Introduction

Thyrotropin (TSH)-secreting adenomas are rare; they account for about 1–2% of all pituitary adenomas [3, 7]. The secretion of TSH by these pituitary tumors results in central hyperthyroidism, the diagnosis of which may be delayed because patients, especially those with long histories of thyroid dysfunction, are often mistakenly treated for Graves’ disease. Consequently, at detection, many TSH-secreting adenomas are large, invasive, and their surgical cure is difficult [2, 3, 7, 15, 16, 17, 18, 19] and—despite treatment by surgery, radiation therapy, medication, or a combination thereof—outcomes tend to be poor [15, 16]. TSH-secreting adenomas are often plurihormonal; approximately 60% concomitantly secrete growth hormone (GH) and/or prolactin (PRL) [21, 22]. While the ultrasensitive immunometric TSH assay facilitates early detection of TSH-secreting adenomas, their pathology and appropriate medical treatment are not well understood. Here we report the clinical and pathological features of 20 surgically treated patients with TSH-secreting adenomas. We also discuss ultrastructural findings on octreotide-treated TSH-secreting adenoma cells.

Clinical materials and methods

Patients

The study population consisted of 20 patients (16 women and four men ranging in age from 23 to 62 years, mean 41.2 years) with TSH-secreting pituitary adenomas. They represent 1.8% of 1,100 patients who underwent transsphenoidal surgery by the first author (AT) between 1983 and 2003. Clinical data on 16 of these patients have been reported elsewhere [24]. All patients underwent complete physical examination including palpation of the thyroid and visual-field testing by Goldman’s perimetry. Serum TSH levels were measured using a highly specific immunoradiometric assay (AxSYM TSH Dainapack/Dainabot Co., Ltd. Japan); the detection limit was 0.005 mU/L. Serum α-subunit of glycoprotein, free T4, and free T3 were determined by radioimmunoassay. The α-subunit/TSH molar ratio was calculated using 28,000 (1 µg TSH corresponding to 4.93 mU) and 14,700 as the molecular weight for TSH and α-subunit, respectively [7]. Other basal hormone levels recorded included GH, PRL, adrenocorticotropin (ACTH), luteinizing hormone (LH), follicle-stimulating hormone (FSH), and serum-free cortisol. Eleven patients underwent an endocrinological loading test with bromocriptine. Serum TSH and PRL levels were measured before and 2, 4, 6, and 12 h after the oral administration of 2.5 mg bromocriptine. The octreotide test was performed in 12 patients; serum TSH and GH levels were measured before and 2, 4, 6, and 12 h after the intravenous injection of 50 µg octreotide. All patients underwent contrast-enhanced CT scanning of the pituitary; 14 patients were also subjected to magnetic resonance imaging (MRI) study at 1.5 Tesla. Coronal and sagittal T1-weighted images were obtained before and after gadolinium injection. Patients 16 and 19 were preoperatively treated with octreotide. They received three subcutaneous injections of 100 µg octreotide daily for the 7 days preceding surgery. Serum TSH, free T4, GH, PRL, and IGF-1 levels were measured before and after octreotide treatment, and MRI study was carried out before and after treatment.

Histopathological studies

Surgical specimens were fixed immediately with 10% buffered formalin or 4% paraformaldehyde and embedded in paraffin for light microscopy, or fixed with 2.5% glutaraldehyde and 2% osmium tetroxide for electron microscopy. For immunohistochemical characterization of the cellular hormone content, the indirect immunoperoxidase method was applied to formalin-fixed paraffin-embedded tissue sections as described previously [21]. The following antisera were used: anti-TSHβ (1:3200 dilution), anti-α-subunit of glycoprotein (1:200), FSHβ (1:200), LHβ (1:200, NIADDK, Maryland, USA), GH (1:800), PRL (1:600), ACTH (1:800, DAKO, CA, USA). The antipituitary-specific transcription factor (Pit-1, 1:200, Immunotech, USA) was another antibody we used with the avidin-biotin complex method and heat retrieval. The immunohistochemical staining of hormone was divided into four grades, according to the number of immunopositive cells, as follows: −, negative; 1+, less than 20%; 2+ 20–50%; and 3+, more than 50% of all tumor cells. As a positive control for the staining of hormones and Pit-1, normal human pituitary glands obtained at autopsy were used. Ultrastructural analysis was performed on 11 TSH-secreting adenomas including two that had been treated with octreotide for 7 days. Ultrathin slices were double-stained with lead citrate and uranyl acetate and inspected under an electron microscope (JEM-100S; JEOL DATUM Ltd., Tokyo, Japan).

Results

Clinical features

As shown in Table 1, all patients manifested elevated serum TSH levels despite elevated thyroid hormones. Serum TSH ranged from 0.46 to 55 mU/L and the α-subunit level was elevated (0.2–9.8 ng/ml). The molar ratio of α-subunit/TSH was greater than 1.0 in all but one patient. Of the 20 patients, four underwent subtotal thyroidectomy and/or radioiodine thyroid ablation earlier, 11 (55.0%) had been treated with antithyroid medication, and 16 had a goiter. Signs of other hormone secretion were present in five patients: patients 2, 3, and 5 manifested acromegalic features and patients 15 and 18 had amenorrhea-galactorrhea. Patients 11, 15, and 18 had elevated PRL levels. The basal serum levels of ACTH, cortisol, FSH, and LH were within normal limits in all patients.

Table 1 Clinical data of 20 patients with thyrotropin (TSH)-secreting pituitary adenoma
Full size table

The administration of octreotide significantly suppressed TSH (<50%) in ten of 14 patients (71.4%) (Fig. 1). Results of the bromocriptine test showed significant suppression (below 50% of the baseline) in two of ten patients; in the other eight, suppression was partial (Fig. 2). The results of other endocrine tests, including the PRL response to TRH (200 µg, i.v.), the ACTH and cortisol response to CRF (100 µg, i.v.), and the FSH and LH response to LHRH (100 µg, i.v.) were normal in all patients (Table 2).

Fig. 1
figure 1

TSH response to 50 µg of octreotide intravenous injection in patients with TSH-secreting adenoma positive for GH (line); immunonegative for GH (dotted line)

Full size image
Fig. 2
figure 2

TSH response to 2.5 mg bromocriptine administration in patients with TSH-secreting adenoma. Immunopositive for PRL (line); immunonegative for PRL (dotted line)

Full size image
Table 2 Results of octreotide and bromocriptine administration test in thyrotropin (TSH) secreting adenomas
Full size table

Radiological studies including CT and MRI scans disclosed that all 20 patients had pituitary mass lesions (two micro- and 18 macroadenomas); cavernous sinus invasion was present in nine patients.

All 20 patients underwent transsphenoidal surgery. On intraoperative inspection, the adenomas were well demarcated, and the anterior pituitary gland was normal; 16 of the 20 tumors were fibrous and firm and thus different from other types of adenomas. Both microadenomas and eight intrasellar-type tumors were totally resected. More than 80% of the tumor mass of seven macroadenomas was resected, and in three patients with larger and invasive tumors, less than 80% of the mass was resected.

Pathological findings

Histopathological study of resected specimens confirmed that all 20 patients had pituitary adenomas. Hematoxylin and eosin staining of tumor tissues disclosed a diffuse sinusoidal pattern; the cells had uniformly oval nuclei, and the cytoplasm was mixed basophilic and acidophilic. The results of immunohistochemical study are shown in Table 3. Tissues from all 20 adenoma specimens were positive for TSH-β and α-subunit. The degree of immunopositivity and the percentage of positive cells varied from case to case. As shown in Fig. 3, 16 adenomas were positive for GH and/or PRL; 15 were GH-positive, and 11 were PRL-positive. None of the 20 adenoma tissues were positive for LH, FSH, or ACTH. All adenomas expressed Pit-1 protein in many adenoma cell nuclei. The MIB-1 labeling index ranged from 0.6% to 1.2%.

Table 3 Results of immunohistochemical study of 20 thyrotropin (TSH)-secreting adenomas
Full size table
Fig. 3
figure 3

A Photomicrograph of the hematoxylin-eosin stain of specimens from patient 9. The tissue demonstrated chromophobe pituitary adenoma. H&E, original magnification ×200. B Immunohistochemistry for TSHβ. TSHβ immunoreactivity is found in cytoplasm of adenoma cells. Avidin-biotin complex method, original magnification ×400. C Immunohistochemistry for α-subunit, D GH, and E PRL, respectively. Positive immunostaining is observed in cytoplasm of adenoma cells. F Immunohistochemistry for pituitary specific transcription factor Pit-1 showing positive staining in nuclei of adenoma cells. Original magnification ×400

Full size image

Electron micrographs of the adenoma were obtained in nine patients, including two who had received preoperative octreotide treatment. The tumor cells were elongated and manifested long cytoplasmic processes and round or ovoid nuclei with prominent nucleoli. There was moderate development of rough endoplasmic reticulum and Golgi complexes. Small secretory granules measuring between 50 and 200 nm in diameter were arranged peripherally under the plasmalemma (Fig. 4A). In patients with preoperative octreotide treatment, the cytoplasm was shrunken and contained diffusely distributed secretory granules (Fig. 4B). In some adenoma tissues, there was remarkable fibrosis between adenoma cells (Fig. 4C).

Fig. 4
figure 4

A Electron micrographs showing elongated cytoplasmic processes and ovoid nucleus with prominent nucleoli. The small secretory granules that measured 50–100 nm arranged peripherally under the plasmalemma. Bar=1 µm. B Electron microgram of adenoma tissue of patient 16 treated with octreotide (300 mg/ day) for 7 days preoperatively. Cytoplasm was shrunken and secretory granules are observed diffusely in cytoplasm. C Other specimen showing massive fibrosis between adenoma cells. (*) Bar=1 µm

Full size image

Discussion

Somatostatin analogs are an important improvement in the medical treatment of TSH-secreting adenomas. Somatostatin inhibits GH secretion by somatotrophs through seven transmembrane G-coupled SRIF receptor subtypes (SSTR 1–5). Normal and neoplastic human pituitary somatotrophs primarily express SSTR2 and SSTR5, both of which have been implicated in mediating the suppression of GH secretion induced by somatostatin analogs [5]. The efficacy of this treatment in the present series is similar to that reported by others [8, 9, 10].

Somatostatin analog treatment is potentially useful for preparing patients for surgery. We found that preoperative octreotide administration resulted in the good control of serum TSH levels and hyperthyroidism and brought about tumor shrinkage. Preoperative octreotide treatment of acromegalic patients is reportedly efficacious because of its ability to reduce the size of the pituitary tumor and the degree of suprasellar extension [11, 26]. In patients with TSH-secreting adenomas, preoperative octreotide administration controls thyroid function and reduces the mass effect of the adenomas [9, 28].

Unlike bromocriptine in PRL-secreting adenomas, octreotide treatment in GH-secreting adenomas does not induce tumor fibrosis. While shrinkage was observed in more than 50% of GH-secreting adenomas following primary treatment with somatostatin analogs [11, 12], this was fewer in TSH-secreting tumors, possibly due to the fibrotic characteristics of these neoplasm [27]. TSH-secreting adenomas tend to be large, invasive, and difficult to cure, and they are more highly fibrotic than other pituitary tumors [24, 25]. Fibroblast growth factor has been implicated in the fibrotic nature of these tumors [13]. These characteristics complicate surgery and may result in unsatisfactory outcomes following surgical and medical treatments. Among a group of 33 patients with TSH-secreting adenomas, 30 (90.9%) responded to short-term treatment with octreotide (100–300 µg daily) by manifesting a decrease in thyrotropin secretion; serum thyroxin levels fell to normal values in 73% percent of these patients [9].

In 10 of 14 of our patients, octreotide administration tests reduced serum TSH levels to below 50% of the baseline. Bertherat et al. attributed this decline to the high concentration of somatostatin receptors on TSH-secreting adenomas and concluded that octreotide effectively decreases the TSH level and controls hyperthyroidism [5]. Following octreotide administration, shrinkage of the adenoma has also been observed, although not as many as in GH adenomas [5, 9], making this agent a candidate for combination therapy in patients where the achievement of remission by surgery alone is difficult.

In our series, dopamine agonists were effective in only one mixed TSH/PRL-secreting adenoma. Similarly, Bevan et al. [6] found that only four of 24 patients with TSH-secreting adenomas who were challenged with dopamine agonists showed a suppression of TSH and thyroid hormone levels. Wood et al. [29] detected dopamine D2 receptors on TSH-secreting adenomas but reported limited treatment success. In our series, a single administration of bromocriptine reduced TSH significantly in only two of 11 patients. We consider patient 7 an exceptional case; bromocriptine was effective in combination with surgery.

Certain pituitary tumors secrete both GH and PRL [21] and the existence of plurihormonal cells that secrete GH and PRL in normal and tumorous pituitaries is known. These dual cells, known as mammosomatotropes, account for 25–50% of all cells in the human anterior pituitary. In situ hybridization studies demonstrated both GH and PRL mRNA in those adenomas [22], and pituitary cell lines have been derived that secret both GH and PRL. The ratio of GH and PRL secreted by GH3 cell is affected by factors such as estradiol, TRH, and cortisol, and physiological and experimental evidence indicates that GH and PRL secretors can undergo functional interconversion in a regulated manner [14]. Ontogenically, GH- and TSH-secreting cells share a common precursor cell expression, and GH-to-TSH cell transdifferentiation has been observed in the hypothyroid rat pituitary [20]. This suggests that TSH-secreting adenomas may derive from early totipotential progenitor cells that may differentiate into GH-, PRL- or TSH cells. This mechanism may play a role in the pathogenesis of the multihormonality of TSH-secreting pituitary adenomas.

The pituitary-specific transcription factor-1 Pit-1 is a member of the POU-domain family that has a role in the development and differentiation of three pituitary cell types: somatotrophs, lactotrophs, and thyrotrophs. Whether Pit-1 is associated with abnormal cell proliferation during the pathogenesis of pituitary adenomas remains controversial [1, 23]. Immunohistochemically, we observed Pit-1 proteins in all TSH-secreting adenomas studied. Their presence suggests that TSH-secreting adenomas are derived from anterior pituitary cell types of Pit-1 lineage.

While many normal and tumorous pituitaries secrete both GH and PRL, the concomitant secretion of TSH is rare. The existence of a distinct genomic region that inhibits the transcription of certain genes may be suggested in efforts to explain the restriction of PRL and GH gene expression in thyrotrophs. Additional regulatory factor(s) that explain the loss of TSH secretion by GH- and PRL cells are expected to be investigated. The switch that creates TSH-secreting adenomas from common multipotential progenitor cells remains to be clarified, and its rarity may explain why TSH-secreting adenomas are also rare.