Kalman Kovacs

Microadenomas of the Pituitary and Abnormal Sellar Tomograms in an Unselected Autopsy Series

PROLACTIN-SECRETING microadenomas of the pituitary gland have been diagnosed on the basis of elevated serum prolactin concentrations and abnormal tomograms of the sella turcica.1 , 2 This diagnosis...

Proliferative Activity and Invasiveness among Pituitary Adenomas and Carcinomas: An Analysis Using the MIB-1 Antibody

Although histologically benign, one-third of all pituitary tumors will be invasive of surrounding structures. In this study, the relationship between the proliferative activity in pituitary adenomas and their invasiveness was investigated. Invasion was defined as gross, operatively or radiologically apparent infiltration of dura or bone. Using the recently developed MIB-1 monoclonal antibody, which recognizes the Ki-67 cell cycle-specific nuclear antigen, the growth fractions of 37 noninvasive adenomas, 33 invasive adenomas, and 7 primary pituitary carcinomas were determined. All tumors were fully classified by histology, immunohistochemistry, and electron microscopy. The mean Ki-67 -derived growth fractions for noninvasive adenomas, invasive adenomas, and pituitary carcinomas were 1.37 +/- 0.15%, 4.66 +/- 0.57%, and 11.91 +/- 3.41%, respectively (mean +/- standard error of the mean). An analysis of variance and then individual pairwise comparisons confirmed significant differences in the mean Ki-67 labeling index between each of the three tumor groups (P < 0.01). The mean growth fraction of hormonally active pituitary adenomas (3.25 +/- 0.26%) was significantly higher than that for nonfunctioning adenomas (2.06 +/- 0.23%) (P = 0.03). Establishing a threshold labeling index of 3% served to distinguish invasive from noninvasive adenomas with 97% specificity and 73% sensitivity and was associated with positive and negative predictive values of 96 and 80%, respectively. Although invasive pituitary tumors exhibited significantly higher growth fractions than did noninvasive tumors, there were individual exceptions, indicating that in a subpopulation of invasive pituitary tumors, factors other than proliferative activity determine invasive potential.

p53 Expression in Pituitary Adenomas and Carcinomas: Correlation with Invasiveness and Tumor Growth Fractions

Although most pituitary tumors are well differentiated, histologically benign neoplasms, their clinical behavior is known to vary greatly. These lesions are relentlessly aggressive in some instances yet biologically indolent in others, but these prognostically relevant differences in behavior are not reflected in their histopathological appearance. As a means of identifying intrinsically aggressive pituitary tumors, we evaluated 70 pituitary adenomas and 7 primary pituitary carcinomas for their expression of the p53 gene product, a nuclear phosphoprotein whose immunohistochemical accumulation has served as an unfavorable prognostic factor for a wide range of human neoplasms. All tumors were fully classified by immunohistochemistry and electron microscopy; adenomas were further stratified on the basis of their invasion status, the latter being defined as gross operatively or radiologically apparent infiltration of dura or bone. Conclusive nuclear immunopositivity for p53 was identified in a total of 12 tumors, all being either invasive adenomas or primary pituitary carcinomas. A clear and highly significant association was evident between p53 expression and tumor behavior, as the proportion of p53-positive cases among noninvasive adenomas, invasive adenomas, and pituitary carcinomas was 0, 15.2, and 100%, respectively (chi 2 = 44.72; degrees of freedom, 2; P << 0.001). A comparison of previously reported growth fraction data with p53 expression indicated that the mean Ki-67-derived growth fraction of p53-positive tumors was significantly higher than that of p53-negative tumors (10.41 +/- 2.20 versus 2.51 +/- 0.28%) (+/- standard error of the mean, two-sample t test for independent samples, P = 0.004). There was no apparent relationship between the functional status of the tumor and p53 expression; positivity was observed among somatotroph, lactotroph, corticotroph, and clinically nonfunctioning pituitary tumors. These data indicate that p53 expression, when conclusively present in pituitary tumors, may be of some diagnostic usefulness as a marker of biologically aggressive behavior.

Ectopic Secretion of Corticotropin-Releasing Factor as a Cause of Cushing's Syndrome: A Clinical, Morphologic, and Biochemical Study

Corticotropin-releasing factor, a hypophyseo-tropic hormone that stimulates adrenocorticotropic hormone (ACTH) secretion, has recently been isolated, characterized, and synthesized in the sheep and rat. We report on a patient with metastatic carcinoma of the prostate presenting with anterior and posterior pituitary hormone deficiency together with ACTH-dependent Cushings syndrome. At postmortem examination, large areas of the median eminence and pituitary stalk were replaced by tumor, but the corticotrophs were markedly hyperplastic. Immunostaining of tumor cells was positive for corticotropin-releasing factor and was negative for ACTH and a wide range of other hormones. Radioimmunoassay and bioassays showed that tumor extracts and further purified fractions were active in corticotropin-releasing factor, and the tumor material coeluted with corticotropin-releasing factor on high-pressure liquid chromatography. These studies demonstrate that ectopic secretion of corticotropin-releasing factor is a cause of Cushings syndrome in human beings. The features of this syndrome include hypercortisolism, pituitary corticotroph hyperplasia, elevation of circulating ACTH levels, and failure to suppress the pituitary-adrenal axis with exogenous glucocorticoids.

Lymphocytic hypophysitis of pregnancy resulting in hypopituitarism: a distinct clinicopathologic entity.

Two patients presented with abnormalities suggestive of pituitary adenoma; one during pregnancy and one in the postpartum period. However, pathologic examination of the pituitary showed extensive destruction by a lymphoplasmacytic infiltrate; no tumor was identified. Both patients developed hypopituitarism. We know of eight additional cases of lymphocytic hypophysitis, seven of which have been reported in the literature. In only three cases, including the two reported here, the diagnosis was established by biopsy. In each of those cases, the entity mimicked a pituitary tumor. This is the first report of electron microscopy of this lesion and the ultrastructural features support the previously suggested autoimmune etiology. The lesion has been described only in women and seven of ten patients were pregnant or postpartum at the onset. This fact and previously reported experimental evidence, including the identification of anti-prolactin cell antibodies, support our suggestion that lymphocytic hypophysitis associated with pregnancy represents a distinct clinicopathologic entity.

Cushing's Disease Associated with an lntrasellar Gangliocytoma Producing Corticotrophin-Releasing Factor

A 58-year-old woman had Cushings disease with elevated plasma adrenocorticotrophin and an intrasellar tumor. Light microscopy showed that the tumor was a gangliocytoma containing immunoreactive corticotrophin-releasing factor accompanied by pituitary corticotroph hyperplasia. Ultrastructural examination identified an intimate association and desmosomal attachments between interdigitating cell processes of neurons and corticotrophs. It is suggested that Cushings disease was due to the effect of corticotrophin-releasing factor on corticotrophs; this case represents a syndrome supporting the concept that, in some patients, Cushings disease may have a hypothalamic origin.

p21Cip1 restrains pituitary tumor growth

As commonly encountered, pituitary adenomas are invariably benign. We therefore studied protective pituitary proliferative mechanisms. Pituitary tumor transforming gene (Pttg) deletion results in pituitary p21 induction and abrogates tumor development in Rb+/−Pttg−/− mice. p21 disruption restores attenuated Rb+/−Pttg−/− pituitary proliferation rates and enables high penetrance of pituitary, but not thyroid, tumor growth in triple mutant animals (88% of Rb+/− and 72% of Rb+/−Pttg−/−p21−/− vs. 30% of Rb+/−Pttg−/− mice developed pituitary tumors, P < 0.001). p21 deletion also accelerated S-phase entry and enhanced transformation rates in triple mutant MEFs. Intranuclear p21 accumulates in Pttg-null aneuploid GH-secreting cells, and GH3 rat pituitary tumor cells overexpressing PTTG also exhibited increased levels of mRNA for both p21 (18-fold, P < 0.01) and ATM (9-fold, P < 0.01). PTTG is abundantly expressed in human pituitary tumors, and in 23 of 26 GH-producing pituitary adenomas with high PTTG levels, senescence was evidenced by increased p21 and SA-β-galactosidase. Thus, either deletion or overexpression of Pttg promotes pituitary cell aneuploidy and p53/p21-dependent senescence, particularly in GH-secreting cells. Aneuploid pituitary cell p21 may constrain pituitary tumor growth, thus accounting for the very low incidence of pituitary carcinomas.

The World Health Organization classification of adenohypophysial neoplasms. A proposed five-tier scheme.

Although numerous attempts have been made, the classification of pituitary neoplasms remains controversial.

'spindle cell oncocytoma' of the adenohypophysis: A tumor of folliculostellate cells?

We describe five primary tumors of the adenohypophysis featuring mitochondrion-rich spindle cells. The patient ages ranged from 53 to 71 years (mean 61.6 years); two were female. All presented with panhypopituitarism. Two also had visual field defect. On neuroimaging all tumors showed suprasellar extension and were indistinguishable from pituitary adenoma. None showed imaging or operative evidence of dural involvement. All were gross totally removed: four by transsphenoidal surgery and one by frontal craniotomy. Follow-up ranged from 2 to 68 months (mean 35.4 months). No recurrences were noted. The clinical workup was noncontributory in all but two patients: one (case no. 4) with an oncocytic thyroid adenoma and another (case no. 5) with squamous carcinoma of both the uterine cervix and of vocal cord. Histologically, the five tumors were composed mainly of fascicles of spindle cells with eosinophilic, granular cytoplasm. Mitoses were rare and necrosis was absent. Neoplastic cells were immunoreactive for vimentin, epithelial membrane antigen, S-100 protein, and galectin-3. Stains for pituitary hormones, synaptophysin, chromogranin, glial fibrillary acidic protein, cytokeratin CAM5.2, smooth muscle actin, CD34, and CD68 were negative. No thyroglobulin immunoreactivity was noted in the tumor of case no. 4. Ultrastructurally, the neoplastic cells contained numerous mitochondria with lamellar cristae. The neoplastic cells were linked by intermediate junctions and desmosomes. No secretory granules were noted. The histologic, immunohistochemical, and fine structural features of these tumors were unlike those of pituitary adenoma or any other primary sellar tumor. A derivation from adenohypophyseal folliculostellate cells is suggested.

Human Fetal Adenohypophysis

The pituitary glands were removed from 63 human fetuses from 5 weeks of gestation to term and studied by electron microscopy and ultrastructural immunocytochemistry to document the development of cell