiditis through the production of IL-10 (187, 188).

Another recent model has used homologous (murine) TPO in an immunization protocol and this method established thyroiditis and TPO antibody production although none of the immunized mice developed hypothyroidism (190). Another unique model is the creation of HLA-DRB1*0301 (DR3) transgenic mice which are susceptible to thyroiditis induced by TG immunization, unlike DR2 transgenics, thus confirming that HLA-DRB1 polymorphism determines susceptibility to autoimmune thyroiditis, and his model has been extended to study of the immune response to TSH-R, with results again showing the importance of the DR3 specificity (191). However, when the modeling has attempted to reproduce Graves’ disease by immunization of mice with adenovirus expressing the TSH-R, it is non-MHC genes which play a major role in controlling the development of hyperthyroidism (192). This concurs with the polygenic susceptibility and rather weak effect of HLA-DR3 in Graves’ disease. This model has also been used recently to show that dietary iodide enhances the development of thyroid disease and depletion of CD4 ⁺, CD25 ⁺Tregs exacerbates this iodide-induced thyroiditis (193a)

Spontaneous thyroiditis in OS chickens more closely resembles Hashimoto’s thyroiditis than the immunization models just discussed, particularly as the birds develop hypothyroidism as a consequence of the autoimmune process. Some evidence suggests that the thyroid of the newly hatched chick is intrinsically abnormal, since its function is partially nonsuppressible by thyroid hormone and this constitutes an important element of the genetic susceptibility of these birds, together with genes controlling T cell responses and possibly glucocorticoid tonus. The MHC conversely has only a limited effect. Iodine plays a critical role in the induction of thyroid injury in OS chickens, most likely through the generation of reactive oxygen metabolites, and this injury is an early event, preceding lymphocytic infiltration (193). Iodination of TG is a second path by which iodine influences disease in OS chickens, as autoreactive T cells respond to the antigen only if it is iodinated (194).

Lymphocytic thyroiditis occurs spontaneously in the Buffalo and BB/W rat strains and the NOD mouse line (195). In both species, there are associated abnormalities in the animals' immune system. As in the OS chicken, administration of excess iodine augments the incidence of rat thyroiditis and iodine depletion reduces it (196).

A third kind of model is produced by manipulation of T cells. The original description of thyroiditis in genetically susceptible rats by sublethal irradiation and thymectomy (197) has been followed by a number of more refined models in which T cell subsets can be perturbed more or less specifically to induce disease (181). For instance CD7/CD28 double-deficient mice have impaired Treg function and such animals develop spontaneous thyroiditis after 1 year of age (198). These experiments clearly demonstrate the recurrence of autoreactive T and B cells in normal animals and show that any of a number of factors which can perturb the regulation of these could result in autoimmune thyroiditis (Fig. 7-10). The most elegant model resulting from T cell manipulation is the generation of transgenic mice expressing a human T cell receptor specific for a TPO epitope, which resulted in a spontaneous destructive hypothyroidism and hypothyroidism (199). The CD8 T cells recognizing the epitope in these animals unconventionally were MHC class II rather than class I restricted and it is unclear whether this atypical behavior is significant to the creation of the model, nor is it yet clear what the mechanism is for thyroid cell destruction.

Another intriguing model is the recent description that necrotic thyroid cells can induce maturation of dendritic cells in vitro, and when injected back into autologous mice EAT is induced, with a lymphocytic thyroiditis and TG-specific IgG (200). It is not clear whether this protocol yields cryptic TG epitopes which can break tolerance. It is possible that such work could be in a sense be reversed to allow attenuation of EAT by pulsing tolerogenic dendritic cells.

Establishing an animal of Graves’ disease has been surprisingly difficult despite the cloning of the TSH-R. However, immunization of AKR/N mice (but not other strains sharing the same MHC haplotype) with murine fibroblasts doubly transfected with the human TSH-R and haploidentical MHC class II genes results in a syndrome similar to Graves’ disease except that thyroid lymphocytic infiltration was not induced (201), whereas thyroiditis is a feature of immunization with the TSH-R (202). A similar approach has been used to generate TPO antibodies using fibroblasts transfected with TPO (203). This is a promising model although its exact physiological parallel remains unclear, particularly as fibroblasts may behave differently to TECs in terms of antigen presentation. This is because the fibroblasts used express the critical costimulatory molecule B7-1 and also because the procedure causes generalized in vivo immune activation. This model is therefore not evidence that thyroid follicular cells (which do not normally express B7) could initiate thyroid autoimmunity.

One unexpected finding has been the observation that mice with a TSH-R knockout do not differ in their response to immunization with TSH-R when compared to healthy animals, whereas the expectation was that such animals would have no tolerance to this autoantigen (as it had been absent throughout development) and therefore a greater immune response would be predicted (204). This suggests that thymic (central) tolerance is not a critical step in self tolerance to this autoantigen. An excellent summary of the current state of the field with regard to animal models of Graves’ disease has recently been published (202).

Balb/c strain mice appeared to develop orbital changes suggestive of ophthalmopathy when given TSH-R primed T cells derived from donor mice immunized with TSH-R protein or cDNA but this model has not proved reproducible by the original authors, for reasons which are not yet clear, although complex histological artefacts may be part of the answer (205). Another novel model has used the hamster rather than mouse-immunization with TSH-R-transfected CHO cells, co-expressing MHC class II molecules, produced mainly blocking antibodies, whereas MHC class II negative cells induced, rarely, TSAb together with a focal lymphocytic infiltrate (206).

One general concept derived from all of these studies is that a genetically controlled balance of helper and suppressor T cell function is needed to prevent autoimmunity, and that a variety of perturbations leads to onset of the disease.