The endogenous fibroblast growth factor-2 antisense gene product regulates pituitary cell growth and hormone production.

Basic fibroblast growth factor (bFGF; FGF-2) is one of 19 related members of a growth factor family with mitogenic and hormone-regulatory functions. In Xenopus laevis oocytes, a 1.5-kb FGF-2 antisense (GFG) RNA complementary to the third exon and 3'-untranslated region (UTR) of FGF-2 mRNA has been implicated in FGF-2 mRNA editing and stability. The human homolog has been cloned, and we localized this gene by yeast artificial chromosome (YAC), somatic cell, and radiation hybrid panels to the same chromosomal site as FGF-2 (chromosome 4, JO4513 adjacent to D4S430), confirming this as a human endogenous antisense gene. The full-length GFG antisense RNA encodes a 35-kDa protein, which is highly homologous with the MutT family of antimutator nucleosidetriphosphatases (NTPases). We show that human pituitary tumors express FGF-2 and its endogenous antisense partner GFG. While normal pituitary expresses GFG but not FGF-2, pituitary adenomas express FGF-2 and have reduced levels of GFG; aggressive and recurrent adenomas expressed more FGF than GFG mRNA. To examine the effects of this antisense gene in the pituitary, we transfected the pituitary-derived GH4 mammosomatotroph cell line with constructs encoding the full-length human GFG cDNA. Transiently and stably transfected cells expressed the 35-kDa GFG protein that was localized to the cytoplasm. These cells exhibited enhanced PRL expression as documented by transiently transfected PRL-luciferase reporter assay and by endogenous PRL protein. GFG expression in these cells did not alter endogenous FGF-2 expression but increased the proportion of the higher molecular mass 22-kDa form of GH. Moreover, GFG expression inhibited cell proliferation as shown by [(3)H]thymidine incorporation, proliferating cell nuclear antigen (PCNA) nuclear staining, and cell cycle analysis. We conclude that the GFG-encoded protein has divergent hormone-regulatory and antiproliferative actions in the pituitary that are independent of FGF-2 expression. GFG represents a novel mechanism involved in restraining pituitary tumor cell growth while promoting hormonal activity.

Hyalinizing trabecular tumor of the thyroid: a variant of papillary carcinoma proved by molecular genetics.

Hyalinizing trabecular tumors of the thyroid are interesting but uncommon neoplasms. They have been classified as benign hyalinizing trabecular adenomas or malignant hyalinizing trabecular carcinomas. They share both epidemiologic and morphologic features with papillary carcinoma, and there has been much speculation about the relationship between these two entities. Because RET/PTC gene rearrangements are specific to papillary thyroid carcinoma, the authors examined the presence of RET/PTC-1, -2, and -3 in eight hyalinizing trabecular tumors using reverse transcription-polymerase chain reaction with Southern hybridization and immunohistochemistry. They detected the presence of a RET/PTC gene rearrangement in six of the eight hyalinizing trabecular tumors. This confirms the long-standing suspicion that hyalinizing trabecular tumors do indeed represent a morphologic variant of papillary carcinoma.

Molecular basis off hurthle cell papillary thyroid carcinoma.

Among thyroid neoplasms, Hurthle cell tumors (HCTs) have traditionally been a distinct diagnostic category. Hurthle cell adenomas are encapsulated follicular lesions with benign behavior. Hurthle cell carcinomas exhibit unequivocal capsular and/or vascular invasion; they are aggressive tumors with a poor prognosis. Recently, Hurthle cell papillary thyroid carcinomas (PTCs) have been identified on morphological grounds. We hypothesize that a subset of HCTs represent PTC with clinical, histological, and immunohistochemical features based on specific molecular events. ret/PTC gene rearrangements give rise to novel oncogenes that are unique to PTC. We studied a group (n = 50) of HCTs for ret/PTC gene rearrangements. Tumors were examined for papillary differentiation by light microscopic evaluation of nuclear features, by RT-PCR for ret/PTC gene rearrangements, and by immunohistochemistry for ret. Among 24 noninvasive tumors, 13 contained ribonucleic acid for ret/PTC-1, -2, or -3, and 9 of these were immunoreactive for ret. Among 19 Hurthle cell carcinomas, 15 had focal nuclear hypochromasia with grooves and/or inclusions; expressed transcripts of ret/PTC-1, -2, or-3; and exhibited ret positivity. Tumors with ret/PTC gene rearrangements tended to have lymph node metastases rather than hematogenous spread. Our results indicate that a subset of HCTs exhibit features of PTC that are attributable to specific gene rearrangements, resulting in expression of ret/PTC oncogenes. These data support subclassification of HCTs into three groups: Hurthle cell adenomas, Hurthle cell carcinomas, and Hurthle cell PTC.

Structure-function correlations of growth hormone or/and prolactin-producing pituitary adenomas: an in vitro study with the reverse hemolytic plaque assay.

The purpose of this study was to detect in vitro growth hormone (GH) and prolactin (PRL) secretion from adenomas clinically associated with GH or PRL hypersecretion. The reverse hemolytic plaque assay (RHPA) was applied in order to reveal possible differences among various morphologic adenoma types, and to examine the inhibitory effects of octreotide on GH release as well. The 20 surgically resected pituitary adenomas studied included 15 from acromegalic patients and 5 from patients with hyperprolactinemia. All adenomas were diagnosed by histology, immunocytochemistry and electron microscopy. Among tumors associated with acromegaly, 5 were densely granulated (DG), 5 were sparsely granulated (SG) somatotroph (SM) adenomas, 2 were mammosomatotroph (MSM) and 3 mixed somatotroph-lactotroph cell (mixed SM-LT) adenomas; tumors causing hyperprolactinemia included 4 lactotroph (LT) adenomas and 1 mixed SM-LT adenoma. GH release assessed by the RHPA corresponded to in vivo hormone secretion and to tissue immunoreactivity. Statistical analysis showed significant differences among all morphologic types of SM adenomas, exclusive of SG-SM adenomas compared to mixed SM-LT adenomas. The mean plaque size in DG-SM and MSM adenomas was significantly greater than that of SG-SM and mixed SM-LT adenomas, indicating higher GH secretion by the former two types during the same incubation time. PRL secretion was documented in 2 mixed SM-LT adenomas. Plaques for PRL, but not for GH were formed in all LT adenomas. In all SM and LT adenomas, cells producing large plaques represented a minority of the plaque-forming cell population, however, they accounted for the largest part of the total plaque area, thus the largest part of hormone secretion. Octreotide effects on GH release were studied in 6 adenomas by the RHPA. Octreotide treatment induced a rapid and significant reduction in GH secretion by SM cells in vitro, with a selective effect on high-secreting cells.

Cell-specific expression of estrogen receptor in the human pituitary and its adenomas.

Estrogen affects the synthesis and release of several pituitary hormones. The estrogen receptor (ER), a member of the steroid hormone receptor family, is thought to mediate transcriptional effects in a cell-specific fashion. We investigated whether ER is expressed in specific hormone-producing cell types in the human pituitary and its adenomas. Pituitary adenomas (n = 34) were collected at the time of surgery, and normal glands were obtained from autopsy. Expression of ER messenger ribonucleic acid (mRNA) was determined by reverse transcription-polymerase chain reaction (RT-PCR) and in situ hybridization. ER was also localized with immunohistochemistry and protein extraction. By RT-PCR, ER mRNA was found in the nontumorous pituitary and in pituitary adenomas expressing only PRL, in those producing GH and PRL, and in adenomas expressing the gonadotropic hormones. No ER mRNA was detected in adenomas expressing only GH without PRL or gonadotropins, nor in tumors producing ACTH without PRL or gonadotropins. In situ hybridization was not as sensitive or specific as RT-PCR. Biochemical analysis performed on seven tumors that were positive for ER mRNA by RT-PCR detected ER protein in only one PRL adenoma and one oncocytoma and yielded negative or equivocal results in one PRL adenoma, three GH-PRL adenomas, and one null cell adenoma. ER protein was localized by immunohistochemistry in scattered cells of the nontumorous adenohypophysis and in a few PRL and gonadotroph adenomas. We conclude that ER expression, as determined by RT-PCR, correlates with the expression of PRL or gonadotropins; in contrast, ER mRNA was not detected in adenomas that express only GH or ACTH. These findings implicate ER as a cell-specific transcription factor that may regulate cytodifferentiation in the pituitary.

Heterogenous in vivo and in vitro expression of basic fibroblast growth factor by human pituitary adenomas.

Basic fibroblast growth factor (bFGF) is a potent mitogenic and angiogenic factor that is known to regulate GH, PRL, and TSH secretion. Sequences within a bFGF gene family member have been detected in transforming DNA samples derived from human PRL-secreting tumors. Furthermore, elevated serum concentrations of bFGF have been noted in patients with multiple endocrine neoplasia-1. To further examine the significance of bFGF in sporadic human pituitary adenomas, we investigated the expression of bFGF by these tumors. Using an enzyme-linked immunoassay that recognizes all 16-24 kilodalton molecular mass forms of bFGF, we measured circulating serum concentrations in 21 patients with sporadic pituitary adenomas; they ranged from less than 0.5-84 pg/mL and declined following surgical adenomectomy. To confirm the pituitary source of this growth factor, we determined in vitro bFGF release from 43 adenomas (10 GH, 7 PRL, 10 ACTH, 14 gonadotrope adenomas/oncocytomas, and 2 silent subtype 3 adenomas). bFGF was present with wide variability (0.75-2100 pg/24 h.10(5) cells) in conditioned culture media of all adenomas examined. The adenohypophysial source of this growth factor was further demonstrated by the reverse hemolytic plaque assay. Variable bFGF messenger RNA expression was identified by the reverse-transcription polymerase chain reaction technique in 9 functional (2 PRL, 5 GH, 2 ACTH) and 7 nonfunctional (1 oncocytoma, 2 null cell, 2 gonadotrope, 2 Silent Subtype 3) adenomas examined. bFGF levels were unaltered in vitro following hypothalamic hormone stimulation/inhibition. The lack of a bFGF signal peptide sequence and hypothalamic hormone-independence suggest that secretion of this factor may be independent of pituitary hormone regulation. Immunocytochemistry failed to localize bFGF in tumors that released this factor in vitro, suggesting that storage of this peptide does not correlate with its synthesis and release. In conclusion, the heterogenous expression of bFGF suggests that it may play a specific and selective role in the tumorigenic process of some pituitary adenomas.

Membrane-anchored expression of transforming growth factor-alpha in human pituitary adenoma cells.

Growth factors induce cell proliferation and are implicated in the multistep process of tumorigenesis. Transforming growth factor-alpha (TGF alpha), a peptide that binds to the epidermal growth factor receptor, is expressed by carcinomas and normal tissues. To investigate the possible role of TGF alpha in adenohypophysial tumorigenesis, we studied its expression in nontumorous human pituitary and different clinically and morphologically characterized human pituitary adenomas. Ribonucleic acid was reverse transcribed and amplified by polymerase chain reaction; transcript signals were identified with marked variation in 14 of 15 adenomas, and a weak signal was detected in nontumorous pituitary. Immunohistochemical positivity was found with variable intensity in all adenoma types, but not all tumors. Ultrastructural immunogold localized TGF alpha in endoplasmic reticulum, in Golgi apparatus, and on cell membranes; surface localization was confirmed by immunofluorescence. To assess possible secretion, the reverse hemolytic plaque assay was performed; small plaques were identified using an antibody that recognizes the extracellular domain of pro-TGF alpha; however, the plaques did not increase in size with time, suggesting that they detected membrane-anchored TGF alpha. Moreover, TGF alpha was undetectable by enzyme-linked immunosorbent assay in pituitary tumor-conditioned culture media. The marked variable expression of TGF alpha, the absence of secretion in measurable quantities, and the preferential membrane localization suggest a specific juxtacrine mechanism for TGF alpha in pituitary tumorigenesis.

Human pituitary null cell adenomas and oncocytomas in vitro: effects of adenohypophysiotropic hormones and gonadal steroids on hormone secretion and tumor cell morphology.

Human pituitary null cell adenomas and oncocytomas are not associated with evidence of excess hormone secretion in vivo; their cellular derivation has not been clarified by morphologic investigation. In this study we examined 41 null cell adenomas and 58 oncocytomas in vitro to determine hormone release and its response to several adenohypophysiotropic hormones and gonadal steroids. In vitro, 96/99 tumors released LH, FSH, and/or alpha-subunit of glycoprotein hormones. TSH was released by 11 tumors. GH, PRL, and ACTH were found in small quantities in 11, 8, and 5 tumors, respectively. Only 3 tumors released no detectable hormones. Incubations with test substances were examined at 2- and 24-h periods for up to 72 h. All but 3 of 53 tumors showed marked and persistent increases in the release of LH, FSH, and/or alpha-subunit in response to GnRH in short and long duration experiments. Secretion of LH, FSH, or alpha-subunit was stimulated to more than 150% of control by TRH in 37/48 tumors, by CRH in 10/20, by GRH in 7/20. Estradiol, progesterone, and testosterone increased release of FSH, LH, and/or alpha-subunit in 23/32, 3/12, and 3/12 tumors, respectively, and reduced their release in 6/32, 5/12, and 7/12, respectively. In tumors which showed no response to gonadal steroids, GnRH in combination with estradiol, progesterone, or testosterone yielded the same result as GnRH alone; in tumors inhibited by gonadal steroids, GnRH in combination with one of those substances reduced the response to GnRH. No secretion of GH, PRL, ACTH, or TSH was detected after incubation with GRH, estradiol, CRH, or TRH except in the tumors which initially released GH, PRL, or TSH. Ultrastructural examination of cultured cells from 15 cases revealed morphologic alterations that correlated with changes in hormone release and could be quantified by morphometry. This study represents the largest analysis of hormone production and release in vitro and morphologic correlation of clinically nonfunctioning pituitary adenomas. The responsiveness of gonadotropin secretion by null cell adenomas and oncocytomas to GnRH and gonadal steroids resembles that of gonadotroph adenomas. However, the unexpected increases in gonadotropin release attributable to several other adenohypophysiotropic hormones and the release of multiple hormones suggests that null cell adenomas and oncocytomas may represent neoplasms derived from uncommitted or committed precursor cells that can undergo differentiation towards several cell lines.