The Pituitary Gland in Hyperthyroidism

BERND W. SCHEITHAUER, M.D., Section of Surgical Pathology; KALMAN T. KOVACS, M.D., Ph.D., Department ofPathology, St. Michael's Hospital, Toronto, Ontario, Canada; WILLIAM F. YOUNG, Jr., M.D., RAYMOND V. RANDALL, M.D.,* Division of Endocrinology/Metabolism and Internal Medicine . The pituitary glands of 33 patients (24 women and 9 men, 18 to 78 years old) who died in thyrotoxicosis (18 with Graves' disease and 15 with toxic multinodular goiter [Plummer's disease]) were examined by histologic and immunocytologic methods. Thirteen patients (39%) died in "thyroid storm." The avidin-biotin-peroxidase complex immunostaining method was used to demonstrate the spectrum of pituitary hormones, including growth hormone, prolactin, adrenocorticotropic hormone, thyrotropin, follicle-stimulating hormone, luteinizing hormone, and a-subunit. .. The most striking finding was a pronounced decrease or loss of immunoreactivity to thyrotropin in all thyrotoxic cases, a consistent change that allowed ready distinction of thyrotoxic from euthyroid pituitary glands. When immunoreactive thyrotrophs were identified, they were sparse and small and demonstrated only faint thyrotropin reactivity. No morphologic differences were noted between the pituitary glands of patients with Graves' disease or Plummer's disease or between sexes. Loss of thyrotropin immunoreactivity was found to be reversible in that thyrotropic cells in the pituitary glands of 16 additional concurrently studied patients, who had thyrotoxicosis but were treated and subsequently had normal thyroid function or hypothyroidism, appeared normal or even hyperplastic. Other types of adenohypophysial cells in both the thyrotoxic and the successfully treated groups exhibited no abnormalities. Pituitary adenomas were incidental findings in 6 of the 33 patients (18%). Their immunotypic spectrum included three prolactin-immunoreactive tumors, two growth hormone-containing adenomas (one of which was plurihormonal), and one tumor witbfolllcle-stimulatlng hormone and luteinizing hormone; no thyrotropin-containing adenomas were noted. No examples of pituitary hyperplasia were encountered in pituitary glands of thyrotoxic patients, and no hypophysitis or fibrosis was noted. We conclude that thyrotoxicosis exerts a profound effect on pituitary morphologic features that is identical in both Graves' disease and Plummer's disease. The reversible loss of thyrotropin immunoreactivity and involution of thyrotrophs are interpreted as exaggerated feedback effects of circulating thyroid hormones. Whether the suppressive effect is direct or mediated through the hypothalamus cannot be determined on the basis of this study.

ism,' pregnancy;' excess of estrogen.i-' excess of growth hormone releasing hormone," hypercortisolism,? dopamine agonist medication.s? and Kallmann' s syndrome.'? In patients with hyperthyroidism, the thyroid gland secretes thyroid hormones in excess; consequently, secretion of thyroidstimulating hormone (thyrotropin) is suppressed. As a result of inhibition, affected pituitary glands are expected to show suppressive effects due to negative feedback by thyroid. hormones. These morphologic changes in patients with hyperthyroidism have not been fully explored with use of specific and reliable immunocytochemical techniques, Murray and Ezrin!' investigated five pituitary glands obtained at

The morphologic features of adenohypophysial cells reflect endocrine activity in that they undergo considerable structural alterations as a result of stimulation and suppression; examples include Addison's disease, 1 primary hypothyroid*Emeritus staff. This work was supported in part by Grant MT 6349 from the Medical Research Council of Canada and by the generous contributions of Mr. and Mrs. Francis A. Wittem, Sr., and Mr. and Mrs. Francis A. Wittem, Jr. Individual reprints of this article are not available. Mayo Clin Proc 67:22-26, 1992

22

Mayo Clin Proc, January 1992, Vol 67

autopsy from patients with various types of hyperthyroidism. Applying routine histochemical stains, including the aldehyde thionin-periodic acid-Schiff method, those investigators found substantial suppression of what they considered thyrotrophs. At that time, however, no immunocytochemical stains were available to provide conclusive evidence about the identity, morphologic characteristics, and hormone content of thyrotropin-producing cells. Herein we describe the histologic and immunocytochemical findings in the adenohypophyses of patients with Graves' disease and Plummer's disease (toxic multinodular goiter). Our findings in these disorders are compared with those in a control population and in patients with hyperthyroidism who subsequently had normal thyroid function or hypothyroidism because of treatment.

PITUITARY GLAND IN HYPERTHYROIDISM

23

dilution 1:200). Staining specificity was verified by substituting primary antisera with phosphate-buffered saline and either normal rabbit or normal mouse sera, as well as by absorption of antisera with their respective antigens. The antigen-antibody binding sites were visualized with use of chromogen 3'-diaminobenzidine.

RESULTS Study Group (Patients With Thyrotoxicosis).-Of the 33 patients who died in thyrotoxicosis, 10 had received no treatment, and 23 had undergone either medical or surgical therapy (or both) but had died in thyrotoxicosis. In the overall study group, 18 patients (55%) had had Graves' disease, including 11 women and 7 men who were 18 to 65 years of age (mean, 51 years), and 15 patients (45%) had had toxic multinodular goiter (Plummer's disease), including 13 MATERIAL AND METHODS women and 2 men who were 46 to 78 years of age (mean, 59 The files of the Mayo Clinic Tissue Registry were searched years). for patients who had had thyrotoxicosis, either untreated or Clinical Findings.-Because of the retrospective nature treated, and had died between 1925 and 1970. In 49 cases, of the series, the accession of cases between 1925 and 1970, pituitary glands recovered at autopsy were available and and the fact that autopsy had been done before 1958 in all but were suitable for immunocytologic study. Of these pituitary one patient, the diagnoses of hyperthyroidism were based glands, 33 were from patients who had died in thyrotoxico- not on currently used, sophisticated serum measurements but sis, and 16 were from patients who had been successfully on clinical features and determinations of basal metabolic treated. All but four patients had died before 1945. rate. In 60% of cases, the basal metabolic rate was more than The control group consisted of a published series of 1,000 +40%; the highest rate encountered was +108%. Only in nonselected patients of various ages." No symptoms of later years were protein-bound iodine, thyroxine, thyroadenohypophysial hypersecretion were apparent, either be- tropin, and other measurements available for biochemical fore death or in the medical records. A few patients had confirmation of the diagnosis. received treatment for diabetes mellitus or hypothyroidism. In 21 of the 23 patients who had received treatment, The autopsy-derived pituitary glands were fixed in for- iodides had been administered orally. In addition, subtotal malin, cut in either the horizontal or the sagittal plane, em- thyroidectomy had been done in 10 patients and total thybedded in paraffin, and sectioned at 5 urn. Histochemical roidectomy in 6. Only two patients had been treated with stains routinely used were hematoxylin-eosin, periodic acid- radioactive iodine. "Thyroid storm," a sudden transition to extreme hyperSchiff, and Gordon-Sweet silver stain for demonstrating reticulin fibers. thyroidism manifested by hyperpyrexia, agitation, and For immunocytochemistry, the avidin-biotin-peroxidase tachyarrhythmia that resulted in cardiac failure and hypocomplex technique of Hsu and associates" was applied. tension, was the factor underlying death in 13 of the 33 Sections were incubated overnight at 4°C with antisera di- patients (39%), including 9 patients with Graves' disease rected toward the following hormones or hormone frag- and 4 with Plummer's disease. Only two of these patients ments: growth hormone (DAKO Corporation, Santa Bar- were untreated. Of the treated patients, 78% died within 1 to bara, California; dilution 1:2,000); prolactin (donated by Dr. 2 days after therapy; the longest interval between treatment H. G. Friesen, Winnipeg, Manitoba, Canada; dilution and the onset of thyroid storm was 13 days. The causes of 1:2,000); adrenocorticotropic hormone (1-39 amino acids) death in the study group of 33 patients were cardiovascular (donated by Dr. Salvatore Raiti, National Institute of Diabe- disease (52%)-including cerebrovascular accident (in 9 tes and Digestive and Kidney Diseases, Bethesda, Maryland; patients), myocardial infarction (in 6), and congestive heart dilution 1:1,000); thyrotropin (~-subunit) (Chemicon Inter- failure (in 2)-infection (21%), cancer (15%), and miscellanational Inc., Temecula, California; dilution I :2,000); fol- neous factors (12%). licle-stimulating hormone (~-subunit) and luteinizing Pathologic Findings.-All pituitary glands, including hormone (~-subunit) (both donated by Dr. Salvatore Raiti; those of patients with hyperthyroidism, patients who had dilution 1:2,000); and human chorionic gonadotropin (a- received treatment, and control patients, were normal in size subunit) (BioGenex Laboratories, San Ramon, California; as well as in the proportion of anterior and posterior lobes.

24

PITUITARY GLAND IN HYPERTHYROIDISM

Mayo Clin Proc, January 1992, Vol 67

Sections stained with hematoxylin-eosin showed no discern- . (45%). Of the remainder, two contained prolactin and ible shift in acidophilic, basophilic, and chromophobic cells. growth hormone, four contained adrenocorticotropic Alterations in cell number were noted only on im- hormone, and one was immunoreactive to luteinizing munostains, and only thyrotropic cells were affected. Spe- hormone. In only four cases (4%) was the pituitary gland cifically, no abnormality of prolactin cells was detected. No sufficiently enlarged to suggest the presence of an adenoma. morphologic differences were noted between the glands of None had produced endocrine symptoms. patients with Graves' disease and those with Plummer's disease. No inflammation, infarcts, fibrosis, or hyperplasia of cells (other than thyrotropic cells) was observed in the DISCUSSION study group. This study indicates that substantial morphologic changes 1. Thyrotropic Cells.-Among the 10 patients with un- affect the pituitary thyrotrophs of patients with excessive treated hyperthyroidism (5 with Graves' disease and 5 with secretion of thyroid hormones. In clinically obvious hyperPlummer's disease), the pituitary glands showed identical thyroidism, regression of thyrotrophs is pronounced; they changes. Thyrotropic cells were small, lacked stainable are small and possess only a minute rim of cytoplasm. Their processes, and in some cases were entirely inapparent (Fig. 1 immunoreactivity for thyrotropin is substantially decreased A). Faint paranuclear specks of immunoreactivity were or almost absent. Indeed, immunopositive thyrotrophs may noted in only occasional cells. No correlation was evident not be evident in severely affected pituitary glands. Because between the duration of symptoms, which (recorded in six these morphologic changes are identical in patients with cases) ranged from 16 months to 2 years, and the degree of Graves' disease and Plummer's disease (toxic multinodular goiter), regression of thyrotrophs obviously depends on the loss of immunoreactivity to thyrotropin. Unlike the pituitary glands of the untreated patients, stain- degree and perhaps the duration of hyperthyroidism rather ing of thyrotropic cells was readily identified in the pituitary than on the cause of excessive thyroid hormone. Morphologic changes in the pituitary glands of patients glands of the 23 treated patients. A sequence of reappearance of immunoreaction was noted. Reactivity was first with hyperthyroidism are selective. Types of adenohypophevident in the paranuclear cytoplasm (Fig. 1 B). In conjunc- ysial cells that produce hormones other than thyrotropin are tion with the continued increase in reactivity, enlargement unaffected; their distribution, immunoreactivity, and morand staining of the cell body (Fig. 1 C) and of processes (Fig. phologic appearance are within the normal range. In patients 1 D) were observed. The earliest return of staining was with successfully treated hyperthyroidism-that is, those in apparent at 6days. whom clinical manifestations disappear and euthyroidism or 2. Adenomas.-Pituitary adenomas were encountered hypothyroidism is achieved-not only do thyrotrophs asas incidental findings in 6 of the 33 patients (18%) in the sume their original size and number but also immunoreactivstudy group. These tumors included three prolactin cell ity to thyrotropin returns to normal. Hyperplasia of thyroadenomas (two microadenomas [defined as a tumor smaller trophs may be noted in cases in which therapy has induced than 1 em in diameter] and one macroadenoma), one growth hypothyroidism. This sequence provides conclusive evihormone cell macroadenoma, one plurihormonal microade- dence that hyperthyroidism-based morphologic changes in noma (with growth hormone, luteinizing hormone, and 0.- pituitary thyrotrophs are completely reversible. Our findings confirm and extend experimental studies subunit follicle-stimulating hormone), and one gonadotropic microadenoma (with luteinizing hormone and follicle-stimu- that have shown appreciable regression of thyrotrophs in the lating hormone). No thyrotropin-containing tumors were pituitary glands of dogs and rats after administration of thyroid hormone.v'-" To our knowledge, morphologic found. Successfully Treated Patients.-In three patients in changes in the pituitary glands of human subjects with whom clinical euthyroidism or hypothyroidism had been hyperthyroidism have been studied only by Murray and achieved, mild diffuse hyperplasia of thyrotropic cells was Ezrin." Those investigators used routine histochemical apparent. Staining of thyrotropic cells was normal in the techniques-specifically, the aldehyde thionin-periodic other 13 patients who had been successfully treated. acid-Schiff method-and achieved results similar to ours, Control Group.-Among the 1,000 unselected control finding massive regression of ~2 cells (thyrotrophs) in papituitary glands, immunoreactivity to thyrotropin was readi- tients with hyperthyroidism. Our results with use of specific ly identified in the adenohypophysis in all cases." The and sensitive immunocytochemical techniques confirm their location and configuration of reactive cells were typical of findings. Indeed, the changes that affect the thyrotrophs are those of thyrotrophs. In the control group, 107 adenomas so extensive that the existence of severe hyperthyroidism can (10.7%) were identified. These tumors were primarily im- be conclusively diagnosed on the basis of pituitary morphomunonegative lesions (50%) or prolactin-reactive adenomas logic findings alone.

Mayo Clin Proc, January 1992, Vol 67

PITUITARY GLAND IN HYPERTHYROIDISM

25

Fig. 1. Sequence of thyrotropin-immunostained sections of pituitary gland, demonstrating lack of reactivity in an untreated case of hyperthyroidism (A) and, in cases of treated hyperthyroidism, reappearance of reactivity illustrated at 6 days (B), 28days (C), and5 months (D) aftertreatment. (Original magnification; x250.)

Despite the demonstration of the regression of thyrotrophs, the mechanism of action has not been clarified by our studies. Specifically, the question of whether the negative feedback effect of thyroid hormone is exerted primarily on the hypothalamus by blocking release of thyrotropin releasing hormone or on the pituitary gland by suppressing secretion of thyrotropin has yet to be resolved: Although it would be reasonable to anticipate that the response of the various types of adenohypophysial cells to functional suppression would be uniform, this expectation has not been confirmed by previous studies. In fact, conclusive evidence shows that the morphologic manifestations of decreased endocrine activity differ substantially among the various types of adenohypophysial cells.' Human corticotrophs respond to glucocorticoid excess-induced functional suppression by cytoplasmic accumulation of keratinimmunoreactive intermediate filaments, the ultrastructural equivalent of Crooke's hyaline degeneration.' Accumula-

tion of lysosomes is the principal morphologic feature of functionally suppressed lactotrophs of the rat pituitary gland." In contrast, human lactotrophs behave differently; after treatment with dopaminergic agonists, they show reduction in size, diminution in cytoplasmic volume density of rough endoplasmic reticulum and Golgi complexes, and decrease in content of immunoreactive prolactin as well as of messenger RNA.8 These changes are readily induced by treatment with dopamine agonists, substances that are known to inhibit the endocrine activity (synthesis and release of prolactin) of lactotrophs. In somatotrophs suppressed by exposure to somatostatin, one notes increase in size and volume density of cytoplasmic secretory granules as well as accumulation of lysosomes." Gonadotrophs in patients with Kallmann's syndrome-that is, those with lack of stimulation by hypothalamic gonadotropin releasing hormoneshow considerable involution and also substantial reduction or almost complete loss of immunoreactivity to follicle-

26

Mayo Clin Proc, January 1992, Vol 67

PITUITARY GLAND IN HYPERTHYROIDISM

stimulating hormone and luteinizing hormone." Similar to . these gonadotrophs, functionally suppressed thyrotrophs in patients with hyperthyroidism show regression in conjunction with decrease in cell size and in immunoreactivity to thyrotropin. In summary, although striking differences have been noted in morphologic responses to suppression among the various types of adenohypophysial cells, the underlying mechanisms are not understood. The various adenohypophysial cells are thus distinctly different and have individual characteristics.

CONCLUSION On the basis of the current study, we conclude that (1) the effects of hyperthyroidism on the pituitary gland are selective, affecting only thyrotropic cells; (2) hyperthyroidism produces reversible involution of thyrotropic cells, the extent of which reflects the severity of hyperthyroidism; and (3) pituitary changes are identical in Graves' disease and Plummer's disease.

6.

7. 8.

9.

10.

11.

ACKNOWLEDGMENT We thank Nancy Ryan for expert assistance in the performance of immunocytologic studies.

12. 13.

REFERENCES 1. Scheithauer BW, Kovacs K, Randall RV: The pituitary gland in untreated Addison's disease: a histologic and immunocytologic study of 18 adenohypophyses. Arch Pathol Lab Med 107:484-487,1983 2. Scheithauer BW, Kovacs K, Randall RV, Ryan N: Pituitary gland in hypothyroidism: histologic and immunocytologic study. Arch Pathol Lab Med 109:499-504, 1985 3. ScheithauerBW,SanoT, KovacsKT, YoungWF Jr, RyanN, Randall RV: The pituitary gland in pregnancy: a clinicopathologic and immunohistochemical study of 69 cases. Mayo Clin Proc 65:461-474,1990 4. Scheithauer BW, Kovacs KT, Randall RV, Ryan N: Effects of estrogen on the human pituitary: a clinicopathologic study. Mayo Clin Proc 64:1077-1084,1989 5. Thorner MO, Perryman RL, Cronin MJ, Rogol AD, Draznin M, Johanson A, Vale W, Horvath E, Kovacs K: Somatotroph

14.

15. 16. 17.

hyperplasia: successful treatment of acromegaly by removal of a pancreatic islet tumor secreting a growth hormone-releasing factor. J Clin Invest 70:965-977,1982 . Sano T, Asa SL, Kovacs K: Growth hormone-releasing hormone-producing tumors: clinical, biochemical, and morphological manifestations. Endocr Rev 9:357-373, 1988 Horvath E, Kovacs K: Fine structural cytology of the adenohypophysis in rat and man. J Electron Microsc Tech 8:401-432, 1988 Kovacs K, Stefaneanu L, Horvath E, Lloyd RV, Lancranjan I, Buchfelder M, Fahlbusch R: Effect of dopamine agonist medication on prolactin producing pituitary adenomas: a morphological study including immunocytochemistry, electron microscopy and in situ hybridization. Virchows Arch [A] 418:439-446,1991 Tindall GT, Kovacs K, Horvath E, Thorner MO: Human prolactin-producing adenomas and bromocriptine: a histological, immunocytochemical, ultrastructural, and morphometric study. J Clin Endocrinol Metab 55:1178-1183, 1982 Kovacs K, Sheehan HL: Pituitary changes in Kallmann's syndrome: a histologic, immuriocytologic, ultrastructural, and immunoelectron microscopic study. Fertil Steril 37:8389, 1982 Murray S, Ezrin C: Effect of Graves' disease on the "thyrotroph" of the adenohypophysis. J Clin Endocrinol Metab 26:287-293, 1966 Mct.omb OJ, Ryan N, Horvath E, Kovacs K: Subclinical adenomas of the human pituitary: new light on old problems. Arch Pathol Lab Med 107:488-491,1983 Hsu S-M, Raine L, Fanger H: A comparative study of the peroxidase-antiperoxidase method and an avidin-biotin complex method for studying polypeptide hormones with radioimmunoassay antibodies. Am J Clin Pathol 75:734-738, 1981 Panciera DL, Atkins CE, Bosu WTK, MacEwen EG: Quantitative morphologic study of the pituitary and thyroid glands of dogs administered L-thyroxine. Am J Vet Res 51:27-31, 1990 D'Angelo SA: Role of the hypothalamus in pituitary-thyroid interplay. J Endocrinol 17:286-299,1958 Farquhar MG, Rinehart JF: Cytologic alterations in the anterior pituitary gland following thyroidectomy: an electron microscope study. Endocrinology 55:857-876,1954 George SR, Kovacs K, Asa SL, Horvath E, Cross EG, Burrow GN: Effect of SMS 201-995, a long-acting somatostatin analogue, on the secretion and morphology of a pituitary growth hormone cell adenoma. Clin Endocrinol 26:395-405, 1987