Severe Obesity Associated With Pituitary Corticotroph Hyperplasia and Neoplasia.

OBJECTIVE: To investigate the incidence of corticotroph hyperplasia (CH) or lymphocyte infiltration in the pituitary of patients with obesity. METHODS: The pituitary and adrenal glands from 161 adult autopsies performed between 2010 and 2019 at our institution were reviewed. The clinical history, body mass index (BMI), and cause of death were recorded. Routine hematoxylin and eosin staining, reticulin staining, and immunohistochemical staining for adrenocorticotropic hormone, CD3, and CD20 were performed. The results were analyzed using the Fisher and chi-square statistics. Decedents were separated into 4 groups based on BMI (kg/m2): (1) lean (BMI, <25.0), (2) overweight (BMI, 25.0-29.9), (3) obesity class I (BMI, 30.0-34.9), and (4) obesity classes II to III (BMI, >34.9). RESULTS: CH/neoplasia was identified in 44 of 161 pituitary glands. Four (9.1%) of 53 lean patients had pituitary lesions, whereas 27.3% (12) of overweight, 22.7% (10) of obesity class I, and 40.9% (18) of obesity class II patients had hyperplasia (P < .0001). Small corticotroph tumors were identified in 15 patients; only 1 was a lean patient, and the tumor was associated with the Crooke hyaline change of nontumorous corticotrophs. The presence of CH and neoplasia was associated with adrenal cortical hyperplasia and lipid depletion. Microscopic foci of T and B lymphocytes were identified in the pituitaries of patients in each weight category; no independent association between BMI and lymphocyte inflammation was found. CONCLUSION: Our data indicate an association between CH/neoplasia and obesity. It remains unclear whether obesity is the cause or effect of adrenocorticotropic hormone and cortisol excess.

Follicular cells in pituitary neuroendocrine tumors.

Follicular cells (FCs) are thought to be agranular, non-hormone-producing stellate cells distributed throughout the adenohypophysis, occasionally arranged around colloid-filled follicles, and thought to be more prominent in the vicinity of necrosis and apoptotic cells. A distinct but similar cell type, the folliculostellate cell (FSC), is a sustentacular cell that is negative for keratins and stains for S100, GFAP, and SOX10. While several studies have examined FSCs in pituitary neuroendocrine tumors (PitNETs), the distribution and derivation of FCs in these lesions is unclear. We examined the presence and distribution of FCs in 104 PitNETs obtained by trans-sphenoidal surgery, using immunohistochemistry for keratins as well as the full complement of immunohistochemical stains for tumor characterization. The tumors included 9 somatotroph, 5 mammosomatotroph, 7 lactotroph, 7 immature PIT1-lineage, 2 acidophil stem cell, 17 corticotroph, 53 gonadotroph, 2 null cell, and 2 unusual plurihormonal tumors. CK-positive FCs were only identified in gonadotroph PitNETs and were found in 12 (23%) of those tumors; all other tumor types were negative for FCs. FCs express keratins identified by CAM5.2, AE1/AE3, CK18, and CK19 antibodies. FCs were identified scattered singly among hormone-producing neuroendocrine cells, in small clusters of 3-5 cells and surrounding colloid-filled follicles, as well as linearly along intratumoral blood vessels. Sequential stains showed that FCs express nuclear SF1 and GATA3, transcription factors of gonadotrophs, and multiplex immunohistochemistry confirmed colocalization of SF1 in the nucleus of keratin-positive FCs. In this series, FCs were exclusively found in gonadotroph PitNETs and occurred in 23% of those tumors. Co-expression of gonadotroph transcription factors in FCs supports the concept of cellular plasticity and transformation of neoplastic hormone-producing neuroendocrine cells to FCs. Further studies are required to determine if and why gonadotrophs alone undergo this transformation, the function of these cells and whether they have prognostic value.