Autoimmunizing mechanisms are very hard to study in humans, so we have focused on vital clues in thymomas and hyperplastic thymuses in myasthenia gravis (MG). According to our multi-step hypothesis: thymic epithelial cells (TEC) present epitopes from the isolated acetylcholine receptor (AChR) subunits they express, and autoimmunize helper T cells; subsequently, these evoke "early antibodies" that then attack rare thymic myoid cells expressing intact AChR; in the resulting germinal centers, autoantibodies diversify to recognize native AChR. We have studied: 1) thymomas, to identify autoimmunizing cell types, focusing on IFN-alpha, against which many patients have high titer autoantibodies, as in another highly informative autoimmune syndrome. Although IFN-alpha is much easier to label than the sparse and delicate AChR subunits, we have not yet located obviously autoimmunizing micro-environments; 2) hyperplastic MG thymuses, where we find (a) upregulation of complement receptors and regulators on hyperplastic TEC and deposition of activated C3b complement component on them, (b) absence of complement regulators from almost all myoid cells, indicating vulnerability to attack, and (c) deposition of C3b, and even of the terminal membrane attack complex, especially on the myoid cells close to the infiltrating germinal centers. The changes are very similar in over 50% of the so-called seronegative patients with generalized MG (SNMG) but without detectable autoantibodies against AChR or MuSK, consistently with other evidence that they belong to the spectrum of AChR-seropositive MG. Together, moreover, our findings implicate both myoid cells and TEC in autoimmunization, and thus strongly support our hypothesis.