Autoimmune hypophysitis: mouse and human studies

Pituitary autoimmunity encompasses a spectrum of conditions ranging from histologically proven autoimmune hypophysitis to the presence of pituitary antibodies in apparently healhy subjects. Autoimmune hypophysitis (AH) is a rare but increasingly recognized disease of the pituitary gland that still needs a full characterization. The autoantigens target of the autoimmune attack are unknown. The diagnosis of AH is difficult because it is often misdiagnosed as a nonsecreting pituitary adenoma. Pituitary antibodies, a marker of pituitary autoimmunity, are measured by non-antigen specific methods so that their determination has limited clinical value. Moreover, a reliable mouse model of AH has not been established. Aims of the study were: 1) to establish a new mouse model of AH 2) to determine the prevalence of pituitary autoimmunity in endocrine patients and in healthy subjects 2) to search for novel antigens in AH using human sera. By immunizing female SJL/J mice with mouse pituitary extracts, we established a new mouse model of experimental AH. Immunized mice developed severe lymphocytic infiltration in the anterior pituitary that closely mimicked the human pathology. In the florid phase of the disease hypopituitarism developed, whereas in later stages of the disease the pituitary gland became atrophic as also demonstrated by the MRI finding of “empty sella”. By measuring pituitary antibodies using indirect immunofluorescence we found that pituitary autoimmunity prevalence is low in normal subjects, is moderately increased in normal pregnancies and in pituitary adenoma patients, and is significantly increased in patients with autoimmune thyroid diseases and in postpartum thyroiditis. By testing the pituitary function in patients with the combination of autoimmune thyroid diseases and pituitary autoimmunity we found a certain grade of GH deficiency in 35% of the patients. By measuring pituitary antibodies in a large cohort of patients with AH, using immunoblotting and a proteomic approach, we were able to identify several proteins that could act as autoantigens in AH. Of these, the most interesting candidate is the chorionic somatomammotropin, that need to be confirmated with further studies. Finally we found that immunoblotting has slighter greater sensitivity and specificity than immunofluorescence in predicting histologically proven AH.