As most autoimmune diseases, inherited predisposition to Graves' disease (GD) is polygenic with the main contributory genes being located in the HLA region. Also, as in other autoimmune diseases, family linkage, candidate gene association, and GWAS studies have identified an expanding number of predisposing genes (CTLA4, CD40, PTPN22...) and 2 of them, TG and TSHR, are thyroid specific. In spite of this expanding number of associated genes, it has been estimated that all together they account for only a 20% of the heritability of GD. TSHR is of special interest as it codes for the target of TSHR stimulating antibodies (TSAbs), which are unequivocally pathogenic and an exception in autoimmunity by being stimulating rather than neutral, blocking, or cytotoxic. This is surprising because the generation of stimulating TSHR antibodies by immunisation of laboratory animals has been remarkably difficult, suggesting an underlying mechanism that favours stimulating over neutral or blocking anti-TSHR antibodies must be operating in GD patients. Besides, after HLA, TSHR is the gene most tightly associated to GD. The TSHR polymorphisms conferring susceptibility are located in the unusually large intron 1. Two mechanisms have been already put forward to explain its association with GD. According to one, the risk alleles determine an increase in the expression of TSHR mRNA splice variants that code for a soluble form of the receptor. The wider distribution of soluble TSHR would favour its immunogenicity and the development of an autoimmune response to it. It does not explain why it becomes immunogenic, as immunogenicity and distribution are not necessarily connected, nor why the immune response focus to the production of stimulating antibodies. According to the second mechanism proposed, the risk alleles determine a lower TSHR expression in the thymus and this would favour the escape of more TSHR reactive T cells, that is, central tolerance failure. The unexpected finding that thymocytes express TSHR and that TSAbs stimulate them lead to postulate that this would accelerate their egress from the thymus and a less efficient deletion of the TSHR self-reactive T cells. It can be envisaged that these autoreactive T cells may enhance the production of TSHR-Abs in the germinal centres of the thyroid draining lymph nodes, especially of those capable of further stimulating the egress of autoreactive T cells from the thymus. This mechanism, which does not exclude the former, provides and insight of the way in which TSAbs are favoured over neutral or blocking antibodies. Finally this would explain the frequent finding of thymic hyperplasia in GD patients.