The exact pathogenic mechanism of thyroid-associated ophthalmopathy (TAO) remains unclear, and extensive studies on this disorder have resulted in often conflicting data. Well-known technical difficulties including the limited access to orbital tissues from patients with active and early disease, lack of an animal model and poor reproducibility of some of the immunological techniques used are in part responsible for this confusing situation. Despite this there is considerable evidence for eye muscle (EM) tissue involvement in the autoimmune reactions of TAO. Although the primary EM antigen(s) recognized by immunocompetent cells and autoantibodies has not been definitely identified, some good candidates, among them a membrane antigen of 64 kD which is also expressed in the thyroid, have been partially characterized. While it is unclear which component of the autoimmune reaction against EM-humoral or cell mediated-plays the more important role, autoantibodies seem to be responsible at least in part for the clinical features of the eye disorder. On the other hand, the orbital connective tissue (OCT) cells, especially the fibroblasts surrounding the EM fibers, seem to be extremely sensitive to stimulation by cytokines and other soluble proteins and immunoglobulins released in the course of an immune reaction in the muscle cells. Fibroblasts secrete large amounts of glycosaminoglycans and also participate in maintaining the autoimmune reaction. It seems likely that the EM is the main and primary target of the orbital autoimmune process whereas inflammation of the OCT is probably secondary.